£?
X
Jks ^J
jr-'
s
<3&Ti
Carnegie
Institution
OF WASHINGTON
Year Book 72
1972-1973
Library of Congress Catalog Card Number 3-16716
The Maple Press Company, York, Pennsylvania
Issued December 1973
Contents
page
Officers and Staff v
Report of the President 1
Reports of Departments and Special Studies 1
Department of Embryology 3
Hale Observatories 97
Department of Terrestrial Magnetism 165
Department of Plant Biology 319
Geophysical Laboratory 415
Bibliography 735
Administrative Reports 737
Report of the Executive Committee 739
Abstract of Minutes of the Seventy-Fifth Meeting of the
Board of Trustees 741
Report of Auditors 743
Articles of Incorporation 757
By-Laws of the Institution 761
Index 767
in
Digitized by the Internet Archive
in 2012 with funding from
LYRASIS Members and Sloan Foundation
http://www.archive.org/details/yearbookcarne72197273carn
President and Trustees
PRESIDENT
Philip H. Abelson
BOARD OF TRUSTEES
Garrison Norton
Chairman
William McChesney Martin, Jr.
Vice-Chairman
William T. Golden
Secretary
Rt. Hon. Lord Eric Ashby
Lewis M. Branscomb
Michael Ference, Jr.
Carl J. Gilbert
Crawford H. Greenewalt
Caryl P. Haskins
William R. Hewlett
Keith S. McHugh
Henry S. Morgan
William I. Myers
Walter H. Page
Robert M. Pennoyer
Richard S. Perkins
William M. Roth
William W. Rubey
Frank Stanton
Charles P. Taft
Charles H. Townes
Juan T. Trippe
James N. White
Alfred L. Loomis
Trustee Emeritus
Former Presidents and Trustees
PRESIDENTS
Daniel Coit Gilman, 1902-1904 Robert Simpson Woodward, 1904-1920
John Campbell Merriam, 1921-1938; President Emeritus 1939-1945
Vannevar Bush, 1939-1955 Caryl P. Haskins, 1956-1971
TRUSTEES
Alexander Agassiz
George J. Baldwin
Thomas Barbour
James F. Bell
John S. Billings
Robert Woods Bliss
Amory H. Bradford
Lindsay Bradford
Omar N. Bradley
Robert S. Brookings
Vannevar Bush
John L. Cadwalader
William W. Campbell
John J. Carty
Whitefoord R. Cole
Frederic A. Delano
Cleveland H. Dodge
William E. Dodge
Charles P. Fenner
Homer L. Ferguson
Simon Flexner
W. Cameron Forbes
James Forrestal
William N. Frew
Lyman J. Gage
Walter S. Gifford
Cass Gilbert
Frederick H. Gillett
Daniel C. Gilman
John Hay
Barklie McKee Henry
Myron T. Herrick
Abram S. Hewitt
Henry L. Higginson
Ethan A. Hitchcock
Henry Hitchcock
Herbert Hoover
William Wirt Howe
Charles L. Hutchinson
Walter A. Jessup
Frank B. Jewett
Samuel P. Langley
Ernest 0. Lawrence
Charles A. Lindbergh
William Lindsay
Henry Cabot Lodge
1904-05
1925-27
1934-46
1935-61
1902-13
1936-62
1959-72
1940-58
1948-69
1910-29
1958-71
1903-14
1929-38
1916-32
1925-34
1927-49
1903-23
1902-03
1914-24
1927-52
1910-14
1920-55
1948-49
1902-15
1902-12
1931-66
1924-34
1924-35
1902-08
1902-05
1949-66
1915-29
1902-03
1902-19
1902-09
1902
1920-49
1903-09
1902-04
1938-44
1933-49
1904-06
1944-58
1934-39
1902-09
1914-24
Robert A. Lovett 1948-71
Seth Low 1902-16
Wayne MacVeagh 1902-07
Andrew W. Mellon 1924-37
John Campbell Merriam 1921-38
Margaret Carnegie Miller 1955-67
Roswell Miller 1933-55
Darius O. Mills 1902-09
S. Weir Mitchell' 1902-14
Andrew J. Montague 1907-35
William W. Morrow 1902-29
Seeley G. Mudd 1940-68
William Church Osborn 1927-34
James Parmelee 1917-31
Wm. Barclay Parsons 1907-32
Stewart Paton 1916-42
George W. Pepper 1914-19
John J. Pershing 1930-43
Henning W. Prentis, Jr. 1942-59
Henry S. Pritchett 1906-36
Gordon S. Rentschler 1946-48
David Rockefeller 1952-56
Elihu Root 1902-37
Elihu Root, Jr. 1937-67
Julius Rosenwald 1929-31
Martin A. Ryerson 1908-28
Henry R. Shepley 1937-62
Theobald Smith 1914-34
John C. Spooner 1902-07
William Benson Storey 1924-39
Richard P. Strong 1934-48
William H. Taft 1906-15
William S. Thayer 1929-32
James W. Wadsworth 1932-52
Charles D. Walcott 1902-27
Frederic C. Walcott 1931-48
Henry P. Walcott 1910-24
Lewis H. Weed 1935-52
William H. Welch 1906-34
Andrew D. White 1902-16
Edward D. White 1902-03
Henry White 1913-27
George W. Wickersham 1909-36
Robert E. Wilson 1953-64
Robert S. Woodward 1905-24
Carroll D. Wright 1902-08
Under the original charter, from the date of organization until April 28, 1904, the following
were ex officio members of the Board of Trustees: the President of the United States, the Presi-
dent of the Senate, the Speaker of the House of Representatives, the Secretary of the
Smithsonian Institution, and the President of the National Academy of Sciences.
VI
OFFICE OF ADMINISTRATION
1530 P Street, N.W., Washington, D.C. 20005
Philip H. Abelson President
Edward A. Ackerman 1 Executive Officer
James W. Boise Bursar; Secretary -Treasurer, Retirement Trust;
Executive Secretary to the Finance Committee
Montgomery S. Bradley 2 Administrative Officer
Marjorie H. Walburn Assistant to the President
Sheila A. McGough Editor
Kenneth R. Henard Assistant Bursar; Assistant Treasurer,
Retirement Trust
Joseph M. S. Haraburda Assistant to the Bursar
Marshall Hornblower Counsel
STAFF MEMBERS IN SPECIAL SUBJECT AREAS
Alfred D. Hershey
Barbara McClintock
Tatiana Proskouriakoff
RESEARCH ASSOCIATE AT LARGE
Harry E. D. Pollock
1 Died March 8, 1973.
2 Resigned June 30, 1973.
vu
Report of
the President
The year 1973 has been one of those periods in which it is necessary for a
nation to reevaluate its outlook and policies. For generations we have acted
as if we enjoyed inexhaustible natural resources. Suddenly we have run headlong
into crippling limitations. The energy crisis is the more painful because it comes
relatively soon after the successful Apollo Program. A few years ago people every-
where said "If we can go to the moon, we can do anything." Now suddenly we
are unable to provide enough fuel to keep warm. Such a primitive technology as
the mining of coal is temporarily beyond us.
We Americans have been smug and careless. We have assumed that our high
standard of living was a just reward for our inventiveness and organization of
mass production. While these components had a role, the real key to prosperity
in the United States was a superior combination of natural resources. We had
abundant supplies of oil, coal, and iron ore, and many other resources. It did not
require great ingenuity to exploit them.
Encountering the new scarcities comes as a shock to many people, but more
than a decade ago it was already obvious that we were headed for trouble. This
was particularly evident with respect to energy.
We were consuming energy, including hydrocarbons, at an ever increasing rate
but not discovering corresponding amounts of petroleum. While possessing huge
reserves of oil shale and coal we made only token efforts to develop potential for
utilizing these reserves to meet our future needs. With domestic production of
petroleum level or declining and consumption growing, we soon fell into depend-
ence on foreign sources.
The environmental movement sharply hastened the growth of dependence.
Fuel oil replaced coal, and automobiles became more inefficient. The use of
imported petroleum, which had been increasing slowly, spurted to an annual
growth rate of as much as 50 percent in September 1973.
The huge added imports of the United States created a tremendous seller's
market for the petroleum exporters. In the course of a year, the producers were
able to roughly double their return on a barrel of oil. A continuation of the growth
of imports into 1974 would have been financially unsustainable. It would have
entailed a cost of $15 billion or more.
The growing dependence of the United States placed a strong weapon in the
hands of the Arabs, and they used it. In the process they gained in many ways.
They conserved a nonrenewable natural resource. By creating scarcity they drove
the price up further. The immediate diplomatic consequences of their act were
obvious — the long term results are not so clear. As for the United States, we will
have a winter and likely much longer to reflect on the necessity of more prudence
in preparing for future needs.
During the past several decades in which we enjoyed a higher standard of
living than other countries, our economy was largely insulated from the rest of the
world. We could build a prosperous economy largely on items derived from
within the United States. In addition, we enjoyed a superiority in technology that
enabled us to compete in world trade and to pay for imports of raw materials.
Our technical superiority has faded at a time when we have become more
dependent on imports of raw materials, particularly of petroleum and petroleum
products, but also such commodities as iron ore, bauxite, chromium, nickel, silver,
and platinum. At the same time, other countries throughout the world are also
seeking to increase their consumption of materials. The buyer's market charac-
3
4 CARNEGIE INSTITUTION
teristic of petroleum is likely to spread to many other items. If we base our
economy too heavily on imports, we shall have no end of troubles — financial and
political.
How can we pay for the oil and other materials? Over the long haul we must
depend on increased exports. But this need collides with the fact that Western
Europe and especially Japan can now produce most goods more cheaply than we
can. Such strength as we have lies in a few items, such as computers, jet aircraft,
chemicals, and agricultural products. In attempting to pay for imports, we face
the need to export huge quantities of agricultural products, leading to an increase
in the cost of food for all our citizens.
Since our economy has become vulnerable to the actions of others, it seems
worthwhile to explore a few trends in other countries.
During the past 15 years there have been profound changes all over the world.
For better or worse, other peoples are trying to adopt what they consider the
best elements of our life style. Fifteen years ago there were few automobiles in
the countries of Europe and Latin America and in Japan. Now many of these
countries manufacture and export automobiles. Today their major cities have
traffic jams rivaling our own.
Traffic jams are an annoying and scarcely desirable manifestation. They are,
however, an indication of something more fundamental — the consumption of
energy generally in an industrial society. Neither we nor the peoples of Western
Europe or Japan can maintain a high standard of living without the use of
substantial quantities of energy. All these peoples are heavily dependent on
oil as a principal source of energy. Moreover, they are much less well endowed
than we in terms of fossil fuels, including oil. Thus our effort to import oil
is a threat to the well-being of many other peoples. They need the oil even more
than we do, and they are likely to strive harder than we will to get it.
The advanced countries are not the only ones that have aspirations to utilize
more energy and to enjoy amenities that technology can bring. The urge to
share some of our way of life is widespread.
Recently I visited Mexico, which is a curious mixture of the completely modern
and the primitive. Many Indians in the mountains continue to live as their
forebears did a thousand years ago. A Mexican friend told me of an incident
which seems significant. The Mexican government decided to try to improve the
standard of living of some of the tribes. The government thought that the
building of a highway would bring many economic benefits — better markets for
Indian handicraft products and cheaper costs of goods the Indians then might
buy. In building the highway it was decided to avoid use of heavy machinery
and instead to employ Indian labor. Thus for the first time in their lives a group
of Indians received a substantial pay-off after the first week's work. Naturally
on Saturday night the Indians went to town. My informant did not reveal all
that happened in town, but he did tell me that when the Indians returned to the
labor camp each brought with him three things : a new sombrero which is essential
to protect against the tropical sun, a string tie which catered to a man's need for
finery, and a transistor radio. It is difficult for us to guess what went on in the
minds of those Indians as they listened to their transistor radios. But one thing
seems certain. An irreversible change occurred. They will never return com-
pletely to their pre-Columbian life. Their appetite for our gadgets has been
whetted, and the advertising and other material that they will receive on their
radios will stir additional desires. What I have just described is only one example
REPORT OF THE PRESIDENT 5
of what has been happening in many lands. The rest of the world wants to enjoy
our standard of living.
As a world leader of the great scientific technological revolution, we have set
patterns that have led to the weakening of centuries-old social systems. Our
example has stirred longings and raised hopes and expectations that probably
cannot be fulfilled.
There are not on this planet sufficient resources to permit everyone to live in
the style in which we have been living. Among other things, it would not be
possible for the world to maintain for long a civilization based on oil. Were all
the world to consume petroleum and its products at our rate, the fabulous re-
sources of the Near East would vanish in about two decades. If the world is to
enjoy for long a comfortable standard of living, substantial adjustment in the
procurement and uses of energy must occur. These adjustments will be slow,
as our own experience shows.
In the coming years we are going to witness a global scramble for oil, other
raw materials, and food. We will witness frustration, starvation, and disorder.
Since we are blessed with great natural resources, a highly productive agriculture,
and a highly competent scientific and technological establishment, we have the
potential of being able to avoid much of the trauma that others will face. Our
most vulnerable feature is our dependence on imports of petroleum.
At present we obtain 78 percent of our energy consumption from hydrocarbons
— natural gas and petroleum. Liquid hydrocarbons are especially important, for
they furnish the energy for agriculture and transportation. We could not exist to-
day without a highly mechanized agriculture and our transportation system — air-
planes, trains, trucks, and automobiles. Without ample supplies of petroleum,
our economy would crumble and we would face the most drastic changes in our
way of life. We need to move to utilize more effectively our reserves of oil shale
and coal. We need to conserve energy.
Liquid hydrocarbons are so useful that we should explore vigorously various
alternatives for obtaining them besides importation. In principle, enormous
quantities could be obtained from coal and from oil shale. Use of either of these
sources is certain to raise some environmental problems. Our best way to identify
and to minimize those problems is by actual experience. Moreover, if we were
to design and build a few large plants quickly, we would make some expensive
engineering mistakes, but we would learn fast how to improve our performance.
The cost of building major plants would be substantial. We should be thinking
in the range of $1000 million for each of several plants. The government
must be prepared to share part of the costs. One method might be for the
government to invite bids to supply over a ten or twenty year period a fixed but
large quantity of hydrocarbons taken from domestic coal or shale. The genius
of the research and engineering staffs of the large chemical and petroleum com-
panies would be turned loose on the problem, and great progress would soon
result.
Our economy is also vulnerable in its supply of raw materials. We need to make
a determined effort to lessen dependence on imported materials by designing
components that utilize locally available substances, and by developing our own
resources.
Given good organization, discipline, and thoughtful use of the tools that
science and technology furnish, the world could probably move toward an en-
O CARNEGIE INSTITUTION
during economy — one based on abundant and renewable sources of energy,
materials, and food. However, in much of the world there is lack of organization
and discipline as exemplified by the continuing population explosion in many
lands. Perhaps the major contribution we can make is one of example. If we
set our house in order and behave with a modicum of discipline, we can enjoy a
prosperous life while leaving our descendants a proper heritage. In future we are
going to continually bump against limitations, the need to weigh trade-offs, for
example, between environmental considerations and energy needs. The public
and the politicians must come to realize the strengths and the limitations of
science and technology when responding to public needs. Given steady support
and encouragement, science can give humanity additional power and tools to
meet needs. But everyone should be aware of the long time span between dis-
covery and application — in the range of ten years. Moreover, even when the
technology of an application is established, considerable time must elapse before
large-scale production can be achieved.
At the moment our greatest weakness lies in lack of ability to mesh our political
institutions with technological facts of life. This weakness has been apparent for
a long time under many administrations. We have shrunk from realistic plan-
ning, perhaps out of ideological considerations. Of course, rigidity in planning
may well be a greater handicap than no plan at all. But we have carried our
reluctance in this matter to extremes and have seemed unable to prepare for the
future. Our political system is poorly equipped to deal with long term undramatic
problems.
The behavior of politicians necessarily reflects the demands of the public. In
large measure we have become a society sharply focused on the here and now. A
generation ago it was necessary for the individual to give thought to the future.
Individuals had to consider not only their long-term future but also how they
would survive the coming winter. But social security and other developments
have removed the necessity for thought and self-discipline. The future will be
taken care of by others. Public attention is on the here and now, and politicians
act accordingly.
As a resident of Washington for more than 30 years and an observer of the
actions of politicians and bureau heads, my experience can be boiled down to an
acid comment which though an exaggeration is not very wide of the mark. The
key officials follow closely press, radio, and TV. They watch "Meet the Press,"
"Face the Nation," and other such shows on Sunday, digest the weekend news-
papers, glance at Monday morning headlines, and then make national policy for
the week. Unfortunately, many of these short term decisions have long term
consequences; for example the Apollo Program. In that case, a momentary
political necessity led to a decade-long commitment and a program that continues.
All of us find it difficult to avoid immersion and entrapment in the excitement
of the moment. But there are a few elements in society whose careers depend on
being future oriented; for example, some engineers, utilities executives, and other
corporate planners. Another group is the scientists. They are continually seeking
new discoveries which will be significant not only to this but to future generations.
Ultimately, our way of life must be transformed from one dependent on non-
renewable resources to one with a more enduring basis. The foundation knowledge
for such a transformation will come from science.
While we are working to solve immediate energy problems and those of the
next decade, we should also be moving toward longer range goals. We should
REPORT OF THE PRESIDENT 7
seek ultimately to derive our energy and materials from renewable resources. We
should arrange to obtain a substantial fraction of our energy from the sun
because that energy is enduring and because the use of solar energy would cause
less damage to the environment than the use of fossil fuels and less danger than
nuclear reactors.
In time the federal government will be geared to overcoming our energy prob-
lems. Past experience, though, is that the government will emphasize a few
aspects of the many long range problems we face while largely neglecting most.
It is important that some future-oriented people be given the freedom, and the.
funding, to identify promising opportunities not in the immediate spotlight.
When viewed in the context of federal research and development expenditures,
the annual budget of the Carnegie Institution is minuscule. Nevertheless, in the
scheme of things we can have a significant function. We are accustomed to
taking a long term view, to looking for leverage and areas that the fashion of
the day is neglecting. The record of past achievements and contributions arising
from the Institution's efforts supports our view that society makes no better invest-
ment than the kind of education and basic research being carried out in the
Institution. Our training and research activities relate to aspects of the great
human needs — energy, materials, and food. Our activities also relate to two other
needs: health and, for many, the greatest need of all, the need to know.
At our Department of Plant Biology we are expanding studies of man's
greatest source of renewable materials. We are studying the process of photo-
synthesis with new and more effective tools. Using a highly instrumented mobile
laboratory, we are investigating the ability of plants to thrive under extreme
conditions, including those in Death Valley. These studies could well lead to
better management of our fields and forests and perhaps to more productive
plants.
Our studies of geochemistry and geophysics relate to such natural hazards as
earthquakes and volcanism, to geothermal energy sources, and to the better
understanding of the evolution of the crust of the earth. These studies also
relate to economic geology and to the materials sciences. The history of the
Institution includes many examples of work that has had practical application in
the steel, cement, ceramics, and glass industries.
Our research in biology is directed to understanding fundamental processes in
growth and development, and this effort too has a proud history of contributing
to the solution of important practical problems while advancing basic knowledge.
In the pages that follow, Ebert provides a summary of some of the contributions
of the Department of Embryology, including a study this year of myasthenia
gravis.
We cannot foresee practical applications from astronomy. But it is apparent
that humans have an urge to know that affects them deeply. Each year hundreds
of thousands of visitors come to the Observatories. In addition, many amateur
astronomers grind their own lenses and conduct observations of the skies. These
people are touched by the mystery and grandeur of the universe. Who can look
out at the stars on a clear cool night and not be moved?
In the pages that follow, some highlights of the Institution's work are presented.
The reader can note many studies that relate to long term societal problems,
such as food, materials, and health. He can also see interwoven our concerns
with education, growth of fellows and staff, and everywhere, an appeal to the
human need to know.
The Year in Review
THE HALE OBSERVATORIES
Astronomy is in the midst of continuing evolution. Once astronomers were
principally concerned with optical measurements of visible light. Now the
heavens are being explored using virtually the entire electromagnetic spectrum
from gamma rays to multicentimeter waves. This has uncovered new phenomena
such as quasars, pulsars, and x-ray emitting stars. The vocabulary of astronomers
has been enriched with words like neutron stars, black holes, and synchrotron
radiation.
Much of the new excitement has originated outside the realm of optical
astronomy, but, as reported by Horace W. Babcock, Director, Staff Members
at the Hale Observatories have found stimulus and challenge from the new
discoveries.
Quasars were first identified by a combination of optical and radio telescopes.
In those days the resolution of radio telescopes was poor. Thomas Matthews
and Allan Sandage were able to pinpoint the location of a quasar, using the
California Institute of Technology facility at Owens Valley and the optical
facilities of the Hale Observatories. The resolution of radio telescopes has since
been greatly improved, using long base lines. Today their resolution is comparable
to that of the large telescopes. We are now entering an era in which there will be
even more intense interaction between optical and radio astronomy. Similar
comments are likewise applicable to other parts of the electromagnetic spectrum.
The x-ray satellite Uhuru has detected more than 200 x-ray emitting objects.
Optical examination of some of these has led to new puzzles.
The Hale Observatories conduct studies of the sun and other objects in our
solar system. But most attention is devoted to other stars and to galaxies,
their structure, evolution, and relationship to quasars. Star formation, stellar
evolution, radio emission, and high energy physical processes are subjects of
investigation.
The controversy concerning origin of the universe continues. Those who ad-
vocate the big bang hypothesis seem to be able to marshal the most evidence.
Nevertheless, some believers in a steady state cosmology continue unconvinced.
This matter is fundamental, and further evidence is important.
8
REPORT OF THE PRESIDENT 9
A contribution by Maarten Schmidt provides additional weight against a
steady state cosmology.
Schmidt has developed a theoretical analysis of the abundance and flux densities
of radio sources to be expected from steady state cosmology. He has applied
this analysis to observational data on strong radio sources provided by the
Third Cambridge Catalogue. Optical identification of many of these sources has
been made and their redshifts determined. The actual number of remote radio
sources observed is considerably greater than can be accounted for by steady
state cosmology.
Sandage's investigations of N galaxies show that a typical object of this type
is a giant elliptical galaxy with a miniquasar in its nucleus, and, furthermore,
that the magnitude-redshift relationship for the N galaxies leaves little, if any,
margin for a possible noncosmological component of the redshift.
Jerome Kristian has extended the search for galaxies underlying quasars by
identifying the conditions under which such galaxies should be detectable. When
the galaxies are relatively distant, one would not expect to resolve the light from
the quasar as against that from the remainder of the galaxy. In contrast, it should
be possible to differentiate between the quasar and its accompanying galaxy
when they are at an intermediate distance. Critical examination of 200-inch
photographs of 26 quasars gave results in good agreement with the predictions.
The evidence supports the increasingly general view that quasars occur in the
nuclei of galaxies.
Increasing attention is being given to the history of star generation in galaxies
and to the processes that trigger large-scale bursts of star formation. Studying
ring galaxies, for example, Leonard Searle and Wallace Sargent have obtained
from multichannel scans evidence that the energy distributions of the rings are
closely similar to those of star clusters a few hundred million years old. The
same age is indicated for these galaxies from the measured expansions of the
rings. Searle and Sargent believe that these galaxies are the first truly young
extragalactic systems to be recognized. They also conclude that the outer reaches
of certain paired galaxies, distorted by mutual interaction, consist of predom-
inantly young stars.
The concept that density waves or compressional effects in galaxies accelerate
the formation of stars from the interstellar material is finding increased support
not only in situations such as the foregoing, but in the spiral structure and
nuclear regions of galaxies in general. Pieter van der Kruit, examining the high-
resolution radio brightness distributions within galaxies, obtained at Westerbork,
The Netherlands, found that the radio brightness distributions were displaced to-
ward the inner edges of the optical spiral arms, thus confirming the concept of com-
pressed regions in the density wave picture of spiral structure. The radio spiral
structure is believed to result from compression of the magnetic field and rela-
tivists electrons in the large-scale "base disk" of a galaxy. Star formation is
triggered in the density wave compression regions, in a manner governed by the
degrees of compression. Van der Kruit concluded that the relationship of specific
angular momentum to mass is the governing parameter. His conclusions were
reinforced by observations at Palomar of the rotation curve for NGC 4321.
Numerous attempts have been made to identify optical counterparts of recently
discovered galactic x-ray sources. One such object is the binary BD + 34°3815,
which is a weak variable radio source and is identified with Cygnus X-l. The
star is a 5.6-day spectroscopic binary of which the primary optical component
10 CARNEGIE INSTITUTION
is of spectral type BO lb, with a velocity amplitude of 74 km s"1. Robert Brucato
and Kristian, observing the star spectroscopically at Mount Wilson, discovered
that the weak emission feature at A4686 (He II) had a velocity amplitude nearly
twice as great as that of the B star and that it varied in antiphase; no other
evidence of a secondary spectrum is seen. After analyzing the evidence, Brucato
and Kristian concluded that one possibility is that the secondary component
of the binary is a black hole, the x rays being produced by the thermalization
of gravitational energy of matter falling in the vicinity of the black hole.
The pulsating x-ray source Her X-l associated with the variable star HZ
Herculis is being studied by J. Beverley Oke and Jesse Greenstein with the
multichannel spectrophotometer and the coude image-tube spectrograph. This
fascinating star shows an optical variation which requires that the side facing
the x-ray pulsar be quite hot, and the opposite side near 8000°. Resolution into
two, or possibly three, different major components on the surface of this optical
object is required by the multichannel observations.
Spectroscopy by Greenstein with the coude image-tube at the 200-inch telescope
has continued to show that the hydrogen-line white dwarfs have quite small
rotational velocities. At present, five of the six stars successfully observed show
cores, formed in nonlocal thermodynamic equilibrium, that are barely wider than
the instrumental resolution. If these stars are viewed as the central regions of
former red giants, an efficient mechanism must operate to reduce the angular
momentum. This occurs in the interiors of whatever stars undergo slow evolution
into the white dwarf region.
A large number of the infrared sources that have been detected in the past
several years have been interpreted in terms of the so-called dust shell model.
In this model energy from a central object at relatively short wavelengths is
absorbed by particles in a surrounding envelope and reradiated at the much
lower temperature characteristic of the grains. In this way the maximum of the
energy distribution can be shifted far into the infrared, perhaps totally obscuring
the original energy distribution of the central star. Unfortunately, direct obser-
vational evidence for these shells other than the observed energy distributions
themselves is largely lacking. For objects sufficiently near the ecliptic, however,
lunar occultations provide a useful method of measuring angular extent of infrared
sources and may provide direct evidence for a shell.
IRC + 1011 (= CIT 3) is a bright infrared source that also emits microwave
lines of the OH molecule. The object has recently undergone a series of lunar
occultations. Three of these events could be observed at Palomar and Mount
Wilson and attempts were made by Robert Zappala, Eric Becklin, and Gerry
Neugebauer to observe these simultaneously in two wavelengths with the 200-inch
and 101-inch telescopes. Successful observations were obtained at 2.2 and 10.1 n
on one occasion, and at 20 n on another.
These observations provide clear evidence for a shell of material around a
smaller central object. If the total luminosity of the object is about 104 LQ, the
distance to the star is about 500 pc and the observed extent of the outer dust shell
corresponds to a diameter of about 60 A.U. The temperature of the dust shell
is about 600 °K.
In discussions of scientific accomplishments, emphasis is commonly placed
on concepts, ideas, and results. Somehow, the means of achieving new insights
gets short shrift. This is unfortunate, for practically every field of science is
heavily dependent on new instrumentation or techniques as a means of creating
REPORT OF THE PRESIDENT 11
new opportunities for examining nature. This is especially true in astronomy.
Over the years, one of the strengths of the Hale Observatories has been their
facilities and instrumentation. That tradition is being maintained. Construction
of the 101-inch Irenee du Pont Telescope at Las Campanas is proceeding and will
result in a splendid observatory in the southern hemisphere.
Under the general supervision of Bruce Rule as Project Officer and with the
collaboration of many members of the scientific, engineering, and administrative
staffs, substantial progress was made in meeting scheduled design requirements,
guidance of manufacturing operations, procurement, and delivery of site con-
struction components for the du Pont Telescope and dome building. Most
mechanical components are in advanced stages of inspection and completion in
preparation for shipment to Chile, with assembly planned at the Las Campanas
site in 1974, after the dome and building are ready.
Following tests at the Optical Sciences Center of the University of Arizona
of the primary mirror support cell and grinding table supports with the mirror
at a spherical figure, the operation of polishing to achieve the final aspheric figure
was started in early summer 1972 and continued throughout the year. The work
is under the supervision of Don Loomis, Chief Optician for the telescope optics,
with close cooperation in testing by Arthur Vaughan and Eugene Fair. Design
and fabrication of optical test equipment and development of computational
procedures were pursued at the O.S.C. and by Vaughan in Pasadena. By April
1973 all test procedures required for evaluation of the final optical figure of the
primary were in operation. The mirror figure was brought to within 0.5 micron
of the final surface by June 1973.
Following a number of general meetings to discuss auxiliary instrumentation,
several working groups were set up by Oke to supervise the design and construc-
tion of specific instruments and facilities. It is the hope that most of the initial
set of instruments will be ready for use when the telescope is completed. The
groups (with the chairman) are listed below: Direct Plate Holder and Darkroom
Processing Equipment (Sandage) ; Instrument Mounting Base (Oke) ; TV Guid-
ing System (Edwin Dennison) ; Image-Tube Spectrograph (Searle) ; Broad-band
Photometer (Kristian) ; Infrared Photometer (Neugebauer) ; 90-mm Image-Tube
(Brucato) ; Coude Spectrograph (Vaughan) ; Cassegrain Chair (Rule) ; Computer
Interfaces (Neugebauer) .
Substantial progress has been made by the Astroelectronics Laboratory under
Dennison toward the design and completion of the control system for the du
Pont Telescope. A large fraction of the effort has gone into ensuring that the
design is satisfactory in two fundamental interface areas: between the control
system and the mechanical parts of the telescope and between the control system
and the operator. This initial study phase was essential to minimize the possibility
of human error in operating the telescope and to eliminate the necessity for
on-site modification of the system when it is installed with the telescope in Chile.
The Hale Observatories are fortunate in benefiting from multiple initiatives in
development and use of electronic instrumentation. Caltech has been the source
of a number of new approaches. The SIT-Vidicon photometer developed by
James Westphal is working out very well, and it provides increased opportunities
for study of far distant or dim objects (magnitude 23).
Paul Harvey, Michael Werner, Jay Elias, and Ian Gatley of Caltech have
constructed equipment for use with the 200-inch telescope that has permitted
observations at 1 mm wavelength.
12 CARNEGIE INSTITUTION
Staff of the Hale Observatories have also been active. Oke and Dennison have
been carrying on a joint effort with Princeton University to develop, build, and
test a digital image recorder. The instrument has been tested successfully at
Palomar.
Vaughan has been developing equipment that has enhanced capability of
detecting and measuring magnetic fields of stars.
The Astroelectronics Laboratory under Dennison emphasizes quality and de-
pendability in equipment and brings to the Observatories engineering skill.
Of considerable significance to astronomy generally has been a new technique
in the use of photographic plates developed in part by William Miller of the
Photographic Laboratory of the Hale Observatories. Using new IHa-J plates
provided by Eastman Kodak, combined with a bake-out procedure conducted
in dry nitrogen, sensitivity to weak light sources has been increased by a factor
of 3.
THE GEOPHYSICAL LABORATORY
Students of the earth sciences have been enjoying a most stimulating period.
At a time of increased emphasis on relevance, they can take pride in the fact that
their science is central to the search for scarce materials. Simultaneously, many
aspects of their intense exploration of the unknown have broad public appeal.
As scholars they are fortunate, for they can contemplate many challenges. New
opportunities and frontiers have been opened by a combination of big science
projects, the concept of plate tectonics, and a host of advances in techniques and
instrumentation, as indicated by Hatten S. Yoder, Jr., Director, in the Geo-
physical Laboratory report.
The Apollo program has been completed, but the astronauts brought back
invaluable samples from the moon, providing some of the best opportunities
to probe the early history of the solar system.
The Joint Oceanographic Institutions for Deep Earth Sampling (JOIDES) has
made available many cores and new information and questions. The concept of
plate tectonics has illuminated many old puzzles while creating new ones. New
instrumentation, largely electronic, has vastly increased the effectiveness of data
gathering and data handling, and in some instances made accessible phenomena
hitherto unexplored. Research at the Geophysical Laboratory emphasizes funda-
mental understanding of the earth and solar system rather than applications.
However, the history of the Laboratory is rich with examples of practical applica-
tions arising out of basic work performed there.
An excellent example of the interplay of the concept of plate tectonics and
the opportunity created by instrumentation may be seen in Francis Boyd
and P. H. Nixon's study of kimberlites of Lesotho, South Africa. The kimberlites
are in themselves very interesting rocks. They are the major source of the
world's gem grade and industrial diamonds. A number of lines of evidence point
to an origin deep in the earth's mantle, and a sudden, violent propulsion to the
surface through a pipe extending from the deep mantle through the crust of the
earth. The rock, the diamonds, and especially the enclosed fragments of ultramafic
rock, usually rounded into nodules, are of great interest to petrologists because
they provide samples and record processes within the mantle. The depths from
which these materials come have been unknown, yet a remarkable variety of
REPORT OF THE PRESIDENT
13
rocks is incorporated and transported upward in the explosive ascent of kimber-
lite to the earth's surface. Boyd and Nixon have been attempting to establish
the depths from which the nodules come by examining the properties of the
coexisting minerals in these rocks. The compositions of the minerals reflect the
temperature and pressure that they last experienced before being torn loose
and carried up in the explosion. By carefully measuring these compositions and
relating them to those determined on similar minerals in the laboratory at
measured temperatures and pressures, the record can be read. Boyd and Nixon
argue that the nodules reflect a series of temperatures and depths that essentially
define an ancient geothermal gradient down to about 250 km (Fig. 1). The
geotherm was disturbed, probably by shearing that resulted from the plate move-
ments accompanying the break-up and dispersal of Gondwanaland. The nodules
that plot below the inflection are sheared (see Fig. 1), and those represented by
points above are granular. They suggest that the point of inflection locates the
top of the low-velocity seismic zone in Cretaceous times (100 million years ago)
and also marks the top of a crystal-mush zone from which the kimberlite magma
originated. This reconstruction is the first experimental attempt to define a geo-
therm at a particular moment in the evolution of the mantle.
600
400-
o* 1200
Q.
E
CD
1000-
800-
600
SHIELD GEOTHERM
PYROXENE GEOTHERM
0
200
250
Depth, kilometers
Fig. 1. An experimental estimate of the Cretaceous (100 m.y. ago) geotherm compared with
the present calculated shield geotherm (dashed) in northern Lesotho, South Africa. Based on
temperatures and depths of equilibration of pyroxenes from lherzolite nodules from kimberlite
pipes calibrated with experimental data on synthetic systems.
14 CARNEGIE INSTITUTION
The new conclusions rest on a solid foundation — accumulated knowledge of
silicate systems that provided the necessary geobarometers and geothermometers
and a highly effective electron microprobe automated by L. W. Finger that
facilitates collection of accurate data on many samples of kimberlite.
Another example of new knowledge gained through powerful equipment and
excellent technique is provided by Ho-Kwang Mao's studies of the behavior of
minerals at very high pressures. He has examined materials thought to be
especially abundant deep in the earth's mantle. The experiments are conducted
in a diamond-faced pressure cell. Optical examination and determination of
electrical conductivity of materials under pressures as high as 300 kbars (« 1000
km depth) at temperatures up to 800° C can be carried out in this equipment.
The measurement of properties of crystals such as compressibility, thermal
expansion, density, electrical resistivity, and index of refraction yields clues
to the nature of the mantle of the earth. For example, the compressibility, thermal
expansion, and density give a measure of the geothermal gradient, thermal con-
ductivity, and seismic velocity. The index of refraction can be directly related
to the dielectric constant, which together with electrical resistivity defines the
behavior of the earth in an electric field and reveals the origin of the earth's
magnetic field. Mao has measured the optical and electrical properties of crystals
believed to be important in the earth's mantle — olivine, spinel, and mag-
nesiowiistite. The crystals absorb an increasing amount of light in the visible
and near infrared wavelength region and thus darken as the pressure is raised
to 150 kbar. The frequency of the light absorbed is directly related to the
radiation transfer and indirectly to the electrical conductivity. His observation
thus suggests that no significant heat transfer in the mantle is likely to be by a
radiative process, as previously thought. In other experiments Mao made direct
measurement of the electrical conductivity, which was sufficiently high to suggest
that heat transfer in these substances may be mainly the result of electronic
thermal conductivity. The new data coupled with other parameters are essential
in constructing models of temperature distribution and of the thermal history
of the earth. Simple physical measurements on crystals are indeed of great value
in ascertaining the properties of the earth.
The nuclear age has produced a wealth of techniques for studying rocks and
minerals. One technique, particle-track mapping, has opened a new approach
to determining partition coefficients of trace elements between crystals and be-
tween crystals and surrounding melt. Martin Seitz has developed a method
whereby uranium, thorium, and boron can be measured in amounts of 0.1 ppb,
1.0 ppm, and 1.0 ppm, respectively. He found, for example, that the activity
coefficients of uranium in the crystal and melt are independent of concentration,
and the results support the view that silicate melts are dilute with respect
to uranium at concentrations found in natural systems. In addition, the parti-
tioning between crystal and melt does not seem to depend appreciably on the
composition of the melt or the temperature and pressure of the system. Those
properties simplify the application of experimental data to studies of geologic
processes. From the very low concentrations of uranium and boron, it appears
that ultramafic rocks and basalts from the oceanic ridge may have formed in the
same fractionation event. Radioactive elements such as uranium and thorium
potentially date the time of their synthesis in stellar environments. Similar
partitioning studies of these elements in meteorites are essential for a better
understanding of the chronology of nucleosynthesis.
REPORT OF THE PRESIDENT 15
Radioactive isotopes are particularly useful not only for measuring the time of
formation of a rock but also for dating major changes in rocks as they take part
in various geological processes. The limitations are no longer imposed by ana-
lytical techniques but by one's ability to unravel the complex response of the
isotopic systems in the constitutents of the rock to the processes. Thomas Krogh
and Gordon Davis realized that much of the confusion in dating geological events
is due to the different response of the various minerals and to the different be-
havior of the elements of the dating systems used during metamorphism. Much
effort was put into comparing these responses in two rock layers about 1850 m.y.
old that underwent metamorphism about 1000 m.y. ago. Each layer was
initially formed with a different Sr-Rb ratio, and large differences developed
in the Sr87/Sr86 ratio during the long time between formation and metamorphism.
The limits of isotopic mixing were uniquely determined: Chemical equilibrium
had not been obtained in distances of even a few centimeters during the meta-
morphism. It was shown for the first -time that a single outcrop can yield two
vastly different whole-rock Rb-Sr isochron ages. The zircons displayed an added
complication. By analyzing the ends of 200 micron grains, it was possible to
prove that overgrowths of new zircon formed during the metamorphism. These
overgrowths are not always evident optically, and therefore the U-Pb dating
method may yield confusing results. Their analyses of these rock layers cast
considerable doubt on the attainment of isotopic equilibrium during regional
metamorphism.
The bulk of the organic matter that accumulates in a sediment or a soil
bears little resemblance to the materials in living organisms from which it is
formed. A significant fraction of this material, extracted by dilute alkaline solu-
tions and precipitated by acidifying the extract, is called humic acid. The chem-
istry of this organic matter has been difficult to study, and an adequately established
theory for its origin, chemistry, and geological fate does not exist. Over 60 years
ago Maillard noted the similarities between natural humic substances and the
products of the reaction between sugars and amino acids. He proposed that these
products, called melanoidins, were the precursors to humic acids. However, most
workers consider that lignin of woody plants is the source of humic acid. Because
the lignin hypothesis entails difficulties in explaining natural occurrences, Thomas
Hoering has followed up on some studies of the melanoidin model initiated
earlier by Abelson and Edgar Hare. Hoering has found a number of additional
chemical and physical properties of synthetic melanoidin that are comparable
to those of natural humic acid and has concluded that the Maillard hypothesis
forms a logical basis for the origin of humic acids. He has further shown that the
small amounts of aromatic and phenolic acids, derived from humic acids, which
are key pieces of evidence for the lignin hypothesis, may arise from natural
diagenetic processes and are not necessarily indicative of lignin. The melanoidin
model is a base for new insights into the structure and reactions of humic acids.
THE DEPARTMENT OF TERRESTRIAL MAGNETISM
The Department of Terrestrial Magnetism carries a name which ill describes
its activities. Yet if one insisted that the designation describe the functions, one
would be faced with the unhappy choice of frequent name changes to reflect the
evolution of programs or of freezing the program to preserve the preciseness of
the name.
16 CARNEGIE INSTITUTION
Over the years most of the Department's activities have been in geophysics,
with substantial contributions in biophysics, nuclear physics, and radio and optical
astronomy.
By its l are, geophysics implies international activities, and the work of the
Department since its inception has included a strong international component.
In this year's report of Ellis T. Bolton, Director, both geophysics and its inter-
national aspects are highly evident.
The Geophysical Laboratory and the Department of Terrestrial Magnetism
are located about a mile apart in Washington. There is considerable interaction
of the two Departments in research in the earth sciences, with the staffs sharing
major equipment, exchanging ideas, and often joining together in projects.
Much of the geophysical and geochemical research at Terrestrial Magnetism
has been stimulated by the concept of plate tectonics. The rim of the Pacific
basin is restlessly undergoing tectonic activity. Together with colleagues in Japan
and in South America, staff scientists are examining some of the features of plates
and plate motions.
Selwyn Sacks and Hiromu Okada of Hokkaido University, now a predoctoral
Fellow at the Department, have undertaken a comparison of the Q structure
beneath the South American continent and the Japanese Archipelago. Similar
spectral ratio techniques were used for Q determinations in the two regions.
When waves travel through a medium, they are attenuated, and the more attenu-
ated the lower the Q. Low frequency waves are attenuated less than those of high
frequency. A solid rock tends to have a high Q ; a partially melted viscous mate-
rial has a low Q. In South America, Qp in the somewhat seismically active wedge
bounded by the seismic plane, the trench, and volcanoes is 1000 to 2000. Beneath
Japan in the corresponding region, which is aseismic, Qp is 400 to 500; this value
is identical to that in the asthenosphere below the trench. In South America, the
high Qp values extend down to 350 km, but there do not seem to be any paths
from deep earthquakes (600 km) with Qp greater than 1000. In Japan, however,
the high Qv' values in the Benioff zone (^3000) persist down to the deep earth-
quakes. Above the deep earthquake zone in Japan and Fiji, very low Qp values
(100) were found, in contrast to a minimum value of 350 in South America. The
tectonic implications of these data are that lithospheric plates have varying thick-
nesses and that continental plates are several times thicker than oceanic plates.
In another study of events accompanying tectonic processes, Stanley Hart and
J. G. Schilling of the University of Rhode Island have examined basalts from
Iceland and the adjacent track of the mid-Atlantic ridge with respect to trace
element composition and Sr87/Sr86 isotope ratios. Tholeiites from a 500 km sec-
tion of the Reykjanes ridge and its extension on Iceland were analyzed for K, Rb,
Cs, Ba, Sr, and Sr87/Sr86 Samples most distant from Iceland show 87/86 ratios
similar to typical ridge basalt (0.7026 to 0.7028) but are more highly depleted in
K and Sr (K < 400 ppm, Sr < 80 ppm) and show less fractionated element ratios
(K/Rb < 900, K/Ba < 65). Concentrations of the trace elements increase pro-
gressively as Iceland is approached (factors of 2 to 5) ; 87/86 ratios increase
stepwise to 0.7030 about 150 km from Iceland. Tholeiites from Iceland show a
considerable range in concentration (K, 700 to 2000 ppm; Sr, 90 to 200 ppm)
which overlaps and is somewhat higher than the nearest ridge samples. Ratios
of 87/86 on Iceland show little spread (0.7030 to 0.7031). The isotopic data show
REPORT OF THE PRESIDENT 17
that the mantle sources under Iceland are chemically different from those under
the southern Reykjanes ridge. The concentration variations between the two
provinces are not well described by mixing models having only two melt types.
A possible model involves variable partial melting and mixing of two mantle
sources.
Nobumichi Shimizu and Hart have developed a sensitive chemical procedure
for the analysis of trace elements in a two-phase system in which one phase is
crystalline and the other, a quenched liquid, is a glass. They call their procedure
the "Differential Dissolution Technique," or "DDT." The procedure takes ad-
vantage of the fact that the glass portion is dissolved in dilute hydrofluoric acid
much more readily than the highly ordered crystalline portion. Thus, in a series
of model experiments, it was found that in a mixture of clinopyroxene (from a
spinel lherzolite from Salt Lake Crater, Hawaii) and glass (tholeiitic basalt glass
from the Juan de Fuca Ridge) the glass dissolved, leaving pure clinopyroxene as
a residue. The technique permits accurate determination of the solubility of very
small amounts of an element in the crystalline phase, for the crystalline phase
can be freed of contaminating glass.
Nuclear physics studies have been part of the program at the Department for
more than four decades. The venerable Van de Graaff accelerator has been up-
dated, and it has features that facilitate experiments such as those with polarized
beams. This year Louis Brown and colleagues have been examining an interesting
phenomenon known as electron promotion. When target atoms are bombarded by
heavy ions, the simple theory which assumes an interaction of two point charges
during Coulomb excitation must be modified to account for the presence of the
projectile ion within the K shell of the target atom. The incoming ion resides
within the target atom for long periods of time compared with the oscillation
times of the K electrons. If the electron shells of ion and target atom overlap,
then a "hyperatom" consisting of the combined nuclei and a new K shell is formed.
During the time in which the two nuclei are close, two of the four electrons which
made up the two originally independent K shells are forced out, or "promoted."
When the two nuclei later separate, there is a high probability that each will have
a K shell vacancy, or even vacancies in the L shell. In turn, there is a high
probability of ionization and a consequent high cross section for x-ray production.
In cooperation with H. A. Van Rinsvelt of the University of Florida, studies were
undertaken to bombard copper with beams of alkali ions which are readily pro-
duced in the DTM Van de Graaff accelerator and offer a wide range of atomic
numbers. It has been found that the characteristic x-ray lines of copper are
shifted in energy as a result of the high probability of ionizing the L shell when
the K shell is ionized.
Activities in Optical Astronomy at the Department of Terrestrial Magnetism
were initiated in about 1960. The impetus came from Merle Tuve and his initia-
tive in pushing development of the Carnegie Image Tube. To implement this
effort, Kent Ford was brought to the Department. He was very helpful in the
successful development of a series of tubes. As part of the development program,
testing of the devices was carried out at Lowell Observatory and elsewhere. Ulti-
mately Ford was joined by Vera Rubin and the two have maintained a modest
but effective observational program. One of their recent discoveries is the aniso-
tropic distribution on the celestial sphere of a class of galaxies known as Sc I —
a special class of spiral galaxies of high luminosity showing photographic qualities
18 CARNEGIE INSTITUTION
of being quite blue in color and having reasonably clearly defined arms in the
spiral structure.
From an all-sky sample of 208 such galaxies, radial velocities, that is, measure-
ments of the speed at which these objects are receding from the sun, have been
determined for 70-odd; 50 velocities range from 4000 to 7000 kilometers per
second. Strikingly, 28 of these have velocities tightly grouped (4966 ± 122 km/
sec) while 22 have a similar grouping (6431 ± 160 km/sec). Each group is posi-
tioned on a different region of the sky (approximately one-third of the total for
each) and there is virtually no overlap in the positions of the two groups. These
observations raise important problems; the most unorthodox is to question the
validity of the widely accepted idea that the Hubble expansion (which relates
recessional velocity to distance) is isotropic throughout the universe.
Much of the work of the Biophysics Section during the past decade has been
devoted to study of nucleic acids and especially to the matching of DNA strands
derived from the same and from different organisms. These studies have illumi-
nated questions about organization of chromosomes and evolution of the species.
Many mechanisms are at play in establishing genetic diversity even among
bacteria. Ultimately, according to present understanding, diversity is traceable
to the message encoded in the nucleic acids. Some portions of bacterial or viral
genomes are highly conserved, as would be required for the production of enzymes
and enzyme systems necessary for accommodation to the environment or changes
in the environment and hence to survival. Enzymes, being proteins and generally
antigenic in foreign species, are subject to immunologic analysis. Modern organic
chemistry has made it possible to couple appropriate antibodies to insoluble sub-
strates such as the polysaccharide sepharose and thus permit antibodies to select
enzymic antigens from crude bacterial extracts with high efficiency. Throughout
the Enterobacteriaceae the distinctive features of the regulation of the enzyme
aspartokinase I-homoserine dehydrogenase I (AKI-HSDI) are conserved even
though the compositions of the DNAs vary widely. It is accordingly possible to
investigate immunologic affinities of the enzyme, the conformational states of the
enzyme as ligands are introduced into the system, and to draw inferences con-
cerning the systematic relationships of specific proteins as contrasted with the
similarities and differences between the total genomes indicated by nucleic acid
interactions among the bacteria. These studies are described in detail by Dean
Cowie and his collaborators, Georges N. Cohen and Paolo Truffa-Bachi of the
Pasteur Institute in Paris.
Roy Britten has continued his studies on the organization of chromosomes.
These studies are being carried out at the Kerckhoff Marine Laboratory in Cali-
fornia where he collaborates with a group of California Institute of Technology
biologists.
THE DEPARTMENT OF EMBRYOLOGY
The interplay of biological and medical research has been a feature and a
strength of the Department of Embryology since its inception. The field of human
embryology was long virtually synonymous with the Department, in which much
of the anatomic basis was established for today's worldwide thrust toward an
understanding of congenital defects. In his report this year, James D. Ebert,
Director, enumerates some of the many interrelations of the Department's work
REPORT OF THE PRESIDENT 19
with medical research and practice. Much of modern reproductive physiology,
one of the fields contributing to the technology of population control, sprang from
the Department and its pioneering monkey colony. In their day, the Lewises
were pioneers in the then budding technology of cell culture. It is on such a base
that much of today's animal virology, vaccine production, and cytogenetic screen
for genetic defects rests. The Department's contributions in the field of develop-
mental immunobiology, notably its role in helping to elucidate the graft-versus-
host reaction, have influenced the course of clinical research in tissue transplanta-
tion. Although it is still too early to predict accurately, it seems likely that
Masako Yoshikawa-Fukada and Ebert's pioneering demonstration of the viral
oncogenic sequences may help in understanding tumorigenesis.
The fundamental research now being carried on in the Department is concen-
trated on (1) the isolation of nuclear and mitochondrial genes and the study of
their structure, evolution, and control, and (2) the development of membranes.
Most of this research at present is obviously not related' closely to medical prob-
lems, though most of the studies will ultimately contribute toward better practice.
Already one of the lines of research on membranes has developed a close relation
to medicine. This is work of Douglas Fambrough and his colleagues on my-
asthenia gravis. This disease is a neuromuscular disorder characterized by mus-
cular weakness and fatigue. It has been thought to involve the neuromuscular
junction. In collaboration with Daniel Drachman and S. Satyamurti of The
Johns Hopkins University School of Medicine, Fambrough has conducted experi-
ments designed to determine whether muscle fibers from patients with the disease
show changes in the number and distribution of acetylcholine receptors. The
answer is unequivocal: The neuromuscular junctions of patients with typical
myasthenia gravis contain only 11 to 32 percent of the control number of ACh
receptors. This finding permits more critical questions to be asked. Is the number
of receptor molecules reduced? Or are many receptor molecules blocked or altered,
resulting in a decrease in functional receptors?
Fambrough and his colleagues will be able to examine these and other questions
pertaining to acetylcholine receptors. During the last few years he has developed
and evaluated an excellent method of measurement of the receptors, based on the
use of a-bungarotoxin combined with 125I.
One of the great miracles that we may possibly never fully penetrate is develop-
ment of the embryo from an egg. The process is perhaps more accurately a whole
series of miracles, one of the most dramatic of which is the beginning of beating
in the embryonic heart and the tendency of embryonic heart cells to beat in
unison. This has long fascinated Robert DeHaan, and he and his associates have
conducted a series of investigations on the topic.
DeHaan and Howard Sachs had demonstrated that spheroidal aggregates pre-
pared from dissociated heart cells beat at rates which were inversely proportional
to aggregate volume; that is, the bigger an aggregate, the slower it would beat.
In the past year, working with an able group of students, DeHaan launched a
multifaceted analysis of this unexpected finding.
Two sets of properties that might, influence pulsation rate change with increas-
ing size of such aggregates are (1) those properties that are related strictly to
number of cells in the system and are independent of spatial configuration, and
(2) those properties that depend on spatial organization of the parts of the system.
To perform experiments designed to distinguish between these broad classes
of rate-controlling phenomena, DeHaan has worked out methods for manipulating
20 CARNEGIE INSTITUTION
spontaneously beating aggregates one at a time under direct visual observation.
When brought into contact, aggregates beating at different rates adhere, and soon
(usually within 30 to 60 min) synchronize to a common rhythm. Thus, pairs or
groups of aggregates of selected rates and sizes can be joined to form systems of
any desired spatial configuration and allowed to synchronize.
One approach to factors governing beat rate was carried out by Linda Moyzis,
an undergraduate student, who built up linear chains of 6 to 20 aggregates and
demonstrated that the synchronous pulsation rate of these chains was only slightly
reduced from the average rate of the component aggregates prior to synchroniza-
tion. In this case, the rate was not a function of total synchronized cell number.
A second approach was taken in experiments performed by Caroline Krueger,
also an undergraduate student. When sheets of cells were prepared at known
densities (in cells/mm2 on the plate) and cut into rectangles of different areal
sizes, Krueger found that the synchronous pulsation rates of the rectangles did
not differ from each other in a systematic way over a 60-fold range of areas
(i.e., number of synchronized cells). Again, rate was found not to be a function
of synchronized cell number. When, however, sheets were prepared at two differ-
ent thicknesses (4 cells or 6 cells deep) , the thicker rectangles beat only half as
fast as the thin ones. The controlling factor appeared to be not cell number but
diffusional restriction between the inside of the mass and the outside, or S/V
ratio.
A major fraction of the Department's activities has been focused on study of
nucleic acid structures. This area of effort is one of the most fundamental and
dynamic of all biological research today. Peter Wellauer and Igor Dawid have
come up with a most interesting series of discoveries relating to structure of RNA.
They initially focused their attention on the secondary structure of 28S ribosomal
RNA. Their electron micrographs showed a reproducible secondary structure
which permits the mapping of these molecules — different regions of each molecule
and their polarities becoming recognizable. They next applied the technique to
the study of the "processing" of the ribosomal RNAs. It has been known for
some time that the 28S and 18S rRNAs originate by the "processing" (or splitting)
of a larger 45S precursor. Wellauer and Dawid have compared the structures of
the 45S, 28S, and 18S molecules and have clarified their relationships. The 28S
region is located at the 5'-end of the 45S precursor. It is followed by a "spacer"
region, then an 18S region, and another "spacer" at the 3'-end. In "processing,"
the "spacers" are removed. This technique, which holds great promise for the
elucidation of nucleic acid structures, comes at an opportune time when the ques-
tion of chromosome structure is coming in for careful attention in laboratories
throughout the world.
Dawid has also been deeply immersed in another investigation of unusual
interest. In collaboration with Ivan Horak, a second-year Carnegie Fellow
recently returned to Wurzburg, and Hayden Coon of the National Cancer Insti-
tute, Dawid has been studying the mitochondrial DNAs (mtDNAs) of somatic
hybrid cells (human-mouse and human-rat). Many readers of last year's Report
will recall that mtDNAs of both species can replicate in the same cell for extended
periods. That much can be said unequivocally. However, a more crucial question
is this: Do they simply "coexist" in the same cytoplasm, or do they interact in
some way? Even more difficult to answer is the further question, do they recom-
bine, forming a hybrid molecule? Molecular hybridization studies, detailed by
Dawid, make it very likely that mouse mtDNA segments become covalently
REPORT OF THE PRESIDENT 21
linked to human mtDNA during growth in the hybrid cells. A number of hybrid
cell strains have been examined. Thus far, five strains have both parental mtDNAs
without recombination, but there are 13 strains in which recombination has
apparently taken place.
Another experimental system in which molecular hybridization techniques have
revealed significant gene recombination is found in Donald Brown and E. Jordan's
study of the origin of ribosomal DNA homogeneity. Brown's earlier studies had
shown that the genes for 28S and 18S rRNA in two species of Xenopus, laevis
and mullein, are quite similar, if not identical, indicating that these sequences
have not evolved significantly. However, the associated spacer regions differ by
as much as 10 percent of their some 7000 base pairs. How do these sequences
evolve — some genes being maintained in a homogeneous state, others changing
rapidly? It has been proposed that unequal crossing-over can maintain homo-
geneity in a family of genes. Brown and Jordan have been examining a number
of possibilities. They have reared hybricl frogs derived from matings of X. laevis
and X. mulleri, and backcrossed one such animal with both laevis and mulleri.
The progeny are now sexually mature and the extent to which the ribosomal
genes have been recombined has been determined. If there were extensive recom-
bination, the backcrossed animals should have laevis and mulleri genes adjacent
to each other. The further question may then be asked, when meiosis occurs, will
there be a "correction"? Brown and Jordan present the several possibilities clearly
in their account. Here we need only say that 20 individuals from each group of
backcrossed frogs have been analyzed. All animals tested in the backcross with
a laevis male contain some mulleri rDNA. Twenty progeny of the mother's back-
cross with a mulleri male were also examined. Although none had as much
laevis rDNA as the mother, 18 out of 20 had some laevis rDNA. Brown and
Jordan believe that this result, while not yet conclusive, is strong evidence for
extensive recombination in the oocyte between the homologus chromosomes
containing rDNA.
Ebert has maintained a personal research program designed to help bridge the
intellectual and biological hiatus between nuclear material and membranes. To-
gether with Peter Stambrook and Sachs, he has been examining the close relation-
ship that exists between cell membrane potential and cell proliferation. In earlier
experiments the membrane potential was manipulated by varying the external
potassium concentration, and the results suggested that the membrane potential
might participate in the initiation or perpetuation of DNA synthesis. The ex-
periments this year were designed to determine whether an increase in membrane
potential is required for cells to begin DNA synthesis and to traverse the S phase
and G2 to their next mitotic division. A second question examined was whether
the high K+ medium blocks cells randomly throughout the cell cycle, or whether
it blocks cells at specific time points in the cycle. To gain experimental evidence
in these matters, Chinese hamster cells were exposed to various external potassium
concentrations.
Three conclusions emerged from a series of experiments. (1) Cells can initiate
and continue DNA synthesis while exhibiting a low membrane potential. Syn-
chronized cells incubated with control medium but transferred to 115 rnikf K+
medium prior to the initiation of DNA synthesis can traverse the S phase and G2
to their next division with little delay. (2) The high potassium medium impedes
the passage of cells through mitosis. (3) There is an event (or events) during Gi
that is particularly sensitive to the 115 mM K+ medium and that may be required
22 CARNEGIE INSTITUTION
for the initiation and perpetuation of DNA synthesis. Mitotic cells suspended
in 115 ml K+ medium exhibit impaired DNA synthetic activity. However, if
115 ml K+ medium is added 2 hours after mitosis, the cells proceed through
DNA synthesis and the remainder of the cell cycle almost normally. Presumably
the high K+-sensitive block is passed during the 2-hour incubation in control
medium, allowing subsequent traverse of the remainder of the cycle in high K+
medium.
THE DEPARTMENT OF PLANT BIOLOGY
The end of the report year was a time of transition for the Department of Plant
Biology. C. Stacy French, who guided the Department for 26 years, retired from
his post as Director and has been succeeded by Winslow Briggs. The new Director
will conserve and build on the present program which contains elements central
to plant biology, including study of the process of photosynthesis and investiga-
tions in physiological ecology. In both these areas the Department has made
distinguished contributions and has helped train many of the leaders. The new
Director will take new initiatives. He will broaden the program to include more
emphasis on plant biochemical studies, particularly on membrane processes.
During French's tenure as Director, the frontiers of plant biology evolved, and
the research program of the Department changed with them. In the past quarter
of a century the idea that the mechanism of photosynthesis could be deduced
solely from kinetic measurements of such effects as oxygen evolution, fluorescence
intensity changes, or rates of carbon dioxide uptake was slowly abandoned. In
place of that procedure, the detailed examination of the overall photosynthetic
process into its individual chemical steps was found to yield more definitive infor-
mation. In addition to the various forms of chlorophyll, preparative isolation
biochemistry has shown the presence of a number of highly reactive compounds
such as cytochromes, plastoquinones, high-energy phosphates, pyridine nucleo-
tides, plastocyanin, and nonheme iron compounds in the photosynthetic system.
Current biochemical work centers on the interactions of such substances when
light is absorbed by the chlorophyll complexes of plant chloroplasts. Kinetic
measurements now usually refer specifically to changes in known intermediate
substances rather than to hypothetical components. Many of these known inter-
mediate electron carriers of the photosynthetic system change color with their
oxidation state, so their participation in photosynthesis can be followed by meas-
uring their absorption changes at appropriate wavelengths.
Close observation of the plant pigments and their changes on illumination has
been and continues to be a fundamental area for investigation. French and other
members of the Department of Plant Biology have long been active in such studies.
In recent years much of French's efforts have been devoted to identifying and
measuring the various forms of chlorophyll in plant material. One of his strengths
has been the development of increasingly powerful and sensitive instrumentation
for such work. This tradition of a flare for instrumentation is being carried on
by David Fork, who has been fortunate in having the collaboration of Tetsuo
Hiyama. They are exploiting the latest techniques in tunable lasers and in fast
responding circuitry to develop instrumentation ultimately capable of observing
photobiological effects occurring in the order of 10"7 seconds. Equipment capable
of meeting their objectives is not commercially available. They have been able to
utilize some component circuits but have found it necessary to design and build
REPORT OF THE PRESIDENT 23
others. While they are developing their equipment, they have been using it. and
have already been able to measure cytochrome changes in times as short as several
microseconds.
Olle Bjorkman and his colleagues have created a very important new facility
at the Department for studies of physiological ecology. This effort is in part a
joint program with the Department of Biological Sciences of Stanford University.
Knowledge of the physiological and biochemical mechanisms that enable plants
to photosynthesize efficiently in the great diversity of environments that exist on
earth is fundamental to understanding primary productivity in different ecosystems
and of plant evolution and distribution. Such knowledge, inevitably derived from
experiments on wild plants occupying extreme environments, is also critical to any
assessment of the potential possibilities of breeding crop plants that can be suc-
cessfully cultivated in areas presently considered unsuitable.
The mechanisms underlying the responses and adaptation of plants to contrast-
ing environments continue to be in the focus of the physiological ecology investi-
gations. Most of the group's efforts this year have been devoted to the establish-
ment of new field transplant gardens, design and construction of new measuring
equipment, controlled growth facilities, and extensive remodeling and modifica-
tion of already existing mobile and stationary laboratory equipment and facilities.
A specially constructed, self-contained mobile laboratory unit will permit
measurements of photosynthetic gas exchange characteristics on the whole
plant and single leaf levels to be performed with the same instrumentation in the
field, in the transplant gardens, and at the controlled growth facilities at Stanford.
As now equipped, the unit houses all of the measuring systems needed for con-
tinuous measurements of the exchange of carbon dioxide, oxygen and water vapor,
and conductance to the diffusive gas transport by whole plants and single leaves
as well as for measurement of external parameters such as radiant energy
and quantum fluxes, air flow rates and pressures, air, tissue and soil temperatures,
and water vapor, C02 and oxygen concentrations. This integrated system also
incorporates an extremely compact computerized data acquisition, data processing
and control system which should greatly improve the speed of data acquisition,
the flexibility and convenience of operation, and the range and accuracy of the
measuring system.
The new transplant gardens, the controlled growth facilities, the mobile labo-
ratory, and other instrumentation will certainly provide an excellent and in several
respects a unique tool in the comparative study of plant adaptation to ecologically
diverse environments and of the underlying physiological and biochemical
mechanisms.
One of French's happy characteristics as a Director has been an ability to
attract very capable visiting investigators to the Department. They have come in
part because of the reputation and opportunities of the place. They have also
come because of the atmosphere that has been maintained there. Over the years
scores of leading plant biologists from many countries have worked at the labo-
ratory. In his report this year, French provides a list of those who have worked
at the laboratory during the past decade. The presence of these guest investiga-
tors has been stimulating to staff, and the guests have generally been successful
in their research efforts. This year was not an exception, as can be seen from a
brief description of some of the results obtained this year at the laboratory.
The accessory pigments of red algae, phycoerythrin and phycocyanin, function
in photosynthesis by transferring their absorbed light energy to chlorophyll.
24 CARNEGIE INSTITUTION
Ulrich Schrieber of Aachen, Germany, who spent part of the summer of 1972 at
the Department, found that hydrostatic pressure greatly reduces the efficiency
of energy transfer from the phycobilin pigments to chlorophyll in red algae.
Atusi Takamiya, former head of the Department of Biochemistry and Bio-
physics at the University of Tokyo, spent five months at the Department on leave
from Toho University. He found that a water soluble and phototransformable
chlorophyll protein is far more widely distributed in wild plants than had previ-
ously been realized.
Ralphreed Gasanov, Assistant Professor of Plant Physiology at Baku, Azer-
baijan, USSR, visited the laboratory for eight months. He found that the photo-
system fraction 1 isolated from chloroplast grana by digitonin differs in its pig-
ment composition from the corresponding material in the chloroplast stroma
particles located between the grana stacks.
With the hope that eventually a more direct use of solar energy may be
possible, Jeanette Brown and Gasanov are attempting to understand the state
of chlorophyll in vivo and how this essential pigment functions in photosynthesis.
They have approached this problem by dividing chloroplasts into the smallest
units that still retain photochemical activity. Successfully applying an improved
and extended fractionation procedure to an alga that has unusually distinct
absorption bands of chlorophyll a, they were able to obtain fractions of Dunaliella
that had activity only for photosystem 1 or only for photosystem 2, the reducing
and oxidizing light reactions of complete photosynthesis.
Curve analyses of the absorption spectra of these fractions revealed the relative
proportions of the forms of chlorophyll present in each photosystem. Chlorophyll
a-694 is lacking in the photosystem 2 fraction and is present in all of the fractions
with photosystem 1 activity. The longer wavelength forms of chlorophyll (>700
nm) seen in spinach chloroplast particles are not found in this alga, suggesting
that these forms are not essential for photosynthesis but may function as addi-
tional light harvesting pigment.
During his return visit to the laboratory last summer, Norio Murata, from the
Department of Biochemistry and Biophysics of the University of Tokyo, collabo-
rated with Fork on a continuation of their study on the function of the copper
protein plastocyanin in photosynthesis. Their previous work, reported in Year
Book 70, showed that little plastocyanin remained in particles prepared by ex-
trusion through a needle valve. These particles nevertheless showed rapid electron
transfer reactions from cytochrome / to the reaction center of system 1 (P700) —
as fast as those seen in chloroplasts in their native state. These observations
suggested that cytochrome /, rather than plastocyanin, functions as the "second-
ary electron donor" to P700. However, the lack of plastocyanin in these particles
was questioned by some workers, particularly by those who used the so-called
bioassay method to determine plastocyanin. Murata and Fork found that, although
the bioassay method is highly sensitive for plastocyanin, it is not necessarily
specific for this substance. After doing extensive studies on the contents of plasto-
cyanin in chloroplasts and in subchloroplast particles prepared by a variety of
methods to release plastocyanin, they confirmed their earlier conclusions that
plastocyanin is absent from particles derived by needle valve extrusion and that
it does not function between cytochrome / and P700. It would seem from these
results that cytochrome / thus plays the role of the secondary electron donor to
P700.
REPORT OF THE PRESIDENT 25
THE PROGRAM OF INSTRUCTION
The Institution has two principal, related objectives. One is to advance knowl-
edge. The other is to foster intellectual growth through educational activities.
The young people who are trained by the Institution are a highly select group.
They receive special individualized tutelage from senior staff and they have
unusually good access to advanced facilities. Predoctoral and postdoctoral- fellows
do not serve lengthy apprenticeships as members of research teams. Rather, they
are encouraged to engage in carefully monitored but independent scholarship. In
consequence both of the major objectives of the Institution are advanced, for as
the individuals grow they are creative in research.
The academic orientation of the Institution is particularly obvious in the three
departments that are closely associated with particular universities.
In the joint operation of the Hale Observatories by the Institution and the
California Institute of Technology, teaching and research are inextricably inter-
twined. Staff Members of the Observatories serve as thesis advisers for Ph.D.
candidates and help train students in observational astronomy by supervising
various research projects. During this report year, of the 120 major papers listed
in the Observatories' bibliography, 28 were either single author papers by post-
doctoral Fellows or joint efforts with Staff Members, and an additional 19 were
written by or were in collaboration with Caltech graduate students.
The training in, and the teaching of, research is a major activity of Staff
Members. Particularly active at the Hale Observatories is the infrared studies
group under Neugebauer. Fellows during the recent past include Harry Hyland,
now returned to Australia to begin his own infrared group, and David Allen and
C. G. Wynn-Williams of the United Kingdom. As is usually the case, these
Fellows provide the seed nuclei for the spread of science in their own countries.
Becklin, a Staff Associate of the Observatories, also a member of the infrared
group, discovered the Galactic Center in infrared radiation while he was a gradu-
ate student at Caltech, preparing his thesis using telescopes on Mount Wilson.
During the current report year, notable student activities within the Carnegie
Institution at the Observatories include (1) a pre-thesis project in which Caltech
student W. G. Bagnuolo is collaborating with Searle and Sargent on the history
of star formation in galaxies as related to questions of chemical evolution of the
galaxies as a whole and the age distribution of blue systems; (2) the mastery by
van der Kruit (Carnegie Fellow) of techniques of spectroscopy of extended
galaxies, which has led him to develop further the analysis of noncircular (ex-
plosive?) motions in some galaxies with active nuclei; (3) the studies of Robert
Kirshner (a predoctoral student at Caltech) in collaboration with Searle and
Sargent in which he developed the interpretation of supernova spectra to an extent
that fundamental distances to these stars might be determined from their spectra
alone, leading perhaps to an independent calibration of the scale of the universe;
(4) the studies of Michael Hart (Carnegie Fellow) with Robert Howard that have
shed light on the mystery of the solar limb redshift of spectral lines, perhaps
solving an enigma of 60 years; (5) the investigations of Ermanno Borra (Carne-
gie Fellow) studying with Vaughan, that helped to develop new methods of
measuring magnetic fields in white dwarfs — a fundamental problem in stellar
evolution; and (6) the finding by Eduardo Hardy (Carnegie-Chilean Fellow,
under Sandage's guidance), that Abell's hypothesis — that there exists a unique
26 CARNEGIE INSTITUTION
feature of the luminosity function of clusters of galaxies that can be used as a
general distance indicator — is untenable.
In addition to the graduate programs, the Observatories have a summer program
for the training of selected undergraduates in methods of research by apprentice-
ships to Staff Members. Particularly noteworthy are the studies of Robert Harri-
son, a Caltech undergraduate in astronomy, under Kristian and Westphal in the
laboratory tests and development of reduction of techniques for the silicon-vidicon
area photometer. Harrison's exposure to research problems at a professional level
and his fine ability have been a benefit both to the Institution and to himself.
The Institution's Department of Plant Biology occupies ten acres in the middle
of the Stanford University campus. Naturally, there is intense interaction be-
tween Staff Members and Fellows of the Department and faculty and students at
Stanford. Joint seminars, joint projects, and a sharing of facilities are customary.
The educational ties of the Institution to Stanford University were reinforced
recently when Briggs, the new Director at Plant Biology, was named Professor
of Biological Sciences at Stanford. Briggs' new appointment includes membership
in the academic senate of Stanford.
Graduate students as well as postdoctoral students benefit from the educa-
tional experiences provided by the Department. William Hagar, a postdoctoral
Fellow, commented recently that his experience was immeasurably enriched by
associating with visiting scholars of such renown as Professor Atusi Takamiya,
formerly from the University of Tokyo, and Professor Hemming Virgin of the
University of Goteborg, and that the opportunity for exchange of ideas between
young and mature scholars is one of the many exciting opportunities available
in the Institution. Charles Weiss, who recently completed his undergraduate
studies in the Biology Department of Whitman College, Washington, was associ-
ated with the physiological ecology group for several weeks during the summer
and found that the Department serves as a testing ground for students who
want to experience research in the plant sciences before committing themselves
to further graduate studies.
The sharing of facilities and programs with Stanford is particularly evident
in the physiological ecology studies. The mobile laboratory owned by Stanford
is docked at the Department where it keys into experimental facilities. The
total enterprise, including field stations in Death Valley and at Bodega Bay, is
operated jointly as a cooperative venture by Bjorkman of Plant Biology and
Professor Harold Mooney of Stanford. The studies involve participation of many
students.
Other facilities maintained by the Department both at Stanford and at the
field stations provide a wealth of educational opportunities not only to those
working in the plant sciences but also to those working in other fields.
The mountain stations, for example, provide a valuable teaching and research
resource. This summer Professor Theodosius Dobzhansky along with Professor
Francisco Jose Ayala and associates from the University of California at Davis
continued their experiments on the genetics of the fruit fly Drosophila at the
Mather Station. Richard Papp, an entomology graduate student from the
University of California at Berkeley, used the Timberline Station as a base f( '
his research on the ecology and distribution of snowfield insect populations.
Another entomology student, John Horn, has used the Mather Station in his
REPORT OF THE PRESIDENT 27
study of insects associated with the Sierra Corn Lily (Veratrum). David Burdin,
a student of Professor Larry Swan from California State University at San
Francisco, was at the Mather Station also working on a research project in
zoology.
Stanford University makes continued use of the mountain stations as part of
its teaching program. Each year Professor John Thomas and Professor Mooney
from the Department of Biological Sciences take field trips with their students
to these stations. Professor Craig Heller and his graduate students are using the
Timberline Station for their study of the distribution of rodents across the Sierra
range.
Professor Clifford Schmidt and associates from the California State University
at San Jose have used the Mather and Timberline Stations in their study of the
effects of overuse by people and livestock on the fragile mountain ecosystems.
This study may be useful in defining what limits must be placed on mountain
wilderness use.
The Department of Embryology, located on the campus of The Johns Hopkins
University, interacts strongly with that institution while maintaining its own
fellowship program. When asked to comment on educational activities at Em-
bryology, Ebert submitted the following brief essay.
How does one evaluate an educational program, one in which no examina-
tions are set and no grades assigned, one whose purpose is to aid in the produc-
tion and training of investigators? There are many yardsticks, of course — the
placement of graduates and fellows and their ultimate rise to positions of
responsibility, their imprint, in turn, on others, and most importantly the
impact of their research on the field. One measure of the last criterion is the
citation index: How frequently does a given article, or the output of a given
individual, reappear in the literature? The approach can be extended to the
output of a given department or institution.
The Twenty-Fifth Anniversary volume of the American Institute of Biolog-
ical Sciences, Challenging Biological Problems,1 was designed to kindle thought
and spark discovery. This collection of articles by distinguished biologists was
meant to be not a series of commonplace reviews, but a series of blueprints for
the future, in which only the most critical building stones were to be used as the
basis for projection. It is especially interesting to see that in F. H. Wilt's
chapter on developmental biology, "The impact of molecular biology on study
of cell differentiation," 23 of the 98 articles cited were the products of Carnegie
students and fellows, and of their students in turn. It is a small sample, to be
sure, but it is reassuring nonetheless to see how significant an impact the Insti-
tution has had in this field in the past decade.
The Institution's Staff Members are teachers, but not in Louis Agassiz's
words "simple machines of study." In fact Professor Agassiz's address on July
8, 1873, the opening day of the Anderson School of Natural History,2 in which
he described his own relationship with his students, epitomizes some of the
relations between preceptor and student in the Department of Embryology.
Agassiz said, in part, "Every morning when I can be here I propose to give you
a piece of advice about the way to employ your time. In proportion as you
have made some progress it will be of different character. I shall constantly
ask you to tell me what you have seen to see how your mind works. I shall never
make you repeat what you have been told, but constantly ask you what you
1 J. A. Behnke, editor, Oxford University Press, New York, 1972.
2 A. H. Wright and A. A. Wright, The American Midland Naturalist, 43(2) : 503-506. 1950.
28 CARNEGIE INSTITUTION
have seen yourselves. What I want you to do in order to profit by this is to
work yourselves."
Today the spirit is much the same, but today's students have a further
advantage. They are constantly challenged not just by one preceptor, but by
a close-knit staff and student group, all anxious to see that every investigation
moves forward, being shaped by criticism at every step.
Thus it should not be surprising that of the highlights in this year's report of
the Director of the Department of Embryology, four research projects involved
fellows or students — three postdoctoral fellows, Horak, Sachs, and Wellauer,
and one undergraduate, Camilla Velez. Moreover, as already observed, these
represent only a sampling of the Department's activities. One sees the influence
of fellows and students at every turn.
One final striking similarity to Agassiz's philosophy should be noted. Agassiz
insisted that his students be his equals in their access not only to problems but
to equipment as welL Speaking of aquaria and related items, he wrote "I have
ordered one for every person admitted to the school, so that each of you will
have means to make these investigations. I have never had in my own labora-
tory better opportunities for work than I place at your disposal." So it is
today: outstanding studies are supported, of Staff Member and student alike.
Though not situated on university campuses, the Geophysical Laboratory and
the Department of Terrestrial Magnetism cooperate with many universities both
here and abroad.
Over the years students from more than a dozen different American universities
have conducted their doctoral thesis work at the Geophysical Laboratory. Post-
doctoral Fellows have come from most of the major universities of this nation
as well as from 16 foreign lands. During the report year, 3 predoctoral and 12
postdoctoral Fellows were in residence. In the current climate of restricted federal
support for fellowships, openings throughout the Institution and, of course, at the
Geophysical Laboratory, are eagerly sought.
Yoder says that the success of the predoctoral and postdoctoral instructional
program hangs primarily on the selection of worthy candidates. All of the
numerous applicants submitting research proposals are carefully reviewed by
a Selection Committee composed of Staff Members and resident Fellows. There
are no known predictive measures of motivation, innovation, aggressiveness,
patience, determination, and all the other personality factors which contribute to
the success of a researcher in the laboratory environment. Yet, the Selection
Committee attempts to advise the Director of those individuals who are most
likely to grow and achieve their goals. How does one compare the overworked
lecturer burning with desire to return to research, the bright-eyed graduate
student bubbling with new ideas, and the established professor craving critical
discussions with his peers. Each wants to achieve that feeling of satisfaction of
contributing new knowledge to the world, and each must do it in his own way.
The selection process, although based mainly on intuition, must single out the
imaginative researcher from the limited facts available. By means of the re-
search proposal and aided by a biography, bibliography, curriculum vitae, personal
interviews with the staff, and three letters of recommendation from persons of
the applicant's choice, the growth potential and capabilities of the applicant are
judged. In spite of the large amount of time required for the selection process,
each applicant is given a fair evaluation by the Selection Committee. Their
selections are judged further and in the light of the scientific goals and economic
factors within the Institution by the Director and finally by the President. The
REPORT OF THE PRESIDENT 29
Geophysical Laboratory has enjoyed an unusual measure of success in choosing
candidates who have grown academically, have been productive, and have become
leaders in their field.
At the Geophysical Laboratory as elsewhere in the Institution predoctoral
Fellows enjoy the same privileges as postdoctoral Fellows and Staff Members.
There is no ranking among the scientists except that which accrues from the
accumulation of successful endeavors. One predoctoral Fellow expressed con-
siderable enjoyment in not being treated as a student. Another found the
environment of a research laboratory more conducive to rapid advancement than
that obtained from the occasional personal contacts with university professors.
Still another predoctoral Fellow had the unusual experience of setting up his
laboratory from scratch — an intensive, and somewhat disconcerting, educational
process. The growth observed in predoctoral Fellows is unusually large and most
gratifying to the other Fellows and Staff who have contributed to their maturation.
The Department of Terrestrial Magnetism has continued its instructional and
research efforts in accord with the decades old pattern of close student-Staff
Member relationships. This year 8 predoctoral and 12 postdoctoral Fellows re-
ceived training at the Department. These Fellows had a variety of backgrounds
and experiences. Three Ph.D.'s whose graduate training was in high energy
nuclear physics or theoretical physics have been broadening their horizons in
a way that will make them especially versatile. Two young men have joined
the Biophysics Section where they are becoming acquainted with and are con-
ducting research in molecular biology. One has already published a significant
report in the Journal of Molecular Biology. The young theoretical physicist has
made effective progress and has submitted for publication a new analytical
approach in the study of strain release in a prestressed medium. With the paucity
of academic positions presently available in physics and the massive turning
away of students from the study of physics, it is highly desirable that young
men such as these be encouraged. Years of experience at the Department of
Terrestrial Magnetism have proved that a solid training in physics provides an
excellent background for the individual to branch out into other areas and to
bring fresh and novel ideas to other subjects. Recent endeavors strongly support
this contention.
The Department has also continued to train postdoctoral Fellows in areas
where the Fellow brought to the Department considerable related background
and expertise. These Fellows are studying in close cooperation with Staff Members
in each of the three major areas of the Department's activities: Astrophysics
(including nuclear and atomic physics), Biophysics, and Geophysics (includ-
ing geochemistry) where they are mastering new techniques and unique instru-
ments, and applying them to problems of mutual interest.
The Department has undertaken to train and advise a considerable number
of predoctoral students both at the Department and off campus. Three (in bio-
physics, nuclear physics, and seismology) study in the DTM laboratories and
under staff guidance will be preparing theses to be submitted to Purdue, Howard,
and Hokkaido Universities, respectively. Three South American students (on
leave from the University of Chile, the Servicio Geologico y Mineria, Peru, and
the Instituto Geofisico del Peru) are being supported for formal graduate course
work at the University of Alberta, McGill University, and the University of
Edinburgh. Close contact is maintained with these students and cooperative
30 CARNEGIE INSTITUTION
studies in geomagnetics, geochemistry, and seismology are planned or are under
way. One of the Department's Staff Members is supervising in part the research
and thesis preparation of a predoctoral student in astronomy at Harvard. While
interim Director of the joint CIW-IAR Observatory in Villa Elisa, Argentina,
Turner taught a formal course in radio astronomy and also supervised the studies
of ten students who used the radio telescope facility and are degree candidates
at the University of La Plata and the University of Buenos Aires. Additionally,
four students from the Department of Physics and Astronomy of the University
of Maryland, in an ongoing cooperative educational program, were taught the
essentials of operating the 60-foot radio telescope at Derwood, Maryland, and
were given the opportunity to carry on a modest research project. Under Aldrich's
sponsorship and guidance, three students (University of Buenos Aires, Cuyo
University at San Juan, Argentina, and the University of Chile in Santiago) were
introduced to field work, data reduction, and interpretation in terrestrial electrical
conductivity studies along the Argentine-Chilean border. This effort is one of a
long series of interactions of the Department with predoctoral and postdoctoral
students of Latin America. Seeds planted earlier are now evident in a flourishing
and first-class geophysical capability of the Andean countries.
PERSPECTIVE
The President's report must of necessity represent only a sampling of the
activities of the Institution which are described more fully in the reports of the
various Directors. Even there, however, the accounting is somewhat uneven.
Were the total activities and contributions to be set forth fully, as much as 3000
printed pages might be required instead of the 700 pages currently employed.
In reporting their activities, the various Departments have differing audiences
and differing objectives.
The Geophysical Laboratory uses the Annual Report for two principal pur-
poses— an accounting to the Trustees of the Institution and an accounting to
scientific colleagues throughout the world. Thus some 2650 copies of the Labora-
tory's section of the report are sent to colleagues in a total of 68 countries. The
recipients tell us that they especially value this collection of the work of the
Laboratory.
In contrast, some of the other Departments report more briefly highlights of
their activities and depend heavily on communication to colleagues by means of
the scientific literature.
Another way in which the report is uneven is in describing some of the
activities that are essential to effective research and to the health of the scientific
and educational enterprise but do not lead directly to scientific publications.
For example, our facilities, such as those of the Hale Observatories, are used
by many visiting students and investigators. Ministering to such guests and
maintaining facilities for them places a substantial burden on staff.
Educational activities outside the Institution's facilities are also only briefly
treated. Our staff annually gives numerous lectures in many universities and
serves as thesis advisers and on Visiting Committees.
Another facet of our activities that is accorded minor attention is public
service. The staff contributes far more than its proportionate share to such efforts.
These include service on various government advisory groups and responsibilities
as officers of scientific societies.
REPORT OF THE PRESIDENT 31
LOSSES . . .
This year the Institution lost the distinguished astronomer Ira S. Bowen, who
died February 6, 1973. Dr. Bowen, Director of the Hale Observatories from 1946
to 1964, was a pioneer in the 1920s in the application of quantum theory to
spectroscopy, and he applied this knowledge to an insightful interpretation of
stellar spectra. Some of his greatest scientific contributions were in instrument
and telescope design. Among many astronomers he is regarded most highly for
his achievement in placing the 200-inch Palomar Mountain telescope into splendid
working order. When delivered to the mountain from the optical shop in
Pasadena the 200-inch mirror was not in finished form. In order to bring it to
final figure, a temporary polishing machine was set up on the floor of the dome.
Then the shape of the mirror was gradually completed. This entailed first taking
star photographs through a Hartmann screen placed over the mirror, careful
measurement of the resulting plates, removal of the mirror from the telescope to
the polishing machine, some judicious polishing, replacing the mirror in the
telescope, and then repeating the whole cycle until no further work was required.
This tedious and demanding job was completed in about a year. The final result
was a telescope of excellent optical quality.
A native of Seneca Falls, New York, Dr. Bowen received his A.B. degree in
1919 from Oberlin College. Shortly thereafter he joined the faculty of the newly
reorganized California Institute of Technology, where he did some of the work
for which he will be best remembered.
When the Carnegie Institution and California Institute of Technology began
joint operation in 1948, Dr. Bowen became the first Director of the Mount Wilson
and Palomar (now Hale) Observatories. After his retirement in 1964, he con-
tinued to apply himself to problems of optical design. He was responsible for the
optical design of the recently completed Palomar 60-inch telescope, the wide-
field Las Campanas 40-inch telescope, and the 101 -inch Irenee du Pont telescope
for the observatory under construction at Las Campanas in Chile.
C. Stacy French retired on June 30, 1973, after twenty-six years as Director of
the Department of Plant Biology. Under his direction the Department has added
to our knowledge of the ways plants adapt to different environments and it has
increased our understanding of that complicated and vitally important chemical
process called photosynthesis.
Dr. French's own work has centered around the role of plant pigments in
photosynthesis. He has measured thousands of fluorescence, absorption, and
action spectra of chlorophyll and other pigments. From one series of experiments
on fluorescence spectra he was able to establish that energy is transferred from
accessory pigments to chlorophyll during photosynthesis. In the course of his
investigations Dr. French invented several useful laboratory devices, such as the
derivative spectrophotometer for studying chlorophyll in vivo. With this machine,
French and his associates confirmed the existence of various forms of chlorophyll
in the living plant and discovered a new form of chlorophyll that absorbs light
at 695 millimicrons.
Another of his inventions automatically plots the rate of photosynthesis
of algae illuminated at different wavelengths, making it possible to obtain action
spectra and enhancement spectra for photosynthetic reactions and to study their
interactions. The French pressure cell, which he designed to rupture plant cells
32 CARNEGIE INSTITUTION
without the damaging effects caused by detergents or sonication, is also widely-
used in the study of yeasts and bacteria and for the preparation of vaccines.
Dr. French has also made a distinct contribution to botanical research in his
influence on the development of younger scientists. About one fourth of the
people working in photosynthesis today have studied at the Department of Plant
Biology, most of them as Carnegie Fellows.
It is with the deepest sorrow that I report the death on March 8, 1973, of
Edward A. Ackerman, Executive Officer. Dr. Ackerman served the Carnegie
Institution for fifteen years. His wise counsel in helping to shape its policies
was invaluable. We have lost a distinguished scholar, and a warm and loyal
friend.
Dr. Ackerman came to the Carnegie Institution in 1958 as Assistant Executive
Officer and became Executive Officer in 1960. The perspective and good judgment
with which he carried out his responsibilities were derived in part from his
broad background in academia, science administration, and consulting work.
Dr. Ackerman was especially interested in environmental systems, particularly
population and water resources. He authored or co-authored eight books and
more than sixty articles on these subjects, and was consultant to the Departments
of Interior and State, the House Foreign Affairs Committee, the Senate Committee
on Natural Resources, the President's Panel on Water Resources Research, the
National Water Commission, the United States Geological Survey, and the United
Nations Economic Commission for Europe.
Dr. Ackerman will be remembered for his patience, unfailing good humor, and
extraordinary human sensitivity. He embodied the spirit and ideals of the
Institution, and he will be missed by everyone who had the privilege of knowing
him.
I must also report the death on April 18, 1973, of Dr. Alfred H. Joy, who was
associated with the Institution for fifty-eight years. Dr. Joy had been a Staff
Member of the Mount Wilson Observatory from 1915 to 1948. He was especially
interested in the chemical spectra of stars, and he participated in the spectral clas-
sification and radial velocity surveys that provided the data upon which much of
the later progress in astrophysics and galactic structure was built. In 1945 he
pointed out a new class of recently formed variable stars designated "T Tauri";
it is thought that these objects may be the key to the understanding of how stars
are formed.
In addition to his scientific work, Dr. Joy assisted in the administration
of the Observatory by serving as Secretary from 1920 until his retirement in
1948. He carried on an active program of postretirement research, and at the
age of 90 still walked two miles a day to study stellar spectra.
Harlow Shapley, who died on October 20, 1972, changed the way all of us
picture the universe. Shapley gave us a truer, more useful model of what lies
beyond our world, and he gave modern astronomy its most important tool for
determining celestial distances, the period-luminosity relationship of Cepheid
variable stars.
It was early in his career, while he was at the Mount Wilson Observatory,
that Shapley pursued the study of Cepheid variables that led to his most im-
portant discoveries. By measuring changes in the temperature and spectra of
REPORT OF THE PRESIDENT 83
the Cepheid variable stars, he was able to calculate their rates of pulsation and
from those learn their intrinsic brightness. Their observed brightness could then
be measured as an index of distance. Using this relationship between the period
and the luminosity of the Cepheids as a kind of measuring stick for the rest of the
galaxy, Shapley plotted the location of swarms of stars called globular clusters.
From the distribution of the globular clusters he calculated that our solar system
is nowhere near the center of its galaxy, as had been supposed, and that the
galaxy is larger than astronomers had believed.
. . . AND GAINS
One of the year's notable gains was the addition of Lewis McAdory Branscomb
to the Board of Trustees. Dr. Branscomb's career has combined outstanding
accomplishments in both science and administration.
He was graduated summa cum laude from Duke University in 1945 and received
the M.S. and Ph.D. degrees in physics from Harvard University in 1947 and
1949. He has taught physics at Harvard, University College (London) , the
University of Maryland, and the University of Colorado.
Branscomb joined the National Bureau of Standards in 1951. He served in
the atomic physics division from 1954 to 1962. For the next seven years he was
chairman of the Joint Institute for Laboratory Astrophysics in Boulder, Colorado,
a cooperative venture with the University of Colorado, until President Nixon
appointed him Director in 1969. As Director, Branscomb led the Bureau to a
major role in stimulating industrial technology.
In May 1972 he was appointed chief scientist of the International Business
Machines Corporation; the following month he was elected vice-president and is
responsible for planning the research program.
Dr. Branscomb's career has included service on a number of boards and com-
mittees, including the President's Science Advisory Committee, the President's
Committee for the Medal of Science, the Council of the National Academy of
Sciences, the Board of Directors of the American Association for the Advancement
of Science, and the Board of Trustees of the Rand Corporation.
Dr. Branscomb has received the Rockefeller Public Service Award, the Gold
Medal for Exceptional Service from the U.S. Department of Commerce, and the
Procter Prize from the Scientific Research Society. He is a Fellow of the American
Academy of Arts and Sciences, the American Philosophical Society, and the
American Physical Society. He holds honorary doctor of science degrees from
Duke, Western Michigan, and Rochester Universities.
To succeed Dr. French as Director of the Department of Plant Biology, the
Institution is fortunate in having one of the leaders in modern developmental
biology, Winslow R. Briggs. Dr. Briggs plans to continue his work in plant
growth and development and to establish this as one of the areas of research
at the Department. At the same time, he has considerable interest in physiological
ecology and the biochemistry of photosynthesis, and he intends to maintain and
strengthen the Department's programs in these fields.
Dr. Briggs comes to the Institution from Harvard, where he has been professor
of botany since 1967. From 1955 to 1967 he was instructor and later associate
and full professor of biology at Stanford University. While at Stanford Dr. Briggs
visited our Department of Plant Biology and participated in several studies there.
34 CARNEGIE INSTITUTION
Dr. Briggs is well known for his research on the pigment phytochrome, the
photoreceptor that conveys the required information to the plant about light in
the environment. Among many other effects phytochrome mediates seed germina-
tion, stem elongation, and the orientation of leaves to best capture sunlight for
photosynthesis. Phytochrome also ensures that flowering, seed formation, and
dormancy in many plants occur at the appropriate times of the year.
Briggs' work is concentrated on the elucidation of the molecular mechanisms
that allow phytochrome to regulate plant metabolism. He has greatly advanced
the techniques for the isolation of this chromoprotein in a highly purified state.
His studies on the protein chemistry, structure, and the photochemistry of phyto-
chrome have aided in the understanding of the mechanism of action of this
substance. Currently he is investigating the possibility that phytochrome trans-
formations may give rise to alterations of membrane properties, which in turn
can lead to profound biochemical changes.
A graduate of Harvard University, Briggs received his bachelor's degree in
1951 and his doctorate in 1956. He is a member of the American Association for
the Advancement of Science, the Botanical Society of America, and the American
Society of Plant Physiologists. Briggs is the author of more than sixty published
articles on such topics as growth hormone physiology, plant tropisms, and plant
pigment function, and he is editor of the Annual Review of Plant Physiology.
I am pleased to announce the following honors to Staff Members.
Donald D. Brown of the Department of Embryology was elected to the
National Academy of Sciences. At the annual meeting of the Academy, Dr.
Brown received the U.S. Steel Foundation Award in Molecular Biology.
James D. Ebert, Director of the Department of Embryology, received the
honorary degree of Doctor of Science from Yale University. Dr. Ebert was
also awarded the President's Medal of the American Institute of Biological
Sciences and was named as one of the first Eminent Scientists in an exchange
program sponsored by the Japan Society for the Promotion of Science.
Vera C. Rubin of the Department of Terrestrial Magnetism was elected to the
Board of Directors of the Association of Universities for Research in Astronomy,
Inc., as a Director-at-Large, to serve until 1976. Dr. Rubin was also elected a
member of the Nominating Committee of the American Astronomical Society for
a three-year term, and was appointed a member of the Space Astronomy Com-
mittee of the Space Science Board of the National Academy of Sciences from
January 1973 to March 1975.
Hatten S. Yoder, Jr., Director of the Geophysical Laboratory, was given the
Abraham Gottlob Werner Medal of the Deutsche Mineralogische Gesellschaft,
the highest scientific award of the society. The American Academy of Achieve-
ment selected him as one of the recipients of the Golden Plate Award for his
pioneering work in the study of rock and mineral synthesis in the earth's crust
and upper mantle.
C. Stacy French, Director of the Department of Plant Biology, received the
Award of Merit from the Botanical Society of America.
Barbara McClintock, Distinguished Service Member of the Institution, had a
laboratory building at Cold Spring Harbor Laboratory named for her in recogni-
tion of her contributions to genetics research.
REPORT OF THE PRESIDENT 35
Stanley R. Hart of the Department of Terrestrial Magnetism was appointed
to the Panel on Orientations for Geochemistry of the U.S. National Committee
for Geochemistry, National Academy of Sciences— National Research Council.
Ronan O'Rahilly and John Tucker of the Department of Embryology won the
silver medal at the annual meeting of the American Academy of Ophthalmology
and Otolaryngology for their exhibit on the embryology of the human larynx.
Fritz Zwicky, who is engaged in postretirement studies at the Hale Observa-
tories, was awarded the Gold Medal of the Royal Astronomical Society.
36
CARNEGIE INSTITUTION
FACULTY, FELLOWS, AND STUDENTS
1972-1973
DEPARTMENT OF EMBRYOLOGY
Baltimore, Maryland
Director
James D. Ebert
Staff Members
Donald D. Brown
Igor B. Dawid
Robert L. DeHaan
Douglas M. Famb rough
Richard E. Pagano
Ronald H. Reeder
Fellows
Joseph F. Albright
Dana Carroll
Tasuku Honjo
Ivan Horak
Paul Lizardi
Terence F. McDonald
Keiko Ozato
Aileen K. Ritchie
Howard G. Sachs
Kazunori Sugimoto
Yoshiaki Suzuki
Peter K. Wellauer
Robert Williamson
Guang-Jer Wu
Masako Yoshikawa-Fukada
Assistant Investigator
Peter J. Stambrook
Students or Predoctoral Fellows
Sandra L. Biroc
Peter Devreotes
Thomas E. Durr
Scott Gilbert
H. Criss Hartzell
Carol Kaushagen
Caroline R. Krueger
Dennis E. Leister
Linda Moyzis
Calvin E. Plitt
Ralph H. Stern
Camilla Velez
GEOPHYSICAL LABORATORY
Washington, D.C.
Director
Hatten S. Yoder, Jr.
Staff Members
Peter M. Bell
Francis R. Boyd, Jr.
Felix Chayes
Gordon L. Davis
David H. Eggler1
Larry W. Finger
P. Edgar Hare
Thomas C. Hoering
T. Neil Irvine 2
Thomas E. Krogh
Ikuo Kushiro
Ho-Kwang Mao 3
Robert N. Thompson
David Virgo
Alain Weisbrod
Emeritus Research Associate
Emanuel G. Zies
Postdoctoral Fellows
Jagannadham Akella
Donald M. Burt
John S. Dickey, Jr.
John D. Frantz
Bastiaan J. Hensen
Floyd N. Hodges
Kenneth King, Jr.
James M. Mattinson
Karlis Muehlenbachs
Yoshikazu Ohashi
Michael G. Raymond
Martin G. Seitz
Predoctoral Fellows
John I. Hedges
D. James Misener
Bjorn 0. Mysen
1 Temporary Staff Member from September 1,
1972.
2 Temporary Staff Member from August 16,
1972.
3 Temporary Staff Member from September 1,
1972; Postdoctoral Fellow to August 31, 1972.
REPORT OF THE PRESIDENT
87
HALE OBSERVATORIES
Pasadena, California
Director
Horace W. Babcock
Associate Director
J. Beverley Oke
Distinguished Service Member
Ira S. Bowen 1
Staff Members
Halton C. Arp
Edwin W. Dennison
Jesse L. Greenstein
James E. Gunn
Robert F. Howard
Jerome. Kristian
Robert B. Leighton
Guido Munch
Gerry Neugebauer
George W. Preston
Bruce H. Rule
Allan R. Sandage
Wallace L. W. Sargent
Maarten Schmidt
Leonard T. Searle
Arthur H. Vaughan, Jr.
Olin C. Wilson
Harold Zirin
Staff Associates
Eric E. Becklin
Robert J. Brucato
Michael W. Werner
James A. Westphal
Fellows
Ermanno F. Borra
Michael M. Dworetsky
Michael H. Hart
Pieter van der Kruit
Andrew Michalitsanos
Ronald Moore
Jay M. Pasachoff
Stephen W. Prata
William J. Quirk
Katsuo Tanaka
Richard E. White
Robert R. Zappala
1 Died February 6, 1973.
Carnegie-Chilean Fellows
Eduardo Hardy
Maria Teresa Ruiz
Student Observers
Marc Aaronson
William G. Bagnuolo
Ian Gatley
Richard Green
Paul M. Harvey
Paul Hickson
James A. Howell
John P. Huchra
Robert P. Kirshner
John Kormendy
Elliot C. Lepler
Steven J. Loer
Jorge Melnick
Augustus Oemler, Jr.
Valdar Oinas
Glenn S. Orton
Paul L. Schechter
Stephen A. Shectman
Barry E. Turnrose
Glenn J. Veeder
Bruce Waddington
William E. Westbrook
Richard A. Wickes
Steven P. Willner
DEPARTMENT OF PLANT BIOLOGY
Stanford, California
Director
C. Stacy French
Staff Members
Joseph A. Berry
Olle Bjorkman
Jeanette S. Brown
David C. Fork
Malcolm A. Nobs
Staff Member Emeritus
William M. Hiesey
Guest Investigators
Georgi Detchev
Ralphreed A. Gasanov
Ulrich Schreiber
Atusi Takamiya
38
CARNEGIE INSTITUTION
Fellows
George Bowes
William G. Hagar
Tetsuo Hiyama
Norio Murata
Helga I. Ninnemann
John H. Troughton
Stanford Students and
Postdoctoral Fellows
Sherry DeRemer
Louis Pitelka
DEPARTMENT OF
TERRESTRIAL MAGNETISM
Washington, D.C.
Director
Ellis T. Bolton
Associate Director
L. Thomas Aldrich
Section Chairmen
L. Thomas Aldrich, Geophysics
Louis Brown, Astrophysics x
Dean B. Cowie, Biophysics
W. Kent Ford, Jr., Astrophysics 2
Distinguished Service Member
Merle A. Tuve
1 From January 1, 1973.
2 To December 31, 1972.
Staff Members
George E. Assousa
Roy J. Britten
Louis Brown
Dean B. Cowie
W. Kent Ford, Jr.
Stanley R. Hart
Albrecht W. Hofmann
Bill H. Hoyer
David E. James
Alan T. Linde
Nancy R. Rice
Richard B. Roberts
Vera C. Rubin
I. Selwyn Sacks
Diglio Simoni
Norbert Thonnard
Kenneth C. Turner
Fellows
Jesus A. Berrocal
Tom I. Bonner
Michael B. Davis
Borwin Grauert
Cidambi K. Kumar
Allan T. Leffler II
Hiromu Okada
Urs C. Rohrer
Nobumichi Shimizu
J. Arthur Snoke
Neil A. Straus
Thangasamy Velusamy
Donald G. Wallace
John W. Warner III
Trainee Fellows
Arturo Cuyubamba
Douglas M. Haefele
Godwin C. Igiri
Jack M. Klitzman
Ricardo Quiroga
Enrique Triep
Reports of Departments
and Special Studies
Department of Embryology
Hale Observatories
Department of Terrestrial Magnetism
Department of Plant Biology
Geophysical Laboratory
Department of Embryology
Baltimore, Maryland
James D. Ebert
Director
Carnegie Institution of Washington Year Book 72, 1972—1973
Contents
Introduction 7
The Isolation of Genes and the Study of Their Structure, Evolution, and Control . 10
The origin of ribosomal DNA homogeneity 11
Transcription of 5S DNA in vitro 14
Analysis of 5S DNA heterogeneity 15
The use of antibiotics in buoyant density gradients 15
Studies on silk fibroin mRNA and its genes 18
A polyadenylic acid sequence in silk fibroin mRNA 19
The Regulation of Fibroin Genes 20
Quantitative measurements of fibroin mRNA synthesis 20
Mutants for fibroin production 22
Fibroin mRNA from Bombyx mandarina 23
The Molecular Weight of Mouse Globin Messenger RNA 24
Control of Gene Transcription in Eukaryotes 25
Transcription of rRNA and 5S RNA genes in amphibian chromatin by homologous
RNA polymerase 26
Transcription of chromatin by bacterial RNA polymerase 28
Repression of Xenopus mulleri rDNA in X. laevis X X. mulleri hybrid frogs . 31
The Time of Ribosomal Gene Replication during the S Phase of Synchronized
Chinese Hamster Cells 35
Secondary Structure Maps of RNA Molecules Derived by Electron Microscopy and
Their Use for the Analysis of the Processing of Hela Ribosomal RNA ... 38
Biogenesis of Mitochondria 39
Recombination of mitochondrial DNA in somatic hybrid cells 40
Protein composition of cytoplasmic and mitochondrial ribosomes of Xenopus laevis 42
Mitochondrial RNA of Drosophila 43
Measurement of mitochondrial RNA and DNA-RNA hybrids in the electron
microscope 45
Transcription of mitochondrial DNA by mitochondrial RNA polymerase ... 46
Structure and Function in Cell Membranes 47
Interactions between lipid bilayers and cell surface membranes 47
Viscoelastic properties and surface tension of membranes 49
Membrane dynamics 49
Studies on Muscle Fiber Membranes 50
Acetylcholine receptors in skeletal muscle fiber membranes 50
Interaction of ACh receptors with a-bungarotoxin 50
Turnover of ACh receptors in myotube plasma membranes 51
Incorporation of new ACh receptors into myotube membranes 55
The neuromuscular junctions in myasthenia gravis: decreased acetylcholine
receptors 56
Electrophysiological properties of the membrane and cholinergic receptor of
developing myotubes 60
Membrane potential in developing myotubes 61
Ionic selectivity of the cholinergic receptor during development 62
Some effects of Triton X-100 on acetylcholine receptors 63
The Effect of Potassium on the Cell Membrane Potential and the Passage of Cells
through the Cell Cycle: A Block in Gx 64
A Cellular Response to Dibutyryl Cyclic AMP: A Cell Cycle Block in G2 . . . 69
Functional Differentiation of the Embryonic Heart 73
Modulation of sodium current channels in embryonic heart cells by tetrodotoxin . 73
Myofibrillar organization during the first beats of the chick heart 80
Control of synchronized pulsation rate in cell groups 83
Cytochalasin B and embryonic heart muscle: Contractility, excitability, and
ultrastructure 86
The Collection of Human Embryos 87
Developmental Stages in Human Embryos 87
Development of the conducting system of the heart 88
Development of the nervous system 88
Development of the larynx 89
Development of bones and joints 89
Staff Activities 89
Bibliography 91
Personnel 93
INTRODUCTION
The Department continued its work recognition by their peers. To the tech-
during the year without serious distrac- nician, who plays an integral part in re-
tions or interruptions. It has always been search, there is the personal satisfaction
a friendly and pleasant environment in of a job well done, but other forms of
which to work; in fact, it is one that is recognition are infrequent. Friends of the
not infrequently inspiring as the day-to- Department know how much all of us,
day regimen of research is punctuated by established Staff Members and incoming
new findings and ideas. Of course many Fellows alike, have relied on trusted col-
— if not most — of them are short-lived, leagues like Thomas Malooly, Margaret
but enough stand the test of criticism by Proctor and John Wiser. However, very
our peers so that our younger Staff Mem- few have appreciated the remarkable ex-
bers and Fellows pursue their ideas with tent to which many members of the tech-
the reassuring feeling that if they, too, nical staff have developed over the years,
persevere, the pieces of their puzzles will It is always difficult to single out exam-
fall into place. pies for special mention, but biologists
If any one factor can be singled out to like to use models to illustrate their argu-
account for the confident attitude and ments, so perhaps I can make my point
reflective mood that pervades the Depart- by saying that Eddie Jordan, Delores
ment, it is the continuity and stability of Somerville and Ernestine Flemmings,
our financial support. It is true that the among others, have shown remarkable
departmental budget has barely kept personal growth during their association
pace with inflation over the past several with the Department,
years, yet we have been able to proceed I began by saying that the year was
in the confidence not only that we have a one of uninterrupted work in the Depart -
stable "floor" for our research, but also ment. Significant progress is reported in
that the Institution has sufficient re- both of our principal streams of activity,
serves, sufficient flexibility, and, above (1) the isolation of nuclear and mito-
all, a sufficiently deep understanding of chondrial genes and the study of their
the scientific process to enable changes in structure, evolution, and control and (2)
course when they are demanded, either the development of membranes. A glance
by unexpected new findings or by the at the Bibliography reveals that the year
"drying up" of what had appeared to be was marked by a tide of published find-
promising leads. The fact that there have ings, but as I remarked on an earlier oc-
been no abrupt swings, no radical budget- casion, to say this gives a false impression
ary ups and downs, has enabled us to of the state of the Department, for the
continue growing in stature without vitality of the Laboratory is reflected less
growing in size and administrative com- in published and already "cold" data
plexity. than in the flow of new ideas and in the
I would mention a second key ingredi- constant search for new approaches to
ent, one that is too often overlooked. I long-resistant problems. In offering the
refer to the rich human resources in the following samples of our activities, I
Department's supporting staff. Investi- shall stress the novel findings and the ex-
gators derive deep personal satisfaction ploratory pathways, cautioning the
from a successful experiment or the distil- reader whose interest is whetted to look
lation and synthesis of a new concept. In carefully at the detailed accounts that
addition, however, they gain the added follow,
satisfaction of national and international No one approach to the structure of
s
CARNEGIE INSTITUTION
RNA has generated more interest than
Peter Wellauer and Igor Dawid's analy-
sis of secondary structure maps of RNA
molecules by electron microscopy. Their
electron micrographs of 28S ribosomal
RNA show a reproducible secondary
structure which allows the mapping of
these molecules — different regions of each
molecule and their polarities becoming
recognizable. The technique has been ap-
plied to a study of the "processing" of the
ribosomal RNAs. It has been known for
some time that the 28S and 18S rRNAs
originated by the "processing" (or con-
trolled fragmentation) of a larger 45S
precursor. Now the structures of the 45S,
28S and 18S molecules have been com-
pared in the electron microscope, and
their relationships clarified. The 28S re-
gion is located at the 5' end of the 45S
precursor. It is followed by a "spacer"
region, then an 18S region and another
"spacer" at the 3' end. In "processing" the
"spacers" are removed. Wellauer, an In-
ternational Fellow of the U.S. Public
Health Service, will continue these highly
promising studies for a second year.
Dawid has also been deeply immersed
in another investigation of unusual inter-
est and potential importance. I say "po-
tential" because, as he observes, "explicit
proof" of his conclusions is still forth-
coming. In collaboration with Ivan
Horak, a second-year Carnegie Fellow
recently returned to Wiirzburg, and Hay-
den Coon of the National Cancer Insti-
tute, Dawid has been studying the mito-
chondrial DNAs (mtDNAs) of somatic
hybrid cells (human-mouse and human-
rat) . It was found, first, that mtDNAs
of both species can replicate in the same
cell for extended periods. This much can
be said unequivocally. However, a more
crucial question is this: Do they simply
"coexist" in the same cytoplasm or do
they interact in some way? Even more
difficult to answer is the further question,
do they recombine, forming a hybrid
molecule? Molecular hybridization stud-
ies, detailed later in the Report, make it
very likely that mouse mtDNA segments
become linked to human mtDNA during
growth in the hybrid cells. A number of
hybrid cell strains have been examined.
Thus far, five strains have both parental
mtDNAs without recombination but
there are 13 strains in which recombina-
tion has apparently taken place.
Another experimental system in which
molecular hybridization techniques have
revealed significant gene recombination is
found in D. D. Brown and E. Jordan's
study of the origin of ribosomal DNA
homogeneity. The genes for 28S and 18S
rRNA in two species of Xenopus, laevis
and mulleri, are quite similar, if not
identical, indicating that these sequences
have not evolved significantly. However,
the associated spacer regions differ by as
much as 10% of their 7000 base pairs.
How do these sequences evolve — some
genes being maintained in a homogeneous
state, others changing rapidly? It has
been proposed that unequal crossing over
can maintain homogeneity in a family of
genes. Brown and Jordan have been ex-
amining a number of possibilities. They
have reared hybrid frogs derived from
matings of X. laevis and X. mulleri and
backcrossed one such animal with both
laevis and mulleri. The progeny are now
sexually mature and the extent to which
the ribosomal genes have been recom-
bined can be determined. If there is ex-
tensive recombination, the backcrossed
animals should have laevis and mulleri
genes adjacent to each other. When
meiosis occurs, will there be a "correc-
tion"? Brown and Jordan pose the sev-
eral possibilities clearly in their account.
It will suffice in this summary to say that
twenty individuals from each group of
backcrossed frogs have been analyzed.
All animals tested in the backcross with a
laevis male contain some mulleri rDNA.
Twenty progeny of the mother's back-
cross with a mulleri male were also ex-
amined. Although none had as much
laevis rDNA as the mother, 18 out of 20
had some laevis rDNA. Brown and Jor-
dan conclude that this result, while not
yet conclusive, is strong evidence for ex-
DEPARTMENT OF EMBRYOLOGY
9
tensive recombination in the oocyte be-
tween the homologous chromosomes con-
taining rDNA.
Progress in the Department's second, or
membrane, "track" has been equally stim-
ulating. Of the several impressive continu-
ing studies of Douglas Fambrough and his
colleagues, I have selected one for special
attention. Myasthenia gravis is a neuro-
muscular disorder characterized by mus-
cular weakness and fatigability.' It has
been thought to involve the neuromus-
cular junction. In collaboration with
Daniel Drachman and S. Satyamurti of
the Johns Hopkins University School of
Medicine, Fambrough has conducted ex-
periments designed to determine whether
the disease involves changes in the num-
ber and distribution of acetylcholine re-
ceptors in muscle fibers from dystrophic
patients. The answer is unequivocal: The
neuromuscular junctions of patients with
typical myasthenia gravis contain only
11 to 32% of the control number of ACh
receptors. This finding represents only a
beginning, of course, but it permits more
critical questions to be asked. Is the
number of receptor molecules reduced?
Or are many receptor molecules blocked
or altered, resulting in a decrease in func-
tional receptors? The writer cannot re-
frain from showing his"immunobiological
bias" to raise a further question: The
older literature speculated (with some
reason) that myasthenia gravis might be
an autoimmune disease, somehow related
to aberrant function of the thymus. Is
it possible that the two ideas might be
brought together? Might the blocking of
ACh receptor sites be accomplished by an
autoimmune reaction?
The Department's capacity to deal
with the structure and function of cell
membranes was enhanced by the arrival
during the year of Staff Member Richard
E. Pagano, who has undertaken the de-
velopment of physical and chemical tech-
niques for examining the surface proper-
ties of natural membranes. His initial
exploratory findings are summarized in
the body of the Report.
The cell membrane has also been a
focal point in Peter Stambrook's research.
With Howard Sachs and Ebert, he has
continued the analysis of the effect of
potassium on the cell membrane potential
and the passage of cells through the cell
cycle, demonstrating that Chinese ham-
ster cells are blocked in mitosis and in Gi.
With a Johns Hopkins undergraduate,
Camilla Velez, he has explored other
techniques of altering the cell cycle by
changes at the membrane, in a further
effort to probe the means whereby mem-
brane changes impinge upon the genetic
apparatus. He has succeeded in showing
that the application of dibutyryl cyclic
AMP [(but) 2 cAMP] blocks cells pri-
marily in G2 and to a lesser extent in
mitosis.
Fellows continue to play an important
role in the life of the Department. In
Donald Brown's laboratory Dana Car-
roll, Paul Lizardi, Kazunori Sugimoto,
and Senior Carnegie Fellow Robert Wil-
liamson contributed significantly. Car-
roll, in his first year as a Carnegie Fellow,
has begun to analyze the heterogeneity of
5S DNA; Lizardi, completing his two-
year stay as a Jane Coffin Childs Fellow,
has synthesized a radioactive reverse
transcriptase product of silk fibroin
mRNA ("anti-messenger" DNA) and
has established its specificity and sensi-
tivity. It should be useful as a tool in de-
tecting fibroin messenger nucleotide
sequences. Lizardi will return to Rocke-
feller University late in 1973. Second-
year Carnegie Fellow Sugimoto, who will
return to Nagoya University in August
1973, has begun to study the fidelity with
which 5S DNA is transcribed in vitro.
Williamson's contribution was an im-
portant one, not only in his specific
investigations (A polyadenylic acid se-
quence in silk fibroin mRNA ; with Well-
auer, the molecular weight of mouse
globin messenger RNA; plus a study in
collaboration with Stambrook to be re-
ported in a future Report) but also in his
generous sharing of ideas and techniques
10
CARNEGIE INSTITUTION
and his lively exchanges here and
throughout the Nation.
Second-year Carnegie Fellow Tasuku
Honjo, working with Reeder, also focused
his attention on the thorny question of
the control of gene transcription in eu-
karyotes. Substantial progress in defin-
ing the experimental system is reported.
On July 1, 1973, Honjo will move to the
National Institutes of Health to work in
association with Philip Leder.
Douglas Fambrough's laboratory was
enhanced by the addition of Aileen
Ritchie, a Fellow of the Muscular Dys-
trophy Association of America ; and Max
Springer of Zurich completed his second
year with Robert DeHaan. Springer's
artistry with the electron microscope, by
both scanning and transmission tech-
niques, is exhibited in Figures 50 and 51.
Among the graduate students, two
completed their studies. H. C. Hartzell
has taken up a fellowship appointment in
the Department of Neurobiology at Har-
vard Medical School, and Dennis Leister
will hold a fellowship in the laboratory of
Hans Kiintzel in the Max-Planck Insti-
tute for Experimental Medicine in Got-
tingen. Ralph Stern will transfer to The
Jackson Laboratory; and Carol Kaus-
hagen, working with Dawid, has begun to
look at the mitochondrial RNAs of Dro-
sophila.
Losses — and Gains. Robert DeHaan
joined the Department on July 1, 1956.
For seventeen years he has been a de-
voted colleague and effective contributor
who has established himself as a leading
student of cardiogenesis — of both mor-
phogenesis and functional differentiation,
including the origin and differentiation of
the pacemaker. In addition, he has con-
tributed importantly to the local com-
munity and to the scientific community
at large. I report his departure on June
30, 1973, with regret, tinged with pleasure
— regret at the loss of a valued colleague,
but pleasure in the fact that in his new
role as Professor of Anatomy at Emory
University he will be able to provide
leadership to a new research group em-
phasizing the developmental physiology
of the heart specifically, and the proper-
ties of developing membranes more gen-
erally.
I am pleased to report that Yoshiaki
Suzuki, currently an Extramural Fellow
of Carnegie Institution working in the
National Institute of Health of Japan
(for a report of his studies, see pp. 20-24) ,
is expected to rejoin the Department as a
Staff Member in the fall of 1973.
In addition, plans are under way to
bring Masako Yoshikawa-Fukada, also
an Extramural Fellow, currently working
in the Institute for Virus Research in
Kyoto (see Bibliography for her most
recent contribution), back to the Depart-
ment early in 1974 to resume her studies
of cell transformation.
It is also good to announce that Ernest
Gardner, Professor of Neurology at the
University of California School of Medi-
cine at Davis, has accepted our invitation
to serve as Associate Curator of the Car-
negie Collection when it is moved to the
Davis campus in the summer of 1973.
Professor Ronan O'Rahilly remains
Curator of the Collection.
THE ISOLATION OF GENES AND THE STUDY
OF THEIR STRUCTURE, EVOLUTION,
AND CONTROL
D. D. Brown, D. Carroll, P. Lizardi, R. Stern, K. Sugimoto, and R. Williamson
Previous Year Books have outlined our their controls in vitro. Experiments in
approach to developmental genetics. We the earlier phase of this project have in-
have isolated genes of known function volved methods for isolating these genes
with the long-range goal of reconstructing from complex mixtures of DNA and anal-
DEPARTMENT OF EMBRYOLOGY
11
yses of their structure and arrangement
along the DNA molecule. More recently
we have undertaken studies of the evolu-
tion of these genes. The repetitive genes
termed rDNA and 5S DNA have been
purified from Xenopus laevis and Xeno-
pus mulleri. Studies in this report de-
scribe analyses of the arrangement and
homogeneity of these highly repetitive
genes. These genes appear to evolve as a
family and the means by which this "par-
allel" or "horizontal" evolution takes
place are under experimental test.
Whereas the genes for 18S and 28S RNA
and their spacers appear to be homo-
geneous within a species, the multiple
genes for 5S RNA are similar to each
other but nonetheless heterogeneous even
within a single animal. What mecha-
nisms cause multiple DNA sequences to
evolve together, and when heterogeneity
exists, how is it arranged in the genome?
A second area of our research deals
with silk fibroin messenger RNA and its
genes. Two kinds of problems are being
pursued. We are studying the length of
DNA encoding the structural fibroin
genes, and methods are being developed
to isolate this DNA component. Secondly,
the structure and metabolism of the
mRNA itself are under investigation. We
are developing an assay system with
which we hope to analyze transcriptional
control of fibroin messenger RNA syn-
thesis. Studies of polyA content of the
messenger RNA are also summarized.
The Origin of Ribosomal DNA
Homogeneity
D. D. Brown and E. Jordan
The DNA containing the genes for 18S
and 28S RNA is clustered at a single
chromosomal locus in Xenopus and con-
tains about 450 extremely similar if not
identical repeating nucleotide sequences.
Half of each repeat includes the genes for
18S and 28S rRNA and these sequences
have not evolved detectably between two
species of Xenopus — mulleri and laevis.
The spacer regions, however, are very
different and must differ by more than
10% of their approximately 7000 base
pairs. How do hundreds of DNA se-
quences evolve together? One group of
explanations, referred to as gene "rectifi-
cation," have adjacent sequences being
corrected to conform to one of a limited
number of these repetitious sequences.
Several mechanisms for this have been
proposed — notably the "master-slave"
model of H. G. Callan and the "expan-
sion-contraction" model. These kinds of
models require unusual molecular mecha-
nisms for which we do not now have evi-
dence. In contrast to this, frequent un-
equal crossing over can maintain a family
of genes homogeneous, as has recently
been pointed out by George Smith of the
University of Wisconsin. This could oc-
cur at meiosis between homologues and/
or between sister chromatids. In addition,
"scrambling" of sequences can occur at
mitosis of germ cells by sister chromatid
exchange.
We have begun a series of experiments
designed to test these different possibil-
ities. It is implicit in these studies that
the mechanism which maintains sequence
homogeneity must correct the DNA
faster than the mutation rate will cause it
to diverge. The gene regions remain
homogeneous due in part to selective re-
straints on rRNA divergence. In order to
observe the correction mechanism experi-
mentally, the process will have to occur
at a relatively high frequency — at least
once per animal per generation. If cor-
rection is less frequent our experiments
will not detect it.
We have raised hybrid frogs derived
from matings of X. laevis and X. mulleri
(Fi) and backcrossed one such animal
with both X. laevis and X. mulleri. The
animals are now sexually mature. We
wish to measure the extent to which
ribosomal and 5S genes have recombined
in the progeny. If this occurs at high fre-
quency, these backcrossed animals should
have genes from one species adjacent to
those of the other. When such a recom-
bined locus goes through meiosis, will it
12
CARNEGIE INSTITUTION
be "corrected"? If not, can we at least
assess the extent to which recombination
occurs within these gene clusters?
When an Fi animal termed Im —
the progeny of a laevis female and a
mulleri male) is backcrossed with a laevis
male (termed 11) , the predicted segrega-
tion of progeny rDNA genotypes is ex-
pected to be 50% Im and 50% 11. The
extent to which the animals deviate from
this ratio is a measure of several possible
factors: selection of certain genotypes in
the progeny, chromosomal abnormalities,
or prior recombination within the rDNA
locus. Selection is always possible, since
most of the progeny from these matings
die from abnormalities due to problems
introduced by the hybrid nature of the
chromosomes. We must, therefore, study
the chromosome composition of each
animal to show that it still contains only
two homologues each with a single nucle-
olar organizer. In addition, a careful
study of the number of ribosomal genes in
each animal together with the relative
abundance of X. laevis and X. mulleri
rDNA assesses the extent to which un-
equal crossing over has occurred in those
animals. If crossing over has occurred in
somatic cells, different tissues of the same
animal may have different numbers of
rRNA genes. However, if exchange oc-
curred in meiosis, individual animals
would vary in gene number, but all the
cells of an individual should have the
same number of genes.
Another possible influence on gene cor-
rection that can be tested experimentally
is the importance of gene amplification.
Gene correction might occur by amplifi-
cation of one or a limited number of
chromosomal copies followed by exten-
sive recombination of the extra copies
with the chromosomal copies. This would
effectively remove any divergence among
chromosome copies. In its extreme form,
this possibility is also testable in hybrid
animals. We know from previous experi-
ments that Fi females only amplify X.
laevis rDNA. If there were efficient re-
combination of all chromosomal rDNA
copies with amplified rDNA molecules,
then progeny of the backcrosses would
never receive any X. mulleri rDNA from
the female; those copies would have been
replaced with X. laevis rDNA.
The DNA of each hybrid animal has
been analyzed for its content of rDNA
from both parents. This is carried out by
molecular hybridization of complemen-
tary RNA mixtures which have been syn-
thesized with E4 coli RNA polymerase
from purified rDNA of X. laevis and X.
mulleri. The two cRNAs are synthesized
with different isotopic precursors (3H and
32P-labeled nucleoside triphosphate),
mixed together, and hybridized with
somatic DNA from individual animals.
The ratio of isotopes bound to each DNA
is a measure of the abundance of the two
kinds of rDNA spacers that are present.
One experiment is shown in Fig. 1.
Twenty individuals from each of the two
groups of backcrossed frogs have been
analyzed for the rDNA content of their
somatic DNA. The results show that the
progeny have not followed the predicted
Mendelian segregation. All animals tested
in the backcross with an X. laevis male
contain some X. mulleri rDNA whereas
half of them should have only X. laevis
rDNA. About half of the animals have
approximately equal amounts of the two
kinds of rDNA like the mother (Im) . The
other half have about two-thirds X. laevis
rDNA. The question then is raised
whether the mother passed on any X.
laevis rDNA to her progeny. This is
answered by her other backcross with an
X. mulleri male. Twenty of these progeny
were examined. None of the twenty prog-
eny had as much X. laevis rDNA as the
mother. As can be seen in Fig. lb, ten of
them have 10 to 30% X. laevis rDNA.
The remainder of the animals which ap-
peared to have pure X. mulleri rDNA by
this test were examined by a more sensi-
tive technique. The somatic DNA from
each animal was centrifuged to equilib-
rium in an actinomycin CsCl gradient.
The X. mulleri and X. laevis rDNAs
separate by this method. Hybridiza-
DEPARTMENT OF EMBRYOLOGY
13
Q_
CVJ
ro
to
A
4
Imxll
2
9wO
1.0
0.8
• /
0.6
0.4
n o
1 1 1
i
■
0
20 40 60 80
Percent X.laevis rDNA
100
CL
CVJ
ro
\
X
to
B
6
4
-
-Em x mm
2
-
^m s'
1.0
0.8
0.6
0.4
-mm
o*
n o
i 1
i i
0
20 40 60
Percent X.laevis rDNA
80
100
Fig. 1. The relative abundance of X. mulleri and X. laevis rDNA in backcrossed hybrid animals.
Somatic DNA was purified from each animal and hybridized with a mixture of cRNAs. (3H)cRNA
and (32P)cRNA were transcribed from denatured X. laevis and X. mulleri rDNA, respectively, with
E. coli polymerase. They were mixed with a 100-fold excess of nonradioactive rRNA and hybrid-
ized. A standard curve was prepared by including filters with known mixtures of X. laevis and X.
mulleri rDNA. The DNA of the mother and father of the progeny was measured also in each ex-
periment. (A) Im X 11. (B) Im X mm.
14
CARNEGIE INSTITUTION
tion of fractions across each gradient
showed that 18 out of the 20 progeny had
some X. laevis rDNA. This result, while
not yet conclusive, is strong evidence for
extensive recombination in the oocyte be-
tween the homologous chromosomes con-
taining rDNA.
Experiments that measure the num-
bers of each kind of gene and their chro-
mosomal location are in progress. Satu-
ration hybridization experiments are be-
ing used to determine the numbers of
each kind of gene; while their chromo-
somal location is being studied by in situ
hybridization of metaphase chromosomes,
in collaboration with Mary Lou Pardue
at Massachusetts Institute of Technol-
ogy.
If we assume that selection of these
genotypes in the backcrossed progeny has
not occurred and they truly resemble the
rDNA genotype of the germ cells in the
F1 female, then there appear to be two
kinds of rDNA-containing chromosomes
in the female which are transmitted to
her progeny — those with pure X. mulleri
rDNA and those with about equal
amounts of each kind of rDNA. This
suggests but does not prove extensive re-
combination between the rDNA of homo-
logues. It further demonstrates that the
amplified rDNA copies in the oocyte,
which in this animal are only X. laevis,
do not recombine with the chromosomal
A", mulleri rDNA extensively. On the
contrary, the progeny inherit X. mulleri
rDNA more abundantly from the female
Fi than they do her A", laevis rDNA.
In summary, we now have populations
of hybrid animals which exhibit unusual
transmission of rDNA genes to their
progeny. The nature of this non-Mendel-
ian segregation is not yet known, but one
possibility is that these genes have under-
gone extensive recombination during
meiotic pairing. Experimental schemes
for testing this and other possibilities
have been detailed.
If extensive recombination exists
within this locus, it prompts us to specu-
late about the origin and evolution of
other highly repetitive regions of the
genome. Frequent unequal crossing over
might account for the process referred to
as "saltatory" replication, which has been
used to account for the origin of multiple
repeats within satellite DNAs. This kind
of mechanism can not only drastically
change the number of tandemly arranged
repeats but also cause them to evolve to-
gether as a family.
Transcription of 5S DNA in vitro
K. Sugimoto
Experiments were begun to study the
fidelity with which 5S DNA is tran-
scribed in vitro by added RNA polymer-
ases. 5S DNA from X. laevis is particu-
larly advantageous for these studies for
several reasons: (1) Since the repeat
length is short (about 0.5 to 0.6 X 106
daltons), individual 5S DNA molecules
contain many repeats and therefore pre-
sumably many initiation and termination
sequences. (2) The exact region of 5S
DNA which is transcribed in vivo is
known. It consists of about one-seventh
of one strand ; the remainder of the strand
and the opposite strand are not tran-
scribed in vivo. Our evidence suggests
that there is no precursor to 5S RNA in
vivo, so initiation occurs exactly at the 5''
end of the molecule. (3) 5S RNA itself
has been sequenced so that proper initia-
tion would be recognized by sequence
techniques. (4) A variety of Xenopus
RNA polymerases have been purified by
R. Roeder and are available for tran-
scription of 5S DNA.
Our studies to date have assayed the
transcripts made by E. coli polymerase
from purified 5S DNA, and recently we
have done similar experiments with X.
laevis RNA polymerase III in collabora-
tion with Roeder. Both kinds of enzymes
transcribe both strands of 5S DNA as
well as the spacer sequences. Various salt
concentrations have been tried with the
same nonspecific results.
Another approach to transcriptional
fidelity has been to study isolated nuclei.
DEPARTMENT OF EMBRYOLOGY
15
Endogenous RNA synthesis proceeds for
only about ten minutes in frog nuclei.
Techniques have not yet been perfected
in which reinitiation of RNA synthesis
by endogenous polymerase can take
place.
Analysis of 5S DNA Heterogeneity
D. Carroll
We are using bacterial restriction en-
zymes to investigate the heterogeneity
and organization of the genes for 5S RNA
in Xenopus. Here we describe experi-
ments designed to ask whether adjacent
repeats on a single molecule of 5S DNA
are always identical in sequence. Let's
construct some hypothetical repeat units
(one gene and one spacer) , use them in
various combinations to represent the
total 5S DNA, and see what sort of pat-
terns of restriction fragments might re-
sult. Suppose that repeat type A has no
sequence which is cleaved by a certain
restriction endonuclease, repeat B has one
such sequence, C has two, D has three,
etc. If the 5S DNA is made up entirely
of repeats of a single type, the enzyme
will produce a homogeneous pattern of
fragments. The number of different frag-
ments will equal the number of cuts per
repeat, and the sum of each kind of frag-
ment will equal the repeat length. For
example, multiple repeats of C would lead
to only two different DNA fragments. If,
on the other hand, the DNA contains sev-
eral types of repeats, and repeats of each
type are clustered, i.e., A's next to A's, B's
next to B's, etc., the overall pattern of
fragments will be a sum of the patterns
for each type alone. DNA made up of
clusters of A's and clusters of B's would
lead to a mixture of long, uncut molecules
(A) and repeat length-sized fragments
(B). If repeats of different sequence are
interspersed randomly, a very heteroge-
neous collection of fragments will result.
But if the interspersion is regular and re-
peating—e.g., . . . BCDBCDBCD . . . —
a unique pattern of fragments will be ob-
served, and the sum of the fragment
lengths (if they can all be distinguished)
will equal the length of the larger repeat
(B + C + D).
If, in addition to differing in sequence,
the repeats differ in length, the distribu-
tion of fragment length will be very
heterogeneous. If there is one cut made in
each repeat, the distribution of fragments
will be identical to the distribution of re-
peat lengths; but this may be practically
indistinguishable from the case in which
both the number of cuts and the repeat
lengths are heterogeneous. If there are
internal repetitions within the spacer re-
gions, the distribution of fragments might
be more uniform than the repeat lengths
themselves.
In our initial experiments, we have
examined in the electron microscope frag-
ments produced by treatment of 5S
DNAs with various restriction enzymes.
The Ri enzyme from E. coli apparently
makes no cuts in either X. laevis or X.
mulleri 5S DNA. Digestion of X. laevis
5S DNA with the H. influenzae restriction
enzyme preparation endo R leads to the
production primarily of repeat length-
sized pieces (approximately 0.5 X 106
daltons) . X. mulleri 5S DNA shows a
heterogeneous distribution of fragments
on treatment with endo R. Since the
known restriction sequences of these en-
zymes are not found within the sequence
of 5S RNA itself, we presume that any
cuts which do occur are made in spacer
sequences. We are now investigating in
detail the exact size distribution of pieces
produced by endo R. This plus other
kinds of experiments should elucidate the
way in which heterogeneity of sequences
is arranged in 5S DNA.
The Use of Antibiotics in Buoyant
Density Gradients
R. Stem
The utility of antibiotics as ligands in
buoyant density gradient centrifugation
has been investigated and actinomycin D
has proven to be a great aid in fractionat-
16
CARNEGIE INSTITUTION
1.8-
ro
£
1.7
u
\
o>
1.6
>N
•4—-,
oj
1.5
c
<D
Q
10
15 5 10
Fraction Number
Fig. 2. CsCl buoyant density gradients of X. mulleri DNA. About 400 fig of DNA was centri-
fuged in 20 ml gradients for 2 days at 35,000 rev/min in a 50.1 fixed angle Spinco rotor. The buoy-
ant density of each fraction was determined from refractive index measurements. (A) control (B)
200 fig of actinomycin.
ing DNA. Actinomycin D binds to DNA thus the highest density bind the most
and makes it less dense. Usually those actinomycin D and become the lightest
DNAs with the highest GC content and fractions in a gradient. In Figure 2A a
ro
E
m
c
(V
Q
c
E
o
1.65
B
1.60
1.55
-
* >
S>
800
J
400
^
J
/
1 \i
\
\
\ 1
• 0
10
Fraction Number
Fig. 3. Hybridization of CsCl buoyant density gradients of X. mulleri DNA with cRNA tran-
scribed from 5S DNA and rDNA. About 100 fig of DNA in 5 ml gradients was centrifuged for 2
days at 35,000 rev/min in a 40-angle rotor. (A) control (B) 50 fig of actinomycin added. Dashed
line, cRNA from 5S DNA; open circles, cRNA from rDNA.
DEPARTMENT OF EMBRYOLOGY
17
20-ml gradient containing 400 fig of
Xenopus mulleri DNA is shown. Two
satellite DNAs are seen in analytical
gradients at 1.683 and 1.686 g/cm3. They
are not resolved here, but their approxi-
mate positions are indicated by the
arrows. In Figure 2B a similar gradient
containing 200 fig of actinomycin D is
shown. The satellite DNAs are now well
resolved from each other as well as from
the main-band DNA.
Figure 3A shows a 5-ml CsCl gradient
containing 100 ^g of X. mulleri DNA. The
fractions have been placed on filters and
hybridized with radioactive cRNA tran-
scribed from 5S DNA and rDNA, and the
hybridization patterns are indicated. In
Fig. 3B a similar gradient containing 50
fig of actinomycin D is shown. The satel-
lite DNAs are pelleted at the bottom of
the tube. The ribosomal DNA is found in
a very low density region of the gradient.
Unexpectedly the 5S DNA is also less
dense than the main-band DNA. Figure
4 shows the results from a series of exper-
iments in which 400 fig of X. mulleri
DNA in 20-ml gradients are banded in
the presence of increasing amounts of
ro
E
u
\
(/)
c
O
50 IOO
Actinomycin jsq
Fig. 4. Effect of actinomycin D concentration on buoyant density of X. mulleri DNA compo-
nents. About 400 fig of A", mulleri DNA in 20 ml gradients was centrifuged for 2 days at 35.000
rev/min in a 50.1 fixed angle rotor. The abscissa indicates the amount of actinomycin D per 100
fig of DNA. Solid circles, satellite I; open squares, satellite II; open circles with solid line, main-
band DNA; open circles with dashed lines, 5S DNA; open triangles, rDNA.
18
CARNEGIE INSTITUTION
actinomycin D. The abscissa indicates
the amount of actinomycin D per 100 /xg
of DNA in each gradient. The buoyant
densities of rDNA, 5S DNA, main-band
DNA, and the two satellite DNAs are
plotted at different actinomycin D levels.
No additional separation of components
is obtained with concentrations of actino-
mycin above 50 /xg. The large buoyant
density differences show the utility of this
technique.
With separation of main-band DNA
from the ribosomal and 5S RNA genes of
X. laevis and X. mulleri as an assay sys-
tem, other DNA-binding antibiotics
studied originally in density gradients by
Kersten et at. {Biochemistry 5, 236, 1966)
were evaluated. Nogalamycin (a gift of
Dr. Paul F. Wiley, The Upjohn Co.),
daunomycin, chromomycin, olivomycin,
mithramycin (a gift of Dr.T. J. McBride,
Chas. Pfizer and Co.), and cinerubin (a
gift of Dr. H. Zahner, Institut fur Mikro-
biologie, Tubingen) all proved inferior to
actinomycin D.
Studies on Silk Fibroin mRNA
and Its Genes
P. Lizardi
As a continuation of our work on silk
fibroin mRNA, we have developed meth-
ods for studying the structure and ex-
pression of the genes which code for this
mRNA in Bombyx mori.
As mentioned in last year's Report
{Year Book 71, pp. 21—22) , we have syn-
thesized a radioactive reverse tran-
scriptase product of fibroin mRNA
("anti-messenger" DNA) . Experiments
were designed to test whether this a-
mDNA could be used as a specific probe
for detection of fibroin genes. We took
advantage of the fact that fibroin genes
have a characteristic buoyant density,
different from that of main-band DNA,
when fractionated in gradients of cesium
salts (Suzuki, Gage, and Brown, J. Mol.
Biol., 70: 637). Using three types of
equilibrium gradients (neutral CsCl,
silver-Cs2S04 and actinomycin-CsCl) , we
have fractionated sheared Bombyx DNA
and have localized the position of indi-
vidual genes by molecular hybridization
assays of each gradient fraction. Profiles
were obtained using ribosomal RNA,
fibroin mRNA, and a-mDNA as hybrid-
ization probes. In each of the three types
of fractionation, the a-mDNA hybridiza-
tion exhibited the same unique distribu-
tion as the mRNA hybridization. There
was little overlap with ribosomal genes
or with main-band DNA.
These results demonstrate the specific-
ity of a-mDNA and imply that it can be
used as a tool for the detection of fibroin
messenger nucleotide sequences, whether
these are in the form of an RNA or a
DNA molecule. Furthermore, the assay
is highly sensitive, being able to detect
c
E
T3
a:
I
o_
ro
o
I
-300
-200
100
Q
E
i
X
ro
10 15
Fraction No.
Fig. 5. Hybridization of anti-messenger DNA
(a-mDNA) with B. mori DNA which had been
fractionated in an actinomycin-CsCl gradient.
The gradient contained 90 fig of sheared carcass
DNA and was centrifuged to equilibrium in a
50.1 Spinco rotor. The immobilized DNA frac-
tions were hybridized together with 3H-a-mDNA
(solid circles) and 32P-ribosomal cRNA from X.
laevis (open circles) ; absorbance at 260 nm
(solid line).
DEPARTMENT OF EMBRYOLOGY 19
fibroin genes in the DNA from a single scribed previously but with DNase and
silk gland. pronase treatment sometimes omitted. All
A technique which proved of consider- preparations used were centrifuged
able value in the aforementioned experi- through at least one sucrose gradient fol-
ments was the fractionation of Bombyx lowed by one formamide-sucrose gradi-
DNA in actinomycin-CsCl gradients, ent. Between 20 and 40% of the fibroin
Figure 5 shows a typical gradient in mRNA was retained by oligo-dT cellulose
which the position of ribosomal and fi- or polyU Sepharose, which specifically
broin genes has been determined by mo- bind polyA regions of messenger RNA.
lecular hybridization. The remarkable The material bound to the oligo-dT cellu-
separation between main-band DNA and lose and that which is not retained both
fibroin genes shows that the technique is have the base composition reported for
potentially useful for studies on the co- fibroin mRNA. The "fingerprints" of a
valent length of fibroin DNA, and ulti- two-dimensional separation of the oligo-
mately, for the isolation of these genes, nucleotides from a Ti digest of the bound
We are presently using the actinomycin- and nonbound mRNA were identical, and
CsCl technique in a study of buoyant the oligonucleotides obtained were those
density changes of fibroin genes as a predicted from the known Ti digest pro-
function of DNA size. Another current file ( Year Book 70, pp. 35-39) .
project involves the use of poly lysine After digestion with Ti plus pancreatic
precipitation and actinomycin-CsCl gra- RNase at high salt, 1.4% of the bound
dients for the isolation of fibroin genes, mRNA and 0.3% of the unbound mRNA
using a-mDNA hybridization as a puri- could then bind to oligo-dT cellulose,
fication assay. This retained fragment consisted of 82%
Other present projects involve struc- A residues. The fragment migrated on
tural studies of fibroin mRNA using elec- polyacrylamide gel at a size of approxi-
tron microscopy, and the development of mately 6-7S, or 200 nucleotides in length,
new methods for fibroin mRNA isolation The proportion of fibroin mRNA which
using oligonucleotide-cellulose affinity bound to polyU-Sepharose rose from 30%
columns. to 40% when the initial salt concentra-
tion was increased from 0.1 to 0.3 M.
A Polyadenylic Acid Sequence in Messenger RNA labeled for 6 hours and
Silk Fibroin mRNA 72 hours in vivo had the same proportion
of molecules which bound to the column.
The inability to detect polyA in fibroin
In the initial studies on the isolation of mRNA resulted from the very large size
silk fibroin messenger RNA from Bombyx 0f the mRNA and the consequent diffi-
mori, the oligonucleotides were analyzed culty in finding the relatively small pro-
after digestion with Ti ribonuclease and portion of tfte sequence which is polyA.
no significant amount of material larger R ig unclear why Qnl t of the mRNA
than 10 residues in length was found , , ■, , ■, , .,
t\7 r> 7 n,n or o^x mi r ., (whose purity can be assessed bv its
{Year Book 70, pp. 35-39). Therefore, it ' . , J , . ... *
„„„„ 4. i A i i A . fingerprint and base composition at
was suggested that a polyA sequence is , ,, «„,* s , .
not present. In view of the recent interest Sreater than 80%) contams Pol^A se"
in polyA sequences, their frequent pres- Quences. The slze of the mRNA whlch
ence in animal mRNAs, and the priming stlcks and that whlch remains unbound
capacity of silk fibroin mRNA for reverse to dT-cellulose are similar on a sucrose
transcriptase with an oligoT primer, this gradient. No functional correlation has
question was reexamined. been found between those mRNA mole-
Fibroin mRNA was isolated from pos- cules which contain polyA and those
terior silk glands of Bombyx mori as de- which do not.
R. Williamson
20
CARNEGIE INSTITUTION
THE REGULATION OF FIBROIN GENES
Quantitative Measurements of
Fibroin mRNA Synthesis
Y . Suzuki
Assisted by E. Suzuki and N. Tsuchida
During the past year, we have studied
the following questions: (1) Are the fi-
broin genes expressed not only in the
fifth instar but also in earlier larval
stages? (2) Is the induction of extensive
fibroin synthesis due to increased fibroin
mRNA content during the fifth instar or
due to translational controls of a large
preexisting population of fibroin mRNA
which had accumulated before the fifth
instar? (3) Is there repeated switching
on and off of fibroin genes during larval
life?
These questions have been answered by
quantitative measurements of fibroin
mRNA at varying stages. Bio-Gel A-50,
originally introduced by Lizardi and
6 -
Brown, efficiently separates 2-3 mg of
fibroin mRNA (Fig. 6). The mRNA
fraction thus obtained is generally about
80% pure without further purification. A
further sucrose gradient centrifugation
step gives an mRNA fraction which is
greater than 90% pure.
During the feeding stage of the fourth
instar both 32P-labeled and nonlabeled
RNAs were extracted from the posterior
silk glands of several hundred larvae.
About 1 mg of fibroin mRNA was iso-
lated easily by the two-step purification
described above. The mRNA was di-
gested by Ti RNase and the digest was
fractionated on DEAE-Sephadex A25
column (Fig. 7). The optical density
and radioactivity profiles are character-
istic for fibroin mRNA (see Year Book
' 70, p. 35) . From this experiment we con-
clude that the mRNA can be measured
u -
o
<
2 -
0
- 2
'o
X
c
E
1 ^
o
0
500
1000
1500 ml
Fig. 6. Large-scale preparation of fibroin mRNA on a Bio-Gel A-50m column. During the feed-
ing stage of the fourth instar both 32P-labeled RNA and nonlabeled RNA were extracted from the
posterior silk glands of several hundred larvae. About 65 mg of the RNA was applied to a Bio-Gel
column (5 X 80 cm) and was fractionated by an elution buffer containing Tris (20 ml, pH 7.5),
NaCl (0.15 M), EDTA (1 ml), SDS (0.5%). Solid circles, 32P cts/min; solid line, A260.
DEPARTMENT OF EMBRYOLOGY
21
0.3 -
0.2 -
S
CM
<
0.1 -
0
- 400
c
E
200 ^
o
0
40 80
Fraction number
Fig. 7. DEAE-Sephadex fractionation of the oligonucleotides produced by Ti RNase digestion
of fibroin mRNA from the fourth feeding stage. The mRNA fraction was obtained by the Bio-Gel
column shown in Fig. 6, and was further purified by a sucrose density gradient centrifugation. About
520 ^g of the pure mRNA was digested with Ti RNase, and the oligonucleotides produced were
fractionated on DEAE-Sephadex column A25. Solid circles, 32P cts/min; open circles, A26o.
quantitatively and comprises about 2%
of cellular RNA by weight. Therefore,
mRNA comprises about the same fraction
of cellular RNA in the fourth instar
larvae as it does in the fifth instar (Table
1). This result also answers the second
question. The mRNA present in the
fourth instar comprises only 2% (0.1 mg)
of the maximum amount which accumu-
lates in the fifth instar (5 mg) (Table 1) .
These results led us to the third ques-
tion: Are the fibroin genes switched on
once in the beginning of larval life, and is
the transcription going on continuously
throughout the entire larval life? Or, is
there repeated switching on and off of
fibroin genes during larval life? In order
to answer the question, studies were car-
ried out on RNA preparations from the
fourth moulting stage. Using the tech-
nique described above, an RNA fraction
which should contain fibroin mRNA was
obtained. Its GC content was 42% DNA-
like in base composition, and its oligo-
nucleotide profile after Ti RNase diges-
tion, both by radioactivity and optical
density, was quite distinct from that of
mRNA. The results suggest two alterna-
tive possibilities. (1) Both the mRNA
which accumulated in the fourth instar
and the newly synthesized mRNA were
mostly degraded. In this case "degrada-
tion" would be selective for fibroin
mRNA because synthesis and accumula-
tion of rRNA, sRNA and DNA-like RNA
are detectable at this time. (2) The tran-
scription of the fibroin genes is abolished
and the preexisting mRNA is degraded
due to its natural half-life in the gland.
At present, we cannot decide between
these possibilities. If the fibroin genes
are repeatedly switched on and off, this
22
CARNEGIE INSTITUTION
TABLE 1. Fibroin mRNA Content during the Fourth and Fifth
Larval Instars of Bombyx mori
Amount of RNA in
a Pair of Posterior
Silk Glands
Fibroin
% of Fibroin mRNA
Bulk
mRNA,*
Radio-
Stage
RNA, mg
Mg
by Weight
activity
4th instar (feeding
stage)
0.1
2
2
6
Middle of 4th
moulting stage
0.15
<0.8
<0.5
<1
5th instar (early
stage)
1
10
1
1
5th instar (middle)
4
80
2
10
5th instar (spinning
stage)
5
150
3
25
KA11 the data for fibroin mRNA have been corrected for its purity.
adds another dimension to the control
mechanism.
In the fifth instar synthesis and ac-
cumulation of fibroin mRNA begin again
as early as the second day and continue
as late as the spinning period. At the be-
ginning of the fifth instar the mRNA
comprises about 1% of the cellular RNA
by weight and 1% by radioactivity
(Table 1). These values increase grad-
ually up to 3% (by weight) and 30% (by
radioactivity) of cellular RNA toward
the end of larval life (Table 1). A new
question is raised if all of the fibroin
mRNA is functional. The results also in-
dicate that the transcriptional and/or
posttranscriptional control for fibroin
mRNA is quite different from that of
rRNA.
Mutants for Fibroin Production
Y . Suzuki and H. Chikushi
Several spontaneous mutants of B.
mori are affected in silk production.
Larvae of these mutants metamorphose
into pupae without cocoons or with un-
usual cocoons composed of only sericin
(Nd, naked pupa mutants). Other mu-
tants spin a very thin cocoon containing
both fibroin and sericin and are desig-
nated flc (flimsy cocoon) .
We have analyzed the Nd mutants.
The mutant gene is dominant over its
normal allele. Thin silk glands have the
same number of cells as normal glands,
and the anterior and middle portions of
the mutant silk glands look almost iden-
tical to those from normal animals. The
posterior gland of the Nd animal, how-
ever, is quite different from that of the
normal animal. It is markedly inhibited
in its growth, is very small in its size
especially in the fifth instar, and does
not contain appreciable amounts of
fibroin in its lumen. The mutant cocoon
(or often just fibers that do not form the
typical cocoon shape) weighs a quarter
of the normal and consists mainly of
sericin with only 0.6% of the cocoon as
fibroin-like material. The sericins from
the normal and the mutant are indis-
tinguishable and the mutant sericin is
suitable for sericine chemistry. Genetic
and biochemical experiments support the
idea that the Nd gene acts only on the
posterior silk gland and not on the an-
terior or middle silk gland. In the fifth
instar the posterior gland is partly de-
generated, its wet weight and nucleic
acid contents being about one-fifth of
the normal. In the cells some usually
large blocks of fibroin-like material are
DEPARTMENT OF EMBRYOLOGY
23
deposited. The deposition might be due
to disturbed secretion of the fibroin-like
material from cell to lumen, which in
turn might cause degeneration of the cell.
Recently, we analyzed 32P-labeled
RNA from this mutant. The RNA
looked intact in spite of the slight de-
generation of the posterior silk gland and
contained about 1% of its total radioac-
tivity as fibroin mRNA. Again the
mRNA was identified by the typical
oligonucleotide profile after RNase Tl
digestion. We conclude that the gene
Nd is not a mutant affecting fibroin gene
transcription, but rather a mutant af-
fecting some translational or posttrans-
lational control mechanism.
Fibroin mRNA from
Bombyx mandarina
Y. Suzuki and E. Suzuki
A wild silkworm, Bombyx man-
darina, which inhabits China, Japan,
and Korea is presumed to be the ancestor
of Bombyx mori. According to Chinese
literature Bombyx mori was domesti-
cated in China for the purpose of silk
production more than 4000 years ago.
If this presumption is correct, then B.
mori and B. mandarina should have in-
distinguishable fibroin genes. The argu-
ment stems from the following fact.
Fibroins are secreted and have an "ex-
tracellular function" as cocoons, webs,
and various attachment threads. Once
produced they do not have contact with
intracellular components of the animal.
It is proposed that this lack of contact
enables animals to tolerate a variety of
these proteins. Actually, a surprising
variety of silk fibroin is produced by
various insects and arachnids, and fi-
broins are classified into several groups
by their extremely different amino acid
compositions and x-ray diffraction pat-
terns, in marked contrast to the con-
served homology of amino acid sequence
12 -
'28
c
I
0
20 40 60
Fraction number
Fig. 8. The oligonucleotide profile of Ti RNase digest of fibroin mRNA from Bombyx man-
darina. 32P-labeled fibroin mRNA was purified and digested with Ti RNase, and the digest was
fractionated on a DEAE-Sephadex A25 column.
24
CARNEGIE INSTITUTION
or nucleotide sequence of functionally
related proteins or RNAs in different
organisms.
The amino acid composition of fibroins
from B. mori and that of B. mandarina
are indistinguishable. Although the
amino acid sequence of the latter fibroin
has not been analyzed, x-ray diffraction
patterns of the two fibroins strongly sup-
port the idea that they have similar
sequences.
The fibroin mRNA from B. mandarina
has been isolated and partly sequenced.
The mRNAs from B. mori and B. man-
darina are indistinguishable in molecular
size under denaturing conditions. Their
nucleotide sequences are the same ex-
cept for some modifications in the third
nucleotides of some codons (compare
Fig. 8 with Year Book 70, p. 35). These
modifications should provide a good ex-
perimental system for the analysis of
functional adaptation of the tRNA pop-
ulation to fibroin synthesis in the pos-
terior silk gland. Mutations in the third
nucleotides of codons of fibroin mRNA
should have been accompanied by a
change in regulation of the tRNA popu-
lation.
THE MOLECULAR WEIGHT OF MOUSE GLOBIN
MESSENGER RNA
R. Williamson and P. Wellauer
The accurate determination of the
molecular weight of mouse globin mes-
senger RNA has been complicated by the
lack of standard reference RNA mole-
cules; ribosomal RNA may migrate dif-
ferently in polyacrylamide gels or su-
crose gradients because of its extensive
secondary structure. Since the electron
microscope visualization techniques de-
veloped by Norman Davidson and now
used in this Department utilize spread-
ing in formamide, which eliminates most
secondary structure, the molecular
weight of the globin mRNA can be
measured directly. Moreover, as in ribo-
somal RNA, any regions of resistant
secondary structure can be seen and
quantitated.
Current estimates of the molecular
weight of globin mRNA vary from 150,-
000 to 230,000. The coding region of the
mRNA has a calculated molecular weight
of 143,000 (a-chain) or 148,000 (0-chain).
There is a polyA sequence approximately
60 nucleotides in length (20,000) and a
noncoding but identified sequence 93 nu-
cleotides in length, at least in human a-
globin mRNA, as shown by study of the
abnormal hemoglobins Constant Spring
and Wayne (30,000). A more accurate
estimate of molecular weight should per-
mit an analysis of the number of nucleo-
tides at the 5'-end of the molecule (prior
to the initiation triplet) and between the
final termination triplet (that for mutated
hemoglobin Constant Spring) and the
polyA region.
Three different globin mRNA prepara-
tions, prepared either by gradient centri-
fugation of EDTA-dissociated polysomes
followed by isolation of the 14S mRNP
or by polyU-Sepharose affinity chroma-
tography, were spread from 80% form-
amide— 4M urea and the lengths measured
using E. coli 16S rRNA as a standard.
The molecular weight determined was
200,000 ± 50,000. This is in agreement
with the values obtained by gel electro-
phoresis but not those from gradient sedi-
mentation or end group analysis. It al-
lows a prediction that for human a-chain
mRNA there is a maximum of 60 nucleo-
tides both at the 5'-terminus and between
the termination triplet for hemoglobin
Constant Spring and the polyA region.
DEPARTMENT OF EMBRYOLOGY
25
CONTROL OF GENE TRANSCRIPTION IN
EUKARYOTES
R. H. Reeder and T. Honjo
with the assistance of R. Handler
A long-term goal of this laboratory
has been the establishment of an in vitro
system in which the control mechanisms
acting on animal cell genes could be re-
produced. The problem of gene control is
one of the central problems in develop-
mental biology and its solution is almost
certain to require reproduction of those
controls in vitro. The first chance for a
truly molecular attack on this problem
was provided by the isolation of a single
eukaryotic gene of known function, the
gene coding for ribosomal RNA from the
frog Xenopus laevis. In order to use this
isolated gene in studying transcription, it
was first necessary to learn more about
its structure and function in vivo and
then to develop assays that would accu-
rately measure the correctness of its
transcription in vitro. Therefore, con-
siderable time was expended in develop-
ing methods for measuring strand selec-
tion, spacer-versus-gene transcription,
and specific initiation. In the course of
setting up these assays, Escherichia coli
RNA polymerase was used because of its
ready availability to transcribe the ribo-
somal genes in vitro. It was shown that
this prokaryotic enzyme is unable to rec-
ognize the proper transcription initiation
signals in a eukaryotic ribosomal gene
{Year Book 68 and Year Book 69). The
bacterial enzyme does transcribe these
genes in a nonrandom manner, but the
pattern of transcription is not the same
as occurs in vivo. At about this time, the
first reliable procedure for isolating eu-
karyotic RNA polymerase was devel-
oped, making it possible to study in vitro
transcription of the ribosomal genes using
their own homologous polymerase. This
work was done in collaboration with R.
Roeder. At first it was not clear which
polymerase to use, however, since Xen-
opus has at least three distinct types of
nuclear RNA polymerases {Year Book
69). Studies on isolated nuclei using the
fungal inhibitor a-amanitin showed that
the most likely candidate was polymerase
I and showed furthermore that ribosomal
RNA was by far the major product of
that enzyme {Year Book 70). Methods
were developed during that study for
measuring rRNA synthesis in the pres-
ence of large amounts of non-rRNA syn-
thesis and to assess its fidelity of tran-
scription. A major disappointment came
when Xenopus polymerase I was used to
transcribe purified ribosomal genes. Poly-
merase I (and later polymerases II and
III) transcribed rDNA just as aber-
rantly, if not more so, as E. coli poly-
merase had done. The obvious implica-
tion was that some factor (s) essential for
specific transcription had been lost dur-
ing purification of the genes, the poly-
merase, or both. The first efforts were
directed toward the enzyme itself to see
if some subunit or cofactor had been lost
during purification. Enzymes were tested
at various stages of purification, crude
fractions were added back to purified en-
zymes, and enzymes from several differ-
ent sources (embryos, liver, cultured
cells, oocytes) were all tested. None
showed any improved fidelity of tran-
scription over the purified enzyme. At-
tention was then shifted to the possibility
that some essential substance had been
lost from the gene itself during purifica-
tion. T. Honjo began experiments in
which partially purified enzyme was
added to crude chromatin to see if this
could stimulate correct transcription of
the ribosomal genes. The results of this
investigation are presented in this year's
Report.
AVhile the experiments on ribosomal
gene transcription were in progress,
Brown isolated another gene, that coding
for 5S RNA of ribosomes. Assays have
also been developed for measuring the
26
CARNEGIE INSTITUTION
fidelity of its transcription under various
conditions.
Corollary to the experiments with Xen-
opus RNA polymerase and chromatin, a
second study has been under way for
about two years in which the ability of
E. coli RNA polymerase to transcribe
chromatin faithfully has been explored.
This work was undertaken largely be-
cause there is a considerable literature
stating that E. coli RNA polymerase
transcribes chromatin faithfully, i.e., it
copies the same genes in vitro as were
transcribed in the living cell from which
the chromatin was derived. Since we had
at hand techniques for examining this
question in much greater detail than had
been done before, we felt it was worth pur-
suing. The findings are also presented in
the present account.
During the year, another, more biolog-
ical, study of ribosomal gene control was
also undertaken by Hon jo — a biochem-
ical description of the dominance of
Xenopus laevis rRNA synthesis over that
of Xenopus mulleri rRNA in Fx hybrids
between these two species.
Transcription of rRNA and 5S RNA
Genes in Amphibian Chromatin by
Homologous RNA Polymerase
T. Honjo and R. H. Reeder
As described previously (Year Book
71, pp. 15-16), we have been testing the
possibility that the use of intact chroma-
tin and RNA polymerase purified from
the homologous animal would provide us
with accurate transcription. We were un-
able to get conclusive results earlier, since
exogenous RNA polymerase stimulated
RNA synthesis only 2- to 3-fold over
endogenous RNA synthesis.
We have overcome this difficulty by
concentrating RNA polymerase up to 40-
fold and isolating chromatin from liver
instead of tissue-culture cells. The chro-
matin preparation from the latter cells
TABLE 2. 5S RNA Synthesis by Chromatin and RNA Polymerases I and II
RNA Hybridized to 5S DNA2,
cts/min
L-strand
H-strand
Total RNA
Experimental System1
Synthesized,
-5S
+5S
-5S
+5S
cts/min
RNA
RNA
RNA
RNA
(A) Chromatin alone
68,000
8
0
0
7
Chromatin +
polymerase I
752,000
233
161
462
507
(B) Chromatin alone
53,776
16
3
5
2
Chromatin +
polymerase II
834,076
475
411
464
384
1. Each reaction mixture contained 50 ml Tris-Cl, pH 8.0, 5 ml MgCl2,
1 mikf MnCl2, 50 ml (NH4)2S04, 0.63 ml each of ATP, UTP, and GTP
50 ^Ci of CTP-32P (specific activity 20 Ci/mmole), chromatin (47 fig of DNA
for exp. A and 31 £ig of DNA for exp. B) and enzymes as indicated (54 units
of polymerase I and 40 units of polymerase II). The reaction mixtures were
incubated at 30° for 60 minutes.
2. Hybridization was performed in 4X SET and 50% formamide at 40°
for 18 hours using 0.1 /".g each of strand-separated 5S RNA bound to a Milli-
pore filter in the presence or absence of 30 £ig of cold 5S RNA. The efficiency
of hybridization was monitored by the addition of (3H)-RNA transcribed from
either H or L strand of 5S DNA by E. coli RNA polymerase. The efficiency
ranged from 14% to 19% for L strand and from 15% to 18% for H strand.
DEPARTMENT OF EMBRYOLOGY
27
J
C
E
o
Q_
M
ro
200
150
100
50-
0
00
50
0
Control
in vivo r RNA
in vitro RNA
V"°--\
*>—A ^...O
Competition
3000
2000
1000 —
c
E
(/>
o
-1000
500
0
Fraction Number
Fig. 9. Fidelity of ribosomal gene transcription by RNA polymerase I. 32P-labeled RNA syn-
thesized by polymerase I with chromatin as template was mixed with 3H-labeled rRNA synthesized
in vivo. The RNA mixture was hybridized to strand-separated rDNA in the presence and absence
of a 100-fold excess of unlabeled rRNA.
has a very high endogenous activity for
RNA synthesis and does not respond well
to exogenous RNA polymerase. As shown
in Table 2, using liver chromatin both
polymerases I and II stimulate total
RNA synthesis at least 10-fold above
endogenous RNA synthesis. The RNA
from each reaction was purified and hy-
bridized to the L and H strands of 5S
DNA in the presence and absence of an
excess amount of cold 5S RNA. Chroma-
tin alone synthesized only a negligible
amount of 5S RNA, while polymerase I
transcribed 5S genes in chromatin in sig-
nificant quantity. However, the products
of 5S genes transcribed by polymerase I
hybridized to both L and H strands.
Since 5S RNA synthesized in vivo is
complementary only to the L-strand, it
is clear that transcription in vitro by
polymerase I is aberrant. Spacer se-
quences were also transcribed by poly-
merase I because only 30% of RNA
hybridized to L strand was competed by
native 5S RNA. Essentially similar ob-
servations were made with polymerase II.
Polymerase II seems not to be responsible
for 5S RNA synthesis in vivo but it tran-
scribes 5S genes in chromatin in vitro. A
similar experiment measuring the fidelity
of transcription of the genes for 18S and
28S rRNA is shown in Fig. 9. In this case
28
CARNEGIE INSTITUTION
also, RNA made by polymerases I and II
hybridized to both strands and to both
gene and spacer sequences.
From the results just described, we can
conclude that (1) both RNA polymerases
I and II can initiate RNA chains in vitro
using chromatin as template; (2) RNA
transcripts from this system contain
RNA sequences not found in living cells.
For both 5S and rRNA genes, therefore,
eukaryote RNA polymerases transcribe
chromatin no more accurately than does
E. coli RNA polymerase.
Transcription of Chromatin by
Bacterial RNA Polymerase
R. H. Reeder
There are a number of experiments re-
ported in the literature stating that bac-
terial RNA polymerase can only tran-
scribe a restricted set of DNA sequences
when given eukaryotic chromatin as a
template. Furthermore, the sequences it
does transcribe are believed to be very
similar, if not identical, to those tran-
scribed in the living cell from which the
chromatin originated. These conclusions
are largely based on the results of RNA-
DNA hybridization experiments that de-
tected only sequences which are highly
repetitious in the DNA.
Our own studies of the transcription of
specific sequences in chromatin seemed to
lead to the opposite conclusion, i.e., that
bacterial RNA polymerase transcribes
eukaryotic genes nonspecifically and the
presence or absence of the chromatin pro-
teins makes little difference. Therefore,
additional experiments have been done to
explore this matter further.
We began with chromatin from mouse
liver, since fairly extensive studies have
already been reported by others using
this tissue. In the first experiment un-
labeled nuclear RNA was tested for its
ability to compete with (1) radioactive
RNA transcribed from chromatin by E.
coli RNA polymerase, (2) RNA tran-
scribed from deproteinized bulk DNA,
and (3) RNA synthesized by the endoge-
nous RNA polymerase in nuclei. The re-
sults are shown in Fig. 10.
In this experiment RNA from chroma-
tin and endogenously labeled RNA from
nuclei were both competed by unlabeled
nuclear RNA in an identical manner. To
this extent, the results agree with previ-
100 150
Unlabeled nuclear RNA (jd.g)
Fig. 10. Hybridization of various RNAs to
mouse DNA and competition with unlabeled
mouse liver nuclear RNA. In two separate ex-
periments a different 3H-labeled RNA was
mixed with 32P-labeled RNA transcribed from
deproteinized mouse bulk DNA, hybridized to
filters containing unfractionated mouse DNA
(15 ^g), and competed with unlabeled mouse
liver nuclear RNA. Since the (32P)-RNA com-
petition curves were superimposable, both ex-
periments have been combined in one graph.
Mouse liver chromatin transcript (25,000 3H
cts/min added, 100% represents 246 cts/min
bound), solid circles. Deproteinized mouse DNA
transcript (61,200 32P cts/min added, 100% rep-
resents 1,100 cts/min bound), open circles. En-
dogenous nuclear RNA (44,000 3H cts/min
added, 100% represents 492 cts/min bound),
solid squares. In a separate experiment mouse
satellite L-strand transcript (18,000 3H cts/min)
was hybridized to niters containing mouse satel-
lite L-strand DNA (3 fig,) and competed with
unlabeled mouse liver nuclear RNA (100% rep-
resents 572 cts/min bound) ; shown by solid
triangles. All points are the average of two
duplicate filters.
DEPARTMENT OF EMBRYOLOGY
29
ous observations. The unexpected result
is that the bulk DNA transcript is also
strongly competed by unlabeled nuclear
RNA. This strong competition disagrees
with previous reports and, if true, greatly
reduces the sensitivity of the assay to
detect differences between RNA and
chromatin transcripts in vivo.
A trivial explanation for the strong
competition of bulk DNA transcripts
would be the presence in the unlabeled
nuclear RNA of some contaminant which
nonspecifically prevented hybrid forma-
tion. To test this possibility 32P-
labeled RNA was transcribed from the
L-strand of purified mouse satellite DNA
and hybridized to mouse satellite L-
strand DNA on filters. Addition of the
same amount of unlabeled nuclear RNA
as used previously caused no reduction in
hybrid formation (Fig. 10). This shows
that the competition between nuclear
RNA and bulk DNA transcript is due to
base sequence homology and not some
nonspecific artifact. This experiment is
additional evidence that, as reported by
other workers, mouse satellite L-strand
is not transcribed in liver nuclei or, at
least, the transcript is not accumulated.
The experiment in Fig. 10 has been re-
peated using the hybridization conditions
described by Smith et al. (1969) with and
without RNase. Under all conditions
tested, nuclear RNA competed more than
50% of the hybridization of bulk DNA
transcripts. We conclude that this assay
procedure is a very insensitive method for
demonstrating that chromatin is a re-
stricted template for bacterial RNA
polymerase.
Mouse satellite sequences comprise
about 10% of mouse DNA and are
known to reassociate very rapidly. There-
fore, it seemed possible that the entire
difference between chromatin and DNA
transcripts in Fig. 10 could be explained
by a failure of E. coli RNA polymerase
to transcribe satellite sequences in chro-
matin. As a test of this possibility, mouse
DNA free of satellite sequences was im-
mobilized on filters and hybridized to a
mixture of 3H-labeled chromatin tran-
script and 32P-labeled deproteinized
DNA transcript exactly as done for the
experiment in Fig. 10. The hybridization
was competed by increasing amounts of
unlabeled liver nuclear RNA. The results
(Fig. 11) show that after removal of
satellite sequences from the filter-bound
DNA (but not from the DNA used for
template) chromatin and deproteinized
DNA transcripts behave identically in
this type of test:
The ability of E. coli RNA polymerase
to transcribe mouse satellite sequences in
chromatin was examined more directly in
the following experiment. 3H-labeled
RNA was transcribed from liver chroma-
100 <
y
remaining (% )
O* 00
o o
i i
O^
BulkDNAcRNA
^ 1
i, 40-
Chromatin ""^
CD
£ 20-
D
i
n
u
1
l l
0 50 100 150
Unlabeled nuclear RNA (jj.g )
Fig. 11. Hybridization of mouse liver chroma-
tin and deproteinized DNA transcripts to satel-
lite-free mouse DNA. Mouse liver chromatin
transcript (100,000 32P cts/min) was mixed with
deproteinized DNA transcript (487,000 3H cts/
min), hybridized to mouse DNA from which
the satellite has been removed (6 //.g/ filter), and
competed with unlabeled mouse liver nuclear
RNA. Chromatin transcript (100% represents
1575 cts/min bound), solid circles. Deprotein-
ized DNA transcript (100c/c represents 14,445
cts/min bound), open circles. All points are the
average of two duplicate filters.
30
CARNEGIE INSTITUTION
TABLE 3. Measurement of Mouse Satellite L-strand Transcription
Template for (3H) RNA
Synthesis
Fraction of (3H) cts/min % of (3H)
(3H) cts/min (3H) cts/min (32P) cts/min in L-strand in L-strand
Added Hybridized1 Hybridized2 Sequences3 Sequences
Endogenous nuclear DNA
224,000
8
.082
98
.04
Deproteinized DNA
68,000
81
.018
4,500
6.6
Chromatin
polymerase4 (units)
42
260,000
47
.039
1,200
0.5
210
1,240,000
197
.028
7,000
0.6
420
1,332,000
358
.037
9,700
0.7
840
1,608,000
381
.032
11,900
0.7
1 Each reaction contained a filter with 3 /ug of mouse satellite L-strand DNA.
2 As an internal standard, 32P-labeled L-strand transcript (10,000-20,000 cts/min) was added to
each reaction.
3 Calculated by dividing (3H cts/min hybridized) by (fraction of 32P cts/min hybridized).
4 Each reaction contained mouse liver chromatin (57 ng of DNA) plus the indicated amount of
E. coli RNA polymerase. After 30 minutes at 30° the total RNA from each reaction was isolated
and its content of satellite sequences was measured.
tin and mixed with a small amount of
(32P)-RNA transcribed from purified
satellite L-strand. The mixture was hy-
bridized to a filter containing satellite L-
strandDNA. The amount of (32P)-RNA
hybridized measures the efficiency of
satellite sequence hybridization in this
particular reaction. Knowing this effi-
ciency and the amount of (3H) chroma-
tin RNA that hybridized, one can calcu-
late the fraction of the total (3H)
chromatin RNA consisting of satellite se-
quences. The results of such measure-
ments when made on several types of
RNA are shown in Table 3. RNA made
in isolated mouse liver nuclei by endoge-
nous polymerases contained no satellite
sequences. When E. coli polymerase was
used to transcribe deproteinized mouse
DNA, 6.6% of the CMP incorporated
was found in satellite sequences. Since
it is known that the mouse satellite L-
strand is about 5% of the total DNA and
has a CMP content of 14% versus 20%
for bulk DNA, one would expect 3.5% of
the CMP incorporated to be in L-strand
sequences if transcription were com-
pletely random. The fact that almost
twice this amount was incorporated sug-
gests that E. coli RNA polymerase tran-
scribes L-strand sequences preferentially
in bulk DNA.
Table 3 also shows the amount of satel-
lite transcription that occurred when E.
coli polymerase was used to transcribe
mouse liver chromatin at varying chro-
matin-to-polymerase ratios. At all poly-
merase concentrations, ranging from be-
low saturation to above saturation of the
chromatin, transcription of satellite se-
quences was uniformly low. The low
level of satellite transcription observed in
chromatin is significant and reproducible
but is about an order of magnitude lower
than that observed for deproteinized
DNA.
Previous workers were unable to detect
any differences between chromatin-
primed RNA and RNA made in vivo in
the same tissue. Therefore, they con-
cluded that the two types of RNA were
largely identical and the bacterial poly-
merase was transcribing chromatin in a
faithful manner. The experiments re-
ported here suggest that the major reason
for not finding differences earlier was the
insensitivity of the hybridization test em-
ployed.
There are several reasons for this lack
of sensitivity. (1) The reiterated DNA
sequences in eukaryotic genomes are
DEPARTMENT OF EMBRYOLOGY
31
composed of large families of related se-
quences, each family cross-hybridizing
within itself. Transcription of only one
member of a family in vivo is sufficient to
cause nuclear RNA to compete with the
hybridization of all members of that
family whether they are transcribed in
vivo or not. The finding that unlabeled
nuclear RNA competes effectively with
deproteinized DNA transcripts (Fig. 10)
suggests that in the mouse at least one
member of most reiterated sequence fam-
ilies is transcribed in vivo. (2) Hybridi-
zations done at low RNA concentrations
tend to be dominated by the most reiter-
ated and rapidly reassociating sequences,
such as the satellite in mouse DNA. Re-
moval of this one sequence caused mouse
chromatin transcript and deproteinized
DNA transcript to behave identically in
hybridization-competition (Fig. 11). It
is likely that previous hybridization ex-
periments were dominated similarly and
no conclusions can be drawn from them
concerning transcription of the rest of the
mouse genome. (3) The previous experi-
ments were also insensitive, since they
were not designed to distinguish wrong-
strand from correct-strand transcription.
It has been shown that unlabeled RNA
homologous to only one strand of the
DNA will compete efficiently for the hy-
bridization of labeled RNA homologous
to both stands. Presumably competition
on one strand occurs through the usual
isotope dilution and on the other through
formation of RNA-RNA hybrids. There-
fore, E. coli polymerase can (and prob-
ably does) transcribe both DNA strands
in chromatin, and the transcripts will still
be efficiently competed by unlabeled nu-
clear RNA.
A reasonable interpretation of these
results, which could also accommodate
earlier observations, is that transcrip-
tional control in eukaryotes occurs on at
least two levels. One level of control
would be a gross control in which certain
blocks of DNA would be complexed with
protein so as to make it inaccessible to
any RNA polymerase, bacterial or eu-
karyotic. It is possible that in many cells
the most highly reiterated sequences
(satellites) are subject to this type of
control and could thus account for the
previously observed similarity between
chromatin-primed and in vivo transcripts.
This possibility is suggested by the occur-
rence of many satellite DNAs in hetero-
chromatic regions of the chromosomes.
Hybridization of such satellite sequences
might be expected to contribute dispro-
portionately to hybridization experiments
that measure only reiterated sequences.
It is clear, however, that a second level of
fine transcriptional control must also
exist in vivo. The endogenous RNA poly-
merases are able to recognize initiation
and termination signals and to select the
correct DNA strand. The evidence shows
that E. coli RNA polymerase recognizes
none of these fine controls in eukaryotic
DNA whether or not chromatin proteins
are present. Thus, it transcribes spacer
as well as gene regions and incorrect as
well as correct DNA strands.
Repression of Xenopus mulleri rDNA in
X. laevis X X. mulleri Hybrid Frogs
T. Hon jo and R. H. Reeder
Wild-type frogs of the species Xenopus
laevis and Xenopus mulleri contain two
prominent nucleoli per cell throughout
most of embryogenesis. Blackler and co-
workers have recently shown, however,
that Fi hybrids between these two species
have only one nucleolus in the large
majority of their cells, suggesting that
the nucleolar organizer of one of the two
species is somehow inactivated by the
presence of the other. Since it is well
established that the nucleolus is the site
of rRNA synthesis in interphase cells, it
has been inferred that the rDNA of one
species is repressed in the hybrid cells. If
this inference is correct, such cells may
provide a useful system in which to ana-
lyze the control of rRNA synthesis in
eukaryotes.
We have undertaken a biochemical de-
scription of the phenomenon of nucleolar
32
CARNEGIE INSTITUTION
dominance in hybrids between laevis and
mulleri. These hybrids have an obvious
advantage for this type of study, since
the rDNA from each species has been
isolated in pure form and their structures
have been studied extensively. Both the
rDNA and the rRNA precursors of the
two species can be distinguished.
In describing the crosses and genotypes
in this report, we follow the convention of
giving the species of the female first. The
diploid genotypes of the rDNA of wild-
type laevis and mulleri are expressed by
l/l and m/m, respectively. For example,
l/l~ indicates laevis heterozygous for the
nucleolar mutation which deletes the
rDNA.
Brown and Blackler originally showed
that hybrid frogs contain rDNA from
both parents, as expected. Our results
have further confirmed this finding.
Knowing that both types of rDNA were
present, we developed a method of deter-
mining which rDNA was being tran-
scribed in hybrid frog nuclei. The rRNAs
of both species of Xenopus are initially
transcribed as large precursor molecules
with a size of about 2.5 X 106 daltons.
About 2.2 X 106 daltons of this molecule
are accounted for by sequences for 18S
and 28S rRNA which cannot be distin-
guished between laevis and mulleri. The
remainder of the precursor molecule
however, differs considerably in sequence
between the two species and hybridizes
about 10-fold more efficiently to homolo-
gous than to heterologous rDNA. To
determine which species of rRNA pre-
cursor is being synthesized in a hybrid
animal, authentic laevis 3H-labeled pre-
cursor was mixed with 32P-labeled pre-
cursor from the hybrid animal and hy-
bridized to purified laevis rDNA and to
purified mulleri rDNA in the presence of
excess unlabeled 18S and 28S rRNA.
From the isotope ratios binding to the
two types of rDNA, the ratio of laevis to
mulleri rRNA in the hybrid animal can
be determined.
To label precursor rRNA molecules to
a sufficiently high specific radioactivity,
nuclei were isolated either from the liver
of a single animal or from a group of
embryos, and the endogenous RNA poly-
merase was allowed to elongate rRNA
chains in vitro. This rRNA has been
shown to be transcribed from the H-
strand of the rDNA, to consist largely of
18S and 28S rRNA sequences, and to
contain additional sequences which are
not 18S or 28S. It thus contains nucleo-
tide sequences similar to those found in
authentic precursor rRNA molecules.
Using rRNA labeled in isolated nuclei,
we constructed a standard curve relating
the isotope ratios bound by the two
rDNAs to the proportion of laevis and
mulleri rRNA in a mixture. In separate
reaction mixtures nuclear RNAs from
embryos of laevis and mulleri were
labeled with 32P. The 32P-labeled RNAs
were then mixed in several proportions.
Nuclei from laevis tissue culture cells
were used to synthesize 3H-labeled nu-
clear RNA as a standard. To each sam-
ple of 32P-labeled RNAs was added an
equal amount of 3H-labeled standard
RNA and lOO^g of unlabeled 18S and 28S
rRNA. Each doubly labeled RNA mix-
ture was hybridized to a filter containing
mulleri rDNA and to a filter containing
laevis rDNA in the same vial.
The result of such an experiment is
that for each artificial mixture of 32P-
labeled mulleri and laevis RNA one ob-
tains a 3H/32P ratio bound to laevis
rDNA (L ratio) and a 3H/32P ratio
bound to mulleri rDNA (M ratio) . When
the (32P)-RNA is 100% laevis, the L
ratio should be the same as the M ratio
and therefore the ratio of the L ratio to
the M ratio (L/M ratio) should equal
unity. At the other extreme, when the
(32P)-RNA is 100% mulleri, the L ratio
is greater than the M ratio. We have
determined empirically that the L/M
ratio is close to 158 at this extreme. A
standard curve obtained by the above
method is shown in Fig. 12.
The sensitivity of the method was
checked by making measurements on dif-
ferent batches of pure laevis and pure
DEPARTMENT OF EMBRYOLOGY
33
o
o
Nf
160
i 1 1 r
0' 20 40 60 80 100
mulleri rRNA(%
10 20 30 40
mulleri rRNA (%)
Fig. 12. Standard curve for the measurement
of mulleri rRNA synthesis in hybrid frogs. Con-
struction of the curve is described in the text.
The points are from individual measurements
on artificial mixtures of laevis and mulleri nu-
clear RNA. Each symbol represents a separate
experiment.
mulleri nuclei. From the measured L/M
ratio and the use of Fig. 12, we calculated
the apparent percentage of mulleri rRNA
synthesized in each batch. The values
range from 0 to 3 for pure laevis and 96
to 100 for pure mulleri nuclei. The lower
limit of detection is therefore about 5%
mulleri rRNA in the presence of 95%
laevis rRNA.
Using this method, we first measured
rRNA synthesis in Fi hybrid embryos
from the cross laevis (I /I) X mulleri
(m/m). The results are shown in Table
4. From the earliest stage tested, stage
15, and thereafter until about stage 43,
the mulleri rDNA was repressed com-
pletely. At stage 43 and beyond there
was detectable mulleri rRNA synthesis in
the population of hybrid embryos.
We also measured rRNA synthesis in
adult tissues from several hybrid frogs
(Table 5) . In liver nuclei from three of
TABLE 4. rRNA Synthesized by laevis X
mulleri Hybrid Embryos
Stage
mulleri rRNA, %
15-17
28-30
37-39
43-46
48-49
0
0
1
8
10
100 Mg of 18S and 28S laevis rRNA and 3H-
labeled standard RNA (191,000 cts/min) syn-
thesized by laevis nuclei were added to (32P)-RNA
of each stage except stage 15-17 where 382,000
cts/min of (32H)-RNA was employed. The RNA
mixtures were hybridized to filters containing
approximately 1 /u-g each of laevis and mulleri
rDNA.
the animals mulleri rRNA synthesis was
either absent or very low. In the other
four animals between 17% and 50% of the
rRNA synthesized was mulleri. The
second type of cross we examined was the
cross laevis (l/l~) X mulleri (m/m).
This type of mating results in an embry-
onic population with a one to one mixture
of the two genotypes, l/m and I'/m. Since
the two genotypes cannot be distinguished
morphologically, mixed populations of
embryos were analyzed for rRNA syn-
TABLE 5. rRNA Synthesized by laevis X
mulleri Hvbrid Adults
mulleri
Crosses
Tissue
rRNA, %
1
liver
0
1
liver
4
2
liver
8
2
liver
17
2
liver
27
2
liver
28
3
liver
34
kidney
49
A separate animal was used for each analysis
except for cross no. 3, in which both liver and
kidney were taken from one animal. Cross no.
1 was l/l~ X m/m. Somatic DNA of the animals
used from cross no. 1 was shown to have the
genotype l/m. Crosses no. 2 and no. 3 were
I /I X m/m. RNA from hybrid nuclei was labeled
with 32P and standard laevis RNA was labeled
with 3H.
34
CARNEGIE INSTITUTION
thesis. It was assumed that any mulleri
rRNA synthesis would be derived from
the l~/m genotype, since the l/m geno-
type makes no mulleri rRNA prior to
stage 43 (Table 3) . The results are shown
in Fig. 13. At stage 13-14 there was al-
most no detectable mulleri rRNA synthe-
sis. After stage 13-14 mulleri rRNA syn-
thesis gradually increased relative to
laevis until stage 43 when mulleri rRNA
accounted for about 50% of the rRNA
synthesized in the total embryo popula-
tion.
In both of the crosses described so far,
the laevis rDNA was donated by the fe-
male parent, which also donated the cyto-
plasm of the zygote. We therefore tested
animals from the reverse cross, mulleri
(m/m) X laevis (l/l) to assess the influ-
50
40
i 30h
<
°= 20
10
0
-
/
/ — @—
l/l x m/m /
V-
/
/
/
/
-
□ ,' l/l x m/m
i i i i i
0
20 40 60 80 100
Time after fertilization (hr )
120
10 1519 26 32 35
_u l L
*■<
V
40 42 4445 46
_l I LJ I
*
^ t
Gastrula Neurula | Hatching Feeding
Heart beat
Nieuwkoop-Faber stages
Fig. 13. Synthesis of mulleri rRNA during
early development in Xenopus hybrids. Nuclei
were isolated from embryos at various stages
and analyzed for their ability to synthesize
rRNA as described in the text. Embryos from
the same mating have the same symbol. Open
symbols with dotted line, embryos from the
cross lll~ X m/m. Solid symbols with solid line,
embryos from the cross l/l X m/m. Data for
this cross were taken from Table 4.
ence of egg cytoplasm on rDNA expres-
sion. The results were essentially the
same regardless of which species was the
female parent.
In one experiment, we also analyzed
embryos from the cross mulleri (m/m) X
laevis (l/l~). As early as stage 15-20
these embryos were making a significant
amount of mulleri rRNA (30%) in con-
trast to the reverse cross shown in Fig.
13. The embryos from this cross are a
mixture of two genotypes, m/l and m/l".
Since at the same stage (16-19) the m/l
genotype embryos make no mulleri
rRNA, it is clear that the m/l' embryos
are engaged in active synthesis of mulleri
rRNA.
The results of these matings can be
summarized as follows. When the hybrid
receives both laevis and mulleri rDNA
(l/m and m/l) transcription of the laevis
rDNA is usually dominant and the mul-
leri rDNA is repressed. Repression of
mulleri rDNA occurs as early as rRNA
synthesis can be measured in embryos
(stage 15) . In most cases the repression
extends throughout embryogenesis and is
found in adult tissues as well. When
laevis rDNA is present, the maternal
cytoplasm has little effect on the severity
of the repression.
When the hybrid receives only mulleri
rDNA and laevis rDNA is deleted, two
different results can occur. (1) If the
zygote receives laevis ooplasm (hybrid
genotype l~./m)} mulleri rRNA synthesis
is transiently repressed and then is grad-
ually turned on. (2) If the zygote re-
ceives mulleri ooplasm (hybrid genotype
m/l~) , turn-on of mulleri rRNA synthesis
is not delayed.
These results suggest that mulleri
rRNA synthesis can be repressed in two
apparently different ways. One type of
repression is due to the presence of laevis
rDNA, is not maternally inherited, and
can be permanent. The second type of
repression is due to some substance in the
laevis cytoplasm, is maternally inherited,
and is transient.
Our biochemical measurements on hy-
DEPARTMENT OF EMBRYOLOGY
35
brid embryos are in agreement with the disappearance of one nucleolus in the hy-
cytological observations by Blackler and brid cells. Synthesis of both species of
co-workers. In general, repression of rRNA correlates with the appearance of
mulleri rDNA correlates well with the two nucleoli per cell.
THE TIME OF RIBOSOMAL GENE REPLICATION
DURING THE S PHASE OF SYNCHRONIZED
CHINESE HAMSTER CELLS
P. J . Stambrook, assisted by D. Somerville and B. Smith
The S phase, during which nuclear
DNA is replicated, occupies a discrete
interval of the cell cycle. Experiments
were initiated to examine the time of
replication of specific genes within the
S phase. The genes investigated in this
study were those which code for ribo-
somal 18S and 28S RNA, and were
selected for two reasons: (1) In eukary-
otic cells, ribosomal genes are repre-
sented by multiple copies and this facili-
tates their identification by RNA-DNA
hybridization. (2) Ribosomal genes from
Xenopus laevis, which have previously
been isolated and purified, can be used
as a template for the in vitro synthesis
of radioactive RNA with very high
specific activity that is complementary
to ribosomal genes (cRNA). Since re-
gions of the ribosomal genes which code
for 18S and 28S ribosomal RNA have
been highly conserved through evolution,
the X. laevis ribosomal cRNA can be
used as a very sensitive probe for detec-
IOOO
o
>-
>
I-
U
<
o
<
500
0.5
1.0
2.0 3.0
RIBOSOMAL RNA(/ig/ml)
Fig. 14. Competition of X. laevis ribosomal cRNA with Chinese hamster ribosomal RNA. Nitro-
cellulose filters containing 50 fig of Chinese hamster DNA were incubated with X. laevis 3H ribo-
somal cRNA (120,000 cpm/ml) in the presence of Chinese hamster ribosomal RNA.
36
CARNEGIE INSTITUTION
tion of ribosomal genes in heterologous
systems, such as the Chinese hamster
cells used in these experiments.
The specificity of hybridization be-
tween X. laevis ribosomal cRNA and
Chinese hamster DNA was established
by two criteria. Hybridization of the
cRNA was performed in the presence of
increasing amounts of unlabeled ribo-
somal RNA from Chinese hamster cells
which acted as an effective competitor
of the cRNA (Fig. 14). When the
cRNA was hybridized to Chinese ham-
ster DNA which had been fractionated
by CsCl equilibrium density centrifuga-
tion, the cRNA hybridized with DNA of
high G-C content, characteristic of ribo-
somal genes, in the dense region of the
gradient (Fig. 16).
100*
50 -
TIME (hrs.)
Fig. 15. Cell synchrony during the mitotic
cycle. Mitotic cells were collected by selective
detachment and seeded into a series of plastic
60-mm culture plates. The cells were labeled for
10 minutes with 3H thymidine (5 /Wml) at
intervals after synchronization and coated with
emulsion for autoradiography. Percentage of
labeled cells, open circles; mitotic index, dashed
line.
>*e«6
1000
500 _
>
I-
O
<
o
a
<
1000
500
10
15 20 25
FRACTION NUMBER
30
FRACTION NUMBER
Fig. 16. Hybridization of X. laevis ribosomal cRNA to BrdUra-substituted Chinese hamster
DNA from different intervals of the S phase. Synchronized cells were labeled with BrdUra (10~5
M ) during successive 2-hour intervals of the S phase. The DNA was extracted, banded in CsCl,
denatured, and immobilized on filters. Filters from all 4 gradients were combined in a single vial
and hybridized with X. laevis ribosomal cRNA (250,000 cpm/ml). Absorbance at 260 nm (open
circles with solid line); radioactivity (dashed line).
DEPARTMENT OF EMBRYOLOGY
37
A population of mitotic cells was col-
lected and their synchronous traverse
through the cell cycle was followed. The
degree of synchrony and the time of
DNA synthesis in each experiment was
monitored by pulse-labeling replicate
plates for 10 minutes with 3H-thymidine
at intervals after synchronization and
then determining by autoradiography
the percentage of cells incorporating 3H
thymidine at each interval (Fig. 15).
The percentage of cells in mitosis at the
start of each experiment, usually 95 to
98%, was always recorded, as was the
mitotic index at each time point after
synchronization. Cells were density
labeled with 5-bromodeoxyuridine
(BrdUrd) for 2-hour intervals extending
over the duration of the S phase so as to
separate DNA replicated within discrete
periods of the S phase from DNA which
has replicated or will replicate at an
earlier or later time. BrdUrd, which is
incorporated into DNA in place of thy-
midine, increases the density of newly
replicated DNA which appears as a
heavy peak after CsCl equilibrium den-
sity centrifugation and which is easily
TABLE 6. Replication of DNA Containing
Ribosomal Genes at Intervals during the
S Phase
100
i
\
50
\
i
i
i
300 J
Q
200 =>
CD
100 5
I
CO
O
6 8
TIME (hrs.
10
Percent Radio-
activity
Time
Hybridizing
Interval,
Percent DNA
to Newly
hr
Replicated
Replicated DNA
3-5
23.1
43.5
5-7
45.0
36.5
7-9
21.0
13.6
9-11
8.4
6.3
Synchronized cells were density labeled with
BrdUrd at 2-hour intervals during the S phase.
The DNA was extracted, banded in CsCl,
immobilized on filters and hybridized with X .
laevis ribosomal cRNA. The percentage of the
DNA replicated at each interval was calculated
from the O.D. profiles in Fig. 16, and the per-
centage of the radioactivity associated with
newly replicated DNA was determined from
the same figure.
Fig. 17. Replication of ribosomal genes dur-
ing the S phase. The relative specific activities of
newly replicated DNA at different times during
the S phase after hybridization with X. laevis
3H ribosomal cRNA. Percentage of labeled cells
(open circles).
separable from bulk DNA. DNA ex-
tracted from cells labeled with BrdUrd
from 3-5 hours, 5—7 hours, 7—9 hours
and 9-11 hours was subjected to iso-
pycnic centrifugation in CsCl. The gradi-
ent was fractionated and the DNA in
each fraction was denatured, trapped on
nitrocellulose filters, and hybridized with
A", laevis ribosomal cRNA.
The relative amount of DNA which is
replicated during each time interval
varies (Fig. 16) and is presented in Table
6. The proportion of ribosomal genes
which replicate during each interval is
reflected by the fraction of the radioac-
tivity which hybridizes to DNA in the
heavy peak (Fig. 16) . By the time repli-
cation of the first 23% of the DNA has
occurred (3-5 hours), almost half of
the ribosomal genes have replicated (Fig.
16 and Table 6). Most of the remaining
ribosomal DNA is replicated during the
next time interval (5-7 hours), indicat-
ing that the multiple ribosomal gene
copies (between 200 and 300) replicate
during a restricted period of the S phase.
The results are summarized in Fig. 17.
The width of each bar represents the time
interval during which cells were labeled
with BrdUrd, and their relative height
indicates the proportion of newly repli-
cated DNA in ribosomal genes.
38
CARNEGIE INSTITUTION
SECONDARY STRUCTURE MAPS OF RNA
MOLECULES DERIVED BY ELECTRON
MICROSCOPY AND THEIR USE FOR THE
ANALYSIS OF THE PROCESSING OF
HELA RIBOSOMAL RNA
P. K. Wellauer and I. B. Dawid
We observed that 28S rRNA from
HeLa cells showed a reproducible sec-
ondary structure of hairpin loops when
spread for electron microscopy from 80%
formamide and 4 M urea. This second-
ary structure allowed the mapping of
these molecules in the sense that different
regions of each molecule and their polar-
ity could be recognized. Therefore, we
used such secondary structure maps to
investigate the processing pathway of
HeLa cell rRNA. Many aspects of this
processing pathway have been character-
ized previously by the work of Darnell,
Penman, Weinberg, Maden, and others.
Figure 18a shows an electron micro-
graph of the 45S rRNA precursor, which
is known to be the primary transcrip-
tion product of the ribosomal genes in
HeLa. A number of loops can be seen
along the molecule. Starting from the
left end (which is the 5' end, as we shall
show below) we see an extended region
leading to a characteristic double loop.
An extended region follows, then a series
of multiple loops, and a longer extended
region. The other end (3' end) of the
molecule is occupied by a large cluster
of loops. We compare to this 45S mole-
cule the 28S rRNA molecule shown in
Figure 18b. It starts with an extended
region, followed by the characteristic
double loop, another extended region,
and finally several small loops. This
series of structures can be found in the
45S RNA molecule in one region only,
at its 5' end, i.e., the left end as shown
in the figure. We conclude that the 28S
RNA must be derived from the 5'-ter-
minal region of the 45S RNA. The 18S
RNA molecule (Fig. 18c) shows no sec-
ondary structure when examined under
these conditions. There is only one ex-
tended region of sufficient length in 45S
RNA to accommodate the 18S molecule
Fig. 18. Electron micrographs of HeLa cell rRNA molecules, (a) 45S RNA; (b) 28S RNA; (c)
18S RNA.
DEPARTMENT OF EMBRYOLOGY
39
in the right half of the molecule as
shown.
The conclusions regarding the location
of the 28S and 18S regions in the 45S
RNA were supported by measuring and
plotting extended and looped regions in
40 to 60 separate molecules of each of
these RNAs. Secondary structure maps
also were prepared for a number of in-
termediates in the processing pathway.
Furthermore, the molecular weights of
precursor, intermediate, and mature
molecules were measured by comparing
their length to the length of E. coli
rRNA (Table 7).
To determine the 5'-to-3' polarity of
the rRNA molecules, we studied the
structure of 45S and 28S rRNA partially
digested with the 3'-exonuclease from
ascites nucleoli. This enzyme was gener-
ously supplied by Dr. Robert Perry.
TABLE 7. Molecular Weight of HeLa rRNA
Molecules as Determined by Electron
Microscopy
5'
RNA
Mol. Wt. ± S.D. X 10-6,
number of molecules
18S
0.661 ± 0.07 (166)
28S
1.757 ± 0.14 (100)
32S
2.266 ± 0.20 ( 95)
41S
3.301 ± 0.33 ( 87)
45S
4.714 d= 0.45 (100)
The molecular weights were determined from
length measurements of the RNA after treat-
ment with 10% formaldehyde (3% for 18S RNA)
at 63° for 15 minutes, followed by spreading
from formamide-urea. E. coli rRNA was used
as internal standard.
f
45S
kVVVVVV\V\VVVVKVVVVVVVV^5^
K^W^xVM
I
41S
24 S
tS^^^^f^^^^SSX^^^XVVH
Cx\\\VV\\VV3 (
I
3'
32S
20S
28S
18S
Fig. 19. Processing pathway of HeLa rRNA.
The molecules are drawn to scale. The 28S and
18S regions are hatched.
Secondary structure maps of the partially
digested molecules clearly showed loss
of material at one end, which was there-
by identified as the 3' end. These ex-
periments placed the 28S RNA region
at the 5' end of the 45S RNA molecule.
From a consideration of secondary
structure maps and the lengths (or molec-
ular weights) of these molecules an un-
ambiguous processing pathway was de-
rived (Fig. 19). This model is in good
agreement with the conclusions from
earlier work but adds considerable de-
tail.
The experiments described above dem-
onstrate the usefulness of secondary
structure mapping of RNA for the study
of the arrangement of sequences in
RNAs with a precursor-product relation-
ship. The method is also likely to be
useful for evolutionary studies: we have
analyzed the secondary structure in
Xenopus 28S rRNA and found it to be
very similar to that of HeLa 28S RNA,
even though the X en opus molecule is
about 15% smaller.
BIOGENESIS OF MITOCHONDRIA
/. B. Dawid, I. Horak, C '. Kaushagen, D. E. Leister, and P. K. Wellauer
Our studies on mitochondrial bio-
genesis (Year Book 71, p. 30) have been
continued in several directions. One
aspect is the study of hybrid somatic
cells in collaboration with H. G. Coon
at The National Cancer Institute, which
aims to develop an approach for the
application of genetic tools to the study
of mitochondrial biogenesis in animal
cells. The other aspect of our work in-
cludes several projects on the character-
ization of various macromolecular com-
ponents of mitochondria which are likely
to be involved in the biogenesis of these
particles. We have investigated the com-
position of mitochondrial ribosomes, the
40
CARNEGIE INSTITUTION
properties of mitochondrial RNA poly-
merase and the interaction of this enzyme
with mitochondrial DNA (mtDNA) , and
the properties of mitochondrial RNA
molecules and the arrangement and
evolution of the genes on mtDNA which
code for some of these RNAs.
Recombination of Mitochondrial
DNA in Somatic Hybrid Cells
/. Horak and I. B. Dawid, in collaboration with
H. G. Coon*
We reported in Year Book 71, p. 24,
that hybrid cells of human and mouse
* Laboratory of Cell Biology, The National
Cancer Institute.
or human and rat origin contained mito-
chondrial DNA (mtDNA) sequences
from both parental species. Since that
time, additional examples were studied,
confirming our conclusion that mtDNA
from two animal species can replicate
in the same cell for extended periods of
time.
One question concerning the state of
mtDNA in hybrid cells asks whether the
molecules derived from the two different
species coexist separately in the same
cytoplasm or whether they interact with
each other. One possible interaction
would be recombination, an event in
which parts of, say, human mtDNA
would attach covalently to mouse
3-
i
O
x
c
E
</>
o
2-
50 60
Fraction number
Fig. 20. Separation of human and mouse mtDNA by banding in a CsCl gradient. Open circles,
3H; solid circles (human cRNA) ; 32P (mouse cRNA). For explanation see text.
DEPARTMENT OF EMBRYOLOGY
41
500
c
*E
o
X*
10
iOOO
500-
30 40
Fraction number
Fig. 21. CsCl gradient centrifugation of the DNA from a hybrid cell strain. Symbols have the
same meaning as in Fig. 20.
mtDNA, forming a hybrid molecule. We
have tested this possibility in the follow-
ing way.
The density of human mtDNA is 1.707,
that of. mouse mtDNA 1.699, and rat
mtDNA 1.701 g/cm3 (assuming the den-
sity of E. coli DNA as 1.710). This
density difference can be exploited to
separate the various mtDNA in CsCl
gradients. Figure 20 shows a gradient in
which a mixture of human and mouse
mtDNAs was banded; the gradient was
fractionated and each fraction adsorbed
to a membrane filter. The entire series
of filters was then hybridized with a
mixture of human 3H-cRNA (that is,
(3H)-RNA transcribed in vitro with E.
coli RNA polymerase from human
mtDNA) and mouse 32P-cRNA. Each
cRNA hybridized with its own template
DNA and gave very little cross-hybridi-
zation with the other mtDNA. Gradi-
ents of this type were used to assay for
recombinant molecules of mtDNA in hy-
brid cells. Figure 21 shows such a gradi-
ent with DNA from a cell strain that
contained about 20% mouse mtDNA
(and 80% human mtDNA) . The human
3H-cRNA detected a broad peak of
human mtDNA sequences. The mouse
32P-cRNA hybridized with DNA mole-
cules which had banded at the same
density as the molecules detected by the
human cRNA; in addition, there were
DNA molecules containing mouse se-
quences which had banded at a lower
density. This result strongly suggests
that a physical link existed between at
least some of the mouse and human
mtDNA sequences. An interpretation of
this gradient implies that the denser
peak consists of molecules which are
predominantly human mtDNA and con-
tain pieces of mouse mtDNA. These
42
CARNEGIE INSTITUTION
molecules would band at or close to
human mtDNA density, but they con-
tain sequences which hybridize with
mouse cRNA. In addition, the dense peak
may also contain "clean" human mtDNA
molecules. The lower-density peak which
binds mouse cRNA could consist of
"clean" mouse mtDNA molecules or of
hybrid molecules containing more mouse
than human sequences. This question
cannot be decided. Human cRNA does
bind to DNA molecules which have
banded at the lower density, but it is
not clear whether this is due to "smear-
ing" of the large peak of human mtDNA
or to the presence of recombinant mole-
cules with a density close to that of
mouse mtDNA.
The experiment in Fig. 21 shows that
the DNA molecules containing mouse
sequences, and possibly those containing
human sequences, are heterogeneous in
the hybrid cell. Furthermore, the ex-
periment makes it very likely that mouse
mtDNA segments became covalently
linked to human mtDNA during growth
in the hybrid cell strain. However, the
latter conclusion is not explicitly proven:
Other interpretations of the observation
are possible, e.g., the presence of con-
catemers or of mouse mtDNA molecules
modified in such a way as to make their
density similar to that of the human
mtDNA. These alternate interpretations
are much less likely than recombination
of human and mouse mtDNAs.
Similar experiments were carried out
with several mouse-human and rat-
human hybrid cell strains. We have
found five examples of cell strains which
contained both parental mtDNAs with-
out detectable recombination and 13 cell
strains in which recombination had ap-
parently taken place. Therefore, recom-
bination of mtDNA in hybrid cells is a
frequent event. These results suggest
that recombination of mtDNA molecules
may occur frequently in the "normal"
cells of animals but would not be de-
tectable by biochemical analysis.
Protein Composition of Cytoplasmic
and Mitochondrial Ribosomes of
Xenopus laevis
D. E. Leister
In Year Book 71, p. 29, we reported the
fractionation of Xenopus laevis cyto-
plasmic ribosomal proteins by two-
dimensional (2D) polyacrylamide gel
electrophoresis. The large number of
spots (at least 60) detected on 2D gel
slabs indicated that greater resolution
would be obtained if ribosomal subunits,
rather than monosomes, were the source
of proteins. Such analyses have now been
carried out with both cytoplasmic and
mitochondrial ribosomal subunits.
Similar strategies were used to isolate
cytoplasmic and mitochondrial ribosomal
subunits. First, undissociated monosomes
were isolated. Then, the monosomes were
dissociated into their constituent sub-
units by exposure to buffers containing
high concentrations of monovalent salts;
subunits were subsequently separated by
centrifugation in sucrose gradients. Iso-
lated cytoplasmic large and small sub-
units were largely free of cross-contami-
nation and, when mixed, were as active
as undissociated monosomes in polyU-
directed polypeptide synthesis. Isolated
mitochondrial large and small subunits
were completely free of contamination by
cytoplasmic ribosomes (RNA extracted
from the mitochondrial particles con-
tained less than 2% cytoplasmic ribo-
somal RNA, as measured by hybridiza-
tion competition) . In addition, these
subunits were free of cross-contamination
and, when mixed, were as active as undis-
sociated mitochondrial monosomes in
polypeptide synthesis.
The composition and other properties
of the ribosomal subunits were studied by
physical and chemical means, in particu-
lar by the separation of the proteins by
two-dimensional gel electrophoresis.
Table 8 summarizes the results. The
mitochondrial rRNA molecules are much
smaller than the cytoplasmic rRNAs, but
DEPARTMENT OF EMBRYOLOGY
43
TABLE 8. Composition of Cytoplasmic and Mitochondrial Ribosomes
of X. laevis
Ribosomal Subunit
Mean Mol.
Number of Wt. of RNA, mol. Particle Wt.,
Proteins Proteins wt X 10~6 106 daltons
Cytoplasmic
large
37
22,000
1.5
2.6
small
34
21,000
0.7
1.4
Mitochondrial
large
41
27,000
0.58
1.8
small
43
33,000
0.32
1.6
the mitochondrial ribosomal subunits
contain a larger number and a larger
total mass of ribosomal proteins. Con-
sequently, the particle weight of the
mitochondrial ribosome is only slightly
lower than the weight of the cytoplasmic
ribosome. This conclusion is unexpected
in view of the low sedimentation coeffi-
cient of the mitochondrial ribosome
( Year Book 69, p. 576) . It appears on the
basis of the data in Table 1 that this low
sedimentation coefficient is related more
to the lower density of the protein-rich
mitrochondrial ribosomes than to their
slightly lower mass.
Figure 22 shows the molecular weight
histogram of mitochondrial ribosomal
proteins. Comparison to histograms of
X. laevis cytoplasmic and E. coli ribo-
somal proteins {Year Book 71, p. 30) re-
veals that the difference in the mean mo-
lecular weights of mitochondrial and
cytoplasmic proteins is not the result of a
bias introduced by a few large mitochon-
drial proteins. On the contrary, all three
patterns are distinct from each other.
CO
z
c54
cr
^ 2
o c _
\h
pnrfyyui.i
10
30
MW x 10
50
Fig. 22. Molecular weight distribution of
mitochondrial ribosomal proteins from X. laevis.
In all physical and chemical properties
known so far, including the size and com-
position of rRNA and the number and
molecular weight distribution of proteins,
mitochondrial ribosomes are clearly dis-
tinct from both bacterial and cytoplasmic
ribosomes. Therefore, these studies nei-
ther support nor speak against any of the
current hypotheses regarding the evolu-
tionary origin of mitochondria.
Mitochondrial RNA of Drosophila
C. Kaushagen
Mitochondrial ribosomal RNAs (mt-
rRNAs) from Xenopus laevis, mouse,
and humans have been characterized, but
the detailed characterization of mtRNA
from an invertebrate has not been re-
ported. We have initiated a study of
mtRNA of Drosophila melanog aster.
Mitochondria were isolated from adult
flies and the RNA extracted by methods
very similar to those used for the isola-
tion of Xenopus mtRNA. The nucleic
acids were analyzed on 2.4% polyacryl-
amide gels at room temperature (Fig.
23). Peaks corresponding to the 28S and
18S cytoplasmic rRNA were present, as
well as two peaks with mobilities similar
to those of E. coli rRNA. When similar
gels were run at 10°C (Fig. 23), the mo-
bilities of the two latter peaks were in-
creased such that both ran faster than
18S. The mobilities at 10 °C correspond
44
CARNEGIE INSTITUTION
Fig. 23. Separation of nucleic acids from a
mitochondrial preparation of D. melanog aster.
The temperature of gel electrophoresis in 2.4%
polyacrylamide gels was room temperature
(top) and 10° (bottom). Lg and Sm refer to
the large and small mt-rRNAs, respectively.
to molecular weights of about 0.55 X 106
and 0.3 X 106 MW, respectively. Be-
cause of the location in the mitochondrial
fraction and the variability of electro-
phoretic mobility with temperature, these
two RNA species were tentatively identi-
10 15 20 25
FRACTION NUMBER
30
35
Fig. 24. Sucrose gradient sedimentation of
mitochondrial RNA of D. melanogaster. Direc-
tion of sedimentation is to the left.
fled as the large and the small mitochon-
drial rRNAs of Drosophila.
Several preparations of Drosophila
mtRNA were run in sucrose-SDS gradi-
ents (Fig. 24). Two RNA components
occur which sedimented at 13S and 10.5S,
in addition to 28S and 18S rRNA, which
contaminated the preparation. These
gradients were fractionated and peak
fractions pooled to separate the two
mtRNA species. Samples of these pools
were analyzed by gel electrophoresis ; the
13S peak contained the large mt-rRNA;
the 10. 5S peak, the small mt-rRNA.
Aliquots of each of the two mtRNA
species so separated on a sucrose-SDS
gradient were spread by the Kleinschmidt
technique and analyzed by electron mi-
croscopy. Molecular weights of 560,000
and 280,000, respectively, were deter-
mined for the two mt-rRNA molecules
(Table 9) . These values are similar to
those measured for HeLa mtRNA by
Robberson and his colleagues and are
slightly but significantly smaller than the
molecular weights of X. laevis mtRNA.
Drosophila mt-rRNA is similar in size
to vertebrate mt-rRNA and also shares
its property of variable electrophoretic
mobility depending on temperature and
gel composition. Drosophila mt-rRNA
appears even more variable in this re-
spect than Xenopus RNA. Furthermore,
the sedimentation rate of the Drosophila
RNAs is also unusual: the Xenopus mt-
rRNAs, which are only slightly larger,
sediment at 13 and 17S, and the HeLa
mt-rRNAs sediment at 12 and 16S. The
Drosophila RNA sedimentation coeffi-
cients of 10.5 and 13S are distinctly lower
and suggest that these molecules assume
an unfolded configuration which slows
their sedimentation rate. We suggested
previously that Xenopus mt-rRNA has a
more unfolded structure than cytoplasmic
or bacterial rRNAs (Year Book 71, p.
30) . Drosophila mt-rRNA appears to be
a more extreme example of such an un-
folded structure.
DEPARTMENT OF EMBRYOLOGY
45
Measurement of Mitochondrial RNA
and DNA-RNA Hybrids in the
Electron Microscope
P. K. Wellauer and I. B. Dawid, with the
assistance of M . Rebbert
Major products of mitochondrial DNA
(mtDNA) are the two RNA components
of the mitochondrial ribosome, the mt-
rRNAs (see Year Book 68, p. 514; Year
Book 69, p. 575; Year Book 70, p. 42).
The physical properties of these mole-
cules from ovaries of Xenopus laevis have
been characterized to some extent. We
have continued these studies in the past
year using electron microscopic tech-
niques.
When different RNA molecules are
spread by a modified Kleinschmidt tech-
nique from 80% formamide and 4 M
urea, most types of RNA appear as fully
extended strands (for important excep-
tions, see below) . The length of RNAs of
unknown size can be compared to an
internal standard, i.e., a different RNA of
known molecular weight. This measure-
ment allows the accurate determination
of molecular weights of various RNA
molecules.
We have measured the length of Xen-
opus mt-rRNAs under these conditions
and calculated their molecular weights,
using E. coli rRNA and HeLa cell 18S
rRNA as standards. The results (Table
9) confirm the previous finding that the
mt-rRNAs are the smallest ribosomal
RNAs known. However, the values ob-
tained are larger by about 15% than the
molecular weights determined previously
by sedimentation (Year Book 70, p. 43).
Table 9 also shows the molecular
weights determined by electron micros-
copy of the mt-rRNAs from Drosophila
melanog aster. These RNAs are slightly
smaller than the corresponding molecules
from Xenopus and are very similar in size
to the mt-rRNAs from HeLa cells which
were analyzed by the same technique by
Robberson and his colleagues.
Each circular molecule of mtDNA car-
TABLE 9. Molecular Weight of Mitochondrial
rRNA Molecules as Determined by
Electron Microscopy
RNA
Moi. wt, x i(r5 ±
S. D., number
of molecules
X. laevis
large mt-rRNA
small mt-rRNA
D. melanogaster
large mt-rRNA
small mt-rRNA
6.15 =fc 0.74 (318)
3.38 d= 0.47 (289)
5.22 ± 0.81 (154)
2.98 ± 0.53 (153)
E. coli 23S rRNA was used as internal stand-
ard. Its molecular weight was taken as 1.07 X
106.
ries one region, or gene, coding for the
small and the large mt-rRNA {Year
Book 69, p. 576) . The arrangement of the
two gene regions with respect to each
other was investigated. A similar study
had been done previously by Robberson,
Aloni, Attardi, and Davidson on HeLa
cell mtDNA. These authors found that
the two rRNA genes are next to each
other, separated by a small gap. We ob-
tained similar results with Xenopus
mtDNA and RNA in the following way.
The two strands of mtDNA were sepa-
rated with the aid of polyUG by a
method introduced by Szybalski. The H-
strand, which is known to carry the
rRNA genes, was then annealed with mt-
rRNA and the hybrid examined in the
electron microscope. When both rRNAs
were annealed with the DNA, two double-
stranded regions were seen which we in-
terpret as DNA-rRNA hybrids. These
regions were separated by a small gap.
The length of each region and of the gap
was measured and the results are pre-
sented as a histogram in Fig. 25. The
sharp distributions of double-stranded re-
gions correspond to the hybrid of the H-
strand with the small and the large mt-
rRNA, respectively. The length ratio of
these two regions is very similar to the
length ratio of the two rRNAs measured
individually (Table 9) . Exact molecular
46
CARNEGIE INSTITUTION
0 0.2 0.4
LENGTH (/xm )
Fig. 25. Length of the duplex region of the
H-strand of X. laevis mtDNA with the large
and small mt-rRNA and of the gap which sepa-
rates the two heteroduplex regions. The length
of the gap is 0.046 fxm. and has been converted
into number of nucleotides. The length of the
duplex with the large mt-rRNA is 0.53 ^m, and
the length of the small rRNA is 0.30 /mi.
weights could not be calculated from
these lengths, since the linear density of
RNA-DNA hybrids is not accurately
known. The length of the gap between
the two rRNAs could be translated into
numbers of nucleotides by the use of an
internal standard of single-stranded
DNA (<f> X 174 DNA) . The length of the
gap of 120 nucleotides (Fig. 25) is very
similar to the corresponding gap between
the mt-rRNA genes of HeLa cells.
It is clear from preliminary experi-
ments (Year Book 71, p. 22) that the nu-
cleotide sequences in mtDNA of animals
of different species vary extensively.
Relatively little homology exists between
the mtDNAs of animals from different
classes of vertebrates. Nevertheless, the
arrangement of the rRNA genes on the
mtDNA has been conserved for a long
period of evolution. This fact suggests
that this arrangement has important
functional implications.
Transcription of Mitochondrial DNA
by Mitochondrial RNA Polymerase
/. B. Dawid and G.-J . Wu* with the assistance
of M. Rebbert
In Year Book 71, pp. 25-29, we re-
ported on the isolation and some proper-
ties of RNA polymerase from ovarian
mitochondria of X. laevis. The transcrip-
tion of mtDNA by this polymerase has
been studied further in the past year.
We observed previously that nascent
RNA chains were associated with the
template in such a way that they traveled
together in SDS-sucrose gradients (Year
Book 71, p. 28). When the template was
digested with DNase or when the com-
plex was heated, the RNA was released.
These facts suggested that an RNA-DNA
hybrid may exist between the nascent
RNA (or part of it) and the template.
This interpretation has now been sup-
ported by the finding that a fraction of
the nascent RNA, amounting to 25 to
30%, is resistant to digestion with
RNase; this material becomes sensitive
to RNase when the DNA is digested or
when the complex is dissociated by heat-
ing.
We investigated the size of the RNase-
resistant region in nascent RNA. For this
purpose the complex was digested with
RNase, the RNase destroyed with pro-
nase, and the resistant RNA released by
digesting the DNA with DNase. This
resistant RNA was then characterized by
chromatography on Sephadex G-100 and
by gel electrophoresis. The resistant
RNA was very heterogeneous in size, from
short fragments of about 50 nucleotides
up to long chains of about 500 nucleotides.
These results suggest that mitochondrial
RNA polymerase "opens" the DNA helix
during transcription in vitro in such a
way that a duplex is formed between the
template strand of the DNA and the
nascent RNA.
* Now at the Department of Biological Sci-
ences, Columbia University.
DEPARTMENT OF EMBRYOLOGY
47
We also investigated the problem of
initiation of RNA chains by mitochon-
drial polymerase on the mtDNA tem-
plate. This was done with nucleoside tri-
phosphates labeled in the gamma-position
with 32P. It is known that the 5'-terminal
nucleotide in any RNA chain retains its
gamma-phosphate while all internal nu-
cleotides lose it. By assaying the incor-
poration of 32P-phosphate from each of
the four nucleoside triphosphates we
found that mitochondrial RNA poly-
merase initiates predominantly with ATP
and at a frequency about two times lower
with GTP. Initiation with CTP to UTP
is very low or absent. This result is in
agreement with previous studies on other
RNA polymerases which also showed
initiation with purine nucleotides only.
We wondered whether initiation with
ATP and GTP each takes place at a par-
ticular site of the DNA or whether initia-
tion could take place at many different
sites. To answer this question, RNA was
synthesized with gamma-labeled ATP
(or GTP), the product digested with
RNase, and the resulting oligonucleotides
separated on DEAE-Sephadex columns.
If initiation had occurred on a single site,
only a single oligonucleotide should have
arisen from the terminal region of the
RNA chain. We found, however, several
oligonucleotides labeled with 32P, both
when ATP and when GTP was used as a
terminal label. These results suggest that
chain initiation of mitochondrial RNA
polymerase on mtDNA occurs in vitro on
several different sites.
It is not known how many initiation
sites exist in vivo for mitochondrial RNA,
but one might expect that there would be
only a few. If this is true, we would have
to conclude that mitochondrial polymer-
ase in vitro does not faithfully reproduce
the in vivo transcription of mtDNA.
Faithful in vitro transcription has not
been obtained so far with any system de-
rived from animal cells.
STRUCTURE
CELL
AND FUNCTION
MEMBRANES
R. E. Pagano
IN
The fundamental problem facing the
membrane biologist is how to unravel the
relationships between the molecular
structure of the membrane and its varied
functions and properties. In order to ex-
plore some of these relationships in bio-
logical membranes, we have undertaken
the development of several physical and
chemical techniques for examining some
of the surface properties of natural mem-
branes. Initially this work will empha-
size the use of phospholipid model mem-
branes, both as an analytical tool for
studying natural membranes and as a test
system for some of the physical tech-
niques under development.
Interactions between Lipid Bilayers
and Cell Surface Membranes
Since numerous cellular properties
(e.g., adhesion, vesicle formation, endo-
cytosis, and secretion) depend upon the
phenomena of adhesion and coalescence
between two continuous membranes, un-
derstanding the interactions between
membranes in terms of their molecular
composition and possible structural varia-
tions is a problem of great interest to the
developmental and the cell biologist. As
one approach to this problem, we have
begun to look at the interaction of arti-
ficially generated lipid membranes of bi-
molecular dimension with each other and
with the surface membranes of cells
grown in culture. A preliminary finding,
and basis for this approach, is outlined
schematically in Fig. 26. Chinese ham-
ster V79 cells were cultured on nylon
fibers strung across petri dishes, using a
modification of the technique described
by Edelman and co-workers. The culture
medium was then exchanged with a sim-
ple balanced salt solution, and a lipid bi-
48
CARNEGIE INSTITUTION
nylon fiber
*////////////////////////* arched ?/////////////////?7?///A
cells
phospholipid
vesicle
micropipette
(a)
Fig. 26. Schematic representation of the interaction between a lipid bilayer vesicle and a Chi-
nese hamster cell grown on a nylon fiber, (a) Bilayer vesicle is manipulated up to a single cell by
means of a micropipette. (b) Several seconds after contact the cell was observed to detach from
the nylon fiber and spread over the surface of the lipid vesicle.
layer vesicle, generated according to
standard techniques, was manipulated up
to a single cell (Fig. 26a). Several sec-
onds after contact, the cell'was observed
to spread over the surface of the bilayer
and spontaneously detach itself from the
nylon fiber (Fig. 26b). Since the molec-
ular composition of the lipid vesicle is
subject to direct experimental control,
and since the distribution of certain of
the macromolecules within the cell sur-
face can be varied by derivatization of
the fiber to which the cells are bound, it
should be possible to gain some insights
into some of the factors responsible for
cell recognition, adhesion, and possibly
membrane fusion.
Elucidation of the structure of the
junction which forms between the bilayer
vesicle and cell membrane, and the
mechanism by which it forms, must await
careful ultrastructural studies. It is ap-
parent, however, that the attractive forces
in this system are considerable, and it is
possible that a partial fusion or coales-
cence of the membranes may exist over
the region of contact. This notion gains
some support from the early studies of
Chambers and Kopac on the spontaneous
coalescence of oil droplets with the sur-
face membranes of various echinoderm
eggs.
If the junction that forms does involve
a partial fusion of the two membranes, or
if the junction is a relatively low resist-
ance one compared to that of the bilayer
vesicle (approx. 1010 ohm-cm2), then
several new kinds of electrophysiological
and biochemical investigations of the cell
membrane may be carried out because of
the unusual intrinsic properties of the bi-
layer. One of the most exciting possibil-
ities will be to look for and study the
electrical properties of individual con-
ducting channels, which by virtue of the
low ionic permeability of the bilayer
vesicle should be detectable if such chan-
nels are free to diffuse across the junc-
tional complex that forms in this system.
Such experiments have already been car-
ried out in lipid bilayers modified by the
addition of chemically well-defined
ionophores, and the statistical properties
of the resulting conducting channels have
been used to investigate and delineate the
possible mechanisms of ion permeability
DEPARTMENT OF EMBRYOLOGY
49
through membranes. A second, and very
intriguing idea, is that the generation of
such hybrid membranes as described here
could form the basis of a procedure for
the isolation and purification of mem-
brane proteins which obviates the use of
detergents or organic solvents. This tech-
nique would have the advantage that the
protein components never leave the mem-
brane phase but would be diluted in two
dimensions as the hybrid is formed, be-
cause of the fluid nature of the mem-
branes. Final purification of the mem-
brane components would take place by
chromatographic methods which have
already been successfully applied to lipid
dispersions and natural membrane prep-
arations.
Viscoelastic Properties and Surface
Tension of Membranes
When a fluid drop is placed in a liquid
of higher density contained in a rotating
horizontal tube, it will tend to find an
equilibrium position on the axis of rota-
tion because of the pressure caused by
the centrifugal force. As the speed of
rotation is increased, the droplet will
elongate along the axis until finally it is
in the form of a long narrow cylinder
with rounded ends (Fig. 27). For each
speed of rotation, the droplet will assume
an equilibrium shape determined by its
density, volume, and interfacial tension.
The quantitative relationships between
these parameters were first discussed by
Lord Rayleigh and have subsequently
been treated in greater detail by others.
To date, this technique has only been
used to measure the interfacial tensions
of simple liquids.
In collaboration with Dr. Norman L.
Gershfeld of the Laboratory of Physical
Biology, National Institutes of Health,
we are constructing an apparatus based
on the principles outlined above, which
can be used to observe the deformation of
particles of cellular dimension rotating in
a horizontal tube at angular velocities up
to about 100,000 rpm. While the immedi-
Jm&m oj
gu2
mmmmm go
cu4
Fig. 27. Schematic representation of the de-
formation of a spherical droplet in a rotating,
horizontally oriented tube, at increasing values
of the angular velocity, a>. The axis of rotation
is given by the dashed line.
ate goal of these experiments is to meas-
ure the surface tensions of a variety of
phospholipid vesicles which are to be
used in the bilayer membrane-cell surface
interaction studies outlined above, it is
also planned to use this method to exam-
ine the viscoelastic properties of cell sur-
faces. A final application of this tech-
nique will be to determine whether, under
the deformation forces of a rotational
field, cells can be induced to undergo
fusion to give viable hybrids.
Membrane Dynamics
One of the central problems in any dis-
cussion of the physical properties of a
membrane is assessment of its thickness.
For artificially generated lipid bilayers
separating two aqueous phases, the trans-
verse dimension of the membrane can be
determined from a measurement of the
intensity of light reflected from the mem-
brane and hence the difference in optical
path length between light rays reflected
from the two surfaces of the membrane.
50
CARNEGIE INSTITUTION
This technique has not been successfully
applied to cell membranes, largely be-
cause a flat membrane of not too small
dimension is needed for the experiment
and because scattering from extraneous
layers of tissue complicates the interpre-
tation of the experimental data. Re-
cently, in collaboration with Cherry and
Chapman (Department of Chemistry,
University of Sheffield), we have shown
that the reflectance technique can be used
not only to determine the transverse di-
mensions of a lipid bilayer but also to
observe membrane phase transitions and
surface inhomogeneities. Furthermore,
we found that even in bilayers formed
from a single, chemically well-defined
lipid, local fluctuations in the intensity of
reflected light could be observed. This
raises the interesting possibility that the
local molecular organization of a mem-
brane may be time dependent, and it sug-
gests that any discussion of the relations
between membrane structure and phys-
ical properties should also include this
dynamic aspect of its organization.
In order to examine this possibility in
greater detail, we are constructing the ap-
paratus described by Cherry and Chap-
man for measuring membrane reflectiv-
ity. Recent advances in the theory of
light scattering from fluctuations in mem-
branes and monolayers make us hopeful
that this type of measurement may also
eventually be applicable to the detection
of structural changes associated with
chemical reactions or other processes oc-
curring in a membrane.
STUDIES ON MUSCLE FIBER MEMBRANES
D. M . Fambrough, H. C. Hartzell, and A. K. Ritchie, in collaboration with
S. Satyamurti* and D. B. Drachman* and with the technical assistance of
N. Joseph
Acetylcholine Receptors in Skeletal
Muscle Fiber Membranes
Acetylcholine (ACh) receptors are of
special interest to us both as membrane
proteins and as elements figuring promi-
nently in neuromuscular interactions. We
have reported (Year Book 70, p. 39) on
the appearance of ACh receptors during
muscle development, the production and
incorporation of ACh receptors into the
surface membranes of myotubes, the clus-
tering of ACh receptors at newly formed
neuromuscular junctions both in vitro
and in vivo, the number and distribution
of ACh receptors in adult rat skeletal
muscle fibers, and the time course of
change in these parameters following de-
nervation. During the past year, we have
continued investigation along these lines,
focusing upon physiological aspects of
receptor function, ACh receptor turnover,
and ACh receptors in pathological states.
We have continued to use two very differ-
* Department of Neurology, The Johns Hop-
kins University Medical School.
ent assays — iontophoresis, described in
Year Book 68, p. 53, and a-bungarotoxin
binding (Year Book 70, p. 62) — to meas-
ure receptor function and the number of
receptors, respectively.
Interaction of ACh Receptors with
a-Bungaro toxin
D. Fambrough
oi-bungarotoxin (BGT) , a component
of the venom of the Formosan banded
krait, forms very stable complexes with
ACh receptors. We have used a radio-
active derivative of BGT in which
125iodine is covalently linked to the toxin
(I-BGT). For practical purposes the
complexes may be considered irreversible
because (1) when toxin is bound to mus-
cle fibers it cannot be competed off with
excess cold toxin or with cholinergic
antagonists such as tubocurarine and (2)
solubilized toxin-receptor complexes are
extremely stable. Our experiments dem-
onstrating this latter point are described
below.
DEPARTMENT OF EMBRYOLOGY
51
Free BGThas a molecular weight of ap-
proximately 8000 daltons. Toxin-receptor
complexes have a molecular weight be-
tween 200,000 and 350,000 daltons. Thus,
toxin-receptor complexes are easily dis-
tinguished from and separated from free
toxin. We prepared toxin-receptor com-
plexes by incubating chick muscle cultures
in a medium containing 0.2 /xg/ml I-BGT.
The unbound toxin was then removed by
repeatedly rinsing the cells with culture
medium, and the toxin-receptor com-
plexes were solubilized by flooding the
cultured muscle with a mixture of 1%
Triton X-100 and 10 ml Tris buffer at
pH. 7.8. This extraction is quantitative,
is complete in about one minute, and re-
sults in substantial purification of the re-
ceptors, since nuclei and much of the
cytoplasm of the cultured cells remain
attached to the petri plate. To the solu-
bilized toxin-receptor complexes we added
a 10,000-fold excess of unlabeled BGT
and, in some experiments, 10"4 M Tubo-
curarin or 10~4 to 10"2 M hexametho-
nium. These solutions were then steril-
ized by passage through a Millipore filter
and stored at controlled temperature
until analysis.
Toxin-receptor complexes were identi-
fied either as approximately 10S material
in sucrose gradient centrifugation or as
material excluded from Bio-Gel P-60
during column chromatography. Dis-
sociation of BGT from receptor was ob-
served as the appearance of free I-BGT,
which has a sedimentation constant ap-
proximately 2S and is partially included
in Bio-Gel P-60 column chromatography.
Proteolytic degradation of BGT did not
occur in any of these experiments.
At 4°C, dissociation of toxin-receptor
complexes was immeasurably slow; after
30 days all of the material still sedi-
mented as a 10. 6S peak. At 25 °C, toxin-
receptor complexes dissociated about 7%
in 6 days, suggesting a half-time for dis-
sociation greater than 6 weeks. At 37 °C
the half-time was 53 hours and at 48 °C it
was 2.5 hours. As shown in Fig. 28, com-
plexes are very rapidly dissociated at
70 °C. All of these measured dissociation
rates most probably reflect the rate of
denaturation of receptor or of BGT, with
concomitant dissociation of complexes.
Neither 10~4 M tubocurarine nor 10~2 M
nor 2 X 10"4 M hexamethonium had any
large effect upon dissociation rate. In
one set of experiments there was less than
15% dissociation in 5 days in the pres-
ence of 10~4 M tubocurarine and less than
4% in the presence of 2 X 10"4 M hexa-
methonium. Short-term experiments with
10"2 M hexamethonium suggest a slight
acceleration of dissociation, amounting to
approximately 0.6% per hour compared
with 0.25% for the control.
Samples of BGT and of I-BGT were
sent to Drs. B. Fulpius and E. Reich
(The Rockefeller University) . They re-
port that complexes of BGT with eel
electroplax ACh receptors are similarly
stable. The homologous toxin from the
cobra (Naja naja siamensis) forms com-
plexes with ACh receptors which readily
dissociate (half-time about 160 minutes
at 25 °C). The difference in stability of
the complexes is surprising in view of the
high degree of homology between the
toxins.
Turnover of ACh Receptors in Myotube
Plasma Membranes
In rapidly growing and differentiating
chick skeletal muscle cultures, we have
described the increase in ACh receptors
(Year Book 71, p. 39). There is an ap-
proximately linear increase in receptors
per culture dish with time for several
days. During this period, the receptor
density (receptors per unit area of myo-
tube surface) increases, and the amount
of myotube surface membrane also in-
creases as the result of both myotube en-
largement and new differentiation. At
early stages in the course of receptor in-
crease, the rate of appearance of new re-
ceptors is so large that turnover (that is,
the degradation of receptors and their
replacement by new ones) is negligible.
During the past year, we have studied
52
CARNEGIE INSTITUTION
Q_
U
n '
i
1
Bio-Gel P-60
600
A
A \
15 x 1 cm
400
-
! a
' t
i \
i \
i \
1 i
/ a-BGT I
200
frf-a-£-a
i i
/ \
i i I i 1
i T i I
10
20
30
FRACTION
Fig. 28. Demonstration of the practically irreversible interaction of a-BGT with the ACh receptor
from embryonic skeletal muscle. Primary cultures of chick muscle were incubated in 125I a-BGT at
0.15 /ttg/ml culture medium for 30 min at 37 °C. Then the unbound toxin was removed by extensive
washing as described in text. The cultures were immediately extracted with 1% Triton X-100 in
10 ml Tris buffer, pH 7.8, to solubilize the toxin-receptor complexes. Unlabeled a-BGT was added
to the extract to a final concentration of 100 /Ag/ml and this solution was sterilized by Millipore
filtration and stored at 25 °C for 5 days. Then an aliquot was applied to a Bio-Gel P-60 column
(15 X 1 cm) and eluted with 1% Triton X-100 in 100 ml Tris buffer, pH 7.8. 0.5-ml fractions
were collected and the radioactivity in each fraction determined by scintillation counting. Elu-
tion profile is indicated by open circles and solid lines. The unheated complexes are excluded from
the column matrix and appear in the void volume. This material was further analyzed by sucrose
gradient velocity sedimentation and shown to move as a single peak with sedimentation coeffi-
cient about 10S as judged by internal catalase and gamma-globulin markers. Another aliquot of
a-BGT-receptor complex was heated to 70 °C for 5 minutes and chromatographed as described
above (open triangles). Heating to 70°C caused release of free, undegraded 125I a-BGT, suggesting
that the toxin was not covalently bound to the receptor.
more developed cultures at which stage
turnover becomes a significant aspect of
new receptor appearance.
Part of the evidence for ACh receptor
turnover comes from measurements of
the release of radioactivity from myo-
tubes after formation of I-BGT-receptor
complexes in chick muscle cultures. In
these experiments the cultures were in-
cubated in medium containing 0.1 to 0.2
/xg/ml I-BGT for 30 min at 37 °C and
then given eight 2-minute rinses with
medium to remove unbound toxin. The
cultures were gently swirled, usually for
6 to 8 hours at 37 °C. Then the medium
was removed, centrifuged, and an aliquot
counted to obtain a determination of
radioactivity released from the cells. The
bound I-BGT was determined by count-
ing a 0.1% Triton X-100 extract of the
culture. When the released radioactivity
was expressed as a fraction of the total
(released plus bound) , the reproducibility
of values for a given set of cultures was
excellent.
The released radioactivity was ana-
DEPARTMENT OF EMBRYOLOGY
53
lyzed by Bio-Gel P-60 column chroma-
tography (Fig. 29). The great majority
of counts were found in the low-molec-
ular-weight included peak and thus rep-
resent proteolytic degradation of the
toxin. Some radioactive material ap-
peared in the excluded peak and may
reflect the liberation of toxin-receptor
complexes from the cells. Only a very
small amount of free toxin is present.
This pattern is independent of incubation
time.
Thus, the release of radioactivity from
the cells does not reflect dissociation of
toxin-receptor complexes but rather re-
flects the degradation of toxin which was
bound to ACh receptors. The degrada-
tion could occur right at the cell surface
(if proteolytic enzymes are present on the
exterior of the surface membrane) or
could occur, as more conventionally
imagined, within the cell cytoplasm after
internalization of small bits of surface
membrane. The destruction of toxin
could result in the reactivation of the re-
ceptors to which the toxin had been bound
or could be accompanied by destruc-
tion of the receptor also. We have per-
formed experiments which help to distin-
guish between these possibilities. These
experiments suggest that toxin-receptor
complexes are internalized, that both
toxin and receptor are destroyed, and
that newly synthesized receptors are con-
tinuously replacing older ones. The data
are presented in Table 10 and summarized
below.
As indicated in Table 10, the release of
radioactivity after I-BGT binding is very
temperature sensitive, is largely inhibited
by dinitrophenol and fluoride, and is par-
tially inhibited by colchicine. The release
is more rapid in the absence of serum and
embryo extract. Release is slightly
affected by cycloheximide and not signifi-
cantly affected by cytochalasin B, 0.01%
Triton X-100, nor is it affected by 10-4 M
tubocurarine or by the following com-
40
t 30
u
20
Bio-Gel P-60
15x1 cm
10
20
30
FRACTION
Fig. 29. Analysis of the radioactive material released into the culture medium for 8 hours fol-
lowing the binding of 125I a-BGT to cultured chick myotubes. Binding and column assay are as
described in text and legend to Fig. 28. Note that very little of the released material chromato-
graphs like free, undegraded toxin. Some material is excluded from the Bio-Gel beads, but the
major fraction is included, thus representing low molecular weight material, the result of enzyma-
tic degradation of a-BGT.
54
CARNEGIE INSTITUTION
TABLE 10. Release of Radioactivity from Myotubes after
125Iodine-labeled a-Bungarotoxin Binding
Experimental Treatment
Percent
126Iodine Released
per hour
Percent
Inhibition of
Release
Compared
to Control
Medium
3.12
0.31
0.00
90%
100%
1.44
46%
2.32
26%
1.26
60%
3.33
2.56
18%
5.23
0.84
0.10
84%
98%
4.26
19%
4.90
6%
3.29
37%
Complete medium*
Complete medium, 26°C
Complete medium, 18°C
Complete medium containing
10~3 M fluoride and 2 X 10~4
M dinitrophenol
Complete medium containing
100 /wg/ml colchicine
Complete medium containing
fluoride, dinitrophenol and
colchicine
Complete medium containing
0.01% Triton X-100
Complete medium containing
100 /ig/ml cycloheximide
BSA medium*
BSA medium, 26°C
BSA medium, 18°C
BSA medium containing
20 Aig/ml BGT
BSA medium containing 2.5
/Ag/ml cytochalasin B
BSA medium containing
100 Mg/ml colchicine
BSA medium containing 10~3 M
fluoride and 2 X 10~4 M
dinitrophenol
0.55
*Complete medium is a modified Ham's F12 containing 15% horse
serum, 2% chick embryo extract, and buffered with 18 ml N-2-
hydroxyethylpiperazine-N'-2-ethanesulfonate. BSA medium lacks
serum and embryo extract but contains all other components of
complete medium, including 0.5% bovine serum albumin. All in-
cubations were at 37°C unless otherwise indicated.
pounds which inhibit certain proteases:
400 ixg/m\ soybean trypsin inhibitor and
saturated solutions of tosylarginylmeth-
ylester, tosylchlorophenyl ketone, and
phenylmethanesulfonyl fluoride. This
spectrum of inhibitor sensitivity is con-
sistent with internalization of toxin-
receptor complexes followed by proteoly-
tic degradation and rapid release of
iodinated proteolytic fragments, but it is
not consistent with extracellular prote-
olytic activity.
Since cycloheximide does not drastic-
ally reduce the rate of release of 125iodine
after I-BGT binding but does block the
appearance of new receptors (after sev-
eral hours treatment) , this inhibitor was
used in experiments designed to distin-
guish between the reactivation of recep-
tors following I-BGT destruction and the
DEPARTMENT OF EMBRYOLOGY
55
replacement of such receptors by turn-
over. In these experiments one group of
cultures was pretreated 3 hours in 100
/xg/ml cycloheximide and incubated in
unlabeled BGT to block all existing re-
ceptors. Then the BGT was removed and
the cells were incubated for 6 more hours,
still in the presence of cycloheximide.
Control cultures were treated in the same
manner except that no cycloheximide was
present. At the end of the 6-hour period,
the number of I-BGT binding sites was
determined. In other controls the rate of
release of 125iodine in the presence and
absence of cycloheximide was determined.
The control cultures lost an average 2530
cpm by release of 125iodine during the
6-hour period and the BGT-pretreated
controls averaged 2660 cpm of new
I-BGT binding after 6 hours. The cyclo-
heximide-treated cultures lost an average
1760 cpm in 6 hours, but the BGT-pre-
treated, cycloheximide-treated cultures
averaged only 570 cpm new I-BGT bind-
ing after 6 hours. In the presence of cyclo-
heximide release of 125iodine continues
but many fewer new BGT binding sites
appear. Thus, the appearance of new
sites must represent new synthesis rather
than reactivation of previously blocked
sites.
Incorporation of New ACh Receptors
into Myotube Membranes
H . C. Hartzell and D. M . Fambrough
When all the ACh receptors on myo-
tubes in culture are blocked with BGT,
the appearance of new BGT binding sites
is the result of the insertion of recently
synthesized ACh receptors into the sur-
face membrane. Whether or not a net
increase in receptor sites occurs depends
upon the relative rates of insertion and
removal of receptors. The removal proc-
ess was described above as a process
wherein I-BGT is destroyed concomit-
antly with receptor removal, and the sen-
sitivity of that process to various inhibi-
tors was described. The insertion process
can likewise be studied using BGT, and
some data on the sensitivity of the proc-
ess to compounds which interfere with
ATP synthesis, to temperature, and to
cycloheximide have been published (De-
velopmental Biology 30, 155-166, 1973).
We have subsequently examined the sen-
sitivity of this process to other condi-
tions: calcium ion concentration, trans-
membrane potential, and osmolarity of
the medium. As shown in Table 11, the
incorporation of new receptors into the
surface membrane is not greatly affected
by any of these conditions.
TABLE 11. Effect of Various External Solu-
tions on ACh-Receptor Incorporation
into Membranes
Relative
Transmem-
Rate of
brane
Appearance
Potential,
of New
External Medium
mV
Receptors
Standard medium
-25
100
Na 152 ml
K 6 ml
Low potassium
medium
-33
87
Na 158 mM
K 0.3 mM
350 mM sucrose
medium
-38
98
9:1
Na 15 mM
K .06 mM
Low sodium medium
-3
115
Na3 mM
K 155 mM
High calcium medium
72
Ca 100 mM
Standard medium plus
124
150 mM sucrose
Standard medium plus
89
350 mM sucrose
Low calcium medium
126
0.001 mM Ca++
Standard medium plus
18
100 yug/ml
cycloheximide
(3 hr pretreatment)
Standard medium plus
6
2 X 10~4 M DNP
and 2 X 10"3 M
-
iodoacetate
56
CARNEGIE INSTITUTION
The cycloheximide inhibition of new
receptor appearance develops slowly be-
cause there exists within the myotubes
sufficient material to allow continued con-
struction of chemosensitive surface mem-
brane for several hours after protein syn-
thesis is blocked. By blocking all of the
surface receptors with BGT and then
measuring the binding of I-BGT to re-
ceptor sites in cell homogenates, we have
identified a set of internal binding sites
associated with material which sediments
like small mitochondria. During rapid
myotube growth, this pool of internal
binding sites contains about one-third as
many sites as do the surface membranes.
Consistent with the hypothesis that the
internal sites are indeed ACh receptors is
the observation that the internal pool of
toxin binding sites is diminished by about
30% by a 3-hour exposure of the cultures
to cycloheximide. During this treatment,
the loss of internal toxin binding sites is
quantitatively balanced by the appear-
ance of new sites in the surface mem-
brane. Another measure of the decay of
the intracellular BGT-binding site pool is
the specific activity of this pool (in I-BGT
binding sites per unit weight protein) in
control and in cycloheximide-treated cul-
tures. BGT-binding material in cell
homogenates was centrifuged at 15,000 X
g for 20 minutes and washed four times
by resuspension and centrifugation. The
final pellets were dissolved in 0.1 iV
NaOH, and aliquots were used for pro-
tein determination and for counting. In
this experiment the specific activity of
the toxin-binding fraction was 14,530
cpm/mg protein for control cultures and
8300 cpm/mg protein for 3-hour cyclo-
heximide-treated cultures — a 43% reduc-
tion in specific activity after cyclohexi-
mide treatment. Presumably the internal
toxin-binding material is preformed
membrane, and the insertion of this mem-
brane into the surface involves an exo-
cytotic mechanism.
The Neuromuscular Junctions in
Myasthenia Gravis: Decreased
Acetylcholine Receptors
D. Fambrough, in collaboration with
D. B. Drachman and S. Satyamurti
Last year we described techniques for
determining the number and distribution
of ACh receptors at neuromuscular junc-
tions and in extra junctional membranes
in rat diaphragm. During the past year,
we have put these techniques to use in the
analysis of muscle biopsy material from
humans with various muscle disorders in-
cluding severe cramps, congenital myo-
tonic dystrophy, myotonia congenita,
myokymia, Eaton-Lambert syndrome,
fascio-scapulo-humeral dystrophy and
myasthenia gravis. The ACh receptor
analytical techniques should allow us to
detect the involvement of denervation in
these muscle disorders and to assess the
state of the receptive postsynaptic mem-
brane at the neuromuscular junction. So
far, we have identified several disorders
in which no gross receptor-related ab-
normality is present, we have identified
denervated fibers in neuropathies and in
congenital myotonic dystrophy and, most
interesting, we have discovered a severe
reduction in the number of functional
ACh receptors at the neuromuscular
junctions of patients with myasthenia
gravis.
Myasthenia gravis is a neuromuscular
disorder manifested by muscular weak-
ness and fatigability. Although the ab-
normality is thought to involve the neuro-
muscular junction, the precise nature of
the defect is uncertain. We measured the
number and distribution of ACh recep-
tors in biopsy material from eight pa-
tients with typical myasthenia gravis. In
each case, the diagnosis was established
by a typical history and physical
findings of muscular weakness and fa-
tigue. The diagnosis was confirmed by a
decline of more than 15% in the third or
fourth muscle action potential in response
to a train of supramaximal stimuli ap-
DEPARTMENT OF EMBRYOLOGY
57
A
D
»M4t1
J
Fig. 30. Autoradiograms of human muscle fibers after incubation in I-BGT and staining for
acetylcholinesterase to reveal neuromuscular junctions. (A) and (B) Normal fibers from patients
HH and NA. Note dense accumulation of silver grains over the junctional area and paucity of
grains elsewhere. (C) Myasthenic fiber (patient AN) with decreased gram density but grains
localized over acetylcholinesterase-stained areas. (D) Myasthenic fiber from patient FS showing
reduced grain density at neuromuscular junction and high density of grains over adjacent extra-
junctional region. (E) Myasthenic fiber from patient FS. There was an elevated grain density
extending for at least 3 mm from the neuromuscular junction on this fiber. Phase contrast micro-
graph in which nuclei appear as dark objects and neuromuscular junction is labeled J. Magnifica-
tion bars represent 50 ixm in (A) and (E), 20 /xm in (B), (C) and (D). Specific activity of the
I-BGT ranged from 2.4 X 104 to 3.4 X 104 Ci/M. Exposure times were 4 days for (A), (D) and
(E) ; 5 days for (B) and (C). On the scale of relative grain density described in the text, (A) and
(B) are 4, (C) is 1 and (D) is 2.
58
CARNEGIE INSTITUTION
plied to a rate of 2 to 5 per second. All
eight patients showed improvement in
muscle strength after injection or oral ad-
ministration of anticholinesterase drugs.
Six of the eight had been treated with
such drugs for a year or more, while the
other two were not under treatment at
the time of biopsy. The control group
consisted of a series of patients with a
variety of problems. Two of these pa-
tients can be considered normal.
Several bundles of muscle fibers were
removed from each patient, stretched to
approximately resting length, and pinned
in dishes in a nutrient medium. Small
groups of fibers were dissected from the
bundles, separately pinned and incubated
for 2 hours at 37 °C in fresh medium con-
taining 0.1 /xg/ml a-bungarotoxin labeled
with 125iodine. After extensive rinsing
and fixation in glutaraldehyde, the fibers
were stained for acetylcholinesterase to
reveal neuromuscular junctions. For
autoradiographic analysis, lengths of sin-
gle muscle fibers including the neuro-
muscular junction were dissected out,
mounted on microscope slides, and proc-
essed for autoradiography. Representa-
tive autoradiographs are shown in Fig.
30. In order to evaluate the autoradio-
graphs without bias, two of us studied all
the autoradiographic slides after they
had been coded in random order, and
graded the grain density over the junc-
tions as follows: 4, normal; 3, possibly
reduced; 2, definitely reduced; 1, mark-
edly reduced. Our estimates were in
close agreement; the grade indicated in
Table 12 is the mean of the grades given
to all the slides by both observers. While
this technique does not yield precise
quantitative data, the basic features of
acetylcholine receptor number and dis-
tribution in myasthenic muscle fibers are
clear: The number of ACh receptors at
neuromuscular junctions is strikingly
subnormal. In most cases, there was no
increase in extra junctional receptors.
For quantitative analysis of ACh re-
ceptors (performed on 4 myasthenic and
4 normal samples), short segments of
muscle bundles containing up to 50 neu-
romuscular junctions were dissected out
and the exact number of junctions counted
at high magnification. Then each group
was hydrolyzed in 100 /xl of 6 N HC1 in a
sealed tube at 115°C for 48 to 72 hours.
The tubes were cut open and inverted in
scintillation counting vials to which was
added a miscible scintillation cocktail.
Samples were counted for at least 40
minutes at 60% efficiency. The results
are summarized in Table 12. The neuro-
muscular junctions of patients with
typical myasthenia gravis contained only
11 to 32% of the control number of ACh
receptors.
Of the eight myasthenic patients, six
were currently being treated with the
anticholinesterase drug pyridostigmine
bromide. One patient had received no
anticholinesterase medication for more
than four months and one had never had
such medication. Since these last two pa-
tients did not differ markedly from the
others in ACh number or distribution and
since two of the "control" patients who
had taken anticholinesterase medication
for more than a year had a normal re-
ceptor number, these results cannot be
attributed solely to an effect of medica-
tion. Nevertheless, to obtain some idea
of the effects of anticholinesterase treat-
ment on BGT binding sites, we gave the
anticholinesterase inhibitor neostigmine
bromide to rats. The doses were 10 to 20
times those recommended for humans.
The data are summarized in Table 13.
Acute administration of neostigmine had
no effect upon measured receptor number.
Chronic administration for 14 to 34 days
led to a decrease in receptor number al-
though not as large a decrease as seen in
myasthenic patients. Due to species and
dosage level differences, we cannot be cer-
tain whether chronic anticholinesterase
treatment of humans can lead to reduced
numbers of ACh receptors, although it
did not in the two "normal" cases re-
ported here. This matter warrants addi-
tional investigation, especially since a re-
duction in ACh receptors would seem
DEPARTMENT OF EMBRYOLOGY
59
'o
Q
d
c3
W
d
C3
CU
m
be
d
-3
d
S
d
o
c3
bfi
d
a
CU
o
CU
cu
d
'o
eg
aj
o
<!
pq
H
g
'e3
a
03
>
■3 G
03 03
d o3
° d
>>
d <J
»-5
Oh 03
bC
b£
d
ry}
eg
d
5.2
■+j ■+■=
a 2
cu d
o ^
d
e3
03
d
o3
CO
cu 73
bfi aj
d m
a o
-d >
OH
^.s£
03
d "53
° d
CK r,
03 "
w
d
-d
'-M
o
e3
o
+3
73
d
a
^S
o
d
bC
c3
X
m
03
bfi
<!
d
en
OS
CO
CD
CO
O TjH <N
lO OS oq p
i— i H r— ( <N
ON H lO H
iO W iO CO N
o d d d o
-H -H -H -H -H
o o o o o
OS O GO CO
C)hOO)
d d d d
-H -H -H -H
!*• r- r*- r-
o o o o
XXXXX XXXX
O CO M lO ©
-^ co id co oi
o
o
Ol
+
lO ^h tJH GO
t— i rt^ lO CO
T-H d o d
d
x
CO
l^-COO^uOOOO _, o
COCOiOCNi— ilOCOCO 5 ^
I I I I I I I In I
a>
03
03
03
03
03
03
03
>
>
>
>
>
>
>
>
'co "go "x 'x "on "cq "od "as
oooooooo
cu
bfi\.
CU
ftftaaaaaac-^a
rK O co
-d >»
o
TfH O
Sis
03
GO
X
d
03
O
C
03
C
o
>.
d
d
03
£
o
d
-J
X
>.
73
d
X
73'
13
X
d
CP
£
d
,d
_C'
d
o3
's
03
,d
o
X
73
d
c
CP
hing
Eaton sy
X
■>
o3
b£
ft
-
i
a
'5
,__,
s3
o3
03
o3
—
03
o
d
03
c
X
X
i
X
c3
fH
03
d
o3
-
"c
>>
03
O
o
X
£
03
£
£*-} ^5 r*~s ^i K^5 d K*i ^*i
■> '> 'S '> '> ■> *> ■>
O3o3o3o3o3o3o3c3
bCbDbfibJDbCbCbCbC
"3 "S 'd '£ '£ 'S '5 "S .
CUCUCDCUCUCUCUCP
(BCBtCCCajCBXX
O3c3o3o3o3u3o3o3
K*"} K*^ P^ r^ P^i r^i r^ K*^
££££££££
^^^^ fe pq ^ fe fe fe fe fe
O C3i lC t^ 00 GO
lO ^ lO iO H CO
H »-S W >-9 »-3 <I
OCOO^CiHCO©
NCOCDiC(N7*NN
60
CARNEGIE INSTITUTION
TABLE 13. Effect of Neostigmine on Number of Acetylcholine Receptors
(a-Bungarotoxin Binding Sites) at Neuromuscular Junctions
in Rat Diaphragm
Treatment
Control (untreated)
Acute prostigminef (24 hrs)
14 days prostigminej
34 days prostigminef
No. of Receptor
No. of
Sites per Junction
Diaphragms
(Mean db SEM)
N*
5
2.81 X 107 ± 0.07
130
3
3.27 X 107 ± 0.17
26
2
1.48 X 107 ± 0.05
22
4
1.83 X 107 ± 0.10
32
*Number of observations.
t50 to 75 Mg injected subcutaneously. One or two injections per animal.
J100 /wg injected subcutaneously twice a day. For the first 5 days, 50 to 75
Mg doses are given, to avoid excessive mortality.
highly undesirable for myasthenic pa-
tients.
The finding of a reduced number of
ACh receptors in patients with myas-
thenia gravis raises questions about the
arrangement of functional receptors at
myasthenic neuromuscular junctions. One
possibility is that the surface area of the
postsynaptic surface is reduced by 70 to
90% in myasthenic junctions. This seems
very unlikely in view of the light and
electron microscopic data currently avail-
able. These data suggest that the folding
of the postsynaptic surface is simplified
but that the junctions are extremely elon-
gated, so that the total surface area may
not be greatly abnormal. In our biopsy
material the mean length of neuromuscu-
lar junctions in myasthenic fibers was
66.5 ^m, while the mean length for con-
trols was 27.7 //.m. A more likely correlate
of reduced ACh receptor number is that
the packing density of functional ACh
receptors in the postsynaptic membrane
is reduced. This could be due to a re-
duced number of receptor molecules or,
alternatively, many of the ACh receptors
may be blocked or altered in some way to
prevent a-bungarotoxin binding. (The
clinical symptoms of myasthenia gravis
suggest the presence of a circulating neu-
romuscular blocking agent and the lim-
ited electrophysiological findings are con-
sistent with the possibility.) A reduction
in the density of functional ACh recep-
tors in the postsynaptic membrane, as
suggested by our results, would be suffi-
cient to account for the clinical manifes-
tations and physiological correlates of
myasthenia gravis.
Electrophysiological Properties of
the Membrane and Cholinergic
Receptor of Developing Myotubes
A. K. Ritchie
Two important specialized functions of
the mature skeletal muscle fiber mem-
brane are the maintenance of a large rest-
ing transmembrane potential and the
ability to generate an electrical signal
upon activation of surface receptors by
the chemical transmitter acetylcholine.
Both the magnitude of the membrane po-
tential and the electrical signal depend
upon the presence of ion selective chan-
nels in the membrane and of ion gradi-
ents, largely sodium and potassium,
across the membrane.
While these properties have been
widely studied in the adult skeletal mus-
cle, little is known about the development
of these properties during cell differentia-
tion. The ion channels responsible for the
passive membrane permeability and the
channel associated with the activated
acetylcholine receptor may conceivably
change, not only in number but in char-
DEPARTMENT OF EMBRYOLOGY
61
acter, as the membrane composition
changes to accommodate the specialized
function of the differentiated muscle
fiber. Furthermore, depending on the
activity and the work required of the
sodium pump, the steady-state distribu-
tion of ions across the membrane may
also be changing. Using electrophysio-
logical techniques, we have examined the
development of these membrane proper-
ties during the maturation of myotubes
maintained in culture.
Membrane Potential in Developing
Myotubes
The large resting potential in adult
muscle is mainly due to the high concen-
tration of K+ ions inside the cell and to
the cell membrane's large K+ permeabil-
ity relative to its Na+ permeability.
In contrast to the mature skeletal mus-
cle fiber, the membrane potential of im-
mature fibers is initially very low and
progressively increases as the fibers ma-
ture. During this period of change, it has
been reported that the internal K+ con-
centration (Ki) remains fairly constant.
Since the transmembrane potential (Em)
is largely determined by the K+ ion dis-
tribution, it seems unlikely that differ-
ences in ion distribution could account for
the large changes in resting potential. It
is more likely that changes in specific
membrane permeability are responsible
for the increase in Em with age. There-
fore, we have attempted to measure rela-
tive ionic permeabilities of the develop-
ing muscle fiber membrane from experi-
ments in which the ionic composition of
the bathing medium was altered.
Using cultured cells derived from em-
bryonic rat muscle, we measured the
membrane potential with an intracellular
microelectrode and found that the Em
increases from a mean low value of — 8
mV in myoblasts to a mean of — 54 mV
in mature fibers (Fig. 31A). These re-
sults thus confirm the data published by
others. The intracellular K+ concentra-
tion and the relative permeability of the
membrane to Na+ and K+ ions, pNa/pK,
were estimated in the following manner.
The Em of a single myotube was con-
tinually monitored through an intracellu-
lar microelectrode while the culture dish
was perfused with an ion gradient in
which the Na+ content in the medium was
completely replaced by K+ within a few
minutes. The extracellular K+ concentra-
tion could be measured within 15 /x of the
myotube by simultaneously recording
with an extracellular K+ selective ion
electrode. A curve of Em vs log K0 can
thus be generated. According to the Gold-
man equation, extrapolation of such a
curve to zero Em provides an estimate of
the intracellular K+ concentration. From
such extrapolation K; is initially 154 mM
and declines to 131 mM as the myotubes
mature (Fig. 31C). If the same data are
plotted as eEmF/RT vs K0, a straight line
is obtained and an estimate of pNa/pK
can be calculated from the y intercept of
such a plot, assuming that Nai is small
enough to be disregarded. Our results
produced such straight lines, or lines
which were only slightly concave at low
K,). The calculated pNa/pK are shown
in Fig. 31D as a function of age. The
data indicate that the low resting poten-
tial in young myotubes is due to a mem-
brane with a large pNa/pK ratio and
that the membrane becomes relatively
more permeable to K+ as the fibers ma-
ture. We hope to confirm these observa-
tions by using flame photometry to meas-
ure the internal Na and K ion content.
Due to the irregular shape of these cul-
tured myotubes, precise measurements of
the specific membrane resistance have
not been made. Therefore, we do not
know if a decrease in pNa or an increase
in pK is responsible for the declining
pNa/pK ratio with age. Similar changes
in Em and permeability ratios have been
reported by others to occur in developing
heart cells and in the fertilized sea urchin
egg. In these cells, it is believed that an
increasing potassium permeability ac-
counts for the change in permeability
ratio.
62
CARNEGIE INSTITUTION
-60r
-50
>
^ -40
C
<x>
o
Q_
<u
c
o
.a
E
-30
-20
-10
oh
iM,4 A
a— §•
II 6
e
■s-i-f B
12 5 9
H c
Q.
O
Q.
D
4 6 8 10 12
Days in culture
Fig. 31. Electrophysiological properties of rat
myotubes as a function of developmental age.
(A) The resting transmembrane potential of pre-
sumptive myoblast (open triangles) and myo-
tubes (solid circles). (B) The acetylcholine re-
versal potential determined by measuring the
ACh response in myotubes held at different
membrane potentials with an intracellular cur-
rent passing electrode. (C) The intracellular po-
tassium concentration and (D) The pNa/pK
ratios estimated by bathing myotubes in media
of increasing external potassium concentrations
while simultaneously recording the membrane
potential. Each point represents the mean ±
S.E.M. The number adjacent to each point
refers to the number of myotubes examined.
The influence of temperature on the
resting membrane potential was also ex-
amined. These studies were performed
on well-developed cells, usually after 7 to
12 days in culture. A marked tempera-
ture dependence of the membrane poten-
tial, greater than could be accounted for
by the Goldman equation, was sometimes
observed. Two likely explanations for
the phenomena are the presence of an
electrogenic Na pump or temperature-
dependent permeability changes. We are
investigating both possibilities.
There is another phenomenon which
may indicate the presence of an electro-
genic sodium pump in developing chick
muscle. After young chick myotubes are
depolarized with a large dose of acetyl-
choline for 20 to 30 seconds, their trans-
membrane potentials, instead of return-
ing to pre-ACh levels, increase to much
larger values. The increase in transmem-
brane potential is accompanied by a de-
crease in input resistance, indicating an
increase in membrane permeability. The
ACh-induced hyperpolarization occurs
when the extracellular medium contains
only NaCl but does not occur in LiCl,
sucrose, or other low-sodium media. The
effect is reversibly blocked at 25 °C and
is blocked by long-term treatment with
10"4 M ouabain (although the blockade is
slow to develop). Curiously, the ACh-
induced hyperpolarization has rarely
been observed in young rat myotubes, al-
though these myotubes also possess low
transmembrane potentials. Our tentative
explanation for the phenomenon in chick
myotubes is that the ACh by depolarizing
the cells allows an increased influx of Na
ions which activate an electrogenic meta-
bolic mechanism for Na ion extrusion.
Ionic Selectivity of the Cholinergic
Receptor during Development
The cholinergic receptor appears as one
of the earliest signs of skeletal muscle
differentiation, and throughout the early
stages of cell development there is con-
tinual insertion of new receptors into the
membrane. Activation of these acetyl-
choline receptors results in a transient
depolarization due to the opening of
channels permeable to both Na+ and K+
ions. A measure of the ionic selectivity of
such channels, that is, the relative change
in conductance to Na and K ions (AgNa/
DEPARTMENT OF EMBRYOLOGY
63
AgK), can be estimated from the acetyl-
choline reversal potential {Er) . This re-
lationship is given by the equation
Er = [EK + (AgNa/AgK)#Na]
[1+ (AgNa/AgK)]1,
where EK and ENSL refer to the K+ and Na+
equilibrium potentials (Takeuchi, J.
Physiol. 167: 128-140, 1963). Er was
measured in cultured myotubes by using
two intracellular microelectrodes, one to
pass a polarizing current and the other
to record membrane potential. Using a
third microelectrode, acetylcholine was
ejected over the surface of the myotube
during the 200 to 300 milliseconds of
polarizing currents. In all cases a re-
versal in the acetylcholine response was
obtained and plots of acetylcholine re-
sponse vs Em produced straight lines.
The membrane potential at which the
straight line intercepted the y axis was
taken as the reversal potential. The re-
sults are shown in Fig. 31B for rat myo-
tubes in different stages of development.
While it is conceivable that a change in
ionic selectivity could have occurred
which exactly compensated for a change
in ion distribution, the constancy of Er
with age suggests that neither the ionic
selectivity of the receptor nor the distri-
bution of Na+ and K+ ions is changing
appreciably with development.
Our results thus far suggest that the
ionic selectivity of the acetylcholine re-
ceptor is a fairly stable property which
does not seem to be influenced by changes
in the composition of the differentiating
membrane. However, the "passive" mem-
brane permeability characteristics of the
myotubes appear to be a variable prop-
erty which is influenced not only by sur-
face membrane differentiation but also
by metabolic state.
Some Effects of Triton X-100 on
Acetylcholine Receptors
D. M . Fambrough
Effect on ACh-induced desensitization :
The non-ionic detergent Triton X-100 is
widely used for extraction of ACh recep-
tors from tissues. In the case of cultured
chick and rat muscle, 0.05% (w/v) Triton
X-100 will solubilize receptors whereas
0.02% or less will not. However, at lower
concentrations Triton X-100 has some in-
teresting effects upon ACh receptor func-
tion and on the distribution of ACh re-
ceptors on cells. In 0.01% Triton,
myotubes fail to respond to acetylcholine
although the myotubes can maintain a
substantial transmembrane potential
(—20 to —40 mV). In 0.001% Triton
ACh sensitivity is reduced by about two
orders of magnitude compared with con-
trols, while the resting membrane poten-
tials will remain in the — 40 to — 60 mV
range for several hours. Even though
ACh receptors are not active in the pres-
ence of Triton, they retain the ability to
form complexes with BGT.
It is perhaps not surprising that recep-
tor function is perturbed by a detergent.
Another aspect of the Triton effect, how-
ever, makes it of potentially great useful-
ness in studying ACh receptor function. It
has long been known that ACh receptors
can become desensitized by high concen-
trations of acetylcholine. The process has
been difficult to study, since desensitiza-
tion is generally a minor phenomenon. In
the presence of 0.001% Triton X-100, de-
sensitization is extremely rapid and com-
plete after approximately 500 msec of
iontophoretic application of ACh. This
desensitization is also readily reversible
with a half-time of resensitization on the
order of 5 seconds. The rapidity and
completeness of desensitization in this
case may allow much more detailed anal-
ysis of the phenomenon.
Effect of ACh receptor distribution:
Triton X-100 also has the effect of over-
coming associative forces at neuromuscu-
lar junctions where ACh receptors are
closely packed and at points of ACh re-
ceptor aggregation on old chick myotubes
in culture. In both cases treatment of the
cells for 4 hours with 0.01 % Triton X-100
results in a redistribution of receptor sites
64
CARNEGIE INSTITUTION
,-.
^
v*'.
"^B.
"-
. ,: J*
■^
"V
B
Fig. 32. Bright field micrographs of autoradiographs of chick muscle cultures after 125I a-BGT
treatment. Binding conditions as described in text. Autoradiographs exposed 7 days. (A) Control
culture, 13-day chick muscle culture. Note clusters of grains, suggesting nonrandom distribution
of ACh receptors on myotubes. (B) Culture as in (A) except treated for 2 hours at 37°C with
0.01% Triton X-100 in culture medium after 125I a-BGT binding. Note the relatively uniform dis-
tribution of grains over myotubes (B) compared with control cultures (A).
as though the individual receptors were and that their redistribution is a result of
now free to diffuse (Fig. 32). Since the their movement in a fluid membrane.
Triton effect on ACh sensitivity is re- This Triton effect is not complete and
versible, it seems likely that the receptors thus far there is no satisfactory method
remain a part of the plasma membrane for quantification of the effect.
THE EFFECT OF POTASSIUM ON THE CELL
MEMBRANE POTENTIAL AND THE PASSAGE
OF CELLS THROUGH THE CELL CYCLE:
A BLOCK IN Gi
P. J. Stambrook, H. G. Sachs, and J. D. Ebert, assisted by D. Somerville and B. Smith
A close relationship exists between the
cell membrane potential and cell prolifer-
ation. Earlier experiments which manip-
ulated the membrane potential by vary-
ing the external potassium concentration
suggested that the membrane potential
might participate in the initiation or per-
petuation of DNA synthesis. In Year
Book 71, p. 34, we described normal
changes in the membrane potential of
Chinese hamster cells as a function of
their position in the cell cycle. Cells in
DEPARTMENT OF EMBRYOLOGY
65
-10 -
200
Fig. 33. Cell membrane potential as a func-
tion of external potassium concentration. Ex-
ponentially growing cells were transferred to
media containing increasing concentrations of
potassium. Five minutes after addition of the
potassium media, cells were impaled and their
membrane potentials recorded. Abscissa: potas-
sium concentration (ml) ; ordinate: membrane
potential (mV).
Gi had a characteristically low mem-
brane potential which increased abruptly
at the onset of DNA synthesis and de-
clined only when the cells entered their
subsequent mitotic division. The coinci-
dent rise in membrane potential and on-
set of DNA synthesis have prompted fur-
ther experimentation. The experiments
were designed to determine whether an in-
crease in membrane potential is required
for cells to begin DNA synthesis and to
traverse the S phase and G2 to their next
mitotic division. A second question ex-
amined was whether the high K+ medium
blocks cells randomly throughout the cell
cycle or whether it blocks cells at specific
times in the cycle.
Chinese hamster cells respond to high
external potassium concentrations with a
decrease in cell membrane potential (Fig.
33) and a reduction in rate of cell prolif-
eration (Fig. 34) . At 115 ml K+ there is
little, if any, increase in cell number.
Although prolonged exposure to 115 ml
K+ is eventually lethal to the cells, there
is complete recovery of cell growth after
12 hours when high K+ medium is re-
placed by fresh control medium (Fig. 35) .
When the mitotic index of the cultures
in various K+ media was determined, we
observed an inverse relationship between
the growth rate and the number of cells
in mitosis (Table 14). With increasing
K+ concentrations there is a reduction in
growth rate yet an increase in the fre-
quency of mitotic cells, suggesting that
high exogenous K+ levels impede the
progress of cells through mitosis.
20 40 60
TIME (hrs)
80
Fig. 34. Effect of external potassium on cell
growth. Cells were seeded into replicate 60-mm
culture plates at 105 cells per plate. After 24
hours the cells were placed into media contain-
ing increasing concentrations of potassium, and
cell counts on duplicate plates were made at
subsequent intervals. 1 raM K+, crosses ; 10 mM
K+, triangles; 30 mM K+, solid circles; 60 rmV
K+, open squares; 100 mM K+, solid squares.
66
CARNEGIE INSTITUTION
TIME (hrs)
Fig. 35. Recovery of cell growth. Cells were
seeded into replicate 60-mm culture plates at 105
cells per plate. After 24 hours, media containing
various potassium levels was introduced. Cells
incubated with 115-mM potassium were divided
into two groups after 12 hours. One series was
transferred to fresh medium containing 10 mM
potassium and the other was incubated further
in 115 mM potassium. The number of cells per
plate was determined in duplicate at different
times during the culture period. 10 mM K+
(open circles, solid line) ; 100 mM K+ (squares,
solid line) ; 115 mM K+ (open triangles, solid
line) ; 115 mM K+ transferred to 10 mM K+
(solid triangles, dashed line).
To further examine the effect of high
levels of potassium on the cell cycle, cells
were synchronized by collection of mi-
totic cells, and their passage through the
cell cycle was followed in the presence or
absence of high K+ medium. A popula-
tion of mitotic cells was divided into four
series of aliquots, each series containing
18 replicate culture plates. The first
series contained cells seeded directly into
medium containing 10 mM K+ and served
as the control (Fig. 36a). In the second
series the cells were placed directly into
medium containing 115 mM K+ (Fig.
36b). The third group consisted of cells
first suspended in 115 mM K+ and after 2
hours transferred to 10 mM K+ (Fig.
36c). The last group was first placed in
10 mM K+ and after 2 hours transferred
to 115 mM K+ prior to the onset of DNA
synthesis. In each series, duplicate plates
of cells were labeled with 3H thymidine
for 10 minutes at varying times after
mitosis, and at every time interval the
percentage of labeled cells was deter-
mined by autoradiography. The mitotic
index was calculated from the same plate.
The control series displayed an approxi-
mate 10-hour generation time with about
a 3-hour Gi and a 6-hour S phase (Fig.
36a). Cells placed directly into 115 mM
K+ medium were retarded in their exit
from mitosis and exhibited a greatly im-
paired DNA synthetic activity (Fig.
36b) . At no time interval examined were
more than 60% of the cells labeled, and
the density of label over nuclei was con-
TABLE 14. The Relationship between Exo-
genous Potassium Concentration and
Mitotic Index
Potassium Concentra-
tion (mM)
Mitotic Index
1.0
1.8
10
2.2
30
2.4
60
3.6
100
6.2
115
8.8
Cells seeded at 105 cells per 60 mm culture
plate were grown for 24 hours in control medium
and transferred to media containing 1.0, 10,
30, 60, 100, or 115 mM K+. After 60 hours of
additional culture (24 hours for the 115 mM K+
sample) the media were aspirated and the cells
washed and fixed in methanol :ace tic acid (3:1).
One thousand cells were counted from several
random fields from each plate and the percent-
age of mitotic cells determined.
DEPARTMENT OF EMBRYOLOGY
67
100*
Q
LU
_l
UJ
CD
<
(/)
UJ
O
u_
O
LU
O
QC
LU
CL
100 -¥
~1 —
24
TIME (hrs)
XI000
-200
-100
0 ?
i
iooo A
h200 <3
100
-100
0
[j 1000
-200
-100
0
LU
o
o
o
o
0 tr
LU
j;l000 CL
200 <f)
LU
O
O
\-
o
Fig. 36. Response of synchronized cells to different exogenous potassium regimens. A synchro-
nized cell population was prepared by selective detachment of mitotic cells and divided into four
different aliquots. Each series consisted of 18 replicate 60 mm culture plates with equal numbers
of cells per plate, (a) Cells incubated in 10 ml K+. (b) Cells incubated in 115 mM K+. (c) Cells
incubated for 2 hours in 115 mM K+ then transferred to 10 mM K+. (d) Cells incubated for 2
hours in 10 mM K+ then transferred to 115 mM K+. In each series, duplicate plates were labeled
with 3H thymidine (1 fic/ml) for 10 minutes at successive intervals after synchronization. The
plates were aspirated and washed, and the cells fixed with methanol: acetic acid (3:1). The plates
were then coated with emulsion for autoradiography. Percentage labeled cells (open circles, solid
line); mitotic cells per 1000 cells (crosses, dashed line).
siderably lower than in the control series.
By 14 hours, 4 hours after the mitotic
peak in the control series, no mitotic cells
were evident. However, at least a frac-
tion of the population was able to tra-
verse the entire cycle, albeit abnormally,
since at 24 hours nearly 20% of the cells
were in metaphase, apparently blocked at
that stage. Cells that were first sus-
pended in high K+ medium and two hours
later transferred to control medium (Fig.
36c) were retarded in their exit from
mitosis. Once fresh medium was added,
the cells progressed normally through the
remainder of the cycle, although all
marker events were delayed by 2 hours
compared to the control series. The
marker events included the times of onset
of DNA synthesis, the minimum in the
DNA synthetic curve, and the mitotic
peak. In the last series (Fig. 36d), the
cells, which were first suspended in con-
68
CARNEGIE INSTITUTION
trol medium and transferred to 115 ml
K+ medium at 2 hours, exited mitosis
normally and began DNA synthesis on
schedule. The duration of the S phase
appeared slightly lengthened and the cells
arrived at mitosis with minor delay. A
block in mitosis was again evident, since
cells accumulated in mitosis at 12 and 14
hours, but by 24 hours the mitotic index
had dropped below control values.
Three conclusions can be drawn from
the above series of experiments. (1) Cells
can initiate and continue DNA synthesis
while exhibiting a low membrane poten-
tial (Figs. 33 and 36d). Synchronized
cells incubated with control medium but
transferred to 115 mM K+ medium before
the initiation of DNA synthesis can tra-
verse the S phase and G2 to their next
division with little delay. (2) The high
potassium medium impedes the passage
of cells through mitosis (Table 14 and
Figs. 36b and 36d) . (3) There is an event
(or events) during Gi that is particularly
sensitive to the 115 mM" K+ medium and
that may be required for the initiation
and perpetuation of DNA synthesis. Mi-
totic cells suspended in 115 mM K+ me-
dium exhibit impaired DNA synthetic
activity (Fig. 36b). However, if 115 mM
K+ medium is added 2 hours after mitosis,
the cells proceed through DNA synthe-
sis and the remainder of the cell cycle
almost normally (Fig. 36d) . Presumably
the high K+-sensitive block is passed dur-
ing the 2-hour incubation in control me-
dium allowing subsequent traverse of the
remainder of the cycle in high K+ medium.
As demonstrated in Table 14 and in
Figs. 36b and 36d, the high K+ medium
interferes with mitosis. The effect of the
high K+ medium on the chromosomes
themselves was examined by making
metaphase chromosome preparations. In
stained chromosome spreads of cells from
the 24-hour time point of Fig. 4b, up to
80% of the metaphase figures consisted of
condensed chromosomes that were dis-
tinctly banded (Fig. 37a). Furthermore,
a number of the metaphase figures dis-
:
Fig. 37. Effect of prolonged exposure to 115
mM K+ on chromosome morphology. Mitotic
cells were collected by mitotic selection and ex-
posed to 115 mM K+ for 24 hours (Fig. 36b).
All cells were trypsinized, washed, and sus-
pended in 1% Na citrate for 15 minutes. After
fixation in methanol: acetic acid (3:1), meta-
phase chromosomes were spread on slides, (a)
Metaphase chromosomes from a control culture,
(b) Metaphase chromosomes from a potassium-
treated culture showing heavy banding, (c)
Metaphase chromosomes from a potassium-
treated culture showing distinct coiling.
DEPARTMENT OF EMBRYOLOGY
69
played a marked supercoiling (Fig. 37b) banding and supercoiling suggest that
that presumably contributes to the ob- aside from the Gi block, one effect of the
served banding. A metaphase figure from high K+ may be to modify the chromo-
control cultures is presented for compari- some architecture and thereby impede
son (Fig. 37c). The induced chromosome the mitotic event.
A CELLULAR RESPONSE TO DIBUTYRYL
CYCLIC AMP: A CELL CYCLE BLOCK IN G2
P. J. Stambrook and C. Velez, assisted by D. Somerville and B. Smith
The above report shows that a high
level of external potassium not only ex-
erts an effect on the cell membrane po-
tential but also blocks Chinese hamster
cells at specific junctures in the cell cycle,
namely, in mitosis and Gi. In an effort
to uncover other cell cycle-specific phe-
nomena, cells were treated with dibutyryl
cyclic AMP [(but)2 cAMP] which is
also known to reduce cell growth rate and
the density to which cells will grow. The
response of exponentially growing cells to
increasing (but) 2 cAMP concentrations
in the medium is shown in Fig. 38. Cells
UJ
<
Ul
Q.
OC
Ul
m
z
UJ
o
I07
-"^^'
^-J7
-
/
/
/
/
/^-»
5xl06
MS^
"~0 0
/ x/
I06
-
5xl05
"
1
l
1
-24
24
TIME (hrs.)
48
72
Fig. 38. Effect of (but)2 cAMP on cell growth. Replicate culture plates (60 mm) were seeded
with 2 X 105 cells. After 24 hours (time = 0 hr), the medium was replaced with media containing
increasing amounts of (but)2 cAMP, and cell counts were made at subsequent intervals. Control
(solid circles, solid line) ; 10"4 M (but)2 cAMP (crosses, solid line) ; 5 X 10"4 M (but)2 cAMP
(squares, solid line) ; 10"3 M (but)2 cAMP (open circles, solid line) ; 10"3 M (but)2 cAMP changed
to control medium at 48 hours (inverted triangles, dashed line).
70
CARNEGIE INSTITUTION
Fig. 39. Cell culture response to removal of (but)2 cAMP. Cells were seeded at 2 X 105 cells per
60 mm culture plate and 24 hours later transferred to medium containing 10-3 M (but)2 cAMP.
After 48 hours in (but)2 cAMP the drug was removed and fresh medium added, (a) A typical field
in the culture plate one hour after removal of the drug, (b) A field in the culture plate 2 hours
after removal of the drug.
DEPARTMENT OF EMBRYOLOGY
71
Fig. 40. Release and progress of a population
of cells blocked by (but)2 cAMP. Exponentially
growing cells in replicate culture plates were ex-
posed to 10-3 M (but) 2 cAMP for 48 hours. At
that time, the drug was removed and duplicate
plates were labeled with 3H thymidine (1 /xc/
ml) for 15 minutes at successive hourly inter-
vals. The cells were fixed with methanol: acetic
acid (3:1) and the plates were coated with
emulsion for autoradiography. Percentage
mitotic cells (open circles, solid line) ; percent-
age labeled cells (solid circles, dashed line).
treated with 10-4 M (but)2 cAMP display
only a slightly altered growth rate,
whereas at concentrations of 5 X 104 M
and 10"3 M (but)2 cAMP, growth rates
and saturation density are visibly re-
duced. After 48 hours in 10"3M (but)2
cAMP there is little further increase in
cell number. The cells, however, are
viable since normal growth resumes as
soon as the (but)2 cAMP is removed
(Fig. 38).
To determine whether (but)2 cAMP
blocks cells randomly in the cell cycle or
at a particular stage of the cell cycle, cells
were incubated in (but)2 cAMP for 48
hours. At that time, the drug was re-
moved and replaced with fresh medium.
Subsequently, at hourly intervals, repli-
cate cultures were labeled for 15 minutes
with 3H thymidine. The cells were fixed
with methanol: acetic acid (3:1) and
coated with emulsion for autoradiogra-
phy. Within one hour after withdrawal
of the drug, between 15% and 20% of the
cell population was in metaphase (Figs.
39a, 40, and 41). The normal mitotic
index for these cells is roughly 4% and
after extended exposure to 10"3 M (but) 2
cAMP is elevated to about 7% (Fig. 41).
The sudden burst of mitotic activity rep-
resents the release of a viable population
of cells which had accumulated at a point
prior to the onset of cell division. The
released cells proceed through metaphase
so that 2 hours after removal of (but)2
cAMP a large fraction of the mitotic
population had progressed from meta-
phase to anaphase and telophase (Fig.
39b). Although there is variability from
one experiment to the next, the mitotic
index drops roughly 4-fold over the next
2 hours (Figs. 40 and 41 ) . The percentage
of cells which incorporate 3H thymidine
at each time interval reflects the progres-
sion of this partially synchronized popu-
lation of cells through the cell cycle.
There is a 40% decrease in the fraction of
labeled cells as the synchronized sub-
population enters and traverses Gi. This
decline in labeled cells is followed by a
greater than 3-fold rise in the fraction of
cells synthesizing DNA as the parasyn-
chronous subpopulation enters the S
phase (Fig. 40).
The rapid appearance of mitotic cells
after removal of the (but)2 cAMP indi-
cates that under these incubation condi-
tions a block exists shortly before the
onset of cell division, either in G2 or at
the end of the S phase. To distinguish
between these two alternatives, cells were
again grown for 48 hours in the presence
of (but) 2 cAMP. After 48 hours the drug
was removed and the cells exposed for
15 minutes to 3H thymidine so as to label
only those cells which were in the S phase
72
CARNEGIE INSTITUTION
CO
Id
O
O
I-
o
z
Id
cc
Ld
Q.
TIME (hrs )
Fig. 41. Localization of the (but)2 cAMP inhibition as a block in G2. Exponentially growing cells
in replicate plates were grown for 48 hours in the presence of 10"3 M (but)2 cAMP. After 48 hours
the drug was removed and the cells pulse labeled for 15 minutes with 3H thymidine (1 ^c/ml).
The radioactive medium was aspirated and the cells were quickly washed. Fresh medium contain-
ing a 100-fold excess of unlabeled thymidine was added to the plates. At subsequent hourly inter-
vals, the cells were fixed with methanol: acetic acid (3:1) and the plates coated with emulsion for
autoradiography. Percentage mitotic cells (open circles, solid line) ; percentage labeled mitotic cells
(crosses, dashed line).
at the time (but)2 cAMP was withdrawn.
The plates were washed and fresh me-
dium containing a 100-fold excess of un-
labeled thymidine was added to the cells.
At subsequent hourly intervals the cells
were washed and fixed with methanol:
acetic acid (3:1) and the plates coated
with emulsion for autoradiographic ex-
amination. For the first 2 hours following
the removal of the drug, approximately
20% the cells were in mitosis (Fig. 41).
None of the mitotic cells at these early
time intervals, however, was labeled (Fig.
41), which indicates that between 30%
and 40% of the cells had accumulated in
G2 during their incubation with (but)2
cAMP. At late time intervals, as the
number of mitotic cells declines, the pro-
portion of mitotic cells that are labeled
increases up to 100%. Since mitotic cells
at later intervals are labeled, they must
have been distributed throughout the S
phase at the time of exposure to 3H thy-
midine. This observation agrees with the
finding that at the time (but)2 cAMP is
removed, 30% of the cells incorporate
radioactivity when labeled with 3H thy-
midine.
In summary, the reduction in cell
growth rate and cell density effected by
incubation with (but)2 cAMP is not due
to random lengthening of the cell cycle
but is the result of an accumulation of
cells in G2 and to a lesser extent in mi-
tosis. The remainder of the cells are
scattered throughout the S phase and
maybe Gi, although this latter possibility
has not been pursued.
DEPARTMENT OF EMBRYOLOGY
73
FUNCTIONAL DIFFERENTIATION OF THE
EMBRYONIC HEART
R. L. DeHaan, T. E. Durr, C. R. Krueger, T. F. McDonald, L. Moyzis, C. E. Plitt,
H. G. Sachs, and M . Springer
During the past year, DeHaan and his
colleagues have focused most of their at-
tention on three aspects of the physiolog-
ical development of heart cells: (1) the
differentiation of tetrodotoxin-sensitive
fast sodium channels in the early heart,
(2) the rate-setting mechanism in syn-
chronized groups of cells, and (3) the
state of myofibrillar organization at the
time of the first heartbeats.
Modulation of Sodium Current
Channels in Embryonic Heart
Cells by Tetrodotoxin
The action potential in adult cardiac
tissue is associated with a rapid increase
in conductance to inwardly flowing ions.
This inward current is known to be car-
ried mainly by sodium ions moving
through specific sodium channels at high
velocity. A second component of the in-
ward current appears to be carried
through different channels, which are less
specific. These permit either sodium or
calcium ions to flow through them, at
slower velocities than the specific sodium
channels.
It has been known for some time that
the fast sodium channels are blocked by
the Puffer-fish poison tetrodotoxin
(TTX). In contrast, the slow calcium
channels are insensitive to TTX, but
these may be blocked instead by manga-
nese ions or by the drug D-600.
In work described in these pages last
year, McDonald, Sachs, and DeHaan
showed that the heart of the early chick
embryo was virtually insensitive to TTX,
unlike adult cardiac tissue, suggesting
that the action potential in the early
heart involved only a slow channel
mechanism. Intact hearts from 2- to 4-
day-old embryos beat spontaneously in
TTX at concentrations up to 10"5 g/ml.
Tetrodotoxin sensitivity increased pro-
gressively with development, however.
Hearts from 7-day embryos were blocked
at a concentration of 10"7 g/ml TTX.
Furthermore, they found that spheroidal
aggregates formed from reassociated
heart cells behaved in similar fashion,
increasing in TTX sensitivity with em-
bryonic age. During the past year, these
workers have shown that the fraction of
such aggregates blocked by TTX also
changes with time after exposure to the
drug. Those aggregates, prepared from
cells of 4-day hearts, whose spontaneous
beat is suppressed upon treatment with
TTX gradually lose their sensitivity to
the drug over a 2—3 hour period of ex-
posure, resume beating in its presence,
and are not blocked by a fresh dose at the
same concentration (10~5 gm/ml). Using
aggregates and analytical techniques
similar to those described previously, De-
Haan and his colleagues have therefore
designed experiments to test whether this
loss of TTX response: (1) resulted from
desensitization of the aggregate cells or
from an inactivation of the drug in their
presence, (2) was associated with changes
in electrophysiological parameters of the
treated cells, and (3) was prevented by
inhibitors of protein synthesis.
Figure 42 shows the effect of TTX on
the beating of aggregates from hearts
aged 4 to 7 days. Fifteen minutes after
the addition of the drug 26% of the 4-day
aggregates were beating, while less than
10% of the 6- and 7-day aggregates re-
mained active. During the next 2 hours
many of the 4- and 5-day aggregates
resumed beating, while there was little or
no increase in the number of active 6-day
and 7-day aggregates.
The pattern of suppression and recov-
ery in 4- and 5-day aggregates treated
with TTX was apparent whether aggre-
gates were formed from ventricular cells,
74
CARNEGIE INSTITUTION
CO
LlI
(3
<
LlI
CD
100 r
80
<
Ld
g 60
<
40
20
TIME (hr)
Fig. 42. The response of heart cell aggregates to TTX (10~5 g/ml). Aggregates were formed from
dissociated heart cells of chick embryos aged 4 to 7 days and maintained in medium 818A con-
taining 1.3 mM/1 K+. Data for 4-day and 7-day aggregates are mean values from 6-8 plates in 4
different cultures; data for 5- and 6-day are mean values of 2 plates each.
atrial cells, or cells from the whole heart.
The response was also essentially the
same in medium containing 1.3 or 4.3
mM/1 K+. The percentage of 4-day ag-
gregates beating after 15 minutes in TTX
varied from culture to culture but was
usually 15% to 35%. Within the next 2
to 3 hours, the percentage of beating ag-
gregates increased to between 50% and
80%. During prolonged incubation with
TTX, the beating percentage after 6
hours was usually the same as after 3
hours. Aggregates from 7-day heart
failed to show a similar suppression-
recovery response even at lower doses of
TTX. For example, at 2.5 X 10"8 g/ml
40% of 7-day aggregates continued beat-
ing. Over a 4-hour incubation at that
concentration none of the suppressed ag-
gregates resumed activity.
To confirm that the toxin was not in-
activated in the culture plates, TTX was
added to one of a pair of plates contain-
ing 4-day or 5-day aggregates. After 3
hours, the TTX-medium was transferred
to the companion plate. Such an experi-
ment is shown in Fig. 43. The two plates
had essentially identical suppression-
recovery responses, indicating that the
toxin remained active. Further, when the
companion plate (Fig. 43B) was washed
and incubated for another 45 minutes
with fresh TTX, the percentage of beat-
ing aggregates did not substantially de-
cline.
The electrical activity of 4-day aggre-
gates was monitored at intervals during
the 3-hour incubation with TTX (Fig.
44). Normal pacemaker activity was
blocked about 3 minutes after the addi-
tion of the drug, resulting in a stable
resting potential (Fig. 44A, arrow) . Ap-
proximately an hour later, in those ag-
gregates that resumed beating, the stable
resting potential gave way to small oscil-
lations and subsequent resumption of
full-scale pacemaker action potentials
(Fig. 44B).
To assess the changes in electrical
properties of cells that might be corre-
lated with desensitization, action poten-
tials were recorded in some 60 aggregates
(in three experiments) before treatment
with TTX and during the second and
DEPARTMENT OF EMBRYOLOGY
75
r5
(A) TTX IO"°g/ml
(B) TTX medium from (A)
100 r
3 0
TIME (hr)
Fig. 43. The response of 5-day aggregates to TTX (10~5 g/ml). (A) One plate of aggregates was
incubated for 3 hours with medium containing TTX. (B) After the 3-hour incubation, the medium
from plate A was transferred to duplicate plate B for 3 hours. Plate B was subsequently washed
three times and immediately treated with fresh TTX.
third hours of incubation with the drug.
The majority of impalements during the
TTX incubation were from aggregates
which had stopped and then resumed
beating. The results are shown in Table
15. There was no important difference in
the maximum diastolic depolarization
(MDP) or overshoot (OS) of aggregates
before and during TTX. However, in the
course of TTX-desensitization, the
Fig. 44. Electrical activtiy in a 4-day heart aggregate before and during incubation with TTX
(10~5 g/ml). (A) Superimposed records showing control action potentials before TTX and a stable
resting potential without action potentials (arrow) 3 minutes after treatment with the drug. (B)
Resumption of pacemaker activity during TTX incubation. Record B was obtained about one
hour after record A. Voltage calibration is 100 mV. The upper horizontal line in both records in-
dicates 0 mV. Time calibration is 400 msec in (A) and 4 sec in (B).
76
CARNEGIE INSTITUTION
TABLE 15. Action Potential Parameters of Beating 4-day Aggregates before and during
Incubation with TTX
Control
Experimental
TTX (g/ml)
K+ (mAf/1)
MDP (mV)
OS (mV)
Vmax (V/sec)
N
0
0
1 X 10~5
1 x 10-5
1.3
4.3
1.3
4.3
67.8 ± 2.4
66.4 ± 1.9
72.2 dz 1.5
70.9 ± 1.6
16.8 ± 1.0
21.2 d= 1.3
20.3 ± 1.0
23.2 d= 1.4
22.8 d= 1.5
33.4 d= 2.1
12.1 =b 0.9
12.4 d= 1.1
20
18
20
19
The maximum diastolic potential (MDP) overshoot (OS) and maximum rate of rise of the action
potential (V max) were measured in aggregates before TTX and during the second and third hours °f
incubation with the drug in medium containing 1.3 or 4.3 milf/1 potassium. Data represent mean
dz S. E. for N number of impalements.
maximum rate of rise of the action poten-
tial (Vmax) was reduced on the average
by more than 50%. Before TTX, Vmax
ranged from 8 to 45 V/sec and exceeded
30 V/sec in about one-third of the im-
palements, while in the presence of the
drug the highest Vmax was 18 V/sec.
The appearance of TTX-insensitive
action potentials in skeletal muscle fol-
lowing denervation has been found by
other workers to be dependent on a pro-
tein synthetic process and was sensitive
to inhibition by cycloheximide. When 4-
day aggregates were pretreated for 4
100 r
n
LxJ
O
Ixl
CC
<D
<
CD
<
LU
CD
TIME (hr)
Fig. 45. Inhibition of TTX-desensitization in 4-day heart aggregates by treatment with cyclo-
heximide (5 jUg/ml). Control aggregates treated with TTX (10-5 g/ml) (solid circles) ; 15 minutes
pretreatment with cycloheximide followed by cycloheximide and TTX (open triangles) ; 4 hours
pretreatment with cycloheximide followed by cycloheximide and TTX (solid triangles) . Aggre-
gates were washed three times after 3Mj hr in TTX. Values are the means of 2 plates each.
DEPARTMENT OF EMBRYOLOGY
77
hours with cycloheximide at a concentra-
tion sufficient to block 90% of amino acid
incorporation (Fig. 45, solid triangles),
the aggregates continued to beat normally
in control medium and did not become
desensitized in TTX. After 3% hours the
percentage of beating aggregates was still
only 17. In contrast, those not exposed to
cycloheximide (solid circles) or treated
for only 15 minutes (open triangles) be-
fore TTX rebounded from 16% to about
75% beating. Fifteen minutes after wash-
ing out the TTX, the percentages of beat-
ing aggregates in the control and those in
the 4-hour cycloheximide plates returned
to 100 and 98, respectively. To test
whether 4-day aggregates responded to
manganese or to D600, which block slow
calcium channels, preparations in tricine-
buffered medium were treated with TTX,
allowed to undergo desensitization for 2.5
hours, and then exposed to manganese ion
(Fig. 46). At a concentration of 0.5 mf,
which had no appreciable effect on the
activity of control aggregates, the per-
centage of desensitized aggregates that
remained beating decreased from 64% to
6%. In one experiment, D600 (1/xg/ml)
Contro
_a A
Fig. 46. Response of 4-day aggregates to
manganese. Aggregates were treated with 10-5
g/ml TTX (circles) or an equal volume of
diluent (triangles), and the percentage of beat-
ing aggregates (% BA) was determined 15, 75,
and 150 minutes later. At that time, 20 /x\ of
50 mM MnCl2 was added to each dish, bringing
the final concentration of 0.5 mM. Thirty min-
utes later % BA was again determined. Each
point represents the mean of 4 plates.
was found to have a similar effect. Im-
palement of nonbeating aggregates in the
presence of either agent showed them to
be electrically inactive.
To determine whether desensitization
is a time-dependent process that might be
triggered by even a brief exposure to
TTX, the experiment illustrated in Fig.
47 was performed. Aggregates were
treated with TTX for 15 minutes, washed,
incubated witl} drug-free medium for 2%
hours, and then exposed to TTX again
(Fig. 47A). The percentage of suppressed
aggregates was similar after both doses.
When companion aggregates were desen-
sitized by a 3-hour exposure to TTX,
washed, and again treated with the toxin,
the percentage of beating aggregates
counted 15 minutes later was similar to
that present at the end of the 3-hour
TTX densensitization period (Fig. 47B).
The response of desensitized aggregates
to the second dose of TTX was in marked
contrast to that of control aggregates,
indicating that a brief exposure to TTX
does not produce desensitization and that
desensitized aggregates remain in that
state after the removal of TTX.
To determine how long the desensitiza-
tion lasts after removal of TTX, DeHaan
and his colleagues employed the following
experimental protocol: aggregates were
incubated with TTX for 3 hours, washed,
and then incubated for 0—4 hours in drug-
free medium before exposure to a second
dose of TTX. After the initial 3-hour
incubation with TTX, the aggregate pop-
ulation was divided into three groups (see
Fig. 47B) : (1) Sensitive aggregates —
those not beating after 3 hours in TTX.
(2) Desensitized aggregates — aggregates
which resumed beating after between 15
minutes and 3 hours in TTX. (3) Insen-
sitive aggregates — those which had con-
tinued beating after the initial 15 minutes
in TTX. In the cycloheximide experi-
ment (Fig. 45, solid triangles) the per-
centage of insensitive aggregates re-
mained nearly constant throughout the
3-hour incubation period with TTX. As-
suming that the percentage of sensitive
78
CO
LlI
(A)
CARNEGIE INSTITUTION
(B)
TTX
lOOr ♦ n
TTX.
80
<
LU
tt 60
e>
o
<
(3
z
|= 40
<
LU
CD
20
0L
> SENSITIVE
> DESENSITIZED
> INSENSITIVE
Fig. 47. The response of 4-day heart aggregates to successive doses of TTX (10~5 g/ml). (A)
Aggregates were incubated with TTX for 15 minutes, washed 3 times, and then incubated in drug-
free medium for 2% hours before the second dose of TTX. (B) Aggregates were incubated for 3
hours with TTX, washed 3 times, and then treated with TTX 15 minutes later. Values are the
means of 2 plates each. Also illustrated is the method used to classify aggregates into 3 groups —
sensitive, densensitized, and insensitive — for the experiments described in Fig. 48.
aggregates also remained constant during
experiments similar to that depicted in
Fig. 47B, any decrease in the total per-
centage of beating aggregates after the
second dose of TTX may be attributed
to a decline in the number of desensitized
aggregates.
The percentage of desensitized aggre-
gates after the second dose of TTX was
expressed as a fraction of the percentage
of desensitized aggregates following the
initial 3-hour incubation with TTX. This
fraction was plotted against elapsed time
between washing and the second dose of
TTX (Fig. 48). The fraction of densen-
sitized aggregates remained nearly con-
stant for 1 hour, declined sharply during
the next hour, and stabilized thereafter at
about zero ; that is, 4 hours after removal
of TTX, just as many aggregates were
sensitive to the drug as had been prior to
the first exposure.
The curve in Fig. 48 illustrates the re-
appearance of response to TTX following
withdrawal from a prolonged exposure to
the drug at 10"5 g/ml. In earlier experi-
ments, during which the dose of TTX was
raised in increments gradually from 10"8
g/ml to 10"5 g/ml, desensitization was
sometimes apparent by the time (about
IV2 hours) the final dose of 10"5 g/ml was
administered, i.e., the large dose did not
provoke as great a depression of beating
as had a smaller dose earlier. Since de-
sensitization occurred upon exposure to
lower doses of TTX, it seemed possible
DEPARTMENT OF EMBRYOLOGY
79
l.2r
0.8
£ 0.6
0.4
0.2
■0.2
TIME (hr)
Fig. 48. Time course of the reappearance of
response to TTX by 4-day aggregates. Aggre-
gates were incubated with TTX (10~5 g/ml) for
3 hours, washed 3 times, and incubated in drug-
free medium for a specified period of time be-
fore receiving a second dose of TTX. The
abscissa denotes the time in drug-free medium
before the second dose. The percentage of de-
sensitized aggregates beating 15 minutes after
the second dose of TTX was compared to the
percentage of desensitized aggregates at the end
of the initial 3-hour TTX incubation and ex-
pressed as a fraction thereof. For example, the
fraction 0.2 indicates that only one-fifth of the
original population of desensitized aggregates
were beating after the second dose of TTX.
Fractions greater than 1.0 or less than 0 oc-
curred when the calculated percentage of de-
sensitized aggregates after the second dose of
TTX either exceeded the original 3-hour per-
centage or was lower than the 15 minute count
of the 3-hour incubation. (For further details
refer to the text and Fig. 47B). Each point
represents one plate containing 100-200 aggre-
gates taken from six different cultures.
that the reappearance of sensitivity might
be prevented by concentrations of the
drug below 10-5 g/ml. After the standard
3-hour desensitization with 10"5 g/ml,
TTX aggregates washed and incubated
for 4 hours with 10"8 g/ml TTX were
found to be still desensitized, as shown by
a second challenging dose of 10~5 g/ml;
in two experiments the mean fraction of
desensitized aggregates was 0.93, com-
pared with 0.05 in the control plates.
Thus, once cells are desensitized to TTX,
they do not regain the ability to respond
as long as they are exposed to the drug,
even at low doses.
From these results, McDonald, Sachs,
and DeHaan conclude that aggregates of
4-day heart cells respond to TTX in one
of three ways. Some (20%— 25%) stop
beating immediately and remain electric-
ally inactive as long as they are exposed
to the drug. In a second group (usually
about 50%), electrical activity ceases
upon exposure to the toxin, but beating
is gradually resumed over the next 2—3
hours in its continued presence. These
aggregates are shown to have been ren-
dered unresponsive to a second dose of
TTX (i.e., desensitized) . The third group
of aggregates appears to be unaffected by
TTX (up to 10"5 g/ml) and continues
beating throughout the period of expo-
sure. With increasing embryonic age a
progressively greater fraction of such ag-
gregates is sensitive to the drug, and
fewer undergo desensitization in its pres-
ence. A reasonable hypothesis to explain
these differences is that action potential
generation in sensitive aggregates in-
volves a "fast" sodium-specific channel,
while those aggregates which remain ac-
tive in the presence of TTX, as well as
those which become desensitized, utilize a
slow channel mechanism.
In last year's report, DeHaan and his
co-workers suggested that the TTX-insen-
sitive, slowly rising action potentials of
the early embryonic heart are produced by
currents carried exclusively through these
slow cation channels and that the transi-
tion with development to rapidly rising
TTX-sensitive action potentials repre-
sents the appearance in the heart cell
membrane of a fast-channel mechanism.
Desensitization of 4-day aggregates by
TTX appears to involve a similar transi-
tion.
By determining the fluxes of several
organic cations across the frog myelin-
ated sciatic nerve membrane, Bertil Hille
has recently been able to deduce certain
properties of the ion-selective channel
that carries this fast inward current. The
model he proposed for the sodium chan-
nel is of a small pore 3 X 5 A wide, lined
80
CARNEGIE INSTITUTION
with oxygens able to form hydrogen
bonds, and haying one anionic site capa-
ble of interacting with cations. The
known structure of TTX, with a posi-
tively charged guanidinium group and
hydroxyls capable of forming hydrogen
bonds, is consistent with the hypothesis
that the drug acts by bonding specifically
in the opening of the fast sodium channel.
This binding has been measured, the den-
sity of binding sites on the membrane has
been estimated, and the site has been
isolated and partially characterized — all
attesting to the reality of the channel as
a definable structure and the specificity
of its association with TTX.
Desensitization was blocked by inhibi-
tion of protein synthesis with cyclohexi-
mide, as long as the aggregates were ex-
posed to the inhibitor for a long enough
time (4 hours) before being treated with
TTX. This suggests that a pool of pro-
tein precursors is somehow involved in
the TTX desensitization and must be
exhausted before the consequences of the
inhibition of amino acid uptake can be
observed. Hartzell and Fambrough have
reported that the incorporation of acetyl-
choline receptors into tissue-cultured
skeletal muscle myotubes is also depend-
ent on protein synthesis, and that recep-
tor incorporation or activation is blocked
only after inhibition of protein synthesis
for 2 to 3 hours by cycloheximide.
It is unlikely that the prevention of
desensitization by cycloheximide re-
sulted from general deleterious effects of
that inhibitor. It is known that cyclo-
heximide does not interfere with oxygen
consumption or oxidative phosphoryla-
tion in chick cardiac cells. Moreover, ag-
gregates continued to beat vigorously
even after 24 hours in cycloheximide with
no significant changes in action potential
parameters. It remains to be resolved
whether desensitization of cardiac aggre-
gates involves the modification of exist-
ing fast channels, activation of dormant
slow channels, or synthesis and/or inser-
tion of new slow channels into the mem-
brane.
Myofibrillar Organization during the
First Beats of the Chick Heart
During the present report year, one of
several studies in which Max Springer
and DeHaan have collaborated has been
an ultrastructural analysis of the state of
myofibrillar organization in the heart at
the time it makes its first few contrac-
tions. The question addressed is whether
the heart begins to beat at a given devel-
opmental stage because it is at that time
that contractile myofibrils first become
organized. It has been known for many
years that the first beats begin in a small
region along the right margin of the form-
ing ventricle. Experiments were there-
fore designed to analyze the level of or-
ganization of myofibrils after the first
few heart beats in these areas on the right
side, at a time when the left side of the
heart tube had not begun to beat.
Chick embryos were incubated for 30-
36 hours to a stage (stage 9) just prior to
the onset of the heartbeat and were then
explanted to paper ring cultures, endo-
derm side up. The embryos were covered
with nontoxic mineral oil to prevent
evaporation of the medium and maintain
optical clarity, and transferred to a mic-
roscope warm stage (37.5 °C) for contin-
uous observation. The first pulsations of
the heart were observed and recorded
continuously with the closed-circuit video
photoelectric system described in previ-
ous reports. Each beat was displayed on
an oscilloscope and recorded on magnetic
tape. Immediately after the first spon-
taneously beating region of the heart was
identified and its contractions recorded,
the embryos were fixed rapidly and pre-
pared for electron microscopy. After the
embryos were embedded in Epon, they
were cut with a razor blade into six pieces
in order to divide the heart into three left
and three right longitudinal segments.
Each embryo fragment was mounted on
double-stick cellophane tape, re-em-
bedded in Epon, and sectioned perpendic-
ular to the long axis of the embryo. To
determine the percentage of cells contain-
DEPARTMENT OF EMBRYOLOGY
81
ing myofibrils at a given degree of differ-
entiation, the ultrathin sections were ori-
ented at low magnification, and counts
were performed directly on the screen of
the electron microscope at a primary
magnification of 30,000.
Springer and DeHaan confirmed that
contraction generally starts along the
right margin of the tubular heart at about
stage 10. By exploring the parts of the
heart with photoelectric sensors on the
video screen over the next 40—60 minutes,
they were able to observe the gradual
spread of contraction to the left side of
the ventricle and the onset of rhythmic
synchronized beats throughout the entire
primitive heart.
Examination of sections of these tubu-
lar hearts, fixed just before they had
started beating (stage 10"), showed that
the myocardium was composed of two to
three layers of cells. These cells were
characterized by abundant free ribo-
somes, mitochondria, an extensive Golgi
system, and sparse granular reticulum
distributed widely. Wavy intermediate
filaments and a few randomly oriented
thick filaments were found throughout
the whole cell, while straight thin fila-
ments seemed to be more frequent near
the cell surface. Long helical chains of
polyribosomes and randomly oriented
bundles of 4-5 microtubules ran through
the cytoplasm.
At stage 10, after the first few beats
along the right side of the heart, the num-
ber of filaments in those cells had sig-
nificantly increased. Single thick fila-
ments were abundant and were now often
surrounded by parallel thin filaments,
forming irregular but identifiable hexa-
gonal arrays. Primitive myofibrils with a
few sarcomeres were occasionally present.
The primitive sarcomeres showed a longer
Z-to-Z spacing than in older stages and
consisted of relatively few, imperfectly
aligned myofilaments.
By stage 11 (13 somites) when each
rhythmic beat involved the entire heart,
some relatively well-organized myofibrils
and bundles of thick and thin filaments
were found in virtually all cells. Z-bands
were sharper, better aligned, and more
prominent. Myofibrils contained more
but shorter sarcomeres.
In order to make a quantitative com-
parison of beating and nonbeating regions
of the heart, 17 of the 42 embryos from
which cardiac contractions were recorded
(Table 16) were cut into six pieces ac-
cording to a pattern designed so that the
right midpiece would include the tissue
that exhibited the first contractions in
each heart. Every cell in a thin section
through each of these segments was ex-
amined at high magnification (30,000X)
and tabulated in three categories: (1)
Cells without detectable thick filaments.
(2) Cells with one or more thick fila-
ments but without obvious myofibrils.
Those showing cross sections through
one or a few primitive myofibrils but
which exhibited no Z-band material were
also counted in this category. (3) Cells
with easily recognizable primitive myo-
fibrils showing distinct sarcomeres and Z-
TABLE 16. Characterization of Embryos Used for Quantitative Study
of Myofibrillar Organization
Mechanic
al Activity
Morphological Features of
the Heart
Number of
Stage
Somites
Left Side
Right Side
Embryos
io-
9
straight primitive heart tube
not yet beating
2
10
10
heart tube slightly bent to
the right side
quiescent
first contraction
i 10
11
13
distinct heart loop
both sides contract vigorously
3
10
10
inverse heart loop; heart tube
bent to the left side
first contractions
quiescent
2
82
CARNEGIE INSTITUTION
lines. These data are summarized in the
histogram of Fig. 49.
Before the hearts started to beat (stage
10") no myofibrils could be observed in
any of the six regions examined. The
right wall of the ventricle showed the
highest percentage of cells with thick fila-
ments. However, even in this region,
many cells contained no myofilaments at
all (Fig. 49-1). About one hour later,
after beating began, the percentage of
cells with thick filaments was about
double that at the previous stage. The
highest degree of differentiation (i.e., the
100%
50%
0%
100%
50 7<
A B C D E F
07c
A B C D E F
100%
HI
50%
07<
I007o
50 %
0%
Fig. 49. Quantitative comparison of six defined regions of the embryonic heart. The regions are
A rostral right side of ventricle, B middle right side, C caudal right side, D caudal left side of ven-
tricle, E middle left side, F rostal left side. The white areas represent cells without detectable
thick filaments; cross hatched areas, cells with one or more filaments, but without obvious myo-
fibrils ; dotted areas, cells with easily recognizable primitive myofibrils showing distinct sacromeres
and Z-lines. The interval of confidence of 95% indicated. I. Heart at stage 10", not yet beating.
II. Heart at stage 10; first contractions in region B. III. Heart at stage 11; both sides contract
vigorously. IV. Inverse heart loop at stage 10; the first contractions are in region E.
DEPARTMENT OF EMBRYOLOGY 83
greatest number of myofibrils) was still (histogram 49-11) , visible contraction of
found in the right wall of the ventricle, the heart may require not only a certain
coincident with the site where contrac- degree of differentiation but also a mini-
tions started. By this time primitive mal number of myofibril-containing cells,
myofibrils could be found in all six re- One of the primary questions the pres-
gions of the heart but were most frequent ent study set out to answer was whether
in the area that was already beating (Fig. the primitive heart starts to beat at the
49-IIB) . Four to five hours later, at stage stage it does because myofibrils first be-
ll, thick filaments were abundant in come organized at that time. In prin-
every cell in all six regions of the heart, ciple, heart tissue must differentiate at
Myofibrils were frequent in all areas but least three systems before it can exhibit
somewhat more abundant in the right visible contractions: (1) organized con-
ventricle (Fig. 49-IIIB, C). In the two tractile myofibrils, (2) spontaneously ac-
hearts with the situs inversus the highest tive pacemaker membranes in some or all
percentage of myofibrils and thick fila- cells, and (3) low-resistance junctional
ments were found at stage 10 in the left contacts between cells to mediate a con-
wall of the ventricle, where the first con- ducted stimulus. The nexus represents
tractions were also visible. Thus, the the morphological site for the low-resist -
histogram produced (Fig. 49-IV) is a ance pathway between cells,
mirror image of the normal (Fig. 49-11). DeHaan and Springer found that focal
When the first contractions occurred nexal junctions were already present be-
(histograms II and IV), the beating re- fore the first heartbeat could be observed
gions showed a significantly higher de- and that no structural changes or striking
gree of cytodifferentiation than the quies- differences in the distribution of the spe-
cent regions. Development in the quies- cialized junctions were observed between
cent regions was very similar to that in beating and quiescent hearts. It has also
the best-developed regions at stage 10" been reported by other workers that elec-
(histogram I) which were soon to beat, trical activity can be recorded from the
Thus, differentiation started in the right heart at stage 9-10, before active contrac-
ventricle and spread out to the other side tions begin. Therefore, the data presented
of the ventricle and to the rest of the here suggest that the final critical event at
heart according to the same pattern as stage 10 is indeed the assembly of myo-
the contractions, which also spread over filaments into functional myofibrils. A
the heart from their site of origin in the reasonable, though still tentative, conclu-
right side of the ventricle. sion seems to be that the embryonic heart
The analysis of myofilament and myo- starts beating as soon as a certain degree
fibril distribution suggests that visible of myofibrillar organization is established
contraction requires a degree of organiza- in a critical number of localized cells,
tion greater than that found in a few cells
at stage 10". The best differentiated cells Control of Synchronized Pulsation
at this stage showed scattered myofila- Rate in Cell Groups
ments and small foci of Z-body densities
similar to cells in the nonbeating areas DeHaan and Sachs had demonstrated
(A, C, D, E, and F) of stage 10 embryos, that spheroidal aggregates prepared from
Primitive myofibrils with sarcomeres, dissociated heart cells beat at rates in-
distinct Z-bands and at least poorly versely proportional to aggregate volume ;
aligned thick and thin filaments seem to that is, the bigger an aggregate, the
be the minimal condition for visible con- slower it would beat. In the past year,
traction of a heart cell. Since a few cells working with an able group of students,
which met these conditions were present DeHaan has launched a multifaceted
in the nonbeating regions of the heart analysis of this unexpected finding.
84
CARNEGIE INSTITUTION
Fig. 50. (A) Scanning electron micrograph of two synchronized cardiac myocytes in tissue cul-
ture. X 1700. (B) Contour map of cell pair shown in Fig. 50A; courtesy of Mr. F. S. Baxter, U.S.
Geological Survey. Contour interval 1.2 fiM, X 2400.
DEPARTMENT OF EMBRYOLOGY
85
Two sets of properties that might influ-
ence pulsation rate change with increas-
ing size of such aggregates: (1) those
properties that are related strictly to
number of cells in the system and are
independent of spatial configuration, and
(2) those properties that depend on spa-
tial organization of the parts of the
system. In the first category would be
such characteristics as total aggregate
volume, total amount of cell membrane,
and total aggregate capacitance. The
second category of properties would in-
clude ratio of external surface to volume,
and intercellular diffusional considera-
tions.
To perform experiments designed to
distinguish between these broad classes of
rate-controlling phenomena, DeHaan has
worked out methods for manipulating
spontaneously beating aggregates one at
a time under direct visual observation.
When brought into contact, aggregates
beating at different rates adhere and soon
(usually within 30—60 minutes) synchro-
nize to a common rhythm. Thus, pairs or
groups of aggregates of selected rates and
sizes can be joined to form systems of any
desired spatial configuration and allowed
to synchronize.
According to the data of Sachs and
DeHaan small spherical aggregates (1 X
105 yu,3 in volume) have a mean pulsation
rate of about 160 B/M, whereas aggre-
gates 20 times that size (2 X 10° /x3) beat
only half as fast. Is the difference in rate
in this case due to the 20-fold difference
in total cell number, or does it result from
other factors only secondarily related to
volume, such as surface-to-volume (S/V)
ratio? One approach to this question was
carried out by Linda Moyzis, who built
up linear chains of 6-20 aggregates and
demonstrated that the synchronous pul-
sation rate of these chains was only
slightly reduced from the average rate of
the component aggregates prior to syn-
chronization. In this case, the rate was
not a function of total synchronized cell
number.
A second approach was taken in experi-
ments performed by Caroline Krueger.
When sheets of cells were prepared at
known densities (in cells/mm2 on the
plate) and cut into rectangles of different
areal sizes, Krueger found that the syn-
chronous pulsation rates of the rectangles
did not differ from each other in a sys-
tematic way over a 60-fold range of areas
(i.e., number of synchronized cells).
Again, rate was found not to be a function
of synchronized cell number. However,
when sheets were prepared at two differ-
ent thicknesses (4 cells or 6 cells deep) ,
the thicker rectangles beat only half as
fast as the thin ones. The controlling fac-
tor appeared to be not cell number but
diffusional restriction between the inside
of the mass and the outside, or S/V ratio.
This possibility was tested directly in a
series of experiments carried out by
Thomas E. Durr. Joining pairs of aggre-
gates of different sizes (and therefore of
different rates) Durr determined that the
change in rate experienced by each mem-
ber of the pair upon synchronization con-
formed to that predicted by the new S/V
ratio rather than to the total volume of
the pair.
The consistency of this finding using
tissue-like multicellular systems has en-
couraged DeHaan and Springer to apply
innovative methods to ask whether at the
single-cell level, pulsation rate is also
controlled by S/V ratio. In collaboration
with Durr, the independent pulsation
rates of pairs of single cells have been
recorded photoelectrically before and im-
mediately after synchronization. The
cells were then fixed and prepared for
scanning electron microscopy, and stereo-
scopic scanning EM photographs were
taken of each pair (Fig. 50A) . Using
photogrammetric techniques (with the
aid of Messrs. F. Baxter, F. Doyle, and
H. Loving of the U.S. Geological Survey)
contour maps of the cell surfaces have
been drawn (Fig. 50B) on a Kern plotter
(model PG2). From these contour maps,
surface areas and volumes of the cells can
be determined with a computerized digit-
izer. At the time of this writing, these
86
CARNEGIE INSTITUTION
calculations are in progress, and thus it is
too early to report the results. If the
hypothesis is borne out, however, it may
lead to a profound generalization con-
cerning the rate-setting mechanism in
heart cells.
Cytochalasin B and Embryonic Heart
Muscle: Contractility, Excitability,
and ultrastructure
M. Springer, T. F. McDonald, and H. G. Sachs
The effect of cytochalasin B (CB) on
spontaneous contraction, electrical activ-
ity, and cytological ultrastructure of em-
bryonic hearts and tissue-cultured heart
preparations was investigated. The in-
hibition of spontaneous pulsation and the
subsequent recovery after drug washout
were found to be dose dependent and to
be a function of the type of preparation
used. In the presence of CB, spontaneous
action potentials can be recorded from all
three (intact hearts, isolated myocytes,
reaggregated cells) after cessation of all
visible contraction, although the action
potential duration is reduced. CB at 0.5
to 10 /Ag/ml was found neither to elec-
f ( r t *. »
^vipf;.
■;- •■*:>■■■■
■ . *
;k , L
m
,....,-. ...
: ■ - ;> ■ ■ : : . ....■: \- ... .
Fig. 51. (A) Electron micrograph of 7-day heart cell aggregate. After 24 hours in culture, CB 2
jug/ml was added for 30 hours. The myofibrils are disrupted in bundles of thick and thin filaments
of approximately sarcomere-length, scattered throughout the cytoplasm. Dense clumps of Z-mate-
rial (Z) seem to be out of alignment with the bundles. X 42,000, calibration bar 0.5 /xM. (B) Con-
trol. Myofibrils with well-developed sarcomeres in an aggregate after 24 hours in culture. X 30,000,
calibration bar 0.5 /xM .
DEPARTMENT OF EMBRYOLOGY
87
trically uncouple heart cells nor to pre-
vent cells from aggregating and estab-
lishing new electrical coupling between
cells. In all three systems, however, CB
does result in disruption of myofibrillar
organization. After 30 hours in cytocha-
lasin B (2 /xg/ml) bundles of myofila-
ments are found to be randomly oriented,
with Z-band material condensed into
amorphous bodies, no longer aligned with
the myofibril but remaining in register
between the A bands (Fig. 51). Within
the bundles the normal hexagonal cross
section of thick and thin filaments is
maintained. The unaligned Z bodies,
however, demonstrate attached thin fila-
ments of very reduced length. Following
washout of the drug, normal sarcomeric
organization is restored in parallel with
resumption of spontaneous beating. The
present data, therefore, are consistent
with the hypothesis that one major effect
of CB is an alteration in myofibril (myo-
filament) stability and that this effect
alone is sufficient to account for the ob-
served inhibition of contractility.
THE COLLECTION OF HUMAN EMBRYOS
R. O'Rahilly and E. Gardner
Owing to the imminent move of the
Carnegie Collection, the material has un-
fortunately been "in mothballs" for most
of the year. However, it is now scheduled
to leave Baltimore during July 1973* for
its new home in the University of Cali-
fornia School of Medicine, Davis, Cali-
fornia 95616. Professor Ronan O'Rahilly
will continue his stewardship as Curator
of the Collection; he has been joined by
Professor Ernest Gardner as Associate
Curator. It is anticipated that the Col-
lection will be "open for business" by
October 1, 1973. It will once again be
available to qualified investigators. In-
quiries may be directed to either Profes-
sor O'Rahilly, whose address is Wayne
State University, Scott Hall of Basic
Medical Sciences, 540 East Canfield Ave-
nue, Detroit, Michigan 48201, or to Pro-
fessor Gardner in the Department of
Human Anatomy at Davis.
Developmental Stages in
Human Embryos
The famous Blechschmidt Collection of
human embryos, housed in the Anatom-
ical Institute of the University of Gott-
ingen, has been given Carnegie numbers
(10,304 to 10,434). Professor Blech-
* Note added in proof : The move was com-
pleted as scheduled.
schmidt is in the process of assigning
these embryos to their appropriate de-
velopmental stages. Because of the ex-
istence of a particularly important col-
lection of reconstructions made over
many years under Professor Blech-
schmidt's supervision, it is believed that
a joint cataloguing of staged embryos
will be advantageous. The Blechschmidt
reconstructions cover 22 embryos and one
fetus (40 mm).
O'Rahilly's account of stages 1 to 9,
that is, embryos of the first three weeks, is
now in press. In addition to the detailed
descriptions of staged embryos, the
monograph will include a specimen index,
75 photographs and drawings, and a bib-
liography of some 325 references.
Although the first three stages (1, uni-
cellular stage; 2, stage of segmentation;
3, free blastocyst) follow Streeter's de-
velopmental horizons I to III, the pro-
posed stages 4 to 9 do not correspond to
horizons IV to IX, respectively. Stages
4 and 5 cover implantation, and stage 6
is characterized by the initial appearance
of chorionic villi, both branched and un-
branched. Stage 7 is marked by the pres-
ence of the notochordal process, and stage
8 by one or more of the following: primi-
tive pit, notochordal canal, neurenteric
canal. The neural folds form during
stage 8. Stage 9 is characterized by the
88
CARNEGIE INSTITUTION
appearance of one to three pairs of
somites.
Progress continues in the tabulation
of the timing and sequence of develop-
mental events in staged human embryos.
Data for the heart, nervous system, and
urinary system have already been pub-
lished. The limbs (O'Rahilly and Gard-
ner) and the respiratory system
(O'Rahilly and Boyden) are now being
readied for publication.
At the end of the embryonic period
proper (stage 23), the human embryo is
approximately 30 mm in C.-R. length and
is about 8 postovulatory weeks in age. It
is surprising that articles and books are
still appearing in which an age (7 weeks)
based on macaque data is being cited.
Development of the Conducting System
of the Heart
Gardner and O'Rahilly are studying
the nerve supply of the human heart and
the conducting system at stage 23 (8
postovulatory weeks).
It has been found that the nerves on
the right-hand side arise from cervical
sympathetic and from cervical and tho-
racic vagal filaments. Out of their inter-
connections a single vagosympathetic
nerve emerges, which (1) sends a branch
in front of the trachea to the aorticopul-
monary ganglion, and (2) as the right
sinal nerve, supplies the sinu-atrial and
atrioventricular nodes and pulmonary
veins.
The nerves on the left-hand side arise
similarly from cervical sympathetic and
from cervical and thoracic vagal filaments.
The cervical filaments form a vagosym-
pathetic nerve, which descends to the
right of the arch of the aorta and enters
the aorticopulmonary ganglion. Fila-
ments leaving the ganglion supply the
pulmonary trunk, ascending aorta, inter-
atrial septum, pulmonary veins, and, as
the left sinal nerve, the fold of the left
vena cava. The thoracic vagal filaments
descend to the left of the arch of the aorta
and supply chiefly the arterial end of the
heart.
No thoracic sympathetic cardiac fila-
ments were found.
The sinu-atrial node is a crescentic
mass in front of the lower part of the
superior vena cava. The left limb of the
node curves downward as the anterior
internodal tract. The right limb curves
downward behind the vena cava and, to-
gether with the crista terminalis, forms
the middle and posterior internodal
tracts. An additional internodal connec-
tion by way of the left venous valve was
observed.
The atrioventricular node is a conspic-
uous mass in the anterior and lower part
of the interatrial septum, from which a
clearly defined bundle leaves to enter the
interventricular septum. Right and left
limbs were observed, the former being a
rounded bundle that passes immediately
in front of the root of the aorta.
Development of the Nervous System
Work by Gardner and O'Rahilly on the
embryonic brain continues, and O'Rahilly
has studied the early development of the
hypophysis cerebri.
The site of the craniopharyngeal in-
vagination is detectable immediately
rostral to the oropharyngeal membrane
at stage 10. The overlying neuroecto-
derm is the infundibular area, and the
hypophysis arises at a single locus, not
from two distinct evaginations which ap-
proach each other and fuse. At stage 13
the basement membrane of the cranio-
pharyngeal evagination and that of the
brain are clearly in contact. The cranio-
pharyngeal pouch becomes elongated by
stage 14, and a prechordal strand, rather
than the notochord itself, is inserted into
its dorsal wall. From stage 14 onward,
blood vessels grow in between the afore-
mentioned basement membranes (c/. the
lentiretinal space) and are later incor-
porated in the stroma (stage 20). At
DEPARTMENT OF EMBRYOLOGY
89
stage 15 mitotic figures are common in
both walls of the pouch and are located
next to the cavity. The infundibular re-
cess, a slight indication of which may be
seen in some embryos at stage 16, dis-
plays a characteristically folded wall,
namely, the neurohypophysis, by stage
17. The juxtacerebral wall of the cranio-
pharyngeal pouch is the thicker. The
lateral lobes (future infundibular, or
tuberal, part) and the anterior chamber
(Vorraum) are clearly visible by stage 17.
The adenohypophysial portion of the
pouch expands during stage 18, when its
pharyngeal stalk may become closed.
The adenohypophysial epithelium adja-
cent to the neurohypophysis constitutes
the beginning pars media by stage 20.
Development of the Larynx
Studies by O'Rahilly and Tucker on
laryngeal differentiation continue, and a
preliminary account was published dur-
ing the year. An exhibit on the embryol-
ogy of the human larynx was prepared
for the annual meeting of the American
Academy of Ophthalmology and Oto-
laryngology, 1972, and a silver medal was
awarded to Tucker and O'Rahilly "for
excellence of presentation, originality of
work, and teaching value."
Development of Bones and Joints
An article on early ossification and an
abstract on the development of the shoul-
der joint appeared during the year.
STAFF ACTIVITIES
A Symposium on Developmental As-
pects of Membrane Biology was held at
the Laboratory on April 12 and 13, 1973.
Organized by Robert DeHaan and Doug-
las Fambrough, the program focused on
four major topics: structure and dynam-
ics of cell membranes, electrical coupling
between cells, physiological differentia-
tion of membranes, and interaction be-
tween cell surface and genome. The roster
of participants included Michael Aber-
crombie (Strangeways Research Labora-
tory), Max Burger (Biozentrum der Uni-
versitat, Basel), Philippa Claude (Harvard
Medical School), Philip Dunham (Syra-
cuse Urliversity) , G. M. Edelman (Rocke-
feller University), Alan Finkelstein (Al-
bert Einstein College of Medicine),
Norton Gilula (Rockefeller University) ,
Victor Ling (The Ontario Cancer Insti-
tute), Philip Nelson (National Institutes
of Health) , J. R. Sheppard and Judson
Sheridan (both of the University of Min-
nesota), S. J. Singer (University of Cali-
fornia at San Diego) , and Hermann Wolf
(Dalhousie University) . A number of
colleagues at The Johns Hopkins Univer-
sity also played leading roles, among
them L. Brand, R. A. Cone, Pedro Cuatre-
casas, M. Edidin, M. Jacobson, R, T.
Johnson, W. Kundig, Saul Roseman, and
Stephen Roth.
International conferences and symposia
in which members of the Department
participated during the year included the
following: Fourth International Confer-
ence "De la Physique Theorique a la
Biologie" (Versailles, France) ; Japanese
Society for Developmental Biology
(Tokyo) ; Fourth Lepetit Colloquium on
the Search for Episomes in Eucaryotic
Cells (Cocoyoc, Mexico) ; Eighth Meet-
ing of the Federation of European Bio-
chemical Societies (Amsterdam) ; Work-
shop of the European Molecular Biology
Organization (Paris) ; and the Fourth In-
ternational Congress of Biophysics (Mos-
cow).
Among the other conferences engaging
the interest of members of the group, the
following should be mentioned : a Sympo-
sium on Synaptic Transmission and Neu-
ronal Interaction (Woods Hole) ; one deal-
ing with Trophic Functions of the Neuron
(New York) , and two dealing with recep-
tors: the American Society for Neuro-
chemistry's Symposium on Cholinergic
Receptors and a related program on the
Current Status of Pharmacological Re-
ceptors, sponsored by the Amercian So-
90
CARNEGIE INSTITUTION
ciety of Pharmacology and Experimental
Therapeutics. The staff found several
Gordon Conferences attractive, among
them conferences on developmental biol-
ogy, heart muscle, molecular pathology,
and nucleic acids. The annual symposia
at Cold Spring Harbor and Gatlinburg
included speakers from the Department,
as did both the University of Minnesota's
program on Repetitive DNA, Chromo-
some Defects and Neoplasia, and the
University of Colorado's Molecular Neu-
robiology Symposium (Given Institute of
Pathobiology, Aspen) .
Lectures were presented on a number
of campuses, including Case Western Re-
serve University, Columbia University,
Cornell University, Duke University,
Georgetown University, Harvard Med-
ical School, Harvard University, Indiana
State University, Kansai Medical School,
Kyoto University, Kyushu University,
Nagoya University, Ohio State Univer-
sity, Rensselaer Polytechnic Institute,
Rockefeller University, Rutgers Uni-
versity, Temple University, Tohoku
University, Tokyo Metropolitan Uni-
versity, The Universities of Cali-
fornia (Berkeley, San Diego, Santa
Barbara), Chicago, Cincinnati, Connecti-
cut, Georgia, Oregon, Osaka, Paris,
Tokyo, Washington, Wisconsin and Wurz-
burg, Vanderbilt University, Wheaton
College, Washington University School of
Medicine, and Yale University.
Special presentations included the
American Institute of Biological Sciences
Silver Anniversary Lecture (Pennsyl-
vania State University), the Carter-
Wallace Lectures (Princeton University)
the Centennial Lecture (Upstate Medical
Center, State University of New York) ,
the Distinguished Scientist Lecture (Tu-
lane University) , the H. V. Wilson Lec-
ture (University of North Carolina), and
the Tamaki Memorial Lecture (Kyoto
University) .
Members of the group also spoke at a
number of hospitals and research centers,
including the Boston University Marine
Program; Max-Planck Institut, Tubin-
gen, Germany; M. D. Anderson Hospital;
National Cancer Center, Tokyo; Na-
tional Institutes of Health; Roche Insti-
tute for Molecular Biology ; Roswell Park
Memorial Institute ; and Scripps Institute
for Medical Research.
Members of the Department took part
in meetings of a number of learned so-
cieties, including, in addition to those al-
ready mentioned, the American Associa-
tion of Anatomists, American Association
for the Advancement of Science, Amer-
ican Society of Biological Chemists,
American Society for Cell Biology, Amer-
ican Society of Zoologists, Biophysical
Society, Federation of American Societies
for Experimental Biology, Japanese So-
ciety for Molecular Biology, Maryland
Academy of Sciences, National Academy
of Sciences, New York Heart Association,
Society for Developmental Biology, So-
ciety of General Physiologists, Society
for Neuroscience, and Tissue Culture
Association.
Advisory and consultative services in-
cluded membership on the editorial
boards of Anales del Desarrollo, Devel-
opmental Biology, In Vitro, Journal of
Biological Chemistry, Journal of Cell
Biology, Journal of Embryology and Ex-
perimental Morphology, Journal of Ex-
perimental Zoology, Excerpta Medica
(section on Human Developmental Biol-
ogy), Current Topics in Developmental
Biology, and Quarterly Review of Biol-
ogy.
Members of the staff also acted in
these capacities: Member of the Board of
Scientific Overseers, Jackson Laboratory;
Trustee, President, and Director, Marine
Biological Laboratory; Member of the
Corporation, Woods Hole Oceanographic
Institution; Vice President, Maryland
Academy of Sciences ; and Member of the
Executive Committee for the Fourth In-
ternational Conference on Congenital
Malformations.
Other posts occupied by members of
the Department include the following: in
the American Institute of Biological Sci-
ences, member, Council of Past Presi-
DEPARTMENT OF EMBRYOLOGY
91
dents; in the American Society for Cell
Biology, member of the Council; in the
Institut de la Vie, Chairman, World
Committee on the Formative Weeks of
Human Life ; in the International Society
of Developmental Biologists, Secretary;
in the National Academy of Sciences,
member, Committee on Science and Pub-
lic Policy, and member, Report Review
Committee; in the National Institutes of
Health, member, Board of Scientific
Counselors, National Institute of Child
Health and Human Development, mem-
ber, Gerontology Research Center Re-
sources Advisory Committee, member,
Cell Biology Study Section, and member,
Physiological Chemistry Study Section;
at the National Science Foundation,
member of the Advisory Committee for
Biological and Medical Sciences; in the
Society for Developmental Biology, mem-
ber of Executive Committee; and in the
Tissue Culture Association, Chairman,
Honor B. Fell Division. Members of the
Department served on a number of visit-
ing committees including those for the
Departments of Biology at Brookhaven
National Laboratory, Harvard Univer-
sity, Massachusetts Institute of Technol-
ogy, and Princeton University.
Formal teaching was again largely con-
fined to the Johns Hopkins Department
of Biology, but during the year lectures
were offered in other departments of the
University as well, among them Biophys-
ics, Engineering, Laryngology and Otol-
ogy, and Pediatrics.
Seminars
The roster of speakers at the seminars
organized by the Department to serve all
those working in developmental biology
in the region included Joyce Benjamins
(Johns Hopkins University School of
Medicine) ; George Brownlee (Medical
Research Council, Cambridge, England) ;
Annette W. Coleman (Brown Univer-
sity) ; Norman Davidson (California In-
stitute of Technology) ; Gary Felsenfeld
(National Institutes of Health) ; Richard
Gelinas (Harvard University) ; K. Gross
(Massachusetts Institute of Technol-
ogy) ; Richard Hallberg (Cornell Univer-
sity); Richard Hendler (National Insti-
tutes of Health) ; W. Hennig (Zurich) ;
Thomas Hollinger (Purdue University) ;
Philip Leder (National Institutes of
Health) ; Harvey Lodish (Massachusetts
Institute of Technology) ; John J. Mar-
chalonis (Walter and Eliza Hall Institute
of Medical Research, Melbourne, Austra-
lia) ; Reiji Okazaki (Nagoya Univer-
sity) ; Ernest Page (University of Chi-
cago) ; John Paul (Beatson Institute for
Cancer Research, Glasgow, Scotland) ;
and Allan Schneider (National Institutes
of Health).
BIBLIOGRAPHY
Barkley, D. S., see Edds, M. V., Jr.
Bladder, A. W., see Dawid, I. B.
Brown, D. D., The evolutionary solution to the
antibody dilemma. In Molecular Genetics
and Developmental Biology, M. Sussman ed.,
Prentice-Hall, Englewood Cliffs, New Jersey,
pp. 101-125. 1972.
Brown, D. D., see also Gage, L. P.; Sugimoto,
K.; Suzuki, Y.
Dawid, I. B., Evolution of mitochondrial DNA
sequences in Xenopus. Develop. Biol., 29,
139-151, 1972.
Dawid, I. B., and A. W. Blackler. Maternal and
cytoplasmic inheritance of mitochondrial
DNA in Xenopus. Develop. Biol., 29, 152-161,
1972.
Dawid, I. B., Mitochondrial protein synthesis.
In Mitochondria I Biomembranes, S. G. van
den Bergh et al., eds. North Holland. Amster-
dam, pp. 35-51, 1972.
Dawid. I. B., see also Wu. G.-J.
DeHaan, R. L., and H. G. Sachs, Cell coupling
in developing systems: the heart-cell para-
digm. Current Topics in Develop. Biol., 7,
193-228, 1972.
DeHaan, R. L., Embryology of the heart. In
The Heart, J. W. Hurst and R. B. Logue, eds.,
McGraw-Hill, New York, pp. 7-21. 1973.
DeHaan, R. L., see also McDonald, T. F. ;
Nason, A.; Sachs, H. G.
Ebert, J. D., Report of the Director of the Ma-
92
CARNEGIE INSTITUTION
rine Biological Laboratory for 1971. Biol.
Bull, US, 7-13.
Ebert, J. D., The molecular basis of gene ex-
pression during devlopment (in Russian).
Yearbook, Science and Humanity, 1971-1972,
pp. 240-253.
Ebert, J. D., Conclusions. Strategie Institution-
nelle. Les Cahiers de L'Institut de la Vie,
number 36, 297-312, 1973.
Ebert, J. D., A. G. Loewy, R. S. Miller, and
H. A. Schneiderman, Biology, Holt, Rinehart
and Winston, Inc. New York, 798 pp., 1973.
Ebert, J. D., see also McDonald, T. F.
Edds, M. V., Jr., D. S. Barkley, and D. M. Fam-
brough, Genesis of neuronal patterns. Neuro-
sci. Res. Prog. Bull, 10, 254-367, 1972.
Edidin, M., and D. Fambrough, Fluidity of the
surface of cultured muscle fibers: rapid lateral
diffusion of marked surface antigens. J. Cell
Biol, 57, 27-37, 1973.
Fambrough, D. M., see Edds, M. V., Jr.; Edidin,
M.; Hartzell, H. C; Rash, J. E.
Gage, L. P., Y. Suzuki, and D. D. Brown, Spe-
cific hybridization of the silk fibroin genes in
Bombyx mori. In Molecular Genetics and
Developmental Biology, M. Sussman, ed.,
Prentice-Hall, Englewood Cliffs, New Jersey,
pp. 127-137, 1972.
Gage, L. P., see also Suzuki, Y.
Gardner, E., The early development of the
shoulder joint in staged human embryos.
Anat. Rec, 175, 503, 1973.
Gardner, E., see also O'Rahilly, R.
Gershfeld, N. L., and R. E. Pagano, Physical
chemistry of lipid films at the air-water inter-
face. I. Intermolecular energies in single-
component lipid films. /. Phys. Chem., 76,
1231-1237, 1972.
Gershfeld, N. L., and R. E. Pagano, Physical
chemistry of lipid films at the air-water inter-
face. III. The condensing effect of cholesterol.
A critical examination of mixed-film studies.
/. Phys. Chem., 76, 1244-1249, 1972.
Gershfeld, N. L., see also Pagano, R. E.
Hartzell, H. C, and D. M. Fambrough, Acetyl-
choline receptors: distribution and extrajunc-
tional density in rat diaphragm after denerva-
tion correlated with acetylcholine sensitivity.
J. Gen. Physiol, 60, 248-262, 1972.
Hartzell, H. C, and D. M. Fambrough, Acetyl-
choline receptor production and incorporation
into membranes of developing muscle fibers.
Develop. Biol, 30, 153-165, 1973.
Honjo, T., and R. H. Reeder, Xenopus mulleri
ribosomal genes are repressed in X. laevis X
X. mulleri hybrid frogs. Fed. Proc, Fed.
Amer. Soc. Exp. Biol, 32, 656 (Abstract), 1973.
Loewy, A. G., see Ebert, J. D.
McDonald, T. F., and R. L. DeHaan, Ion levels
and membrane potential in chick heart tissue
and cultured cells. J. Gen. Physiol, 61, 89-109,
1973.
McDonald, T. F., H. G. Sachs, C. W. Orr, and
J. D. Ebert, Multiple effects of ouabain on
BHK cells. Exp. Cell Res., 74, 201-206, 1972.
McDonald, T. F., see also Sachs, H. G.
Miller, I. R., see Pagano, R. E.
Miller, R. S., see Ebert, J. D.
Muecke, E. C, see O'Rahilly, R.
Nason, A. and R. L. DeHaan, The Biological
World, John Wiley and Sons, New York, 736
pp. 1973.
Nojima, T., see Yoshikawa-Fukada, M.
O'Rahilly, R., Guide to the staging of human
embryos. Anat. Anz., 130, 556-559, 1972.
O'Rahilly, R., The early development of the
hypophysis cerebri in staged human embryos.
Anat. Rec, 175, 511, 1973.
O'Rahilly, R., and E. Gardner, The initial ap-
pearance of ossification in staged human em-
bryos. Amer. J. Anat., 134, 291-307, 1972.
O'Rahilly, R., and E. C. Muecke, The timing
and sequence of events in the development of
the human urinary system during the embry-
onic period proper. Z. Anat. Entwicklungs-
gesch., 138, 99-109, 1972.
O'Rahilly, R., see also Tucker, J. A.
Orr, C. W., see McDonald, T. F.
Ozawa, E., The role of calcium ion in avian
myogenesis. Biol. Bull, 143, 431-439, 1972.
Pagano, R. E., and N. L. Gershfeld, Physical
chemistry of lipid films at the air-water inter-
face. II. Binary lipid mixtures. The princi-
ples governing miscibility of lipids in surfaces.
/. Phys. Chem., 76, 1238-1243, 1972.
Pagano, R. E., and N. L. Gershfeld, A Milidyne
film balance for measuring intermolecular
energies in lipid films, /. Colloid Interface
Sci., 41, 311-317, 1972.
Pagano, R. E., J. M. Ruysschaert, and I. R.
Miller, The molecular composition of some
lipid bilayer membranes in aqueous solution.
/. Membrane Biol. 10, 11-30, 1972.
Pagano, R. E., and T. E. Thompson, Spherical
lipid bilayers: Surface tension and contact
angle measurements. /. Colloid Interface Sci.,
43, 209-210, 1973.
Pagano, R. E., see also Gershfeld, N. L.
Polinger, I. S., Identification of cardiac myo-
cytes in vivo and in vitro by the presence of
glycogen and myofibrils. Exp. Cell Res., 76,
243-252, 1973.
Polinger, I. S., Growth and DNA synthesis in
embryonic chick heart cells, in vivo and in
vitro. Exp. Cell Res., 76, 253-262, 1973.
DEPARTMENT OF EMBRYOLOGY
93
Rash, J. E., and D. Fambrough, Ultrastructural
and electrophysiological correlates of cell
coupling and cytoplasmic fusion during myo-
genesis in vitro. Develop. Biol., 30, 166-186,
1973.
Reeder, R. H., Transcription of chromatin by E.
coli RNA polymerase. Biophys. J., 13, 189a,
1973.
Reeder, R. H., and R. G. Roeder, Ribosomal
RNA synthesis in isolated nuclei. J. Mol.
Biol, 67, 433-441, 1972.
Reeder, R. H., see also Honjo, T.
Roeder, R. G., RNA polymerases during am-
phibian development, In Molecular Genetics
and Developmental Biology, M. Sussman, ed.,
Prentice-Hall, Englewood Cliffs, New Jersey,
pp. 163-172. 1962.
Roeder, R. G., see also Reeder, R. H.
Ruysschaert, J. M., see Pagano, R. E.
Sachs, H. G., and R. L. DeHaan, Embryonic
myocardial cell aggregates: Volume and pulsa-
tion rate. Devel. Biol, 30, 233-240, 1973.
Sachs, H. G., and T. F. McDonald, Membrane
potentials of BHK (Baby Hamster Kidney)
cell line: Ionic and metabolic determinants.
J. Cell Physiol, 80, 347-358, 1972.
Sachs, H. G., T. F. McDonald, and R. L. De-
Haan, Tetrodotoxin sensitivity of cultured
embryonic heart cells depends on cell inter-
actions. /. Cell Biol, 56, 255-258, 1973.
Sachs, H. G., M. Springer, and T. F. McDonald,
Cytochalasin B and embryonic heart cells.
Anat. Rec, 175, 433 (Abstract), 1973.
Sachs, H. G., see also DeHaan, R. L.; Mc-
Donald, T. F.
Schneiderman, H., see Ebert, J. D.
Springer, M., see Sachs, H. G.
Sugimoto, K., and D. D. Brown, Evolution of a
family of genes: Comparison of 5S DNA
from Xenopus laevis and Xenopus mulleri.
Fed. Proc, Fed. Amer. Soc. Exp. Biol, 32, 663
(Abstract), 1973.
Suzuki, Y., The messenger RNA and the genes
for silk fibroin — A model system for the study
of differential gene action in eukaryotes. Ab-
stracts of papers presented at the 1972 Molec-
ular Biology Meeting of Japan, pp. 62-63,
1972.
Suzuki, Y., Studies on the mechanisms of gene
expression in development. I. Isolation and
identification of the messenger RNA for silk
fibroin from Bombyx mori. Protein, Nucleic
Acid and Enzyme, 17, 79-91, 1972.
Suzuki, Y., The genes in development, Science
oj Japan, 42, 590-599, 1972.
Suzuki, Y., Studies on the mechanisms of gene
expression in development. II. The genes for
silk fibroin in Bombyx mori. Protein, Nucl
Acid, Enzyme, 18, 673-685, 1973.
Suzuki, Y., L. P. Gage, and D. D. Brown, The
genes for silk fibroin in Bombyx mori. J. Mol.
Biol, 70, 637-649, 1972.
Thompson, T. E., see Pagano, R. E.
Tucker, J. A., and R. O'Rahilly, Observations
on the embryology of the human larynx. Ann.
Otol Rhinol Laryngol, 81, 520-523, 1972.
Wu, G.-J., and I. B. Dawid, Purification and
properties of mitochondrial deoxyribonucleic
acid dependent ribonucleic acid polymerase
from ovaries of Xenopus laevis. Biochemistry,
11, 3589-3595, 1972.
Yoshikawa-Fukada, M., and T. Nojima, Bio-
chemical characteristics of normal and virally
transformed mouse cell lines. J. Cell Physiol,
80, 421-430, 1972.
PERSONNEL
Year Ended June 30, 1973
(including those whose services began or ended during the year)
Research Staff
Donald D. Brown, Biochemistry
Igor B. Dawid, Biochemistry
Robert L. DeHaan, Experimental Embry
ology
James D. Ebert, Director
Douglas M. Fambrough, Biochemistry
Richard E. Pagano, Biophysics
Ronald H. Reeder, Biochemistry
Assistant Investigator
Peter J. Stambrook
Research Associates (extramural)
Bent G. Boving, Detroit, Mich.
Louis B. Flexner, Philadelphia, Pa.
Ernest Gardner, Davis, Cal.
Arthur T. Hertig, Boston, Mass.
Irwin R. Konigsberg, Charlottesville, Va.
H. Robert Misenhimer, Chicago, 111.
Ronan O'Rahilly, Detroit, Mich.
Fellows
Joseph F. Albright, Senior Fellow of Car-
negie Institution of Washington
94
CARNEGIE INSTITUTION
Dana Carroll, Fellow of Carnegie Institu-
tion of Washington
Robert J. Hay, Fellow of Carnegie Institu-
tion of Washington
Tasuku Honjo, Fellow of Carnegie Institu-
tion of Washington
Ivan Horak, Fellow of Carnegie Institution
of Washington
Paul Lizardi, Fellow of Jane Coffin Childs
Memorial Fund for Medical Research
Terence F. McDonald, Fellow of Carnegie
Institution of Washington
Keiko Ozato, Fellow of Carnegie Institution
of Washington
Aileen K. Ritchie, Fellow of Muscular Dys-
trophy Association of America, Inc.
Howard G. Sachs, Fellow of Carnegie Insti-
tution of Washington
Kazunori Sugimoto, Fellow of Carnegie In-
stitution of Washington
Yoshiaki Suzuki, Fellow of Carnegie Insti-
tution of Washington
Peter K. Wellauer, International Fellow of
the U.S. Public Health Service
Robert Williamson, Senior Fellow of Car-
negie Institution of Washington
Guang-Jer Wu, Fellow of Carnegie Institu-
tion of Washington
Masako Yoshikawa-Fukada, Fellow of Car-
negie Institution of Washington
Students
Sandra L. Biroc, Graduate, Johns Hopkins
University
Peter Devreotes, Graduate, Johns Hopkins
University
Thomas E. Durr, Undergraduate, Johns
Hopkins University
Scott Gilbert, Graduate, Johns Hopkins
University
H. Criss Hartzell, Graduate, Johns Hopkins
University
Carol Kaushagen, Graduate, Johns Hopkins
University
Caroline R. Krueger, Undergraduate, Johns
Hopkins University
Dennis E. Leister, Graduate, Johns Hop-
kins University
Linda Moyzis, Undergraduate, Johns Hop-
kins University
Calvin E. Plitt, Undergraduate, Johns Hop-
kins University
Ralph H. Stern, Graduate, Johns Hopkins
University
Camilla Velez, Undergraduate, Johns Hop-
kins University
Visiting Investigators and Extramural
Collaborators
Antonie W. Blackler, Ithaca, N.Y.
George Brownlee, Cambridge, England
John W. Chase, Boston, Mass.
A. Chikushi, Tokyo, Japan
Hayden G. Coon, Bethesda, Md.
Daniel B. Drachman, Baltimore, Md.
Michael Edidin, Baltimore, Md.
J. W. S. Harris, London, England
Edward C. Muecke, New York, N.Y.
T. Nojima, Kyoto, Japan
Mary Lou Pardue, Cambridge, Mass.
Robert Roeder, St. Louis, Mo.
Glenn C. Rosenquist, Baltimore, Md.
S. Satyamurti, Baltimore, Md.
J. E. Sisken, Lexington, Ky.
Max Springer, Zurich, Switzerland
Kenneth D. Tartof, Philadelphia, Pa.
John A. Tucker, Philadelphia, Pa.
Ifermann K. Wolf, Halifax, Nova Scotia,
Canada
Clerical and Technical Staff
Aldon L. Alston, Recorder
Elaine S. Asch, Technician
James H. Blackwell, Custodian
Julia L. Certain, Librarian (part time)
William J. Cleary, Recorder
Laura L. Costello, Technician
William H. Duncan, Senior Technician
Ernestine V. Flemmings, Laboratory Helper
Richard D. Grill, Photographer
Richard F. Handler, Technician
Norman Handy, Custodian
Virginia S. Hicks, Laboratory Helper
James B. Iglehart, Recorder
Eddie D. Jordan, Senior Technician
Nancy J. Joseph, Senior Technician
June M. Lam, Secretary
Alice H. Mabin, Laboratory Helper
Thomas F. Malooly, Business Manager
Thomas F. Miller, Custodian
Joyce I. Oaks, Laboratory Helper
Alex Ovcharek, Custodian
John Pazdernik, Jr., Building Engineer
Betty Lou Phebus, Bookkeeper-Clerk
Diane M. Prather, Technician
Margaret J. Proctor, Secretary
Martha L. Rebbert, Technician
Adrienne Robinette, Secretary
Phillip E. Rose, Custodian
Carroll Sears, Custodian
DEPARTMENT OF EMBRYOLOGY 95
John J. Sexton, Custodian Student Assistants
Bessie H. Smith, Laboratory Helper Jan c Culbertson, Southampton College
Delores V. Somerville, Technician Andrew Scott Klein, Duke University
Leroy G. Williams, Recorder Nancy Kolzak, Smith College
John L. Wiser, Machinist Andrew Cader, Park School
Hale Observatories
Operated by Carnegie Institution of Washington
and California Institute of Technology
Pasadena, California
Horace W. Babcock
Director
J. Beverley Oke
Associate Director
OBSERVATORY COMMITTEE
Horace W. Babcock, Chairman
J. Beverley Oke, V ice-Chairman
Robert B. Leighton
Guido Munch
George W. Preston
Allan Sandage
Leonard Searle
Harold Zirin
Carnegie Institution oj Washington Year Book 72, 1972—1973
Contents
Introduction 101
Observing Conditions 103
Physics of the Sun 104
Active regions ........ 104
Rotation and large-scale velocity fields 105
Large-scale magnetic fields .... 105
Polar magnetic fields 105
Solar limb shift 105
Solar flares 106
Solar System Studies 106
Saturn 106
Titan 107
Uranus 107
Comet 1972h 108
Stellar Spectroscopy and Photometry . 108
White dwarfs . 108
Halo population 108
Emission-line star MWC 349 .. . 109
Spectrophotometry of B-type stars . . 109
Mercury lines in Si stars 110
Helium-weak B stars 110
Early A stars 110
Photometry of Pleiades stars ... 110
Stellar magnetism Ill
Photoelectric measurements of stellar
magnetic fields 112
Magnetic stars 112
Stellar Chromospheres 112
Stellar cycles 112
Absolute magnitudes of late-type giants 112
Variable Stars 112
Infrared Sources 113
Occupations of infrared stars ... 113
Size measurements of extragalactic ob-
jects 113
2-20 micron band 114
Millimeter-Wave Observations ... 115
1-mm photometry 115
Interstellar Gas and Gaseous Nebulae . 116
Interferometry of nebulae with low sur-
face brightness 116
Pulsars 116
X-Ray Sources 116
Cygnus X-l 116
Hercules X-l 117
Supernovae 118
Supernova search 118
Astrophysical studies 118
Review of spectra 119
Galaxies 119
Redshifts of fainter galaxies in the
Shapley-Ames Catalogue ... 119
Relative distance to nearby E and SO
galaxies 119
Motion of the sun relative to the
nearby galaxies 120
Markarian galaxies 120
History of star formation in galaxies . 121
Bright spiral galaxies 121
Synthesis of the stellar population of
galaxies 123
Photometry of OJ 287 and M82 . . 123
Chains of galaxies 123
Radio sources and spiral galaxies . . 125
Multiple interacting galaxies . . . 126
Deep photography of galaxies with con-
nected companions 126
Superposition printing 126
Nebulosity near 3C120 and BL La-
certae 126
Redshifts in small groups of galaxies . 126
Compact galaxies 126
Radio Sources 127
Identification of 3C sources .... 127
Radio source structure 127
Search for the radio source near
Cygnus X-3 127
Quasars 127
Quasar surveys 127
Spectroscopic observations .... 127
Quasars as events in nuclei .... 128
Association of quasars and clusters of
galaxies 128
Small quasars at the center of N
galaxies 128
Observational limits to nonvelocity
redshifts for the mini-quasars in N
galaxy nuclei 129
Properties of quasars 129
Angular diameters of quasars . . . 130
Observational Cosmology 131
Colors of E galaxies as an indication
of age 131
Absolute magnitude of first-ranked E
galaxies in clusters and groups . . 132
Distant clusters 133
Radio galaxy source counts .... 134
Theoretical Studies 134
Chemical history of galaxies . . . 134
Diffuse x-ray background 135
Magnetic field of the Crab Nebula . 135
Penumbral waves and umbral flashes in
sunspots 135
Photospheric oscillations 135
White dwarfs 136
Magnetic field of 53 Camelopardalis . 136
Zeeman-broadened line profiles . . . 136
Instrumentation 136
Prime-focus corrector 136
SIT-vidicon area photometer . . . 137
200-inch Cassegrain spectrograph . . 137
Prime-focus image tube 137
Coude image tube 137
Data system for 150-foot tower ... 138
Vacuum telescope 138
Guest Investigators 138
Las Campanas 149
Astroelectronics Laboratory .... 151
Digital image recorder 151
Computer systems , . . 151
Electronic instrument utilization . . 152
Photographic Laboratory 152
100-Inch du Pont Telescope .... 153
Mechanical parts and building . . . 154
Optics 154
Auxiliary instruments 154
Control system 155
Computer and data system .... 155
Las Campanas Observatory .... 155
Site development 156
Bibliography 156
Staff and Organization 161
INTRODUCTION
Twenty-five years have elapsed since for the radio galaxy 3C295; of the first
the 200-inch Hale Telescope was dedi- optical identification of a quasar by
cated on Palomar Mountain. This great Sandage and Matthews in 1960; and of
instrument has been an unqualified sue- the elucidation of quasar spectra by
cess. It has fully justified the hopes of Schmidt in 1963, with recognition of the
its planners in expanding man's concep- enormous redshifts of some of them.
tion of the universe and it has partici- The faith of George E. Hale in the
pated in opening a new array of challeng- future of astronomy as he gauged it in
ing problems, some of which are of a kind 1928 and the generosity of the Rocke-
and magnitude quite unimagined at the feller sources that made possible the
time the telescope was built. construction of the 200-inch telescope
If this instrument had done nothing have thus been amply justified. The
else, its construction would have been present generation of astronomers can
abundantly justified by the progress it take satisfaction in the fact that 45 years
has made possible on the cosmological after the 200-inch was funded and 25
problem: the determination of the scale years after it was dedicated, the optical
of the universe, the refinement of the exploration of the discernible universe
Hubble constant that defines the dis- seems to offer more opportunities for new
tance-redshift relationship, and the plac- knowledge than ever before. This tele-
ing of limits on the deceleration param- scope remains our most powerful instru-
eter. In the hands of capable observers, ment for conducting optical investiga-
te 200-inch telescope has produced tions, although it will soon be joined, at
thousands of direct photographs, tens of various other observatories throughout
thousands of photometric measurements, the world, by a number of other tele-
and more than 20,000 spectrograms of scopes that approach it in light-gathering
stars and extragalactic objects. Such power.
data have provided the basis for the In Hale's 1928 article in Harper's Mag-
refinement of the distance scale by Baade azine, in which he proposed the 200-inch
and his associates, and for correcting the telescope, he referred first to the great
expansion rate of the universe (the Hub- light-gathering power that would pro-
ble constant) by a factor of 10 from its vide an immense gain for exploring re-
1935 value to approximately 50 km s"1 mote space and the structure of the uni-
Mpc"1, as recently determined by Sand- verse, then to the value of the instru-
age and Tammann. While the decelera- ment for studying stellar evolution;
tion parameter of the universal expan- thirdly he mentioned the "greatest of
sion, q0, cannot yet be specified with any these problems: that of determining with
certainty, the extension of observations certainty the successive stages in the
to still fainter clusters of galaxies, to- development of spiral nebulae [galaxies]
gether with the application of evolution- . . . ." It is no accident that this year's
ary corrections, promises steady progress Report shows the results of increasing
toward this goal as well. attention to problems of the large subject
In this brief introduction, one can of galaxies — their structure, evolution,
scarcely even attempt to list the other and relationship to quasars. Star forma-
major investigations that the 200-inch tion within galaxies, age effects, radio
telescope has made possible. Mention emission, and high-energy physical proc-
must be made, however, of Minkowski's esses are increasingly rewarding subjects
1959 measurement of a redshift of 0.46 of study.
101
102
CARNEGIE INSTITUTION
Investigations of the distribution of
galaxies and clusters of galaxies in the
universe are proceeding. Advances in our
understanding of the general expansion
are promised by new observations of red-
shifts and more precise distance cali-
brations, as well as through statistical
studies. In terms of steady-state cos-
mology, Schmidt has analyzed counts of
strong radio sources in the Third Cam-
bridge Catalogue for which optical iden-
tifications have been made and redshifts
observed. His method allows the indi-
vidual properties of the identified sources
to be represented in the predicted source
counts. The predicted curve falls below
the observed curve by a factor of 5 at the
lower radio flux densities. His conclusion
is that radio source counts are incom-
patible with the steady-state theory.
Kristian has extended the search for
galaxies underlying quasars by identify-
ing the conditions under which such ga-
laxies should be detectable. Critical
examination of 200-inch photographs of
26 quasars gave results in good agreement
with the predictions. The evidence sup-
ports the increasingly general view that
quasars occur in the nuclei of galaxies.
Sandage's investigations of N galaxies
show that a typical object of this type is
a giant elliptical galaxy with a mini-
quasar in its nucleus; furthermore, that
the magnitude-redshift relationship for
the N galaxies leaves little, if any, margin
for a possible noncosmological component
of the redshift.
Increasing attention is being given to
the history of star generation in galaxies
and to the processes that trigger large-
scale bursts of star formation. Studying
ring galaxies, for example, Searle and
Sargent have obtained from multichannel
scans evidence that the energy distribu-
tions of the rings are closely similar to
those of star clusters a few hundred mil-
lion years old. The same age is indicated
for these galaxies from the measured ex-
pansions of the rings. Searle and Sargent
believe that these galaxies are the first
truly young extragalactic systems to be
recognized. They also conclude that the
outer reaches of certain paired galaxies,
distorted by mutual interaction, consist
of predominantly young stars.
The concept that density waves or
compressional effects in galaxies acceler-
ate the formation of stars from the in-
terstellar material is finding increased
support not only in situations such as the
foregoing, but in the spiral structure and
nuclear regions of galaxies in general.
Van der Kruit, examining the high-reso-
lution radio brightness distributions
within galaxies, obtained at Westerbork,
found that the radio-brightness distribu-
tions were displaced toward the inner
edges of the optical spiral arms, thus
confirming the concept of compressed re-
gions in the density-wave picture of
spiral structure. The radio-spiral struc-
ture is believed to result from compres-
sion of the magnetic field and relativistic
electrons in the large-scale "base disk"
of a galaxy. Star formaton is triggered
in the density-wave compression regions
in a manner governed by the degree of
compression. Van der Kruit concluded
that the relationship of specific angular
momentum to mass is the governing pa-
rameter. He found that, for a particular
mass, galaxies with strong compressional
effects have a lower specific angular mo-
mentum than galaxies with weak com-
pression. Statistically, this results in a
larger central concentration of mass for
strong compression galaxies, which ex-
plains the result that such galaxies
generally have better developed nuclei.
Other evidence suggests that the char-
acteristics of a galaxy such as the visi-
bility of its spiral structure and the de-
gree of concentration toward the nucleus
are not determined by the details of the
density generating mechanism but rather
by the characteristics of the disk, that is,
by the specific angular momentum and
the mass. Given these governing param-
eters, the compressional strength is de-
termined just as is the form of the
rotation curve as measured by optical
spectroscopy. These conclusions were re-
HALE OBSERVATORIES
103
inforced by observations at Palomar of
the rotation curve for NGC 4321.
Compact chains of galaxies, particu-
larly apparently linear chains of a few
members, have been under investigation
for some time by Sargent. Studies of the
colors of member galaxies provide no
indications that these are young systems,
although typical dissolution times com-
puted by others are about 5 X 10s years
— a relatively short interval on the cos-
mic scale. A new and comparatively
simple explanation advanced by Sargent,
which has now found support from com-
puter analysis by himself and Turner,
is the idea that apparent linear forma-
tions in compact groups of galaxies repre-
sent states (nearly coplanar) in the
normal dynamical evolution of small
clusters that on the average are spherical.
Numerous attempts have been made to
identify optical counterparts of recently
discovered galactic x-ray sources. One
such object is the binary BD+34°3815,
which is a weak variable radio source and
is identified with Cygnus X-l. The star
is a 5.6-day spectroscopic binary of which
the primary optical component is of
spectral type BO lb, with velocity ampli-
tude of 74 km s"1. Brucato and Kristian,
observing the star spectroscopically at
Mount Wilson, discovered that the weak
emission feature at A4686 (He II) had a
velocity amplitude nearly twice as great
as that of the B star and that it varied
in anti-phase; no other evidence of a
secondary spectrum is seen. After ana-
lyzing the evidence, Brucato and Kristian
concluded that one possibility is that the
secondary component of the binary is
a black hole, the x rays being produced
by the thermalization of gravitational
energy of matter falling in the vicinity of
the black hole. Other interpretations,
however, still remain open.
The brightest supernova in many
years, 1972e, was discovered by Kowal in
NGC 5253 as reported in Year Book 71,
p. 677. Collaboration by Kirshner, Oke,
Penston, and Searle has resulted in a
report on the spectra of supernovae that
describes and interprets the scans of
several such objects of both Types I and
II and includes a comprehensive series of
observations of SN 1972e covering 230
days. Spectrum scans of this object are
expected to be obtainable for at least one
additional year. Important new conclu-
sions regarding supernova spectra have
been reached from a study of the new
data.
Observations with the telescopes at
Mount Wilson and Palomar have never
been limited to the visible region of the
spectrum. Significant infrared observa-
tions were conducted locally, long before
the advent of sensitive modern detectors.
C. G. Abbot used the 100-inch Hooker
reflector as early as 1924 for determining
infrared energy curves of stars with his
special radiometer. P. W. Merrill made
important observations of stellar spectra
in the infrared, using Eastman Z plates
in the 1930s. The infrared solar spectrum
was measured in detail, and its photo-
graphic limit was extended to 13,495 A in
1935. The past decade has seen remark-
able advances in the 1 to 20 micron range
through the application of lead-sulfide
and mercury-doped germanium detectors
in the hands of Neugebauer, Becklin, and
their associates. Now with the 200-inch
an extension of photometry at a wave-
length of 1 millimeter has been effected
by Harvey, Werner, Elias, and Gatley.
They have found it feasible to observe at
this wavelength during extended periods
of twilight, thus utilizing the telescope
for six hours or more per day when it
would otherwise be idle. Objects studied
included planets, galactic H II regions,
and the QSO 3C273.
OBSERVING CONDITIONS
At Mount Wilson the 100-inch tele-
scope was used for observations on 203
complete nights and 91 partial nights
for a total of 2566 hours. Sixty-nine
104
CARNEGIE INSTITUTION
TABLE 1. 200-Inch Observations
Number of
1972-73
Complete
Partial
Hours
Nights
Nights
Observed
July
21
9
272.5
August
23
8
267.5
September
22
5
263.4
October
19
4
219.9
November
20
6
263.8
December
18
4
221.7
January-
22
2
288.4
February
12
6
175.8
March
9
7
146.9
April
23
4
295.0
May
27
3
263.3
June
17
12
239.4
Totals
233
70
2,917.6
nights were lost because of the weather.
Total precipitation was 40.5 inches, with
63 inches of snow.
The 200-inch Hale Telescope at Palo-
mar Mountain was used for a total of
2917.6 hours, as shown in Table 1.
The figures include a substantial amount
of twilight time that was used for obser-
vations not requiring a dark sky. Total
precipitation at Palomar was 38 inches,
with 85 inches of snow.
PHYSICS OF THE SUN
Observations of the sun on a daily
basis continue at Mount Wilson under
the supervision of Howard. Between
June 1, 1972, and May 31, 1973, the fol-
lowing observations were obtained at
Mount Wilson:
Direct photographs 293
Ha spectroheliograms, 30-foot focus 527
K2 spectroheliograms, 30-foot focus 535
Full-disk magnetograms 389
Integra ted-light magnetic-field
measurements 190
Sunspot drawings 290
Observations were made on a total of
318 days.
The analyses of the digital data from
the magnetograms and the integrated-
light magnetic-field measurements are
carried out on the Raytheon 704 digital
computer. The magnetograms continue
to be published monthly in the National
Oceanic and Atmospheric Administration
bulletin Solar Geophysical Data, and the
computer-plotted synoptic charts con-
tinue to be published in the Quarterly
Bulletin on Solar Activity of the Interna-
tional Astronomical Union. Partial sup-
port for these observing programs comes
from the National Aeronautics and Space
Administration, the Office of Naval Re-
search, and the Air Force Cambridge
Research Laboratories.
Active Regions
From a study of active regions of the
last few years, Howard and S. Edberg
have concluded that chance eruptions of
magnetic flux can explain the occasional
renewal of activity in an active region.
It is not necessary to assume that such
resurgences of activity represent a prop-
erty of an active region; magnetic flux
erupts at the solar surface randomly with
respect to existing active regions. This
does not bear on the phenomenon of ac-
tive longitudes, but it relates to the
problem of the origin of solar active re-
HALE OBSERVATORIES
105
gions because it appears that whatever
is the mechanism to produce the eruption
of magnetic flux to form an active center,
it does not affect the probability that
another region will form at nearly the
same location.
Rotation and Large-Scale Velocity Fields
The velocity data from more than 300
full-disk magnetograms made in various
spectrum lines within the last year have
been studied by Howard. The Fe I and
Fe II lines show identical equatorial ro-
tational velocities, but the Fe II line
shows a more rapid decline of angular
velocity with latitude than does the neu-
tral line. The redshift parameters for the
two lines at the limb differ by nearly
a factor of 2. The ionized line has the
larger value. Variations with time of the
equatorial velocity measured with the
two lines are essentially identical. The
average equatorial velocity measured
during the year was significantly greater
than for a comparable period five years
ago. This may be associated with a lower
level of solar activity, but further obser-
vations are required to be certain of this
point.
Large-Scale Magnetic Fields
From a study of the solar magnetic
flux data from September 1966 to Sep-
tember 1972, Howard has been able to
draw conclusions about the average in-
clination of magnetic lines of force and
related matters. These observations con-
sist of about 1200 individual digitized
magnetograms. Over the whole period,
the average distribution of total mag-
netic fluxes (\F+\ -\- \F.\) over the vari-
ous latitude zones into which the data
were divided shows that in all latitude
zones, and in the north and south sepa-
rately, the fluxes west of the central
meridian were greater than those in the
east by roughly 1%. This indicates an
average inclination of field lines of a few
degrees in a direction to trail the rota-
tion. About 90% of the magnetic flux is
confined to latitudes below 40°, not
taking account of projection effects. If
one assumes radial field lines, 85% of the
flux is equatorwarcl of 40°. In both
hemispheres the preceding and following
polarities are inclined toward each other
by a few degrees. In the north, the
following-polarity flux is inclined slightly
to lead the rotation, and the preceding-
polarity flux trails the rotation. In the
south, where the total magnetic flux is
significantly weaker than in the north,
both polarities trail the rotation.
Polar Magnetic Fields
Over the past decade, the polar mag-
netic fields of the sun have been weak
and variable. At times the fields at the
poles have shown a tendency to vary in
unison. Howard has found that near the
middle of 1971 the north heliographic
polar fields changed sign to become con-
sistently positive, first at lower latitudes,
then above 70°. The south pole remained
weak but predominantly positive until
about April 1973, when it appeared to
change sign. The north polar field is
measured to be about 0.5 gauss. This
polar field polarity reversal has occurred
about three years later in the phase of
the cycle than the polarity reversal of the
preceding cycle.
Solar Limb Shift
With the 150-foot tower telescope,
Hart made measurements of the center-
to-limb variation in the wavelengths of
spectral lines. Most lines, when observed
at the limb of the sun, are shifted slightly
to the red compared with their wave-
lengths at the center of the disk. These
shifts, which are generally of the order of
a few milliangstroms, vary greatly from
line to line.
Hart then suggested a theoretical ex-
planation for this effect, whose cause
has been unknown since the shift was first
observed over 60 years ago. According
to his hypothesis, the apparent redshift
toward the limb is the result of pressure-
106
CARNEGIE INSTITUTION
induced blue shifts of spectral lines com-
ing from the center of the disk. (These
pressure shifts are much smaller at the
limb.) Detailed computations based on
this theory predict a shift of 3.39 mA
for the Na I doublet at AA6154, 6161, and
a shift of 2.24 mA for the Na I doublet
at AA5683, 5688. These results are in
good agreement with observed values.
Solar Flares
A series of unusually intense flares took
place in August 1972 and were exten-
sively observed at Big Bear. Zirin and
Tanaka completed an extensive study of
the different observations, based on
multiwavelength filter cinematograms,
spectra, and radio and x-ray data. They
found the principal cause of the flares to
be the extreme shearing, due to sunspot
motions, of magnetic fields along the
neutral field line. The field on the sur-
face was generally force-free, running
parallel to the boundary of opposite
polarities, while after each flare a loop
system appeared above the neutral line,
tracing out a lower energy potential sys-
tem. The energy in the Ha radiation of
the great flares of August 4 and August
7, 1972, was found, on the basis of ex-
tensive spectra reinforced by focal-plane
observations, to be only 2 X 1030 ergs,
approximately equal to the energy ob-
served in cosmic rays above 10 Mev from
these flares. This number is substantially
lower than previous estimates of optical
flare emission but is probably a better
value because of the high quality of the
observations. Most of the emission from
the flares came from small intense ker-
nels, and most of the bright flare areas
showed much weaker emission than these
kernels. Just after the August 7 flare, a
sharp discontinuity of radial velocity was
detected near the neutral line by means
of spectra, corresponding to a shear flow
of several kilometers per second. This
flow appears quite important in the flare
phenomenon. On August 2, two flares
were observed with a broad-band (15 A)
filter centered on A3835. This filter shows
phenomena in the low chromosphere very
well. An impulsive flare was detected
which produced a number of very rapid
brightenings of individual points along
the neutral line. Each point was about
1 arc sec in diameter and had a lifetime
of 5 or 10 sec. These brightenings are
interpreted by Zirin and Tanaka as due
to the dumping of energetic electrons at
the feet of different lines of force coming
from the flare, which is somewhat higher
above the atmosphere. The actual emis-
sion is in the Balmer line H9. This is the
first time such rapid and small phe-
nomena have been observed in a flare.
Emission in this wavelength has been ob-
served in other, less implusive flares, but
in these it corresponds to a slower,
broader band of emission, also produced
by conductive heating from the flare.
Zirin and Bernard Lazareff, a gradu-
ate student, concluded another study of
activity in a large sunspot group in
August 1971. It was found that the
major activity in this group coincided
with the growth and westward motion of
a new sunspot. So long as this new sun-
spot plowed ahead through the solar at-
mosphere, a number of flares occurred
at its leading edge. A careful comparison
was made between these flares and Type
III radio emission observed at the Uni-
versity of Colorado, showing that Type
III bursts occurred only after flares in
this part of the active region, whereas
other flares in the sunspot group were un-
able to produce the fast electrons that
give rise to Type III bursts.
SOLAR
Saturn
SYSTEM STUDIES
The brightness of the equatorial belt
of Saturn was measured by Neugebauer
and Munch in the M-band (4.6-5.2 /mi)
and through interference filters centered
at Ai = 4.805 and A2 = 4.720 ^m, with
HALE OBSERVATORIES
107
widths at half-peak transmission of 0.11
/Am. An absorption band due to CH3D
appears in the spectrum of Jupiter
centered at A2 (Year Book 70, p. 399;
Beer et al, Science, 175, 1360, 1972),
while the pass band at Ai is essentially
free of the CH3D absorption. Thus it
was hoped that a color Ci2 = m(X1) =
m(A2) would provide evidence for the
presence or absence of CH3D in Saturn.
If the transmission curves of the Ai and
A2 filters are folded into spectral scans
(A/AA = 180) of Jupiter and Mars
(Year Book 71, p. 657), a color excess
C12(Mars) — d2( Jupiter) = +0.16 mag
is obtained. The photometry of Saturn
and of the moon, a Tauri, a Orionis, and
(3 Pegasi was carried out at the Casse-
grain focus of the Hale telescope on
November 23, 1972. The brightness tem-
perature of Saturn in the M-band,
measured over a central circular area of
7.5 arc sec diameter, is 180° K, and its
color excess with respect to that of the
moon is Ci2(moon) — Ci2 (Saturn) =
+0.08 mag. This value equals that ex-
pected for blackbodies at the observed
M-brightness temperatures, namely, 180°
K and 245° K for Saturn and the moon,
respectively. The color measurement
therefore does not provide evidence for
the presence of CH8D absorption in
Saturn. However, it should be noticed
that the observed mean color excess
Ci2 (stars) — Ci2(moon) = 0.02 mag
does not agree with the value +0.20 mag
calculated for blackbodies. It follows
that either the moon or the reference
stars are not radiating as blackbodies in
the (Ai,A2) range. The effect of stellar
CO absorption in its fundamental band
is a possible explanation for the apparent
discrepancy.
Titan
The spectrum of the brightest satellite
of Saturn has been studied by Munch on
material obtained with an image intensi-
fier (S25 cathode) at the 72-inch camera
of the 200-inch coude spectrograph. With
resolutions of 0.3 and/or 0.6 A, over the
range AAO.63-0.88 /mi, no proper spectral
feature that could not be attributed to
CH4 has been detected. In particular,
the NH3 bands at AA0.65 and 0.79 /mi
and the H2 quadrupole lines of the (3, 0)
band have not been seen on these plates.
In comparison with the CH4 bands in
Saturn and Uranus, however, those in
Titan show a noticeable curve-of-growth
differential effect. The weak absorption
lines in Titan (equivalent widths smaller
than about 0.1 A) are as strong as cor-
responding lines in Saturn, but the strong
absorption features are relatively weaker
in Titan. This effect can be qualitatively
understood if the effective column den-
sity in Titan is as large as in Saturn but
the gas pressure is lower.
Uranus
The quadrupole spectrum of H2 in
Uranus has been under continued study
by Munch. The strength of the SO and
SI lines of the (3, 0) band has been
measured photoelectrically with the
Perot-Fabry scanner of the 100-inch
coude spectrograph. Under a resolving
power of 41,000, the equivalent widths
of these two lines, obtained from scans
involving 3000 counts at sampling in-
tervals of 0.087 A, are 0.166 and 0.160
A, respectively. An uncertainty of about
10% in these values arises from difficul-
ties in locating the reference continuum.
The equivalent width of the SO line in
the (4, 0) band, also measured inter-
ferometrically with a resolving power of
44,000, is 0.023 A. These measurements
are in essential agreement with those re-
ported by Trafton {Bull. Amcr. Astron.
Soc, 5, 290, 1972) but do not confirm the
larger values given by Price (Bull. Am.
Astron. Soc, 5, 291, 1972) for the (4, 0)
band.
The spectrum of Uranus in the region
108
CARNEGIE INSTITUTION
of the (5, 0) band of H2 has also been
explored by Munch on image-tube spec-
tra obtained with the 200-inch coude at
6.8 A mm-1. The measurement of two
Uranus spectrograms has revealed the
presence of two nonsolar lines at
AA5327.37 and 5286.36, which are only
about 1.0 A redward of the predicted
wavelengths for the SO and SI lines, re-
spectively, of the (5, 0) Ho vibrational
band. The identification of these lines
appears certain, but because of blends
with weak lines in the Fraunhofer spec-
trum their equivalent widths can be esti-
mated to be no more than 20 mA.
Further work at higher resolving powers
will be carried out to make a precise
measurement of their strengths.
Comet 1972h
This new comet was found on June 9,
1972, by Sandage [Intern. Astron. Union
Circular No. 2413) during a patrol of
quasar fields for variable objects with
the Palomar 48-inch schmidt. The comet
was of magnitude 13 at the time of dis-
covery and remained near that magni-
tude for most of the report year because
of the unusual nature of its orbit. The
comet has a large perihelion distance
of 4.28 A.U. according to elements cal-
culated by B. G. Marsden (Intern.
Astron. Union Circular Aro. 4^6) , and the
distances between the comet, earth, and
sun changed slowly for several months.
Perihelion passage occurred near Novem-
ber 15, 1972.
STELLAR SPECTROSCOPY AND PHOTOMETRY
White Dwarfs
Spectroscopy by Greenstein with the
coude image tube at the 200-inch tele-
scope has continued to show that the
hydrogen-line white dwarfs have quite
small rotational velocities. At present,
five of the six stars successfully observed
show cores, formed in nonlocal thermo-
dynamic equilibrium, that are barely
wider than the instrumental resolution.
If these stars are viewed as the central
regions of former red giants, an efficient
mechanism must operate to reduce the
angular momentum. This occurs in the
interiors of whatever stars undergo slow
evolution into the white dwarf region.
Halo Population
In Year Book 69 (p. 84), Figure 1 dis-
played the schematic Hertzsprung-Rus-
sell diagram of the faint blue stars in
the halo of our Galaxy. The large num-
ber of high-temperature model atmos-
pheres computed by various investigators
has made it necessary to reanalyze the
observed energy distributions, line pro-
files, and line intensities in the approxi-
mately 200 stars that Greenstein and
Anneila Sargent have been discussing.
Particularly the high-temperature end of
the halo population has been extensively
studied. The multichannel spectrophoto-
meter provides reliable temperature cali-
bration to approximately 40,000°, the
line widths provide a spectroscopic esti-
mate of the surface gravity, and the
Balmer discontinuity at a given tempera-
ture provides some indirect evidence of
low hydrogen-to-helium ratio in a few
stars. Fundamentally, the general fea-
tures of the H-R diagram for halo stars
as newly rederived by Greenstein and
Anneila Sargent are the same as pictured
earlier. A nearly horizontal group of
stars exists over a range of temperatures
from below 10,000° to at least 40,000°.
If it is assumed that all these stars have
identical masses, the fit to the horizontal
branches of globular clusters permits de-
termination of the mean mass of these
stars ; this is approximately 0.50 9TZ0. The
large extension to higher temperatures,
not normally seen in globular clusters,
and the extremely high temperatures in-
dicated, requires that the helium core
of such stars be surrounded by a hydro-
gen envelope of negligible mass. Above
HALE OBSERVATORIES
109
a temperature of 30,000° the surface
helium abundance appears to be high in
certain stars, which will be classified as
0 subdwarfs. If they have the same
mass as the horizontal-branch stars, they
form a connecting link between the hot
end of the horizontal-branch sequence
and the hotter white dwarfs. Hot O-type
subdwarfs of high luminosity, related to
the nuclei of planetary nebulae, are rare
among the halo population. The wide
variety of strengths of absorption lines
of other elements (ionized carbon and
silicon) in the halo stars suggests a sur-
face-composition variation like that in
the globular-cluster giants at lower tem-
peratures, which have a large diversity
of abundances of the metals.
Emission-Line Star MWC 349
Greenstein has studied the Be star
MWC 349, which proves to be one of the
most highly reddened stars known, with
an absorption of over 9 mag at visual
wavelengths. The star shows variations
from night to night in the hydrogen
emission lines, which are formed in an
extremely compact, dense H II region.
D. Allen had observed this object in his
infrared survey of Be stars. MWC 349
is particularly notable in its nonthermal
radio-frequency emission spectrum. At-
tempts to fit the spectrum, from 0.4 to
1.0 microns, with model atmospheres re-
sult in a redetermination of the reddening
law which lies within 5% of the Whit-
ford reddening law. The total energy
emitted by the B star is sufficient to heat
interstellar grains and provide the in-
frared emission observed by Allen. If
the radio-frequency source is as small as
the optical He II region, the star pro-
vides sufficient energy to maintain that
radiation also. MWC 349 thus becomes
one of the most completely observed
radio stars. It shows many resemblances,
over the range from 21 cm to 0.4 microns,
to the nuclei of Seyfert galaxies and to
some quasars. Typical electron densities
of 106 to 107 per cubic centimeter are
required in a region only a light-day in
diameter.
Spectrophotometry of B-Type Stars
Preston is engaged in an extensive spec-
trophotometric study of some 400 bright
B5— A0 stars. The coude scanner of the
100-inch telescope is being used to meas-
ure relative intensities in 14 contiguous
4 A bands in the longward wing of HS.
The measurements provide a profile of
HS and equivalent widths of Si II A4128-
30 and He I A4144. If temperature esti-
mates (as from Q) are available, a fit to
D. Peterson's theoretical HS profiles
yields a gravity and, in principle, mass
and age estimates for stars of normal
composition. Low-dispersion spectro-
scopists have reported numerous "mild"
or marginal Si stars. High-dispersion
studies have found a few of these to be
normal (e.g., 21 Aquilae), but C. J. Dur-
rant recently has confirmed the existence
of such marginally peculiar stars and also
has found evidence for stars with ab-
normally weak silicon lines. The purpose
of this investigation is to provide a sub-
stantial body of quantitative information
on the distribution of silicon and helium
abundances among stars of different tem-
peratures, gravities, and rotational ve-
locities in the temperature domain of the
Si stars.
Rotational velocities, required for the
comparison of observations with theo-
retical HS profiles, are not known for the
great majority of stars in this program.
They are being determined from coude
spectrograms obtained concurrently, and
these will greatly improve the statistics
of rotation for upper main-sequence
stars, particularly in the range V sin i
< 50 km s"1. These spectrograms provide
a wealth of additional information. A
number of new Hg-Mn stars have been
discovered as well as occasional double-
lined spectroscopic binaries. Finally,
some 70 of the 200 spectrograms obtained
to date possess interstellar K lines. These
will provide new information about the
110
CARNEGIE INSTITUTION
motion and distribution of interstellar
gas within 200 parsecs of the sun.
Mercury Lines in Si Stars
The presence of the Hg II line A3984
has come to be regarded as a common
characteristic of the spectra of the Mn
stars, but the line has been identified in
a few Si stars as well. Preston has ex-
amined coude spectrograms (mostly 4 A
mm"1) of some 70 Si stars with Te(Q)
> 10,000° K and V sin i < 50 km s'1 for
the presence of Hg II A3984. He con-
cludes that the line is definitely present
in 10 stars and undetectable in 15 others.
No conclusion can be drawn from visual
inspection of the remainder of the sample
because of the likely contributions due
to Cr I A3983.90 and Fe I A3983.96. Thus
it appears that the Hg anomaly is an
occasional but not ubiquitous character-
istic of the Si stars, and it is not accom-
panied by Mn enhancement in this group.
The spectrum of HD 200311, however,
contains the three strongest lines of PT
II in addition to Hg II A3984, and re-
solved Zeeman patterns indicate the pres-
ence of a mean surface field of 10 kilo-
gauss. To judge from its colors (B — T7 =
0.11, U — B — —0.50), it is easily the
hottest of all known Pt stars.
Helium-Weak B Stars
Dworetsky has continued the survey
of B8-B9 III stars listed in the catalog
of Cowley, Cowley, Jaschek, and Jaschek
{Astron. J., 74, 375, 1969), and has ex-
tended it to include A0 III stars as well.
The survey is being conducted with the
100-inch coude spectrograph; the 60-inch
x-spectrograph is being used to observe
stars too far north for the coude con-
figuration. The purpose of the survey is
to see whether or not additional peculiar
stars (especially Hg and Hg-Mn types)
can be found in this manner. So far about
10% of the stars observed are peculiar
in some way. Among the more interest-
ing objects found is HD 191110, a double-
lined spectroscopic binary Hg star. Al-
though both stars share the Hg anomaly,
only the primary has enhanced Mn lines,
while the secondary is overabundant in
platinum. The system closely resembles
another DLSB Hg star, 46 Draconis (P.
S. Conti, Astron. Astrophys., 7, 213, 1970) .
Early A Stars
Dworetsky has completed a 40 A mm-1
survey of all A0 main-sequence stars in
the Cowley, Cowley, Jaschek, and Jas-
chek catalog with the 60-inch x-spectro-
graph. All normal stars of MK types
B9.5-A1, IV-V, were included. The data
will be used to obtain an improved rota-
tional velocity distribution and to search
for new sharp-lined bright early A stars
to be observed with higher dispersions.
It is hoped that new information about
the compositions of AO-type stars with
subtle peculiarities (e.g., early Am stars
and cooler Hg stars) will help in under-
standing the processes that cause these
stars to have anomalous compositions.
Photometry of Pleiades Stars
Zappala has now obtained observations
at 1.6 and 2.2 /x for about 30 Pleiades
main-sequence stars of spectral types K
through early M with reliable UBV
photometry. Most of these stars have
been observed at the two wavelengths on
at least two good nights with the 100-
inch Mount Wilson reflector. Although
many of these objects are known to be
flare stars, no obvious variations were
detected either visually at the telescope
or in the photometry. A standard rela-
tion between the V — [2.2 /x] and B — V
colors has been derived for the lower
main sequence through observations of
field and Hyades late-type dwarfs.
The result suspected earlier that the
Pleiades stars are systematically redder
than the standard relation has been con-
firmed. Between B — V colors of about
+ 1.00 and + 1.50, the Pleiades have an
excess averaging nearly 0.2 mag in V —
[2.2 p] and as large as 0.7 mag in one
case, independent of interstellar redden-
HALE OBSERVATORIES
111
ing. Furthermore, the excess seems to be
independent of the position of the star in
the reflection nebula surrounding the
cluster, which might be held responsible
for producing erroneous UBV measure-
ments.
Interpretation of these results is cur-
rently unclear. Cool companions could
produce excesses in V — [2.2 /x] ap-
proaching 0.1 mag in favorable cases,
but this would be inadequate even in the
unlikely event that all the stars are
binaries. If young stars with consider-
able convective activity are heavily cov-
ered with cool regions similar to ordi-
nary sunspots, an apparent excess in
V — [2.2 [x~\ could be produced; how-
ever, the excess would be no larger
than 0.2 mag even for a star whose photo-
sphere is half spotted. The most likely
explanation is that these comparatively
young objects are still surrounded by
circumstellar materials of low tempera-
ture and moderate optical depth, perhaps
remnants of the period of star formation.
There is support for this hypothesis in
the fact that the excess in T7 — [2.2 /a]
seems to correlate with a deficiency in
the visual magnitude, i.e., the distance
the star appears to lie below the standard
zero-age main sequence. However, there
is no general excess in the observed
[1.6 /x] — [2.2 fx] color index as large
as 0.05 mag, so that shells, if present,
must have temperatures less than ^500°
K. The actual thermal emission of such
shells would be observed at wavelengths
longer than 3 [x] such observations will
be very difficult, since these stars are all
fainter than magnitude 10 at 2.2 (x.
The K-type star BD— 10°4662 (FK
Serpentis), recently discovered to have
flared on several occasions, has been
studied by Zappala both spectroscopic-
ally and in the infrared. Both com-
ponents of this close visual binary have
been shown to have high lithium abun-
dances by Herbig, who suggests that their
similar spectroscopic and photometric
properties place them in the post— T
Tauri phase of evolution near the bottom
of the convective portion of their Ha-
yashi tracks. The combined object has
been observed on several occasions in
the infrared with the Mount Wilson tele-
scopes with the following results: [2.2
/x] = 6.55, [1.6 fx] —[2.2 fxl = +0.27,
[2.2^] — [3.5/*] =+0.59. These colors
approach those of conventional late-type
T Tauri stars and are quite different
from ordinary K stars (both indices
about +0.1). A single 40 A mm-1 spec-
trogram has also been obtained at the
100-inch coude. This blue plate shows
absorption features similar to those of
an ordinary K4 dwarf, with the B aimer
series weakly in emission and Ca II (H
and K) moderately strong in emission.
Thus — 10°4662 indeed appears to be
closely related to the T Tauri stars and
is perhaps in the final part of that stage
of stellar evolution.
Stellar Magnetism
Borra and Vaughan continued their
development of high-resolution photo-
electric polarimetry for investigation of
the Zeeman effect in magnetic Ap stars.
The instrumentation, used with a Fabry-
Perot interferometer in the coude spec-
trum scanner of the 100-inch telescope,
employs a KDP electro-optic crystal
modulated at 0.5 kHz between two ortho-
gonal states of birefringence and yields
simultaneously the degree of polari-
zation (Zeeman signature) and the rela-
tive spectral intensity (line profile) with
a resolution of about 0.08 A. Observa-
tions of circular polarization due to the
longitudinal component of the magnetic
field have been carried out at various
phases on a1 Canum Venaticorum, (3
Coronae Borealis, 78 Virginis, and 52
Herculis during the first quarter of 1973.
So far the line of Fe II at 4520.2 A has
been observed exclusively. The Zeeman
signatures, which are revealed in con-
siderable detail by these observations,
show many of the attributes generally
predicted on the rotating dipole model.
Further observations, including the at-
112
CARNEGIE INSTITUTION
tempt to study linear polarization, are
being pursued as the basis for detailed
comparison with theoretical models of
these and other magnetic Ap stars.
Photoelectric Measurements of Stellar
Magnetic Fields
Borra has interfaced polarizing optics
and electronics with the Mount Wilson
low-resolution scanner. The instrument
has been used with the 60-inch telescope
to measure circular polarization in the
wings of Ha in bright, rapidly rotating
Ap stars whose lines are too broad to be
measured photographically for magnetic
fields. None of the stars so far observed
has been found to show detectable cir-
cular polarization or a magnetic field
within the observational errors (typical
standard deviation is a few hundred
gauss).
Magnetic Stars
Borra and Dworetsky have examined
all of the Zeeman-analyzed plates of /?
Coronae Borealis in. the plate files of the
Observatories. Many of the plates have
been remeasured or measured for the first
time. There appears to be a correlation
between the measured longitudinal mag-
netic field and the photographic density
of the plates at negative polarity ex-
tremum but not at positive polarity.
Numerical models, based on the mag-
netic-field configuration of (3 CrB de-
termined by Wolff and Wolff (Astrophys.
J., 160, 1049, 1970), show simple sym-
metrical Zeeman-analyzed line profiles
at the positive extremum and complicated
asymmetrical profiles at the negative ex-
tremum. Heavily exposed plates tend to
emphasize the core of the spectral lines,
while lightly exposed plates tend to em-
phasize the wings. On the basis of those
profiles, one can expect a dependence of
the measured longitudinal field strength
on photographic density at the negative
extreme but not at the positive extreme.
This effect might explain the large scat-
ter seen at negative polarity when ob-
servations of many years are plotted
in the 18.5-day period of the magnetic
variation. This scatter had previously
been interpreted as a secular variation
in the field strength.
STELLAR CHROMOSPHERES
Stellar Cycles
Wilson has continued his measurement
of fluxes at the center of the H and K
lines in a number of main-sequence stars,
using the coude scanner at the 100-inch
telescope. The earlier types, F5-G5, con-
tinue to show no real evidence of cyclical
behavior. However, among the later-
type objects several have reversed the
direction of change of the flux values
during the past year. For most or all of
these stars it now appears that another
two to four years' observation may yield
completed cycles with adequate overlap
to render the results convincing.
Absolute Magnitudes of Late-Type
Giants
Photographic observation of the spec-
tra of late-type giants at 10 A mm-1 is
being continued by Wilson to determine
absolute magnitudes from measures of
the widths of the chromospheric K lines.
About 100 spectrograms were obtained on
this program during the year.
VARIABLE STARS
Douglas Duncan and Preston have (e '— ' 0.7) of the orbit, the lines are double
completed a study of the double-lined for only 3 days preceding periastron.
spectroscopic binary 8 Delphini (Forb = Series of spectrograms (Lick 120-inch
40.58 days). Because of the eccentricity coude) obtained during those phases in-
HALE OBSERVATORIES
113
dicate that both components have small
amplitude velocity variations character-
istic of 8 Scuti-type variables. The pul-
sational periods of the primary and
secondary are Pi = 0.d158 and P2 =
0.d134, respectively. The mass ratio
(Mx/M2 = 1.2) and the luminosity ratio
(L1/L2 = 1.05) inferred from the equiva-
lent widths of Fe II lines do not seem
to be compatible with either pulsation
theory or stellar evolution theory.
INFRARED SOURCES
Occultations of Infrared Stars
A large number of the infrared sources
that have been detected in the past
several years have been interpreted in
terms of the so-called "dust shell" model.
In this model energy from a central ob-
ject at relatively short wavelengths is
absorbed by particles in a surrounding
envelope and reradiated at the much
lower temperature characteristic of the
grains. In this way the maximum of the
energy distribution can be shifted far into
the infrared, perhaps totally obscuring
the original energy distribution of the
central star. Unfortunately, direct ob-
servational evidence for these shells other
than the observed energy distributions
themselves is largely lacking. For objects
sufficiently near the ecliptic, however,
lunar occultations provide a useful method
of measuring angular extent of infrared
sources and may provide direct evidence
for a shell.
IRC+10011 (= CIT 3) is a bright
infrared source that also emits micro-
wave lines of the OH molecule. On the
basis of CO and H20 absorption features
in its 2.0-2.5 /x spectrum and its vari-
ability at 2.2 /x, it has been classified as
a late-type Mira star. The object has
recently undergone a series of lunar oc-
cultations. Three of these events could
be observed at Palomar and Mount Wil-
son and attempts were made by Zappala,
Becklin, and Neugebauer to observe
these simultaneously in two wavelengths
with the 200-inch and 100-inch tele-
scopes. Successful observations were ob-
tained at 2.2 and 10.1 /x on one occasion
and at 20 ^ on another. The 2.2 /x ob-
servation is particularly interesting, for
it cannot be interpreted in terms of a
simple single source. A good fit to the
data is provided by the Fresnel pattern
of a uniform disk ^0."013 in angular
diameter, contributing 65% of the total
luminosity, superimposed on a disk
~O."070 in diameter, contributing the
remaining 2.2 /x flux. The 10 /x observa-
tion is consistent with a single uniform
disk slightly larger than 0."120 in extent.
At 20 jx the observed diameter is roughly
equivalent to that at 10 /x.
These observations thus provide clear
evidence for a shell of material around
a smaller central object. If the 0."013
source is interpreted as the central star
seen directly through the obscuring ma-
terial, an effective temperature of about
2100° K may be derived, since the total
flux of the object has been measured.
This value will be somewhat smaller if a
limb-darkened model is assumed and is
in satisfactory agreement with the tem-
perature derived from the observed spec-
tral features. If the total luminosity of
the object is about 104 LQ, the distance
to the star is about 500 pc and the ob-
served extent of the outer dust shell
corresponds to a diameter of about 60
A.U. The temperature of the dust shell
is about 600° K.
Size Measurements of Extrag atactic
Objects
It is by now well established that
several extragalactic sources show anom-
alously strong infrared emission. Al-
though the effect is most noticeable in
peculiar galaxies of the Seyfert and
Markarian types, large excesses have
also been observed in some apparently
114
CARNEGIE INSTITUTION
normal galaxies such as NGC 253. How-
ever, the physical mechanisms producing
the infrared emission have not been es-
tablished. In particular, observations in
the visible band of many of the galaxies
in question show evidence for nonthermal
emission. Therefore, a central question
is whether or not the infrared emission
is also nonthermal in origin.
One means of deciding this question is
to search for rapid variability in the in-
frared flux, since rapid variations would
denote small sizes which, coupled with
the observed luminosities, would pre-
clude thermal emission. A continuing
part of the long-term program of Becklin
and Neugebauer has been to monitor the
intensity of about 10 extragalactic
sources. Another method of discrimi-
nating between thermal and nonthermal
emission is to find the size of the emitting
region. During the last two years, the
sizes of two strong infrared sources have
been measured directly. The infrared
energy from NGC 253 from 2 to 20 p has
been shown to come from a core 10"
(about 150 pc) in diameter. In addition,
slit scans of the very strong infrared
source NGC 1068 show that it consists of
a core about 0."5 (or 50 pc) in diameter.
The data are not sufficiently good to de-
termine the profile of the core; models
with both gaussian and cylindrical geo-
metries fit the data. For both NGC 253
and NGC 1068 these emitting regions
are too large to be single nonthermal
sources. It seems, therefore, that the
radiation is probably thermal. It is
possible the radiation arises from heated
dust, as in many galactic infrared
sources.
2-20 Micron Band
During the past two years Wynn-Wil-
liams, Becklin, and Neugebauer have
been studying the 1-20 ^ radiation from
regions where star formation is thought
to be taking place. In particular, studies
of the infrared emission (1.6 fx < A < 20
jx) from H II regions have been continued
and extended; 12 regions have so far
been studied in detail (W3, W49, Orion
Nebula, M8, G333.6— 0.2, H2-3, W51,
NGC 7538, OH 0739—14, Sh 269, W75,
K3-50), leading to the discovery of a
large number of new infrared sources.
The H II regions studied have in general
been those that have displayed fine struc-
ture at radio wavelengths and are asso-
ciated with OH/H20 masers. Insofar as
it is possible to generalize at this stage,
it is found that:
1. Almost all compact (<20") H II
regions detected at radio wavelenghs
have strong infrared emission at 20 /x.
This emission probably arises from dust
mixed with the ionized gas and heated
to 100-200° K by the hot central star
of the condensation. Abnormal physical
properties or abundances of the dust do
not appear to be necessary to explain
the emission.
2. Most of the same objects show
emission at 2 /a, weaker than that pre-
dicted from the radio emission, a result
which we interpret as indicating the pres-
ence of very large amounts of extinction
between the sources and the sun. A
value of Av = 50 mag has been found in
one case.
3. Most, but not all, OH or H20
maser sources in H II regions are infra-
red sources. There seems to be a ten-
dency for these infrared sources, whether
or not they show radio continuum emis-
sion, to have a hotter energy distribution
and a stronger 10 /x silicate absorption
feature than those sources without maser
emission. There is no obvious relation-
ship between the flux densities in the in-
frared continuum and in the microwave
maser lines.
4. Some H II regions contain infrared
sources without radio counterparts. Of
these, two may be protostars. These
objects, in the Orion Nebula and in W3,
both have approximately blackbody
energy distributions, strong silicate ab-
sorption features, and OH or H20 maser
emission; they are less than 3" in di-
ameter.
HALE OBSERVATORIES
115
MILLIMETER-WAVE OBSERVATIONS
1-mm Photometry
A liquid-helium-cooled bolometer sys-
tem designed for 1-mm photometry at
the prime focus of the 200-inch telescope
has been developed. Successful observa-
tions with this system have been carried
out by Paul Harvey, Michael Werner,
Jay Elias, and Ian Gatley during twilight
hours.
A number of objects, including planets,
galactic H II regions, and the QSO 3C
273, have been observed. The most in-
teresting results to date are extensive
maps of the H II regions W51 and M42
(the Orion Nebula) . The map of the
Orion region is shown in Fig. 1.
The data show a ridge of enhanced
brightness extending some 3' to the north
and east from the central position. This
4'
3'
AS 0
-2'
-3'
— i 1 1 1 r
I mm Continuum Emission
Relative To K L Nebula
,0.2
2'
0
Aa
-I'
-2'
Fig. 1. One-mm continuum emission from the
Orion Nebula region, measured at the prime
focus of the 200-inch Hale Telescope. Positions
and fluxes are given relative to position M,
which coincides with both the Kleinmann-Low
infrared nebula and the center of the molecular
cloud. Position H is the center of the H II
region.
feature is similar in shape, size, and
position to features seen in maps made
of the Orion region in the emission lines
of H2CO (2 cm) and 13C1GC (3 mm).
The flux density at the central position
of the map is — 250 ± 100 f.u. at 1 mm.
From hydrogen recombination within the
H II regions alone, a flux of only ^25 f.u.
is expected at 1 mm — much less than is
seen. Furthermore, the distribution of
1-mm emission resembles more closely
that of the molecular cloud than of the
H II region. It is suspected, therefore,
that the 1-mm radiation is associated
with the prominent molecular cloud in
the Orion region.
Several mechanisms for producing this
1-mm radiation from the molecular cloud
have been explored. Molecular emission
from the cloud in the 1-mm region prob-
ably fails by at least an order of magni-
tude to produce this high a density within
the bandwidth used. On the other hand,
it seems very likely that the dust ex-
pected to be associated with the large
column density of gas in the cloud is
capable of thermally emitting the ob-
served energy.
Identification of this 1-mm radiation
with dust in the molecular cloud has
some interesting consequences. The in-
tensity observed at 1 mm is proportional
to the product of the grain temperature
and the column density of matter; its
distribution across the source presumably
represents the spatial variation of these
quantities. These two parameters of the
cloud are not accessible to the line spec-
troscopists, whose results, because of the
necessity of accounting for the excitation
of the molecular emission, reflect some
complicated function of the gas density
and temperature. The 1-mm observations
are therefore complementary to those
of the radio spectroscopists and should
advance our understanding of these ob-
jects.
The objects studied at 1 mm are not
visible against the bright twilight sky.
116
CARNEGIE INSTITUTION
A technique has been developed for morning and evening twilight periods,
pointing the telescope by offsetting from the 200-inch telescope can be utilized six
bright stars that can be seen through an to seven hours per day in this manner,
eyepiece on the photometer. Including thus effectively extending its usefulness.
INTERSTELLAR GAS AND GASEOUS NEBULAE
Interferometry of Nebulae with Low
Surface Brightness
Emission nebulae with low surface
brightness have been observed interfero-
metrically with the 40-inch telescope of
Las Campanas Observatory by Munch
and P. Cruvellier. The interferometer,
constructed for the purpose at the Ob-
servatoire de Marseille, produces Fabry-
Perot fringes of an area in the sky 25
arc min in diameter. The fringes are
focused on the cathode of an image in-
tensifler tube by means of an f/1.5 focal
reducer. With an S25 image tube, Ha
interferograms of faint emission regions
in directions close to the galactic center
of the Vela-Puppis nebular complex and
of some extended planetary nebulae were
obtained. The material is yet to be meas-
ured for detailed radial velocity studies.
However, visual inspection of the fringes
reveals large internal motions in the
planetaries NGC 246, 5189, 6164-65, and
6853. In contrast, the planetaries Abell
35 and 36 show small velocity disper-
sions. Interferograms in the line He I
A10830-29 were also obtained with an
SI image intensifier, but only of the
bright galactic nebulae Messier 8, 16, 17,
and 20 and that around -q Carinae.
PULSARS
Kristian is continuing a program of
monitoring the optical intensity of the
Crab pulsar at the 0.3% level to search
for possible secular changes due to the
slowing down of the spin.
X-RAY SOURCES
Cygnus X-l
Over the past year, a number of in-
dividual stars have been suggested as
candidates for the optical counterparts
of the galactic x-ray sources. A signif-
icant number of these are binary systems,
generally containing a luminous early-
type star. An important member of this
class is BD+34°3815 (HDE 226868),
which is a weak variable radio source
and is identified with Cyg X-l.
The star is a spectroscopic binary with
a period of 5.6025 days in which the
primary optical component has a spec-
tral type of BO lb and a velocity ampli-
tude of 74 km s-1. Brucato and Kristian
noted that the radial velocity of a weak
emission feature of He II at A4686 did
not have the same radial velocity as the
stellar absorption lines. During the past
year, observations made at the coude
focus of the Mount Wilson 100-inch in-
dicated that the emitting region is nearly
180° out of phase with the B star with a
velocity amplitude of around 140 km s"1.
Brucato and Kristian estimated the
mass of the unseen secondary (presumed
to be the x-ray source) by assuming that
the B star is normal for its spectral type
( — 22 371©). Combined with the mass
function
= 971/sinH = Q 2256
(Wlx + SfTCo) 2
the lower limit for the mass of the sec-
ondary is 5.5 gn0 (i — 90°). If it is
HALE OBSERVATORIES
117
assumed that the He II emission arises
in the vicinity of the secondary and thus
provides an estimate of its velocity, then
3TC.r "-' 16 9TC0. These estimates are sig-
nificant, since several workers have sug-
gested that x rays are produced by the
thermalization of gravitational energy of
matter falling onto the surface of a star.
To produce temperatures sufficiently high
to generate x rays, the star must be a
collapsed object. Because the lower limit
of 5.5 3TZ0 exceeds the maximum mass
of a white dwarf or a neutron star, four
possibilities exist: (1) The binary sys-
tem is the source of the radio emission
but not of the x rays; (2) the x rays are
produced by the mechanism outlined
above and the collapsed object is a black
hole; (3) the x rays are produced by
some means, as yet unspecified, which
does not require a collapsed secondary;
and (4) the basic assumption that the
mass of the B star is normal is incorrect.
To evaluate this last point, Brucato
and Kristian assumed that the x rays
are produced by infall to a white dwarf
or neutron star with mass less than
2 9frc0. In this case, the mass of the
primary is extremely peculiar, for it has
a mass of 4 2fH0 or less and yet success-
fully mimics the spectrum of a BO lb star.
The absence of eclipses in the x-ray band
and the assumption that the the B star
fills its Roche lobe further limits its
mass to ^3.15 9TC0. At the present time,
there is no evidence favoring any of these
models over the others.
Hercules X~l
The identification of the pulsating
x-ray source Her X-l with the vari-
able star HZ Herculis, made by
several observers at Princeton and
Harvard, was based on the 1.7-day
period observed at both optical and
x-ray wavelengths. Shortly thereafter, it
was reported that the 1.24-sec pulses
were also detected in the optical band.
Using the 60-inch reflectors at Palomar
and Mount Wilson, Zappala and Brucato
engaged in a program to verify this re-
sult. Fourier analysis of four nights of
photometric data did not show the pulsa-
tion with upper limits as low as 0.11%
of the total signal (at the 2a confidence
level ) . Two of these nights corresponded
to orbit phase 0.2, reported to be pre-
ferred for the appearance of the optical
pulsation. It is now known that, even
at this preferred phase, the variations are
only irregularly present. The nature of
this object is still obscure, and further
observations (particularly spectroscopic)
have been and are being obtained.
HZ Her is also being studied by Oke
and Greenstein with the multichannel
spectrophotometer and the coude image-
tube spectrograph. This fascinating star
shows an optical variation which requires
that the side facing the x-ray pulsar be
quite hot, and that the opposite side be
near 8000°. Resolution into two, or pos-
sibly three, different major components
on the surface of this optical object is
required by the multichannel observa-
tions. The amplitude of the light curve
in the ultraviolet is largest because of
the increased strengths of the Balmer
jump on the far side of the star. Coude
spectra by Greenstein show a spectrum
characterized almost completely by ab-
sorption lines. The lines are of diverse
widths, and the profiles of the hydrogen
lines seem complex and variable. Broad-
ening corresponding to velocities of from
300 to 500 km s"1 is indicated by He I
and Mg II at maximum light. However,
this may not, in fact, be due to rotation
but rather to stream motions. The spec-
trum at minimum is that of a late A-
type star and has relatively sharper lines.
This object will provide major astro-
physical puzzles, since the heating of the
side facing the x-ray pulsar may be
expected to have drastic results, such as
winds along the surface.
118
CARNEGIE INSTITUTION
SUPERNOVAE
Supernova Search
Twelve supernovae were discovered at
Palomar during the report year in the
course of the continuing search, which
is directed by Sargent and Searle. Ten
of the objects were found by Kowal and
two by graduate student J. Huchra, all
with the 48-inch schmidt telescope. In
addition, F. Zwicky discovered 18 super-
novae by comparing accepted and re-
jected plates obtained in the early 1950s
for the Palomar Sky Survey, making a
total of 68 found in this way. He found
that the Sc spiral NGC 5668 produced a
supernova in 1952 in addition to, that of
1954, which was discovered by P. Wild.
The survey of Sc I galaxies, which are
thought to have about one super-
nova every ten years per galaxy, con-
tinued with the 18-inch schmidt tele-
scope. No supernovae in the Sc I galaxies
have been discovered during the 2x/>
years since this program was begun. A
paper by Sargent, Searle, and Kowal on
the statistical results of the Palomar
supernova search in the interval 1958-
1971 was presented by Searle at the In-
ternational Conference on Supernovae
held in Lecce, Italy, in May 1973. The
supernova rate in bright galaxies is about
one every 30 years.
Substantial progress was made during
the report year on the project to set up
UBV photoelectric sequences in each of
the supernova search fields in order to
enable light curves to be determined from
old plates.
Astrophysical Studies
Searle and Oke continued their pro-
gram of observations of the spectra of
supernovae and, in particular, continued
to obtain spectrum scans of SN 1972e
in NGC 5253, the brightest such object
of recent years. Kirshner, Oke, Penston,
and Searle have prepared for publication
a report on the spectra of supernovae that
describes and interprets the spectrum
scans of several supernovae of both
Types I and II and includes a comprehen-
sive set of observations spanning 230
days of SN 1972e. These observations
represent the first major extension of
knowledge regarding spectra of super-
novae since Minkowski's great series of
observations on SN 1937c in IC 4182,
which were obtained 35 years ago. Three
general conclusions have been reached
from a study of the new data which pro-
foundly modify earlier conceptions of
the supernova spectrum. These are: (1)
that supernovae of Types I and II have
a continuum which changes slowly and
uniformly with time and which carries
the bulk of the radiated flux; (2) that
there exist lines in the spectra of both
Types I and II which have P Cygni pro-
files, i.e., broad and strong emission lines
flanked on the violet edge by broad and
deep absorptions; and (3) that lines exist
which persist through the evolution of
the spectrum and which are common to
Types I and II.
The most recent scans of SN 1972e
were obtained by Searle and Oke more
than one year following maximum light.
They show that while the supernova con-
tinues to fade the rate of spectral evolu-
tion is now very slow. The observations
cover the entire spectrum without gaps
from A3200 to Al0,500 in 40 A bands
shortward of A5800 and in 80 A bands
longward of this. Over the last 130 days
of the supernova's first year, the light
curve in each of these band passes can be
described, within the accuracy attained,
as an exponential decline at a steady rate
of 0.017 mag per day. This decline rate
is typical of other Type I supernovae, but
this is the first evidence that the decay
rate characterizes the entire spectrum
rather than one dominant spectral fea-
ture or an arbitrary band pass. It is
anticipated that scans of SN 1972e can
be obtained for at least one more year.
HALE OBSERVATORIES
119
Review of Spectra
The rapid developments in the spectros-
copy of supernovae, using the multi-
channel spectrophotometer, as reported
by Oke and others, made it desirable to
provide as complete a review as possible
of the earlier photographic data on spec-
tra. Such a review of the best available
data has just been published by Green-
stein and Minkowski; more than 20 dia-
grams show typical spectra, at various
phases, of supernovae of Types I, II, and
III, and IV. In general, line profiles are
of the P Cygni character, showing both
emission and absorption. No narrow fea-
tures are present except in Type V.
There is a small amount of fine structure
within the absorption dips, notably that
near A3800 ascribed by Oke et al. to Ca
II. The spectra of supernovae of Type I
seem to be remarkably uniform; Types
II and III are essentially gigantic ordi-
nary novae with high expansion velocity.
Some lines are common to Types I, II,
and III, but there are substantial differ-
ences, notably in the absence of hydro-
gen in Type I.
GALAXIES
Redshifts of Fainter Galaxies in the
Shapley-Ames Catalogue
In a continuation of the program to
map the velocity field for local galaxies,
Sandage spent seven weeks in Australia
during October and November 1972. He
observed redshifts of southern galaxies
with the Mount Stromlo 74-inch reflector
using a Carnegie image-tube spectro-
graph. Exceptionally good weather and
the cordial hospitality and cooperation
of the Mount Stromlo staff resulted in
completion of the southern galaxy pro-
gram for the remaining objects in the
&hapley-Ames Catalogue south of $ =
-30°.
During the three sessions on this pro-
gram (14 months in 1968-1969, six weeks
in February-March 1972, and the ses-
sion in this report year) , about 330 meas-
urable spectrograms on 290 different
galaxies were obtained. These have now
been measured and reduced and are being-
analyzed for the uniformity of the ve-
locity field.
A similar redshift program for bright
galaxies in the north that have no data
was carried on by Sandage as standby,
using the image-tube spectrograph at the
200-inch during partly cloudy nights.
At the end of the report year, there re-
mained 230 galaxies north of 8 = — 30°
to be observed before the Catalogue is
complete with at least one plate at-
tempted for each galaxy.
Relative Distance to Nearby E and
SO Galaxies
After six months of preparation, San-
dage and Visvanathan began routine
photometric observations with the Palo-
mar 60-inch of the energy distributions
of all bright E and SO galaxies to obtain
relative distances. More detailed cali-
bration of the color— absolute magnitude
effect than was available last year was
secured by observing most E and SO gal-
axies in the Virgo cluster brighter than
mpg = 13, using the Visvanathan auto-
matic scanner described in Year Book
71, p. 691. The color change with abso-
lute magnitude becomes steeper as one
proceeds farther toward the ultraviolet,
changing from A(U — V)/&MV = 0.08
mag per mag for (\v) = 3650 A to
A (3460 — V)/AMV = 0.15 mag per mag
and using a narrow-band interference
filter at (A) = 3460 A.
The goal of the investigation is to
combine distances determined in this way
with the new redshifts to test how nearly
the local flow is Hubble-like.
120
CARNEGIE INSTITUTION
TABLE 2. Space Densities* Log *(Afp)
Markarian
All Markarian
Seyfert
Field
Mp
Galaxies
Galaxies
Galaxies
-14
-1.74 (2)
-0.42 (1)
-15
-2.32 (1)
-1.12 (2)
-16
-2.32 (5)
-1.13 (4)
-17
-2.53 (14)
-1.58 (6)
-18
-2.97 (17)
-1.84 (15)
-19
-3.34 (31)
-4.43 (3)
-1.98 (41)
-20
-3.53 (74)
-4.52 (6)
-2.28 (77)
-21
-4.20 (59)
-4.80 (16)
-3.17 (33)
-22
-5.35 (12)
-5.75 (5)
-4.46 (6)
-23
-6.56 (4)
-6.92 (2)
(-6.65)
*Space densities are in Mpc~3 mag-1. H0 is assumed to be 75 km
s_1 Mpc-1. The number in parentheses is the number of galaxies on
which each estimate is based.
Motion of the Sun Relative to the
Nearby Galaxies
In a related program, G. A. Tammann
spent four weeks at the Observatories
working with Sandage's new southern
redshifts for supergiant Sc galaxies in
the first step for a new solar motion
solution of the sun relative to all nearby
galaxies. The results are promising and
show that the kinematic field for spirals
can be mapped in this way once the
necessary direct plates are available to
classify the galaxies into the van den
Bergh luminosity classes. Brucato and
later Babcock began to obtain direct
plates of the desired southern galaxies
using the Las Campanas 40-inch reflector
with a Carnegie image tube mounted at
the f/7.5 Cassegrain focus.
Markarian Galaxies
Sargent and graduate student John
Huchra obtained spectra of 26 galaxies
in Markarian's fourth list of galaxies
with an ultraviolet continuum, continu-
ing work done by Sargent in previous
years. All of the galaxies had emission-
line spectra. They included two new
Seyfert galaxies, Markarian Nos. 315 and
358. Sargent and Huchra then used data
obtained by various spectroscopists on
242 galaxies brighter than mp = 15.55
mag in Markarian's first four lists in
order to obtain an estimate of the space
density of these objects as a function
of absolute magnitude. This work is an
improvement over the estimates reported
earlier by Sargent (Astrophys. J., 173, 7,
1972) , which were based on data for only
about half as many objects. At the same
time, Sargent and Huchra redetermined
the space density of normal field galaxies,
using a list of galaxies brighter than 11.0
mag in the Reference Catalogue of Bright
Galaxies that had been prepared by grad-
uate student Paul Schechter. The results
are given in Table 2.
Markarian's fourth list is the only one
that includes a region of the south
galactic hemisphere. This region, unlike
those covered by lists I— III, does not
contain any nearby clouds or groups of
galaxies. There are indications in the de-
tailed statistics (not presented in Table
2) that Markarian galaxies fainter than
Mp = — 18 are preferentially concen-
trated in clouds and groups of galaxies
and that, in consequence, the high value
for their space density may be an over-
estimate.
Bright Spiral Galaxies
HALE OBSERVATORIES 121
History of Star Formation in Galaxies from both regions is starlight, but evi-
,.,,,. t , ,, dently the history of star formation has
Very little is yet known about toe been gi ificantl different in the tails
history of star formation in galaxies nor frQm ^ [q ^ m&in fa R
about what processes trigger the arge- ^ ^ formation of interaction tails
scale bursts of star formation that give inyolveg mQre than & A h d ica]
rise to such diverse phenomena as mi- perturbation of stars that exlsted before
clear hot spots the patchy appearance ^ laxieg interacted It ig h d that
of irregular galaxies, and to the very gtudies Qf guch turbed gtemg wiu
blue galaxies that have been extensively ;de & ugefu, meang Qf lon the
discussed in recent Year Books. In an gituationg that tri gtar formation in
attempt to throw some light on these mope near, norma] laxieg
problems, Searle has continued to obtain
multichannel scans of star clusters, blue
galaxies, and of the nuclear, disk, and
halo regions of spiral galaxies. Inde- Van der Kruit started his appointment
pendent parameters of the energy distri- as Carnegie Fellow with the completion
butions are being measured to study the of the discussion of his survey of the
stellar content and history of star forma- radio continuum brightness distribution
tion in these sample stellar populations, of bright spiral galaxies that he obtained
In connection with this work, Searle ob- with the Westerbork Synthesis Radio
tained at Las Campanas multicolor Telescope. In a number of galaxies the
photometry of 40 open clusters in the radio observations had shown that the
Large Magellanic Cloud. This work is radio ridges in the brightness distribu-
an extension of the study by Searle, Sar- tions were displaced toward the inner
gent, and Bagnuolo which was reported edges of the optical spiral arms, thus con-
in Year Book 70 (p. 423) and Year Book firming the concept of compressed regions
71 (p. 679). in the density-wave picture of spiral
In a related program, Searle and Sar- structure. This permitted determination
gent have obtained multichannel scans of of the strength of the density-wave corn-
ring galaxies. They find that the energy pression in a dozen galaxies by compari-
distributions of these are closely similar son of the radio-flux density of the
to those of star clusters a few hundred large- and small-scale structure, since
million years old. The same age is indi- the radio-spiral structure results from
cated for these galaxies from the meas- compression of the magnetic field and
ured expansions of the rings. Searle and relativistic electrons in the large-scale
Sargent believe that these galaxies, of "base disk." This property was related to
which VII ZW 466 is the best studied other integral properties and a few cor-
example, are the first truly young extra- relations were established and discussed,
galactic systems to be recognized. Although other relations were also found,
Searle and Sargent are applying similar these appear to be most fundamental,
techniques to the study of tidally inter- Strong compression galaxies turned out
acting galaxies and have found that the to be early (bright) van den Bergh lumi-
outlying fragments of interacting pairs nosity classes and to have relatively low
are systematically bluer than the main mass-luminosity ratios, which indicates
bodies. A good example is NGC 4676 that strong compression results in for-
(Arp 242) ; this interacting pair of gal- mation of either more stars or stars of
axies has been named "The Mice" by lower M/L. Also, these galaxies had a
Vorontsov-Velyaminov. Searle and Sar- small radius of the maximum in the ro-
gent find that the mice have red bodies tation curve compared to the radius of
and blue tails! They find that the light the outermost H II region, which natur-
122
CARNEGIE INSTITUTION
ally results in a relatively high velocity
between gas and wave pattern.
Three other results are:
1. Seyfert nuclei are the strongest
radio emitters at 1415 MHz (1020 to 1022
WHz-1 ster-1), while most normal spirals
have nuclei of 1018 to 1020 WHz"1 ster1.
2. A strong correlation between the
powers at 1415 MHz and 10/x for the Sey-
fert nuclei may persist (more weakly)
for nuclei of normal galaxies.
3. All disks with brightness tempera-
tures above 1°K have nuclei stronger
than 1019 WHz-1 ster"1, which may be
interpreted as meaning that although
generally cosmic rays in the disks are
accelerated by supernovae, the stronger
nuclei may also contribute substantial
amounts.
Having related the compression
strength to rotational properties, the next
step was to investigate those galaxies
with well-determined optical rotation
curves. From a sample of 26 spirals, the
relation with the luminosity class was
confirmed and a possible correlation with
the development of the nucleus ( Byurakan
type) was indicated. It was then shown
that the compressional strength did not
depend on the total mass or specific
angular momentum (angular momentum
per unit mass) , but on the ratio of these
properties. For a 'particular mass, strong
compression galaxies have a lower spe-
cific angular momentum than weak com-
pression galaxies. Statistically, this re-
sults in a larger central concentration of
mass for strong compression galaxies,
which explains the result that such gal-
axies have generally better developed
nuclei. If the nuclei are more active,
there may be important consequences in
the context of the possibility that nuclear
explosions regenerate density waves. This
suggests a closed cycle because strong-
compression results in transport of more
angular momentum, with enhancement of
nuclear activity.
To continue this line of research, spec-
tra were taken of NGC 4321 (M100)
with the Cassegrain spectrographs of the
Palomar 60-inch and 200-inch telescopes
for determination of the rotation curve.
This could serve as a confirmation of the
relations mentioned above, since the
Westerbork data suggested that NGC
4321 most likely has strong density-wave
compression. The characteristics of the
rotation curve and the derived mass-
luminosity ratio plus the luminosity class
did confirm the relations and the strong
compression in this galaxy. Evaluation
of all possible integral properties re-
vealed a close similarity with M51. In
view of the possibility that the spiral
structure of M51 is regenerated by the
passage of NGC 5195 and the absence of
equally massive companions to NGC
4321, this similarity is truly remarkable.
This means that the details of the gen-
erating mechanism, or possibly even the
mechanism itself, do not determine the
characteristics of the galaxy but that
they are determined by the character-
istics of the disk. Since apparently gal-
axies can be specified completely by only
two parameters (P. Brosche, Astron.
Astrophys., 23, 259, 1973), we then have
the following scheme to explain such
similarity. The parameters, mass and
specific angular momentum, initially
specify a galaxy. This means that the
compression strength is determined just
as is the form of the rotation curve.
Through star formation triggered in the
density-wave compression regions, the
development of other properties is now
governed by the compression strength or
is in that way determined by the same
two initial conditions independently of
the generating mechanism of the density
waves. Such a scheme explains the pres-
ence of only two dimensions in many-
dimensional galaxy-parameter space, as
well as the close similarity of NGC 4321
and M51.
An extensive spectroscopic study was
started by van der Kruit of those galaxies
for which the Westerbork survey indi-
cated strong recent activity in the nu-
cleus, in particular NGC 3310 (extremely
bright radio nucleus), NGC 4258 (com-
HALE OBSERVATORIES
123
plicated radio structure and Ha arms) ,
and NGC 4736 (triple radio source in the
central region).
Synthesis of the Stellar Population
of Galaxies
Gunn has obtained high-resolution
scans of four giant galaxies (NGC 4889,
4874, 6166, and 4472) and some sixty
stars to be used as material for stellar
population synthesis of these systems, in
collaboration with Dr. Beatrice Tinsley
of the University of Texas. The observa-
tional program is nearing completion, but
virtually no analysis has been done as
yet.
Photometry of OJ 287 and M82
The absolute-energy-distribution ob-
servations of OJ 287 were obtained by
Visvanathan with 20 A bandwidth in blue
and 40 A in red from 3300 A to 10,000
A on January 18, 19, 20, and 21, 1972.
The continuum is featureless without any
absorption or emission lines. Further,
the continuum can be represented by a
straight line with a slope of 1.25 in log
Fv — logv plot. Polarization observations
on January 19, 1972, in the wavelength
range 3300 A to 8000 A show that the
continuum is linearly polarized and P
and 0 are independent of wavelength.
The above observations are interpreted
to indicate that the continuum of OJ 287
is of synchrotron origin. A theoretical
synchrotron spectrum based on a simple
model of high-energy electrons having a
single power-law energy distribution is
computed and found to fit the observed
spectrum of OJ 287 from 3.5 mm to
visible wavelengths.
Visvanathan has obtained narrow-
band photoelectric scans (20 A in blue
and 40 A in red) of the region RD in
the filaments of M82 with the 200-inch
telscope. The intensity of emission lines
[O II], H7, H/?, [O III], Ha, [N II],
and of some absorption features has been
obtained. From the Balmer emission
lines of hydrogen, an interstellar absorp-
tion of Av ~ 2.0 mag is derived for the
region RD.
Chains of Galaxies
For the past several years, Sargent has
done observational work on compact
chains of galaxies. During the report
year work was completed on Markarian
6, a chain of six galaxies of 15 to 16 mag
at R.A. (1950) 16h55m0, Decl. — 81°39'.
All the galaxies in this chain have rough-
ly the same redshift; the mean is
cAA/A0 = 11,660 km s_1 with a dispersion
of 300 km s"1. Markarian 6 is about 340
kpc in length. Its dissolution time, if it
is really a linear formation in space, is
about 5 X 108 years, a typical value for
compact chains.
The apparent existence of chains of gal-
axies has in the past led to several more or
less bizarre explanations of their nature
and origin. For example, Burbidge, Bur-
bidge, and Hoyle (Astrophys. J., 138,
873, 1963) concluded that the chains
must be young systems which condensed
within the past 108 years from tubelike
magnetic formations in intergalactic
space. Both Vorontsov-Velyaminov (As-
iron. J., 66, 551, 1961) and Arp (Publ.
Astron. Soc. Pacific, 80, 129, 1968) have
expressed the view that certain chains
result from the ejection of galaxies from
the nuclei of parent galaxies. The idea
that true chains of galaxies are unstable
and therefore must have ages which are
less than the Hubble time seems un-
doubtedly correct. However, studies of
the UBV colors of member galaxies by
Sargent over the past few years have
led to no indications that they are young
systems. (Galaxies that have ages of
a few times 108 years should be recog-
nizable from their integrated colors, ac-
cording to the calculations by Searle,
Sargent, and Bagnuolo, mentioned in
Year Book 71, p. 679.)
A fairly mundane hypothesis which
does not seem to have been considered in
the past is the idea that linear formations
124
CARNEGIE INSTITUTION
N
I
Arp 330
E
. D
f C
* B
, A
2'
60kpc
Fig. 2a. A sketch of the chain Arp 330. It is
comprised of six member galaxies lying along a
somewhat broken line. The elongated form of
the large central galaxy (C) connects the AB
and DEF subchains.
in compact groups of galaxies represent
transient states in the normal dynamical
evolution of small clusters of galaxies
that are, on the average, spherical. This
hypothesis has presumably been ignored
because, intuitively, the chance that an
apparent chain of galaxies on the plane
of the sky results from such an accidental
alignment seems very small. However,
it is in practice hard to calculate this
probability because one's notion as to
what kind of formation constitutes chains
is very subjective. Accordingly, Sargent
and Edwin Turner, graduate student in
astronomy at Caltech, have investigated
the problem by carrying out extensive
numerical calculations of the evolution
of gravitationally bound clusters com-
posed of six galaxies represented as mass
points. In these calculations the initial
conditions were defined by randomly dis-
tributing 6 mass points in a sphere; each
mass point was randomly allocated 3
components of velocity arranged so that
the cluster was gravitationally bound.
The dynamical evolution of several such
systems was studied, using a computer-
printed picture of the cluster as seen from
three fixed, mutually orthogonal direc-
tions each crossing time. (This ensures
that the configuration of the "galaxies"
changes considerably between pictures.)
The resulting pictures were then in-
spected for chainlike formations in a
manner similar to that in which astron-
omers examine pictures of the sky. Con-
sequently, although the judgment as to
what constitutes a chain is still subjec-
tive, at least the same subjective criteria
are applied to the real and the simulated
pictures (see Fig. 2a, Fig. 2b).
Several series of computations pro-
duced a total of 297 pictures of 6-body
clusters: chains similar to the well-known
examples in the sky (a good collection is
6 D. 4 3Z (SP)
VIEW
t:
Do
o'
VIEW 2
tL
5
4
100 KPC
03
VIEW 3 l_^2
ofc
O
OS
Q4
Fig. 2b. Views of an assimilated cluster of
galaxies as it would appear from three ortho-
gonal directions. The very striking chain visi-
ble when the system is viewed along the second
axis disappears completely in the other two
orientations. A sixth object (No. 2) was ejected
to a large distance during previous dynamical
activity and is not shown.
HALE OBSERVATORIES 125
contained in Arp's A tlas of Peculiar Gab This work was inspired by Sargent's
axies) were found in 8% of these pictures, study of the cluster ZwCl 0152+33,
The chains arise by chance during the which is a highly elongated system on the
clusters' evolution. They appear and plane of the sky. Sargent suggested
dissolve in a fraction of a crossing time {Astrophys. J., 176, 581, 1972) that this
and from this point of view they are un- cluster might be a flattened system seen
stable, as astronomers have supposed, edge-on, but rejected the idea because
although the clusters are not unstable, no cluster with abnormally small velocity
The three-dimensional views produced in dispersions, which would be expected of
the calculations showed that all the simu- such systems if seen face-on, had been
lated chains were configurations of gal- discovered. However, almost immedi-
axies instantaneously in a plane; no ately Robinson and Wampler {Astrophys.
genuine three-dimensional chains were J. [Lett.], 179, L135, 1973) reported that
found. the compact cluster Shakhbazian I has
The proportion of projected chains the smallest known radial velocity dis-
found in these calculations is much higher persion, less than 60 km s"1. Sargent and
than one would intuitively expect. As a Turner's computer simulations show that
result, Sargent and Turner conclude that this cluster and ZwCl 0152+33 can
the most likely explanation of the well- barely exist as flattened systems for one
known chains of galaxies in the sky is Hubble time, if that is indeed their shape,
that they are due to transient projection In arriving at this conclusion, they found
effects in clusters that are on average it necessary when treating close en-
spherical. More exotic phenomena need counters to perform calculations making
not be invoked. a crude allowance for the fact that gal-
Sargent and Turner have also used the axies are not mass points,
computer simulations to investigate the
accuracy with which the virial theorem Radio Sources and Spiral Galaxies
can be applied to small clusters of gal-
axies. It is not possible to account for Tt has long been known that radio
the abnormally high mass-to-light ratios sources are often associated or paired
inferred for many small clusters. They with elliptical galaxies. Arp is now con-
are also investigating the extended dy- ducting investigations of the neighbor-
namical evolution of clusters that eject hoods of nearby spiral galaxies with the
one or two members because of close en- aim of cataloging all radio sources and
counters. In such systems the remaining optical objects that might be associated
objects assume a more compact configur- with the spirals. He used the 210-foot
ation. This process may form a basis Goldstone Telescope* for mapping areas
for explaining systems such as Seyfert's of UP to 4 square degrees around each of
Sextet, in which the galaxies are so close six sPiral galaxies in the magnitude in-
together that they should be torn apart terval 10.0 to 12.7. He reports that there
by mutual encounters on a time scale appears to be an excess of radio sources
much less than the Hubble time. The over normal background counts in the
galaxies in such systems show no spectro- 01 to °-2 flux-unit range around these
scopic or other indication of being young galaxies. The excess is about 80% at a
systems. 70% confidence level. To confirm the
Finally, Sargent and Turner have suggested association of radio sources
studied the dynamical evolution of flat-
tened clusters of galaxies in which the * The NASA^L Deep Space Network is
, ., ,. • • ii i. •• * operated by the Jet Propulsion Laboratory,
velocity dispersion in the direction of California Institute of Technology, under Con-
flattening is initially considerably less tract NAS 7-100 sponsored by the National
than that in the perpendicular plane. Aeronautics and Space Administration.
126
CARNEGIE INSTITUTION
with spiral galaxies, two team investiga-
tions have been organized with collabora-
tion from the Jet Propulsion Laboratory
and from the National Radio Astronomy
Observatory.
Areas around the six spirals referred to
are being analyzed by Arp on 48-inch
schmidt plates for the purpose of cata-
loging all peculiar galaxies therein.
Multiple Interacting Galaxies
Arp has listed all systems of three or
more bright galaxies that appear to in-
teract strongly on the basis of mutual dis-
tortion. He believes that these tend to
occur in aligned configurations, some-
times near galaxies of much larger ap-
parent dimensions and redshift.
Deep Photography of Galaxies with
Connected Companions
Arp continued the study of luminous
connections between galaxies. Plates of
NGC 4319 and NGC 4151 were analyzed
by means of the automatic isodensity
tracing machine (GALAXY) at the
Royal Observatory in Edinburgh, Scot-
land, in collaboration with N. M. Pratt.
Addition of density information from
three separate plates of NGC 4139 led
Arp and Pratt to suspect a connection of
uniform width between the galaxy and
the quasar Markarian 205, even though
the connection is not visible on any one
of these three moderately short-exposure
plates.
Additional objects analyzed in the
same way were the suspected galaxy-
quasar association of NGC 7413 and 3C
455, as well as 3C 33 and 4C 31.32.
Superposition Printing
In a joint project connected with John
Kormendy, a graduate student at Cal-
tech, Arp has obtained extremely deep
baked Illa-J exposures of the NGC
7331-Stephan's Quintet area of the sky
with the 48-inch telescope at Palomar.
The superposition printing of six such
plates produced a composite photograph
that reached farther than any previous
photograph with the 48-inch schmidt
telescope and about as faint as even deep-
limiting 200-inch exposures.
A network of luminous filaments was
revealed in the region between the large
Sb spiral NGC 7331 and the small inter-
acting group of galaxies in Stephan's
Quintet. Similar plates are being ac-
quired by Arp in the region north of NGC
7331 in an attempt to determine the fre-
quency of occurrence of such faint fila-
ments.
Nebulosity Near 3C120 and BL Lacertae
A project that, like the preceding one,
involves very low light levels of nebu-
losity as recorded by the 48-inch schmidt
is a study of the peculiar radio galaxies
3C 120 and BL Lac. Faint, fairly linear,
luminous filaments have been recorded
in the vicinity by Arp and Barbieri for
3C 120 and by Arp and Penston for BL
Lac. For 3C 120, spectra with the 200-
inch show emission nebulosity at galactic
redshift. It is desirable, therefore, to ob-
tain direct photographs in Ha emission
of 3C 120 in order to study the form of
this peculiar nebulosity.
Redshifts in Small Groups of Galaxies
As part of a continuing program, Arp
has obtained additional spectra and red-
shifts of galaxies in small groups and
associations. Most recent results are for
companions to Markarian 94, 459 (com-
panions E and W) , and 474. Spectra of
companions to NGC 4151, 2974, 2146,
3166, 7448, 157, 210, 1199, and 864 were
also obtained.
Compact Galaxies
Inspection of National Geographic So-
ciety—Palomar Observatory Sky Survey
plates by Zwicky has resulted in the
identification of some hundreds of com-
pact galaxies and "post-eruptive" gal-
axies.
HALE OBSERVATORIES
127
RADIO SOURCES
Identification of 3C Sources
Kristian, Sandage, and Katem are con-
tinuing a program begun in 1970 to es-
tablish the nature of all 3CR sources to
the 200-inch photographic limit, using
the very accurate radio positions being
measured by several groups. The pro-
gram involves position measurements of
optical candidates on 48-inch and 200-
inch direct plates, and photometry and
spectroscopy of new identifications. A
review by Kristian at the end of 1972
shows that one-third of all 3CR sources
are still not identified.
Radio Source Structure
Gunn and Malcolm Longair of Cam-
bridge have begun a program of image-
tube photography of all high-latitude 3C
radio sources, both to obtain identifica-
tions to the limit of the 200-inch tele-
scope and to make an effort to correlate
radio structure on the scale of a few
seconds of arc obtained with the 5-kilo-
meter Cambridge radio telescope with
optical structures on the same scales.
The most difficult part of the program
turned out to be the necessary astrom-
etry; serious distortion in the image-tube
plates and measuring difficulty with the
48-inch plates used to obtain secondary
standards limited the overall accuracy to
slightly less than 1 arc sec, whereas some-
thing better than twice as good had been
expected. A serious component in the
error is lack of good positions for the
primary standards. The AGK3 or some-
thing equivalent is badly needed for this
work.
Search for the Radio Source Near
Cygnus X-3
On 2 September 1972, this source un-
derwent an unprecedented outburst, in-
creasing its radio brightness by a factor
of 1000 within 2 days. During the next
week, a search was made for a visible
counterpart of the source by Westphal,
Kristian, Huchra, Shectman, and Bru-
cato, using the silicon vidicon photo-
meter at the 200-inch prime focus and
direct plates at the 48-inch. No visible
object was detected in the range AA3500-
8000 to a limit of 23 mag. A month after
the outburst, however, an infrared search
by Becklin, Kristian, Neugebauer, and
Wynn-Williams yielded an identification
at 1.6 and 2.2 /a.
QUASARS
Quasar Surveys
Spectroscopic studies reported by
Schmidt in previous annual reports have
established the redshift distribution of
quasars of 18th magnitude. Such infor-
mation is lacking for brighter quasars
since no substantial complete samples are
available. Richard Green, a graduate
student, has started a survey with the
18-inch schmidt telescope that will
eventually cover about 10,000 square de-
grees. Blue and ultraviolet exposures on
the same film are used to search for ob-
jects brighter than 16 or 17 mag with an
ultraviolet excess. Spectroscopic work
will be undertaken to eliminate the white
dwarfs and hot blue stars that constitute
a large fraction of the candidates.
Spectroscopic Observations
Spectroscopic observations of quasi-
stellar sources identified by D. Shaffer in
the 6-cm National Radio Astronomy Ob-
servatory survey are carried out by
Schmidt. This work will yield a com-
plete sample of quasi-stellar sources
selected at high radio frequency. Such
a sample is essential for the interpreta-
tion of high-frequency radio source
counts that contain a large fraction of
quasi-stellar sources.
128
CARNEGIE INSTITUTION
Quasars as Events in Nuclei
Kristian has extended the search for
galaxies underlying quasars and has
studied the general problem of detecting
underlying galaxies on direct plates. If
a galaxy is present, it will not be seen
if its image is smaller than that of the
quasar. The image size of the quasar, as
of any point source, depends only on its
brightness, while the image size of a
galaxy depends on the actual size and
distance of the galaxy and is known to
have a well-defined upper limit at a given
redshift. By measuring stellar image
sizes as a function of brightness and using
Sandage's measurements of bright gal-
axy sizes, Kristian divided the Hubble
diagram (magnitude versus redshift) for
quasars into regions in which a quasar
image is larger or smaller than a giant
galaxy image, and he was able to predict
in individual cases whether or not an
underlying galaxy should be seen if it
were present. An examination of 200-
inch plates of 26 quasars gave results in
good agreement with the predictions:
Those quasars predicted to show a galaxy
did so, and those quasars predicted not
to show a galaxy did not. Galaxy images
concentric with the quasars were clearly
seen in five cases, and some evidence was
seen in six additional cases. Four other
quasars showed off-center structure.
The available evidence is consistent
with the hypothesis that all quasars occur
in the nuclei of giant galaxies, but the gal-
axies are not seen in most cases because
their light is overwhelmed by that of the
much brighter quasar.
The original definition of a quasar in-
cluded the requirement that it have a star-
like image on a direct plate. This require-
ment, however, was intended at the time
to distinguish quasars from large nearby
radio galaxies, and it appears that
whether or not it is met literally depends
on such accidental factors as atmospheric
seeing and telescope focal length.
Association of Quasars and Clusters of
Galaxies
Forty-eight-inch plate material is es-
sentially complete for the sample of
quasars with redshifts z less than 0.36.
Of the sample of 28, good plates now exist
for 26. For more than half of these,
"suspect" galaxies exist that are candi-
dates for redshift determination. Thus
far, six objects are known to be associ-
ated with groups or clusters of the same
redshift: B264 (0.06), Ton 256 (0.13),
PKS 2251+11 (0.32), 3C323.1 (0.27),
PHL 1093 (0.27), and PKS 1049—69
(0.33). All have been published except
the last two, which were measured in late
1972. Gunn and French, an undergrad-
uate, have undertaken a statistical sur-
vey of the plates to determine whether
QSOs are significantly correlated with
the general distribution of galaxies to the
plate limit, using randomly selected con-
trol fields. The counting is virtually com-
plete, but the analysis has not yet begun.
Small Quasars at the Center of
N Galaxies
N galaxies, as a class, are intermediate
between Seyfert galaxies and quasars in
the strength of their central bright blue
nucleus relative to the total light. The
spatial distribution of light was studied
in a number of N galaxies during the
report year in an effort to separate the
two luminous components of these sys-
tems.
Sandage made UBVR photoelectric
color measurements of 12 N galaxies,
using a number of different sky-blocking
apertures at the 200-inch telescope with
its large focal-plane scale. The measure-
ments showed a marked reddening and an
increase in intensity as the apertures
were enlarged on each N galaxy. The
color and intensity gradients permitted
a separation to be made of the two com-
ponents to the light. It was found that
the central blue source had the photo-
HALE OBSERVATORIES
129
metric properties of a quasar, while the
extended source had colors and the in-
tensity distribution of a normal giant
elliptical galaxy.
This composite model of N systems,
with a mini-quasar in the nucleus of the
giant E galaxy, resembles the model of
quasars themselves that is required from
the data reported last year that all
quasars lie on the bright side of the Hub-
ble line for giant E galaxies in the red-
shift-magnitude diagram, and from
Kristian's direct observations of the un-
derlying galaxies. Calculations were
made by Sandage for the intensity and
color contamination by the central
quasar on the combined light of quasar
and galaxy. Good agreement was ob-
tained between the observations and the
predictions for the place in the two-color
U — B, B — V diagram of both quasars
themselves and for N galaxies as the
contamination ratio is changed. The re-
sults support the conventional interpre-
tation that quasars are the nuclei of
some galaxies.
Observational Limits to Nonvelocity
Redshifts for the Mini-Quasars in
N Galaxy Nuclei
The separation of the galaxy and
quasar components in the total light of
N galaxies permits a unique test for a
limit to any nonvelocity redshifts for
the compact quasarlike central source.
The redshifts that are measured in
N-galaxy spectra are always emission-
line shifts from the central mini-quasar;
the spectrum is not from the underlying
E galaxy. Hence, if there were a large
nonvelocity component to this redshift,
the redshift-apparent magnitude dia-
gram for the underlying galaxy, plotted
using the measured redshift for 'the cen-
tral mini-quasar, would show a displace-
ment relative to the points for normal
giant E galaxies, whose absorption-line
redshifts are clearly cosmological. San-
dage made the test with the 12 N galaxies
previously studied, using the apparent
magnitudes of the underlying E galaxy
that were separated out. The Hubble
diagram, given as Fig. 3, shows no devia-
tion of the N galaxy E components from
the mean relation of similar E galaxies
in the field. Any nonvelocity redshift is
smaller than the intrinsic dispersion of the
sample. The limit to such a nonvelocity
effect AZ, put by analysis of the scat-
ter, is v(AZ/Z) < 0.1, which means that
any such postulated component is less
than 10% of the cosmological redshift,
and from these data it could be zero.
Sandage concludes that there is no evi-
dence from N galaxies to support non-
cosmological redshifts in quasars, in
view of their absence in the compact
mini-quasars of the N systems. Details
of the test were published during the
report year (Astrophys. J., 180, 687,
1973).
Properties of Quasars
Simultaneous observations of blue
magnitude, linear polarization, wave-
length dependence of polarization, abso-
lute energy distribution of the con-
tinuum, and strength of the emission lines
obtained in the years 1968-1971 for the
variable quasars 3C279, 3C345, 3C446,
and 3C454.3 were discussed by Visva-
nathan. He found that the continuum
was linearly polarized in all these objects
during outburst and nonoutburst periods
and when the objects reached minimum
brightness. AVhenever the total flux in-
creased, an increase in polarized flux was
observed. The position angle of the elec-
tric vector changed during the burst. The
ratio between the polarized flux and the
total flux does not exceed a value of about
0.16. This low value is interpreted as
being due to the polarization caused by
several active spots in the continuum-
emitting region. The absolute energy
distribution of the continua both in active
and quiet periods shows that the power-
law representation of the continuum is
valid. This fact and the finding that the
polarization is independent of wave-
130
CARNEGIE INSTITUTION
5.4
O 4.6
<
o
o
3.8
3.0
• RADIO GALAXIES
A UNDERLYING GALAXY
TO N SYSTEMS
3C 287.1
3C 390.3
3C37I —
3C 445
3C 120
3C303
3C227
8
16
20
V26-Kv-Av
Fig. 3. A comparison of the Hubble diagram for the underlying galaxy component of N sys-
tems with aormal radio galaxies. The agreement of the two sets of data indicates that the quasar
components of the N systems have no nonvelocity components of redshift to better than 10%.
length support the view that the radia-
tion in these objects in the observed fre-
quency range is of synchrotron origin
in both active and quiet periods. Varia-
tions in the spectral index of the continua
were observed, but no simple relationship
between color and brightness was found.
During the period of the observations,
the absolute intensity of the emission line
Mg II in 3C345 and 3C454.3 and the C
IV line in 3C446 remained constant
within 10% to 20%. The emission line
Mg II in 3C345 was unpdlarized, indicat-
ing that the regions from which the
emission line originates are randomly ar-
ranged.
Angular Diameters of Quasars
Van der Kruit discussed the inferences
drawn from the distribution of (radio)
angular diameters of quasars. It was found
that the distribution with redshift is en-
tirely compatible with evolving universes
and with the Rees and Setti (Nature,
219, 127, 1968) evolution of physical di-
ameters ; it was further indicated that an
apparent gap in the linear size distribu-
tion (Reinhardt, Astrophys. Lett., 12,
135, 1972) can very well be caused by an
observational effect, namely, the distri-
bution of the interferometer baselines
used in the observations. Rowan-Robin-
son's {Nature, 236, 112, 1972) proposition
HALE OBSERVATORIES
131
that some quasars are local, derived from
the uniform V/Vm distribution of small
diameter 3CR quasars and their associa-
tions with peculiar galaxies and Zwicky
clusters, was examined and it was found
that "compactness" of the spectra is a
better criterion than linear size. The uni-
form V/Vm distribution is a property of
only a 3CR sample with respect to its
optical and radio detection limits. All
associations between 3CR sources and
Zwicky clusters on the sky are examined.
From the number of such associations of
radio galaxies with clusters of a different
redshift, it is shown that the agreement
in position between 3CR quasars and
Zwicky clusters at a different redshift
is accidental. It is also shown that there
is no correlation between compactness of
the spectra and membership of Arp's
pairs of radio sources across peculiar
galaxies. It is concluded that the angular
diameters of quasars in no way contra-
dict the cosmological interpretation of the
redshifts.
OBSERVATIONAL COSMOLOGY
Colors of E Galaxies as an Indication
of Age
It has been known since the early
work of Stebbins and Whitford at Mount
Wilson in the late 1940s that the giant E
galaxies show only a small dispersion in
color distribution. This fact was used by
Baade and others as an argument in
favor of nearly equal ages for E galaxies,
in contrast to the continuous age distri-
bution required by a steady-state cosmol-
ogy where galaxies of all ages would
be found.
New BVR observations of giant E gal-
axies in clusters and groups were begun
by Sandage in 1967 for a study of the
redshift-apparent magnitude relation ;
the distribution of colors is a by-product
of the photometry. Analysis of the data
was completed during the report year,
and comparison has been made of the ob-
served color distributions with three
models for the age distribution of giant
E galaxies.
The observations, made with the
Mount Wilson 100-inch Hooker reflector
for the brighter galaxies in groups and
with the Palomar 200-inch for the fainter
clusters, were corrected for the effects
of latitude reddening due to our Galaxy,
redshift effects on the colors, and meas-
uring errors. An upper limit to the dis-
persion of the color distribution for giant
E galaxies is a{B — V) = 0.028 mag,
with a distribution about the mean that
is closely gaussian.
To what extent does this mean similar
ages for E galaxies? Stellar aggregates
become redder as they age. An earlier
study (Astrophys. J., 138, 863, 1963) of
galactic clusters of known age and known
integrated B — V color shows that the
rate of reddening for these simple struc-
tures is &(B — V) ~ 0.3 mag between
ages of 10° and 1010 years. Rates for E
galaxies are expected to be less but not
smaller than '—0.1 mag in the same time
interval. The variation of color index
with time will follow closely the form
B — V csl a log t -\- b, where a is the
reddening rate.
Using this reddening rate, Sandage pre-
dicted the shape of a color distribution
curve for E galaxies in the situations
where (1) E galaxies have formed con-
tinuously over the past 10 billion years
at a constant rate, so that there would be
old and young galaxies in the sample;
(2) the age distribution is that required
by the steady-state model where new
galaxies are born at the rate needed to
keep the space density of galaxies con-
stant in the presence of expansion that
moves galaxies of all ages out of the
volume; and (3) E galaxies formed over
a small interval of time near 1010 years
ago, as in a Friedman-Lemaitre big-bang
model.
The predicted color distributions of
132
CARNEGIE INSTITUTION
models (1) and (2) for reddening rates
between 0.3 mag/dex and 0.1 mag/dex
are strongly nongaussian and have a
much wider spread than a(B — V) =
0.028 mag of the observations. Conse-
quently, in agreement with earlier con-
clusions of Baade, Sandage believes that
the similarity of colors of giant E gal-
axies requires that they formed over
some small but definite interval early
in the life of a Friedman-like big-bang
universe. The results are given in Paper
V of the redshift-distance series (Astro-
phys. J., 183, No. 1, 1973).
Absolute Magnitude of First-Ranked
E Galaxies in Clusters and Groups
Richness effect. The effect of the rich-
ness of an aggregate of galaxies on the
absolute magnitude of the first-ranked
E galaxies was further investigated by
Sandage, extending the results reported
last year. Photometry of brightest gal-
axies in certain small groups that were
isolated in the 1956 study of redshifts
by Humason, Mayall, and Sandage was
reduced and corrected for aperture effect,
iv-dimming, and galactic absorption. The
data were plotted in the Hubble diagram
for first-ranked E galaxies in great
clusters (Year Book 71, pp. 689-690) and
the deviations from the mean line were
analyzed as a function of group or cluster
richness. Surprisingly, there was no sig-
nificant correlation: The absolute magni-
tude of first-ranked galaxies in groups as
small as 5 to 10 members is the same
(to within la of the distribution) as in
great clusters such as those in Coma or
Virgo, which have hundreds of members.
From this it would appear that there is
something special about the formation of
first-ranked group members: A galaxy
grows to a maximum size (luminosity)
initially, limited by some unknown effect
that produces an upper bound in mass
that is sharp at the <t(Mv) ~ 0.3 mag
level (i.e., = 30% in mass). This upper
limit is independent (or nearly so) of the
total amount of matter (richness of the
cluster) available at formation.
Effect of Bautz-Morgan cluster form.
Sandage and Hardy investigated the ab-
solute magnitude of the first three-
ranked galaxies in 97 groups and clusters
of galaxies as a function of the form of
the cluster given by the Bautz-Morgan
classification scheme. This classification
is based on the luminosity contrast of the
brightest cluster member with the rest of
the cluster.
The absolute magnitude of the first-
ranked galaxy is well correlated with
Bautz-Morgan class. In Class I clusters,
where the contrast is greatest, the first-
ranked galaxy averages 0.4 mag brighter
than the mean Hubble line in the magni-
tude-redshift line. However, startlingly,
the second- and third-ranked members in
Class I clusters are fainter than normal.
This inverse effect appears to be well
determined and suggests that the domi-
nance of first-ranked galaxies in clusters
occurs at the expense of the fainter mem-
bers.
The Bautz-Morgan contrast classes are
not correlated with cluster richness. From
this, Sandage and Hardy argue that the
dominance of a first-ranked galaxy was
not acquired by growth during the life-
time of the cluster by accumulation of
matter from the less massive members
(by tidal encounters, for example) be-
cause the frequency of such encounters
would depend on richness, contrary to the
observations. They believe, therefore,
that the Bautz-Morgan effect is more
probably related to an initial condition
of cluster formation rather than to later
evolutionary processes.
Corrections for the variation of <Mr>
with Bautz-Morgan cluster type were
made to all data, and new correlations of
<Afvc> with cluster richness again showed
that <Mvcy for first-ranked cluster gal-
axies had no significant variation (there
was a la effect) with richness as the
cluster size changes from 5 to 250 mem-
bers in the first three magnitudes of the
cluster. However, <Mvcy for second-
HALE OBSERVATORIES
133
and third-ranked cluster members was
strongly dependent on richness.
The effect of the Bautz-Morgan cor-
rection, and the slight (lo-) richness cor-
rection on the redshift-magnitude dia-
gram, was analyzed by making a new
solution for the apparent value (no
evolutionary correction) of the decelera-
tion parameter after applying the new
corrections. The change of q0 due to
these corrections was negligible. The new
value is q() = 1 ± 1 (the quoted error is
a 2<r value) , which is the same as reported
last year. Only the grossest alternatives
to this value (such as q() = — 1 or
g0 > 3) can be discarded by the pres-
ent data. New data for many clusters
with large redshift (z > 0.4) are needed
for a finer solution, but Sandage and
Hardy believe that the prediction of
steady-state cosmology {q0 = — 1) is
clearly at variance with the present data.
Distant Clusters
As part of the collaborative effort by
Kristian, Westphal, and Sandage to ex-
tend the redshift-magnitude diagram to
very distant clusters with Westphal's
new Silicon Intensified Target vidicon
system, the first part of a catalog of
distant cluster-candidates was completed
by Sandage during the report year. A
deep Illa-J photographic survey with
the Palomar schmidt has been in progress
over both the north and south galactic
polar caps since 1969. To date, some 25
fields have eben obtained with 2^/2-hour
exposures that reach 2 mag fainter than
the red Sky Survey plates. Six fields have
been inspected for candidate clusters that
have estimated redshifts larger than z ^
0.25. Many faint groupings are present.
Only the best candidates have been re-
tained in the working list for the SIT
vidicon. The number of these candidate
clusters averages 15 per field. Test
plates on 3C295 show that the fainter
one-third of the candidate clusters should
have redshifts larger than z ~ 0.46.
Routine observations using the SIT
vidicon were begun on the candidate
clusters by Kristian, Westphal, and
Sandage during the report year. Other
clusters for which conventional photo-
metry exists were also observed in order
to connect the program to existing data
(cf. Instrumentation).
The ultimate aim of the program is to
obtain magnitudes, redshifts, colors, and
diameters of galaxies in about 100 faint
clusters in order to distinguish among
world models. The first part of the pro-
gram, which is now under way, is broad-
band (BVR) photometry, yielding mag-
nitudes, diameters, and colors directly.
In addition, the V — R color provides
a rough redshift as a guide for the second
stage of the program, which calls for ob-
taining line spectra for redshifts and
energy distributions to investigate pos-
sible evolutionary effects.
Gunn and Oke have pursued an inde-
pendent program to find faint clusters
of galaxies. Their photographic search,
made with the 48-inch schmidt and deep
Illa-J plates, in practice appears to be
capable of finding clusters with values of
z up to about 0.30. The 90-mm image
tube has been used at the prime focus of
the 200-inch telescope to photograph
about 5 to 6 square degrees of sky with
deep Illa-J plates. This involves about
100 plates so far. A number of very
distant clusters of galaxies have been
found on these plates.
Oke and Gunn have continued to ob-
tain spectral energy distributions and
redshifts of galaxies in distant clusters
discovered in the survey referred to
above. Objects observed during the year
that have redshifts z > 0.20 are given in
Table 3 below. The program has pro-
duced data for a total of 18 galaxies with
redshifts greater than 0.20. Energy dis-
tributions have also been obtained of
galaxies in nearer clusters, including sev-
eral Abell clusters and the Hydra cluster.
These data will be used for synthesis
work.
134
CARNEGIE INSTITUTION
TABLE 3. Magnitudes and Redshifts
Object
V
z
1600+42, 48" field
Cluster 46
18.7
0.23
Galaxy Gl near PKS 1049-09
20.4
0.33
Galaxy Gl near PHL 1093
20.6
0.27
SSC J4
Galaxy G1+G2
19.6
0.26
Abell 2317
Galaxy Gl
18.3
0.21
Abell 1961
Galaxy Gl
17.8
0.23
Cluster 1319+30
Galaxy G2
19.2
0.27
Analysis of the new redshift material
has begun; as yet no firm results have
emerged because of the difficulty of esti-
mating corrections for the large selection
effects apparent in the data.
Radio Galaxy Source Counts
Schmidt has studied radio source
counts in steady-state cosmology on the
basis of identifications and redshifts of
strong 3CR sources. The source counts
corresponding to each of the identified
sources are derived from the relation be-
tween volume distance and flux density
generated if the source is hypothetically
moved to different redshifts. The method
allows all the individual properties of the
identified sources to be represented in the
predicted source counts. Quasars were
assumed to be local, since a cosmological
interpretation of their redshifts leads to
<y /ymy = 0.75 in steady-state cosmol-
ogy, a value incompatible with steady
state in which no evolutionary effects are
admitted. The predicted total source-
count curve is a factor of 5 below the
observed counts at low radio flux densi-
ties. The conclusion is that radio source
counts are incompatible with steady-
state cosmology.
THEORETICAL STUDIES
Chemical History of Galaxies
Searle continued theoretical studies of
element enrichment in galaxies. It is gen-
erally believed that galaxies collapse
from clouds of primordial material and
that the heavy elements are formed in
massive stars and are ejected from them
into the interstellar gas. Each succeeding
generation of stars is formed from pro-
gressively more enriched interstellar gas.
Ten years ago Schmidt found that if the
distribution over mass of newly formed
stars is the same at all times, it is difficult
to reconcile this hypothesis with the ob-
served scarcity of metal-poor stars in our
own Galaxy. He proposed that in early
times the stars that formed were mainly
short-lived massive ones.
Searle has reexamined this conclusion
in an attempt to account for the observa-
tion that the interstellar gas in all normal
galaxies has a composition very similar
to that in the solar neighborhood. He
finds that this observation is very hard
to understand unless the distribution over
mass of newly forming stars is every-
where nearly the same, averaged over
sufficiently large volumes and times.
Searle has shown that Schmidt's conclu-
sion follows only if it is assumed that the
interstellar gas of the collapsing galaxy
is homogeneous at all times. He has
proposed as an alternative model for the
early evolution of galaxies that they are
chemically inhomogeneous during their
collapse and that the stars form prefer-
entially in regions where the metal con-
tent is unusually large. With models of
this type, Searle has been able to ac-
count both for the scarcity of metal-poor
HALE OBSERVATORIES
135
stars in the Galaxy and for the nearly
uniform chemical compositions found in
the interstellar gas of galaxies that have
otherwise grossly different integral prop-
erties.
Diffuse X-Ray Background
Gunn has constructed mathematical
models of the distribution of clusters of
galaxies and their evolution in order to
compute the expected contribution of
clusters to the diffuse x-ray background.
The conclusion is that clusters must con-
tribute between a quarter and a half of
the total observed background in the 0.5-
10 keV range and may account for essen-
tially all of it. A satisfactory theory of
the fluctuations has also been developed,
and definite predictions can be made that
can be checked by x-ray detection of
moderate spatial resolution ; existing data
are not quite good enough.
Magnetic Field of the Crab Nebula
Gunn and Martin Rees of Cambridge
have developed a model for the genera-
tion of the magnetic field of the Crab
Nebula and other supernova remnants
which combines the wave ideas of
Ostriker, Gunn, and Pacini and the uni-
polar inductor picture of Goldreich and
Julian. Explained in a natural way are
the field geometry as inferred from polar-
ization and the cavity -wisp structure ob-
served near the pulsar. The work is
proceeding.
Penumbral Waves and Umbral Flashes
in Sunspots
Observed properties of umbral flashes
and running penumbral waves were re-
ported last year by Zirin, Tanaka, and
Stein. During the present report year,
Moore concluded that a likely source of
these periodic phenomena is the oscilla-
tory convection which is expected to oc-
cur in the superadiabatic subphotospheric
layers of sunspot umbras due to the
presence of a strong (2000-3000 gauss)
vertical magnetic field. To test this idea,
Moore computed the periods and growth
rates for oscillatory modes arising in a
simple two-layer model umbra. The re-
sults indicate that umbral flashes result
from disturbances produced by oscilla-
tory convection occurring in the upper
subphotospheric layer of the umbra
where the superadiabatic temperature
gradient is much enhanced over that in
lower layers, while running penumbral
waves are due to oscillations in a layer
just below this upper layer. The oscil-
lations in the lower layer may be driven
by the oscillations in the upper layer,
which would explain the phase-locking
between umbral flashes and running pen-
umbral waves.
Photospheric Oscillations
The well-established cause-and-effect
notion for the 5-minute oscillations ob-
served in the solar photosphere is that the
oscillations are excited by disturbances
generated by the motions in the convec-
tion layer just below the photosphere.
Moore has investigated the photospheric
oscillations by analyzing the steady re-
sponse of an isothermal atmosphere to
pressure fluctuations at its base. This
work was stimulated in part by the work
of G. Worrall (Astrophys. J.\ 172, 749,
1972), who concluded that the maximum
response of the atmosphere occurs at a
fixed frequency (og, the maximum cutoff
frequency for gravity waves (buoyancy
waves), regardless of the horizontal scale
of the perturbation applied to the base
of the atmosphere. Moore has found that
this behavior occurs only for downward
traveling waves and hence does not apply
to the case at hand, since causality re-
quires the response of an atmosphere per-
turbed from below to be due to waves
traveling up from the disturbance rather
than downward from infinity. When
downward traveling waves are excluded,
Moore finds that the response does de-
pend on the horizontal scale of the ap-
plied perturbations. For horizontal scales
136
CARNEGIE INSTITUTION
much longer than about 10 H, where H
is the scale height of the atmosphere,
the maximum response occurs at wa(^
o)g) , the minimum cutoff frequency for
compression (acoustic) waves. Only for
perturbations of horizontal scale much
shorter than 10 H does the maximum re-
sponse occur at <og. These results imply
that the photospheric oscillations of ob-
served horizontal scale ^ 4000 km are
compression modes oscillating at <oa
rather than buoyancy modes oscillating
at wg. Either wa or u>g gives an oscillation
period in the vicinity of 5 min, so it is the
horizontal scale rather than the period
that indicates the physical nature of the
oscillations.
White Dwarfs
Borra has computed models for the
Balmer lines of DA white dwarfs with
strong magnetic fields. It appears that
the interpretation of quadratic Zeeman
wavelength shifts to obtain, the mag-
netic field is not simple. Different parts
of the line give significantly different
shifts. In particular, emphasizing the
core of the line leads one to underestimate
the violet shifts and therefore the mag-
netic field. The line profiles and the
shifts depend also on the geometry of the
magnetic field.
Magnetic Field of 53 Camelopardalis
Borra has computed models for the
magnetic field of the star 53 Cam. It is
found that the linear polarization ex-
pected from the transverse Zeeman effect
in H/3 is two orders of magnitude lower
than the linear polarization reported re-
cently in H/3 by Kemp and Wolstencroft
(Astrophys. J. [Lett.], 179, 153, 1973).
It is unlikely that errors in the magnetic
field configuration used or the approxi-
mations in the models can explain the
discrepancy.
Zeeman-Broadened Line Profiles
Borra has done extensive computations
of the Stokes parameters (intensity,
linear, and circular polarization) as a
function of wavelength across a spectral
line formed in the presence of a magnetic
field. The models are oblique rotator
models with dipolelike configurations of
the magnetic field. Rotation is taken into
account. The polarization profiles will be
used in the interpretation of the photo-
electric observations currently being
made.
Theoretical Zeeman-analyzed line pro-
files are obtained from the Stokes param-
eters. They are used to simulate photo-
graphic observations of longitudinal mag-
netic fields. It is found that the longi-
tudinal fields inferred from those profiles
can differ substantially from the true
longitudinal field when only parts of the
lines are used to compute the Zeeman
displacements. In particular, very de-
centered dipolar fields give profiles such
that photographic spectroscopy will not
reveal a reversal of polarity. This might
explain the excess of stars that do not
reverse sign of the longitudinal field.
Many of the magnetic curves obtained
from the models derived from only the
line-profile cores mimic very well the
curves of known magnetic stars, while the
true curves from the models utilizing
complete line profiles do not.
INSTRUMENTATION
Prime-Focus Corrector
For the prime focus of the 200-inch
telescope, a new field correcting lens de-
signed by C. G. Wynne has become avail-
able for use. The lens, having four ele-
ments all of UBK-7 glass, was con-
structed by the Optical Shop. It provides
a field of excellent definition 25 minutes
of arc in diameter. Between the limits
3650 and 10,000 A, the aberrations of the
lens were computed to be less than 0."3.
HALE OBSERVATORIES
137
With either photographic plates or the
90-mm image tube, it provides very sig-
nificant gains.
SIT-Vidicon Area Photometer
The SIT-vidicon photometer, described
in Year Book 71 (p. 705), has been fur-
ther developed by Westphal, Kristian,
and Sandage and has been advanced from
a developmental to an operational stage.
Much of the effort this year has gone into
adapting it for use at the prime focus of
the 200-inch, where it is being used regu-
larly, and in developing techniques for
handling the large amount of data. Raw
data (65,000 elements per exposure) are
collected and stored digitally at the tele-
scope and reduced later. The reduction
ultimately requires the use of a large
computer, but special hardwired equip-
ment has been designed and built to
facilitate and simplify early stages of the
reduction. A playback system reads and
stores the digital data tapes, either com-
plete frames or selected parts of frames,
processes them, and displays them in
analogue form on an oscilloscope from
which photographs of conventional ap-
pearance can be made, intensity profiles
examined, star positions located, etc. The
playback system has proved especially
valuable in early stages of examination
and editing of the data and has increased
the efficiency of using the large computer.
200-Inch Cassegrain Spectrograph
A new spectrograph has been designed
and built for use at the Cassegrain focus
of the 200-inch telescope. It is designed
primarily for use with the television-type
area sensors and will be capable of ob-
taining slit-type spectra or absolute
spectrophotometric data. The spectro-
graph has both sky and object apertures
and uses a cross-dispersing quartz prism
to separate grating orders. The spectro-
graph is being tested and is expected to
be ready for use on the telescope in the
fall of 1973.
Prime-Focus Image Tube
A 90-mm diameter single-stage mag-
netically focused image tube has been
put into operation at the prime focus of
the 200-inch telescope with either the
Ross or new Wynne corrector and at the
Cassegrain focus of the new Palomar 60-
inch telescope. The cathode of the tube
is an extended S-20 type and the output
is via a fiberoptic plate. Limiting expo-
sures with 500 to 1000 A bandpasses on
baked Illa-J plates are obtained in 5
minutes or less with the 200-inch and in
40 minutes or less with the 60-inch. The
image quality does not depend on the
image tube except for the faint honey-
comb pattern produced by the fiberoptic
output plate.
Coude Image Tube
A new image-tube system has been put
into operation by Zappala at the 32-inch
camera of the Mount Wilson 100-inch
coude spectrograph. This system uses a
40-mm diameter single-stage electrostat-
ically focused image intensifier with
fiberoptic input and output faceplates
manufactured by Varo Electron Devices.
The faceplates have been found to be of
high quality and are relatively free of
the shear and dislocations commonly
found in such tubes. An excellent ex-
tended-red-response S-20 photocathode
allows spectra to be obtained in the near-
infrared to beyond A8800. During opera-
tion the cathode is cooled to approxi-
mately 0°C to allow exposures longer
than 3 hours without objectionable back-
ground fog.
Preliminary observational tests with
unbaked Ila-D emulsion pressed into con-
tact with the output screen show that
speed gains of about 90 can be expected
at A8400 compared to ammoniated IN
emulsion, while gains at Ha are about
10 and 20 compared to 098 and 103a-F
emulsions, respectively. Although the
tube can be used in the AA4000-9000 re-
gion, little real information gain can be
138
CARNEGIE INSTITUTION
expected in the blue because of degrada-
tion in resolution by the intensifier.
Data System for 150-Foot Tower
The digital data system for the 150-
foot solar tower magnetograph was
nearly completed during the report year.
This system and the Raytheon 704 com-
puter will be moved to the 150-foot tower
telescope sometime during the late sum-
mer or autumn of 1973. The computer
will operate the telescope and magneto-
graph and do the data reduction.
Vacuum Telescope
Through a grant to Zirin from the Na-
tional Aeronautics and Space Adminis-
tration for support of the Skylab pro-
gram, construction was begun and largely
completed on a new 26-inch vacuum
telescope for the Big Bear Solar Observa-
tory, to be placed in the existing fork
mount in conjunction with two smaller
vacuum telescopes. Final assembly is
scheduled for the summer of 1973, with
the principal remaining problem the com-
pletion of the vacuum window. It is
hoped that the new instrument will make
possible full exploitation of the seeing at
Big Bear.
The new telescope will utilize two new
birefringent filters recently purchased:
one 0.3 A wide for the A10830 helium line,
and a Universal Birefringent Filter capa-
ble of use at any wavelength between
A4000 and A7000, with bandpass ranging
from 0.1 to 0.25 A in that range. The
latter is being fabricated by Zeiss and
will be delivered shortly.
GUEST INVESTIGATORS
During the report year, the Hale Ob-
servatories accommodated 56 guest in-
vestigators representing 34 different in-
stitutions. Fourteen guest investigators
were from countries other than the
United States.
Dr. Saul J. Adelman of the NASA
Goddard Space Flight Center has con-
tinued his studies of sharp-lined magnetic
Ap stars and of suitable main-sequence
comparison standards. His current pri-
mary interests are silicon and B5-B7
stars. In March 1972, he obtained for
several of these stars 4.3 A mm-1 IIa-0
spectrograms with 100-inch coude and
energy distributions in the region AA3200-
7400 with the 60-inch Cassegrain scanner.
Dr. J. R. P. Angel of the University of
Texas and Dr. J. D. Landstreet of the
University of Western Ontario used the
multichannel spectrophotometer on the
200-inch telescope together with a Pock-
els cell modulator as a narrow-band
multichannel polarimeter during two
nights in August 1972. Observations of
the circular and linear polarization spec-
trum of the magnetic white dwarf Grw
+70° 8247 were obtained. The new cir-
cular polarization spectrum, obtained
with 80 A resolution in the blue, is simi-
lar to the one reported by Angel, Land-
street, and Oke (Astrophys. J. [Lett.],
171, Lll, 1972) but is more accurate.
The linear polarization spectrum, ob-
tained with 160 A resolution in the blue,
has never been studied at this resolution
before. In the blue some structure is ap-
parent, with a polarization maximum of
about 4% at 4000 A. The polarization
falls to near zero at 5000 A, rotates
smoothly through 90°, and rises again to
a value of 3-4% near 1/x. This material
has not yet been published. Preliminary
measurements were obtained with 20 A
resolution of the wavelength dependence
of linear polarization of the Be star y Cas
and of <f> Cas, which has strong inter-
stellar polarization. Further observa-
tions of this sort are planned in future
observing runs.
Dr. K.-H. Bohm and Dr. R. Schwartz
of the University of Washington, using
the multichannel spectrometer of the
200-inch telescope, made observations of
the condensations G and H (designations
according to G. H. Herbig in Nonperiodic
HALE OBSERVATORIES
139
Phenomena in Variable Stars, Reidel,
1969, p. 75) in Herbig-Haro object No. 2,
as well as observations of the brightest
part of Herbig-Haro object No. 1 and of
the small nebula near T-Tauri. A band-
width of 20 A was used in the whole spec-
tral range from —3200 to 11,000 A. The
line intensities are being used to deter-
mine the electron density, electron tem-
perature, and the degree of ionization in
individual condensations in order to find
out whether the physical parameters in
the condensations depend on their pre-
vious brightness changes. The ratio of
the intensities of the infrared [S II] lines
AAl0317/10336tothe [S II] AA4068/4076
lines will be used to determine the red-
dening of the spectra of the individual
condensations. The energy distribution
of the continuum (or quasi-continuum)
and the B aimer decrement in individual
condensations are also being studied.
Dr. Clark G. Christensen of Brigham
Young University, making use of the
Cassegrain scanner at Mount Wilson, has
obtained energy distributions in the
wavelength range AA3448-8050 for the
nuclei of globular clusters NGC 6229,
NGC 6934, and M13, and of the Irr I
galaxy NGC 4449. He also obtained
energy distributions of the metal-defi-
cient stars HD 122563, HD 151937, HD
170153, M92 111-13, and M92 VII-18.
These data will be used in globular clus-
ter and galaxy population syntheses.
In July 1972, Dr. S. V. M. Clube of
the Royal Greenwich Observatory ob-
tained direct plates, with the Mount Wil-
son 60-inch telescope, of twelve Kapteyn
selected areas. These are matched by
the old photometric plates taken in 1911-
1912 and will be used to determine proper
motions of faint red stars.
Observations for studying the be-
havior of optical interstellar lines in dark
clouds have been obtained by Dr. Judith
G. Cohen of the University of California
at Berkeley. The equivalent widths and
radial velocities of the interstellar ab-
sorption lines of Ca II, Na I, CH, and
CH+ were measured in 30 stars. The
stars were selected to be within or be-
hind dense clouds with up to 3 magni-
tudes of absorption. The interstellar lines
in the cloud stars are compared with those
in a group of highly reddened supergiants
whose reddening arises from a large
distance, i.e., a long path length through
relatively thin material. The atomic in-
terstellar lines of the cloud stars are
weak compared to those of supergiants
with the same color excess; the molecular
lines are of comparable strength in the
two groups, except that the ratio
ncu/^cn is larger in the cloud stars. The
ionization equilibrium in the two cases
can be approximated, and it is concluded
that the deficiency of atomic Ca and Na
relative to hydrogen in the clouds is
about a factor of 100 larger than in the
supergiants. A new series of observations
is planned to study interstellar lines in
stars at high galactic latitudes.
Dr. Peter S. Conti of the Joint Insti-
tute for Laboratory Astrophysics of the
University of Colorado has employed the
coude spectrographs of both the 100-inch
and the 200-inch telescopes to study the
spectra of O and Of stars. Spectra ob-
tained of HDE 226868, the supergiant
O-type star in the binary system which
contains the x-ray source Cyg X-l, have
been discussed in association with Mar-
gon and Smith of Berkeley (Astrophys.
J. [Lett.], 179, L125, 1973). They find
the emission at A4686 of He II to be vari-
able in intensity, and they suggest it is
coming from material flowing in the sys-
tem rather than from the primary star or
secondary object. They also observe
slight changes in spectral class in the
supergiant, probably due to distortion
effects in this close binary system. Conti
obtained spectra of two anomalous O
plus Of binary systems. In both cases,
emission lines had previously been de-
tected in one component, and absorption
lines in the other; the mass ratios indi-
cated that the Of star was the less mas-
sive. For BD+40° 4220 the spectra ex-
hibit absorption lines in both components,
indicating that the previously derived
140
CARNEGIE INSTITUTION
mass ratio was incorrectly estimated.
The lines are very broad and shallow and
it has not proved possible to measure
them accurately on a Grant machine.
However, attempts are being made to
obtain velocities, using microphotometer
tracings that have sufficiently accurate x
recording. For the other system, HDE
228766, no certain absorption lines are
detected in the Of component, but there
are certain variations in line profiles in
the 0 star's absorption lines. Studies of
this system are continuing with tracings.
One excellent spectrogram of the 06f
star, HD 153919, associated in a binary
system with the x-ray source 2U 1700-37,
has been obtained. Its spectral appear-
ance suggests a high luminosity and a
large mass, but this cannot be proven
with certainty. Emission lines of Si IV,
N III, C III, and A4686 He II are seen,
although they are not unique to this Of
star. A few spectra in the blue of some
early Of stars in the h and x association
have also been secured. These will be
used to determine accurate line profiles
to compare with theoretical models,
which will include the physical effects of
extended atmospheres. The reduction
procedure from the microphotometer in-
volves minimizing the noise due to the
grain fluctuations by inverse Fourier
transform filtering. The profiles are
thereby smoothed without bias. The
computer programs for this method are
presently being developed.
Conti also obtained two very good
spectrograms of the supernova in NGC
5253, which was detected just before his
first coude run at Mount Wilson com-
menced. Both spectra show the inter-
stellar H and K lines in our own Galaxy,
and in NGC 5253 itself. This has been
discussed by Conti, Greenstein, and Wal-
lerstein in Astrophysical Letters, 12, 101,
1972.
In February 1973 Dr. Kyle Cudworth
of the University of California at Santa
Cruz used the 48-inch schmidt at Palo-
mar to obtain second-epoch plates to be
compared with the Sky Survey to deter-
mine proper motions in the Praesepe
cluster and the galactic anticenter. This
study of Praesepe is in collaboration
with Dr. Burton Jones of Greenwich.
Dr. D. G. Currie, Mr. S. L. Knapp, and
Mr. K. M. Liewer of the University of
Maryland have used the amplitude inter-
ferometer on the 60-inch and the 100-
inch telescopes on Mount Wilson and the
200-inch telescope on Palomar Mountain
to conduct observations of the apparent
angular size and the structure of late-
spectral-type giant stars. These observa-
tions were conducted using a modern
version of the Michelson stellar inter-
ferometer. The measurements are made
in a sufficiently short interval of space,
time, and wavelength to guarantee that
any atmospheric degradation of the
measurements is less than some predeter-
mined value. Initial tests were conducted
on the 100-inch at Mount Wilson in June
1972, followed by a series of measure-
ments in June and July. During Decem-
ber 1972, a second series of measurements
was conducted at Mount Wilson and
Palomar Mountain. The diameters of
four stellar disks have been measured:
a Ori (Betelgeuse) , a Tau (Aldebaran) ,
a Boo (Arcturus), and ft Peg (Scheat).
The measurements, their accuracy, and
the wavelength regions studied have been
reported elsewhere. The better sets of
measurements have an internal precision
of 0.002 arc seconds (standard deviation
of the mean). The wavelength depend-
ence of the effective diameter (which de-
pends upon both the geometric diameter
of the stellar disk and the degree of limb
darkening at that wavelength) was also
measured. Finally, preliminary measure-
ments were made on the apparent oblate-
ness of a Ori and on the orbital param-
eters of a Aur (Capella). Further obser-
vations on a Boo, a Her and a Sco were
conducted in June 1973. A continuing
program of observations and instrument
improvement is planned. In particular,
a new instrument, the multi-aperture
HALE OBSERVATORIES
141
amplitude interferometer, is planned
which will increase the sensitivity by
about ten magnitudes.
Dr. S. J. Czyzak of Ohio State Uni-
versity and Dr. L. H. Aller of the Uni-
versity of California at Los Angeles have
continued their spectrophotometry of
planetary nebulae. Observations were
secured of 29 nebulae in the range
AA4400-7800. They also observed thfe
diffuse nebula NGC 604 in M33. Reduc-
tions have been completed for 12 of these
objects. For most planetary nebulae,
photoelectric measurements give the
monochromatic fluxes integrated over the
object. Because of the necessity of using
wide apertures, the spectral resolution is
often low. In some instances, it has been
possible to secure higher resolution data.
These higher resolution data invariably
refer to regions of small angular size in
the nebula and can be compared with
photoelectric measurements only if the
relative spatial energy distribution in the
monochromatic images is known. A spe-
cial effort has been directed to NGC 7027.
Here the photoelectric scanner measure-
ments must serve as fundamental cali-
bration for other, higher spectral resolu-
tion measurements. Reductions covering
the spectral range from A3100 to A8500 A
are now completed. The spectrograms of
NGC 6543 secured at Palomar in 1971
have been reduced and utilized to inves-
tigate the physical conditions (electron
density, temperature, and level of excita-
tion ) at various points in the filamentary
pattern of this object. Results for NGC
40 {Astrophijs. J.] 172, 361, 1972), NGC
6778 (Astrophijs. J., 181, 817, 1973) have
been published; those for NGC 6445 are
in press. The discussion of NGC 6803
has been written and is ready to be sub-
mitted.
Dr. R. J. Dickens of the Royal Green-
wich Observatory used the Palomar 60-
inch telescope to obtain spectra of gal-
axies in the clusters Abell 1367 and
I311 53m _j_ 390 In conjunction with
spectra taken on other telescopes, the
measured redshifts will be analyzed to
yield velocity dispersions and estimates
of the virial masses of the clusters.
Mr. Dainis Dravins of the Lund Ob-
servatory carried out solar observations
simultaneously with the magnetograph
at the Mount Wilson 150-foot tower and
with a photographic telescope at Big
Bear. These studies aim at determining
the magnetic features of network ele-
ments in quiet regions on the sun. At Big
Bear an attempt was made to detect
possible transient electric fields in the
energy buildup phase preceding a flare
by measuring linear polarization in the
wings of a line sensitive to the Stark
effect. However, no such fields were seen.
Dr. Yoshio Fujita of the University of
Tokyo is continuing his study of carbon
stars. In late carbon stars the shorter
wavelength region below A4380 is masked
by strong absorption considered to be
due to C3. However, carbon stars of
comparatively high temperature do not
show such a remarkable feature and well-
exposed spectrograms are extended to the
violet region far beyond the H and K
lines. The purpose of this investigation
is to identify the spectral feature short-
ward of A4380 in such a star as HD
182040 (RO, CL>). The comparative
study in connection with the problem
described above was carried out using
spectrograms of HD 182040 (RO, CL>),
HD 156074 (RI, CD) and RU Cam (RO,
COi) obtained with the 100-inch and the
200-inch telescopes. The coude spectro-
graph of the 100-inch telescope has been
used to obtain spectrograms of carbon
stars with a dispersion of 13.6 A mm"1 in
the spectral region from A6400 to A8900.
One purpose is to estimate the intensity
of the forbidden [CI] A8727 line in some
carbon stars. A preliminary investiga-
tion (Y. Fujita, Proc. Jap. Acad., Ifi,
295, 1970) revealed the presence of this
line in some late carbon stars. It is de-
sirable to extend this investigation to
some early carbon stars and to establish
the intensity variation of atomic carbon
along the spectral sequence; this should
give some knowledge on the dissociative
142
CARNEGIE INSTITUTION
equilibrium of carbon species in stellar
atmosphere. The second purpose is to
estimate the abundance ratio 12C/1:5C by
comparing the intensity of 12CN satellite
lines with some isolated rotational lines
of 1HCN in the spectral region from A7750
to A8060. This was discussed in a pre-
liminary paper (Y. Fujita, T. Tsuji, and
H. Maehara, Proc. Jap. Acad., 4-5. 484,
1969). Following this method, the abun-
dance ratio of 12C to 13C will be deter-
mined with regard to late carbon stars as
well as early carbon stars. The value of
12C/13C in the spectral sequence of car-
bon stars may give information on the
chemical elements' synthesis in these
stars.
Dr. Tom Gehrels of the University of
Arizona used the Palomar 48-inch
schmidt for a survey of asteroid and
comet populations; the limiting magni-
tude was near 21.0. New Mars-orbit
crossing asteroids were discovered; the
survey is continuing for statistical de-
termination of the magnitude-frequency
relation. A high density was found for
the Hungaria asteroids, which have near-
circular orbits immediately beyond that
of Mars. Trojan asteroids in both La-
grangian points of Saturn were searched
for, but none was found; the same pre-
liminary conclusion was reached for
Neptune, but that search is continuing.
A preliminary result also has been
reached regarding the occurrence of dis-
tant comets (perihelion distance q > 3
AU). These comets are of particular in-
terest because of their relative freedom
from nongravitational effects and be-
cause they may be original condensations
on their first passage into the solar sys-
tem. Three such comets were discovered,
one of which was subsequently identified
by Dr. B. G. Marsden at the Smithsonian
Astrophysical Observatory as Comet
Swift 1889 IV.
Dr. Alexander F. H. Goetz of the Jet
Propulsion Laboratory of the California
Institute of Technology is using a silicon
vidicon photometer and the Leighton
Image Motion Compensator to produce
high quality, high dynamic range pic-
tures of Saturn. These digital pictures
are further analyzed and enhanced by
the JPL Image Processing Laboratory.
During this spring's run he was plagued
by TV interference and poor seeing con-
ditions. The TV problems have been cor-
rected and he is looking forward to ob-
taining suitable data in the fall.
Dr. R. F. Griffin of the Cambridge
Observatories and Gunn used the radial
velocity spectrometer at the 200-inch
coude focus on three nights during the
report year. Observations were made of
stars in the regions of the Hyades, M67,
and M3. A photographic spectrogram of
BD+12°1919, newly demonstrated by
radial velocity to be the brightest mem-
ber of M67, was obtained with the 36-
inch camera of the coude spectrograph.
Dr. F. D. A. Hart wick of the Univer-
stiy of Victoria and Dr. Robert D. Mc-
Clure of Yale University Observatory
have done intermediate-band photoelec-
tric photometry of giant stars in several
outlying old stellar systems, using the
200-inch Hale telescope. The ultimate
goal is to intercompare the surface gravi-
ties, effective temperatures, and metal
abundances of the giant stars in these
systems and some of the bright halo
globular clusters. Giants in the remote
globular cluster NGC 7006 were found to
have a cyanogen excess compared to
giants in other clusters with the same
ultraviolet excess. Previously, Sandage
and Wildey found this cluster to be
anomalous in having too red a horizontal
branch for its metal abundance. These
results suggest that the CNO abundances
may be the second parameter affecting
the color-magnitude diagrams of globular
clusters. The dwarf spheroidal galaxy in
Draco, which has a similarly anomalous
horizontal branch, was also observed.
The giants in this system exhibit an ultra-
violet excess comparable to that of the
most metal-poor globular clusters in the
Galaxy. The observations also suggest
that the surface gravities of giants in
this system may be significantly different
HALE OBSERVATORIES
143
from globular cluster giants in the Gal-
axy. Further observations, using the
multichannel spectrometer, of giants in
the Ursa Minor dwarf galaxy are in
progress.
During the year Dr. T. V. Johnson and
Dr. D. L. Matson of the Jet Propulsion
Laboratory of the California Institute of
Technology used the Mount Wilson 60-
inch telescope to carry out spectropho-
tometry of asteroids in the visual and
near infrared (0.3-2.5 /xin) wavelengths.
During July 1972 the earth-crossing
asteroid (1685) Toro came unusually
close to the earth and was bright enough
for spectrophotometric study in the 0.4
to 0.8 fxm wavelength range. This was
the first detailed spectral information ob-
tained for a member of this important
asteroid group which is believed to be
the immediate source of some meteorites.
In October 1972, observations of asteroid
(43) Ariadne were made in five colors
from 0.34 to 0.95 tim. It was discovered
that this asteroid has a much larger light-
curve amplitude (~0.7 mag) than had
been reported previously and that it also
has a significant solid state absorption
band near 0.95 /xm. The spectral reflec-
tance of this asteroid matches that of or-
dinary chondrite meteorites and that of
Toro, but is not similar to other asteroids
in the main belt. Thus, it may be part of
the sought-after source or reservoir of
chondritic material. Observations at 1.2,
1.6, and 2.2 /xm of four large asteroids
(Ceres, Pallas, Juno, and Vesta) were
made on several occasions in cooperation
with G. Veeder and S. Loer of the Cal-
tech Infrared Group. This is the first
time observations of asteroids have been
made at these wavelengths, and the re-
sults will extend the wavelength cover-
age of their spectral reflectivity by more
than a factor of 2. Preliminary results
have indicated large differences between
these four asteroids and have suggested
that the surface composition of Ceres,
and perhaps that of Pallas, resembles
Type-I carbonaceous chondrites. If so,
then the surfaces of these two asteroids
may be very old and unmodified by endo-
genic forces sinCe the time of accretion.
In one run with the 200-inch coude
spectrograph at Palomar, Dr. Philip C.
Keenan of the Perkins Observatory ob-
served several very luminous red super-
giants with 098-01 plates. The purpose
was to test the effectiveness of the widths
of metallic absorption lines as measures
of luminosity for G, K, and M stars of
luminosity classes lab and brighter. At
the moderate scale of 13.5 A/mm it was
found possible to confirm the high lumi-
nosity of four supergiants by comparison
with BS 8752 (G5 0) and RW Cep (KO
0-Ia) . Most interesting was the very red
star, no. 1 in the cluster Tr 27, which
appears to be CoD— 33°12241. In the
study of this cluster by The and Stokes
they had measured the color of no. 1 as
B — V = 2.83, and since its position co-
incides with a very heavily absorbed re-
gion at one edge of the cluster, they
thought no. 1 might be an enormously
reddened OB star. Albers, however, ob-
tained an IR objective-prism spectro-
gram on which he could see weak CN
bands, showing the star to be of late
type. In spite of the southern declina-
tion, several degrees below the galactic
center, the 098-01 spectrogram was well
exposed and shows greatly widened lines
and indicates a type of about K5 and a
luminosity class near la. Although no. 1
is less than 4' distant from the four
brightest blue stars in the cluster, its
membership is not certain and it may lie
behind the cluster.
Mr. Richard G. Kron and Air. David
C. Koo of the University of California at
Berkeley have used the 48-inch schmidt
to search for highly obscured stars in the
Taurus clouds. Because of poor weather,
the only plate obtained of the region
reached only m = 16 at A(,ff = 8300 A and
was badly affected by differential refrac-
tion. A plate of the dark clouds in Ophiu-
chus was also obtained, but the exposure
was limited by dawn. A very red star was
found at 16i;22»\3, — 24° *22'.5 (1950).
144
CARNEGIE INSTITUTION
for which Luis Carrasco has observed
about 15 magnitudes of extinction in the
visible. At the request of Francois
Schweizer, Kron and Koo obtained a 1-N
plate of the interacting galaxies NGC
4038-39. At this wavelength the nuclei
stand out above the emission regions, and
the identification agrees with that de-
duced spectroscopically by Rubin et al.
(Astrophys. J. 160, 801, 1970).
Dr. A. Labeyrie of the Paris Observa-
tory has developed the speckle interfer-
ometry program at the 200-inch tele-
scope, which has a unique resolution
advantage in addition to the luminosity
advantage. For efficient use of the ob-
serving time, a television system was
specially built and used at the prime
focus in June, October, and April. Ap-
proximately two million images were
recorded. This has created a processing
problem requiring some technological de-
velopment, which is being carried out at
Meudon. However, preliminary process-
ing work has evidenced color-dependent
limb-darkening features in o Ceti, a
Orionis (Bonneau and Labeyrie, Astro-
phys. J. [Lett.], 181, LI, 1973) and a
Scorpii (to be published) , as well as pre-
viously unknown companions in 8 Scorpii,
o- Hercules, i Serpentis, BSC 6927, (3
Corona Borealis and R Lyrae. The limb-
darkening observations suggest the pres-
ence of a blue-scattering envelope around
cool stars. It was recently found that
speckle interferometry observations can
in principle be extended to objects much
fainter than previously assumed. Cur-
rent plans are to add an image-intensifier
stage to the present recording system;
this should permit a limiting magnitude
of the order of mv = 22 to be reached.
Mr. Howard H. Lanning of California
State University at San Diego has been
studying the photometric variability of
the short-period binary Humason-
Zwicky 22 (VX Canum Venaticorum).
Differential UBVr and no-filter observa-
tions of the 13th magnitude object have
been obtained using an S-20 photocath-
ode on the Mount Wilson 60-inch tele-
scope. Poor weather conditions severely
limited the total number of acceptable
nights. The light curve was, however,
successfully obtained and will be ana-
lyzed in detail shortly. Lanning expects
to continue the survey of the two-color
48-inch schmidt plates of the galactic
plane begun recently at Hale Observa-
tories as part of the program on optical
identification of x-ray sources. Arrange-
ments have been made for loan of the
plates, and equipment is presently being
constructed at San Diego to aid in their
examination.
Dr. Craig D. Mackay and Mr. R. J. E.
Kraft of the Institute of Astronomy,
Cambridge, England, have used the 48-
inch schmidt telescope to observe fields
centered on the clusters of galaxies Abell
1367, Abell 2151, and Abell 2199 in six
colors. The plates have been calibrated
using Gunn's wedge sensitometer, and
the reduction will be done with the GAL-
AXY mk. 2 measuring machine at Edin-
burgh. These data will allow a detailed
study of the color structure and morphol-
ogy of the cluster over a range of nearly
7 mag.
Dr. T. B. McCord of the Massachusetts
Institute of Technology employed the
60-inch and 100-inch telescopes with the
MIT silicon vidicon to record images of
Jupiter, Saturn, and Uranus through
narrowband interference filters. Images
were obtained in methane and in some
cases ammonia-absorption bands and at
nearby continuum wavelengths. Maps of
methane absorption were obtained. The
same system was used to obtain visible
and 1.0 /iiii images of several galaxies.
Despite bad weather, a series of high-
resolution images of geologically inter-
esting regions of the moon was obtained
using interference filters. These images
are extremely important, for they show
the boundaries of compositional units
and are in fact geologic maps of a sort.
The work is described in a publication in
preparation. The spectrometer was first
used with the vidicon during the run. A
series of calibration and sensitivity tests
HALE OBSERVATORIES
145
were run. Spectra 0.3-1.2 jum were ob-
tained for a selection of standard stars
and for Saturn and several areas of the
moon. The devices worked quite well.
Dr. Walter E. Mitchell, Jr., of the
Perkins Observatory utilized the com-
puting facilities at Caltech and to a
lesser extent the Raytheon 703 computer
at Santa Barbara Street to reduce data
obtained at the Snow Telescope during
1971. These data were obtained for two
programs: (1) the construction of a
spectrum map from 3000 A to the atmos-
pheric limit, and (2) the limb-darkening
measurements for the determination of
the run of solar opacity in the wave-
length range 4000 A to 3000 A. In the
first program, the digitally recorded
spectra were "spliced" and combined into
a single high-resolution intensity map.
In the second, the numerical corrections
for the instrumental profile were worked
out and the limb-darkening coefficients
for the gray case were calculated for a
first approximation.
During the past year, two types of
television sensors were used at Palomar
by Dr. D. C. Morton of Princeton Uni-
versity Observatory. In October 1972 the
SEC vidicon integrating system was
mounted on the coude spectrograph to
obtain further high-resolution spectra of
quasars and galaxies. The objects ob-
served included PHL 957 for a detailed
analysis of the absorption lines, and the
central regions of M31 and M32 for fur-
ther studies of velocity dispersions and
rotation curves. An earlier TV spectrum
of M31 (Astrophys. J., 180, 705, 1973)
showed that the line-of-sight dispersion
in the nucleus was only 120 ± 30 km s"1.
In March and May 1973, in collabora-
tion with Oke, a new photon-counting
television system was tested for direct
imagery at the Cassegrain focus of both
the 60-inch and 200-inch telescopes.
Amplifications of 2 X 105 are possible
from a two-stage tube that has a silicon
target, so that individual photoelectron
events can be detected and counted by a
fast digital memory with a dynamic
range of 216 for each pixel in a 256 X 256
format. The system was operated suc-
cessfully for several nights, but further
work is needed to calibrate spatial irreg-
ularities in the tube efficiency and reduce
significant losses of counts in the circuits.
Dr. Thomas W. Noon an of the State
University College at Brockport, New
York, continued his study of counts of
galaxies in clusters of galaxies. He ob-
served two clusters (Abell 234 and Abell
732) and obtained data from existing
plates of five other clusters (1306+2716;
1448+2617; 0024+1654A; 0024+1654B;
and 1410+5224) . The radii of the seven
clusters, which range in redshift from
0.17 to 0.46, are being studied, along with
radii from published counts of other
clusters, for use in the angular-size
cosmological test.
Using the coude spectrograph of the
100-inch telescope, Dr. John Norris of
Yale University Observatory has ob-
tained high-dispersion spectra of the sub-
dwarf B-type star HD149382 and the
high-velocity B star HD125924. Five
analyses of these spectra are under way
to study the heavy element abundances
in the groups of objects to which the
stars belong.
Dr. Valdar Oinas of the University of
Nebraska at Lincoln employed the Casse-
grain scanner at the Mount Wilson 60-
inch telescope to obtain energy distribu-
tions for a selection of late-type stars.
The purpose of the scans was twofold:
(1) to provide temperatures for an anal-
ysis of a discrepancy between ion and
neutral line strengths that appeared dur-
ing his thesis work at Caltech; (2) to
delineate more accurately a suspected
wide absorption feature near 8800 A.
Two of the three nights assigned were of
photometric quality and good results
were obtained.
Dr. B. E. J. Pagel of the Royal Green-
wich Observatory used the coude of the
100-inch telescope during April and May
1973 to secure several 2.9 A mm-1 spectra
in the yellow and red of the "super-
metal-rich" candidates fx Leonis and /?
146
CARNEGIE INSTITUTION
Oph and standard stars e Vir and a Ser.
Dr. Sidney B. Parsons of the Warner
and Swasey Observatory observed the
classical Cepheid rj Aquilae and several
nonpulsating supergiants of similar tem-
perature at the coude focus of the 100-
inch telescope. Five out of the seven
scheduled nights were clear and gave a
good sampling of phases for the 7.2-day
variable. Observations were made with
the 32-inch and 73-inch cameras on am-
moniated I-N, Ila-D, and baked IIa-0
emulsions at reciprocal dispersions of 20,
4.5, and 3.0 A mm"1, respectively. The
spectra are intended primarily for work
on the effective temperatures and surface
gravities of F- and G-type supergiants
and Cepheids, although radial velocities
will also be measured. In the near-
infrared, profiles of Ca II and Paschen
lines are sensitive primarily to tempera-
ture and gravity and are not blended
badly into other lines. Lines of N I are
also present and will be a test of the in-
creased abundance expected after mixing
in the red-giant phase of evolution. In
the near-ultraviolet, blending of numer-
ous absorption lines seriously hampers
the study of the Balmer jump, which is
quite sensitive to temperature and grav-
ity for these stars. The new observations
down to about 3300 A will give a much
better calibration of synthetic spectra
being computed in collaboration with Dr.
R. A. Bell (at the University of Mary-
land), and the computations will allow
good model-atmosphere analysis of pho-
tometric measurements of the Balmer
jump. H/3 profiles for a few stars were
photographed in order to provide a check
on the temperatures; H/3 is not sensitive
to surface gravity in cool stars, is not
blended too badly, and is not as sensitive
as Ha to the presence of a chromosphere.
Dr. S. Eric Persson and Dr. Jay A.
Frogel of Harvard College Observatory
began a spectrophotometric study with
the Mount Wilson 100-inch and Palomar
60-inch telescopes of northern H II re-
gions that are known to be infrared
sources (Sh 138, 156) . The planetary
nebula BD-|-30o3639 was shown to have
an optical continuum at H/3 slightly in
excess of that expected, although the re-
sult is quite sensitive to the electron
temperature assumed. The explanation
is that the particles believed responsible
for the known large infrared excess also
scatter a small fraction of the light from
the central star. They used the 200-inch
multichannel spectrometer in September
1972 to measure emission lines in the
planetary nebulae NGC 7027 and NGC
7662 at 2 and 3.5 arc sec resolution. A
complete study of the level of ionization,
temperature, reddening, and He I-line
ratio variations from point-to-point is
under way. One interesting result for
NGC 7027 is that the interstellar extinc-
tion at H/3 varies by a factor of 3 across
the face of the nebula, consistent with the
very different appearance of the object
at optical and radio frequencies. Some
time was spent measuring the energy dis-
tributions of two H II regions, K3-50 and
Sh 228, which are infrared sources. Meas-
urements of several suspected young
planetary nebulae having large 10 ^ ex-
cess emission (IC 4997, MHa 328-116,
VV 8, HBV 475) were obtained. The
data show evidence for high densities,
^lO6 to 107 cm-3, and will be compared
with previous spectrographic data.
During the summer of 1972, Dr. Daniel
M. Popper of the University of Cali-
fornia at Los Angeles carried out UBV
photometry of eclipsing binaries with the
20-inch reflector on Palomar Mountain.
The purpose of the observing program
was to obtain orbital inclinations, frac-
tional radii, and colors of systems for
which he had spectrographic orbits but
for which adequate photometry was not
available. These systems are double-
lined binaries, and complete spectro-
graphic and photometric analyses will
provide fundamental masses and radii of
some precision. The systems observed
most intensively were V805 Aql, TV Cet,
V478 Cyg, BS Dra, and BK Peg. The
observations have not yet been analyzed.
HALE OBSERVATORIES
147
Dr. Diane M. Pyper of the University
of Wisconsin at Parkside has made a
study of periodic magnetic stars, using
the scanner on the Mount Wilson 60-inch
telescope. The low-resolution spectrum
scans (20 A band pass) are accurate
enough to reproduce the variations pre-
viously published for the program stars
(this was expected, of course) . So far,
scans of t Cas, 56 Ari, 53 Cam, a CVn,
HD 153882, 73 Dra and HD 224801
(CGAnd) have been studied. The varia-
tions in M(5556) for these stars have
been compared with published V varia-
tions and in all cases the amplitudes of
variation and the scatter in the observed
points agree with the published data. In
some cases it may be possible to improve
upon the published periods. There is
evidence for variation of the Balmer
jump index [M(4255) — M(3636)] in
several stars; the Paschen continuum
index [M(4255) — M(6055)] also seems
to change in some stars. The amplitudes
of these variations are from about 0.08
mag to 0.03 mag from the present data.
Line blocking effects have not yet been
calculated.
Dr. Paul H. Richter of California State
University at Northridge made observa-
tions at the Cassegrain focus of the
Mount Wilson 60-inch telescope to study
the effects of atmospheric turbulence on
the optical contrast transfer function for
the atmosphere as a function of integra-
tion time and zenith angle, and to at-
tempt some short-exposure photography
of Jupiter and Ganymede through the
use of a magnetically focused, single-
stage image tube. This program is part
of an ongoing effort to utilize digital
image processing techniques to enhance
atmospherically degraded images of arbi-
trary, extended objects and so resolve
detail not visible in photographs ob-
tained by conventional means. The data
gathered during the above observing
schedule has aided in the understanding
of the effect of turbulence on astronom-
ical images and has permitted some very
substantial improvements to be made in
the resolution of selected images of
Jupiter.
Dr. W. L. Sanders of New Mexico
State University has a program to use the
100-inch Newtonian nebular spectro-
graph to obtain classification spectra of
faint Be stars (B = 12-14.5). The ob-
ject is to identify the la Supergiants
among those stars for use as spiral arm
tracers. Due to bad weather no results
have been obtained so far.
A search program for new Apollo-
Amor asteroids was initiated in January
1973 by Dr. Eugene M. Shoemaker and
Mrs. Eleanor Helin of the California In-
stitute of Technology, using the 18-inch
schmidt telescope at Palomar Mountain.
Although adverse weather has reduced
the anticipated sky coverage, 174 fields
had been obtained by early July. As
each field is photographed twice, the
number represents approximately 87 in-
dependent fields. An exceptional Apollo
asteroid has been discovered (Int. Astron.
Union Circular No. 2555) within the
statistically predicted number of fields
necessary for discovery of a new object.
Preliminary results of the new discovery,
1973 NA, are the following: it had the
greatest apparent motion at discovery of
any known Apollo asteroid (greater than
ll°/day); it has an inclination of ap-
proximately 70°, the highest of any
known asteroid; and, at discovery mag-
nitude of 10, it has an absolute magni-
tude of 14, which represents one of the
brightest Apollos. It is anticipated that
this unique Apollo will provide new in-
formation for the analysis of the dynam-
ical evolution of these objects. A second
set of observations of 1973 EA (a new
Apollo asteroid) was obtained by Mrs.
Helin, which permitted still further ob-
servations to be secured in Europe and
led to the establishment of its orbit.
Helin 's observations and positions were
announced in the Int. Astron. Union Cir-
culars, Nos. 2507 and 2509. In addition,
several objects were detected but were
unconfirmed because of periods of in-
148
CARNEGIE INSTITUTION
clement weather and the fast motion of
the objects.
Dr. R. W. Shorthill of The Boeing Com-
pany and Dr. T. F. Greene of The Boeing
Company and the University of Washing-
ton have continued their study of the at-
mosphere of Jupiter and of the Galilean
satellites by observing an ingress and
egress of Callisto in late 1972. This event
enabled probing of the extreme south
polar region of Jupiter. No refractive tail
was observed, indicating a significant
aerosol content in the south polar Jovian
atmosphere at the 1018— 1019 cm-3 level.
This result stands in contrast to the pres-
ence of a refractive tail (reduced rate of
light intensity falloff) on the latter por-
tion of several other observed satellite
light curves. A complete discussion of
the nature of the aerosol haze at the
levels probed during the observed eclipses
is in preparation. The aerosol haze is
stratified and apparently not uniformly
distributed in longitude even on the size
scale of the region sampled by the beam
illuminating the satellite. The variation
of the integrated aerosol extinction cross
section with altitude has been determined
at 1.05 fi. At the time of this writing,
other shorter wavelength data are also
being interpreted. The wavelength de-
pendence of the aerosol extinction cross
section and a number of other properties
of the atmosphere are being sought in
this study.
Dr. Sidney van den Bergh of the David
Dunlap Observatory reports that the
dwarf spheroidal galaxy And III, which
appears to be a companion to the Androm-
eda nebula, has been resolved into stars
with the 200-inch Hale Telescope. Since
resolution in this galaxy occurs at the
same magnitude level as does the resolu-
tion of NGC 185, it follows that these
two objects are at approximately the
same distance. In collaboration with
Herbst and Pritchet, an area of 6.2
square degrees near M31 has been
searched for variable objects down to the
limit of the 48-inch schmidt telescope.
Thirteen faint variable objects were dis-
covered close to the plate limit. The ob-
jects detected in this survey might be
either (a) quasars of intermediate color,
(b) very distant galactic or possibly
intergalactic variable stars, or (c) in
some cases very distant galaxies that
increased their total luminosity because
of a supernova outburst. The ten-year
program on the study of slow variables
in galaxies of the Local Group and some
members of the M81 group and South
Pole Group was continued with the 48-
inch schmidt telescope on Palomar
Mountain.
An optical atlas of galactic supernova
remnants, based in part on plate mate-
rial obtained with the 200-inch and 48-
inch telescopes, has been produced by
van den Bergh, Marscher, and Terzian
and will appear as Astrophys. J. Supp.
No. 227. The atlas presents the best
available photographs of all known op-
tical supernova remnants. Comparison
of Baade's 1950 plate of Kepler's super-
nova with 200-inch plates taken twenty
years later shows that the fan-shaped
remnant of this object is expanding with
a proper motion of ^- O."03 per year.
Combining this result with Minkowski's
radial velocity observations and assum-
ing a distance of 10 kpc yields an expan-
sion velocity of ^1400 km s_1 for the
optical remnant of Kepler's supernova.
This value is very similar to the expan-
sion velocity of the Crab Nebula. Since
Kepler's supernova was almost certainly
a supernova of Type I, the low expansion
velocity of the Crab can no longer be
used as an argument against the hypoth-
esis that the Crab Nebula was produced
by a supernova of Type I. Additional
spectrograms and direct plates were ob-
tained of the optical remnant of Cassi-
opeia A. It it hoped that these ob-
servations will lead to an improved
understanding of the physical processes
taking place in this supernova remnant.
A study was done on the hard x-ray
and microwave components in the impul-
sive phase of flares by Dr. Joan Vorpahl
of Sacramento City College. In 36 ran-
HALE OBSERVATORIES
149
domly chosen events the value for the
slope in the differential electron power
spectrum, E~8 electrons cm-2 sec-1 keV"1,
was related to the 20—32 keV spike rise-
time as trise = 0.6 exp (0.878), i.e., as
the rise-time becomes shorter, the frac-
tion of higher electron (and photon) en-
ergy increases. A good correlation was
also found between the peak flux at 8800
MHz in an event and the number of
electrons above 100 keV: F(8800) aN1-7,
when thin target theory was used. Only
13 out of 173 impulsive x-ray events had
no microwave flux above 2800 MHz;
N(>100) for these 13 flares was less than
1033 electrons. Conversely, N(>100)
was greater than 1033 electrons for 18
random events with flux at 8800 MHz or
above. We conclude from this that 1033
determines the lower threshold for detec-
tion of impulsive microwave emission
(assuming thin target models hold).
Vorpahl and Zirin examined two multi-
ple impulsive events in May 1969 and
compared their impulsive H-alpha, hard
x-ray and microwave components with
the observed Type III emission. No good
correlation was observed between meter
intensity and the degree of simultaneous
Ha outflow, although almost every Type
III burst had some associated Ha activ-
ity. Correspondingly, an inverse rela-
tionship existed between x-ray and Type
III emission, i.e., a strong meter burst
accompanied weak x radiation. This,
along with the fact that sufficient ener-
getic electrons were normally present to
produce the x rays, implies that the tra-
jectory, rather than electron spectrum or
supply, determines whether Type III
noise is observed during an impulsive
event.
Dr. George Wallerstein of the Univer-
sity of Washington completed a program
of searching for interstellar Ti II, using
the resonance line at A3383 A. From
spectrograms of 15 stars combined with
data for 8 stars reported by others, Wal-
lerstein and D. Goldsmith have deter-
mined column densities for Ti II for 11
stars and upper limits for 12 others. Ob-
servations and theory show that little or
no Ti I or Ti III is expected in inter-
stellar gas. The derived Ti/H ratios
show a deficiency of from 10 to 200 with
respect to the sun and solar system. By
investigating the correlations among col-
umn densities of Ca II, Ti II, H I and
color excess, they can show that the
missing titanium (as well as calcium) is
tied up in the grains.
An abundance analysis by Wallerstein,
Snow, Roberts, and Baird of two low-
velocity bright giants with very strong
CH, HR 7606 and HR 8626, yields a
small metal deficiency and a deficiency
of oxygen but normal carbon and nitro-
gen. HR 8626 has a substantial lithium
line while HR 7606 does not. These
abundances are difficult to understand in
terms of stellar evolution and nucleo-
synthesis. Wallerstein's program of ob-
servation of OH- and H^O-emitting stars
is near completion. Spectra near mini-
mum of R Cas, R Aql and RT Cyg were
obtained in October 1972. Extensive
radial velocity measurements of most of
the spectra have been completed and the
data are being analyzed in terms of mass
loss from these outermost layers. Ob-
servations of stars behind the Monoceros
Supernova remnant have been started.
One star shows displaced Ca II com-
ponents at ±70 km sec-1, thus yielding
an expansion velocity at one point. Ad-
ditional stars will be observed in the
autumn of 1973.
Las Campanas
Mr. Gonzalo Alcaino of Santiago,
Chile, has obtained six blue and six yel-
low plates of the globular clusters NGC
362, 1904, 3201, 5927, and 6809. Expo-
sures range from 40 sec to 40 min. Photo-
electric sequences of 15 stars between
visual magnitudes 11 and 16 have been
set up for the globular clusters NGC 288,
1904, 3201, 5927, 6121, and 6809.
Mr. Eduardo Hardy of the University
of Chile, using the 40-inch telescope,
150
CARNEGIE INSTITUTION
started work on a photographic atlas of
the Large Magellanic Cloud, a joint proj-
ect with Sandage. He also began a study
of the "old" stellar content of the LMC.
A number of deep B and V plates were
obtained on the Bar and selected areas of
the Cloud. Photoelectric UBV integrated
photometry was performed on special
regions of the Bar.
Dr. Edward P. Ney of the University
of Minnesota and Dr. Roberta M. Hum-
phreys of the University of Arizona used
the 40-inch telescope to observe super-
giants identified by Roberta Humphreys.
Approximately 24 supergiants in the
Norma-Sagittarius region were measured
at wavelengths from 1.2 to 18 microns.
Four very unusual stars with large in-
frared excesses were identified. Measure-
ments of Caltech Anonymous Objects in
the — 10° to — 30° declination range
were also made. Approximately 250 un-
identified IRC objects were available
during nighttime hours. A total of 235
of these were observed in 92 clear hours'
observing. The goal was to identify the
true infrared stars. Ney and Humphreys
had previously studied 230 objects in the
northern hemisphere and identified two
classes of infrared stars, which they
called Taurids (like NML Tauri) and
Cygnids (like NML Cygni). About 10%
of the northern hemisphere objects were
infrared stars and 20 were Taurids, 8
Cygnids. In the sample of 235 southern
hemisphere stars, again 10% (25 objects)
were infrared stars, but the preliminary
analysis indicates that they are all
Taurids. With only two exceptions, the
infrared stars were not visible in the 40-
inch telescope and had visual magnitudes
>16 mag. They were found by scanning
a 4 square-minute area around the IRC
coordinates. The two visual objects had
mv = 13 and mv = 14, and Humphreys
plans to get spectra of these on the
Steward Observatory 90-inch telescope.
Dr. John Norris of Yale University
Observatory has used the image-tube
spectrograph of the 40-inch telescope to
obtain radial velocities of stars which fall
above the horizontal branch and to the
left of the giant branch in the HR dia-
grams of the globular clusters w Centauri,
M4, and M22. Seven new members have
been found in w Cen, and five have been
found in M22. These stars are being in-
vestigated to determine their atmospheric
parameters so that a comparison can be
made with theoretical predictions of
postasymptotic giant-branch evolution.
Dr. S. Eric Persson and Dr. Jay A.
Frogel of Harvard College Observatory
have continued their infrared studies of
H II regions in the wavelength range 1.2
to 20 /x. Thirteen southern thermal radio
sources having high-emission measure
were found to be strong 10 and 20 \x
emitters and were mapped at several
wavelengths. One of these has been de-
scribed in detail by Becklin et al. {Astro-
phys. J. [Lett.], 182, L137, 1973). The
data show that these sources have the
following properties: The energy spectra
from 1.2 to 20 /x are at least as steep as
Fv a v~3 ; the observed flux at 20 /* is
greater by a factor of 100 than that ex-
pected by extrapolating the radio obser-
vations, while the flux shortward of 2.2 ^
is less than that expected; observations
with intermediate band-width filters in the
10 jx region show a spectral feature in
several of the objects which has been
associated with silicate absorption; sev-
eral of the sources are barely resolved
with a 14. "5 beam at 10 and 20 /x, while
others (such as RCW 38 and RCW 57)
are highly resolved and show complex
structure. At least one of the objects
(RCW 57) contains an unresolved infra-
red source with a spectrum similar to
that of the source IRS 5 in W 3 and the
point source in the Orion nebula. An
interpretation which has been advanced
for these two sources is that they are
pre-main-sequence objects. The extended
infrared emission is believed to arise
from radiating dust particles that are
heated by the Lyman-a radiation field,
and perhaps also by direct starlight. In
HALE OBSERVATORIES
151
addition, Persson and Frogel found the frared study (made in collaboration with
peculiar optical nebula NGC 6302 to I. J. Danziger) show that NGC 6302 and
have excess emission at 10 and 20 p. The the peculiar planetary nebula NGC 7027
results of a combined optical and in- are similar in several respects.
ASTROELECTRONICS LABORATORY
Digital linage Recorder
This project has been carried out as a
joint effort with Princeton University
Observatory. The instrument was used
on both the 60-inch and 200-inch tele-
scopes at Palomar during the spring of
1973. The Princeton group was respon-
sible for the high-gain television camera
and its associated controls; AEL was
responsible for the digital memory, the
system controls, the computer interface,
and the operational software.
The memory is a 65,536-word, dynam-
ically shifted Metal Oxide Semiconduc-
tor array. Each word consists of 16
binary bits, or 65,536 counts. For its first
application, the system was used to re-
cord the data from a 256 X 256 pixel
(picture element) format at a pixel rate
of 4.8 MHz. The memory array can be
set to between 8 and 2,048 pixels per scan
line and between 8 and 2,048 scan lines
per picture. The memory size cannot be
less than 4,096 words or more than
65,536 words. The memory word rate
can be adjusted from 1 sec to 9,999 sec.
Other controls allow the observer com-
plete flexibility in his data collection
process.
The memory is scanned in synchro-
nism with the camera sweep circuits to
ensure that the memory words and image
pixels maintain a one-to-one correspond-
ence with each other. When a photoelec-
tron pulse is detected in any pixel, a cor-
responding memory word is incremented.
Although the memory was designed pri-
marily for use with photoelectron-
counting-type detector systems, it can
also be used with camera tubes that use
target integration and slow-scan readout.
The memory data read into the Hale
Observatories' computer system is for-
matted and transmitted to an industry-
compatible magnetic tape recorder. The
computer is also used for recording time,
telescope coordinates, and object identi-
fication data. Other telescope control
functions are also handled by the com-
puter.
The first observing runs clearly dem-
onstrated that this type of image recorder
is successful and that it will be an effec-
tive new observing tool. The initial data
tapes are now being analyzed to deter-
mine the overall quantum efficiency and
system characteristics that may require
future improvements.
Computer Systems
In addition to the computer systems
on the Palomar 60-inch and 200-inch
telescopes, the computer systems for the
Mount Wilson 60-inch telescope and the
150-foot solar tower were completed this
year. These latter two systems are under-
going software checkout.
After more than a year of experience
with the computer systems, it is interest-
ing and helpful to assess their effective-
ness. On the positive side, it is very clear
that the computer systems provide new
capability to the observers. This is de-
rived from more flexible control, more
efficient data collection, and more accu-
rate data displays. Most observers feel
that the basic computer concept is good
and that it offers further potential for
increases in effectiveness. On the nega-
tive side, the observers have raised two
serious objections to the present system.
These objections are directly related to a
failure to adhere strictly to two of the
most important philosophical guidelines
for astronomical instrument applications.
These guidelines were not followed be-
cause limited resources led to an attempt
to compromise these principles. It is now
152
CARNEGIE INSTITUTION
clear that the compromises have caused
serious problems and that these guide-
lines are of fundamental importance to
astronomical instrument design and ap-
plication.
The first principle is that any instru-
ment must be extremely reliable and free
from even minor defects. The number of
components that have failed to give per-
fect performance has been very small
and the actual observing time lost due to
these failures has also been small. But
these occasional failures have led to frus-
tration and irritation for the observers.
This small and infrequent failure rate
has also increased the problems of
troubleshooting and of making corrective
changes in the system design. The AEL
priorities have been changed to ensure
early elimination of these problems.
The second principle that was com-
promised is that the observer should at
all times feel that he is fully in control
of his instruments. This concept was
completely followed in the design of the
observer controls, but it was neglected in
the design of the software system. The
operational programs were written in
assembly language to minimize the cost
of program generation, the program ex-
ecution time during critical phases of the
operation, and the required size of the
computer memory. The software system
is now being redesigned to permit the
data collection programs to be written in a
modified FORTRAN language. This new
software system will use almost all of the
existing assembly language subroutines
as well as the computer manufacturer's
FORTRAN programs. As a result, this
effort will require only minimal additional
cost. It is anticipated that the user pro-
gram system will be available in the early
fall of 1973. The computer memory will
have to be enlarged from 16,000 to 32,000
words and a card reader will be added to
facilitate program editing. This new ap-
proach will restore to the observers full
control over the operation of the com-
puter systems.
Electronic Instrument Utilization
Electronic instrumentation has con-
tinued to play an important role for the
Hale Observatories' telescopes. The fol-
lowing table indicates the number of
nights for which AEL had at least some
degree of responsibility for the observing
instruments. The actual utilization of
electronic equipment was somewhat
higher because a few of the observers
used their own electronic equipment with-
out the asisstance of AEL.
No. of
Observatory
Telescope
Nights
Percent
Palomar
200-inch
279
76
Palomar
60-inch
251
69
Mount Wilson
100-inch
217
60
Mount Wilson
60-inch
176
49
PHOTOGRAPHIC LABORATORY
Miller continued tests of methods of
reducing low-intensity reciprocity failure
by outgassing oxygen and water vapor
from photographic emulsions. Experi-
ments using vacuum, reported in Year
Book 71 (p. 707), were followed by tests
using dry nitrogen without recourse to
vacuum. The results closely paralleled
those found for vacuum, with several
advantages in favor of the inert gas. The
nitrogen system was adopted for general
use on Mount Wilson and Palomar.
Existing baking equipment is being con-
verted to the use of nitrogen as rapidly
as possible.
For 2-hour exposures the increase in
efficiency of Ha and 103a emulsions
baked optimally in nitrogen amounts to
around 2 times, and for the new Ilia
emulsions the gain factor ranges from 3
times to nearly 10 times for different
batches. Optimal baking times at 65° C
in nitrogen range from 4 hours for the
Ilia emulsion to 15 hours for the Ha and
HALE OBSERVATORIES
153
24 hours for the 103a emulsions. Current
efforts are concentrated on establishing
optimum baking times for all types of
emulsions to facilitate application of the
system by observers.
A major plate-salvage program was
instituted to reverse image deterioration
in old plates resulting from faulty proc-
essing in past years before the chemistry
involved was fully understood. Many
plates in the files were developing brown
stain caused by silver sulfide, a product
of slow decomposition of a silver thio-
sulfate complex left in the processed
emulsion by inadequate fixation. The
treatment applied is a chemical bath
which not only returns the photographic
image to neutral silver, but forestalls any
further deterioration. Improved under-
standing of the chemistry of fixation has
resulted in new darkroom techniques that
will prevent deterioration of plates being
taken now and in the future.
The Observatories' long-standing pol-
icy of close cooperation with the East-
man Kodak Company in improving
existing types of astronomical photo-
graphic plates and devising new types
was continued. Miller made many labo-
ratory tests of samples submitted by the
Kodak Research Laboratories. Tele-
scope tests of the most promising samples
were made to confirm laboratory tests
and to give quantitative results regard-
ing the astronomical value of the experi-
mental emulsions. The objective of cur-
rent experiments is the production of a
red-sensitive counterpart of the high
Detective Quantum Efficiency Illa-J
emulsion that has proved so popular. It
is hoped that such a red-sensitive emul-
sion will help significantly in the detec-
tion of very distant objects whose large
redshifts leave very little detectable en-
ergy in the blue region of the spectrum.
Making use of a greatly improved
image-registration device, Bedke has
produced impressive "integration" prints
that combine the information contained
in several separate negatives of astro-
nomical objects in a single print with
greatly improved signal-to-noise charac-
teristics. These prints increase visibility
of very faint details by increasing the
limiting magnitude beyond that avail-
able from any one of the negatives alone.
Miller has continued routine quality-
control tests on all shipments of
photographic plates received to ensure
maximum efficiency of photographic ob-
servations at the telescopes.
The Photo Lab had several visitors from
other observatories during the year, who
came to receive training in astronomical
photographic techniques. These were:
Mr. William Schoening (Kitt Peak Na-
tional Observatory), Miss Elizabeth Sim
(Royal Observatory, Edinburgh), Mr.
Hans Schuster (European Southern Ob-
servatory, La Silla, Chile), Mr. Hans
Vogel (European Southern Observatory,
Santiago, Chile) , and Mr. Robert Wilson
(Lick Observatory).
100-INCH DU PONT TELESCOPE
Under the general supervision of Rule
as Project Officer, and with the collabora-
tion of many members of the scientific,
engineering, and administrative staffs,
substantial progress was made in meeting
scheduled design requirements, award of
contracts, guidance of manufacturing op-
erations, procurement and delivery of
site construction components for the du
Pont Telescope and dome building. Some
15 major fabrication contracts for tele-
scope and dome items were committed by
April 1973. These, together with most
other components, are in advanced stages
of inspection and completion in prepara-
tion for shipment to Chile, with assembly
planned at the Las Campanas site in
1974, after the dome and the building
are ready.
154
CARNEGIE INSTITUTION
Mechanical Parts and Building
All telescope heavy parts are complete,
with subassemblies of drives, bearings,
and auxiliary mechanisms undergoing
acceptance tests now for plant trials of
the complete telescope assembly during
August and September, after which the
telescope will be dismantled and packed
for shipment.
Dome construction work at the site
was started in December 1972. Excava-
tion is complete, with concrete poured for
pier, foundations, basement, and sub-
walls. Some contractor delays were ex-
perienced, but work is progressing to
complete the building up to the circular
dome rail to be ready for structural steel
dome erection by about November 1973.
Fabrication of the rotating dome was
started in February, with most heavy
parts complete in June. Fit-up and as-
sembly of the completed dome structure
will be made at the contractor's plant by
about September.
Other building equipment, hardware,
mechanical supplies, and interior em-
bedment steel have been obtained and
shipped as needed. Major equipment
items such as dome crane, hoists, passen-
ger elevator, central ram platform, and
all equipment for heating and air con-
ditioning systems are all in process,
nearly complete, or ready for shipment.
The aluminizing system for mirror
coating to accommodate primary and all
secondary mirrors was engineered in de-
tail. Fabrication was started in April
by High Vacuum Equipment Corpora-
tion for coating tests to be made by
December 1973.
Optics
Following tests at the Optical Sciences
Center of the primary mirror support cell
and grinding table supports with the
mirror at a spherical figure, the operation
of polishing to achieve the final aspheric
figure was started in early summer 1972
and continued throughout the year. This
work is under the supervision of Mr. Don
Loomis, Chief Optician for the telescope
optics, with close cooperation in testing
by Vaughan and Fair. Design and fabri-
cation of optical test equipment and de-
velopment of computational procedures
were pursued at O.S.C. and by Vaughan
in Pasadena. By April 1973 all test pro-
cedures required for evaluation of the
final optical figure of the primary were
in operation. The mirror figure was
brought to within % micron of the final
surface by June 1973, with the work ex-
pected to continue for several weeks.
The coude elliptical flat and Cassegrain
f/7.5 and f/13.5 secondary blanks were
delivered in late fall to Loomis Custom
Optics after preliminary polishing in
Pasadena. By June 1973 the f/13.5 sec-
ondary had been completed and awaited
final acceptance tests. Preliminary pol-
ishing of the Gascoigne-Bowen corrector
plate and fabrication of a spherical test
mirror for the corrector were in progress
in Pasadena.
Auxiliary Instruments
Following a number of general meet-
ings to discuss auxiliary instrumentation,
several working groups were set up by
Oke to supervise the design and construc-
tion of specific instruments and facilities.
It is the hope that most of the initial set
of instruments will be ready for use when
the telescope is completed. The groups
(with the chairman) are listed below:
Direct Plate Holder and Darkroom Proc-
essing Equipment (Sandage) ; Instrument
Mounting Base (Oke) ; TV Guiding Sys-
tem (Dennison) ; Image-Tube Spectro-
graph (Searle) ; Broad-Band Photometer
(Kristian) ; Infrared Photometer (Neuge-
bauer); 90-mm Image Tube (Brucato) ;
Coude Spectrograph (Vaughan) ; Casse-
grain Chair (Rule) ; Computer Inter-
faces (Neugebauer). Design work has
started on direct plate holders, instru-
ment mounting bases, and the Cassegrain
observer's chair.
HALE OBSERVATORIES
155
Control System
During the report period, substantial
progress has been made by the Astro-
electronics Laboratory under Dennison
toward the design and completion of the
control system for the du Pont Telescope.
Much of the effort has gone into ensuring
that the design is satisfactory in two fun-
damental interface areas: between the
control system and the mechanical parts
of the telescope and between the control
system and the operator. This initial
study phase was essential to minimize
the possibility of human error in operat-
ing the telescope and to eliminate the
necessity for on-site modification of the
system when it is installed with the
telescope in Chile.
To increase its reliability and to facil-
itate telescope checking, the system has
been designed and built to enable all
telescope functions to be operated, on oc-
casion, from portable control boxes that
are totally independent of the basic tele-
scope control system. This facilitates
checking of the telescope drive at the
fabrication plant and also guarantees
that all telescope functions can be in-
dividually operated as soon as the tele-
scope is erected in Chile. This concept
also gives an independent means of oper-
ating the telescope in case of control sys-
tem failure. The telescope-control sys-
tem is being designed with solid state
logic and operational interlocks and with
mechanical relay safety and limit inter-
locks, which give maximum operational
lifetime and protection for the telescope
and provide an efficient computer inter-
face.
Computer and Data System
A substantial beginning has been made
on the computer and data system. Many
of the purchased system components have
been delivered and are operational. The
individual interface units that are to be
built by AEL will generally be identical
with the existing interface units now in
operation at several telescopes of the
Hale Observatories. By using these
existing designs, substantial savings in
time and cost should be achieved in both
hardware and software. The software
savings will be particularly outstanding
because it should be possible to use most
of the 250 subroutines that now exist in
our currently operating system.
The only remaining major problem is
to provide the observers with an easily
modifiable high-level programming sys-
tem. The user's software system must
provide for almost complete independ-
ence between the control and data sys-
tems. There are many practical solutions
to this problem, and it is expected that a
decision, based on both cost and effec-
tiveness considerations, will be arrived at
shortly.
LAS CAMPANAS OBSERVATORY
The average temperature for the year
was 18° C with a maximum of 30° C on
January 6, 1973, and a minimum of
— 11° C on August 26, 1972. Total pre-
cipitation for the year was 76 millimeters,
of which 36 millimeters occurred in the
form of rain. Total snowfall was meas-
ured at 46 centimeters. Approximately
280 nights were suitable for observing.
Following the stormy months of July
and August 1972, the 40-inch telescope
was in regular use for all dark runs and
some of the light runs during the year.
Staff members or staff associates of the
Hale Observatories who observed with
the 40-inch during the year ending June
30, 1973, were Becklin, Brucato, Munch,
Neugebauer, Searle, Vaughan, and Bab-
cock. Guest investigators included Mr.
G. Alcaino of the Catholic University,
Santiago ; Dr. P. Cruvellier of the Obser-
vatoire de Marseille; Dr. Jay A. Frogel
of Harvard College Observatory; Mr.
Eduardo Hardy of the University of
156
CARNEGIE INSTITUTION
Chile; Dr. Roberta M. Humphreys of the
University of Arizona; Dr. Edward P.
Ney of the University of Minnesota ; Dr.
John Norris of Yale University Observa-
tory; Dr. S. Eric Persson of Harvard
College Observatory; and Dr. John
Warner of the Department of Terrestrial
Magnetism, Carnegie Institution.
The 24-inch telescope of the Univer-
sity of Toronto was in regular operation
throughout the year.
The working group of approximately
32 Chilean employees and laborers as-
sisted in the operation of the telescopes,
provided logistical services, maintained
equipment, assisted the contractor in
various phases of construction, and con-
tinued the work of site development, all
under the general supervision of Adkison
in Pasadena and Wagner in Chile.
Site Development
For the water system, an additional 7
kilometers of pipeline was laid and the
contractor completed the major reservoir
of 200,000-gallon capacity on the slope
of Las Campanas. The structure of the
new powerhouse was largely completed
and the new 300 kilowatt diesel genera-
tor was placed in position. Satisfactory
electrical service is being provided tem-
porarily by two gasoline generators. A
materials storage yard was provided and
a much needed warehouse was well on
the way to completion. Construction of
the Lodge was begun. Three temporary
dormitory rooms were provided on the
mountain. In La Serena, three new
dormitory rooms were built adjacent to
the El Pino office.
Regular shortwave communications
were maintained between Pasadena and
the El Pino office and between El Pino
and Las Campanas. Frequent negotia-
tions between the Pasadena office and the
shipping lines and air freight carriers
have achieved an acceptable if not com-
pletely predictable level of transport
capability.
BIBLIOGRAPHY
Adelman, Saul J., A model atmosphere abun-
dance analysis of the B9V star v Capricorni.
Astrophys. J., 182, 531-538, 1973.
Adelman, Saul J., and Wallace L. W. Sargent,
A peculiar B star in the old disk population.
Astrophys. J., 176, 671-676, 1972.
Agnew, Duncan, and Halton Arp? A list of
quasi-stellar sources and quasi-stellar source
candidates from 3C and 4C catalogs between
Dec —7° and +40°. Publ. Astron. Soc.
Pacific, 85, 162-173, 1973.
Allen, David A., Near infrared magnitudes of
248 early-type emission-line stars and related
objects. Mon. Notic. Roy. Astron. Soc, 161,
145-166, 1973.
Allen, David A., The method of determining
infrared diameters, in Physical Studies of the
Minor Planets, NASA SP-276, pp. 41-43, 1972.
Allen, David A., and F. C. Porter, Infrared
photometry of southern Wolf-Rayet stars.
Astron.. Astrophys., 22, 159-160, 1973.
Allen, David A., P. M. Harvey, and J. P.
Swings, Infrared photometry of northern
Wolf-Rayet stars. Astron. Astrophys., 20,
333-336, 1972.
Allen, David A., see also Swings, J. P.
Altschuler, Martin D., see Newkirk, Gordon, Jr.
Arp, Halton, and Frank Brueckel, Diameters of
H II regions in M31 and comparison to the
largest regions in M33. Astrophys. J., 179,
445-451, 1973.
Arp, Halton, and John Kormendy, Faint sur-
face-brightness features between NGC 7331
and Stephan's quintet. Astrophys. J. (Lett.),
178, L101-L103, 1972.
Arp, Halton, see also Agnew, Duncan.
Babcock, Horace W., Stellar magnetic fields:
5i-5. Spectral observations, in American In-
stitute of Physics Handbook, 3d ed., pp.
5-306-5-310, Dwight E. Gray coord, ed.,
McGraw-Hill Book Company, New York,
1972.
Babcock, Horace W., Ira Sprague Bowen, 1898-
1973. Eng. Sci., 36, 26-27, 1973.
Bagnuolo, W. G., see Searle, Leonard.
Bahcall, J., see Spinrad, Hyron.
Becklin, E. E., E. B. Fomalont, and G. Neuge-
bauer, Infrared and radio observations of the
nucleus of NGC 253. Astrophys. J. (Lett.),
181, L27-L31, 1973.
Becklin, E. E., G. Neugebauer, and C. G.
Wynn-Williams, On the nature of the infra-
red point source in the Orion Nebula. Astro-
phys. J. (Lett.), 182, L7-L9, 1973.
HALE OBSERVATORIES
157
Becklin, E. E., G. Neugebauer, and C. G.
Wynn-Williams, Infrared emission from the
OH/H2O sources in W49. Astrophys. Lett.,
13, 147-149, 1973.
Becklin, E. E., J. Kristian, G. Neugebauer, and
C. G. Wynn-Williams, Discovery of infrared
emission from the radio source near Cygnus
X-3. Nature, 239, 130-131, 1972.
Becklin, E. E., J. A. Frogel, G. Neugebauer,
S. E. Persson. and C. G. Wynn-Williams, The
H II region G333 .6-0.2, a very powerful 1-20
(x source. Astrophys. J. (Lett.), 182, L125-
L129, 1973.
Becklin, E. E., see also Epstein, E. E.; Kirsh-
ner, R. P.; Neugebauer, G.; Spinrad, Hvron;
Wilson, W. J.; Woolf, N. J.; Wynn-Williams,
C. G.
Bhatnagar, Arvind. W. C. Livingston, and J. W.
Harvey, Observations of sunspot umbral
velocity oscillations. Solar Phys., 27, 80-88.
1972.
Bowen, I. S., and A. H. Vaughan, Jr., "Non-
objective" gratings. Publ. Astron. Soc. Pacific,
85, 174-176, 1973.
Braccesi, A., L. Formiggini, I. Gioia, and
W. L. W. Sargent, Two galaxies dominated
by bright ultraviolet knots. Publ. Astron. Soc.
Pacific, 84, 592-593, 1972.
Brucato, Robert, and Jerome Kristian, Spec-
troscopic observations of the optical candi-
date for Cygnus X-l. Astrophys. J. (Lett.),
179, L129-L133, 1973.
Brucato, Robert J., and Jerome Kristian,
BD+34°3815, Int. Astron. Union Circular
No. 2421, July 6, 1972.
Brucato, Robert J., see also Westphal, J. A.
Brueckel, Frank, see Arp, Halton.
Chaldu, R., R. K. Honeycutt, and M. V. Pen-
ston. The extinction curve for Cygnus OB2
No. 12. Publ. Astron. Soc. Pacific, 85, 87-90,
1973.
Conti, Peter S., see Wallerstein, George.
Cragg, Thomas A., The classical Cepheid pro-
gram, J. D. 2,440,000-2,441,000. /. Amer.
Assoc. Var. Star Observers, 1, 9-15, 1972.
Crutcher, Richard M., see Quirk, William J.
Dennison, Edwin W., An integrating television
system for visual enhancement of faint stars,
in Advances in Electronics and Electron
Physics. Vol. 33B, pp. 795-800, J. D. McGee,
D. McMullen, and E. Kahan, eds., Academic
Press, London. 1973.
Dravins, Dainis, A possible solar electrograph.
Astrophys. Lett., 13, 243-245, 1973.
Dworetsky, Michael M., The orbit of 53 Tauri
(B9p-Mn). Publ. Astron. Soc. Pacific, 84,
652-655, 1972.
Dworetsky, Michael M., and Arthur H.
Vaughan, Jr., Wavelength shifts and probable
istotopic structure of Pt II X4046 in mercury
stars. Astrophys. J., 182, 811-816, 1973.
Edberg, Stephen J., see Howard, Robert.
Epstein, E. E., E. E. Becklin, G. Neugebauer,
and C. G. Wynn-Williams, and 20 coauthors,
3C120, BL Lacertae, and OJ 287 : coordinated
optical, infrared, and radio observations of
intraday variability. Astrophys. J. (Lett.),
178, L51-L59, 1972.
Fomalont, E. B., see Becklin, E. E.
Formiggini, L., see Braccesi, A.
Foukal, Peter, Measurement of the electron
temperature of small 3-cm radio bursts. Solar
Phys., 24, 411-413, 1972.
Foukal, Peter, and Harold Zirin, Reply to "The
relations between chromospheric features and
photospheric magnetic fields" by E. N.
Frazier. Solar Phys., 26, 148-150, 1972.
Frogel, J. A., see Becklin, E. E.
Gillett, F. C, see Woolf, N. J.
Gioia, I., see Braccesi, A.
Glackin, David L., Spatial distribution of
emerging flux regions. Publ. Astron. Soc.
Pacific, 85, 241-248, 1973.
Goad, L. E., L. Goldberg, and Jesse L. Green-
stein, High-n. Balmer transitions in gaseous
nebulae. Astrophys. J., 175, 117-125, 1972.
Goldberg, L., see Goad, L. E.
Gott. J. Richard, III, see Gunn, James E.
Grandi, Steven, The expansion of NGC 7293.
Publ. Astron. Soc. Pacific, 85, 200-202, 1973.
Greenstein, Jesse L., A highly evolved, low-
mass binary, HZ 22. Astron. Astrophys., 28,
1-7, 1973.
Greenstein, Jesse L., The future of astronomy.
Eng. Sci., 36, 18-23, 1972.
Greenstein, Jesse L., Preface, in Astronomy
and Astrophysics for the 1970s, Vol. 2, Re-
ports of the Panels, pp. vii-x, National Acad-
emy of Sciences, Washington, D.C.. 1973.
Greenstein, Jesse L., and R. Minkowski, An
atlas of supernovae spectra. Astrophys. «/.,
182, 225-243, 1973.
Greenstein, Jesse L., and Deane M. Peterson.
Line profiles and rotation in white dwarfs.
Astron. Astrophys., 25, 29-34, 1973.
Greenstein, Jesse L., and Virginia Trimble, The
gravitational redshift of 40 Eridani B. Astro-
phys. J., 175, L1-L5, 1972.
Greenstein, Jesse L., see also Goad, L. E.;
Trimble, Virginia; Wallerstein, G.
Gunn, James E., and J. Richard Gott, III, On
the infall of matter into clusters of galaxies
and some effects on their evolution. Astro-
phys. J., 176, 1-19, 1972.
Gunn, James E., see also Oemler. Augustus, Jr. ;
Spinrad, Hyron.
158
CARNEGIE INSTITUTION
Harvey, J. W., see Bhatnagar, Arvind.
Harvev, P. M., see Allen, David A.; Wilson,
W.J.
Honeycutt, R. K., see Chaldu, R.
Howard, Robert, General magnetic field of the
sun, in American Institute of Physics Hand-
book, 3d ed., p. 5-304, Dwight E. Gray, coord,
ed., McGraw-Hill Book Company, New York,
1972.
Howard, Robert, Sunspot fields, in American
Institute of Physics Handbook, 3d ed., p.
5-305, Dwight E. Gray, coord, ed., McGraw-
Hill Book Company, New York, 1972.
Howard, Robert, Recent solar research. Science,
177, 1157-1163, 1972.
Howard, Robert, Polar magnetic fields of the
sun: 1960-1971, Solar Phys., 25, 5-13, 1972.
Howard, Robert, Photospheric magnetic fields,
in Solar Wind, NASA SP-308, pp. 3-9, C. P.
Sonett, J. P. Coleman, Jr., and J. M. Wilcox,
eds., National Aeronautics and Space Ad-
ministration, Washington, D.C., 1972.
Howard, Robert, and Stephen J. Edberg, On
the random nature of the eruption of magnetic
flux at the solar surface. Solar Phys., 28, 73-
75, 1973.
Howard, Robert, see also Newkirk, Gordon, Jr.;
Schatten, Kenneth H.
Huchra, John, and S .P. Willner, UBV photom-
etry of selected Ap stars. Publ. Astron. Soc.
Pacific, 85, 85-86, 1973.
Huchra, John, see also Westphal, J. A.
Hunten, Donald M., and Guido Munch, Helium
abundance in Jupiter. Space Sci. Rev., 14,
433-443, 1973.
Kirshner, R. P., S. P. Willner, E. E. Becklin, G.
Neugebauer, and J. B. Oke, Spectrophotom-
etry of the supernova in NGC 5253 from 0.33
fi to 2.2 /i. Astrophys. J. (Lett.), 180, L97-
L100, 1973.
Kormendy, John, Calibration of direct photo-
graphs using brightness profiles of field stars.
Astron. J., 78, 255-262, 1973.
Kormendy, John, see also Arp, Halton.
Kowal, C. T., Comet Gehrels (1972e), Intern.
Astron. Union Circular No. 2469, December
12, 1972.
Kowal, C. T., W. L. W. Sargent, L. Searle, and
F. Zwicky, The 1971 Palomar supernova
search. Publ. Astron. Soc. Pacific, 84, 844-849,
1972.
Kristian, Jerome, Quasars as events in the nu-
clei of galaxies: the evidence from direct
photographs. Astrophys. J. (Lett.), 179, L95-
L99, 1973.
Kristian, Jerome, A summary of optical pulsar
observations, in Proceedings of Symposium
on Pulsars and High Energy Activity in
Supernova Remnants, pp. 113-120, L. Grat-
ton, ed., Accademia Nazionale dei Lincei,
1972.
Kristian, Jerome, see also Becklin, E. E.; Bru-
cato, Robert; Spinrad, Hyron; Westphal,
J. A.
Kruit, P. C. van der, Spiral structure and nu-
clear activity in galaxies. Nature, 243, 127-
130, 1973.
Kuiper, T. B. H., and Jay M. Pasachoff, De-
tailed correlation of type III radio bursts
with Ha activity, I, Active region of 22 May
1970. Solar Phys., 28, 187-196, 1973.
Lanning, Howard, A finding list of faint UV-
bright stars in the galactic plane. Publ.
Astron. Soc. Pacific, 85, 70-84, 1973.
Leighton, Robert B., see Schatten, Kenneth H.
Livingston, W. C, see Bhatnagar, Arvind.
McKee, Christopher F., and Wallace L. W.
Sargent, The absorption-line spectrum of the
bright QSO Markarian 132. Astrophys. J.
(Lett.), 182, L99-L101, 1973.
Merrill, K. M., see Woolf, N. J.
Minkowski, R., see Greenstein, Jesse L.
Munch, Guido, see Hunten, Donald M.
Neugebauer, G., and Eric E. Becklin, The
brightest infrared sources. Sci. Amer., 228,
28-40, 1973.
Neugebauer, G., see also Becklin, E. E.; Ep-
stein, E. E.; Kirshner, R. P.; Spinrad, Hy-
ron; Wilson, W. J.; Woolf, N. J.; Wynn-
Williams, C. G.
Newkirk, Gordon, Jr., Dorothy E. Trotter,
Martin D. Altschuler, and Robert Howard,
Atlas of Magnetic Fields in the Solar Corona,
35 mm film, High Altitude Observatory,
Boulder, Colorado, 1972; Research note,
Solar Phys., 24, 370-372, 1972.
Oemler, Augustus, Jr., The cluster of galaxies
Abell 2670. Astrophys. J., 180, 11-23, 1973.
Oemler, Augustus, Jr., James E. Gunn, and J.
Beverley Oke, 3C232.1 : A QSO in a rich clus-
ter of galaxies. Astrophys. J. (Lett.), 176,
L47-L50, 1972.
Oke, J. B., and E. M. Shoemaker, Object
Kowal, Int. Astron. Union Circular No. 2509,
March 19, 1973.
Oke, J. B., see also Kirshner, R. P;. Oemler,
Augustus, Jr.; Shields, G. A.
Pasachoff, Jay M., see Kuiper, T. B. H.
Penston, M. J., see Penston, M. V.
Penston, M. V., Multicolor observations of stars
in the vicinity of the Orion Nebula, in Stellar
Ages, International Astronomical Union Col-
loquium No. 17, pp. IX-l-IX-6, G. Cayrel de
Strobel and A. M. Delplace, eds., Paris-
Meudon Observatory, France, 1973.
HALE OBSERVATORIES
159
Periston, M. V., and M. J. Penston, New obser-
vations of two compact galaxies. Mon. Notic.
Roy. Astron. Soc, 162, 109-116, 1973.
Penston, M. V., see also Chaldu, R.
Pepin, T. J., see Woolf, N. J.
Persson, S. E., see Becklin, E. E.
Peterson, Deane M., and H. L. Shipman, Helium
abundances in NGC 2264, II Scorpii, and I
Lacertae. Astrophys. J., 180, 635-645, 1973.
Peterson, Deane M., see also Greenstein, Jesse L.
Pope, Thomas, see Vorpahl, Joan.
Porter, F. C, see Allen, David A.
Prata, Stephen W., Some aspects of flare prop-
erties versus magnetic boundary morphology.
Solar Phys., 25, 136-140, 1972.
Preston, George W., Stellar astronomy objec-
tives, in Astronomy from a Space Platform,
Vol. 28, Science and Technology, 1972, pp.
95-104, G. W. Morgenthaler, ed., American
Astronautical Society, Tarzana, California,
1972.
Preston, George, The unique magnetic and
spectrum variations of HD 24712. Astrophys.
J., 175, 465^72. 1972.
Quirk, William J.. On the gas content of gal-
axies. Astrophys. J. (Lett.), 176, L9-L14,
1972.
Quirk, William J., Do interstellar gas clouds
exist between spiral arms? Astrophys. J.,
180, 25-30, 1973.
Quirk, William J., and Richard M. Crutcher,
Evidence for a shock associated with the local
spiral arm. Astrophys. J., 181, 359-362, 1973.
Richstone, Douglas, and Wallace L. W. Sargent,
A new determination of the mass of M32
Astrophys. J., 176, 91-98, 1972.
Sandage, Allan, Absolute magnitude of E and
SO galaxies in the Virgo and Coma clusters
as a function of U — B color. Astrophys. J.,
176, 21-30, 1972.
Sandage, Allan, The redshift-distance relation,
II, The Hubble diagram and its scatter for
first-ranked cluster galaxies: a formal value
for q0. Astrophys. J., 178, 1-24, 1972.
Sandage, Allan, The redshift-distance relation,
III, Photometry and the Hubble diagram for
radio sources and the possible turn-on time
for QSOs. Astrophys. J., 178, 25-44, 1972.
Sandage, Allan, The redshift-distance relation,
IV, The composite nature of N galaxies, their
Hubble diagram, and the validity of the
measured redshifts as distance indicators.
Astrophys. J., 180, 687-697, 1973.
Sandage, Allan, Opening the last frontier. Eng.
Sci., 36, 4-10, 1973.
Sandage, Allan, and Basil Katem, Positions of
comet Sandage 1972h. Int Astron. Union Cir-
cular No. 2416, June 21, 1972.
Sandage, Allan, see also Visvanathan, N.
Sargent, Wallace L. W., The cluster of compact
galaxies ZW CI 0152+33. Astrophys. J., 176,
581-587, 1972.
Sargent, Wallace L. W., Redshifts for 51 gal-
axies identified with radio sources in the 4C
catalogue. Astrophys. J. (Lett.), 182, L13-
L15, 1973.
Sargent, Wallace L. W., The redshifts of ten
clusters of galaxies. Publ. Astron. Soc. Pacific,
84, 281-282, 1973.
Sargent, Wallace L. W., Lectures on compact
galaxies, in Scuola di Fisica Cosmica, Sum-
mer Course in High-Energy Physics, pp. 1-37,
at the "Ettore Majorana" Center for Scien-
tific Culture, Erice, Italy, June 1971.
Sargent, Wallace L. W., The properties of extra-
galactic X-ray sources from visible light ob-
servations, in X and Gamma Ray Astronomy,
Int. Astron. Union Symp. No. 55, pp. 184-198,
H. Bradt and R. Giacconi, eds., D. Reidel
Publishing Co., Dordrecht-Holland, 1973.
Sargent, Wallace L. W ., and Leonard Searle,
Supernova in NGC 2841, Int. Astron. Union
Circular No. 2498, February 27, 1973.
•Sargent, Wallace L. W., see also Adelman, Saul
J.; Braccesi, A.; Kowal, C; McKee, Christo-
pher F.; Richstone, Douglas; Searle, Leon-
ard; Shields, G. A.; Spinrad, Hyron.
Schatten, Kenneth H., Robert B. Leighton,
Robert Howard, and John M. Wilcox, Large-
scale photospheric magnetic field: the diffu-
sion of active-region fields. Solar Phys., 26,
283-289, 1972.
Schmidt, Maarten, Statistical studies of the
evolution of extragalactic radio sources, I,
Quasars. Astrophys. J., 176, 273-287, 1972.
Schmidt, Maarten, Statistical studies of the
evolution of extragalactic radio sources, II,
Radio galaxies. Astrophys. J., 176, 289-301,
1972.
Schmidt, Maarten, Statistical studies of the
evolution of extragalactic radio sources, III,
Interpretation of source counts and discus-
sions. Astrophys. J., 176, 303-314, 1972.
Schmidt, Maarten, Comet Huchra (1973h).
Int. Astron. Union Circular Nos. 2423, July
21, 1972; 2533, May 7, 1973; 2534, May 10,
1973.
Schmidt, Maarten, Supernova in an anonymous
galaxy. Int. Astron. Union Circular Nos.
2U7," September 28, 1972; 2485, February 1.
1973; 2487, February 7, 1973.
Schmidt, Maarten, Supernova in NGC 735.
Int. Astron. Union Circular No. 2448. October
3, 1972.
Schmidt, Maarten, Supernova in NGC 3656.
Int. Astron. Union Circular No. 2491, Feb-
ruary 12, 1973.
160
CARNEGIE INSTITUTION
Schmidt, Maarten, 1973 EA. Int. Astron. Union
Circular No. 2532, May 1, 1973.
Schmidt, Maarten, Radio source counts and
redshifts in steady state cosmology. Nature,
240, 399-400, 1972.
Schmidt, Maarten, Reply to Hoyle. Nature,
242, 109, 1973.
Schwartz, P. R., see Wilson, W. J.
Searle, Leonard, Star formation and the chem-
ical history of galaxies, in Stellar Ages, Inter-
national Astronomical Union Colloquium No.
17, pp. LII-l-LII-15, G. Cayrel de Strobel
and A. M. Delplace, eds., Paris-Meudon Ob-
servatory, France, 1973.
Searle, Leonard, Wallace L. W. Sargent, and
W. G. Bagnuolo, The history of star forma-
tion and the colors of late-type galaxies.
Astrophys. J., 179, 445-451, 1973.
Searle, Leonard, see also Kowal, C. T. ; Sargent,
Wallace L. W.
Sheet-man, Stephen A., Clusters of galaxies and
the cosmic light. Astrophys. J., 179, 681-698,
1973.
Shectman, Stephen A., see also Westphal, J. A.
Shields, G. A., J. B. Oke, and W. L. W. Sargent,
The optical spectrum of the Seyfert galaxy
3C120. Astrophys. J.t 176, 75-89,* 1972.
Shipman, H. L., Masses and radii of white
dwarfs. Astrophys. J., 177, 723-743, 1972.
Shipman, H. L., see also Peterson, Deane M.
Shoemaker, E. M., see Oke, J. B.
Smith, H., see Spinrad, Hyron.
Spinrad, Hyron, J. Bahcall, E. E. Becklin,
James E. Gunn, J. Kristian, G. Neugebauer,
W. L. W. Sargent, and H. Smith, Optical and
near infrared observations of the nearby
spiral galaxy Maffei 2. Astrophys. J., 180,
351-358, 1973.
Stein, W. A., see Woolf, N. J.; Zirin, Harold.
Swings, J. P., and D. A. Allen, Photometry of
symbiotic and W Cephei stars in the near
infrared (with a note on MWC 56). Publ.
Astron.. Soc. Pacific, 84, 523-527, 1972.
Swings, J. P., see also Allen, David A.
Trimble, Virginia, and Jesse L. Greenstein, The
Einstein redshift in white dwarfs, III. Astro-
phys. J., 177, 441-452, 1972.
Trimble, Virginia, see also Greenstein, Jesse L.
Trotter, Dorothy E., see Newkirk, Gordon, Jr.
Vaughan, Arthur H.. Jr., Ira Sprague Bowen,
Phys. Today, 26, 77-78, 1973.
Vaughan, Arthur H., Jr., see also Bowen, I. S.;
Dworetsky, Michael M.
Visvanathan, N., and Allan Sandage, Linear
polarization of the Ha emission line in the
halo of M82 and the radiation mechanism of
the filaments. Astrophys. J., 176, 57-74, 1972.
Vorpahl, Joan, and Thomas Pope, Solar bright
points in 3840 A and Ha. Solar Phys., 25,
347-356, 1972.
Wallerstein, George, Peter S. Conti, and Jesse
L. Greenstein, Interstellar lines in the spec-
trum of the supernova in NGC 5253. Astro-
phys. Lett., 12, 101-102, 1972.
Weart, Spencer, What makes active regions
grow? Astrophys. J., 177, 271-276, 1972.
Weistrop, Donna, The luminosity function and
density distribution of disk population stars.
Astron. J., 77, 849-862, 1972.
Westphal, J. A., J. Kristian, J. P. Huchra, S. A.
Shectman, and R. J. Brucato, Search for a
visible counterpart of the 2 September 1972
radio outburst in Cygnus. Nature, 239, 134-
135, 1972.
Wilcox, John M., see Schatten, Kenneth H.
Willner, S. P., see Huchra, John; Kirshner, R. P.
Wilson, Olin C., Ira Sprague Bowen (1898-
1973). Sky and Telescope, 45, 212-214, 1973.
Wilson, W. J., P. R. Schwartz, G. Neugebauer,
P. M. Harvey, and E. E. Becklin, Infrared
stars with strong 1665/1667-MHz OH micro-
wave emission. Astrophys. J., 177, 523-540,
1972.
Woolf, N. J., W. A. Stem, F. C. Gillett, K. M.
Merrill, E. E. Becklin, G. Neugebauer, and
T. J. Pepin, The infrared sources in M8.
Astrophys. J. (Lett.), 179, L111-L115, 1973.
Wynn-Williams, C. G., E. E. Becklin, and G.
Neugebauer, Infrared sources in the H II
region W3. Mon. Notic. Roy. Astron. Soc,
160, 1-14, 1972.
Wynn-Williams, C. G., E. E. Becklin, and G.
Neugebauer, New infrared sources associated
with H II regions, in Interstellar Dust and
Related Topics, Int. Astron. Union Symp. No.
52, J. M. Greenberg and H. C. van de Hulst,
eds., D. Reidel Publishing Co., Dordrecht-
Holland, in press, 1973.
Wynn-Williams, C. G., see also Becklin, E. E.;
Epstein, E. E.
Zirin, Harold, Response to Dr. Frazier's re-
search note. Solar Phys., 26, 145-147, 1972.
Zirin, Harold, The absence of flares in X3835 in
the heating of the chromosphere. Solar Phys.,
26, 393-396, 1972.
Zirin, Harold, High-resolution solar observa-
tions from space, in Space Research XII, pp.
1751-1763, S. A. Bowhill, L. D. Jaffe, and
M. J. Rycroft, eds., D. Reidel Publishing Co.,
Dordrecht-Holland, 1972.
Zirin, Harold, and Alan Stein, Observations of
running penumbral waves. Astrophys. J.
(Lett.), 178, L85-L87, 1972.
Zirin, Harold, see also Foukal, Peter.
HALE OBSERVATORIES
161
Zwicky, Fritz, Radiations in outer space. Astro-
naut. Acta, 17, 129-135, 1972.
Zwicky, Fritz, Morphology of rapid cosmic
processes, Astronaut. Acta, 17, 307-314, 1972.
Zwicky, Fritz, Circular Letter No. 12 on super-
novae, California Institute of Technology,
May 1972.
Zwicky, Fritz, Morphologisches Denken und
Vorgehen, in Die Neuen Method der Ent-
scheidungsftndung, pp. 130-143, Verlag Mod-
erne Industrie, Munich, 1972.
Zwicky, Fritz, In memory of Milton L. Huma-
son. Eng. Sci., 36, 30-31, 1972.
Zwicky, Fritz, 100 supernovae — the reward of a
40-year search. Eng. Sci., 36, 20-21, 1973.
Zwicky, Fritz, and Margrit A. Zwicky, Cata-
logue of Selected Compact Galaxies and of
Post-Eruptive Galaxies, Guemligen (Berne),
Switzerland, 1971.
Zwicky, Fritz, see also Kowal, C. T.
Zwicky, Margrit A., see Zwicky, Fritz.
STAFF AND ORGANIZATION
Ira S. Bowen, Director of the Mount
Wilson and Palomar Observatories from
1946 to 1964, died suddenly on February
6, 1973.
As a graduate student, Bowen came to
the California Institute of Technology in
1921 to begin a productive decade in lab-
oratory spectroscopy, physical optics,
and the study of cosmic rays. His solu-
tion of the enigma of the "nebulium
lines" in 1928 earned him international
renown; it also helped to swing his re-
search interests irrevocably to the field
of astronomy and astrophysics.
Bowen assumed the Observatory di-
rectorship at a time when the 200-inch
telescope was nearing completion and
when the Carnegie Institution and the
California Institute were joining forces
to establish and build a unique observa-
tory organization for the advancement of
research and teaching in astronomy.
During a period of remarkable growth,
Bowen presided over the expansion of the
staff, the formulation of Observatory
policy, and the guidance of research. The
excellent performance and productivity
of the 200-inch telescope reflect the high
scientific and technical standards that
he set for himself and, by example, for
the staff. Since putting aside administra-
tive concerns in 1964, Bowen had been a
Distinguished Service Member of the
Carnegie Institution and had continued
to be active at the Observatories. In re-
cent years his interests had centered on
the design of optical systems for astron-
omy. He made major contributions to
the optical design of such wide-field
telescopes as the Palomar 60-inch, the
Las Campanas 40-inch, and the 100-inch
Irenee du Pont Telescope that is now
under construction for Las Campanas.
Over the years his counsel on the optical
design and technical aspects of telescopes
was widely sought and freely given, to
the benefit of practically all of the
world's large telescopes of the past
quarter-century.
Alfred H. Joy, a staff member from
1915 until his formal retirement in 1948,
died on April 18, 1973, at the age of 90.
He was the last survivor of the group
that constituted the scientific staff prior
to 1930. Internationally known and re-
spected for his contributions to stellar
spectroscopy, especially of variable stars,
he pursued this field with characteristic
energy and devotion for nearly 60 years.
Dr. Joy was Secretary of the Mount Wil-
son Observatory from 1920 until his
retirement.
Dr. Robert J. Brucato was appointed
Staff Associate on September 15, 1972.
Research Division
Distinguished Service Member, Carnegie
Institution of Washington
Ira S. Bowen 1
Staff Members
Halton C. Arp
Horace W. Babcock, Director
Edwin W. Dennison
1 Died February 6, 1973.
162
CARNEGIE INSTITUTION
Jesse L. Greenstein 2
James E. Gunn 3
Robert F. Howard
Jerome Kristian
Robert B. Leighton 4
Guido Munch 3
Gerry Neugebauer 5
J. Beverley Oke, Associate Director 3
George W. Preston
Bruce H. Rule, Chief Engineer
Allan R. Sandage
Wallace L. W. Sargent 3
Maarten Schmidt 6
Leonard T. Searle
Arthur H. Vaughan, Jr.
Olin C. Wilson
Harold Zirin 7
Members Engaged in Post-Retirement
Studies
Alfred H. Joy 8
Alexander Pogo
Henrietta H. Swope
Fritz Zwicky
Visiting Associates
Clark G. Christensen
N. Visvanathan
Staff Associates
Eric E. Becklin 9
Robert J. Brucato
Michael W. Werner 10
James A. Westphal n
2 Lee A. DuBridge Professor of Astrophysics,
California Institute of Technology.
3 Professor of Astronomy, California Institute
of Technology.
4 Professor of Physics ; Chairman, Division of
Physics. Mathematics, and Astronomy, Cali-
fornia Institute of Technology.
5 Professor of Physics, California Institute of
Technology.
6 Professor of Astronomy ; Executive Officer
for Astronomy, California Institute of Tech-
nology.
7 Professor of Astrophysics, California Insti-
tute of Technology.
8 Died April 18, 1973.
9 Senior Research Fellow in Physics, Cali-
fornia Institute of Technology.
10 Assistant Professor of Physics, California
Institute of Technology.
11 Associate Professor of Planetary Science,
California Institute of Technology.
Carnegie Fellows
Ermanno F. Borra
Michael M. Dworetsky
Michael H. Hart
Pieter C. van der Kruit
Richard E. White 12
Robert R. Zappala
Research Fellows
Andrew Michalitsanos
Ronald Moore
Jay M. PasachofT 13
Stephen W. Prata 13
William J. Quirk 13
Katsuo Tanaka 14
Carnegie-Chilean Fellows
Eduardo Hardy 13
Maria Teresa Ruiz
Librarians
Eleanor G. Ellison
Marjorie A. Henderson
Helen Z. Knudsen
Senior Research Assistants
Grace V. Knox
A. Louise Lowen
Research Assistants
John M. Adkins
John E. Boyden
Frank J. Brueckel 15
Ken D. Clardy
Thomas A. Cragg
Thomas S. Gregory
Basil N. Katem
Margaret Katz
Charles Kowal
Howard H. Lanning 16
Susan B. Mellerup
Malcolm S. Riley
Student Observers
Marc Aaronson
William G. Bagnuolo
Ian Gatley
12 Resigned August 15, 1972.
13 Resigned August 31, 1972.
14 Resigned March 31, 1973.
15 Retired December 31, 1972.
16 Resigned October 24, 1972.
HALE OBSERVATORIES
163
Richard Green
Paul M. Harvey
Paul Hickson
James A. Howell
John P. Huchra
Robert P. Kirshner
John Kormendy
Elliot C. Lepler
Steven J. Loer
Jorge Melnick
Augustus Oemler, Jr.
Valdar Oinas
Glenn S. Orton
Paul L. Schechter
Stephen A. Shectman
Barry E. Turnrose
Glenn J. Veeder
Bruce Waddington
William E. Westbrook
Richard A. Wickes
Steven P. Willner
Photographic Department
John R. Bedke, Photographer
William C. Miller, Research Photographer
Instrument Design and Construction
David A. Bell, Electronics Specialist
Lawrence E. Blakee, Supervisor, Electronic
Services
Carroll Parrott Blue, Junior Photographic
Laboratory Technician
Robert E. Cadman, Senior Electronics Spe-
cialist
Ertugrul Cepni, Electronics Engineer 17
Cynthia V. Cable, Laboratory Technician
Maynard K. Clark, Electronics Operations
Manager
John P. Cowley, Engineering Assistant
Martin A. Danihel, Junior Programmer
Floyd E. Day, Optical Shop Superintend-
ent18
Theresa M. DiBella, Junior Electronics
Specialist 19
Stephen Doro, Machinist
Raymond Dreiling, Machinist
Earle B. Emery, Research Engineer
Eugene B. Fair, Optician
Bradley R. Feinner, Senior Electronics
Specialist
Wiley P. Fowlie, Electronics Engineer
17 Resigned August 14, 1972.
18 Retired June 30, 1973.
19 Resigned March 16, 1973.
Jerry T. Fridenberg, Chief Electronics En-
gineer
Robert D. Georgen, Machinist
Richard M. Goeden, Engineer
Simon Groesz, Senior Electronics Techni-
cian
Donn M. Hall, Senior Programmer
Robert Stephen Hayes, Associate Engi-
neer 20
Robert J. Hippard, Photographic Labora-
tory Aide 21
Melvin W. Johnson, Optician
Herman F. Kelderman, Research Engineer
Leroy M. Kimoto, Senior Electronics Spe-
cialist
John H. Knowles, Junior Electronics Tech-
nician 22
Wilfred H. Leckie, Senior Draftsman
Steve A. Macenka, Senior Engineer
William B. Mainland, Electronics Engi-
neer 23
William H. McLellan, Senior Engineer
Charles R. McMahan, Electronics Engineer
Martin J. Olsiewski, Senior Electronics
Specialist
Frederick G. O'Neil, Shop Foreman
Emilio B. Perez, Junior Electronics Tech-
nician
Kenneth C. Perry, Junior Draftsman
Rudolf E. Ribbens, Design and Shop Su-
perintendent
Alexander Rittenhouse, Junior Electronics
Technician
Robert G. Stiles, Optician
Merle R. Sweet, Electronics Technician
David F. Thompson, Technical Assistant
Virgil Z. Vaughan, Supervisor, Electronics
Construction
Robert V. Webber, Laboratory Techni-
cian 24
Madeleine B. Williams, Draftswoman 25
Felice Woodworth, Draftswoman
Maintenance and Operation
Mount Wilson Observatory and Offices
Fern V. Borgen, Typist-Telephone Opera-
tor
Clyde B. Bornhurst, Mountain Superin-
tendent
20 Resigned February 9, 1973.
21 Resigned April 6, 1973.
22 Resigned October 6, 1972.
23 Resigned November 3, 1972.
24 Resigned November 17, 1972.
164
CARNEGIE INSTITUTION
Herman E. Carpentier, Carpenter
Hugh T. Couch, Superintendent, Buildings
and Grounds
Helen S. Czaplicki, Typist-Editor
Sue H. DeWitt, Secretary
James E. Dittmar, Night Assistant
Raquel E. Ferrer, Translator-Secretary
Hazel M. Fulton, Head Stewardess
Eugene L. Hancock, Night Assistant
Mary Hark, Stewardess
John Kondratowicz, Jr., Night Assistant
Jose Lopez-Tiana, Accounting Clerk
Peter Mastrosimone, Custodian
Jane Newcomb, Bookkeeper 26
Glen Sanger, Senior Custodian
Clair E. Sharp, Accountant
William D. St. John, Chauffeur
Frank Trylko, Custodian
Warren Weaver, Assistant Mountain Su-
perintendent 27
Frederick P. Woodson, Assistant to the
Director
Palomar Observatory and Robinson
Laboratory
Ranney G. Adams, Night Assistant
Mayme L. Adkins, Secretary 28
Albert R. Andrews, Junior Maintenance
Mechanic
Ray L. Ballard, Senior Administrative As-
sistant
Mary E. Bauer, Secretary
Kathleen S. Brown, Secretary
Jan Adriaan Bruinsma, Painter and Gen-
eral Maintenance
Maria J. Bruinsma, Housekeeping Aide
Juan R. Carrasco, Night Assistant and
General Mechanic
Rita M. Duprey, Accounting Clerk
Gail Fuqua, Accounting Clerk 29
Beulah Greenlee, Housekeeping Aide
Frank Greenlee, Senior Custodian
Delia Harris, Secretary 26
Liselotte M. Hauck, Secretary
Victor A. Hett, Maintenance Mechanic
Helen Holloway, Secretary
Taras Kiceniuk, Mountain Superintendent
J. Luz Lara, Mechanic
25 Terminated June 30, 1973.
26 Resigned July 31, 1972.
27 Resigned June 30, 1973.
28 Resigned February 23, 1973.
29 Resigned April 4, 1973.
Patricia McKay, Secretary 30
Dorothy J. Mulkey, Records Clerk
Carl D. Palm, Night Assistant
Marilynne J. Rice, Secretary
Tamsin S. Schulte, Secretary
Sidney R. Shannon, Assistant Mountain
Superintendent 31
Gabrielle South, Accounting Clerk 32
Nadia Tesluk, Clerk-Typist
Elsa B. Titchenell, Secretary
Gary M. Tuton, Senior Night Assistant
Paul Van Ligten, Maintenance Mechanic
and Electrician
Ruth E. Weaver, Clerk-Typist
David K. Williams, Relief Night Assistant
and Maintenance Mechanic
Big Bear Solar Observatory
Jack R. Klemroth, Laboratory Assistant
Eugene H. Longbrake, Superintendent and
Senior Electronics Specialist
Charles F. Mason, Custodian
Walter M. Nagao, Jr., Mechanic
Thomas P. Pope, Sr., Chief Observer 33
Georg Schacht, Laboratory Assistant
Las Campanas Observatory
Bruce Adkison, Associate for Administra-
tion, Pasadena
Wilma J. Berkebile, Secretary
S. Thomas Couch, Shipping and Receiving
Clerk
Wayne Thomas, Accounting Clerk
Manfred Wagner, Project Supervisor, La
Serena
Irenee du Pont Telescope Project
Louis E. Beidler, Engineer
Betty A. Browne, Secretary
Oscar D. Dubon, Field Engineer, Las
Campanas
C. L. Friswold, Instrument Designer
Richard C. Haskell, Project Engineer
Harald Scott Jensen,. Engineer 27
Charles W. Jones, Engineer 31
Roger L. Minnix, Engineer
Bruce H. Rule, Chief Engineer and Project
Officer
Edward H. Snoddy, Engineer
Albert Villanueva, Electrical Engineer
30 Resigned June 15, 1973.
31 Resigned April 30, 1973.
32
33
Resigned April 13, 1973.
Resigned September 22, 1972.
Department
of Terrestrial Magnetism
Washington, District of Columbia
Ellis T. Bolton
Director
L. Thomas Aldrich
Associate Director
Carnegie Institution of Washington Year Book 72, 1972—1973
Contents
Introduction 169
Astrophysics 173
Introduction 173
Optical astronomy 173
Radio astronomy 179
Equipment development 179
1.35-cm water-line observations 180
Search for 21-cm hydrogen absorption during flare-ups of Cygnus X-3 ... 181
Survey of neutral hydrogen in the vicinity of 17 galactic supernova remnants . 182
Accurate radio positions for Ohio radio sources, using the NRAO three-element
interferometer at 3.7 cm and 11.1 cm 184
Institute Argentino de Radioastronomia activities 185
Atomic physics 186
Nuclear physics 190
Improvements in the Atomic Physics Observatory 190
Boron-10 191
A study of (p, n) reactions 194
Biophysics 194
Introduction 194
Studies of memory 195
The mouse satellite in two subspecies of Mus musculus 197
Relatedness among several hamsters 200
Instantaneous binding fractions of DNA 204
Enzyme evolution in the Enterobacteriaceae 207
Sonication of DNA to produce fragments suitable for reassociation experiments . 214
Expression and divergence of repeated DNA sequences 217
Labeling of DNA with 125I 217
Cooperative research at California Institute of Technology 221
Interspersion of sea urchin repetitive and nonrepetitive DNA sequences . . . 222
Size distribution of interspersed repetitive sequences in sea urchin DNA by
reassociation and single-strand-specific nuclease digestion 222
Hyperpolymers and another approach to repetitive sequence organization . . 223
General assessment of advances in the knowledge of DNA sequence arrangement 224
Geophysics 224
Introduction 224
Cosmic-ray research 225
A comparison of the anelasticity structure between western South America and
Japan 226
Empirical models for anomalous high-frequency arrivals from deep-focus earth-
quakes in South America 233
Converted P phase from the ScS phase at the inclined deep seismic zone . . . 238
Source spectral relations for earthquakes 245
Borehole strainmeter installation in Japan 246
Project Naririo 247
Electrical conductivity studies 249
Geochemistry and geochronology 252
Geochemical trends in Andean rocks 252
The geochemistry of basalts from Iceland and the Reykjanes Ridge .... 259
Geochemistry of ultramafic inclusions from Salt Lake Crater, Hawaii .... 262
Differential dissolution technique (DDT) : chemical separation of crystals from
glass 268
Trace-element content of liquid formed by partial melting of a garnet lherzolite
at high pressures : a preliminary report 270
Trace-element contents of clinopyroxenes from garnet lherzolites in kimberlites 272
Metasomatic theory : some examples of diffusion and infiltration 276
Ion microprobe analysis of an isotopic diffusion experiment on biotite . . . 280
Direct determination of rock rheology from wavelength selection in folding . . 283
U-Pb isotopic studies of zircons from the Baltimore Gneiss of the Towson Dome,
Maryland 285
U-Pb isotopic studies of zircons from the Gunpowder granite, Baltimore County,
Maryland 288
U-Pb isotopic analyses of zircons from granulite and amphibolite facies rocks
of the West Chester Prong and the Avondale Anticline, southeastern
Pennsylvania 290
Age and origin of zircons from metamorphic rocks in the Manhattan Prong,
White Plains area, southeastern New York 293
Old radiogenic lead components in zircons from the Idaho batholith and its
metasedimentary aureole 297
Effect of regional metamorphism on whole-rock Rb-Sr systems 299
Uranium gain of detrital zircons studied by isotopic analyses and fission track
mapping 302
Special Activities 304
References Cited 306
Bibliography 312
Personnel 314
INTRODUCTION
No field of thought can be properly laid out by men who are merely
measuring with a ruler. Sections of history are liable to be transformed
— or, even where not transformed, greatly vivified — by an imagination
that comes, sweeping like a searchlight, from outside the historical pro-
fession itself. Old hunches are then confirmed by fresh applications of
the evidence or by unexpected correlations between sources. New matter
emerges because things are joined together which it had not occurred to
one to see in juxtaposition. New details are elicited, difficult details be-
come relevant, because of a fresh turn that the argument has taken.
Herbert Butterfield
The Sleepwalkers, by Arthur Koestler,
Grosset & Dunlap, New York, 1959
This report illustrates the thrust of the
quotation from Professor Butterfield, an
historian of science. It is no accident
that, though we measure with astonish-
ingly sophisticated "rulers," new matter
has emerged. The reader will perceive
this fact in our efforts this year.
Rubin and Ford have brought to frui-
tion their studies of the internal motions
in Andromeda — a galaxy generally be-
lieved to be a homologue of our own. It
can be stated that our "sister" is dynam-
ically and structurally complex, and still
full of puzzles. Particularly, old stars
and (presumably) young gas have similar
motions. Perhaps the gas and associated
dust are shed outwrard from stars which
are undergoing evolution — losing mass,
and changing their chemical composition
through nuclear transmutations. Most
exciting at present is the recent discovery
of the anisotropic distribution of veloc-
ities on the celestial sphere of a class of
galaxies known as Scl — a special class of
spiral galaxies of high luminosity.
From an all-sky sample of 208 such
galaxies, radial velocities, i.e., the speeds
at wdiich these objects are receding from
the Sun, have been determined for 70-odd ;
50 velocities range from 4000 to 7000 kilo-
meters per second. Strikingly, 28 of these
have velocities tightly grouped about 5000
km/sec (actually, 4966 ± 122 km/sec)
wdiile 22 have a similar grouping about
6500 km/sec (actually, 6431 ± 160 km/
sec) . Each group is positioned on a differ-
ent region of the sky ( approximately one-
third of the total for each) and there is
virtually no overlap in the positions of
the two groups. These observations raise
most important cosmological problems;
the most unorthodox is to question the
validity of the wridely accepted idea that
the Hubble expansion (which relates re-
cessional velocity to distance) is isotropic
throughout the universe. Other alterna-
tives are discussed in the Astrophysics
Section report.
In the report of this section are also in-
cluded observations and conclusions in
radio astronomy, atomic and nuclear
physics. They include the work at the
NRAO, Greenbank, West Virginia, at our
Derwood, Maryland, Observatory, and
the cooperative effort writh the joint CIW-
Argentine installation outside La Plata
where Dr. Kenneth Turner has served as
interim director for the past two years.
The focus at NRAO has been on 21 -cm
hydrogen line observations of large
galactic supernova remnants in order
to study the effect of supernova rem-
nants on the interstellar medium. The
South American studies have centered
169
170
CARNEGIE INSTITUTION
principally upon neutral hydrogen radia-
tion of selected objects. Many Argentine
students and scientists have engaged in
this almost unexplored area and they are
beginning to yield some of the harvest
envisioned a decade and more ago. It is
our conviction that continuing coopera-
tion will prove fruitful.
At our local Derwood Observatory a
large investment of talent and time has
been expended in improving the facility
for work at higher frequencies than 1420
MHz in order to study molecules in
our Galaxy and also to have at hand
a readily available facility for Fellows
and local students, especially from the
University of Maryland and The Johns
Hopkins University. Nevertheless, the
H-line work, which has been our main-
stay, continues with much improved effi-
ciency, with continuing emphasis on the
study of high velocity neutral hydrogen
clouds.
In the areas of atomic and nuclear
physics our work, largely under Dr.
Louis Brown's aegis, has taken on
new significance. Twenty years ago
Drs. Norman Heydenburg and Georges
Temmer studied low-lying excited states
of nuclei through Coulomb excitation
with a-particles. In recent years it has
become possible in this and other labora-
tories to study atoms by bombardment
with heavy ions which induce both nu-
clear and atomic excitations over a wide
range of energies, and a new phenomenon,
electron promotion, has been uncovered.
In the case of bombardment of target
atoms by heavy ions, the simple theory
which assumes an interaction of two
point charges during Coulomb excitation
must be modified to account for the pres-
ence of the projectile ion within the K
shell of the target atom. The incoming
ion lingers within the target atom for
relatively long periods of time compared
with the oscillation times of the K elec-
trons. If the electron shells of ion and
target atom overlap, then a "hyperatom"
is formed consisting of the combined nu-
clei and a new K shell. During this for-
mation two of the four electrons which
made up the two originally independent
K shells are forced out, or "promoted."
When the two nuclei later separate there
is a high probability that each will have
vacancies in their K and L shells. In turn,
there is a high probability of ionization
and a consequent high cross section for x-
ray production. In cooperation with Dr.
H. A. Van Rinsvelt of the University of
Florida, studies were undertaken to
bombard copper by beams of alkali ions,
which are readily produced in the DTM
Van de Graaff accelerator and offer a
wide range of atomic numbers. It has
been found that the characteristic x-ray
lines of copper are shifted in energy as a
result of the high probability of ionizing
the L shell when the K shell is ionized.
The energy shifts and the cross section
for ionizing the K shell result from two
effects: Coulomb excitation and electron
promotion.
Studies of brain biochemistry, the
evolution of DNA in rodents, the phys-
ical chemistry of nucleic acid interactions
in vitro, and the immunochemistry of en-
zyme evolution have occupied the atten-
tion of the Biophysics Section.
In recent months there have been reports
which implicate peptides in learning and
memory. N-acetyl aspartic acid (NAA) ,
which is found only in the brain, stimu-
lates peptide synthesis. Consequently,
experiments were set up to test whether
NAA might affect learning and memory
in mice. Performance in the avoidance of
a mild electric shock was improved by
nearly twofold. The improvement per-
sisted for at least a week after training
and NAA withdrawal, indicating that
NAA may well work on the consolidation
of memory rather than as a response to
the continued administration of the drug.
Studies of the evolution of DNA in
closely related mice and other rodents
DEPARTMENT OF TERRESTRIAL MAGNETISM
171
have continued. This is a most curious
area in the bio-sciences, and some of the
potential for understanding evolution at
the molecular level would seem to reside
here. It has been known for some time
that the common laboratory mouse (Mus
musculus musculus) contains a "satellite"
DNA which appears upon centrifugation
in CsCl density gradients and which con-
sists of highly repeated sequences (about
a millionfold) 300 or so nucleotides in
length. Satellite DNAs have been found
in other Mus species (M. caroli and M.
cervicolor) , yet these DNAs appear only
loosely related to those of M. m. musculus,
while the satellite DNA from a Japanese
subspecies (M. musculus molossinus) is
very similar by DNA cross-hybridization
tests even though the Japanese species
contains much less of the satellite.
Phenotypic differences are obvious: The
species differ in size, coloration, and
chromosomal rearrangements and they
interbreed to form fertile hybrids. Thus,
tracking down the functional difference
in the DNA of these species with differ-
ent satellite contents remains a challenge.
Many mechanisms are at play in
establishing genetic diversity even among
bacteria. Ultimately, diversity is trace-
able, according to present understanding,
to the message encoded in the nucleic
acids. Some portions of bacterial or viral
genomes are highly conserved as would
be required for the production of enzymes
and enzyme systems necessary for ac-
commodation to the environment or
changes in the environment and hence to
survival. Enzymes, being proteins and
generally antigenic in foreign species are
subject to immunologic analysis. Mod-
ern organic chemistry has made it possi-
ble to couple appropriate antibodies to
insoluble substrates such as the polysac-
charide sepharose and thus permit anti-
bodies to select enzymic antigens from
crude bacterial extracts with high effi-
ciency. Throughout the Enterobacteri-
aceae the distinctive features of the regu-
lation of the enzyme aspartokinase I-
homoserine dehydrogenase I (AKI-
HSDI) are conserved even though the
compositions of the DNAs vary widely.
It is accordingly possible to investigate
immunologic affinities of the enzyme, the
conformational states of the enzyme as
ligands are introduced into the system,
and to draw inferences concerning the
systematic relationships of specific pro-
teins as contrasted with the similarities
and differences among the total genomes
indicated by nucleic acid interactions
among the bacteria. These studies are
described in detail by Cowie and his col-
laborators, Georges N. Cohen and Paolo
Truffa-Bachi of the Pasteur Institute in
Paris.
In his introduction Dr. L. T. Aldrich,
Associate Director and Chairman of the
Department's Geophysics Section, ap-
propriately describes the appreciation
we have for the opportunity to ex-
press our creativity on a global scale.
His words — substituting only the names
of different sets of fellow scientists —
apply equally to our activities involving
cooperation in research in Astrophysics
and Biophysics and to the generous sup-
port provided to the Department by the
National Science Foundation, the U.S.
Public Health Service, and numerous
agencies throughout the world. This
spirit of cooperation in which we have
most frequently taken the initiative to
stimulate new ventures on new frontiers,
wherever these may lie, has been a hall-
mark of the Department's way of life
since its very beginning. Indeed, the
Department's first proposed name was
"International Magnetic Bureau of the
Carnegie Institution of Washington."
A few paragraphs illustrate the wide
sweep of the work of the Geophysics
Section.
Dr. Sacks and Mr. Hiromu Okada of
Hokkaido, now a predoctoral Fellow at
the Department, have undertaken a com-
parison of the Q structure beneath the
172
CARNEGIE INSTITUTION
South American continent and the Japa-
nese Archipelago. Similar spectral ratio
techniques were used for Q determina-
tions in the two regions. In South Amer-
ica, Qp in the somewhat seismically active
wedge bounded by the seismic plane, the
trench and volcanoes, is 1000-2000. Be-
neath Japan for the corresponding region,
which is aseismic, Qp is 400-500; this
value is identical to that in the astheno-
sphere below the trench, which indicates
there is no evidence for extra low Q
values beneath the volcanoes. In South
America, the high Qp values (up to
3000) extend to 350 km, but there do not
seem to be any paths from deep earth-
quakes (600 km) with Qp greater than
1000 ; therefore, there must be a Q inver-
sion below 350 km. In Japan, however,
the high Qp values in the Benioff zone
(«3000) persist down to the deep earth-
quakes. Above the deep earthquake zone
in Japan and Fiji, very low Qp values
(100) were found, in contrast to a mini-
mum value of 350 in South America.
Ray paths through these very low Q
zones have positive arrival time anoma-
lies.
The tectonic implications of these
structures are that lithospheric plates
have varying thicknesses and continental
plates are several times thicker than
oceanic plates.
In another study relating to massive
events concerning tectonic processes, Dr.
Hart, in cooperation with Dr. J. G.
Schilling of the University of Rhode
Island, has examined basalts from Ice-
land, and the southerly track of the mid-
Atlantic ridge has been analyzed with
respect to trace element composition and
87Sr/86Sr isotope ratios. Tholeiites from
a 500-km section of the Reykjanes ridge
and its extension on Iceland were ana-
lyzed for K, Rb, Cs, Ba, Sr, and 87Sr/86Sr.
Samples most distant from Iceland show
87/86 ratios similar to typical ridge ba-
salt (.7026-.7028) but are more highly
depleted in these L1L elements (K < 400
ppm, Sr < 80 ppm) and show less frac-
tionated element ratios (K/Rb < 900,
K/Ba < 65). Concentrations increase
progressively as Iceland is approached
(factors of 2-5) ; 87/86 ratios increase
stepwise to .7030 about 150 km from Ice-
land. Tholeiites from Iceland show a
considerable range in concentration (K,
700-2000 ppm; Sr, 90-200 ppm) which
overlaps and is somewhat higher than the
nearest ridge samples. 87/86 ratios on
Iceland show little spread (.7030-.7031).
The isotopic data show that the mantle
sources under Iceland are chemically
different from those under the southern
Reykjanes ridge. The concentration
variations between the two provinces are
not well described by mixing models in-
volving only two melt types. A possible
model involves variable partial melting
and mixing of two mantle sources.
Drs. Shimizu and Hart have developed
a sensitive chemical procedure for the
analysis of trace elements in two-phase
systems in which one phase is crystalline
and the other, a quenched liquid that is
a glass. They call their procedure the
"Differential Dissolution Technique," or
"DDT." The procedure takes advan-
tage of the fact that the glass por-
tion is dissolved in diluted hydrofluoric
acid much more readily than the highly
ordered crystalline portion. Thus, in a
series of model experiments, it was found
that in a mixture of clinopyroxene (from
a spinel lherzolite from Salt Lake Crater,
Hawaii) and glass (tholeiitic basalt glass
from the Juan de Fuca Ridge) the glass
dissolved, leaving pure clinopyroxene as
a residue. Milligram amounts of crude
starting material containing only parts
per billion of trace elements can be freed
of contaminating glass and analyzed
accurately.
In these vignettes of the work of the
past year, and especially in the body of
the Report, it is hoped that the reader
will find "new matter emerging and new
details elicited, difficult details becoming
relevant, because of fresh turns that the
arguments have taken."
DEPARTMENT OF TERRESTRIAL MAGNETISM
173
ASTROPHYSICS
G. E. Assousa, L. Brown, E. T. Ecklund, W. K. Ford, Jr., C. K. Kumar, C. A. Little,
U. Rohrer, V. C. Rubin, N. Thonnard, K. C. Turner, M . A. Tuve, and J. W. Warner
Introduction
The preliminary results of an investi-
gation begun two years ago by Rubin and
Ford have brought into question a crucial
assumption under which observations of
distant galaxies are interpreted as meas-
urements of an expanding universe. They
have found an anisotropic grouping over
the sky of the redshifts of galaxies whose
magnitudes and angular sizes lead one to
expect them to lie at about the same dis-
tance. As they point out, this does not
necessarily imply that the expansion is not
isotropic at these distances, but it does
require a more critical interpretation of
redshift observations than has heretofore
been made.
Observations of the water-line signals
from W-49, of the hydrogen line emission
near supernova remnants, and of an ob-
ject whose spatial velocity and structure
suggest its ejection from the galactic cen-
ter illustrate the varied activities in radio
astronomy of the Department's staff at
the Derwood Experimental Laboratory
(Derwood, Maryland) , the National
Radio Astronomy Observatory (Green
Bank, West Virginia) , and the Institute
Argentino de Radioastronomia (Villa
Elisa, Argentina), respectively. Recent
studies of x rays produced with fast,
heavy ions point out the versatility of our
Van de Graaff machine, now 35 years of
age and the second oldest operating Van
de Graaff (atom smasher) in the world.
It has recently been used for optical,
x-ray, and nuclear spectroscopy, thereby
spanning microscopic structures with en-
ergies from less than 1 eV to several MeV.
Improvements of the instrumental ca-
pabilities, some of them major, took place
throughout the section. One can see over
the years a continuous trail of instru-
mental improvements, which have ab-
sorbed a large fraction of the staff's time
but which are indispensable for observa-
tional progress. Particularly noteworthy
this year was the use of nitrogen-baked
Illa-J plates, a simple procedure that
greatly enhances the value of the image
tube— photographic plate combination.
Work was begun to exploit image tube
capabilities at infrared wavelengths, and
major improvements were made at the
Derwood and the IAR-CIW telescopes,
resulting in a much improved system
noise figure. The Van de Graaff machine,
our Atomic Physics Observatory, re-
ceived its third accelerator tube and a
magnetic tape recording system for its
pulse-height analyzer.
Optical Astronomy
W. K. Ford, Jr., C. K. Kumar, Vera C. Rubin,
M. A. Tuve, and J. W. Warner
This report year has seen the comple-
tion of one major program, the study of
internal motions in the Andromeda Gal-
axy, and the early stages of two others:
the investigation of the isotropy of the
Hubble expansion and a program to ob-
tain direct photographs and spectra of
nebulae and galaxies at a wavelength
near one micron. These are described
below.
The isotropy of the Hubble expansion.
Rubin and Ford discussed last year the
start of an observing program to deter-
mine radial velocities of Scl galaxies
(spiral galaxies of high luminosity and
well-defined arms) with 14.0 < m < 15.0.
From an all-sky sample of 208 galaxies
they now have velocities for 75; 50 ve-
locities are in the range 4000 < V < 7500
km/sec. There is a striking nonrandom
distribution of these velocities on the sky,
as shown in Fig. 1. One-third of the sky
contains 28 galaxies, with <V7j> = 4966
± 122 km/sec; one-third contains 22 gal-
axies, with <Tn> = 6431 ± 160 km/sec,
with almost no overlap between the two
velocity groups. The mean magnitudes
174
CARNEGIE INSTITUTION
+90°
O 4000 <V< 5400 km/s (28)
• 5400<V<6100km/s (2
■ 6100 <V<7500 km/s (20)
Fig. 1. Distribution on sky of Scl program galaxies, 14.0 < m < 15.0, with observed velocities
between 4000 km/s and 7500 km/s; equatorial coordinates. Note separations on sky of gal-
axies with different velocities.
and mean diameters of galaxies are simi-
lar on both parts of the sky.
There are several possible explanations
for this curious distribution, which they
are presently investigating while continu-
ing to gather radial velocities of sample
galaxies. These include: (a) Significant
intergalactic or excess galactic absorp-
tion ( — 0.5-1 mag.) over large regions of
space, (b) A large (^1000 km/sec) mo-
tion of our Galaxy and the local group of
galaxies toward the lower velocity gal-
axies. Both in velocity and direction,
such a motion would violate the conclu-
sions from the observed isotropy of the
3°K background relict radiation, (c)
Large-scale clumping of galaxies or dif-
fering luminosity properties of Scl gal-
axies over regions of 100-200 Mpc, such
that the galaxies observed with <Fi> =
4966 km/sec are actually 30 Mpc closer,
as their velocities would imply, but in-
trinsically (0.5 magnitude) fainter than
those observed with < Vn> = 6431 km/sec.
(d) A difference of ^20% in the value of
the Hubble constant in regions I and II.
Such an effect might be due to local gravi-
tational distortions, remnants of the big
bang, or second-order effects in the
Hubble expansion.
In collaboration with Dr. M. S. Roberts
of the National Radio Astronomy Ob-
servatory, Rubin has started a program
to observe neutral hydrogen in the Scl
program galaxies, using the NRAO 300-
foot telescope. They expect to determine
both HI content and velocity, at least for
the nearer objects in the sample.
Internal motions in the Andromeda
Galaxy. Rubin, Ford, and Kumar com-
pleted the analysis of stellar motions near
the nucleus of M31. There is a complex
pattern of stellar motions, generally simi-
lar to the complex motions observed
earlier in the ionized gas. Why the mo-
tions of the old stars and young gas
should be similar is a puzzle, perhaps sug-
gesting that the gas and dust arise from
the material shed by the evolving stars.
Since 1967, Rubin, Ford, and Kumar
have studied the rotation of the outer
regions of M31, gas motions and stellar
motions near the nucleus, and the varia-
tion of chemical abundance or ionization
processes across the nucleus and spiral
disk.1'2-3-4
DEPARTMENT OF TERRESTRIAL MAGNETISM
175
Infrared investigations. Warner is
working on several programs involving
the infrared. After testing and evaluating
14 Carnegie SI image tubes, he built two
image tube systems, one to be used with
the DTM spectrograph and the other as
a direct camera. The spectrographic sys-
tem is being used to obtain strengths of
the [S III] 9069, 9532, He I 10830, and
Paschen lines in external galaxies and
galactic emission regions for chemical
abundance studies.
The direct camera will be used for the
study of extended objects in highly ob-
scured regions and for photography in
the light of infrared emission lines.
In collaboration with R. F. Wing of
Ohio State University, Warner is continu-
ing a program of identification and photo-
electric photometry of sources from the
California Institute of Technology Two-
Micron Sky Survey. To date, 50 identifi-
cations and spectral types have been ob-
tained in the —20° and —30° declina-
tion zones near the galactic center. All
but one are M stars, the exception being
IRC 20385 = Terzian 5, a heavily red-
dened globular cluster at approximately
9 kiloparsecs distance in the general di-
rection of the center of our Galaxy. A
number of supergiants have been found
as well as one object (IRC 30350) with
the properties of VX Sagittarius. The
purpose of this work is to make statistical
studies of the distribution of cool stars in
this region of our Galaxy.
Far ultraviolet variability of quasi-
stellar objects. Quasi-stellar objects
(QSOs) are known to vary in the radio,
optical, and near-ultraviolet spectral re-
gions. Observations have not yet been
made in the far ultraviolet (FUV, Arest
< 912 A) . A knowledge of the variability
in the FUV is important if one is to
properly interpret the spectra of QSOs.
Kumar has interpreted the observed hy-
drogen line intensities in the spectra of
quasars to indicate that these objects
could vary by about 2.5 magnitudes in
the far ultraviolet.
The following simple model is adopted:
A central source emitting the continuous
radiation is surrounded by the line-
emitting region (nebula). The nebula
consists of two parts: one, optically thin
to Lyman continuum radiation, and the
other, optically thick. Observations of
high-redshift objects indicate the pres-
ence of the optically thin part. Measure-
ments of 4C 05.34 (Oke, 1970) 5 and PHL
957 (Lowrance et al., 1972) 6 indicate
that the optical thickness (to) of the
nebula at Arest = 912 A is of the order of
unity. This optically thin part subtends
a solid angle 4tt£ at the central source.
The factor e is found to have the value
1/2 from the observation that nearly one-
half of the 36 high-velocity (Z = AA -f- A
> 1.8) objects show Lya in absorption.
Similarly, observations of low redshift
objects (0.25 < Z < 0.75) show that only
4 out of 70 objects have Mg II in absorp-
tion. Since the Mg II ion is abundant
mainly in the H I zone, the optically
thick part of the nebula subtends a solid
angle of 0.24 it (= 4/70 X 4tt) at the
central source. Consequently, the two-
component model is defined to be ei =
0.5,
T"0i
2 and eo =
TOo
OO
Assume that the nebula is primarily
photoionized, that the spectral index of
ionizing radiation is given by a power law
with spectral index a = 1.5, and that the
line is formed mainly by recombination.
Then the ratio of the number of Ha pho-
tons emitted by the nebula r0 to the num-
ber of Lyman continuum photons emitted
by the central source N(Ha)/N(LyC)
can be computed.
From the measured H/3 fluxes in a
number of QSOs (Oke et al, 1970) 7 and
the predicted Ha/H/3 intensity ratio of
2.84, Kumar has calculated values of the
Ha flux. The "observed" N(Ha)/N(LyC)
ratios are plotted in Fig. 2, along with the
model prediction shown as a solid line.
It is clear that in most cases the observed
values exceed the prediction and the
average observed value is twice the pre-
dicted value. This excess emission line
176
CARNEGIE INSTITUTION
i.o-
7
0.5-
0.4-
£ 0.31-
0.2
0.1
h
o
o
-O— O-
i I I
0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6
Fig. 2. Comparison of observed (circles) and
predicted (solid line) N(Ha)/N(Ly C) ratios of
quasi-stellar objects.
strength is attributed to line formation at
a time when the FUV intensity was higher
than the average intensity.
The time-dependent behavior of line
intensities under the influence of a vari-
able flux of ionizing photons has been dis-
cussed (Bahcall et al, 1972). 8 The work
of Angione (1970)9 shows that on time
scales larger than 10 years the QSOs are
incessantly active and are above their
minimum luminosity at least half the
time. This information and the relevant
formulae given by Bahcall et al. shows
that on the average the amplitude of
variation of QSOs in the far ultraviolet is
about 2.5 magnitudes for this model.
The deduced variability of 250 S is
based on two rather important assump-
tions. They are (1) that the far UV con-
tinuum is given by the spectral index 1.5,
and (2) that interstellar extinction is
negligible. These are the assumptions
generally made by investigators attempt-
ing to explain line intensities in 250 S. It
is proposed to continue this investigation
by removing the above assumptions and
trying to explain the hydrogen and
helium line intensities in 250 S.
Anticenter blue stars. A search for blue
stars near the galactic plane at large dis-
tances from the sun in the galactic anti-
center direction is under way by Tuve
and Rubin using three-color plates taken
with the 48-inch Palomar schmidt in
1965. This is a part of our interest in
determining the rotation of the outer
parts of the galaxy which can only
broadly be determined by neutral hydro-
gen observations.
Image tube development. The cas-
caded image tube that has been used so
successfully in the Carnegie image tube
systems was designed ten years ago, early
in the 1960's. In 1969 the same basic type
of tube was built at RCA with more
modern fabrication techniques, including
the use of a ceramic rather than a glass
envelope. These tubes were operable at
higher voltages and hence had higher
brightness gains than the older, low-
voltage tubes. The general performance,
however, was inferior to the older tubes
because of low-quality phosphor screens
and, in some tubes, excessive ion emis-
sion. Consequently, with financial sup-
port from the National Science Founda-
tion, we instigated a two-year develop-
ment program at RCA on the RCA
C33063 image tube. This development
program has now resulted in an improved
version of the Carnegie cascaded tube.
The work at RCA has been directed
toward (1) reducing the ion scintillations
and other sources of spurious background,
(2) reducing the geometrical distortions
in the tube by redesigning the electron
optics, (3) improving the quality and
uniformity of the phosphor screens, and
(4) improving the contrast in the in-
tensified image by reducing the amount
of internally scattered light in the tube.
The higher gain of these tubes makes
the ion scintillations a particular prob-
lem. Unless the system is operated with
DEPARTMENT OF TERRESTRIAL MAGNETISM
177
the transfer lens stopped down, each ion
scintillation is recorded on the photo-
graphic emulsion as a discrete image.
After a long exposure, the plate appears
to be covered with these random faint
star images as though the plate had been
sprinkled with black pepper. To reduce
the number of the scintillations, the RCA
engineers have relocated the alkali chan-
nel electrode used in cathode formation to
a more nearly field-free region. The
chromoxide coating technique used to
provide an even voltage distribution in
the tube was improved. A better vacuum
technique was introduced to reduce the
chance of high-energy collisions of photo-
electrons with free gas molecules. Better
assembly and clean-room procedures were
also introduced. These steps have not
eliminated the ion scintillations but have
greatly reduced their number per unit
time.
The geometrical distortions in a mag-
netically focused tube depend on the uni-
formity of the externally generated mag-
netic field and the electric field generated
internally by accelerating electrodes. To
improve the electronic imaging, the
flanges on the tube bulb, or envelope,
and the accelerator spacers were rede-
signed, resulting in a reduction of image
rotation and spiral and linear distortions.
There are three accelerating electrodes
per stage in the C33011 and four acceler-
ating electrodes per section in the C33063.
The new C33063 is 6.375 inches in overall
length, or 1.5 inches longer than the old
Carnegie tube. Hence a longer, lower
field-strength focusing magnet and a
modified voltage divider are required to
operate the new tube.
The performance of the cascaded image
intensifiers has been limited by the pres-
ence of mottling and fine structure in the
phosphor screens. Attempts to use the
tube to detect weak spectral lines or
faint objects in a field photograph were
often frustrated by the uncertainty intro-
duced by mottling in the screens. In the
course of the improvement program at
RCA, considerable attention was directed
to identifying and controlling the sources
of structure in the screens. The average
particle size of the phosphor material
is many times smaller than the average
size of mottle pattern. Many test screens
were prepared utilizing a variety of depo-
sition, drying, filming, and aluminizing
techniques. Microphotographs were taken
of the screens at each step of processing.
It was found that the organic filming
used to isolate the phosphor from the
aluminum back coating played a much
more critical role in determining screen
uniformity than was previously supposed.
Various procedures have now been intro-
duced to control the filming and, conse-
quently, much-improved screens are in-
corporated in our recent intensifiers.
An attempt has been made to reduce
the amount of light internally reflected
and scattered about by bright surfaces
within the tube. Also, we are experiment-
ing with thin cover slides having an effi-
cient antireflection coating that can be
oiled to the tube faceplate to reduce re-
flections at the air-glass surface. It is
hoped that these techniques will improve
the contrast in the image tube system.
One of these improved RCA C33063
tubes has been mounted with a Bowen
f/1.4 Cassegrain schmidt camera for use
on the Kitt Peak 84-inch Cassegrain
spectrograph. This system is proving to
be a substantial improvement over the
original Carnegie system.
The development of a digitized spec-
trum scanner based on the cascaded tube
has continued. The approach we are in-
vestigating uses the persistence of the
phosphor screens in the tube to store
single photoelectron events until they can
be detected with an image dissector. The
phosphor screen is, of course, an inefficient
storage element, as each scintillation is
decaying exponentially. The problem,
then, is to determine the optimum scan-
ning rate to minimize the probability of
missing photoelectron events. Much of
178
CARNEGIE INSTITUTION
the effort has been to develop fast linear-
sweep amplifiers to drive the deflection
yoke of the image dissector. This has
been done by Mr. Scuderi and Mr. Doak.
Experiments designed to measure the
counting efficiency of the system are now
in progress.
Figure 3 illustrates the impressive in-
crease in information content of image
tube spectra of the night sky due to in-
creasing excellence of the optics and
photographic plates. The first spectrum
(1965) is a 60-minute exposure using
baked IIa-0 plates and taken with the
69-inch Perkins telescope of Ohio State
and Ohio Wesleyan Observatory, using a
Super Faron spectrograph camera and an
RCA C33011 image tube. The middle
spectrum (1967) illustrates the increased
resolution arising from the use of a
Bowen f/2.25 schmidt camera; the spec-
trum is a 180-minute exposure on a baked
IIa-0 plate, taken with the 84-inch Kitt
Peak telescope. The third spectrum
(1973) shows the highest resolution; it
was taken with an f/1.4 schmidt camera,
an RCA 33063 image tube, and used a
nitrogen-baked Illa-J plate at an expo-
sure time of 150 minutes on the 84-inch
telescope.
We thank Drs. Babcock, Blanco, Gold-
berg, Hall, and Heeschen for making
telescope time available at CARSO,
Cerro Tololo Inter-American Observa-
tory, Kitt Peak National Observatory,
Lowell Observatory, and National Radio
Astronomy Observatory.
We also wish to thank Drs. Goldberg
and Blanco for assistance in travel ex-
penses of observing trips to Kitt Peak
and Cerro Tololo.
The development of the high-gain
image tube and the work on this digital
system have been supported in large part
by a grant from the National Science
Foundation.
Fig. 3. Increasing resolution and gain of image-tube night sky spectra: (a) RCA C33011 cas-
caded image tube and f/0.87 Super Faron spectrograph camera. Exposure 60 min on baked IIa-0
plates, 69-inch Perkins telescope of Ohio Wesleyan and Ohio State Universities at Lowell Observ-
atory; August 3, 1965. (b) RCA C33011 cascaded image tube and f/2.25 Bowen schmidt camera.
Exposure 180 min on baked IIa-0 plates, 84-inch Kitt Peak telescope ; March 8, 1967. The in-
creased resolution from spectrum a is due principally to improved spectrograph optics, (c) RCA
C33063 cascaded image tube and f/1.4 Bowen schmidt camera. Exposure 150 min on nitrogen-
baked Illa-J plates, 84-inch Kitt Peak telescope; March 5, 1973. Increased gain from spectrum b
is due to higher gain of type C33063 tube. Increased resolution is due principally to use of slower,
finer grained Illa-J plates.
DEPARTMENT OF TERRESTRIAL MAGNETISM
179
Radio Astronomy
G. E. Assousa, E. T. Ecklund, C. A. Little,
N. Thonrmrd, K. C. Turner, and M. A. Tuve
Equipment Development
E. T. Ecklund, C. A. Little, and N. Thonnard
The old gray trailer, which for many
years had housed our frequency analyz-
ing spectrometer, has finally been decom-
missioned. This was an occasion for us,
as the trailer had served well in the early
1960's to make our receiver portable
when NRAO had no spectrometers. How-
ever, the increased stability and flexibil-
ity gained by having all of the electronics
in a temperature-controlled environment
has more than compensated for this loss.
The move out of the trailer allowed us to
make many modifications to enhance the
stability and versatility of the multi-
channel spectrometer. Figure 4 is a block
diagram of the present spectrometer.
The solid-state broad-band amplifier
described in Year Book 71, p. 224, was
completed this year. The new system has
27 filter drivers arranged in three sets.
Each driver is capable of driving at least
six filters. Therefore, the two present sets
of filters have independent drive systems,
eliminating any interaction between filter
sets and allowing the filter outputs to be
set independently. The remaining set of
drivers will soon be used in conjunction
with the very narrow crystal filters now
under construction.
The wide-band filters discussed in
Year Book 71 are now complete. To fully
utilize the versatility of a multiple filter
system" without transferring the cables
and resetting all the gains, two special
three-position, 60-pole switches were con-
structed. One switch connects the R. F.
output from the filter system in use to the
R. F. detectors; the second switch selects
one of three sets of 60 presettable gain
30 MHz IF
30 MHz
&\ amp
From control room (BW = IO)
^ Mixer
[25 MHzh>,
I
I
J28MHz|~Q
VI
Video amp
(BW = 12)
System A
'«»
System B
Filter drivers
(9 in each system)
System C
1
f
Pre-drivers
System C
Crystal oscillators
(total of 6)
System B
System A
C9)
Extra
Extra
Audio
amplifiers
and phase
dete
(total of 56)
Audio
preamps
and gain
M> «o6\- equalization *-(56)— detectors ^6V
C,0r ' - V^1(total of 56)V
switch
(3 position,
60 poles'
RF
RF
selector
switch
[ 3 position,
60 poles)
Wide band
filters
total of 54 K
spaced 75
KHz)
RC
integrator
+
stopping
switch
Narrow
band filters
(total of 56
Spaced 19 kHz)
> Electrometer
Voltage to
frequency
converter
.to chart recorder
to frequency counter
and card punch
NOTICE-. Encircled numbers indicate number of separate lines in this circuit
Fig. 4. Block diagram of present frequency analyzing spectrometer at Derwood.
180
CARNEGIE INSTITUTION
controls, which are in the feedback loop
of 60 IC preamplifiers. The presettable
gain controls were initially adjusted so
that the detected output level from each
filter (both narrow-band and wide-band)
was the same. This enables one to go
from narrow-band to wide-band observa-
tions by simply rotating two switches.
Two additional advantages were obtained
by using the IC preamplifiers. First, with
the increased audio gain it was possible
to reduce the R. F. level in the detectors
and optimize it for the best power-law
response. Second, by having higher levels
in the phase detectors and on the inte-
grating capacitors, the electrometer gain
was reduced by a factor of 10, which has
completely eliminated drifts due to capa-
citor leakage and diode balance changes.
This system has been in use for more
than four months. Its stability has been
much greater than our old tube system:
Since the initial gain adjustment, no ad-
ditional adjustments have been neces-
sary; the old system required adjustment
every few weeks.
A new local oscillator system with bet-
ter power and frequency stability and
much wider frequency coverage has been
completed for the hydrogen-line receiver.
The system uses a solid-state phase-
locked source that accepts a reference
signal of about 106 MHz and delivers ap-
proximately 220 mW at 1390 MHz. The
reference source uses an oven-stabilized
crystal oscillator and a temperature-
compensated variable frequency oscilla-
tor so that the local oscillator can be
varied by ±10 MHz.
Owing to the much wider bandwidth
of the new filters, a new gain modulator
has been designed and constructed. It
uses recently introduced hybrid inte-
grated circuit modules having 500-MHz
bandwidths. We are using a voltage-
controlled attenuator module and a 15-db
amplifier module, together with an FET
switch and IC amplifier, to synchronously
modulate the system gain by as much as
6 db. The linearity is better than we can
measure over a frequency range wider
than we anticipate using. The gain mod-
ulator system has also been useful as a
very flat, synchronously varying signal
source to adjust the individual filter gains
in the spectrometer.
Last August we finally had to return
the K-band mixer-preamplifier loaned to
us by NRAO for our 1.2 to 1.5 cm receiver
front-end. We now have our own mixer
and IF amplifier, which we placed in
operation in February 1973. Even though
the IF amplifier noise figure is 2.0 db,
the overall mixer-IF noise figure was 5.0
db at 22 GHz. The overall single side-
band system noise temperature during
active observations has been about
1400 °K, a tremendous improvement over
the 2400 °K system temperature we were
obtaining with the NRAO system. The
IF bandwidth is 500 MHz, which will
make this system also very useful for
continuum observations.
1.35-cm Water-Line Observations
N . Thonnard
Owing to the major equipment modi-
fications and additions to our line spec-
trometer and the installation of our new
K-band mixer-IF system, observations of
H20-line emitters were not resumed until
this February. It was decided to under-
take an intense study of W-49, by far
the brightest and most variable water-
line emitter. So far, ten continuous weeks
of observations have been made on this
source. Most observations were made
twice a week, and for one week the source
was observed daily. All observations
were first taken with the wide-band fil-
ters and covered a velocity range of at
least 300 km/sec; then the narrow-band
filters were used to study the fine struc-
ture of the central emission features of
W-49. Some of the features are so nar-
row that they should be studied with even
finer resolution than our present 0.25 km/
DEPARTMENT OF TERRESTRIAL MAGNETISM
181
Ld
LT
D
h-
<
LU
Q_
LJ
<
360-
320
280 -
240-
200-
160-
120
80
40
I I I I I I I I
I I I I I I I I
I I I I I I
i i i r
/!
HIGH RESOLUTION PROFILES
OF CENTRAL VELOCITY
PEAKS;
= 2.0°K
i I , i , i i i i i i i i l i i i i i i i l i i i i i i i l i i i l i i 1 1 i
-4 0 4 8 12 16 20 24 28
RADIAL VELOCITY (KM/SEC)
FEBRUARY 23, 1973
H20 LINE EMISSION FROM
W-49 oC -- I9h 07m50s 8= 9°l.r095O)
W°-8 °K
*&m,**S^—m+<m*IH——m**w**<m»»** >' » ■*■
MARCH 23, 1973
.■^'^/^■■^■llWfl n>i
J_l I I I I L
-220 -200 -180 -160 -140 -120 -100 -80 -60 -40 -20 0 20 40 60
RADIAL VELOCITY (KM/SEC)
80
i l l l I L
100 120 140 160 180
Fig. 5. H20 line emission from W-49 on February 23, 1973, and March 23, 1973.
sec. In a few instances, the velocity cov-
erage was from — 450 km/sec to +400
km/sec.
A typical illustration of water-vapor
emission from W-49 on two occasions is
given in Fig. 5. These data were taken
with the wide-band filter. An expanded
view of the velocity spectrum taken with
the narrow-band filter is shown for a
small section to illustrate the complexity
of the emission features. Analysis of
these data and further observations are
now in progress. It is hoped that by care-
ful monitoring of these features it will be
possible to understand the nature of the
large variation and, in particular, the
mechanisms giving rise to the large ve-
locity spread in the observed emission
features.
Search for 21 -cm Hydrogen Absorption
during Flare-ups of Cygnus X-3
G. E. Assousa, N. Thonnard, and M. A. Tuve
A sudden rise of three orders of mag-
nitude in the intensity of the radio source
associated with the x-ray source Cygnus
X-3 was first noted by Gregory et al.10 on
September 2, 1972, at the 46-m radio tele-
scope of the Algonquin Radio Observa-
tory. On August 31, 1972, measurements
taken at NRAO indicated intensities of
0.015 f.u. at 3.7 cm and 0.016 f.u. at 11.1
cm (1 f.u. = 10"26 Wm"2 Hz"1). The in-
tensity on September 2-3 rose sharply to
18.4 f.u. and 9.4 f.u., respectively. B.
Balick11 of NRAO notified us of this
event and asked us to look for 21-cm
hydrogen-line absorption in the direction
of the source.
182
CARNEGIE INSTITUTION
Unfortunately, our Derwood receiver
was undergoing major modification when
the Cygnus X-3 flare-up occurred. After
very hurried and makeshift reconnect-
ing of the receiver, observations were
started on the evening of September 5,
1972. By that time, the intensities ob-
served by NRAO had dropped to 5.0 f.u.
at 3.7 cm and 7.3 f.u. at 11 cm. Observa-
tions were also in progress during the
second flare-up of Cygnus X-3 on Septem-
ber 20. No absorption could be detected
from our data. French observers at the
Observatoire de Paris, Meudon (Lauque
et al., 1972) 12 were successful in detect-
ing absorption of the source by the inter-
vening neutral hydrogen at 21 cm and so
were able to place Cygnus X-3 in our
Galaxy at a distance between 8 and 11
kpc.
Our failure in observing absorption in
the direction of the source is not too sur-
prising. At 21 cm our Derwood dish has
a beamwidth of 50' and hence a sensi-
tivity of only 0.05°K per f.u. This would
give an absorption of at most 0.5°K.
Two methods were used to search for
absorption. First, data were taken on the
source, and at four points one-half beam-
width and four points one beamwidth
from the source. An attempt was made to
construct a background hydrogen emis-
sion for the region by interpolating these
data, but, owing to the large beamwidth
and the complexity of the Cygnus region,
no absorption could be detected. An at-
tempt was also made to compare data
taken at the Cygnus X-3 position on days
when the source was intense with data
taken when the source was weak. This
required reproducibility of the receiver
gain of better than 0.5%. Our typical
results of 3-5% rendered this method
ineffective.
Survey of Neutral Hydrogen in the
Vicinity of 17 Galactic Supernova
Remnants
G. E. Assousa, B. Balick, and J. W. Erkes
Recent 21 -cm observations in the vicin-
ity of the supernova remnant HB21 (As-
sousa and Erkes, Year Book 71, p. 240)
showed evidence of an expanding con-
centric shell. At the best distance esti-
mate for HB21, the accreted shell of
interstellar material has a calculated
radius of 26 pc and a total mass of 6300
M0 . Based on the HB21 results, an ex-
tensive survey of the neutral hydrogen
near 17 large galactic supernova rem-
nants was completed at the NRAO.*
This included CTA1, HB3, CTB13, HB9,
VRO 42 05 01, OA 184, S147, P 0607+17,
W41, 3C3961, W50, CTB 72, CTB 63,
W56, Cygnus Loop, W63, and CTB1.
Observations were made at beamwidth
intervals (0.33°) over and in the imme-
diate vicinity of each remnant, and at
0.5° or 0.66° intervals elsewhere. Pre-
liminary longitude versus velocity maps
and latitude versus velocity maps for all
17 remnants have been completed.
There is evidence for the presence of
anomalous velocity neutral hydrogen
near the Cygnus Loop (Fig. 6) and S147
(Fig. 7). The contour levels in both fig-
ures are those of antenna temperature in
°K. The center of the Cygnus Loop is at
lu = 74.0° and b„ = —8.6°. In Fig. 6
there is a long feature between — 12°
< frn < —7° which may well be the
remnant-associated hydrogen. Bright
optical filaments in the Loop are known
to have velocities of up to 115 km/s
(Minkowski, 1958) ,13 Here, the veloc-
ities observed extend to roughly the same
value. In S147 (Fig. 7), a velocity ex-
tension to — 53 km/s appears at the op-
tical center, lu = 179.6°, 6„ = —2°; no
filament velocities are available. In both
places there is a strong hint that the neu-
tral hydrogen is associated with the rem-
nant.
Integrated hydrogen maps over fixed
velocity intervals give the best evidence
for any shell structure (Assousa and
Erkes, Year Book 71, p. 240) . Such maps
* The National Radio Astronomy Observa-
tory is operated by the Associated Universities,
Inc., under contract with the National Science
Foundation.
DEPARTMENT OF TERRESTRIAL MAGNETISM
183
o
o
CO
ID
z
o
>-
u
CO
c3
p
p
p
CD
. (-■ ro
^ o
. t»
^-v CO
CO °3
I o3
,, P
►©
CD
CO X!
CO =3
,, o
w o
O xn
O --1
J 8
in P
Q o
0) i— I
-P I
CD W)
-JS p
o P
CD
CD +s
M X
aaniutn
a
CD h
P -M
p CD
O «+H
'-M CD
S-i
22 P
P 13
oj ^
2 »
,P bC
p— i ^
d T3
03 ao
P CD
-P d
H >
fab B
w s-.
CD
ft
^Ik
CD
CO
CD
P
O
o
CD
03
CO
bJD
P
'-3
p
CD
+3
X
CD
CD
-P
GO
P
CD
O
CD
-P
33
CD
P
P
o
+3
p
-a
p
CD
bfi
o
S-I
-p
p
CD
P
CD
-P
3aniu«i
184
are now being prepared for the survey.
Detailed analysis of the Cygnus Loop,
S147, and the other remnants must thus
be deferred until all integrated hydrogen
maps are available.
Accurate Radio Positions for Ohio Radio
Sources, Using the NRAO Three-Element
Interferometer at 3.7 cm and 11.1 cm
J. W. Warner, G. E. Assousa, and B. Balick
Recent studies for the identification of
the optical counterparts of radio sources
from the 1415-MHz Ohio survey have
led to the discovery of unusual astro-
nomical objects. In several instances
identifications have been made with op-
tically distinctive sources which show
high redshifts of variable luminosity.
The radio source OH471, recently ob-
served by Carswell and Strittmatter
CARNEGIE INSTITUTION
(1973) 14 at Steward Observatory to have
the largest known redshift, Z = 3.4
(Zeb
1 + V/C
[1_ (y/c)2]i/2
•1)
is a result of such comparative studies.
Statistical studies of positions of Ohio
sources, in conjunction with lists of com-
pact galaxies and "blue" stellar objects,
have shown a significant number of these
objects to have radio counterparts. Radio
source positions, accurate to several sec-
onds of arc, are essential for positive
identifications. We have used the
NRAO three-element radio interferom-
eter at 3.7 cm and 11.1 cm to obtain
accurate positions and fluxes at both
wavelengths. Sources we have already
observed appear in Table 1. Positive
identification of optical counterparts is in
progress.
TABLE 1. List of Sources Observed with the NRAO Three-Element Interferometer
OB-001
OB-199
OC-077
OD-378
OK492
OX130
OY-148
OB-100
OC-211
OC083
OD392
ON119
OX029
OY051
OB-015
OC415.5
OC+085
OE-113
OP-303
OX-041
OZ-020
OB-016
OC528
OC-286
OE-115
OP-342
OX-055
OZ022
OB-026
OC036
OC287
OE419
OP-053
OX-260
OZ-029.1
OB-078
OC-147
OC390
OE-025
OQ532
OX-179
OZ533
OB-081
OC-051
OC-294
OE-244
OQ-139
OY110
OZ036
OB-281
OC364
OD315
OE-256
OR450
OY120
OZ441.5
OB-089
OC-069
OD-049
OG086
OT546
OY033
OZ-256
OB488
OC268
OD372
OH404.1
OX018
OY-038
OZ064
OZ090
DEPARTMENT OF TERRESTRIAL MAGNETISM
185
Instituto Argentino de Radioastronomia
Activities
E. Arnal, R. Colomb, R. Dugatkin, E. Filloy,
S. Garzoli, M. Gil, D. Goniadzki, M. Gordon,
I. Mirabel, R. Morras, W. Poppel,
R. Quiroga, and K. C. Turner
This organization, founded to utilize
the 100-foot radio-telescope and receivers
provided by the Carnegie Institution,
continued during the year under the
temporary direction of Dr. K. C. Turner
of our Department. As in the past, the
major emphasis has remained on studies
of galactic hydrogen distribution.
Equipment development. The new sur-
face for our existing 30-meter-diameter
telescope was completed in September.
It consists of a solid steel surface from
the center to a radius of 10 meters and a
perforated steel surface on the remainder.
The antenna efficiency at 21 cm is now
essentially equal to that obtained before
the hailstorm of last year.
The new parametric amplifier front
end, utilizing one of the four paramps
kindly donated by S. Colgate of the New
Mexico Institute of Mining and Tech-
nology, has been operated for several
5 Km/sec
-5 Km/sec
• • •
$- 35 Km /sec
25 Km/sec
20 2
Scale: 1 =10 at/cm
Fig. 8. Integrated brightness temperature for Vr > 30 km /sec. The contour units are 1020
atoms/cm2. The arrow indicates the direction of the galactic center. Small circles represent the
points observed. The dashed lines are lines of constant radial velocity, corrected for the effect of
galactic rotation.
186
CARNEGIE INSTITUTION
months and is proving much more stable
and reliable than our old system. The
measured system noise temperature is
now 190°K.
The mount and carriage for the second
telescope are now finished, and the parab-
oloid is ready to be mounted. This will
be done by three large cranes as soon as
the ground is dry enough to support their
weight; it should be completed by the
time this report is published.
The bank of thirty 100-kHz-wide fil-
ters has been modified to eliminate a
serious problem of spurious responses
well away from the band centers.
21 -cm neutral hydrogen observation.
K. C. Turner and I. F. Mirabel have been
studying an object with peculiar prop-
erties illustrated in Fig. 8. Its elongation
suggests the possibility that it was
ejected from the galactic center, whose
direction is indicated by the solid arrow.
However, the intrinsic velocity field of
the object (indicated by the dashed lines)
is that which would be expected if the
object were moving parallel to the galac-
tic plane in the direction toward the cen-
ter. Analysis of this material is not yet
complete.
W. Poppel, working with E. R. Vieira
of the Institute de Fisica of the Universi-
dade Federal do Rio Grande do Sul, has
completed the data reduction of their
very large survey of the region 248° < I <
12°, 3° < b < 17°. It is hoped that pub-
lication of this material will begin shortly.
These investigators have also com-
pleted an extremely accurate study of a
set of calibration points for H I observa-
tions, and have observed, as well, the
neutral hydrogen distribution near NGC
2287, NGC 6231, and 6 Carinae.
S. Garzoli and I. F. Mirabel are cur-
rently studying H I structures in the outer
parts of the galaxy between 288° < I <
310°, —7° < b < 2°.
I. F. Mirabel and W. Poppel are in-
vestigating an anomalous velocity dis-
covered by Cugnon near I = 349°, 6 = 3°.
F. R. Colomb and D. Goniadzki are
investigating the neutral hydrogen dis-
tribution near the supernova remnants
G 261.9 + 5.5, G 296.3 + 10, G 327.6 +
14.5, and G 330 + 15.
F. R. Colomb, M. Gil, and R. Morras
are observing an extensive region near
the south celestial pole.
R. Morras is beginning a program to
study small-scale angular structure of
neutral hydrogen in the galaxy using the
moon as an occulting screen.
E. Arnal has begun a study of the dis-
tribution of neutral hydrogen near very
early type stars.
F. Strauss and Z. Strauss of the Insti-
tute de Fisica of the Universidade do Rio
Grande do Sul visited to search for possi-
ble "fossil" supernova remnants near
pulsars.
S. Garzoli has just published her Atlas
of Galactic Neutral Hydrogen for the
Region 270° < I < 310°, —7° < b < 2°
as a Carnegie Institution of Washington
monograph (Publication No. 629) .
Atomic Physics
L. Brown, G. E. Assousa, and
H. A. Van Rinsvelt
In Year Book 53, p. 67, Norman Hey-
denburg and Georges Temmer described
their pioneer work in studying the low-
lying excited states of nuclei through
Coulomb excitation with a-particles.
These investigations included measure-
ments of the excitation of the K x-ray
lines of the target atoms as well as the
y-rays from its nucleus, since their scin-
tillation counters recorded both kinds of
photons for the heavy elements. These
data are the first in a research field that
has grown very active in the past few
years, the study of atoms by bombard-
ment with heavy ions. The theory of
Coulomb excitation has proved capable
of describing not only a wide range of
nuclear excitations but atomic excitations
as well. However, a new phenomenon,
called electron promotion, is known to
exist; its theory is as yet only qualitative.
Coulomb excitation assumes a simple
interaction of two point charges. It can
DEPARTMENT OF TERRESTRIAL MAGNETISM 187
be calculated in a straightforward man- work of others who had observed shifts
ner with quantum mechanics, the theory in photon energies when targets were
being applicable to atoms or nuclei. For bombarded with particles of tens of MeV;
the atoms ionized by heavy ion bombard- it thus seemed reasonable to follow this
ment, the original theory must be modi- to the lowest energies capable of exciting
fied to account for the presence of the the lines. The shifts result from a high
heavy ion within the K shell of the target probability of ionizing the L shell when
atom. Typically, it remains there for the K shell is ionized. Both the energy
periods that are long compared with the shifts and the cross sections for ionizing
oscillation times of the K electrons. With the K shell from the experiment can be
this binding energy correction, Coulomb divided into two groups: effects resulting
excitation accounts accurately for a wide from Coulomb excitation, and effects
range of excitation phenomena, provided from electron promotion,
that the incident particle has no electrons Figure 9A shows the spectrum of the K
of its own or, if it does, that they do not lines of copper excited by bombardment
overlap the region occupied by the K with 1.75 MeV Rb+ ions. It was observed
shell of the target. If the electron shells with a lithium drifted silicon detector,
do overlap, the electrons try to accom- which is a solid-state device made of a
modate themselves to the structure of a single crystal of silicon with lithium dif-
new hyperatom, having as its nucleus the fused through certain portions. It func-
combined nuclei of projectile and target, tions as would an ionization counter of
Thus a new K shell forms, and two of the high density and low energy loss per ion
four electrons, which made up the two pair. The open circles are from the rubi-
independent K shells, are forced out or dium beam; the smooth curves are from
"promoted." When the two nuclei sepa- copper fluorescence radiation, multiplied
rate later, there is a high probability that by 0.1, which was used to locate the un-
each will have a K-shell vacancy. The shifted lines. The intensity is plotted as
same kind of arguments can be applied to the ordinate, the photon energy as the
the L shells. Electron promotion gives a abscissa. The Ka line, shown at the left,
very high probability of ionization, hence is made up of two unresolved lines of
a very high cross section for x-ray pro- energies 8027.8 and 8047.8 eV, and the
duction. K/3 line, shown at the right, has an energy
During the past summer we returned of 8902.9 eV. The resolution of the de-
to the study of the excitation of x rays tector is 220 eV. The shifts in energies
by heavy ion bombardment. We used the of the lines excited by rubidium are
beams of alkali ions, used previously in apparent.
the study of the lifetimes of optical The energy shifts proved to be almost
states, because they offered a wide range independent of particle energy, as shown
of atomic numbers. Our first bombard- by the results plotted in Fig. 9B. Lithium,
ments were the K lines of copper. the incident ion that had electrons with
The experiment proposed to answer little overlap of the L shell of copper,
two questions: (1) Are the characteristic produced almost no energy shifts. The
lines of copper shifted in energy when other alkalis, which have overlap of L
excited by the heavy ion beams of a few shells, had shifts that were close to what
MeV, and, if so, how do the shifts vary one expects theoretically from a single
with the kind of ion and the energy? (2) vacancy produced in the L shell. This
How do the cross sections for the lines, high probability easily follows from the
as observed by us through the yields, de- idea of electron promotion,
pend on the kind of ion and the energy? The yields of the K x rays can be cal-
The first question was suggested by the culated for Coulomb excitation, so
188
CARNEGIE INSTITUTION
2000K
50
Channel no
50
Fig. 9 A. Spectrum of the K lines of copper observed with a lithium drifted silicon detector.
The open circles are from bombardment with a Rb+ beam of 1.75 MeV; the smooth curves are
from copper fluorescence radiation, multiplied by 0.1, which was used to locate the unshifted posi-
tion of the centroids.
DEPARTMENT OF TERRESTRIAL MAGNETISM
189
50
0
OOh
>
Z 0
jC
CO
100
0
100-
0
_ Li
1
o
o
o o
N1
$..0
„• o o o a
1
1
1
—
" Na
1
i
■ 1
♦
*
■0
■ a
4
»M
*
i
> ♦
*
" K
7
j-j-M
-1 i
*
i,
$
-a
} I M
*
V
~ Rb
1-
-I
} M ♦
♦ <
f v
*
0
•a
I^S
I
> 6
1
*
1
♦ ♦ ♦ '
i
0
2 E (MeV) 4
Fig. 9B. Shifts in the energies of the copper Ka and Kg lines in eV as functions of incident ion
energy in MeV. Data on nitrogen and cesium were taken at only one energy and are plotted with
the results for lithium and rubidium, respectively. The horizontal lines shown for sodium, potas-
sium, and rubidium are calculated values of the shifts produced by a single hole in the 2p shell of
copper.
190
CARNEGIE INSTITUTION
Fig. 9C. Thick target yields of the K lines of
copper as a function of atomic number of the
incident ion at 1.6 MeV. The smooth curve is
the prediction of Coulomb excitation with bind-
ing energy correction and shows the enormous
increase for the ions heavier than sodium.
comparison with experiment is possible.
Figure 9C shows the measured values for
the thick target yields for incident par-
ticles of 1.6 MeV plotted as a function of
the atomic number of the incident ion.
At the left is shown a smooth curve that
connects values calculated for Coulomb
excitation with a binding energy correc-
tion. The agreement is good for lithium
and sodium but misses by many orders
of magnitude for potassium, rubidium,
and cesium. The high probability of ion-
ization is evident if the K shells of the
colliding atoms overlap.
Nuclear Physics
L. Brown and U. Rohrer
Improvements in the Atomic
Physics Observatory
The Van de Graaff accelerator — in
spite of its 35 years, which make it the
second oldest operating atom smasher in
the world — continues to be a most useful
instrument for observing nuclear and
atomic phenomena. The heroic size of its
high-voltage terminal allows the use of
the polarized ion source of the University
of Basel, which gives the machine capa-
bilities still unmatched. During the past
year, two changes increased the efficiency
of this instrument: the accelerator tube
was replaced, and the pulse height ana-
lyzer was connected to a magnetic tape
output.
During the past few years the vacuum
tube of the Van de Graaff has cost users
increasing amounts of time for locating
and sealing leaks. During the last two
years alone 150 detectable leaks occurred.
When a decision was made a year ago to
replace the tube, a new design, which
had the same electrode geometry as the
previous tube Norman Heydenburg built
and several elements common to the tube
Merle Tuve built for the 1934 machine,
evolved from studies with Joseph Peoples
of Potentials, Inc., Austin, Texas, who
has considerable experience in bonding
accelerator tubes. The new tube is made
up of seven identical sections, each made
of cylinders of Pyrex bonded to the alloy
Kovar with vinyl acetate. The machine
is subjected to large changes of tempera-
ture, which caused leaks in the old tube.
The near identity of the thermal expan-
sivities of Pyrex and Kovar should pre-
vent such leaks. The aluminum elec-
trodes on the interior of the tube, which
shape the electrostatic field for accelerat-
ing the ions, are connected to the high-
voltage machine through the Kovar. The
new tube is not strong enough to be self
supporting, as were its predecessors, the
tubes of 1938 and 1955, so it is connected
to the main structure at eight points
through an assembly designed to isolate
it from vibration and shock. Selwyn
Sacks aided in this part of the design by
measuring the amount of the vibration
present with a seismometer.
The new tube first operated in Decem-
ber 1972 and has surpassed the design
goals. No leaks have opened after four
DEPARTMENT OF TERRESTRIAL MAGNETISM
191
months of operation during a period when
changes of temperature were severe. Fur-
thermore, the new tube delivers a much
steadier and more intense beam than the
old one. This was not expected but is
very welcome. It is evidently the result
of better electrical connection between
the electrodes in the vacuum and the high
voltage machine.
In most experiments with the APO the
data are acquired in a pulse-height ana-
lyzer. At the end of an experimental run,
they are read out of the memory of the
analyzer for reduction and interpretation.
In the past this has been done with a line
printer, and therefore part of the data
reduction had to follow by hand. This
was reasonable for the simple spectra of
the charged particle reactions observed in
the past. It proved onerous in the x-ray
work reported elsewhere in this report
and would have been increasingly so in
the neutron work just started. An oppor-
tune purchase by the Department of five
digital tape recorders at bargain prices
took place in happy coincidence with
these new requirements. One of these
new recorders was interfaced to the pulse-
height analyzer, so all data reduction now
takes place with the computer. For the
experiments under way, this removes
much tedious labor and improves the
reliability of the extracted data by allow-
ing many mathematical operations to test
the measurements that otherwise would
not have been made.
Boron-10
Much of our effort in nuclear physics
has applied phase-shift analysis to meas-
urements of elastic scattering made with
the polarized beam. Last year we re-
ported the results of our studies of
7Li(p,p)7Li, our first attempt with a
spin-1/2 on spin-3/2 interaction which
gave information about the structure of
8Be. A similar study of 9Be(p,p)9Be,
an identical spin system, has led us to
results about 10B. This isotope, consist-
ing of five protons and five neutrons, has
both of its Is shells filled with two pro-
tons in one shell and two neutrons in the
other, but has only three nucleons in each
of the lp shells (protons and neutrons fill
separate shells) . The nuclear shell model,
our best picture of nuclear structure to
date, works best when the nucleons fill a
shell completely, when there is one nu-
cleon in excess, or when there is a hole in
the shell. Boron-10, with its lp shells
only 3/8 full, is difficult to handle but
offers a relatively simple case for theorists
to attempt an explanation of intermedi-
ate structure. Accurate identification of
the angular momentum and parity of the
excited states is required, of course.
The five bound states of 10B appear to
have been completely mapped: Their
energies are accurately determined and
the quantum number assignments are
not in dispute. Even the seven states
between the thresholds for cx-par-
ticle (4.461 MeV) and proton (6.587
MeV) emission have but three uncertain
spin-parity assignments. This harmony
disappears above the proton threshold,
however, where the states encountered
are generally broad, overlapping, and of
impure iso-spin (refer to Fig. 10) . These
conditions lead to data that are difficult
to interpret, and there is as yet no agree-
ment about the number and location of
states, much less about their angular mo-
mentum and parity.
We have measured the angular dis-
tribution of the polarization analyzing
power at 20 energies and have subjected
them, together with the previously meas-
ured cross-section data taken at the Cali-
fornia Institute of Technology and the
University of Maryland, to a phase-shift
analysis. Our first attempts to fit the
data naturally employed the structure
proposed on the basis of cross-section
measurements and (p,y) reactions. The
result was a failure, the discrepancies be-
ing too great to be resolved by small
modification of the resonating phase
shifts. Eventually, the set of phase shifts
192
Ex
(MeV)
8.89
r
(MeV)
0.038 & 0.085
2+8 3 +
8.65
0.3
(It2+)
7.82
0.3
r
7.56
0.003
o+ i
7.47
o.io a o.n
l+82~
6.96
(l+)
6.88
r
6.57
(2 + )
CARNEGIE INSTITUTION
- 3
Ex
Jt
8.438
9B + n
-
8.252
8Be+p+n
6.587
3/2"
- 2
- I
'Be + p
10
B
Fig. 10. Limited energy level diagram of 10B. The only levels shown lie above or near the pro-
ton threshold in the energy range spanned by this experiment. Ex denotes the excitation energy,
which is the energy of the state above the ground state ; r denotes the width of the state ; J17, the
angular momentum and parity; and T, the iso-spin. At the right is shown the information neces-
sary to relate the proton energy, Ep, to the energies of the excited states; three particle disinte-
gration thresholds are shown.
shown in Fig. 11 developed. The exist-
ence of an excited state is characterized
by the phase shift of a particular partial
wave resonating, which in most cases
shows an increase of about 180°. The
structure that satisfies the data is readily
apparent from Fig. 11: there are two
levels at proton energy 0.930 MeV (7.47
MeV excitation energy) : a 1+ (designated
5Pi in the figure) of width 0.10 MeV and
a 2- (designated 5S2) of width 0.11 MeV;
a third level at 1.37 MeV (7.82 MeV ex-
citation energy) is 1" (designated 3Si) of
width 0.30 MeV. This structure is simpler
than what has generally been supposed
necessary to satisfy the cross-section
DEPARTMENT OF TERRESTRIAL MAGNETISM
193
180'
120'
60<
0 -
120°
8
60'
0
'S, -
-8ooo0°ooOo? OOo0o_
o o o„
j0-^o-o-°-o-
•
/
#>.•
,od
1.0
Ep(MeV)
Fig. 11. Phase shifts as functions of energy. The curves show the phase shifts that were ob-
tained as the best fits to the experimental data. Closed circles denote energies at which there are
cross-section and polarization data; open circles indicate cross-section data alone.
data. The key to the matter came from
polarization, which forced a change of
spin assignment of one of the levels at
0.98 MeV and of the one at 1.37 MeV.
Reference to Fig. 10 discloses a 0+ state
at 1.086 MeV (7.56 MeV excitation en-
ergy) that is not seen in Fig. 11. This
state is very narrow, about 3 keV, and
shows up strongly in elastic scattering.
It is narrower than the energy resolution
of our experimental set-up, which pre-
cluded our observing it with the polarized
beam; more to the point, we carefully
avoided it so that its effects did not enter
into our measurements. We did look into
the spin-parity assignment of this state,
since we could bring a complete phase-
shift analysis to bear on the cross-
section data. The old assignment of 0+
withstood all our efforts to change it.
One aspect of the analysis of 7Li(p,p)7Li
was the demonstration of the need of
channel spin mixing in order to fit the
measurements for the two 1+ states en-
194
CARNEGIE INSTITUTION
countered (see Year Book 71, p. 247).
The same amount of mixing was also
needed for the 1+ state at 0.98 MeV in
9Be(p,p)9Be. Information on channel
spin mixing is difficult to obtain by other
methods. We have now seen three cases
of 1+ states requiring the same amount of
energy independent mixing. The theo-
retical implications of this are yet to
come.
Our measurements extended beyond
the energy limits shown in Fig. 11, but
the energy region from 1.7 to 2.8 MeV
had usable cross-section measurements at
only one angle. This proved to be a seri-
ous handicap. The polarization data
alone were insufficient to determine the
phases; in particular, the two s-phases
could not be determined, nor could the
effects of the 5Pi and 5P2 phases be sepa-
rated. Two levels in this region attracted
our interest (Year Book 71 , p. 250) orig-
inally because they are both at the same
energy, Ep = 2.56 MeV. Although we
could not determine a satisfactory set of
phases, we were able to test the assign-
ments, which had surprised us at first, of
the two levels degenerate in energy to
within 1 keV. The assignments seem
quite reasonable as seen through the
polarization measurements. The two
levels at 0.98 MeV are also degenerate
in energy, so our surprise at this phenom-
enon has been diluted.
A Study of (p,n) Reactions
The nuclear physicist, probably more
than any other kind of scientist, is locked
to the experiments that can be done with
a given set of equipment. It is his lot to
find that the pursuit of an interesting and
potentially fruitful aspect of his work
must terminate because further means of
observation have not been developed.
The converse requires him to study nu-
clear interactions because their observa-
tion is virtually a consequence of his
equipment. Our interest in the two over-
lapping 3+ states of 8Be (see Year Book
71, p. 247) led to building equipment for
studying them through the 7Li(p,n)7Be
reaction. Two pairs of neutron detectors
are now functioning. One pair uses recoil
protons to produce pulses in scintillation
counters that use fast electronics to dis-
tinguish the recoil protons from fast elec-
trons produced by y rays, an old and
serious problem of neutron detection.
This method allows only about 1% of the
y-ray photons to be counted as neutrons.
Another pair of detectors uses the 6Li
(n,t)4He reaction to produce the needed
scintillations. This technique is used for
neutron energies too low for the proton
recoil method. It is, fortunately, insensi-
tive to y rays but presents some shielding
problems against background neutrons.
Thus our interest in the structure of
8Be caused us to build equipment for ob-
serving neutrons. It is obvious that we
should observe the polarization effects in
two other reactions that fit almost per-
fectly into this experimental setup:
9Be(p,n)9B and 3H(p,n)3He. The former
reaction may well furnish information
about the structure of 10B, although just
what information is not immediately evi-
dent. The latter reaction is a study of the
first excited states of the a-particle, a
subject of fundamental importance.
BIOPHYSICS
E. T. Bolton, T. I. Bonner, D. B. Cowie, M. B. Davis, B. H. Hoyer, N. R. Rice.
R. B. Roberts, and D. G. Wallace
Introduction
A very broad range of experimental
problems has been pursued by members
of the Biophysics Section in the past
year, and real progress has been made in
each area. Immunochemical studies of
enzyme evolution in bacteria have been
brought to the Section. This program,
initiated last year at the Institut Pas-
DEPARTMENT OF TERRESTRIAL MAGNETISM
195
teur, Paris, has continued, and a col-
laborative program, both here and in
Paris, has been projected. Studies of
brain biochemistry have also been vigor-
ously pursued, yielding new and exciting
findings of a chemically induced en-
hanced rate of learning in mice. Investi-
gations concerning nucleic acids, long our
dominant research interest, have centered
on satellite and nonrepeated DNAs in
rodents and on the nature of DNA that
appears to reassociate instantaneously.
In addition, we are pleased to be able to
report technical advances in the produc-
tion of DNA fragments of defined piece
size as well as the in vitro labeling of
DNA with 125I. Two new postdoctoral
fellows and one graduate student joined
us late in 1972 and early 1973, bringing
backgrounds in nuclear physics, protein
evolution, and RNA synthesis. Their re-
search is now well under way.
Studies of Memory
R.B. Roberts
Peptides are synthesized in the brain
and are known to act in releasing hor-
mones from the pituitary gland. In ad-
dition, there are several indications that
peptides may play some role in learning
and memory. Lande, Witter, and de
Wied 15 have shown that hypophysec-
tomized rats cannot learn the conditioned
avoidance response in the shuttle-box un-
less they are injected with one of a num-
ber of peptides related to vasopressin.
We have observed that large injections
of crude ACTH improve the rate of
learning of normal mice in the shuttle-
box. De Wied has reported large effects
of small quantities of ACTH and vaso-
pressin on extinction of an acquired re-
sponse. Flexner found that injections of
vasopressin conferred an immunity to the
amnesia caused by injection of puro-
mycin.16 This amnesia (caused by puro-
mycin) is probably due to puromycin
peptides which remain in the brain for
long periods of time. Finally, in the ex-
periments which indicate a transfer of an
acquired behavior pattern after injections
of extracts of the brains of trained ani-
mals, the active material seems to be a
peptide.
With these thoughts concerning a pos-
sible role of peptides in memory forma-
tion in mind, we were intrigued by read-
ing that N-acetyl aspartic acid (NAA),
a compound found only in the brain, can
cause a stimulation of peptide synthesis
in brain slices. It seemed worthwhile to
determine whether this compound would
have any effect on learning and memory.
Mice were given ten training runs per
day in the shuttle-box and the training
was continued for five days. Control
groups were injected with water or not
injected and the experimental animals
were injected (i.p.) with NAA immedi-
ately after training each day.
On the first two days the mice learned
to escape the shock which was delivered
after 2% seconds of a warning tone. On
the third day and thereafter the mice
scored an increasing number of success-
ful avoidances of the shock by running
at the sound of the tone. The perform-
ance of the individual mouse was meas-
ured by the sum of the avoidances on
days 3, 4, and 5.
The dramatic and consistent improve-
ment caused by NAA injections is shown
in Table 2. This table includes tests on
ten different sets of DBA mice (different
ages, different sex, some purchased, some
home grown) and 17 groups of control
animals. In all cases, the performance
of the NAA-treated mice was better,
usually by a factor of 2. The improved
performance persisted when the mice
were tested again a week after the con-
clusion of the training runs. Thus, the
effect seems to be on the consolidation
of the memory of the training experi-
ence and is not simply due to improved
performance in the presence of the drug.
Dosage experiments indicate the NAA
is effective down to a level of 10 mg/kg.
Three mice injected with that dosage
after each day's training showed an aver-
age score of 10.3, compared to 5.6 for the
196
CARNEGIE INSTITUTION
TABLE 2. Performance of Mice after Treatment with NAA
Group
Size of
Group
Average Scores of Groups
NAA
Injected
3
4
5
9.7
10.7
9.4
10.3
10.0
10.3
11.7
9.7
13.7 12.7
Controls
3
4
5
37
61
5.0
3.75
7.6
4.0
8.0
4.2
6.0
8.5
7.0 2.33 4.67
4.25 4.5 5.5
Percent Correct Responses
4.33
NAA
Control
day 3
20.0
9.9
day 4
38.0
17.5
day 5
64.2
27.8
4.33 10.0
The upper part of the table shows the average scores of ten groups of NAA-treated animals and
seventeen groups of control animals; all were injected immediately after training on days 1-5. The
lower part shows the day-by-day average performance of 37 NAA-treated and 61 control animals.
All were DBA strain, but differed in age, sex, and source.
control animals. Animals given daily in-
jections of 200 mg/kg of NAA immedi-
ately after training showed marked im-
provement in performance. Four groups
of 4 mice each, had average scores of
10.0, 10.8, 11.7, and 12.7 again compared
to 5.6 for the controls. A group of 3
animals injected at the 200 mg/kg level
showed an average score of 13.7, and
a group of 5 mice averaged 9.4. The
normal level of NAA in the brain is re-
ported to be 5.4 milf; thus a dose of 200
mg/kg is sufficient to triple the normal
level in the brain even if it is distributed
uniformly through the animal. If the
compound is concentrated exclusively in
the brain (approximately %0 of the total
weight), then 5 mg/kg would suffice to
triple the concentration in the brain. The
compound is toxic when given daily at
400 mg/kg.
A few experiments were done using
other related compounds in place of, or
in addition to, the NAA. Each drug was
injected for five days immediately after
training runs, at a dosage equimolar to
NAA at 200 mg/kg. N-acetyl alanine
appeared to have similar effects in two
tests. After its injection in a group of
3 mice, an average score of 9.3 was ob-
served, compared to 5.0 for a control
group; in another group of 3, the average
score was 9.2, while controls averaged 4.2.
N-acetyl glutamic acid injections did not
produce an increase over control group
scores (average = 4.6 in a group of 3
animals). Aspartic acid gave ambiguous
results — increasing average scores to 9.2
in one group of 4 mice, while not effect-
ing any improvement in another group
of 3, which averaged 4.3. L-dopa may
counteract the effect of NAA ; it reduced
the average score from 11.7 to 5.5. The
statistical probability of a score as low
as 5.5 in a group of four is only 15/1000.
In some experiments the NAA was in-
jected at different times before training
sessions. The interval varied from 3
hours to 3 days. Injection before the
training period seems effective even when
a single dose is given 3 days in advance.
For two groups of 3 mice each, scores of
10.7 and 7.3 were observed after such
an injection. When 3 mice were injected
on three successive days before the start
of training, their average score was 8.3.
For the 4 animals injected 3 hours before
each training session, the average score
was 10.4. As these tests were carried out
with 200 mg/kg (20 times the minimum
effective dose) and the half-life is re-
ported to be 12 hours, it is not surprising
that its effect could persist.
DEPARTMENT OF TERRESTRIAL MAGNETISM
197
Some work was also done on the effect
of NAA injected after the completion of
the usual training period. When injected
on days 5, 6, and 7, it seems to have no
effect on the performance a week after
the last injection. When injected on days
11, 12, and 13, however, it has little effect
on the performance on day 14, but does
seem to improve the performance of day
15. This is an indication that NAA af-
fects not performance but the consolida-
tion of the additional training of day 14.
NAA was tested in several other learn-
ing situations. Dr. Rake found no effect
of NAA on the time to step into a cham-
ber where the mouse had previously
received a shock. Dr. Flexner observed
that treatment with NAA did not confer
immunity to the amnesia caused by puro-
mycin. We found no improvement in
maze learning in a small number of tests.
Thus, NAA seems quite limited in its
effects on various types of learning.
However, the effect on shuttle-box
learning is large and reproducible, so it
offers an opportunity to compare a
change in behavior with the concurrent
changes in biochemistry. At present, this
effort is in its initial stages. No change
was found in the level of norepinephrine
in the brain when NAA was injected
either 1 or 4 hours before the measure-
ment. Amino acid incorporation into
brain and liver does, however, seem to
be altered, but further checks are needed
before any definite conclusions can be
reached.
The Mouse Satellite in Two
Subspecies of Mus musculus
Nancy Rice and Patricia Esposito
About ten percent of the DNA of all
tested strains and cell types of the labor-
atory mouse Mus musculus musculus ap-
pears as a satellite in CsCl density gradi-
ents.17 It has been found to consist of
a rather short sequence (about 300 nu-
cleotides) which is highly repeated
(about a millionfold) ,18 Related satellite
families have been found in two other
Mus species, M. caroli and M. cervicolor.
Each of these identified families displays
quite limited heterogeneity (i.e., reasso-
ciates to high thermal stability Te50 over
80°C in 0.14 M PB) and yet is only very
loosely related to the others (Te$Q of
cross-product as low as 60°C in 0.14 M
phosphate buffer).19-20 Satellite DNA
from a much closer relative of the lab-
oratory mouse, the Japanese subspecies
M. musculus molossinus, is virtually in-
distinguishable from that of M. m. muscu-
lus (cross-product of very high thermal
stability). Nevertheless, as demonstrated
below, there are differences in the satel-
lites of even these two close relatives:
M. m. molossinus DNA contains signif-
icantly less satellite DNA than does M.
m. musculus DNA.
Difference in satellite content of the
two DNAs is clearly seen in preparative
Ag+-Cs2S04 density gradients. In the
examples of two such gradients shown
in Fig. 12, about 7.5% of the total optical
density (260 m/*) of M. m. musculus
DNA is found in the satellite band; the
M. m. molossinus satellite, of identical
density to the M. m. musculus satellite,
comprises about 4.5% of the total optical
density. As seen in Table 3, this differ-
ence in satellite content has been ob-
served consistently in various prepara-
tions of these two DNAs. Over all prepa-
rations, a mean of 7.9% satellite was
found for M. m. musculus DNA and
5.0% for M. m. molossinus DNA. These
values are significantly different (t test,
p < 0.01).
Difference in satellite content as de-
fined by a density gradient does not
necessarily mean difference in sequence
content. For example, if an altered ar-
rangement within the genome of many of
the M. m. molossinus satellite sequences
resulted in their migrating with main-
band DNA in the density gradient, the
percentage of DNA banding as a satellite
might significantly underestimate the
total number of satellite sequences pres-
ent. However, there are two indications
that M. m. musculus and M. m. molos-
198
CARNEGIE INSTITUTION
1 1
M. M. MUSCULUS
1
M.M.MOLOSSINUS
i
P
E 1.0
o
10
^^
1.50
CVJ
>■
H
to
z
Ld
Q
1
1.40
«*0.5
o
i-
Q.
O
i 1
i I i
I
i I i
20
40
FRACTIOI^
20
1 NUMBER
40
Fig. 12. M. musculus DNA in preparative Ag+-Cs2S04 gradients. 355 tig of each DNA was in-
cubated with AgNo3 at Ag+/DNA — P = 0.3 in 6.9 ml 0.01 M sodium borate buffer, pH 9.1, for
2 hours at room temperature.21 Solid Cs2S04 was added to give a density of 1.5 g/cc, and the sam-
ples were centrifuged at 33,000 rpm for 64 hours at 21 °C in the L40 rotor. Gradients were collected
(by tube puncture) into about 60 fractions; optical density was assayed after addition of 0.6 ml
0.1 M Tris, pH 8.0. The M. m. musculus satellite comprises about 7.5% of the total OD; the
M. m. molosinnus satellite, about 4.5%.
sinus DNA do in fact differ in the num-
ber of their copies of the satellite se-
quence.
First, the percentage of 3H-M. m. mo-
lossinus DNA (derived from an adult
kidney primary cell culture grown in
3H-thymidine) which behaves as satel-
lite in reassociation experiments is lower
than observed for radioactive 3H-M. m.
musculus DNA. For example, when a
mixture of 3H-M. m. molossinus DNA
and 14C-M. m. musculus DNA is incu-
bated with M. m. musculus DNA to C0t
5 X 10"2 (sufficiently low that much of
TABLE 3. Satellite Content of M . musculus DNA in Preparative
Ag+-Cs2S04 Density Gradients
Preparation
Percent
Number of
Number
Source of DNA
Satellite
Determinations
1
M. m. musculus liver
7.9%
8
2
M. m. musculus liver
8.0
1
3
M . m. musculus liver
8.9
1
4
M. m. musculus liver
8.6
1
5
M. m. musculus liver
5.9*
1
6
M. m. musculus brain,
kidneys, spleen
7.9
1
7
M. m. molossinus liver
6.0
4
8
M. m. molossinus carcass
4.5
2
9
M. m. molossinus liver,
kidneys, spleen
4.6
3
*A11 DNAs were prepared by the urea-phosphate method of Britten et at.,
Year Book 68, p. 400, and, except for Prep. No. 5, all were treated identically.
A modification of the same method was used in Prep. No. 5, resulting in impure
DNA; it was subsequently RNase treated, ethanol precipitated, and repassed
through HAP in urea-phosphate. Whether this additional handling accounts
for the low percent satellite in the purified product is not known.
DEPARTMENT OF TERRESTRIAL MAGNETISM
199
the observed reassociation is due to satel-
lite), about 12% of the 14C, but only
about 8% of the 3H, elutes from HAP
above 76°C (Fig. 13). This difference is
also apparent when the same radioactive
DNAs are allowed to reassociate with M.
m. molossinus DNA to C0t 10~2. Here,
about 10% of the 14C, but only 6% of
the 3H, elutes from HAP above 76°C.
Similarly, over 6% of radioactive M. m.
musculus DNA, but only about 4% of
3H-M. m. molossinus DNA, is able to
reassociate with the isolated L strand of
the M. m. musculus satellite. In these
reassociation experiments, there is no de-
tectable difference (< 1°C) in the
thermal stability of the heterologous and
homologous products (Tm from HAP in
0.14 MPB = 82°C). Thus, the only
observable difference in the satellite in
these two radioactive DNAs is in its
relative abundance.
The second indication that M. m. mo-
lossinus DNA contains fewer satellite
sequences than does M. m. musculus
DNA comes from reassociation kinetics
experiments: a 3H-M. m. musculus DNA
fraction composed primarily of satellite
reassociates more slowly with M. m. mo-
lossinus DNA than with M . m. musculus
DNA. The 3H-DNA fraction was incu-
bated with M. m. musculus DNA (a
preparation which exhibited 7.9% satel-
lite in density gradients) or M. m.
-3
—
SlOO
Q_
1
1 1
I IV'
—
PERCENT TOTAL
o
- /Vl4c -
7
1
1
•/
/*\ 1 1
/tl4c
\>A
80 90
TEMP.(0C)
'1
^r \\
1 \ \
(m ^^
/ SELF
ji REACTIONS
1 -^~^==*f-#-^.
^\ 1
-4
-2
60 70 80 90
TEMPERATURE. °C
100
Q_
(J
_l
<
O
h-
h-
LU
O
LT
LJ
Fig. 13. Reassociation of tracers with M . m. musculus DNA. A mixture of sheared UC-M. m.
musculus and 3H-M. m. molossinus DNA was incubated with unlabeled M. m. musculus DNA at
50°C in 0.14 M PB to Cot 5 X 10~2. The sample was applied to HAP at 50°C, and the bound
DNA (22.1% of 14C, 18.5% of 3H) was eluted with increasing temperature. A parallel incubation
of tracers alone measured their self-reaction (5.4%; Cot 3 X 10"5). Inset: Material eluting above
76 °C has been corrected for self-reaction and normalized to 100% reaction.
200
CARNEGIE INSTITUTION
molossinus DNA (a preparation which
exhibited 4.6% satellite in density gradi-
ents) to identical C0t at 60 °C and the
percentage of reassociation of the tracer
was measured. The difference in reac-
tion rate of the tracer with these two
DNAs can be calculated from the data
shown in Table 4 to be about 1.6-fold, in
very good agreement with the 1.7-fold
difference in satellite content seen in the
density gradients.
Except for their satellite contents, the
DNAs of these two subspecies appear
indistinguishable in all respects tested.
Their complements of repeated DNA are
extremely similar (3H-M. m. musculus
DNA reassociates to identical extent and
thermal stability with repeated DNA of
either M. m. musculus or M. m. mo-
lossinus) , and any sequence differences in
their single-copy DNAs have been esti-
mated at less than 1% (Year Book 70,
p. 366). The mice do differ in at least
several phenotypic traits (size, colora-
tion, some chromosomal rearrange-
ments,22 etc.) ; nevertheless, they inter-
breed to form fertile hybrids. Whether
their differing satellite content has any
functional significance remains to be
seen.
TABLE 4. Reassociation of a Satellite-Enriched
3H-M. m. musculus DNA Fraction
Cot
Percent
(added
Reassociation
DNA Added
DNA)
of Tracer
None
2.6
M . m. musculus
1.9 X10-3
35.7
M. m. molossinus
1.9 X10-3
25.7
M . m. musculus
7.6XKT3
63.8
M. m. molossinus
7.6XIO-3
54.0
The 3H-DNA fraction was prepared by in-
cubating sheared DNA at 60°C in 0.14 M PB to
C0t 10~2 and collecting all molecules which eluted
from HAP above 70°C (13% of total 3H).
Instantaneous binding material was eliminated
from this fraction by heating it to 100°C and
repassing it through HAP; the nonbinding DNA
(93%) is the 3H- tracer used in the subsequent
experiments. Tracer was incubated with 5 or
20 p-g nonradioactive sonicated DNA in 11 ml
0.14 M PB at 60°C for 20 minutes; binding of
the 3H-DNA to HAP at 60°C was measured.
Relatedness among Several Hamsters
Nancy Rice and Patricia Esposito
Interspecies comparisons of both
single-copy and repeated DNA fractions
are often helpful in determining the de-
gree of relatedness among the species
tested. Toward this end, we have com-
pared the DNAs of four species of ro-
dents: Mesocricetus auratus (the
"golden," or "Syrian," hamster) , Cri-
cetulus griseus (the "Chinese" hamster) ,
Mystromys albicaudatus (the "white-
tailed rat," or "South African" hamster) ,
and Mus musculus (the "house mouse").
M. musculus is classified as a member
of the family Muridae. Though Eller-
man 23 expressed uncertainty as to the
best classification for Mystromys, both
Simpson 24 and Arata 25 have grouped all
three hamsters in the Tribe Cricetina
(true hamsters) of the subfamily Crice-
tinae, family Cricetidae. Both the Syrian
and Chinese hamsters are native to parts
of Europe and Asia ; the white-tailed rat
is found only in southern Africa.26
Relatedness among these four rodent
DNAs has been assessed according to
their ability to reassociate with Syrian
hamster single-copy 3H-DNA. In these
experiments, the tracer has been incu-
bated with a large excess of nonradioac-
tive DNA to high Cot; both the extent
of cross-reaction and the thermal sta-
bility of the cross-product were then
measured. The results of one such ex-
periment, in which Syrian hamster single-
copy 3H-DNA was allowed to reassociate
with Chinese hamster DNA, are shown
in Fig. 14. About 55% of the tracer was
able to reassociate with the heterologous
DNA in the time allotted (compared to
over 70% reassociation of the tracer with
homologous DNA under similar condi-
tions), and the thermal stability of the
cross-product (Te50 about 70 °C) is much
reduced relative to the homologous prod-
uct (Te50 81-83°C). This Te50 of the
cross-product is roughly comparable to
that observed between single-copy DNAs
of mouse and rat (Year Book 71, pp.
DEPARTMENT OF TERRESTRIAL MAGNETISM
TEMPERATURE, °C
50 60 70 80 90
3H
201
a.
o
E
O
CD
C\J
-0.200 >
CO
LU
Q
o.ioo <3
Q_
O
20 40
FRACTION NUMBER
Fig. 14. Thermal elution profile of Syrian hamster single-copy 3H-DNA reassociated with Chi-
nese hamster DNA. Syrian hamster single-copy 3H-DNA (of average length about 170 nucleo-
tides) was prepared by incubation to Cot 120 and collection of nonreassociated molecules (52% of
total) by passage through HAP at 50 °C in 0.14 M PB. For the experiment shown above, about 2
fig (4000 cpm/Vg) of single-copy 3H-DNA was incubated with about 2.5 mg denatured Chinese
hamster DNA (length about 450 nucleotides) and 2.2 /ig denatured E. coli 14C-DNA (which serves
as an internal standard for column operation) in 1 ml 1 M PB, 0.005 M EDTA at 63 °C for 7 days.
One-half of the sample was diluted to 0.14 M PB and applied to a water-jacketed HAP column at
45°C; 55% of the 3H, 85% of the OD, and 86% of the 14C bound. Reassociated DNA was eluted
during a linear temperature gradient. Column buffer, 0.14 M PB, 0.06% SDS; volume per frac-
tion, 6.5 ml.
262-264) and between single-copy DNAs
of man and New World monkey.27
Quite different results are obtained
when the Syrian hamster single-copy 3H-
DNA is allowed to reassociate with either
white-tailed rat or mouse DNA (Table
5) . In neither case is there more than
30% reaction of the tracer (actually an
overestimate, since some of the observed
reassociation occurs when tracer and non-
radioactive DNA are incubated to low
C0t) , and the Te50 of the cross-product is
even lower than that observed with
Chinese hamster DNA (about 65°C).
Thus, Syrian hamster DNA appears no
more closely related to white-tailed rat
DNA than it does to mouse DNA, a re-
sult which indicates that lines leading
to Syrian hamster and white-tailed rat
diverged considerably earlier (perhaps
up to a factor of 2) than those leading
to Syrian and Chinese hamsters.
The same pattern of relatedness is
seen when repeated DNAs are compared.
Over 25% of Chinese hamster 14C-DNA
(a preparation of 250 nucleotides aver-
age size) will reassociate with unlabeled
Chinese hamster DNA at C0t 10 (50°C,
0.14 M PB); roughly one-third of this
reassociated material is of high thermal
202
CARNEGIE INSTITUTION
TABLE 5. Reassociation of Syrian Hamster Single-copy 3H-DNA with Other Rodent DNAs.
DNA added
Cot*
(added
(DNA)
% bound to HAP
OD 3H
14Q
Te5ot
3H
A
Te50
Syrian hamster
1000
3300
87%
92
71%
74
87.7°
88.0
82.0°
80.3
5.7°
7.7
Chinese hamster
1800
5800
16
79
85
50
55
9
88.1
88.0
69.2
69.6
63.0
18.9
18.4
White-tailed rat
1600
25
77
22
6
88.7
65.2
62.5
23.5
Mouse
1600
27
88
29
7
87.7
64.0
63.0
23.7
*No correction for the high salt concentration has been made in calculation of Cots. For comparison,
E. coli DNA of similar size reassociates under these conditions with Coti/i about 1.3.
fT^o is the temperature at which one-half of the bound material has eluted. Syrian hamster single-
copy 3H-DNA was incubated with variable amounts of nonradioactive DNA to the indicated Cots
in 1 M PB, 0.005 M EDTA at 63°C. Denatured E. coli 14C-DNA was included in all incubations
and serves as an internal standard for column operation; at least 65% bound in each case. Details
in legend to Fig. 14.
Q_
O
-I
<
O
h-
LU
O
cr
Ld
Q_
4.0
2,0
SYRIAN HAMSTER (25%)
MOUSE (18%)
WHITE TAILED RAT(I5%)
SELF REACTION (5%)
+ -- + -- +--+---+--_._+
100
TEMPERATURE,°C
Fig. 15. Thermal elution profile of Chinese hamster 14C-DNA reassociated with other DNAs.
Chinese hamster 14C-DNA (average length about 250 nucleotides) was incubated with each of the
indicated DNAs (average length about 450 nucleotides) to Cot 6-11 at 50 °C in 0.14 M PB. Sam-
ples were applied to HAP (percent binding of the tracer is indicated in the figure) and eluted in
4°C increments.
DEPARTMENT OF TERRESTRIAL MAGNETISM
203
4.0
(J
P
en
LU
a.
2.0
*\ .CHINESE HAMSTER (15%)
xr
MOUSE (14%)
+rr.j7rrSELF(2%)
i i i
60
70
80
90
100
TEMPERATURE, °C
Fig. 16. Thermal elution profile of white-tailed rat "C-DNA reassociated with other DNAs.
White-tailed rat "C-DNA was incubated with each of the indicated DNAs to Cot 9 at 50° C in
0.14 M PB. Samples were applied to HAP (percent binding of the tracer is indicated in the fig-
ure) and eluted in 4°C increments.
stability (elutes from HAP above 75°C).
The results of incubating the same tracer
with other DNAs to comparable Cot are
shown in Fig. 15. The most extensive
cross-reaction is observed with Syrian
hamster DNA (20% net tracer reaction) ,
though less than half as many duplexes
of high thermal stability are formed as
with Chinese hamster DNA. Cross-reac-
tion between the tracer and either mouse
or white-tailed rat DNA is lower (less
than 15% net) , and little or no material
of high thermal stability is observed.
A reciprocal experiment, reassociation
of white-tailed rat 14C-DNA with other
DNAs, gives the same results: white-
tailed rat DNA appears no more closely
related to Chinese hamster DNA than
to mouse DNA (Fig. 16). The extent of
cross-reaction is low (about 13% net),
and there is little or no cross-product of
high thermal stability. Surprisingly, re-
association of the tracer with homologous
DNA to comparable C0t (9) gives results
that differ only slightly from those of
the heterologous incubations. About 12%
of the tracer (and of the unlabeled DNA)
TABLE 6. Repeated Sequence Content of
Various Rodent DNAs
Percent
Percent
of Total
of Total
DNA
DNA
Bound
Eluting
DNA
Cot
to HAP
above 75 °C
Mus musculus
10
40
25
Rattus norvegicus
20
40
18
Peromyscus
maniculatus
8
42
25
Microtus
pennsylvanicus
1
21
7
Cricetulus griseus
25
30
10
Mystromys
albicaudatus
9
12
3
63
17
7
All DNAs are of 450 nucleotides average
length. Incubation in all cases was performed in
0.14 M PB at 50°C; binding to HAP was mea-
sured under these same conditions.
204
CARNEGIE INSTITUTION
reacts, and only about one-fourth of this
elutes from HAP above 74°C. Slightly
more reassociation is observed (about
17% of total cpm or OD) when the DNA
is incubated to higher C0t (63) ; here,
about one-third of the product elutes
from HAP above 75°C. These levels of
repeated sequences are the lowest we
have observed in any rodent DNA tested,
a list which includes various species of
Mus, Rattus, Mastomys, Cavia, Pero-
myscus, Oryzomys, and Microtus, as well
as those employed in the present experi-
ments (Table 6). The significance, if
any, of this variable repeated sequence
content (previously noted in optical re-
association studies of rodent DNAs by
Hennig and Walker),28 as well as the
identity of a closer relative of Mystro-
mys, is still unknown.
Instantaneous Binding Fractions of
DNA
T. I. Bonner
In most DNA reassociation experi-
ments using tracers, one finds that even
at extremely low C0t values (e.g., 10-6
for mouse DNA) a few percent of the
labeled DNA binds to hydroxyapa-
tite,29' 30- 31 suggesting that it has some
double-stranded structure. The C0t val-
ues rule out the possibility of reassocia-
tion by bimolecular collision. There are,
however, several possible sources of the
binding of the DNA. Some are of a tech-
nical nature, such as the possibility that
the DNA was not completely denatured
or that the hydroxyapatite randomly
binds a small amount of DNA independ-
ently of any double-stranded structure.
On the other hand, the binding could be
specific to a fraction of the DNA for
several reasons. First, some of the pieces
of DNA could be cross-linked so that
complementary strands never completely
separate in the denaturation step. The
existence of this effect has been demon-
strated for some bacterial DNAs.32,33
The second possible cause of specific bind-
ing would be the existence of partially
complementary sequences on a single
strand of DNA. After denaturation and
return to reassociation conditions, such a
strand could rapidly reassociate with it-
self without requiring a bimolecular col-
lision. There is reason to believe that
such sequences may be common. A num-
ber of RNAs are known to have large
amounts of secondary structure, e.g.,
tRNA and ribosomal RNA, which is
about 50% base paired at a normal re-
association condition (50°, 0.14 M PB).
The DNA sequences from which these
RNAs are transcribed would be expected
to also have large amounts of intrastrand
base pairing. It also appears that a large
amount of secondary structure is present
in some RNA viruses and that the sec-
ondary structure plays a role in control-
ling the order of transcription of contigu-
ous genes. Thus, sequences of this type
could be present either as the result of
the integration of viral DNA or of their
evolutionary advantage as a mechanism
for the control of gene expression. A
third possibility is that some fragments
have an enhanced affinity for HAP which,
while specific to those fragments, does
not depend on the presence of secondary
structure. It is possible, for example,
that longer fragments are more likely to
stick to HAP. In order to determine the
nature of the rapidly binding portion of
the DNA, we have performed a number
of experiments using 3H-thymidine—
labeled mouse DNA.
We have measured the "instantaneous
binding" for two sizes of DNA, sonicated
DNA with a length of about 450 nucleo-
tides and DNA which results from the
MUP extraction process and which is 5-
10 times as long. Sonicated tracer gives
an instantaneous binding of about 4%
with two-thirds of the material having
high thermal stability, i.e., eluting from
hydroxyapatite above 75 °C (see Fig. 17).
The thermal elution profile of sonicated
native tracer is nearly identical to the
profile of the instantaneously binding
DEPARTMENT OF TERRESTRIAL MAGNETISM
205
4.0-
(-
Z>
o
(J
<
O
o
I-
en
Q_
55 60 70 80 90 95 .4
TEMPERATURE OF ELUTI0N f°C)
Fig. 17. Thermal elution profiles of sonicated
(about 450 nucleotides long) tracer (circles)
and tracer which is estimated to be 5-10 times
as long (plusses). The tracer was denatured by
boiling for 5 minutes in 0.14 M PB, chilled in
ice water for 2 minutes, placed on a 50° C hy-
droxyapatite column and eluted 2 minutes later
with 0.14 M PB. The total Cot prior to elution
is less than 5 X 10"5. Incubation of the long
tracer at 50 times greater Cot indicates that less
than 1% of the mouse satellite should be re-
associated. The amount eluting in 5°C temper-
ature steps is expressed in terms of percentage
of total counts loaded on column. The total
binding of the short tracer is 4.4% and that of
the long tracer is 21.2%.
(IB) tracer above 80 °C except that it
may melt about 1°C lower. The high
stability of IB tracer is highly suggestive
of cross-linking which should allow the
DNA to reform to essentially native sta-
bility. The amount of binding is, however,
much greater than was observed where
cross-linking was demonstrated32,33 for
bacterial DNA. In those cases the
amount of binding, when extrapolated
linearly to the sonicated size, is about
0.01%. Using MUP length DNA, the
amount of IB is about 20% and more
than half of the IB tracer has low sta-
bility (elutes below 75°C). Low thermal
stability appears to be quite unlikely for
cross-linked DNA but would be expected
to result when a single strand folds back
on itself to form either short or imper-
fectly base-paired regions.
In order to determine whether the
binding is the result of a failure to de-
nature the DNA completely, a sample
was boiled in 0.035 M PB instead of 0.14
M PB, which should have lowered the
melting temperature by 10 °C. No dif-
ference in binding was observed. In ad-
dition, measurements of binding follow-
ing denaturation for 5, 10, and 15 minutes
show a small decrease in binding which is
compatible with estimates of the reduc-
tion in size due to strand scission.
The specificity of the binding to hy-
droxyapatite was checked by measuring
the rebinding of fractions which initially
bound (15% total binding). The fraction
which initially bound with high stability
(greater than 75 °C) , rebound about 65%.
Since less than 1% of the initially non-
binding material will bind with high
stability on a second pass, this fraction
appears to be quite specific. Specificity
of the fraction which binds with low sta-
bility (less than 75°C) is less clear. Al-
though it also rebinds about 65%, the
low-stability binding on a repass of
initially nonbinding material has been
measured to be about 1% or about 6%,
using different tracer preparations and
different HAP batches. Thus, the low-
stability fraction may contain a moder-
ate amount of nonspecific binding. On
the other hand, it would be difficult to
reconcile the 65% rebinding (which is
identical to the high-stability rebinding)
with more than a small amount of non-
specific binding.
In order to determine whether cross-
linking is present, alkaline sucrose gradi-
206
CARNEGIE INSTITUTION
ents were run, using non-IB, low-stability
IB and high-stability IB fractions. In
these denaturing conditions, cross-linked
fragments should have twice the molecu-
lar weight of single-stranded noncross-
linked fragments. The results are not
yet conclusive but indicate that the low-
and high-stability IB fractions have mo-
lecular weights about 1.7 to 2.0 times as
large as the nonbinding fraction. This
result suggests that cross-linking is in-
volved, but it is possible that the IB frac-
tions are selected on the basis of greater
length.
In order to determine whether the in-
stantaneous binding sequences are rep-
resentative of a specific fraction of the
genome or are randomly distributed
throughout the genome, two experiments
were performed. In the first experiment,
the IB fraction from DNA several times
as large as sonicated DNA was sonicated
along with a large excess of unlabeled
mouse DNA. By measuring the reasso-
ciation of tracer compared to the optical
density of the cold DNA at various Cot
values, one can learn whether the IB
sequences are contiguous with sequences
representing a kinetically defined frac-
tion of the genome. As shown in Fig. 18,
when one subtracts the 27% instantane-
ous binding component, the tracer has a
reassociation curve quite comparable to
the whole mouse genome. There appears
to be about 8-10% less tracer reacting in
the intermediate Cot range of 1—100 with
a corresponding increase in the single-
copy region, but there are no repetition
frequencies that are dramatically en-
hanced or suppressed. This is similar to
the result of Davidson et al.sl who find
no difference in the reassociation of simi-
larly prepared Xenopus tracer and un-
labeled Xenopus DNA.
The second experiment sought to de-
termine whether there was any variation
in IB with base composition. Fractions
were taken across a CsCl gradient of
mouse tracer, and the IB character of
these fractions was determined. As
shown in Fig. 19, there is a definite varia-
tion in the amount of tracer IB across the
icr3
10-2
10"' 0° 10
C0t (WHOLE MOUSE)
10 2 I03
10 4
Fig. 18. Reassociation kinetics of sonicated tracer in the presence of a 104-fold excess of un-
labeled mouse DNA (except for Cot < 10-4 points which are tracer alone). The instantaneously
binding fraction (12%) of a longer tracer was sonicated together with the cold DNA. The frac-
tion of tracer binding to HAP is shown as plusses, the fraction of total optical density which
binds is shown as circles. The optical density readings of nonbinding and binding fractions were
corrected for hypochromicities of 12.9% and 23.7%, respectively, as determined by melts of such
material after incubation to Cot ^ 2000.
DEPARTMENT OF TERRESTRIAL MAGNETISM
207
1.76
£
u 1.72
>-
t 1.68
CO
.64
1200
1000-
800-
O
O 600
Q.
O 400
200
"i — I — r
1 — i — I — r
i ~ t — r+ l i I i i i i i i i
10 14 18 22 26
FRACTION
30
Fig. 19. The instantaneous binding of various
fractions resulting from a CsCl gradient of
mouse tracer estimated to be 1200-4000 nucleo-
tides long. The tracer was used in amounts
ranging from 0.025 ml to 0.100 ml in 15 ml of
0.14 M PB. The instantaneous binding is shown
as open circles, the number of counts per min-
ute for 0.010 of gradient fractions is shown as
plusses, and the density of gradient fractions is
shown as closed circles.
gradient. There are two sources of varia-
tion which would occur even if the
amount of DNA with IB properties did
not vary with base composition. The first
is that the specific activity of the DNA
varies linearly across the gradient (in-
creasing with increasing fraction num-
ber) since the label is only in thymine.
Thus, part of the increase in tracer bind-
ing of fractions 12-15 could be due to
increasing specific activity. However,
the decrease in binding of fractions 16-21
must be due to another effect. The sec-
ond source of variation in tracer binding
is the possibility that the extremes of the
distribution of tracer in the gradient may
be enriched in short fragments relative to
the peak of the distribution. Shorter
fragments would be expected to bind
smaller amounts of tracer. However, the
size variation would be expected to be
symmetrical about the peak, whereas the
amount of tracer binding peaks well
down the heavy side of the gradient peak.
It is therefore apparent that the amount
of IB DNA has a dependence on base
composition. Preliminary results indi-
cate that the variation is more pro-
nounced in the low-stability fraction.
The present results suggest that fold-
back, cross-linking, and some nonspecific
binding all occur. The binding does not
seem to be uniformly distributed through-
out the genome and is specific enough
that it may prove a useful means of frac-
tionating the genome. It remains to be
seen what the functions of these fractions
are.
Enzyme Evolution in the
Enterobacteriaceae
Dean B. Cowie, Paolo Truffa-Bachi*
and Georges N. Cohen*
The extent of homology among nucleo-
tide sequences of the DNA of related
organisms is a measure of their relative
evolutionary histories. In bacteria, evo-
lutionary divergence may arise by ran-
dom single-base changes ; by gene fusion ;
or by the deletion, addition, or substitu-
tion of large DNA segments. Such
changes may be indicated by determining
the extent of DNA homology and the
thermal stability of the reaction products
formed by the hybridization of the DNAs
of any two related species.
Since more than one mechanism is
involved in genetic divergence, portions
of the genome might be expected to have
different rates in the extent of fixation of
base sequence change. These rates, how-
ever, will depend upon the number of
changes tolerable without damaging por-
tions of the DNA which code for the
enzymes essential for survival.
* Service Biochimie Cellulaire. Institut Pas-
teur, Paris, France.
208
CARNEGIE INSTITUTION
Such DNA-DNA hybridization studies
have already indicated that certain por-
tions of bacterial or viral genomes are
highly conserved. Other portions show
marked divergence and the loss or addi-
tion of large contiguous segments of nu-
cleotide sequences. These differences
among the DNA sequences may result in
differences in amino acid sequence and
protein structure. Indeed, recent im-
munological comparative studies of sev-
eral enterobacterial enzymes34,35 gave
results that indicated the pattern of in-
tergeneric relationships agrees with that
derived from DNA hybridization inves-
tigations.36*37,38
On the other hand, certain proteins —
despite species variations — are known to
be conserved, tolerating little change.
Cohen et al. (1969) 39 provided strong
biochemical evidence that throughout the
Enterobacteriaceae distinctive features
of the regulation of aspartokinase I-
homoserine dehydrogenase I (AKI-
HSDI) are conserved in spite of wide
variations in the guanine-plus-cytosine
(G + C) content among these bacteria.
Last year another immunological tech-
nique was developed (Year Book 71, p.
253) to test whether the known genetic
divergences among these enterobacteria
are reflected in the structure or composi-
tion of this essential enzyme. Sepharose
columns were prepared containing anti-
bodies specific for purified E. coli K12
AKI-HSDI. When crude extracts from
E. coli K12 were passed through the col-
umn (in the presence of 2 X10"3 M L-
threonine) , AKI-HSDI was bound to the
antibodies and removed from the crude
extract.
A number of enterobacterial strains
(shown in Table 7) were selected to test
whether the homoserine dehydrogenase
(HSD) contained in crude extracts from
these cells would bind to the E. coli K12
AKI-HSDI antibody column. Table 7
also shows the G + C content of their
TABLE 7. Bacterial Strains Employed43-44
Percent
Relative
Percent
Binding
Strain
G+C
at 60°C
E. coli K12 Tir 8
50-52
100
Serratia marcescens
55-58.5
24
Proteus vulgaris
37-39
<10
Providencia sp. 7/67
40-42
11
Aeromonas hydrophila
55.5
?
DNAs and the degree of genetic related-
ness as determined by DNA-DNA hy-
bridization studies (using E. coli DNA
as the reference material). For these
enzyme tests, homoserine dehydrogenase
was assayed in the direction homoserine
— > aspartate semialdehyde by following
NADP+ reduction spectrophotometric-
ally.40
The sepharose-antibody column util-
ized in these experiments completely
binds all the AKI-HSDI contained in
crude extracts of E. coli K12, strain Tir
8 * as shown in Fig. 20. The capacities
of the immunoadsorbents were deter-
mined by adding successive loads of crude
extract containing known quantities of
AKI-HSDI to the column. After each
load of protein had passed through the
column, it was washed with buffer A f
and a new load added. When enzyme
activity appears in the eluate, the col-
umn is saturated to determine the capac-
ity.
Usually, after each experiment the
column was washed with 4 or 6 M urea
(containing .001 N hydrochloric acid),
removing all the enzyme bound to the
antibodies. The column was then washed
with buffer A and was ready for further
use.
Figure 21 shows that at least a part of
the AKI-HSDI remains active while on
the column; elution with 1 M urea re-
* Strain Tir 8 is a genotypically derepressed
mutant of E. coli K12.
t Buffer A : 2 X 10"2 M potassium phosphate,
pH 7.2, containing 0.15 M KC1.
DEPARTMENT OF TERRESTRIAL MAGNETISM
209
TIR 8
2X I05
cc
UJ
Q.
CO
O
U
>
O
<
o
<
ENZYME
I04
E. coli TIR 8
TOTAL
PROTEIN
LOAD
LOAD
1000
800
3
o
<
00
600
J L
>-
N
400
200
20 40 60
FRACTION NUMBER
Fig. 20. Immunoadsorption of AKI-HSDI on antibody column. Open circles, appearance of
labeled protein in eluant fractions; closed circles, appearance of AKI-HSDI activity in eluant
fractions. Proteins labeled by growing cells in synthetic medium containing C14 L-arginine.
leases active enzyme and, when followed
with 2 M urea, additional enzyme activ-
ity is eluted. Additional protein is re-
leased by 4 M urea, but at this concen-
tration the enzyme is denatured.
When crude extracts from Serratia
marcescens or Proteus vulgaris are added
to the antibody column, essentially all of
the HSD enzyme activity is retained by
the column. Saturation occurs at about
the same level as that observed with the
AKI-HSDI contained in the crude ex-
tract from E. coli Tir 8 cells. The aver-
age G + C content of Serratia DNA is
55-58.5%, while that of Proteus ranges
between 37 and 39%, quite different from
that of the E. coli cells (50-52%). The
degree of genetic interrelationship as
measured by DNA hybridization meth-
ods is 24%, compared with ~ 9% in E.
coli. It is clear that, despite these dif-
ferences, HSD contained in the crude ex-
tracts of both Serratia and Proteus has a
high affinity for the AKI-HSDI antibody
210
CARNEGIE INSTITUTION
column. Only a few percent of the HSD
activity is observed in the eluate after
passage of the crude extract through the
column (until saturation occurs) .
Another enterobacterium tested, Provi-
dencia sp. 7/67, gave different results.
No retention of the HSD activity was
observed when crude extracts of fiese
bacteria were loaded on the antibody
column. Two loads of 1600 enzyme units
were loaded on the column. The HSD
activity contained in the eluant fractions
was observed to be 1690 and 1600 units.
This was the only enterobacterium tested
in which the HSD had no affinity for the
antibody column.
Cohen et al.,39 assessing taxonomic in-
terrelationship by studying regulatory
mechanisms among the enterobacteria,
came to the conclusion that Aeromonas
hydrophila is closer to the coliform bac-
teria than to the aerobic pseudomonads.
When crude extracts of Aeromonas hy-
drophila were passed through the anti-
body column, 70% was retained by the
column. Some of the HSD activity that
passed through did not completely track
with the bulk of the Aeromonas protein,
I M UREA
2 M
UREA
<-4M UREA-
-io_-
-8 ^-4
O
ro
I-
<
•6 -
V)
3
UJ
-4 S -
M
UJ
-2
ro
O
X
5 ~
UJ
Q_
CD
Z>
o
o
2 f
<
o
Q
<
or
10 20 30
FRACTION NUMBER
Fig. 21. Release of antibody-bound protein with urea. Open circles, labeled protein released in
eluant fractions by the addition of 1 M and 2 M urea to antibody column ; closed circles, labeled
protein released in eluant fractions with 4 M urea; X's, appearance of AKI-HSDI activity in
eluant fractions. Proteins labeled as in Fig. 20.
DEPARTMENT OF TERRESTRIAL MAGNETISM
211
E
O
00
\-
<
</)
z
Q
<
O
Q_
O
10 20
FRACTION NUMBER
Fig. 22. Passage of proteins from Aeromonas hydrophila through antibody column. Open cir-
cles, total protein as determined by optical density of eluant at 280 nm; X's, HSD activity in
eluant fractions.
as shown in Fig. 22. This result would
indicate that a portion of the HSD had a
different chromatographic characteristic
than that retained by the column or that
which moved with the passage of bulk
protein. It would then appear that this
organism has some HSD not specific to
the AKI-HSDI antibodies.
Exchange of antibody -bound homo-
serine dehydrogenase. Last year exchange
reactions were observed when crude ex-
tracts of an E. coli mutant Gif 108 (con-
taining aspartokinase I but devoid of
homoserine dehydrogenase I) were loaded
on the antibody column previously sat-
urated with wild-type AKI-HSDI. The
properties of eluted enzyme were unus-
ual: When tested for threonine sensitiv-
ity, HSD activity was enhanced rather
than inhibited as it is for wild-type en-
zyme. This enhancement was always
greater than 100%.
This result could be due to either the
displacement of wild-type enzyme or the
formation of hybrid molecules containing
monomers or dimers from both mutant
and wild-type enzymes. Whereas the
enhancement by threonine would favor
the hypothesis that the molecules eluted
are different from the wild-type AKI-
HSDI, the possibility remains that wild-
type enzyme is eluted in a stable, less
active conformation and that the en-
hancement by threonine reflects a reac-
tivation to the normal configuration.
This year further exchange reactions
were investigated. Experiments were
carried out to determine whether HSD
previously bound to the antibody col-
umn would exchange with enzyme con-
212
CARNEGIE INSTITUTION
tained in crude extracts of other entero-
bacteria. In one experiment the antibody
column was saturated with radioactive
HSD by the passage of crude extract
from E. coli strain Tir 8 labeled with C14
L-phenylalanine. Two loads, one ml
each, containing 12,000 units of HSD
activity per ml were passed through the
column (capacity ^20,000 units) and the
column washed with buffer A until no
radioactivity appeared in the eluate. The
total radioactivity passed through the
column was 179,000 counts per minute;
at column saturation with HSD 2%4ths,
or 150,000 counts per minute, would have
been put on the antibody column.
The addition of nonradioactive crude
extract from Serratia marcescens sp.
7/72 cells caused the release of radio-
active protein, as shown in Fig. 23. The
first Serratia extract contained 13,500
units of HSD activity; roughly 13,000
units were recovered in the eluate, indi-
cating that the column had been pre-
viously saturated with the E. coli enzyme.
The radioactivity eluted by the Serratia
extract was 1000 counts per minute.
A second load of crude extract from
Serratia cells (2800 units of HSD) re-
sulted in the further displacement of
radioactive protein (750 counts per min-
ute) and the recovery of 2600 units of
enzyme activity in the eluate. Six molar
urea added to the column released the
balance of the bound radioactivity on the
column (1800 counts per minute). No
HSD activity appeared in this eluate
because the enzyme is denatured by the
high concentration of urea.
The total amount of radioactivity
bound (displaced by the urea or ex-
changed with Serratia protein) was 3550
counts per minute, or 2.4% of the total
protein load at saturation with the E. coli
enzyme. Approximately 50% of the
bound protein had been displaced by the
UJ
Q.
C/>
\-
z
3
O
o
<
o
o
<
l 1
1 <
1 1
1 I
1 1 1 1
IOOOO
(
: 1
-
'
^
1
1000
— <o
L
Q _
ki
|
1 .fc
*
s
\ 6M
Q UREA
£
00
b °^
I 5
S
I-
\ h-
\
<0
<
UJ
1
Q
T <D
9 Q
p\ Q
n CC
1
<
! <
h (? <
[b <
00 3
6
3[
! O
1 ft> "J
\ A o
oil - '
b O
/ \ -1
IV 5
1
IOO
"55
1 l\ "O
O \ c
D 1 "^
r \ "o
/ A. c
1 to *
T
"J
\p \ ^
I —
J ^ ^
\
¥
%aaA
^°*fc%co
fl \,r
7 \o *
V
/
i i
1 1
1 1
* w*fi h BACKGROUND
l 1 1
10 20 30 40 50 60
FRACTION NUMBER
70
80
90
Fig. 23. Release of labeled, antibody-bound E. coli protein with Serratia marcescens proteins
and 6 M urea. Open circles, appearance of labeled protein in eluant fractions. Proteins labeled by
growing cells in synthetic medium containing C14 L-phenylalanine.
DEPARTMENT OF TERRESTRIAL MAGNETISM
213
20 30 40 50
FRACTION NUMBER
60
Fig. 24. Release of labeled, antibody-bound Serratia protein with E. coli protein and 8 M urea.
Open circles, appearance of labeled protein in eluant fractions. Proteins labeled as in Fig. 23.
Serratia enzymes prior to the addition of
the urea.
In another experiment a crude extract
of C14 L-phenylalanine labeled Serratia
marcescens was loaded on the antibody
column. Figure 24 shows that the release
of antibody bound Serratia HSD can be
produced by the addition of E. coli strain
Tir 8 enzyme.
The quantity of HSD in Serratia cells
is less than a tenth of that found in the
derepressed strain of E. coli Tir 8. Con-
sequently, much more Serratia protein
must be loaded on the column in order to
approach saturation conditions. Further-
more, when heavy loads of crude extracts
are added to the column, some of the
HSD does not bind and is eluted along
with the bulk of the cellular proteins,
even though the column is unsaturated.
Only when the crude extract is diluted
and the flow rate reduced (1 ml/10
minutes), will complete retention of the
HSD (99%) occur.
In the above experiment, 5 ml of a
crude extract of Serratia cells were
loaded on the column. This preparation
contained 13,000 units of HSD activity
and 670,000 counts per minute of radio-
activity. About 60% of the HSD activity
was retained by the column (7480 units).
After washing with buffer A, a non-
radioactive preparation of E. coli Tir 8
containing 18,400 unite of HSD was
214
CARNEGIE INSTITUTION
added to the column, which retained
12,400 units of HSD activity. The total
bound enzyme (12,400 E. coli plus 7500
Serratia) was 19,900 units, a value in
good agreement with column saturation
results (^20,000 units) obtained when
either E. coli or Serratia enzymes are
used singly. A second and third load of
crude extracts of E. coli Tir 8 (each
18,400 units HSD) resulted in the recov-
ery of 18,700 and 17,600 units, respec-
tively, in the eluate fractions.
The passage of unlabeled E. coli HSD
released 1760 counts per minute, or '—'50%
of Serratia protein. The total bound Ser-
ratia protein was determined by eluting
the balance of the radioactivity bound to
the column with 8 M urea (3580 counts
per minute) and adding this amount to
the E. coli displaced or exchanged pro-
tein. This total (5340) represents 0.8%
of the Serratia radioactivity.
These preliminary exchange experi-
ments suggest that the relative antibody
binding efficiencies of the two enterobac-
terial enzymes are similar. More quanti-
tative determinations will depend upon
comparing Serratia HSD with a strain of
E. coli having about the same ratio of
HSD to total protein content (an E. coli
strain that is not derepressed). These
experiments do demonstrate that ex-
change or displacement does occur, and
they confirm the conclusion that the
number of binding sites on the antibody
column for Serratia and for E. coli HSD
are about equal.
Studies reported last year outlined
some of the characteristics of the AKI-
HSDI antibody column {Year Book 71,
p. 253, Cowie et al*1 1973). The studies
emphasized the capacity of the column
to distinguish between two conforma-
tional states of this enzyme. The form
predominating in the presence of threo-
nine (T form) has antigenic determi-
nants exposed to react with correspond-
ing sites on the antibodies, whereas the
enzyme is less available for binding in
the presence of aspartate or other ligands
that shift the equilibrium from the T to
the R form of the enzyme.42 In the T
form, all of the AKI-HSDI in E. coli is
bound by this specific antibody column.
Studies carried out this year employed
this column characteristic to study en-
zyme evolution in Enterobacteriaceae.
In the presence of L-threonine, HSD con-
tained in crude extracts of Serratia and
Proteus bound to the antibody to the
same extent as that observed with the
E. coli enzyme. This result indicates
that among these bacteria, despite evolu-
tionary divergences (Table 7) , HSD is
more conserved than the DNAs.
Another enterobacterium, Providencia,
contained HSD that did not react at all
with the E. coli antibody column. In ad-
dition, HSD contained in crude extracts
from Aeromonas hydrophila appears to
be in two forms: — 70% binds to the
antibody column, and the rest does not.
This result indicates the usefulness of the
antibody column in the detection and
separation of isofunctional enzymes in
crude extracts of genetically related cells.
Thus, Aeromonas hydrophila seems to
have an HSD with immunological fea-
tures resembling those of enterobacteria
and some material endowed with the
same enzymatic activity but not able to
bind to the antibody column.
SONICATION OF DNA TO PRODUCE FRAG-
MENTS Suitable for Reassociation
Experiments
Bill H. Hoyer, N. R. Rice, and
N. W. van de Velde
Pressure-cell shearing has convention-
ally been used to prepare DNA frag-
ments for reassociation experiments. The
50,000 lbs/in"2 (50K) pressure apparatus
is not used by many laboratories because
of its cost and the mechanical upkeep
required. In addition, pressure cells are
difficult to clean and to decontaminate
when exposed to highly radiolabeled
preparations. Ultrasonic devices are
relatively inexpensive, readily available,
and very easy to clean. Because of these
advantages we have investigated "sonica-
DEPARTMENT OF TERRESTRIAL MAGNETISM
215
tion" as a means of preparing DNA frag-
ments and have found it to be satisfac-
tory. However, a number of variables
influence the size and properties of the
DNA fragments produced by sonication,
and we will discuss some of these.
An obvious variable is the ultrasonic
device and its attachments. Table 8 gives
the specifications and calibration of the
instrument we used. The type and diam-
eter of the tip used, as well as the effi-
ciency of energy transfer, are important
variables. The properties of the instru-
ment we used may be approximated from
these simple measurements and compar-
able micro tip. We have found that
power-meter indications are not consist-
ent from instrument to instrument and
therefore do not include such readings in
our data.
Table 9 gives the results obtained
TABLE 8. Specifications and Calibration of
Sonifier Cell Disruptor *
Setting
Output (gm)t
3
0.21
4
0.9
o
3.05
G
4.55
*Heat Systems Ultrasonics, Inc., Plainview,
Long Island, N.Y.; Model WHO with 1/8 inch
diameter micro tip.
fSonic output was measured on a Met tier top-
loading balance : 50 ml of water was placed in a
100-ml Pyrex No. 1000 beaker on a foam pad
on the balance pan and the micro tip was
immersed halfway (16 mm) between the meniscus
and the bottom of the beaker. The downward
force, in grams, was measured at the indicated
settings.
when time of sonication and volume of
the preparation sonicated are changed.
Both time and volume are important
variables. The size distribution of the
DNA fragments becomes narrower with
increased time of sonication, and the
TABLE 9. Time of Sonication of Native DNA and Sedimentation of the
Single Strands on Alkaline Sucrose Gradients *
Volume and
Setting
Sonication
50K
Time,
Marker
min
"C-DNA
Sonicated
3H-DNAf
Width/ Width/
Distance! Distance § Distance! Distance §
2-ml volume
setting 2
10-ml volume
setting 4.5
1
10
100
1
10
100
4.3
.053
5.0
.086
4.3
.055
3.8
.083
4.1
.055
2.9
.071
4.4
.051
4.6
.092
4.3
.053
4.0
.072
4.6
.049
3.4
.065
*Escherichia coli randomly labeled 14C-DNA (50K sheared) was mixed with
sonicated 3H green monkey DNA and 50/ul of the mixture was layered onto 11
ml (8 cm) linear 5-20% sucrose gradients into 10~2 M EDTA, 0.1 M NaCl,
0.25 M NaOH. The gradients were centrifuged 21 hours at 39,000 rpm in a
Te-41 horizontal rotor. Gradients were sampled from the top with a Buchler
auto-Densiflow and a peristaltic pump.
fFor sonication, the 3H-DNA in 0.1 M NaCl, 0.01 M HEPES buffer, pK
7.4, was placed in either a Cat no. 915 Schwarz/Mann 3.5 ml vial (2 ml volume)
or a Packard glass scintillation vial (10 ml volume). The micro tip was placed
8 mm from the bottom of the small vial and 10 ml from the bottom of the scin-
tillation vial, Cooling was done with magnetically stirred crushed ice-water
mixture, and temperatures within the vials did not exceed 5° C.
{From the meniscus to the peak of the radiolabeled DNA distribution (in
cm).
§The width of the DNA distribution at 1/2 height (in cm) was divided by
the distance the DNA sedimented at the peak of its distribution. This number
indicates the fragment size distribution and may be used to compare that of the
50K marker to that of the sonicated DNA.
216
CARNEGIE INSTITUTION
average fragment size also becomes
smaller. Less energy input is necessary
to fragment DNA in the smaller volume.
It is apparent that any set of conditions
derived to produce DNA fragments of a
consistent size must take into account
time, volume, and geometry. The 50K
fragments (about 450 nucleotides single
stranded) produced by the pressure cell
have a somewhat narrower size distribu-
tion than those produced by sonication.
All sonications described have been
done continuously rather than in bursts
because it was possible to keep the tem-
perature at 5°C or less in vessels im-
mersed in stirred ice- water mixtures.
Both native and denatured DNA may
be sonicated, and the fragment size ob-
tained is influenced by the physical prop-
erties of the DNAs that are processed.
Native DNA fragments (prepared in
neutral solutions) have the advantage
that they may be characterized by their
optical and chromatographic properties.
(Evidence suggests that a small amount
of cross-linking may occur and that son-
icated native DNA maintains its helical
structure.) On the other hand, introduc-
tion of cross-links is less likely in de-
natured (prepared in alkaline solutions)
DNA fragments.
Table 10 indicates that lower energies
are required to fragment single-stranded
alkaline DNA than are required to frag-
ment native DNA comparably. Also,
fragments considerably smaller than the
50K marker may be prepared by both
means, but even smaller fragments may
be obtained from alkaline DNA solutions
at easily attainable energy inputs.
Small DNA fragments, when reassoci-
ated, have lower thermal stabilities than
larger fragments.45 In agreement with
their sedimentation properties, the Te50
values 46 of reassociated DNA fragments
in Table 10 indicate that the DNA frag-
ments prepared in alkaline solutions are
smaller at a given energy input than
those prepared from native DNA sonic-
ated in neutral solutions.
The fragments produced by sonication
at setting levels 2 and 4 (Table 10) in
0.25 M NaOH were reassociated to about
Coti/2, adsorbed to hydroxyapatite in
0.05 M phosphate buffer (PB) at 50°C
and eluted by a linearly ascending con-
centration of PB. The single-stranded
Escherichia coli DNA fragments both
eluted with 0.14 M PB; the reassociated
level-four fragments eluted with 0.24 M
PB ; and the reassociated level-two frag-
ments with 0.28 M PB. Thus, the size of
TABLE 10. Sedimentation and Thermal Elution Properties of Escherichia coli DNA
Sonicated in Neutral and Alkaline Solutions *
Sonicator
Setting
Tesot of
reassociated
sonicated
DNA (°C)
50K
14C-DNA
Sonicated
32P-DNA
Range
Distance!
Width/
Distance §
Distance!
Width/
Distance §
Neutral
Alkaline
2
4
2
3
4
91
89
89
88
87
4.9
5.0
4.8
4.4
5.1
.051
.054
.051
.050
.046
4.4
4.0
3.8
3.3
3.2
.082
.079
.087
.077
.068
*Marker and sonicated DNAs were mixed and layered onto alkaline sucrose gradients as described
in Table 9. The neutral solution contained native DNA 0.1 M NaCl, 0.01 M HEPES (pH 7.4), and
10-3 M EDTA. The alkaline solution was 0.2 M NaOH and 10~3 M EDTA. All sonications were
done in 2 ml with conditions described in Table 9; the time of sonication was 10 minutes.
fThe Teb0 (midpoint of the thermal elution curve) was determined by eluting adsorbed, reassociated,
sonicated DNA from hydroxyapatite. A linear thermal gradient of 0.14 M PB was used as the eluant.
^Distance was determined as described in Table 9.
§Width/distance ratio was determined as described in Table 9.
DEPARTMENT OF TERRESTRIAL MAGNETISM
217
the DNA fragments influences their
phosphate elution properties and may
therefore influence fractionations based
upon PB concentrations.
It has been determined that native
DNA in a 10-ml volume under the con-
ditions described in Table 9 is frag-
mented to about 50K size in 5 min at a
setting of about 4.5. However, it is es-
sential that one calibrate his own ultra-
sonic device to produce the required
DNA fragment size. Volume, geometry,
energy input, state of the DNA, and time
of sonication appear to be the important
variables. The concentration of DNA
and the salt concentration of neutral
solutions appear to have little influence
within reasonable limits.
Expression and Divergence of
Repeated DNA Sequences
Michael J. Byers
Repeated sequences are found in the
DNA of eukaryotic cells, but little is
known of their function. Previous ex-
periments have investigated the expres-
sion and divergence of single-copy DNA
(Year Book 70, p. 376). Experiments
now in progress are directed at the fol-
lowing questions: Are repeated nucleo-
tide sequences transcribed into messenger
RNA? If so, are these transcribed se-
quences more highly conserved than re-
peated sequences as a whole or than those
transcribed from single-copy DNA?
It has been demonstrated in several
laboratories that cytoplasmic mRNAs
have a sequence of polyadenylic acid
covalently attached at the 3' end of the
messenger molecule. The use of affinity
chromatography on oligo-dT cellulose
allows separation of these poly A at-
tached molecules from the bulk of the
cellular RNA. DNA sequences comple-
mentary to this class of RNA molecules
can be isolated by hybridization. The
degree of conservation of these tran-
scribed sequences can be determined by
thermal chromatography of the double-
stranded molecules formed in reactions
with DNA sequences from homologous
or heterologous species.
Efforts to date have concentrated on
developing a method of isolation that
minimizes the probability of enzymatic
degradation and physical loss of RNA
during extraction procedures. One
method that has shown promise is the use
of dithiothreitol (DTT) in the lysing
buffer. As a reducing agent, DTT re-
duces the disulfide linkages of pancreatic
RNase and destroys its activity. DTT,
in conjunction with 8 M urea, solubilizes
most of the protein in a tissue sample,
and CHC13 extraction does not precipi-
tate the solubilized protein. Loss of
RNA at the interface is thus eliminated,
but the solubilized proteins are difficult
to remove. As a compromise, DTT has
been included in a lysing medium without
urea, and resulting precipitates at or-
ganic-aqueous interfaces are reextracted.
This modification shows promise for the
extraction of undegraded RNA.
Labeling of DNA with 125I
M.B. Davis
Comparison of DNAs by reassociation
techniques has been hindered by the
difficulty of obtaining high specific ac-
tivities using in vivo isotopic labeling.
Commerford 47 has recently described a
method for labeling polynucleotides in
vitro with 125I. In order for 12rT-labeled
DNA to be useful in hybridization ex-
periments, it is necessary to determine
the effects of the iodination reaction on
piece size, thermal stability, and ability
to form hybrids.
A series of iodination reactions was
carried out in 0.5 ml of buffer A* (pH 5) ,
with 2.5 X 10"5 M KI (carrier) and
1.5 X 10"4 M TICI3 (oxidizing agent).
In one case the reaction was carried out
in buffer A adjusted to pH 6. Each re-
action mixture contained 100 /xg of
double-stranded E. coli DNA (sonicated
♦Buffer A: 0.1 M sodium acetate, 0.04 M
acetic acid (pH 5).
218 CARNEGIE INSTITUTION
to an average length of approximately tities of non-TCA precipitable counts.
760 nucleotides) and 0.05 mCi of 125I as To test for the presence of precipitable
Nal. The reaction mixtures were heated labeled products other than DNA (or
for various periods of time at 40 °C or RNA), a phenol-chloroform extraction
60 °C, cooled to 0°C, combined with 0.025 was done on one of the samples. Less
ml of 0.1 M Na2S03 (to reduce excess than 5% of the counts remained in the
T1C13) and 0.1 ml of buffer B f (to raise phenol phase. For the further testing
the pH to 9), and heated for 20 minutes described below, DNA from batch 1
at 60 °C (to remove any noncovalently (optimal labeling conditions) was used
bound iodine). The fraction of 125I unless otherwise stated. In all cases, the
which was stably bound to the DNA was unlabeled reference was E. coli DNA
determined by measuring the ratio of that had been sonicated to an average
TCA precipitable counts to total counts length of 760 nucleotides,
for 10A samples of each mixture. The The stability of the final product was
results are shown in Table 11. The per- tested by measuring the TCA precipi-
centage of cytosine that was iodinated table counts before and after heating to
was calculated, using the assumption 100 °C for 5 minutes. There was no sig-
that all precipitable counts were due to nificant difference in the number of pre-
the binding of iodine to cytosine. The cipitable counts before and after the
specific activity of batch 1, measured treatment.
under optimal conditions in a Packard The effect of the iodination reaction on
well-type scintillation counter (approxi- DNA piece size was determined by com-
mately 25% efficient), was 1.5 X 104 paring the sedimentation velocities (at
cpm per ^g of DNA. The specific activ- pH 13) of E. coli before and after iodina-
ity could be greatly increased by increas- tjon> jn one set 0f measurements, the
ing the ratio of 125I to carrier I; during sedimentation velocities of both treated
these reactions the ratio of 125I to carrier ancj untreated DNA were determined
I was approximately 1:500. Commer- optically, using a Spinco Model E ultra-
ford's results also indicate that the spe- centrifuge with ultraviolet optics. The
cific activity could be further increased average length of the iodinated DNA was
(perhaps tenfold) by denaturing the approximately 85% of the length of the
DNA before iodinating. untreated DNA. Another set of measure-
The DNA from each reaction mixture ments was made in an alkaline sucrose
was precipitated with ethanol and dis- gradient on a Spinco Model L ultra-
solved in 0.2 ml of water. The final centrifuge. The relative distances trav-
solutions contained no significant quan- ded by the ^ DNA and by the labeled
t Buffer B: 1 M ammonium acetate, 0.5 M P^ces of DNA were determined by UV
ammonium hydroxide (pH 9). absorbance (OD) and radioactivity
TABLE 11. Extent of Iodination of Sonicated E. coli DNA under
Various Labeling Conditions
Percentage of
I Which is
Percentage of
Batch
TCA
Cytosine Which
Number
Conditions
Precipitable
Is Iodinated
1
60°C 15 min pH 5
7.5
1.3
2
60°C 5 min pH 5
4.5
0.8
3
60°C 15 min pB. 6
2.5
0.4
4
40°C 15 min pH 5
3
0.5
DEPARTMENT OF TERRESTRIAL MAGNETISM
219
E
0.5
1
1 1 1 1
o
1
c
A1 1
O
r a
<£
/o\ •
CO
0.4
-
> 4 1
-
>-
/ ' 1°
r-
l\\
co
/ / lo
2
LU
0.3
—
/ ° \
—
Q
/ ' ,°
_J
J 1 1 \
<C
0.2
—
/ / \ \
—
O
/ o I .
h-
/ / \ ft
Q_
/ \ o
O
0.1
n
o- -
^" 1 1 1
o~
1
r^r^^
0
10
20 30
40 50
60
FRACTION NUMBER
- 10000
-8000
LU
Z>
6000 S
00
-4000
-2000
o
u
0
Fig. 25. Distribution of 125I-labeled (open circles) and unlabeled (closed circles) E. coli DNA
in an alkaline sucrose gradient. A mixture of iodinated and untreated DNA, in 0.2 ml of 0.1 A7
NaOH, was layered on top of a 5%-20% alkaline sucrose gradient in a 5-ml centrifuge tube, and
the tube was spun in a SW39 head on a Spinco Model L ultracentrifuge at 30,000 rpm for 19 hours
at 5°C. The bottom of the tube was pierced, and 67 fractions of 6 drops (0.07 ml) were collected.
After adding 0.6 ml of water to each fraction, the UV adsorbance and counting rate were meas-
ured. Fraction 67, at the top of the tube, contained the lightest molecules.
measurements, respectively (see Fig. 25).
The labeled DNA was found to be ap-
proximately 80% as long as the un-
treated DNA, in general agreement with
the figure given above. There was no
evidence for the existence of large num-
bers of small labeled fragments.
The thermal stabilities of the labeled
and the untreated DNA were compared
by applying a mixture of labeled and un-
labeled E. coli DNA to a hydroxyapatite
column at 60°C in 0.12 M PB and meas-
uring the amounts of radioactivity and
OD which were eluted as the temperature
was raised stepwise. The thermal elution
curve is shown in Fig. 26. It was found
that 59% of the counts and 7% of the
OD passed through the column at 60 °C
without binding. Approximately 15% of
the bound counts were associated with
DNA which was eluted at low tempera-
ture (less than 80°C). The difference in
Te50 (the temperature at which half the
material has been eluted) between the
labeled and unlabeled samples was ap-
proximately 1°C.
The ability of the iodinated DNA to
form stable hybrids with normal DNA
was investigated. A mixture of iodinated
and untreated E. coli DNA (in the ratio
2:100) was heated at 100°C for 5 min-
utes and was incubated in 0.12 M PB
at 60°C to C0t 300. The DNA was then
put on a HAP column in 0.12 M PB at
60 °C, and the bound DNA was eluted
with increasing temperature as before.
Approximately 80% of the counts and
95% of the OD bound to the column at
60 °C. The thermal elution curve is
shown in Fig. 27. Nearly all the labeled
DNA was able to form stable hybrids,
including the labeled DNA which had
previously failed to bind to the column.
The difference in Te50 between the
labeled and unlabeled products was
approximately 1°C.
A further study was made to determine
why more than half of the counts from
the iodinated DNA failed to bind to the
HAP column in 0.12 M PB at 60°C, yet
most were later able to form stable hy-
brids with noniodinated DNA. A sample
220
CARNEGIE INSTITUTION
O
CO
00
Q
O
(T
O
^
Q_
O
_l
<
h-
O
I-
u_
o
<
o
en
LJ
70 80 90
TEMPERATURE, °C
100
Fig. 26. Thermal stability of iodinated E.
coli DNA. A mixture of native (closed circles)
and iodinated (open cirles) E. coli DNA was
applied to a HAP column in 0.12 M PB at
60 °C, and the bound DNA was eluted with in-
creasing temperature. Approximately 59% of
the counts and 7% of the OD did not bind.
Te50 was 93.5°C for the native DNA and 92.5°C
for the iodinated DNA.
of labeled E. coli DNA (iodinated under
the same conditions as batch 1) was
applied to a HAP column at 60°C in 0.12
M PB, and the amounts of radioactivity
60
70 80 90
TEMPERATURE,°C
00
Fig. 27. Thermal stability of hybrids. A mix-
ture of iodinated and normal E. coli DNA
(ratio 2:100) was heat-denatured and incubated
at 60° C in 0.12 M PB to Cot 300. The sample
was then applied to a HAP column at 60° C,
and the bound DNA was eluted with increas-
ing temperature. Approximately 12% of the
counts and 5% of the OD did not bind. Te50
was 93.5°C for the homologous product and
92.5°C for the heterologous product.
and UV absorbance eluting in 0.12 M PB
were measured at 60°, 85°, and 100 °C.
The results are shown in Table 12. Note
TABLE 12. Differences in the Extent of 125I-Labeling for Fractions of
Labeled E. coli Eluting from HAP at Different Temperatures
in 0.12 M PB
cpm//Ag
Elution
OD,
OD,
cpm,
cpm,
of DNA,
Temperature
260 nm
% of total
thousands
% of total
thousands
60°C (nonbind-
ing fraction
0.033
6
136
43
94.0
85°C
0.183
32
58
18
7.2
100°C
0.352
62
122
39
7.9
Total
0.568
100
316
100
12.6
DEPARTMENT OF TERRESTRIAL MAGNETISM
221
that the nonbinding fraction (that elut-
ing at 60°C) contained 43% of the radio-
activity but only 6% of the UV-
absorbing material. This fraction ap-
peared to be labeled more than 10 times
as extensively as the rest of the DNA.
To test for the possibility that the non-
binding counts were associated with
RNA contamination, aliquots of the
60°C and the 100°C fractions were in-
cubated at 60° for three hours at pH
12.5. The decrease in TCA precipitable
counts was the same for both fractions,
approximately 10% ; RNA associated
counts were therefore negligible. It was
therefore concluded that the nonbinding
fraction was single-stranded DNA which
had been labeled approximately 12 times
as extensively as the double-stranded
DNA. This conclusion is consistent with
Commerford's finding that denatured
DNA is labeled more than 10 times as
effectively as native DNA.
The source of the single-stranded
DNA is uncertain. It is unlikely that the
denaturation occurred during the iodina-
tion reaction, since Tm for E. coli DNA
in buffer A was measured to be 81.7 °C,
considerably higher than the temperature
of the iodination reaction mixtures. The
single-stranded DNA was probably pro-
duced during sonication, since, as men-
tioned above, approximately 6% of the
unlabeled sonicated DNA did not bind to
HAP at 60°C in 0.12 M PB. This per-
centage of denatured DNA would prob-
ably not be detected during an optical
melt.
Preliminary studies of the extent of the
single-stranded sections have been made,
using the single-strand-specific nuclease
Si, prepared by the method of Sutton.48
A quantity of 125I-labeled E. coli DNA
(from batch 2) was applied to a HAP
column at 60°C in 0.12 M PB. The
bound fraction (containing 88% of the
OD and 31% of the radioactivity) was
eluted with 0.4 M PB, dialized against
water, lyophilized, and dissolved in 1 ml
of 0.05 M sodium acetate buffer (pH
4.5). To this solution, 10A of 10"2 M
ZnCl2 and 5A of Si nuclease preparation
were added. The solution was then in-
cubated at 50 °C for one hour and
chromatographed on a Sephadex G-100
column with 0.12 M PB. The DNA
which was digested (18% of the bound
fraction, based on OD measurements)
contained 54% of the 125I; the digested
DNA was therefore labeled approxi-
mately five times as extensively as the
intact DNA. Although further studies
of the properties of the Si preparation are
necessary before quantitative results are
possible, it is evident that single-stranded
sections are present in sonicated DNA
and that these sections are labeled con-
siderably more effectively than the
double-stranded sections.
In conclusion, Commerford's iodina-
tion procedure is a convenient method
for obtaining labeled DNA of high spe-
cific activity and of near native size,
thermal stability, and ability to hybrid-
ize. To obtain uniformly labeled DNA
with a well-defined specific activity, how-
ever, precautions are necessary. The
DNA sample being labeled must not
contain both single- and double-stranded
sections, since the single-stranded sec-
tions will be preferentially labeled.
Cooperative Research at
California Institute of Technology
A research program on the mechanism
of regulation of genetic activity is being
carried out at the California Institute of
Technology under the joint direction of
Eric Davidson, Associate Professor of
Developmental Biology at Cal Tech, and
Roy Britten, member of the Biophysics
Section, Department of Terrestrial Mag-
netism.* The work is being done both at
* Other members of the group are : Senior
Research Fellow Barbara R. Hough; Research
Fellows Robert B. Goldberg. Dale E. Graham.
William H. Klein, Berney R. Neufeld, Michael
J. Smith; graduate students Christopher S.
Amenson and Glenn A. Galau ; and research
staff, Margaret E. Chamberlin. Dennis B. Coch-
ran, Charles F. Collett, Michael Henerey, Marta
Rico. Jane Rigg, Vasile Sapatino, Michelle T.
West, and Paul H. Yancey.
222
CARNEGIE INSTITUTION
the main campus in Pasadena, California,
and at the Kerckhoff Marine Laboratory
at Corona del Mar, California. This
program is in its second year and the first
report of this program appeared in Year
Book 71, p. 270.
The focus of interest in these labora-
tories continues to be on the molecular
basis of gene regulation in higher orga-
nisms and the operation of the genomic
apparatus in oogenesis and early devel-
opment. Much of our effort this year
was devoted to analysis of DNA se-
quence organization, since understanding
sequence order is clearly prerequisite to
understanding gene regulation.
maining single stranded are single-copy
sequences. These data indicate that
there are two classes of interspersed
repetitive sequences: a closely spaced set
averaging about 1000 nucleotides' spac-
ing and a more widely spaced set with a
4000-nucleotide or greater spacing. The
repetitive sequence elements themselves
appear to average about 300 nucleotides
in length. These observations are almost
identical to those described last year for
Xenopus DNA, even as far as the spac-
ings and the quantities of DNA involved.
It is astonishing and probably significant
to find such a detailed and quantitative
similarity between the genomes of an
echinoderm and an amphibian.
Interspersion of Sea Urchin Repetitive
and Nonrepetitive DNA Sequences
Dale E. Graham, Berney R. Neufeld, and
Roy J. Britten
Measurements of sequence intersper-
sion have been made for the DNA of the
sea urchin, Strongylocentrotus purnura-
tus, by methods essentially the same as
those used previously for Xenopus ( Year
Book 71, p. 273). Labeled fragments of
DNA of many different lengths were re-
associated with an excess of short (450
long) fragments to a Cot at which most
of the repetitive DNA but no nonrepeti-
tive DNA had paired. The fraction of
the labeled fragments with paired re-
gions, i.e., repetitive sequence elements,
was measured (by binding to hydroxy-
apatite) as a function of fragment length.
About 25% of the DNA is actually made
up of repetitive sequences so that for
very short fragments about 25% of the
DNA binds. Fractions containing repeti-
tive sequences increase rapidly with
length— up to about 55% for 1000-
nucleotide-long fragments. Above that
length the amount increases more slowly
until 80% of 4000-nucleotide-long frag-
ments contain repetitive sequences. In
each case the parts of the fragments re-
size Distribution of Interspersed Repeti-
tive Sequences in Sea Urchin DNA by
Reassociation and Single- Strand-
Specific Nuclease Digestion
Dale E. Graham, Michael Henerey,
Roy J. Britten, and Eric Davidson
In these experiments sea urchin DNA
was sheared to about 2000 nucleotides in
length, denatured, and reassociated to a
C0t which would pair most of the repeti-
tive DNA elements but not the nonrepet-
itive DNA. The DNA preparation was
then digested with S-l nuclease (Ando,
T., Biochim. Biophys. Acta, 114, 158,
1966) under conditions such that duplex
structures that would not be melted at
60 °C, 0.12 M phosphate buffer were pre-
served and less stable structures were
digested by the enzyme. At C0t 20, about
25% of the DNA was enzyme resistant.
One quarter of this DNA (6% of the
total) was in fragments greater than
2000 nucleotides long. A smaller amount
of DNA was distributed in a range of
sizes between 500 and 200 nucleotide
pairs, with the majority of the resistant
DNA fragments present in a peak of
about 300 nucleotide pairs. The single
strands that make up the 300-long du-
plexes were also about 300 nucleotides
DEPARTMENT OF TERRESTRIAL MAGNETISM
223
long. Thus, relatively few of the individ-
ual strands had been attacked at internal
regions of imperfect base pairing. The
long fragments consisted of single-
stranded DNA with the same size dis-
tribution as the original sheared DNA
preparation.
These size distributions show that
there are extensive regions of repetitive
sequences which we have termed "clus-
tered repetitive regions." This DNA
gives rise to the long, enzyme-resistant
DNA fragments. The short (300 nucleo-
tide, average) enzyme-resistant fragments
represent the repetitive sequence ele-
ments that are interspersed throughout
the genome. Their size, when isolated by
S-l nuclease treatment, is about the same
as that deduced for the short-period
interspersed repetitive sequences men-
tioned above.
Similar treatment with S-l nuclease of
Xenopus DNA reassociated to a repeti-
tive C0t confirmed that, in this organism
also, the repetitive sequence element is
about 300 nucleotides long, measured
either as duplex or after denaturation.
Samples of each of the size classes of
enzyme-resistant, reassociated, double-
stranded sea urchin DNA were melted in
the spectrophotometer. As expected, the
hyperchromicity was approximately that
of native DNA. The thermal stability
of the enzyme-resistant regions was cor-
related with their length. The fraction of
greatest length (excluded peak) melted
within one or two degrees of the tempera-
ture at which long native DNA melts.
Each succeeding smaller size fraction
melted at a lower temperature, down to
the 300-base-pair, long-peak regions,
which melted as much as 15° below na-
tive long DNA. The reduction of melting
temperature observed is too great to be
due simply to the length of the duplex
region and must result from increasing
sequence divergence. There is a clear
correlation between the amount of se-
quence divergence among the repetitive
sequences and their length.
Hyperpolymers and Another Approach to
Repetitive Sequence Organization
Dale E. Graham and Roy J . Britten
It has been known for many years that
the reassociation of repetitive DNA leads
to the formation of extended structures
termed "hyperpolymers." This happens
because the DNA fragments are ran-
domly broken and two complementary
strands that pair seldom terminate at the
same place. Thus, after the first pairing
there remain single-stranded ends which
can pair again with other fragments. The
resulting structures may be very large
after a reassociation reaction is near
completion.
Recent observations indicate that in-
terspersed repetitive sequences do not
form extended hyperpolymers. In these
tests 300-nucleotide-long fragments were
dissociated and incubated to a C()t 50 or
100 times that required for half reasso-
ciation. The size of the reassociated
DNA was assayed on an Agarose A50
column. Treated in this way, T4 viral
DNA is mostly excluded, indicating that
most of the DNA is incorporated in
structures containing at least 20 frag-
ments. The same experiment was done
with total sea urchin DNA and with iso-
lated highly repetitive and middle repet-
itive DNA fractions. In each case very
little of the DNA was excluded from the
A50. The excluded DNA quantity cor-
responded to that known to be in clusters
from the measurements described above ;
clustered repeated DNA is expected to
form hyperpolymers. The remainder of
the duplex structures consists of only
pairs or tetramers of the original frag-
ments. This result confirms the fact that
most of the repetitive sequences are pres-
ent as interspersed short-sequence ele-
ments whose length is about that of our
fragments, i.e., about 300 nucleotides.
Furthermore, it shows that these ele-
ments are internally nonrepetitive with
homologous elements beginning and end-
ing at homologous positions.
224
CARNEGIE INSTITUTION
General Assessment of Advances in the
Knowledge of DNA Sequence
Arrangement
Dale E. Graham, Berney R. Neujeld, and
Roy J. Britten
During the year, a number of related
observations have been made of sequence
arrangement in the DNA of several
species, including the marine worm
Urechis, a sea urchin, an amphibian and
a mammal. Although there are differ-
ences in genome size and repetition fre-
quency among these animals, neverthe-
less there are common features and a
pattern emerges from the comparisons:
in all of these species (1) the majority of
the repetitive sequences are interspersed
with nonrepetitive sequences; (2) a
minor component of clustered repetitive
sequences exists; (3) the clustered re-
gions extend for at least several thousand
nucleotides, while (at least for the three
deuterostomes) the length distribution of
the interspersed repetitive sequences has
a peak at about 300 nucleotides; (4) the
clustered sequences display relatively
precise sequence repetition, and the inter-
spersed sequences show divergence rang-
ing from 5% to 20% or more.
This evidence shows that there are
general patterns of organization of the
repeated and single-copy sequences in
living systems and helps to establish the
foundation for our current investigations
of nuclear RNA and polysomal messen-
ger RNA. These patterns appear almost
certain to lead us to a deeper understand-
ing of the functional organization of the
genome.
GEOPHYSICS
L. T. Aldrich, L. Beach, K. D. Burrhus, R. C. Fletcher, A. Cuyubamba, S. E. Forbush,
B. Grauert, S. R. Hart, A. Hofmann, D. E. James, A. T. Linde, H. Okada,
I. S. Sacks, N. Shimizu, D. Simoni, T. J . Smith, and A. Snoke
Collaborators
C. A . Andersen, J. Bannister, F. R. Boyd, C. Brooks, I, Casaverde, D. Comajord, G. L. Davis,
M. L. Crawford, S. del Pozo, D. W. Evertson, A. Flores, A. A. Giesecke, B. J. Giletti,
L. Hall, J. R. Hinthorne, T. E. Krogh, I. Kushiro, Y. Motoya, J. Mattinson,
L. Ocola, G. Olajsson, R. Quiroga, J. E. Ramirez, A. Rodriguez B.,
R. Salgueiro, J. G. Schilling, M. G. Seitz, H. Sigtryggsson,
S. Suyehiro, L. Tamayo, E. Triep, M. A. Tuve,
V . Volponi, M. E. Wagner, and Y. Yamagishi
Introduction
L. T. Aldrich
For several years there have been
comments in this introduction extolling
the opportunities at Carnegie for express-
ing one's creative concerns in earth sci-
ence on a global scale. We have implicitly
and, on occasion, explicitly expressed
both our appreciation for these opportu-
nities and our hope that we are using
them fruitfully. This year I would like
to outline briefly some of the collabora-
tive efforts making up the fabric of the
work in earth science at the Department.
The most extensive program is that of
Selwyn Sacks, whose broad-frequency-
band, high-dynamic-range seismometers
occupy a seven-station net, ranging from
New Guinea to Iceland with two stations
each in Peru and Japan. He even has one
on the DTM campus! Without the help
of devoted scientists and technicians in
these remote locations, these observations
would be impossible. Among these col-
laborators, Dr. S. Suyehiro in Japan
should be specially mentioned for his
continued interest. The contributions of
another of our collaborators, Ing. D.
Simoni of Peru, have been recognized by
his appointment as a Staff Member of
the Department.
DEPARTMENT OF TERRESTRIAL MAGNETISM
225
Dr. Suyehiro and Mr. Dale Evertson,
of the University of Texas, Austin, are
full partners with Sacks in the installa-
tion and testing of a tripartite net of
Sacks' volume strainmeters in Japan, the
fruits of which are yet to be appreciated
fully. David James includes among his
collaborators not only those listed in his
reported geochemical studies below but
also numerous geologists in the geological
surveys of Peru and Bolivia. The inter-
est of local men in these studies provided
physical and logistic support during the
collecting activities this year. My elec-
trical conductivity activities are based
on the commitments of Mateo Casaverde,
Lupe Tamayo, John Bannister, and Rey-
naldo Salgueiro — all of whom are install-
ing and operating magnetic variographs
in their home countries as this is written.
Stanley Hart has for several years
been interested in samples of ridge ba-
salts as indicators of process in the mod-
ern theory of global tectonics. Such
samples are collected only by large-scale
oceanographic expeditions. A measure of
the value he places on these samples is
expressed by the fact that he will act as
co-chief scientist on such an expedition
in the coming report year.
Finally, James, Sacks, Linde, Simoni,
and I were a part of a five-nation group
that organized and executed a major geo-
physical study of the region between
2° S and 4° N latitude, and 74° and 82°
W longitude this year. We and our col-
leagues in North and South America were
joined by a German team of specialists
in explosion seismology to complete
seismic studies all over the area and
magnetic and gravity observations at
sea. Of special importance to the suc-
cess of our plans were the efforts of Dr.
J. E. Ramirez, S. J., in Colombia.
It is anticipated that we will all be
interested in continuing our far-flung
activities in earth science. We again ex-
press our appreciation for the support of
these studies from both the Carnegie
Institution and the National Science
Foundation.
Cosmic-Ray Research
S. E. Forbush and L. Beach
The variation, with -period twice that
of the sunspot cycle, in the cosmic-ray
diurnal anisotropy. As described in pre-
vious Annual Reports, the annual means
of the total cosmic-ray diurnal aniso-
tropy vector from Carnegie Institution
of Washington ionization chambers are
well represented as the sum of two com-
ponents, V and W, where V and W are in
the asymptotic directions 90° and 128°
east of the sun, respectively. The com-
ponent V varies with the period of the
sunspot cycle and is well correlated with
magnetic activity. When data were avail-
able only through 1965, annual means of
W were found to be well fitted by a sine
wave with a period of two solar cycles,
which amounted to twenty years during
the interval covered by the data. This
wave passed through zero in 1958, which
is near the time that Babcock found,
from regular measurements begun in 1953
at Mount Wilson Observatory, that the
north polar magnetic field of the sun
changed from positive to negative.
The interval between successive sun-
spot maxima (or minima) varies from
about 9 to 13 years. Thus, within the
interval from 1937 to about 1965, it is
fortuitous that this sunspot period re-
mained close to 10 years and that a wave
with a period of 20 years fitted W.
Thus, not unexpectedly, the observed
points (squares) in Fig. 28 after about
1966, definitely depart from the indicated
sine wave, Ws, that fitted quite well the
observed points (circles) until about
1965. Figure 28 shows a "guesstimated"
curve, Wg, fitted to observed values after
1965. WG shows that the component W
changed sign shortly after the beginning
of 1971.
From continuing measurements at
Mount Wilson Observatory, Howard
found that the north polar field of the
sun, having been negative since 1958, be-
came significantly positive about the
226
CARNEGIE INSTITUTION
OS
6 2>
~ E
5 -
-5
O /r
\ O
," ' \
' \
/ \o
/
1940
/
/
/
/
/ o
/o
/ °
■Wc
V) o
\
\
I
I
I °
/
/ o
r
O
LEGEND \
o, o .Annual means of W *
Wg : Sine wave with period
10 years \
Wq :"Guesstimated"
variation in W
after 1966
1950
I960
1970
Fig. 28. Annual means, 1937-1972, of the component W of the diurnal anisotropy in the asymp-
totic direction 128° east of the sun.
middle of 1971. Thus the two observed
reversals in the sun's polar magnetic field
since observations began in 1953 occurred
near the time of the last two reversals in
the sign of the component W of the
cosmic-ray diurnal anisotropy.
Although the mechanism responsible is
still not understood, this variation of W
with a period twice that of the sunspots
is probably the first terrestrial effect ob-
served to be related to the variation in
the direction of the sun's polar magnetic
field. The components V and W combine
to give an excellent fit to the observed
annual mean diurnal anisotropy. It is
evident that this satisfying aspect could
only have been uncovered from a long
series of records.
Observations and reduction of data.
Cosmic-ray ionization chambers were
operated throughout the report year at
Huancayo, Peru; at Fredericksburg, Vir-
ginia; and at Christchurch, New Zealand.
Scalings and reduction of records have
been maintained on a current basis for
these three stations. The reductions have
been greatly facilitated by the use of the
IBM 1130 computer.
Cooperation in operation of cosmic-ray
meters. Grateful appreciation is ex-
pressed to the National Oceanic and
Atmospheric Administration and the
staff of its magnetic observatory at
Fredericksburg for efficient operation of
the cosmic-ray meter there during the
past report year, and to the Government
of Peru and the Director and staff of the
Instituto Geofisico del Peru for making
cosmic-ray records from Huancayo avail-
able. Grateful appreciation is also ex-
pressed to the Director and staff of the
Geophysical Observatory at Christ-
church, New Zealand, for excellent main-
tenance of the equipment there and for
the fine records obtained.
A Comparison of the Anelasticity
Structure between Western
South America and Japan
/. S. Sacks and H. Okada
Although the driving forces for sea-
floor spreading and continental drift are
largely unknown, these motions are cer-
tainly influenced by the viscosity struc-
ture of the crust and upper mantle and
perhaps also by that of the lower mantle.
DEPARTMENT OF TERRESTRIAL MAGNETISM
227
Viscosity in the earth is difficult to de-
termine directly but may be estimated
from the anelasticity (Q_1) even though
at this time the formal relationship be-
tween these quantities has not been
established. Anderson (1966)49 and
Sacks (1972) 50 have proposed empirical
relationships.
South America and Japan, the two re-
gions studied, have broad similarities
(ocean trenches, dipping seismic planes,
volcanoes) and important differences
(South America is a continent, whereas
the Japan region is an island-arc type
structure). In addition to these differ-
ences, the seismicity is different in the
two regions. In Japan the seismicity is
restricted to a dipping seismic plane and
shallow activity west of the ocean trench.
In South America this dipping seismic
plane is more (northern Chile) or less
(central Peru) obvious, but the wedge
above the dipping plane contains earth-
quakes, relatively few in northern Chile
but many in central and northern Peru.
In Japan the seismic activity is con-
tinuous down the dipping plane, though
the most energy is released by the shal-
low quakes and the least energy by
events of intermediate depth.
Above the deep earthquakes in South
America, however, there is an aseismic
zone which extends at least from Ecuador
to central Argentina (some 30 degrees of
latitude) and covers the depth range 350
to 530 km. The energy released by deep-
focus events is comparable to that from
shallow events. Thus it is likely that a
comparison of their anelasticity struc-
tures should yield information pertinent
to large-scale earth movements.
In order to make valid comparisons of
Q structures, the fewest possible assump-
tions were made and identical techniques
were used in the two regions. Since the
radiated spectrum of an earthquake de-
pends on its magnitude and location, this
variable has been eliminated by use of
spectral ratio techniques for reliable
quantitative Q determinations (Sacks,
1968). 51 Nevertheless, spurious effects
(e.g., different crustal transfer functions
at different stations) necessitate calcula-
tion of the spectral ratios over a wide
frequency range. The results presented
here derive from recordings by specially
developed wide frequency range seismo-
graphs (Sacks, 1966) r>2 which have been
installed in the two regions.
Summary of results. Q results from
the South American region, areally
(Sacks, 1969) 53 and as a function of
depth (Sacks, 1971) ,n4 have been re-
ported previously. In essence, the South
American Q structure consists of a
high-Q continental block about 350 km
thick overlying a variable low-Q layer
of uncertain dip. The Qs of the thick
block increases with depth from a value
of 500 for the upper 100 km to 1500 at
350 km. There are, however, important
variations in this block. Some values in
the northern Chile— northwestern Argen-
tina region are less than one-half of those
farther north. The presence of a down-
going lithospheric slab has not yet been
detected from the Q study or from time
residual studies. This does not mean that
such a slab does not exist in the South
American region but that its Q structure
is similar to that beneath the continent
since possibly neither region contains sig-
nificant partial melt. We believe that
deep continental roots are not confined to
South America though anelasticity evi-
dence is confined to this continent at
present. Studies of surface wave disper-
sion across shield areas in Africa and
South America by Alexander (1972)"'
show an absence of velocity reversal in
the uppermost mantle (circa 200 km) in
both cases. A velocity reversal is found
commonly in oceanic areas and occurs at
depths of approximately 70 km.
The situation beneath Japan is some-
what different. Figure 29 shows two
seismograms from earthquakes at depths
of about 285 km. The shear arrival at
MAT in Japan shows little high-fre-
quency energy, i.e., the path from the
228
CARNEGIE INSTITUTION
MAT, 29 AUG. 1970, M*5.2, DEPTH = 284KM, A=7.I0
0.1 sec
CUZ, 16 DEC 1968, M«4.l, DEPTH* 283, A=6.4
Fig. 29. Analogue filtered seismograms indicated the effects of anelasticity in the upper 280 km.
Both seismographs, MAT in Japan and CUZ in Andean South America, are situated well above
the dipping seismic plane (160 to 200 km) in roughly equivalent positions relative to the tecton-
ics. The earthquakes are at the same depth, circa 285 km. The MAT seismogram (top) shows
that the shear arrival, while of similar amplitude to the compressional (P) at lower frequencies
(upper traces), has little energy at the higher frequencies. In contrast, the record from CUZ
(bottom), shows considerable energy at the high frequencies for the shear arrival. This is because
the region above the dipping seismic plane in South America is generally of high Q.
DEPARTMENT OF TERRESTRIAL MAGNETISM
229
earthquake to the seismograph must be
low-Q. The shear arrival recorded at
CUS in South America, which is simi-
larly situated relative to the dipping
seismic plane and the volcanic ranges,
shows high frequencies and therefore a
high-Q path from the same depth. The
gross features of the anelasticity struc-
ture can be seen by comparing the areal
distribution of average Q from the deep
earthquakes to the surface. The results
can be seen in Fig. 30. For Japan, mean
intensity contours from deep earth-
quakes, determined by Utsu (1967) ,56
have been plotted on a map showing the
volcanic front. There is substantial simi-
larity between the shape of the intensity
contours and the form of the volcanic
front. The correlation between the inten-
sities and average Q values has been con-
firmed by a study similar to the one done
for the South American region. No ex-
ceptions were found to the observation
that there are high-Q paths to the seis-
mographs on the east coast and low-Q
paths to the stations on the west coast,
i.e., to those 150 to 200 km above the
seismic plane.
In South America, on the other hand,
high-Q and low-Q paths from deep earth-
quakes are not related in any simple way
to the tectonics as expressed by surface
features. This is because the average Q
of these paths is dominated by variations
in a low-Q region between depths of
350 and 530 km. The structure between
the deep earthquakes and the surface is
therefore more regular beneath the Japan
region than beneath South America. The
Q structure of the dipping seismic zone
(descending slab) was determined mainly
from the broad-band seismograph at
Kamikineusu (KMU) on the Pacific coast
of the north island, Hokkaido. All paths
from earthquakes in the dipping plane
north of central Honshu to this sta-
tion have high Q. By studying earth-
quakes at various depths in the seismic
plane, it was found that the average Q
to the seismograph increases with depth
Average
ntensrty Anomalies
Caused by
Deep Earthquakes
Quaternary
Volcanoes •»•
and
Trenches Jzw
Fig. 30. Areal distribution of the average Q structure between the deep earthquakes and the
surface beneath Japan and South America. Intensity anomaly contours (after Utsu, 1967) are
drawn. Intensity anomalies reported from deep earthquakes (after Utsu, 1967) have been con-
toured on a map of Japan (left). The higher numbers indicate greater vibration and therefore
higher average Q. The correlation of the intensity contours with the form of the volcanic front
is apparent. Beneath South America no such simple pattern exists (Sacks, 1969). The variations
in average Q are caused by variations in a region of generally low Q in the depth range 350 to
530 km.
230
CARNEGIE INSTITUTION
WEST
EAST
SEISMIC ENERGY
FROM OISTANT
EARTHQUAKE
Fig. 31. Sketch showing ray paths and seismic geometry beneath Japan. The two ray paths
shown were found to have identical average Q, even though the path to MAT is much longer in
the aseismic wedge than that to KMU. Therefore, the absorption in the section "A" in the aseis-
mic wedge must be exactly compensated by that beneath the seismic plane "B," i.e., the average
Q structure on either side of the high-Q seismic plane must be the same.
of the earthquake. Inversion to actual
Qp shows that it increases from a value
of approximately 1000 near the surface
(100 km), to 2000 at about 200 km, to
more than 3000 below 400 km. We do
not yet have a good constraint on the
thickness of this dipping high-Q zone,
but, from a simple ray path considera-
tion, it must be either fairly thick (more
than 100 km) or have special interfaces
at upper and lower boundaries to effi-
ciently reflect the seismic energy at near
grazing incidence.
The lithosphere beneath Japan has not
been studied yet, but the appearance of
seismograms of elastic waves traveling in
it suggests that it is of high Q.
The Q values of the aseismic wedge
between the surface and the dipping seis-
mic plane were determined mainly from
the broad-band seismograph in Matsu-
shiro, central Honshu (MAT) , again by
studying earthquakes situated at various
depths in the seismic plane. The average
Qp of this region was found to be 400.
We have not yet determined how much
structure there is in the anelasticity of
this aseismic wedge. It is of interest to
note that no exceptionally low-Q regions
were found in the vicinity of the vol-
canoes. Asamayama, one of the nearest
(30 km) to the seismograph (MAT), is
presently active. This is consistent with
no observations of low Q in the vicinity
of volcano El Misti (near Arequipa) in
southern Peru, South America. Farberov
et al. (1971 )57 actually measured the ab-
sorption of elastic waves beneath vol-
canoes in the Kamchatka region and
found a rather narrow high-absorption
zone. This suggests that there is no
great magma pool beneath the volcanoes.
Instead, there must be a narrow pipe
transporting the magma from depths.
Comparison of the Q values on either
side of the dipping seismic plane yields
the interesting result that they are iden-
tical. Figure 31 shows a schematic of the
method used for this comparison. Spec-
tra from deep events below Fiji were
compared at the two broad-band instal-
lations in Japan. The observed spectra
whose ratio is shown in Fig. 32 were
identical over a considerable frequency
range (0.1 to 5.0 cps), indicating that
absorption in the longer path to MAT in
the aseismic wedge is exactly compen-
sated by that in the path to URA below
the seismic plane. This is confirmed by
comparison of spectra of near vertical
shear waves ScS at MAT and URA. The
spectra are essentially identical in slope.
DEPARTMENT OF TERRESTRIAL MAGNETISM
231
10.0
o
h-
<
<
\-
u
Ld
CL
V)
1.0
0.1
-
_
SPECTRAL RATIO
9 Oct. 1967 Fiji
—
P(MAT) / P(KMU)
M(ISC)= 6.2 H = 654 KM
-
1/3 OCTAVE AVERAGED
—
o
o^o o ABOVE NOISE
~ o
o
0 00
CD CcF° O
o
o o
° °o o
° o o ° °
° \ • BELOW NOISE
0 o
Do &
°° o °
-
-° ocP ^o
o
—
O o
o
- °o°
_ o
—
•/' ■ W- :&
ajy^^
1*
/ ■./fj^zmimb
\' "'
_
X mA.t^^ 'K
1 1
JAP
-
from
Fiji
1
1
1
1.0
2.0
FREQUENCY
3.0
(CPS)
4.0
5.0
Fig. 32. The spectral ratio as a function of frequency of the compressional wave (P) arrival at
MAT to KMU from a deep earthquake in the Fiji region. The ray path geometry is shown in Fig.
31. A gross slope of the spectral ratio would indicate a difference in the Q values of the two paths
compared. Undulations in the curve are caused by dissimilarities in the crustal structure of the
two seismograph stations.
Thus, at least for the upper 250 km or so,
materials which may be released from the
downgoing slab are not of a type or
quantity to affect the gross anelasticity
of the wedge above the slab.
At greater depths, however, much
lower Q values are found. Spectral ratios
of PP vs. P from the Fiji deep earth-
quakes recorded at MAT and URA gave
Qp values of 100. A deep earthquake in
the Okhotsk Sea region recorded at the
broad-band seismograph at Port Mores-
by, New Guinea (PMG) also gave a Qp
of 100 from a PP vs. P spectral ratio cal-
culation. The spectral ratio of sScS vs.
ScS at MAT from the same earthquake
yielded a Qs of 50, which is in good agree-
ment with the Qp values. Above the
South American deep earthquakes, Qp is
350 from PP/P and Qs is 160-200 from
sScS/ScS and also from ScS/SKP. These
generally higher Q values are due to the
very thick lithospheric section (or ab-
sence of the low-Q wedge) which has a
high Q for the 350 km below the surface.
Very low (70-200) Qp values were found
in the 350 to 530 km aseismic zone be-
neath South America, but these appear
to be of lesser extent.
Discussion. The anelasticity struc-
tures of the two regions are compared in
Fig. 33. The similarities allow us to
make certain conjectures. We have found
that Qp values in regions which have
earthquakes, even if the seismicity is low,
are above 1000. These regions must be
able to store strain energy, i.e.. have very
high viscosity and probably very little, if
any, partial melt. The Q of such a mate-
rial increases with pressure (depth) and
232
CARNEGIE INSTITUTION
cus
Trench ^ ^^
200 400V^600 800 - 1000 1200 1400 KM
Fig. 33. The QP structure beneath Japan (left) and South America (right). Dots indicate areas
of high Q (1000-3000); slant lines, intermediate-Q values (300-500); and cross-hatching, low-Q
values (50-100). The small numbers in the dotted region indicate earthquake hypocenters. Seis-
mograms to the stations indicated, MAT and CUS, are shown in Fig. 29. The region above the
dipping seismic plane is of intermediate Q in Japan, whereas in South America it has high Q. The
high-Q slab in Japan is continuous down to the deepest earthquakes; in South America there is
a substantial decrease in Q below 350 km even though there are earthquakes 600 km deep.
is probably not much affected by rock
type. The increase of Q with depth in the
dipping slab beneath Japan was similar
to that in the continental block beneath
Cuzco in Peru, which is well east of any
possible slab. Absence of earthquake
activity does not necessarily imply low Q
since sufficient tectonic stress must also
be present.
Qp values for the aseismic wedge be-
neath Japan are as high as 400, and we
assume that the viscosity is too low to
store strain energy, presumably due to
some partial melt. Far lower Q values
are possible, of course, and have been
found above the deep earthquakes in Fiji
and Japan and in pockets in South Amer-
ica. The tectonic significance of the very
low Q's above the deep earthquakes is
not clear to us at this stage, but its com-
mon existence suggests that it is in some
way fundamental.
Some tectonic implications suggested
to us by these results are illustrated in
Fig. 34. For island arcs, such as Japan,
the normal plate tectonic model of a
relatively thin (—70 km) slab dipping
into the upper mantle from the vicinity
of an oceanic trench in an "oceanic plate"
(or possibly "oceanic plate— marginal sea
plate") interaction is adequate and is
shown on the right-hand side of the
Fig. 34. Tectonic implications of the Q struc-
ture. Oceanic lithosphere is 70-100 km thick,
subducting at trenches as confirmed by the Q
structure beneath Japan. Continental litho-
sphere, however, is much thicker (350 km be-
neath South America) and penetrates below
what is traditionally considered asthenosphere
(the weak layer). It is probable, therefore, that
continental roots are more strongly coupled to
the higher rigidity (mesophere) region of the
mantle than are the oceanic plates. Whole
mantle convection would therefore drive con-
tinents strongly (e.g., South America), just as
lack of whole mantle convection beneath a
continent (possibly Africa) would tend to
anchor it.
DEPARTMENT OF TERRESTRIAL MAGNETISM
233
figure. However, for "oceanic plate-
continent" interactions, the difference is
that the continent has a very thick (350
km), high rigidity "root." The shape of
the subduction zone is forced by the
shape of the continent before subduction
starts and will in general not be the ac-
curate shape found commonly in ocean-
ocean interactions.
Empirical Models for Anomalous
High-Frequency Arrivals from
Deep-Focus Earthquakes in
South America
A. Snoke, I. S. Sacks, and H. Okada
Anomalous high-frequency arrivals
(AHFA) from deep-focus earthquakes
have been observed for epicentral dis-
tances of 13-21° in Japan,58 Tonga,59-60
and South America.61,62 Their arrival
time is up to a minute later than the
direct S; their path has a higher Q than
that of direct S; their duration is longer
than that of direct P or direct S; and
they have an apparent phase velocity of
4.7 km/sec. If Jeffreys' velocity-depth
model for the mantle is used, a phase
velocity of 4.7 km/sec corresponds to an
S phase which has a maximum possible
depth of 270 km. The fact that the rele-
vant foci are as much as 300 km below
this depth and that the path has anomal-
ously high Q implies a radically different
path for an AHFA from those for normal
arrivals. A feature common to all the
geographical regions mentioned above is
that lithosphere descends towards the
focal region. On the basis of seismic ac-
tivity and Q studies, the lithosphere is
believed to be continuous down to the
relevant focal depths in all regions except
in South America where there is no evi-
dence of lithosphere between a depth of
about 350 km and the focal depths of
540-600 km.
Isacks and Barazangi60,62 have in-
troduced a model which satisfies the
travel time and Q data for AHFA ob-
served at NNA. In this model the de-
scending lithosphere must be taken as
continuous down to the focal depths. Ac-
cording to their model, an AHFA results
from a phase which propagates from focus
to station by traveling inside the litho-
sphere as a guided wave and then con-
verting to an Sn phase near the earth's
surface. The wave guide is obtained by
postulating that the upper surface of the
lithosphere is a perfect reflector for *S
phases and by assuming a velocity gradi-
ent increasing from this surface inwards.
Assuming the uniqueness of their model
in its ability to fit the data, Isacks and
Barazangi have used their model as
evidence for the existence of structure
and phenomena heretofore not observed
in South America, such as the continuous
lithosphere, a totally reflecting upper
surface of the lithosphere for S phases
and the corresponding wave guide.
We have also been studying South
American AHFA and are seeking a model
which can provide a satisfactory expla-
nation for them. Our data come from
the DTM broad-band stations CUS and
TRU and from the WWSSN stations
NNA and ARE (see Table 13). The
NNA data are the same as those used by
Isacks and Barazangi, and the ARE
data are cited by them but not used. The
locations of these stations and the loca-
tions of events that gave rise to clear
AHFAs observed at at least one station
are shown in Fig. 35. A graph of travel
time versus epicentral distance for ob-
served AHFA is given in Fig. 36. For
five of these twelve events, AHFAs are
observed at two stations. For some
events, including three of these twelve,
there are potentially usable data from
stations for A less than 13° or greater
than 21°, but in these cases we observe
no AHFA near the anticipated travel
time. It is difficult to distinguish between
the direct S and an AHFA at the shorter
distances because the time difference be-
tween S and the AHFA is so short, while
for the longer distances there is simply no
observable high-frequency arrival within
234
CARNEGIE INSTITUTION
TABLE 13. AHFA-Producing Events as Plotted on Figure 36
Date
Magnitude Depth, km Station
AHFA
Travel Time,
A, degrees sec
Dec. 9, 1964
5.9
586
NNA
20.06
481
Mar. 5, 1965
5.5
573
NNA
19.63
472
May 13, 1965
5.1
589
NNA
14.52
351
Dec. 11, 1965
4.2
579
ARE
13.41
325
Dec. 20, 1966
5.7
586
CUS
14.91
361
NNA
19.05
456
Jan. 17, 1967
5.6
588
CUS
15.92
380
NNA
19.91
483 ±2
Sept. 9, 1967
5.8
578
NNA
20.27
486
Jan. 31, 1968
4.9
580
ARE
13.57
327
CUS
16.24
382
May 11, 1968
5.2
602
ARE
14.53
347
Aug. 23, 1968
5.8
537
NNA
16.14
383
TRU
20.23
491 ±5
5.2
541
NNA
16.05
382
July 25, 1969
5.5
579
CUS
14.40
352
NNA
18.59
450
30 S
80 W
Fig. 35. Map showing the locations of stations and AHFA-producing events in South America.
Also included are projections of the conversion region for Models A and B (see text) and projec-
tions of seismic activity for depths of 100-250 km (numbers) and for depths of greater than 500
km (open triangles). The AHFA-producing events are solid triangles.
DEPARTMENT OF TERRESTRIAL MA
500
GNETISM
235
450
400-
o
350
300
1 1 1
EVENTS
i 1
' ' 'l /
• NNA
X ARE
y
■ TRU
A CUS
/*'
/ x NORMAL S
( DEPTH = 580 KM)
X* /
r i i
i i
1 1 1
14 16 18
TOTAL A IN DEGREES
20
Fig. 36. Total travel time versus A for AHFA
observed at NNA, ARE, CUS, and TRU. The
straight line drawn through the data points has
a slope corresponding to a phase velocity of 4.7
km/sec.
a couple of minutes of the arrival of
direct S.
We have attempted to formulate
models that require no geological struc-
ture not observed independently. Our
two most successful models to date have
the following common features: a phase
(phase 1) travels from the focus to some
point where it is converted to another
phase (phase 2) which travels to the ob-
serving station. The paths of these
phases are restricted by the condition
that the resultant arrival should have a
path of higher Q than that of the direct
S. We have been guided by Sacks' find-
ings 63 of a low-Q region in South Amer-
ica between depths of 350 km and 500
km. Accordingly, phase l's trajectory
through this low-Q region should be
steeper than that for direct S, and phase
2 should have a maximum depth of less
than 350 km. A long duration for the
AHFA can be explained by the possibil-
ity of nonplanar paths and irregularities
in the interface producing the conversion.
This interface is discussed below. Our
two models differ in the proposed paths
for phase 1 : In Model A, phase 1 travels
directly from the focus to the conversion
point; while in Model B, phase 1 is re-
flected from the earth's surface before
arriving at the conversion point (see Fig.
37) . In both models we have left the
N DEGREES
NNA
STATION
FREE -SURFACE
REFLECTION
)/ 15
* 300
x
Q_
600
0°
:'*fe%i/T\_~l
f
-*j?w;>
[iA. ' .
15°
nnaT;
o
30° S
l^fe™ to
FOCUS
Fig. 37. The paths according to Models A and B for an AHFA observed at NNA from the
event on May 13, 1965, superimposed on a seismicity plot for the region between the event and
NNA.
236
CARNEGIE INSTITUTION
conversion process unspecified at present.
We chose as the first test for the ac-
ceptability of a model its ability to pro-
duce total travel times which fall along a
straight line whose slope is the observed
phase velocity for an epicentral range of
13-21° (see Fig. 36). In Table 14 the
results for our most successful models are
presented for a Model A case of a P
phase converting to an S phase and for a
Model B case of a P phase that remains
a P phase after reflection from the earth's
surface but is then converted to an S
phase at the conversion point. The final
column in Table 14 gives the time differ-
ence between the total travel time cal-
culated according to the model (using
Jeffreys' velocity-depth model for S
phases and Herrin's model for P phases)
and the time calculated from the straight
line through the data points in Fig. 35.
We note that for both models this time
difference is only a few seconds (out of
about 400) throughout the range of in-
terest. This result is fairly insensitive to
the choice of the particular velocity
models chosen or to the choice of the
depths for the focus and the conversion
point; the depths used in Table 14 are
589 km and 150 km, respectively.
If we assume that the boundary of the
lithosphere is near the boundary of the
seismically active region, a test of our
models derives from the correlation be-
tween the location of conversion points
predicted by the models and the bound-
aries of the seismically active region.
Projections of possible conversion regions
for Models A and B are shown in Fig.
35 along with the seismicity for depths of
100-250 km between the AHFA-produc-
ing events and the stations, and Fig. 37
TABLE 14. Results from Models A and B for a Focus at 589 km and a Conversion at 150 km
Departure
Emergence
Departure
Maximum
Angle
Angle at
Angle from
Depth of
Time
from
Conversion
Conversion
Phase 2,
Difference,
Focus
Point
Point
km
A2
A total
sec
Model A: P to S
56.8
65.7
-25.4
347
18.2
19.8
3.3
51.1
61.8
-22.4
304
16.8
19.6
-4.0
55.4
64.8
-25.4
347
18.2
20.6
0.3
54.0
63.8
-23.9
326
17.5
20.0
0.2
54.0
63.8
-22.4
304
16.8
19.3
0.2
54.0
63.8
-20.8
282
16.0
18.5
1.0
52.6
62.8
-19.0
261
14.9
17.5
-0.2
52.6
62.8
-17.1
240
14.1
16.7
0.1
52.6
62.8
-14.9
219
13.2
15.8
0.4
52.6
62.8
-12.4
197
11.9
14.5
0.4
52.6
62.8
- 9.2
Model B
176
: P to P to S
10.1
12.7
0.0
48.5
-60.1
-19.0
261
14.9
19.3
4.8
44.5
-57.6
-17.1
240
14.1
19.0
-3.2
46.5
-58.9
-23.9
326
17.5
21.5
-0.9
46.5
-58.9
-22.4
304
16.8
20.6
-0.1
46.5
-58.9
-20.8
282
16.0
19.6
0.7
46.5
-58.9
-19.0
261
14.9
18.8
1.1
46.5
-58.9
-17.1
240
14.1
17.9
1.3
46.5
-58.9
-14.9
219
13.2
16.5
1.4
46.5
-58.9
-12.4
197
11.9
14.7
09
46.5
-58.9
- 9.2
176
10.1
12.6
-0.2
Angles in degrees are measured from the horizontal and are positive for upward-directed rays.
DEPARTMENT OF TERRESTRIAL MAGNETISM
237
includes a vertical seismicity plot for the
region between a particular event and
NNA. These figures show that the con-
version points are indeed in the vicinity
of the eastern boundary of the inter-
mediate-depth seismic region.
A further test of our models derives
from an examination of the Q along the
proposed path for each event producing
an AHFA. Figure 38 contains a photo-
graph of a seismogram showing the
AHFA at NNA from the event of Fig.
36 along with a photograph of a seismo-
gram showing the direct S arrival at
NNA for an event, rj, which lies along the
path of phase 2 for either model (as in-
dicated by the arrow in Fig. 36). The
path length for the direct S from the
event -q includes about 40% of the total
path length, resulting in an AHFA for
either model. It is essential to our models
that this be a high-Q path; the high-
frequency content of the S arrival from rj,
compared to that of the AHFA, is con-
sistent with this requirement.
Because the data for AHFA, such as
travel time and Q, can be fitted by widely
differing models, further tests are re-
quired to find the operative model for a
given geographical region. (It is not in-
NNA 13 MAY 1965
h = 589 KM A =14.52°
NS COMPONENT
NNA 26 MAY 1970
h= 159 KM A = 7.2°
NS COMPONENT
— ONE MINUTE
Fig. 38. Photographs of seismograms from NNA for an AHFA and for an event, v. occurring
along the proposed path of the AHFA-producing phase (see Fig. 37).
238
CARNEGIE INSTITUTION
conceivable that no single model will
prove universal.) An obvious further
test for our models involves calculating
expected amplitudes resulting from re-
fraction or reflection from interfaces sug-
gested by the seismicity. Although
Isacks' and Barazangi's model also in-
volves a conversion, more restrictive tests
for that model should result from impli-
cations of their heretofore unverified
assumptions.
We have found that a number of dif-
ferent models fit travel time and Q data
for AHFA. Therefore, in the absence of
other restrictions, no conclusions based
on uniqueness of any one model can
validly be drawn.
Converted P Phase from the ScS Phase
at the Inclined Deep Seismic Zone
Hiromu Okada
A longitudinal forerunner to the ScS
phase has been observed on short-period
seismograms for nearby deep earth-
quakes at Kamikineusu (KMU) and
Urakawa (URA), both stations located
on the Pacific coast of Hokkaido, north-
ern island of Japan. At these stations a
time difference of eight seconds between
the forerunner and the ScS phase is es-
sentially constant. The particle motion
for the ScS phase showed horizontal
polarization (SH). The longitudinal
forerunner for the clear arrival, however,
showed a different incident azimuth
which corresponds to the dip direction of
the deep seismic zone. The proposed
interpretation for these forerunners was
an ScSp phase, the P phase converting
from the ScS phase at the upper bound-
ary of the dipping seismic zone.64 Figure
39 shows the schematic ray paths and
conversion interfaces discussed.
Additional data were found at Matsu-
shiro (MAT) in central Honshu, Japan.
Figure 40 shows a traced record example
(left below). The amplitude of the ScSp
S°UTH
4M£R
w Station
• Hypocenter
++ Earthquakes
i a J ?r schematlc model showing the paths of the ScSp phase (solid line) and the ScS phase
(dotted line). The ScSp phase, a P phase converted from the ScS phase at the inclined seismic
zone, exists at certain stations in the Japanese and South American regions.
DEPARTMENT OF TERRESTRIAL MAGNETISM
0 200 400
0
DEPTH
100
200
400
200
I
MAT
1 43 r
4351 1 1 1
600
I
800
l_
239
1000 KM
y>t ,
121
M
s s
„- 1
A.
TAX
15/ .///
20 25
30
35 sec
1970 Sep. 5 OKHOTSK^ s
m = 5.7 H = 580 KM
A =18.2° AzB= 26.6°
Fig. 40. Traced seismograms (left below) of the ScSp and ScS phases at MAT. The vertical
cross section of the earthquake activity is shown by the numbers of the events located inside the
oblong on the map (USCGS and NOAA, 1964-1971). Assuming the normal incidence for the
ScS phase, the ScS-ScSp time referred to J-B travel times is indicated by the dotted lines. The
time difference contours at 21 seconds are hatched and just cover the upper boundary of the dip-
ping seismic zone.
phase is about half that of the ScS phase.
The time difference, however, is 21 sec-
onds. There are no significant differences
for this lead time. Sometimes the ScSp
phase is hard to identify because of poor
signal-to-noise ratio.
The seismic activity has been projected
onto a vertical cross section in Fig. 40.
Assuming normal incidence for the ScS
phase, the time-difference (ScS-ScSp)
contour is calculated using J-B travel
times and is shown by the dotted lines.
The general pattern of the time differ-
ence curves suggests a dip angle of about
32°. The hatched zone — 21 seconds for
MAT — covers the upper boundary of the
dipping seismic zone. In this case the
interpretation referred to above is ade-
quate. Evidence suggests that for MAT
the interface at the deeper part of the
dipping zone might be sharp enough to
produce a clear conversion.
In South America it was found that
the ScSp phase exists at some stations,
but the relationship of this phase to the
seismicity appears more complex. Exam-
ination of the ScSp phase on the short
period WWSSN seismograms for Are-
quipa (ARE), southern Peru, La Paz
(LPB), Bolivia, Antofagasta (ANT),
northern Chile, and Nana (NNA) , cen-
tral Peru, revealed the following:
240
c
0
DEPTH
100
200
300
KM
200
L
1
400 ARE
_i I_
600
CARNEGIE INSTITUTION
1000 KM
800
i
1
11 1
1 11 11% v
V 1
1 1
V. NN 1
1 V*
1
1
1
1
1 12J2T 111 1\ 1\ x.
^ N
^ 1 N-%cN s-
25 20 15 10 5 0 sec
1965 Nov. 3
m = 6.2 H = 593 KM
A = 7.3° AzB = 0.7°
PERU -BRAZIL BORDER
W
ScS
ScSp
va^W^^V'
10
i
20 sec
_J
Fig. 41. Vertical cross section of earthquake activity and the traced seismograms (right below)
of the ScSp a ad ScS phases at ARE. Time difference contours as for Fig. 39.
Fairly clear longitudinal ScSp phases
were observed at Arequipa about 11 sec-
onds before the ScS phase (Fig. 41). In
southern Peru the deep seismic zone is
clearer and steeper than in central Peru.
The interpretation previously referred to
is also adequate in this case. No clear
phase is recognized at La Paz and Anto-
fagasta. The La Paz station is east of
the dipping seismic activity, and there-
fore no dipping interface is expected. For
the coastal station, Antofagasta, since the
precise form of the deep seismic zone is
convex, the converted wave from the
deeper part could arrive above the inter-
mediate activity but not near the coastal
region.
There is a 10-second precursor at
Nana. The seismicity is more compli-
cated (Fig. 42). The ScSp phase looks
similar to the ScS phase. A seismicity
cross section is shown in Fig. 42. Three
regions show relatively dense activity:
inland shallow activity, ocean-side shal-
low activity, and intermediate activity.
Isacks and Molnar (1971) 65 suggest that
the slab-like seismic zone has a dip of
only 10° to 15° (inset in Fig. 42) (Fu-
kao, 1972,66 Sykes, 1972 67). The hatched
zone around 10 seconds, indicating the
probable interface at which the ScS to
ScSp conversion takes place, is below the
slab. If this slab is a good representation
of the deep seismic zone, the ScSp phase
is not converted at the upper boundary
and the proposed hypothesis apparently
fails. The conversion might occur at the
bottom of the slab. Some other explana-
tion might be possible, considering that
the hatched part can be a good indication
DEPARTMENT OF TERRESTRIAL MAGNETISM
0
241
o
DEPTH
100
200
300
KM
200 NNA
_J I L
1
1 1 1 11
2 1112122 1
2436 14433984625343 21
12132235C3331 v
2 112 2444513 21 ^
1 1444622111 1
400
L
600
■J
800
1000 KM
I 1311 1
1 21 111
II 141
1 1
1
1
14A41 1
1131 11 1
1 1111 1222141444 1 1
21 12 2 1221 4 412247AC71321111 1
K 111 2 221 1421 1 1
V 1 \ 1 11 12 11 1
1 1 11 1
1 11 1
v< 1 1 1 11
i A v 3 \ 1 1 1 11311 1 1
TInA 11K3 \1 11 1 11 1 1 1 1
11<SL 1 \ \ 1 1 113 2 11 1
TV \ \1 11 1 1111 21 1
1 V 1 \ 1 V 1 11 222 21 1
1 %. ~\ \1 11 1 12 12311
31434 31
32311
% 111
%. ^ 1 31
1 1
11
%.B V 1\1
\ 1
\
\
\
'€
1 \
\ %
\
\
20sec15 10
\
\
\
5 0
200
NNA
1
400
600
800
1000 KM
1965 Mar. 5 SANTIAGO DEL
m = 5.5 H = 573 KM ESTERO
A = 19.6° AzB = 141.5°
ScS
ScSp
10
l_
20 sec
Fig. 42. Vertical cross section of the earthquake activity and the traced seismograms (right be-
low) of the ScSp and ScS phases at NNA. Isacks and Molnar's gently dipping slab model (inset,
left below) also indicates only large events for the same region, with magnitude greater than
about five (USCGS, 1961-1967). The upper boundary of the slab is far above the 10-second time
difference zone (hatched zone), and the hypothesis of wave type conversion at the top of the
inclined seismic zone apparently fails.
of the seismic zone. There is clear evi-
dence that the ocean-side shallow activ-
ity shows significant bending to a steeper
angle (Fig. 42) . The extrapolated zone
of that activity is in contact with the
hatched zone (Fig. 42).
Zoeppritz's equations describe the am-
plitude relations in terms of elastic pa-
rameters and geometry.68'69 For ver-
tically polarized shear waves (SV) inci-
dent from the high velocity side, the am-
plitude of the refracted P wave gradually
increases with incident angle and forms a
small minimum at the first critical angle
(reflected P critical). After the first cri-
tical angle, considerably larger amplitude
is expected up to the second critical angle
(refracted P critical). The larger ampli-
tude corresponds to greater velocity con-
trast. Figure 43 shows these amplitude
ratios ( refracted P/refracted S). When
the velocity of the incident side is 10%
higher, the amplitude ratio can reach
nearly one-fifth.
Assuming normal incidence for the
ScS phase, and using the Nafe-Drake
velocity-density relation (Talwani et al,
1959) 70 and a Poisson's ratio of 0.25, the
velocity ratio R (transmitted/incident)
and the amplitude ratio are calculated
for the incident angle i8 and the take-off
angle 6P at the conversion point (Fig.
242
CARNEGIE INSTITUTION
10
20° 30° 40
INCIDENT ANGLE
50'
Fig. 43. The transmitted amplitude ratios (refracted P/refracted S) for the incident angle is.
The velocity contrast R (transmitted /incident) = 0.7, 0.8, 0.9, 1.1, 1.2, 1.3, and the Poisson's ratio =
0.25. In the high velocity side, the P wave velocity is constant at 8 km /sec.
44) . The P wave velocity in the high
velocity side is constant at 8.0 km/sec.
For the normal incidence, is is equal to
the dip angle 8 of the dipping seismic
zone. The converted P to transmitted S
amplitude ratio is indicated by the
broken lines in percent. The velocity
ratio R (transmitted/incident) is shown
by the solid lines. The zones that satisfy
the time differences in Figs. 40 to 42 cor-
respond to the dotted columns and those
marked A, B, C. For NNA, two cases are
examined: NNA-1 for the time difference
zone and NNA-2 for the bottom conver-
sion of the gently dipping slab (Fig. 44) .
It is not easy to determine dip angle, espe-
cially in South America. The incident
angle (equal to the dip angle) for KMU,
MAT, ARE, and NNA-1 is about 30 to
35°, and for these cases is is slightly less
than the first critical angle. From studies
of travel-time anomalies in typical island
arcs, it is believed that the velocity in the
dipping seismic zone is 5 to 10% higher
than the surrounding material. Al-
though this velocity anomaly is a path-
integrated averaged value, it is appro-
priate to constrain the velocity ratio R
at about that value. For shallower con-
version, A (in Figs. 40 to 42) lies near the
R = 0.9 line, and the amplitude ratio is
about 5 to 15% (Fig. 44). Since the
velocity of the dipping seismic zone is
higher than the underlying part, the inci-
dent angle is might be several degrees
larger than the normal incident case.
After correction, the incident angle is
may become larger than the first critical
DEPARTMENT OF TERRESTRIAL MAGNETISM
243
30
INCIDENT
ANGLE
Fig. 44. The geometry, the velocity contrast, and the amplitude ratio for S to P conversion.
Under certain assumptions, the velocity ratio R (transmitted/incident) is calculated (solid lines),
and the amplitude ratio (converted P/transmitted S) is shown in percent by the dotted curves,
from the Zoeppritz equations. The hatched zones in Figs. 39, 40, and 41 correspond to the dotted
columns. The marks A, B, and C indicate the arbitrary position corresponding to the same marks
in Figs. 40, 41, and 42. See text for further discussion.
angle, and more efficient conversion will
occur in a slightly deeper part between A
and B (Figs. 40 to 42). For the con-
version at the bottom of the slab (inset,
Fig. 42), R = 1.3 and the amplitude
ratio is less than 10% (Fig. 44, NNA-2).
This velocity ratio (R = 1.3) is unrea-
sonably large and, for more probable
velocity ratios, the conversion will be too
inefficient to cause the large observed
arrivals. Conversion on the upper bound-
ary could explain not only the time dif-
ference but also the geometry and the
amplitude tendency.
Some spectra from the Carnegie broad-
band seismograms have been analyzed
(Fig. 45) . At KMU, spectra of the ScS and
the ScSp phases show the same form and
slope in the frequency range 0.2 to 0.9
Hz. The spectral ratio ScSp/ ScS is about
unity. Because of the low signal-to-noise
ratio at MAT (for the example shown
in Fig. 45) and the considerable differ-
ence in two horizontal spectra, it is not
easy to compare them. In other cases at
MAT and other stations the apparent
wave characteristics show no significant
differences between the ScSp and ScS
phases. Regional differences in Q may
affect spectral characteristics, but it ap-
244
CARNEGIE INSTITUTION
100
0.5
1.0
Frequency
Fig. 45. The spectral characteristics of the ScSp and ScS phases. Traced seismograms (top), the
spectra (middle), and the spectral ratio (bottom). The 10-sec boxcar window was applied to
KMU (right) and a 16-sec window to MAT (left). The digitizing rate was 40 points per second.
The spectral ratios are calculated after 1/3 octave averaging. For KMU, the three spectra show
the same slope and the ratio is about unity.
pears more likely that conversion is fre-
quency dependent because of velocity
structure at the interface.
Discussion
1. The ScSp phase exists in a certain
area in Japan and in South America
where there is a clear seismic zone. The
time difference, the geometry, and the
amplitude tendency agree with the con-
version at the upper boundary of that
zone (Fig. 39).
2. NNA's case is especially interesting
because when the gently dipping slab
model 65 is taken, the proposed hypothe-
sis apparently fails. This suggests that
DEPARTMENT OF TERRESTRIAL MAGNETISM
245
at intermediate depths the slab itself
is aseismic and may be continued from
the ocean-side shallow active seismic
slab.
3. The amplitude ratio ScSp/ScS can
be quite large, and for these cases the
conversion must be at nearly critical
angles. The velocity contrast might also
be quite large. Averaged velocity anom-
alies from the travel-time studies may
not be enough to produce a clear conver-
sion. Considering the amplitude, the
geometry, and the spectral characteris-
tics, it is likely that the boundary is a
thin transition zone instead of a simple
boundary. Application of a Thomson-
Haskel matrix method to this problem is
under study.
4. The ScSp phase can be a new tool to
investigate a boundary in the upper
mantle and provides a new approach to
understanding the tectonic processes at
island arcs or island arc systems.
Source-Spectral Relations for
Earthquakes
A. Snoke, A. T. Linde, and I. S. Sacks
Archambeau's elastodynamic source
theory 71,72 postulates an infinite medium
subject to a uniform shear field from
which strain energy is released as a con-
sequence of the creation of an expand-
ing rupture zone in which the rigidity
vanishes. The only geometry published
to date is a spherical zone which expands
to some maximum radius R() with a con-
stant radial velocity VR (although Ar-
chambeau and Minster have also found
solutions for various propagating-sphere
geometries73). In some calculations it
appeared convenient to truncate integrals
over the prestressed medium to the vol-
ume of a sphere containing the source
with finite radius Rs. This seemed justi-
fied by Archambeau's finding 72 that
more than 95% of the total radiated
energy comes from a region surrounding
the rupture zone of radius equal to five
i^o and by Linde's finding 74 that for
Ro, VR, and the signal velocity TTa held
constant, the far-field displacement spec-
trum was independent of Rs for Rs large.
Using an equivalent but computation-
ally simpler formulation of the same
theory,72 we have redone Linde's calcula-
tions and have found some of his results
to be in error. We now find that the
shape and the position of the peak of the
far-field displacement spectrum have a
strong dependence on RH. Typical results
are shown in Fig. 46, where we have
plotted the far-field displacement spec-
trum for Rs/Ro equal to 5, 10, 20, and 40
while Ro, VR, and Va are held constant.
(For comparison we have included the
spectrum as calculated previously by
Linde). As Rs increases, the frequency
at the peak, vpeak, decreases so that Rs X
vpeak remains constant and the curve gets
broader. As pointed out by Randall,75
in the limit that Rs becomes infinite, the
FREQUENCY IN Hz
Fig. 46. Far-field displacement spectra for
Ro = 10 km, VR = 4.5 km /sec. Va = 10 km/sec
and RJRo = 5. 10. 20 and 40. Also included
(dashed line) is the spectrum as calculated by
Linde for Ro, Vf, and T"a the same as above
and RsIR, > 8.
246
CARNEGIE INSTITUTION
spectrum approaches the flat-to-d.c.
shape predicted by dislocation mod-
els.76,77-78 Thus, Archambeau's theory
produces peaks only for noninfinite val-
ues of Rs- Since the model assumes an
infinite uniformly prestressed medium,
it is inadmissible to choose a finite value
for Rs.
These findings have some bearing on
the previous study by Linde and Sacks 79
in which dimensions of South American
deep earthquakes were obtained by inter-
preting observed spectra in terms of the
theoretical spectrum labeled "Linde" in
Fig. 46. Although this interpretation is
not correct, it might be noted that these
dimensions were also consistent with
measurements of radiation time, and so
we do not expect serious errors in these
estimates. It also bears mention that
energy calculations in that study are
independent of the source theory, as is
the argument which leads to the conclu-
sion that radiation efficiency increases
with magnitude.
In order to derive any relations be-
tween source dimensions and displace-
ment spectra, we require a theory that
reproduces observed spectra. When con-
sidering possible modifications of existing
source theories, we examine assumptions
implicit in the current formulations of
those theories which, if altered, would
make the theory more physically realistic
but would not make it mathematically
intractable, and we then ask if altering
these assumptions could alter the low-
frequency behavior. In Archambeau's
theory we are examining assumptions
concerning the geometry of the rupture
zone and the size of the prestressed me-
dium. A spherical geometry for the rup-
ture zone is a very unrealistic model for
an earthquake, but it is believed that
going from a spherical to a more planar
geometry will have little effect on the
low-frequency part of the spectrum. A
radial propagation may not be bad for
describing bilateral earthquakes, but it
is a very poor approximation of a uni-
lateral event. Archambeau and Minster
report 73 that linear propagation brings
in higher-order multipole contributions
which can affect the spectrum near the
corner frequency. In the earth, techn-
ically stressed regions are definitely
finite, and for the South American events
we have been studying, the observer is
far outside any prestressed region. The
spectra in Fig. 46, in which Rs is treated
as finite, are probably meaningful to the
extent that they indicate that increasing
the size of the prestressed region tends to
increase the amplitude of the low-
frequency part of the far-field displace-
ment spectrum.
Borehole Strainmeter Installation
in Japan
/. S. Sacks and S. Suyehiro
The major emphasis of the initial phase
of the operation of the DTM borehole
strainmeters in an active earthquake re-
gion in Japan was an assessment of the
capabilities of this type of instrumenta-
tion. Initial tests using explosive-gen-
erated vibrations showed that the instru-
ments and their immediate environment
could tolerate the severe accelerations
caused by near earthquakes without
spurious behavior (Sacks et al., 1971 ).80
What remained to be confirmed was
whether strain measurements made very
near the surface were reliable, since rock
at shallow depths is often fractured and
jointed, particularly in the Matsushiro
area.
We have already recorded the strain-
field changes of a number of earthquakes
which occurred within the net, and we
find that consistent strain steps are re-
corded at the different sites. For example,
an earthquake of magnitude 3.9 on 26
October 1972, which was located about
midway between the Nagano and Matsu-
shiro installations (10 km), gave strain
steps of 1.4 X 10"8 at Nagano and
2.1 X 10"8 at Matsushiro. Body waves
and surface waves from distant earth-
quakes write almost identical records at
the three sites.
DEPARTMENT OF TERRESTRIAL MAGNETISM
247
We now know that strain steps, body
waves, and surface waves can be re-
corded with fidelity by borehole strain-
meters in shallow holes (50 km), but
in the period range of 10 minutes to
several hours, the local noise is domi-
nated by pressure changes in the aquifer.
This is a very local effect; noise in this
period range differs by more than a fac-
tor of 100 in two instruments just 300
meters apart. Because of this low co-
herence, we expect that installation of a
strainmeter well below the level where
the aquifer saturation changes due to
such actions as pumping would reduce
this 10-minute to 2-hour noise.
At very long periods (of about a year) ,
the results from the borehole instrument
agree well with the 100-meter quartz ex-
tensometer at Matsushiro. Further anal-
ysis of the data will be done during a
forthcoming visit by Sacks to Japan.
Project Narino
L. T. Aldrich, D. E. James, A. T. Linde, L. Ocola,
G. R. Poe, I. S. Sacks, and D. Simoni
Collaborating Institutions
Bolivia: Comision de Geofisica; Instituto Pan
Americano de Geografia e Historia. Colombia:
Instituto Geojisico de Los Andes, Universidad
Javeriana; Instituto Geogrdfico "Agustin" Co-
dozzi; La Armada Colombiana; Texas Petro-
leum Corp.; International Petroleum Corp.;
INGEOMINAS. Ecuador: Instituto Geogrdfico
Militar Ecuatoriano; Direccion National Geo-
logico; Escuela Politecnica; Instituto Hidro-
grdfico de La Armade Ecuatoriana. Germany:
German Seismic Group. Panama: Inter-Amer-
ican Geodetic Survey. Peru: Instituto Geojisico
del Peru; Instituto Geojisico, Universidad Na-
tional de San Agustin; Centro Regional de
Sismologia. United States: University oj Texas;
University oj Wisconsin; University oj Hawaii;
University oj Washington.
The Andean cordillera, an active tec-
tonic belt that extends thousands of kilo-
meters along the west coast of South and
Central America, has been a major focus
of geophysical research by the Depart-
ment of Terrestrial Magnetism (DTM)
for more than a decade. Most of these
studies have been confined to the region
of the central Andes — southern Peru,
Bolivia, and northern Chile. The central
Andes, dominated by two mountain
chains — the eastern and western ranges
— and an intervening high plateau, lies at
the heart of the Andean system and stud-
ies of the region have never failed to
reveal unique and extraordinarily com-
plex structures. Nonetheless, there are
other sectors of the Andean cordillera
that present equally fascinating geolog-
ical and geophysical anomalies and com-
plexities. Such a region is southwestern
Colombia. Here the Andes, which to the
south form a compact chain of mountain
belts trending parallel to the coast, di-
verge to form three distinct north-north-
easterly trending ranges separated by
broad low-lying valleys (see Fig. 47).
The point of confluence of these three
ranges near the city of Pasto marks the
approximate latitude north of which the
Peru-Chile trench shallows and nearly
disappears.
Gravity investigations by Case and
others (1971,81 1973 82) demonstrate that
this region is not only physiographically
unusual but tectonically complex as well.
The Bouguer anomalies indicate that
parts of the Andean cordillera within this
zone are either out of isostatic equilib-
rium or are compensated by unusual
crustal and upper mantle structures. For
example, the Cordillera Occidental of
western Colombia — some peaks of which
attain elevations of 10,000 feet or more —
is characterized by position Bouguer
anomalies. Case and his co-workers have
suggested that the Cordillera Occidental
has been built atop oceanic crust, possi-
bly as an island arc that has subse-
quently collided with and been incor-
porated into the South American conti-
nent, and that the crust is therefore dense
and comparatively thin. Other parts of
this region of the Andean cordillera are
similarly complex, and the magnitude of
the gravity anomalies suggests that con-
trasts in crustal structure may be large.
To examine the crustal structure of
southwestern Colombia and northwestern
248
CARNEGIE INSTITUTION
OCEAN SHOTS
MALPELO
ISLAND^
PACIFIC
OCEAN
COLOMBIA
+
78°
O BOGOTA
^
+
EQUATOR
V_
+
74°
PERU
PROYECTO NARINO
********* LAND SEISMIC STATIONS
Fig. 47. Location map showing position of shot lines and observing profiles (asterisks) for Proj-
ect Narifio. Land shot point is situated in Laguna de la Cocha, east of Pasto.
DEPARTMENT OF TERRESTRIAL MAGNETISM
249
Ecuador, the DTM participated in the
organization of a multinational study
carried out early in 1973. The program
was coordinated by Fr. J. E. Ramirez,
S.J., of the Instituto Geofisico de los
Andes Colombianos.
Shot points, shot lines, and observing
profiles are shown in the map given in
Fig. 47. The only genuinely fixed shot
point was that in Laguna de la Cocha
near Pasto where some 20 explosions of
one-half to two tons were fired by DTM.
A series of shots was fired from the ship
Gorgona near Tumaco and Buena-
ventura by the University of Texas at
Dallas. Two-ship seismic profiling be-
tween Tumaco and Buenaventura was
carried out using both the Colombian
vessel Gorgona and the oceanographic
ship Kani-Keoki of the University of
Hawaii. Finally, a series of shots about
50 km apart was fired by the Kani-Keoki
along two profiles, Malpelo Island to
Tumaco and Buenaventura to Malpelo
Island. Seismic observations along short
profiles near Malpelo Island were made
by the University of Washington oceano-
graphic ship, the T. G. Thompson, which
further carried out heat flow, bathymetry,
and magnetic measurements.
Recording sites on land were organized
along four principal profiles, as shown
in Fig. 47. Profile I was occupied by ob-
serving groups from the University of
Wisconsin and the German Seismic
Group along that part of the profile
north of Pasto; south of Pasto and in
Ecuador the observations were primarily
the responsibility of DTM. Four DTM
and IGP recording groups observed along
profile II which trended along the valley
of Huila. Observations along profile III
were carried out by the University of
Texas at Dallas and the University of
Hawaii. All groups observed along pro-
file IV.
Preliminary results are still fragmen-
tary and contradictory. There is some
indication that no apparent velocities
greater than 7.5 km/sec are observed
along the north-south profiles even at
distances well in excess of 300 km. If
this conclusion holds, it will mean that
we have yet to observe definite "normal"
mantle velocities anywhere along the
Andean system. This work was sup-
ported in part by the National Science
Foundation and the Harry Oscar Wood
Fund of the Carnegie Institution.
Electrical Conductivity Studies
L. T. Aldrich, M. Casaverde, J. Bannister,
L. Beach, T. J . Smith, L. Tamayo, E. Triep,
R. Quiroga, R. Salgueiro, and S. del Pozo
Observations to study the electrical
conductivity structure across the Andean
cordillera in central Chile and Argentina
had just been completed at the end of
the previous report year. These observa-
tions were obtained using a 15-station
net of magnetic variographs operated at
stations located as shown in Fig. 48. Five
nighttime events have been digitized to
provide good values for the AZ/AH and
AD/ AH relationships among the stations.
Two or three daytime events are needed
to complete the data required to give
AZ/AD and AH /AD as a function of fre-
quency. This work has been and is being
carried out both at the University of
Chile in Santiago and here at the De-
partment. Preliminary results from the
nighttime events substantiate those found
earlier with two stations. Any abnormal
conductivity with an E-W component, if
it exists, lies west of the net (under the
ocean). The daytime storms with varia-
tion in D are needed to examine struc-
tures with an alignment parallel to the
Andean cordillera (N-S).
Of major importance to future analy-
sis are two developments this year. An
interpolating algorithm developed by
Professor Shelton Alexander of the Penn-
sylvania State University has diminished
the time required for digitizing our rec-
ords by at least a factor of 2. This
250
CARNEGIE INSTITUTION
70° W I
LA SERENA
Guanaqueros
Tololo
r-
V
/
A,
/
Huintil
+
IHuentilaquen
Z7^
CHILE
VALPARAISO
El Quisco
Disputada
+
,'
+
Mario Pinto
SANTIAGO
-f-
30° S
""^Paso de las Aguas Negras
/+ +
+ Jachal
Las Flores
Alumbrera
+
Barreal
«fl
SAN JUAN
Cerro Negro
32° S
A RGENTINA
+ Uspallata
+f MENDOZA
34° S
70° W
Fig. 48. Map showing locations of magnetic variograph stations in Chile and Argentina for elec-
trical conductivity studies in 1972.
algorithm requires digitization only at
points where the first and second time
derivatives of the record are zero. Com-
parison with records digitized at one-
minute intervals showed that the two
digitizing systems produced identical
records. Initial steps have been taken to
adapt new techniques of frequency anal-
ysis to our problem. These techniques
look most promising and would not only
further reduce our digitizing require-
ments but also increase the precision of
our analysis.
The magnetic variographs are cur-
rently deployed in Peru, Bolivia, and
northern Chile to study in more detail
the anomalies in southern Peru and cen-
tral Bolivia. Revisions of the DTM
variographs to make them even more
easily used have been completed. Figure
49 shows a record obtained at Las Flores,
Argentina, of an event recorded on May
16, 1972. The characteristics of the rec-
ord which make it easy to analyze are:
(1) one-minute digitization rate, (2)
independent time marks at one hour and
thirty minutes, and (3) twelve-hour
marks on the Z (vertical) trace to help in
time keeping and trace identification.
The unit operates at 7.5 volts with aver-
DEPARTMENT OF TERRESTRIAL MAGNETISM
251
Is
or
<
or
o
1 K
i %
II /
i
/
i
Ha
*4 K
<r i
1/
Q
I I I
I i
! I
N
Q
>^ 1 \
-
J *
% 1
* f
m f
I 1
I m
il I
N
I
!
l
>
f
I
#• '
C7T
XL
01
<
UJ o
UJ
C/)
<
00
I
*i — I— I
CD
~ 3 ^
3 C •/.
S45 O
~C - ~
:- x —
O ^ ■-
•^ '— ~
'- T- ~
'"_ 5 o
■a § I
X - -
~ 2
bp a
E-i -■
W — '
o —
S 5
— -5
3 w
O
i-i c
—
5 §
g s
X! BD .2 b
O 3 O
« 2 f3
;- — .
X X
•- 3
<M O - O
Oi ^ g '-
CD r3 — S
3 O C
r 3 o
x - "~
£ ™ O
3 ©
bC
N
o3
«~ 5
O x « '
— X
2 ~C *-■ ~
2 g ~
H -H
J3 ^
■f r- !> _z:
■SP § £ 2
o S
252
CARNEGIE INSTITUTION
age current of 60 ma, which is easily
provided by aircells, which in turn re-
quire no other source of power during
their lifetime.
Geochemistry and Geochronology
Geochemical Trends in Andean Rocks
D. E. James, C '. Brooks, and A. Cuyubamba
We have reported in previous Year
Books on our efforts to understand the
evolution of the Andean orogen. Over
the years of studying the central Andes
Mountains, the direction and emphasis of
our research have been altered to pursue
promising new avenues to fresh informa-
tion. As the implications of the role of
plate tectonics in continental evolution
have become better recognized and un-
derstood, the Andean system has emerged
as a classic example of an evolving
orogen produced by ongoing subduction
along a continental margin.
Our previous studies of the Andes were
confined largely to geophysics, with an
emphasis upon crustal and upper mantle
structure. Those early studies form the
necessary underpinnings of an under-
standing of evolution as read from the
geologic record. Plate tectonic theory,
however, has led to a confluence of for-
merly distinct disciplines within earth
science, and there now exists clear com-
plementarity between geophysics, geo-
chemistry, and geology in the study of
orogenic processes. Thus, while geo-
physics tells us much about the existing
tectonic regime of the central Andes, it
can only hint at the tectonic evolution
which has led to the present-day moun-
tain system. To read the record of past
events it has been necessary to turn to
geology and geochemistry.
The kinds of geological or geochemical
studies which can be fruitfully pursued
by one or two individual research men
must be carefully framed. It is clear, for
example, that geological field studies are
pursued best within the context of large,
organized programs such as those car-
ried out over decades by national geo-
logical surveys. Thus, we at DTM have
concentrated our efforts on geochemical
studies which heretofore have constituted
a virtually unexplored domain in the
central Andes. By focusing our attention
upon those geochemical parameters that
have been shown to be sensitive indica-
tors of plate tectonic processes, we hope
to shed light on the early tectonic history
of the Andes.
A model for Andean evolution from
Paleozoic times to the present was de-
scribed in Year Book 70, pp. 344-349.
That model was based largely upon geo-
physical and geological data but may be
at least partially tested by studying cer-
tain aspects of the geochemistry of the
igneous rocks that form the Andean
cordilleran arc. In Year Book 71 we out-
lined our program of geochemical study
and our specific objectives. The results
reported below represent preliminary
data now emerging from these studies.
Sr87/Sr86 ratios for igneous rocks of the
central Andes. One hypothesis related to
plate tectonic theory holds that the ande-
sitic magmas extruded from the vol-
canoes of island and cordilleran arcs
originate by partial melting of the upper
parts of the descending plate or of the
mantle immediately above the plate.
Partial support for this hypothesis comes
from the measurement of Sr87/Sr86 ratios
in igneous rocks of the island arcs. In
these rocks the ratios are uniformly
around 0.7038 and it is generally thought
that ratios this low imply magma origins
from within the mantle or oceanic crust.
As regards island arcs, this hypothesis
has met little opposition, for there ap-
pears to be no other probable source of
magma generation.
For cordilleran arcs such as the Andes,
this conclusion, though following logic-
ally by analogy with island arcs, is not
so simple or so readily verified. It is
possible, for example, that cordilleran
arcs sitting astride sialic crust are fed
DEPARTMENT OF TERRESTRIAL MAGNETISM
253
by anatectic melting of preexisting crus-
tal rocks. Indeed, until the advent of
plate theory, such an origin was generally
assumed; Pichler and Zeil (1967,83
1969,84 1971 85) still maintain that the
late Cenozoic volcanism of northern
Chile is caused by anatectic melting of
lower crustal rocks. Arguments like
these are kept alive by the fact that
Sr87/Sr86 ratios in rocks of the cordilleran
arcs such as the Sierra Nevada are con-
sistently higher than those of the island
arcs.
We have previously concluded, on the
basis of geophysical evidence indicating
that the crust of volcanic arc has nearly
doubled in thickness over the past 50 to
100 m.y., that the magma source of the
late Cenozoic volcanism must lie beneath
the crust. We sought to further discrimi-
nate between various possible magma
sources by measured Sr isotopic ratios in
the igneous rocks of the arc. Clearly,
nothing in the Sr ratios can bear either
for or against a lower crustal origin, for
we have no knowledge of the isotopic
composition of those rocks. Nonetheless,
isotopic ratios near those of island arcs
should lend credence to the hypothesis
that the magma originates subcrustally.
If isotopic values are very high, in excess
of 0.710, for example, they could suggest
anatectic origin by melting of preexisting
crustal rocks.
Strontium 87/86 present-day ratios
uncorrected for age are given in Table 15.
Radiometric ages have yet to be deter-
mined and hence a correction for radio-
genic Sr87 has not been made. The rock
ages given in the table are on the basis
of field relations and provide only a gen-
eral guide to the relative chronology. For
late Cenozoic volcanic rocks, the present-
day ratios are equivalent to initial ratios.
For certain late Cretaceous-early Terti-
ary rocks, the Rb/Sr ratio is sufficiently
high that the correction for radiogenic Sr
may be quite large. For example, sample
86 has a Rb/Sr ratio of greater than 1.0;
thus if the rock is 100 m.y. old, the iso-
TABLE 15. Sr87Sr86 Present-Day Ratios for Rocks of the Central Andes of Southern Peru
Sample
No.
Sr87/Sr!
Rb/Sr
Depth to
Benioff
Zone, km
Description'
PE 26 0.7071 ±0.0001 0.0547 155 Quaternary andesite
PE 27 0.7070 ±0.0003 0.0547 155 Quaternary andesite
PE 30 0.7072 ±0.0002 0.0548 155 Quaternary andesite
PE 46 0.7076±0.0001 0.0883 160 Quaternary andesite
PE 71 0.7052±0.0002 0.1570 200 Quaternary andesite
PE 1 0.7069 ±0.0002 1.095 130 Cretaceous-early Tertiary granite
PE 13 0.7070 ±0.0003 0.3569 78 Cretaceous-early Tertrary diorite/gabbro
PE 17 0.7057 ±0.0002 0.1341 145 Cretaceous-early Tertiary tonalite
PE 58 0.7057 ±0.0002 0.2456 1155 Cretaceous-early Tertiary tonalite
PE 85 0.7052 ±0.0002 0.0455 140 Cretaceous-early Tertiary gabbro
PE 86 0.7088±0.0002 1.234 140 Cretaceous-early Tertiary granodiorite
PE 87 0.7082 ±0.0003 0.2283 140 Cretaceous-early Tertiary diorite
PE 90 0.7060±0.0001 0.1854 140 Cretaceous-early Tertiary granodiorite
PE 92 0.7056±0.0003 0.0811 140 Cretaceous-early Tertiary diorite
PE 12 0.7046±0.0001 0.0168 105 Jurassic basalt (altered)
PE 114 0.7052±0.0002 n.d. 120 Jurassic basalt (altered)
PE 115 0.7058±0.0001 n.d. 120 Jurassic basalt (altered)
PE 117 0.7049±0.0002 n.d. 120 Jurassic basalt (altered)
PE 118 0.7040 ±0.00007 n.d. 120 Jurassic diabase
PE 18 0.7749 ±0.0001 0.971 110 Precambrian banded gneiss
PE 19 0.7724 ±0.0002 0.765 110 Precambrian banded gneiss
PE 37 0.7324±0.0001 0.264 145 Precambrian banded gneiss
PE 111 0.7379±0.0002 0.962 70 Precambrian banded gneiss
* Ages not radiometric.
254
CARNEGIE INSTITUTION
topic correction will be of the order of
0.004.
Histograms of 87/86 ratios for samples
within age categories are presented in
Fig. 50. This series of histograms illus-
trates that the 87/86 ratios tend to be
higher for younger rocks than for older
ones. The oldest rocks of the arc, Juras-
sic extrusives, exhibit the lowest isotopic
ratios and the youngest extrusives, late
Cenozoic in age, exhibit the highest
ratios.
The change in isotopic ratio correlates
not only with time but with depth to the
present-day Benioff zone and crustal
thickness. In detail, however, these cor-
relations fail to survive close scrutiny.
Fig. 51 shows, for example, 87/86 ratios
of late Cretaceous-early Tertiary and
late Cenozoic samples plotted against
depth to the present-day Benioff zone.
Within any given age group, no correla-
tion exists between isotopic ratio and
depth to the Benioff zone. The same lack
of correlation holds for crustal thickness.
All of the samples, with the possible
exception of some Jurassic extrusives, ex-
hibit higher 87/86 ratios than the aver-
age for island arcs. This result is in ac-
cord with other data that show that rocks
of cordilleran arcs consistently exhibit
higher Sr87/Sr86 ratios than rocks of is-
land arcs. Table 16 gives some repre-
sentative averages for island arc and
cordilleran arcs. The Andean rocks have
87/86 ratios that lie comfortably within
the range of values for cordilleran arcs.
If we grant that island arcs represent a
comparatively simple system of ande-
sitic magma genesis, then it is reasonable
to suppose that the isotopic ratios meas-
ured are of uncontaminated rocks. The
consistency between rocks of various arcs
supports this notion. Thus, if we are to
assume that the andesites of cordilleran
arcs are of similar origin to those of
island arcs, it is necessary to postulate
that the strontium isotope ratios have
been altered by contamination in the
passage of magma through sialic crust.
5
4h
3
2"
1-
LATE CENOZOIC
703 704 765 706 707 708 709 JlO
L.CRETACEOUS/ E. TERTIARY
.704
v 5
u
m 4
o J
« 2
.706
JURASSIC
.708
.710
.703 .704 .705 .706 .707 .708 .709 .710
Sr87/Sr86
Fig. 50. Histogram of Sr87/Sr86 ratios for
igneous rocks of various ages of coastal and
western cordilleran regions of southern Peru.
Sample ages are on basis of field determinations
and are not precise. All ratios are uncorrected
for age.
To explore at least one facet of the con-
tamination problem, and at the same
time to examine the question of possible
crustal origins for these rocks, we meas-
ured isotopic ratios of the Mollendo
gneiss of southern Peru — the only wide-
o
00
to
.709
.708
NV -707h
oo
£.7061-
.705-
.704-
70 80
100
120 140
h(km)
160 180 200
Fig. 51. Sr87Sr86 ratios plotted against depth
to the present-day Benioff zone for samples of
late Cretaceous— early Tertiary age and late
Cenozoic age.
DEPARTMENT OF TERRESTRIAL MAGNETISM
255
TABLE 16. Sr87Sr86 Ratios for Andesitic and Granite Rock Suites of the
Circum-Pacific Region *
Area
Rock Type
Average
Sr87/86
Reference Source
Marianas
andesite-dacite
0.7038
Hedge (1966); Pushkar (1968)
New Britain
andesite-dacite
0.7036
Peterman et al. (1970)
Cascade Range
Quaternary basalt and
andesite, N. Cal.,
Oregon, and Wash.
0.7035
Hedge et al. (1970)
Peterman et al. (1970a)
Central America
Quaternary andesite
0.7038
Pushkar (1968)
California
Cretaceous granitic average
of Sierra Nevada
batholith
0.707
Hurley et al. (1965)
Northern Chile
latite-andesite
0.7062
Hedge (in press) cited in
Pichler and Zeil (1972)
Southern Peru
Quaternary andesite
Cretaceous/E. Tert.
granitic rocks
Jurassic basaltic rocks
0.7068
0.7066
(0.7062)
0.7049
present study
present study
present study
15 Adapted from Dickinson, 1970, p. 825, Table 2.
spread pre-Mesozoic rocks exposed in the
cordillera. The isotopic ratios are plotted
against Rb/Sr ratios in Fig. 52. Shown
in parentheses beside each point is the
model age (in b.y.) for an initial ratio
of 0.710. The Sr87/Sr86 ratios of the
gneiss establish two important points:
1. The gneiss is almost certainly Pre-
cambrian in age and, thus, predates any
of the Paleozoic marine sedimentary
rocks of the eastern ranges.
2. The Sr87/Sr86 ratios are so high that
they preclude the possibility that the
Fig. 52. Sr87Sr86 ratios vs. Rb/Sr ratio of
samples of Precambrian (?) Mollendo gneiss.
Model ages (in b.y.) assuming initial ratios of
0.710 are shown beside each point.
gneiss has participated in any significant
way in the genesis of the igneous rocks of
the post-Paleozoic arc. The gneiss, or
rocks of similar isotopic composition,
does pose imporant possibilities for con-
tamination, for even minor assimilation
of such rocks by a rising andesitic magma
could increase isotopic ratios signifi-
cantly.
The rocks of Jurassic age exhibit pres-
ent-day ratios much closer to those of
island arc extrusives. Most of the Juras-
sic extrusives exhibit alteration which
may have increased the 87/86 ratio.
Moreover, no correction has been made
for radiogenic Sr87, although, in general,
Rb/Sr ratios appear to be low. The least
altered rocks typically give the lowest
87/86 ratios and analysis of compara-
tively fresh plagioclase separates from
the altered samples is now being under-
taken to ascertain whether isotopic com-
position varies. In general, however, the
Sr87/Sr86 ratios of the Jurassic rocks,
even without consideration of the effects
of alteration and radiogenic strontium,
are not substantially higher than those
ratios measured for rocks of modern
256
CARNEGIE INSTITUTION
island arcs. This conclusion is consistent
with what we believe to be the origin of
the Jurassic arc — namely, as an island
arc formed partly under water off the
coast of Mesozoic South America. The
arc may have been underlain by a thin
sialic crust or it may have formed en-
tirely upon simatic crust. Further meas-
urement of Sr isotopic compositions may
give us confidence to answer that ques-
tion more fully, for the key is whether or
not the ratios are within the range of
island arc values or whether, as now,
they lie between island arc and cordil-
leran arc averages.
The strontium isotope ratios that have
been presented provide for a few tenta-
tive conclusions:
1. The Jurassic arc may have formed
on simatic or thin sialic crust, and the
Sr ratios of the Jurassic extrusives are
close to those measured for rocks of
island arcs.
2. Rocks of post-Jurassic age exhibit
Sr isotope ratios significantly above those
for island arcs but very close to the
mean value of rocks of other cordilleran
arcs, such as the Sierra Nevada.
3. The very high (0.732 to 0.775) pres-
ent-day 87/86 ratios of Precambrian (?)
gneiss of the western cordillera exclude
the possibility that these rocks or any
older sialic rocks of similar isotopic com-
position have directly participated in the
genesis of the igneous rocks of the arc.
Potash variation within the western
cordillera. Dickinson and Hatherton in
a series of papers86'87,88 have shown
that potash concentration at given Si02
levels in modern volcanic rocks of island
arcs increases roughly linearly with
depth to the Benioff zone. This empirical
correlation has served as one of the bases
for the hypothesis that island arc mag-
mas originate in or near the Benioff zone.
It is further asserted by some, ourselves
included (Year Book 70, pp. 344-349),
that the magma has formed by partial
melting of the oceanic crust of the de-
scending lithospheric slab. The defense
of these speculations is not our objective
here, but they show the direction of our
thinking as regards potash variation, for
if potash varies with depth to the
Benioff zone, then it may prove possible
to reconstruct positions of past Benioff
zones from the measurement of potash
in igneous rocks. Before proceeding on
such an optimistic undertaking, how-
ever, it is instructive to consider the pot-
ash content of Cenozoic volcanic rocks of
the western cordillera with a view toward
determining how much significance can
be attached to K20-/i (depth to Benioff
zone) relationships within a given arc.
Following Dickinson, we first obtain
variation diagrams of K20 vs. Si02 for
various suites, then plot K20 at fixed
Si02 values for the appropriate depths
to the Benioff zone. All analyses re-
ported here are XRF and were com-
pleted at the University of Montreal
with the assistance of B. Gunn.
The late Cenozoic suites have been
selected to represent sampling sites that
lie at different elevations above the
Benioff zone. It should be noted, how-
ever, that all samples included here were
collected at or west of the crest of the
western cordillera, a distance spanning
considerably less than half of the total
east-west extent of Andean volcanic
rocks.
Variation diagrams of K20 vs. Si02
are presented in Fig. 53 for analyses of
late Cenozoic rocks of southern Peru, in
Fig. 54 for analyses of late Cenozoic
rocks of northern Chile,83, 84'90 and in
Fig. 55 for analyses of late Cretaceous-
early Tertiary intrusives of the southern
Peruvian part of the Andean batholith.
Jurassic volcanic rocks are not included
here because of paucity of analyses and
general uncertainty as to the extent of
change in initial K20 concentrations due
to later alteration.
The variation diagrams are self-
explanatory, but it is appropriate to
mention a few anomalies. The andesitic
series of Fig. 53(a) was collected from
the flanks of volcano Misti, a large ande-
sitic cone near Arequipa. The slope of
DEPARTMENT OF TERRESTRIAL MAGNETISM
257
§ 5
O 4
* 3h
2-
1 -
45
O
CM
6-
5
4
3-
2
1
50
. QUATERNARY
ANDESITE
O,' QUATERNARYC?)
IGNIMBRITE
50
55
60" 65"
Si02(o/0)
70
75
QUATERNARY VOLCANICS
hx150km
V*
* h=165 km
■ h=190km
55
60
65
70
75
% Si02
LTERT. VOLCANIC*
Fig. 53. (a) Variation diagram K20 vs. Si02 for late Cenozoic volcanic rocks of southern Peru.
Andesitic volcanic rocks were all taken from the flanks of Misti volcano near Arequipa and dis-
play a slope greater than that observed over the larger range of compositions. Ignimbrites form
the more silicic end members of this series, (b) Variation diagram K20 vs. Si02 for late Cenozoic
volcanic rocks situated along a zone east of sample localities of Fig. 53a and overlying a deeper
part of the Benioff zone.
increasing potash with silica is much
higher than for any other suite collected
in southern Peru and is much greater
than would be expected on the basis of
typical increases even within cordilleran
arcs. A second anomaly may be analyt-
ical, for Guest's 90 analyses show consist-
ently higher potash than those of Pichler
and Zeil for northern Chile (see Fig. 54).
The potash measurements of Guest are
consistent with our results for southern
Peru.
A plot of K20 vs. depth to the Benioff
zone is shown in Fig. 56 for each of these
suites at Si02 of 55%, 60%, 65%, and
70%. Where rock compositions within a
given suite did not extend to a given Si02
concentration, those points were not in-
cluded. Shown for comparison are the
approximate best (eyeball) fit K20-h
lines from the accumulated data pre-
sented by Dickinson (1970). 91 It is im-
mediately clear that no simple relation-
ship exists for these various suites
between potash and depth to the Benioff
zone. Indeed, within the limits of error
of the variation diagrams, there is no
significant difference in potash concen-
tration between the included suites.
There is some indication that the late
258
CARNEGIE INSTITUTION
4-
O 3
CN
h = 175
▲ GUESTU969)
• Z. & P. (1967)
P. & Z. C1969)
50
55
60 65
% Si00
70
75
Fig. 54. Variation diagram K20 vs. Si02 for
late Cenozoic volcanic rocks of northern Chile
(Zeil and Pichler. 1967; 83 Pichler and Zeil,
1969; 84 Guest, 1969 90). Guest's analyses tend to
show consistently higher KL>0 values than those
of Pichler and Zeil and could indicate analyt-
ical differences. Guest's analyses are chiefly of
ignimbrites and tend to lie along the best fit
line of Fig. 53a. Trend lines of Fig. 53a are
plotted for comparison. In general, potash
values, as measured by Pichler and Zeil, tend to
be lower than those observed for late Cenozoic
volcanic rocks of southern Peru.
Cenozoic suite at 190 km above the
Benioff zone exhibits slightly higher pot-
ash, but the increase is slight. In addi-
tion, potash concentrations for Si02 >
65% diverge markedly from the trend
lines of Dickinson's island arc volcanic
rocks. Interestingly, the late Cretaceous-
early Tertiary intrusive rocks exhibit
potash concentrations nearly identical to
those measured for late Cenozoic vol-
canic suites. This fact may indicate that
5
O 3
2- !*#*""• • ••
45
50
55
70
6 0 65
Si02(%)
L. CRET. / E TERT. INTRUSIVES
Fig. 55. Variation diagram K20 vs. Si02 for
intrusive granitic rocks of late Cretaceous-
early Tertiary (?) age for region west of Are-
quipa.
5
4
o
1 2 3
*
© e
' - " _ - 65 -
r©
_-©- -<5
6
*
e
__ „ »-
_ --o-
Si02
O 55
■ 60
© 65
* 70
100
150
200
h(km)
Fig. 56. Plot of K20 vs. h (depth to Benioff
zone) for various suites samples in southern
Peru and northern Chile. Values for h are
taken from James (1971). 92
the Benioff zone has not altered its posi-
tion dramatically since Cretaceous time.
The fact that the potash content of
Cenozoic volcanic rocks 150 km above
the Benioff zone does not differ signifi-
cantly from that measured for rocks 200
km above the Benioff zone suggests that
the extrusives are all derived from
roughly the same source region. The dif-
ferent extrusive sites may be a function
more of available fracture systems along
which magma can be transported than of
varying source region. The Cenozoic
volcanic rocks of northern Chile exhibit
the same phenomenon. Although not
shown here, we sought initially to split
the Chilean analyses into groups for
rocks 150 ± 25 km above the Benioff
zone and for those 200 ± 25 km above
the zone. The best fit lines for the two
variation diagrams proved identical.
We have not yet examined trends for
other trace elements across the arc. One
feature of special importance does arise
from the major element analyses: The
intrusive rocks of the Andean batholith
exhibit average compositions less silicic
than the average rock of the Cenozoic
volcanic arc. Mean compositions of the
intrusive rocks collected are presented in
Table 17, together with some mean com-
positions for representative andesitic and
intrusive suites. It should be noted that
the Andean volcanic suites do not include
any rhyolitic ignimbrites, and their mean
silica content is thus lower than would
DEPARTMENT OF TERRESTRIAL MAGNETISM
259
TABLE 17. Mean Chemical Compositions of Andesitic and Granitic Suites
1
2
3
4
5
6
7
8
9
Si02
57.1
59.3
58.3
60.4
57.8
62.2
64.6
69.3
59.7
A1203
17.3
17.4
17.5
16.7
17.1
16.5
16.7
15.5
17.1
Ti02
0.8
0.8
0.9
0.8
0.9
0.7
0.6
0.4
0.7
Fe203+FeO
7.9
7.7
7.8
6.9
7.8
6.3
5.1
3.3
7.5
MnO
0.2
0.1
0.1
0.9
0.2
MgO
4.3
3.5
3.3
2.3
2.8
2.0
2.4
1.2
2.6
CaO
8.4
7.2
7.4
5.4
6.3
4.9
5.3
3.2
6.1
Na20
3.1
2.9
3.2
4.0
4.2
4.0
3.3
3.4
3.5
K20
1.1
1.2
1.4
3.0
2.3
3.1
2.0
3.7
2.5
P2O5
0.2
0.3
0.2
1. Cenozoic volcanic rock of Little Sitkin Island, Aleutian Islands (Snyder, G. L., 1959, Geology
of Little Sitkin Island, Alaska, U. S. Geol. Surv. Bull. 1028-H, pp. 169-210; cited in Dickinson. 1970).
2. Average analyses of Cenozoic rocks from Kurile Islands (Markinin, E. K., 1968, Volcanism as
an agent of formation of the earth's crust, in The Crust and U^pper Mantle of the Pacific Area, Mono-
graph 12, 413-423, AGU, Washington, D. C; cited in Dickinson, 1970).
3. Average analyses of andesites from Kurile Islands (Gorshkov, G. S., 1970, Volcanism and the
Upper Mantle, Plenum Press, New York, 385 pp.).
4. Average composition of Cenozoic volcanic rocks (exclusive of ignimbrites) of Arequipa area of
southern Peru (28 analyses).
5. Average andesite (13 analyses) from Misti volcano near Arequipa, southern Peru.
6. Average dacite (15 analyses) of western flanks of the western ranges near Arequipa, southern
Peru.
7. Weighted average of Mesozoic plutonic rocks of southern California batholith (Larsen, E. S.,
1948, Batholith and associated rocks of Corona, Elsinore, and San Luis Rey quadrangles, southern
California, 182 pp., Geol. Soc. Amer. Mem. 29; cited by Dickinson, 1970).
8. Weighted average of Mesozoic plutonic rock of the Sierra Nevada batholith (F. W. Dodge,
unpublished data, cited by Dickinson, 1970).
9. Average of 34 analyses of plutonic rocks of the Arequipa area of southern Peru.
be obtained by averaging volcanics of the
late Cenozoic series. From Table 17 it is
apparent that the mean composition of
the intrusive rocks collected (some 34
specimens not representing a weighted
sample) is closer to a typical andesitic
series than a granitic series. As some
effort was made to sample rocks in ap-
proximate proportion to their occurrence,
we suggest that the Andean intrusives
are compositionally more akin to ande-
sitic suites than to batholithic granitic
suites. In particular, the composition of
Andean intrusives is significantly more
mafic than those of the Sierra Nevada
batholith. The similarity of composition
between the intrusive rocks of the
Andean batholith and the extrusive rocks
of the volcanic chain lends support to the
concept that granitic plutons are the
deep-seated comagmatic equivalents of
the extrusive rocks of the arc. Moreover,
it appears possible that the late Cenozoic
volcanic rocks of the central Andes will
prove to have average compositions more
silicic than the andesitic suites of island
arcs, possibly quite comparable to the
batholithic means for other cordilleran
systems.
The Geochemistry of Basalts from Ice-
land and the Reykjanes Ridge
Stanley R. Hart and Jean-Guy Schilling*
Basalts from spreading ocean ridges
differ in a number of ways from basalts
erupted on oceanic islands and island
arcs. While all three tectonic environ-
ments may produce low-K tholeiites, the
spreading ridge tholeiites are generally
characterized by unusually low concen-
trations of elements such as Rb, Cs, Ba,
* Narragansett Marine Laboratory. Univer-
sity of Rhode Island, Kingston, Rhode Island.
260
CARNEGIE INSTITUTION
and Sr, and by relatively low Sr87/Sr86
ratios.
The distinctive character of spreading
ridge basalts is generally ascribed to deri-
vation from mantle which is depleted in
these elements relative to the mantle which
acts as a source for oceanic island and
island arc basalts. In Year Book 70, pp.
353-355, we commented on the problem
of how two distinct mantle sources can
maintain their identities when both are
in convective motion and are in close
proximity. For example, many oceanic
islands occur near the crests of spreading
ridges and yet the basalts from both re-
gions are quite distinct.
In a study of one such transition
zone between an oceanic island and
a spreading ridge environment, Schil-
ling 93 studied a suite of basalts collected
from Iceland and dredged at intervals
along the Reykjanes Ridge south of Ice-
land. The basalts showed only minor
variations in .major element chemistry,
with most of the ridge basalts being
quartz tholeiites and the Iceland samples
being dominantly olivine tholeiites. The
basalts were quite different in minor and
trace element chemistry, however, with
K, P, Ti, and the rare-earth elements
(REE) decreasing fairly regularly from
Iceland to a point some 550 km south on
the Reykjanes Ridge. This variation was
accompanied by a progressive depletion
of the light REE relative to the heavy
REE. We report here an extension of
this study, to include the elements Rb,
Cs, Ba, and Sr and the isotopic composi-
tion of Sr. There is currently much inter-
est in the volcanology of Iceland because
of plume convection or mantle hot spot
models94,95 in which Iceland is consid-
ered the surface volcanic expression of
one such mantle plume. The geochemical
results reported here obviously have an
important bearing on the concept of
mantle plumes.
Figure 57 shows the location of sam-
ples from Iceland and the Reykjanes
Ridge which were analyzed for this
study. Sr87/Sr86 data (Fig. 58) show two
POST-GLACIAL
VOLCANIC ZONES
SAMPLE LOCALITIES
Fig. 57. Map of Iceland showing location of
the Reykjanes ridge spreading axis and loca-
tions of analyzed samples.
general groupings of values, with a rela-
tively sharp transition between them.
Samples from Iceland and for 200 km
south along the ridge show remarkable
uniformity, with an average value of
0.70304 (the standard deviation of these
13 samples is ±0.003%). This is the
most precise demonstration of regional
uniformity of Sr isotope ratios in vol-
canic rocks yet achieved. In the zone
from 200 to 250 km south on the ridge,
some fluctuation in Sr87/Sr86 values is
noted, whereas beyond 250 km south on
the ridge the values are consistently
lower by 0.05%, averaging 0.70270. Thus
the Sr isotope ratios from this area con-
form to the pattern observed throughout
the ocean basins: island volcanics in-
.7031
7030
7029
<? .7028
id
in .7027-
-
1 1
• •
1
•
1
•
••
•
•
•
•
1
•
•
•
1
1
-
1
•
-
-
95%
ERROR
LIMITS
1
1
1
•
•
• •
•
•
1 1
1
1
l
1
1
1
300
•ICELAND
IOO O 100
300
RIOGE -
Fig. 58. Plot of Sr87/Sr86 ratio as a function of
distance from southern tip of Reykjanes Penin-
sula on Iceland. Values are relative to 0.70800
for E and A Sr standard.
DEPARTMENT OF TERRESTRIAL MAGNETISM
261
variably having Sr87/Sr86 greater than
0.7030 and spreading ridge basalts show-
ing, with a few exceptions, Sr87/Sr86 less
than 0.7030 (the average Sr87/Sr86 for 18
spreading ridge basalts analyzed in this
laboratory is 0.70265). The isotopic data
demonstrate that there are two discrete
mantle sources being tapped for magmas,
one feeding Iceland and the northern 200
km of the Reykjanes Ridge, and the
other feeding the ridge from 250 km on
south, with perhaps some mixing or over-
lap in the region between 200 and 250 km.
The trends of selected trace elements
and trace element ratios are shown in
Figs. 59 and 60 as a function of distance
from the southern end of the Reykjanes
Peninsula in Iceland. The samples far-
thest south on the ridge (60° N) show
the depleted trace element characteris-
tics (for example, K and Sr, Fig. 59)
which are typical of spreading ridge ba-
salts. In fact, these samples have lower
concentrations of these elements than
any other spreading ridge basalts yet
analyzed. Spreading ridge basalts aver-
age about 1000 ppm K and 130 ppm Sr,
compared to 330 ppm K and 70 ppm Sr
for the southern Reykjanes Ridge ba-
salts. The elements Rb, Cs, and Ba and
the REE are also unusually low. Various
trace element ratios, on the other hand,
Fig. 59. K and Sr concentration as a function
of distance from tip of Reykjanes Peninsula.
Fig. 60. K/Sr and K/Ba ratio variation as a
function of distance from tip of Reykjanes
Peninsula.
are not as highly fractionated as average
spreading ridge basalt. For example,
spreading ridge basalts average — 1000 for
K/Rb, 75,000 for K/Cs, and 120 for
K/Ba, whereas the basalts from 60-61 ° N
on the Reykjanes Ridge average ^770,
59,000, and 60 for these ratios. If the low
trace element concentrations of typical
ridge basalts are ascribed to previous
depletion processes operating on the
mantle source, it seems that more ex-
treme operation of this depletion process
is much less effective in producing ele-
ment fractionation. It is also interesting
to note that mantle which has under-
gone this severe depletion in the LIL
(large-ion-lithophile) trace elements is
still capable of generating a tholeiitic
basalt that is normal as far as major ele-
ment chemistry is concerned.
The trace elements show a fairly regu-
lar increase in concentration ( and change
in ratios such as K/Sr and K/Ba, Fig.
60) going north toward Iceland. At 200
km, where the step in Sr87/Sr86 occurs, a
small but noticeable step also occurs in
the trace element trends. On Iceland
most of the samples closest to the ridge
have higher concentrations of the LIL
trace elements than any of the ridge
262
CARNEGIE INSTITUTION
samples. However, concentrations ap-
pear to decrease on going north across
Iceland, and the final range in concentra-
tions on Iceland is about the same as on
the ridge, though the average is some-
what higher on Iceland. Though some
of the Iceland tholeiites have element
concentrations as low as spreading ridge
basalt (for example, K = 700 ppm, Sr =
90 ppm) , the element ratios are quite un-
like those for basalts derived from de-
pleted mantle (for example, the K/Rb,
K/Cs, and K/Ba ratios of the Iceland
basalts are about 500, 45,000, and 32,
respectively, and are relatively uniform
and independent of actual element con-
centrations) . Thus, the delineation of
two discrete mantle source regions in the
Iceland area on the basis of Sr isotope
data is corroborated by trace element
ratios such as K/Rb and K/Ba (and
La/Sm1).
While the Sr isotope data would seem
to suggest only limited mixing of these
two mantle sources ( or the derived mag-
mas), the trace element concentrations
do show a progressive change with dis-
tance from Iceland which is suggestive
of a mixing process. This possibility can
be evaluated with two-element mixing
plots (Fig. 61 is one example) . This plot
and others like it suggest that the surface
and dredge samples are aligned on sepa-
rate trends and that mixing of a high-
concentration Iceland basalt with a low-
concentration ridge basalt is too simple a
model to explain all the variations ob-
0.06
0.04-
U 0.02
Ba/Cs~IIOO
20 30 40
Ba , ppm
50
60
Fig. 61. Cs content vs. Ba content of Iceland
basalts (open circles) and ridge basalts (solid
circles). Note approximate fit to trends with
different Ba/Cs ratio.
served along the ridge. Other, more com-
plicated mixing models have not been
evaluated but may be allowed by the
data. Note that two-element plots (such
as Fig. 61) suggest a division of the sam-
ples at the shoreline, not at the 200 km
break shown by the Sr isotope data.
A model which we are in the process of
evaluating quantitatively includes the
following features:
1. Existence of both depleted and un-
depleted mantle, with a fairly sharp
boundary about 200-250 km south of
Iceland. This boundary coincides with a
concave break in slope of the ridge crest.
2. Variable degree of partial melting
of these mantle sources, with the greatest
degree of melting occurring under both
the north edge of Iceland and the south-
ern end of the ridge.
3. Change in mantle mineralogy, with
the mantle underlying the submarine
ridge containing a phase such as plagio-
clase which is absent under Iceland itself.
This feature is required to explain the
two different trends shown by the trace
elements (such as those shown in Fig. 61)
and to explain the large scatter of Sr con-
centrations for Iceland basalts as com-
pared to the ridge basalts.
4. Near-surface crystal differentiation
will produce minor variations in trace-
element concentrations.
Geochemical data such as these ob-
viously cannot prove the existence of a
deep-mantle plume under Iceland; how-
ever, the clear delineation of two mantle
source regions is consistent with a plume
model for Iceland. These data show that
such a plume does not feed significant
quantities of material into the spreading
ridge beyond about 200 km of Iceland.
Geochemistry of Ultramafic Inclusions
from Salt Lake Crater, Hawaii
N. Shimizu
Trace elements in basalts have been
studied by many investigators because
they offer important information on the
chemistry of the upper mantle. Another
DEPARTMENT OF TERRESTRIAL MAGNETISM
263
possible link to upper-mantle chemistry
is the ultramafic inclusions in basalts
and kimberlites. Judging from the major-
element compositions, some of these in-
clusions appear to be potential sources of
basaltic magma; others appear to be
residua left after partial melting or crys-
tal accumulates from basaltic magma.
To better understand the trace element
chemistry of these rocks and especially
to understand the possible correlation be-
tween major element and trace element
chemistry, the samples to be analyzed
should be petrologically well documented.
Preferably, a suite of samples collected
from a single locality should be studied.
Ultramafic inclusions found in an
olivine nephelinite tuff of Salt Lake
Crater, Hawaii, are among the best sam-
ples in these respects. Seventeen rock
samples have been made available for
the study, all of which were described by
Kuno (1969) 9G and Kuno and Aoki
(1970).97 On petrological grounds they
are classified into three groups: lherzolite
(olivine + orthopyroxene + clinopyrox-
ene + spinel), olivine eclogite (olivine
-j- clinopyroxene + garnet ± spinel ±
orthopyroxene) and olivine-free eclogite
(clinopyroxene + garnet ± orthopyrox-
ene ± spinel). Olivine-free eclogite is
better called garnet pyroxenite, because
clinopyroxene in this rock type is not
quite omphacitic (White. 1966; 98 Jack-
son and Wright, 1970 ") .
K, Rb, Cs, Sr, and Ba abundances and
Sr87/Sr86 were measured in separated
clinopyroxenes because in these mineral
assemblages the above trace elements are
largely concentrated in clinopyroxene.
The mineral separation was made only
by handpicking. Grains surrounded only
by fresh surfaces and free from any
visible inclusions or alteration product
were carefully selected under a binocular
microscope, and this process was repeated
several times with successive grinding
into finer grain size. The separates were
then washed with acetone, wrarm 2.5 N
HC1 with ultrasonic agitation, and dis-
tilled water prior to processing. The sam-
ples were processed with low-level blank
chemistry procedures. The total blanks
are: 7 ng K; 0.01 ng Rb; 0.3 pg Cs; 0.2
ng Sr; 0.8 ng Ba. These low-level blanks
made it possible to handle a small
amount of sample, for example, 10-20
mg, which, in turn, made it easy to obtain
extremely pure mineral separates. The
chemistry blanks are negligible, except
for Ba in a few cases.
In general, concentrations of these ele-
ments in clinopyroxenes are low com-
pared with those in the same mineral
from the same locality reported in the
literature (Griffin and Murthy, 1969; 10°
Philpotts et al, 1972 101 ) . Figures 62 and
63 show Rb-Sr and K-Rb relationships.
Arrows in these figures indicate the
trends of increasing whole-rock MgO/
SFeO ratios. From these figures, the
trace element variations appear to be
interrelated, and particularly in Fig. 62,
three groups of rocks fall in separate re-
gions of different Rb-Sr characteristics.
Rb
ppm
.2
05
02
.01
-[ 1 1 1 — i — r-r
1 1 r
LHERZ
OL ECL
i i i i i
50 100 200
Sr, ppm
Fig. 62. Rb vs. Sr plot for clinopyroxenes
from ultramafic inclusions of Salt Lake Crater.
Open circles, lherzolites; solid circles, olivine
eclogites; crosses, olivine-free eclogites. Ar-
rows indicate the trend of increasing whole-rock
MgO/2FeO.
264
K
ppm
100
50 -
20
10
-i 1 1 1— -i — i i i
CARNEGIE INSTITUTION
-i 1 1 1 — i — i — r~r
2000
1000
500
.----o
LHERZ.
j i i i i i i
J I L
I I
.01 .02 .05 .1 .2 Rb, ppm
Fig. 63. K vs. Rb plot. Symbols are same as those in Fig. 62.
The relationship between major and
trace elements is better illustrated in Fig.
64, in which trace element abundances in
clinopyroxenes are plotted against whole-
rock MgO/^FeO ratios. The correlation
is particularly good in olivine eclogites.
It is rather surprising that even at sub-
ppb concentration levels (Cs in olivine
eclogites) , a trace element appears to
behave quite systematically relative to
major elements. This may indicate that
contamination from the enclosing magma,
if any, has been very minor. Correlation
also exists between the trace elements
and other major elements because the
oxide components in these rocks vary in
a systematic manner with respect to
MgO/SFeO ratio (Kuno, 1969). The
whole-rock MgO/3FeO ratio was chosen
here because it should be a good indicator
of magmatic processes ( such as partial
melting or fractional crystallization).
According to Kuno (1969) 96 and Kuno
and Aoki (1970), 97 the lherzolites would
represent residua left after partial melt-
ing, and the eclogites (both olivine-
bearing and olivine-free) would represent
crystal accumulates from the basaltic
magma. Following the idea that these
rocks represent the solid phases which
were once equilibrated with liquid, the
trace element variations illustrated in
Fig. 64 may be explained in terms of
partition relations of these elements be-
tween solid and liquid. An important
feature in this context is the increase of
Sr with increasing MgO/^FeO observed
in the lherzolites and olivine eclogites.
If MgO/^FeO represents the degree of
partial melting (the higher the ratio, the
greater the degree) or the degree (or
stage) of crystallization (the higher the
ratio, the earlier the stage), the increase
of Sr can be interpreted in two ways:
that Sr was more enriched in residual
clinopyroxene left after the larger degree
of partial melting or that Sr was more
enriched in cumulative clinopyroxene
crystallized in the earlier stage of frac-
tionation. These interpretations implic-
itly assume that the present mineral
assemblage represents the primary as-
semblage at the time of solid-liquid
equilibrium. Both interpretations re-
quire Sr to be partitioned more into
clinopyroxene than liquid. However, this
is quite unlikely in the light of the avail-
able data on Sr partitioning (Philpotts
DEPARTMENT OF TERRESTRIAL MAGNETISM
265
| |_|C-R7p)| ITC
— - /'"M C~/'~" 1
K, Sr
LnL i\l W L_ 1
2 45 6
8910 12
UL t.(wL.
17 21
27 30
ppm
( III
I I \(
| |
1 ^
4 00
-
9
.00 3'"-* 1
/
Cs /
/
/
/
/
/
200
/
/
0021 Y
1
1
^ Sr
i
I 1
cr
7
/
' 1
100
^
* -—
1
. /
Rb
\
x '
/
/'>
- /
o"'
/
/
r •". 0004
/
/
.' /
/
1
1
1
1
1 _
/ f
50
Rb
-4 :-
r—
i >
K
i
i
j
K
i :~
i ■
4
K" v —
'••x' "'■•.
20
i i
\.
-4 --—
/
i
,' -x
xx
x Rb
i i
>
Rb
ppm
- .2
. 1
05
I .02
4.5 4 3.5 3
MgO / FeO"
2.5
1.5
Fig. 64. Variations of trace element contents of clinopyroxenes with respect to whole-rock
MgO/2FeO. Sample numbers for lherzolites and olivine eclogites are given at the top of the
figure. Cs in no. 30 is 0.0039 ppm, in no. 12 is 0.0004 ppm.
and Schnetzler, 1970; 102 Shimizu and
Akimoto, 1971 103). In addition, Sr con-
tents (as high as 140 ppm) and low K/Sr
ratios (about 0.3) are difficult to recon-
cile with clinopyroxene-liquid relation-
ships.
Therefore, it is indicated that the pres-
ent mineralogy is a recrystallized assem-
blage and that there must have been a
phase at the time of the primary solid-
liquid equilibrium into which Sr was pre-
ferentially partitioned. This phase must
have been stable only at higher tempera-
ture and must have disappeared during
recrystallization at subsolidus tempera-
tures. One possibility for this phase from
the standpoint of trace element geochem-
istry is plagioclase, because Sr is strongly
partitioned into plagioclase relative to
liquid (Berlin and Henderson, 1968; 104
Philpotts and Schnetzler, 1970; 102 Drake
and Weill, 1972 105). A possible positive
Eu anomaly due to plagioclase eval-
uated by a model calculation based
on the normative plagioclase contents
and the partition coefficients was
266
CARNEGIE INSTITUTION
found to be about 5%, being practically
invisible in the chondrite-normalized
REE diagram. The problem, then, is
whether this is plausible in the light of
phase petrology.
The stability of plagioclase at high
pressures has been studied in both syn-
thetic and natural systems of basaltic
composition in connection with the
basalt-eclogite transition (Kushiro and
Yoder, 1966; 106 Green and Ringwood,
1967; 107 Ito and Kennedy, 1971 108) and
in systems of lherzolitic composition in
connection with defining the mantle min-
eral facies (Green and Hibberson,
1970; 109 Herzberg, 1972 110).
There is no way to estimate the P and
T of the primary equilibrium, but, as-
suming that the subsequent recrystalli-
zation took place by isobaric cooling, the
estimated P and T for these rocks should
be reasonably close to the stability field
of plagioclase-bearing assemblages re-
ported by those investigators. Pressure
and temperature conditions for reequilib-
rium were therefore estimated for lherz-
olites by the major element compositions
of pyroxenes. The major element analy-
sis was made by a Material Analysis
Corporation electron probe at the Geo-
physical Laboratory.
The method used here is the same in
principle as that discussed by Boyd
(1973).111 Temperature was estimated
from the diopside composition on the
diopside-enstatite solvus determined for
the pure binary system by Davis and
Boyd (1966) 112 at 30 kb. As shown in
Fig. 65, the effect of iron was considered
by extrapolating the diopside composi-
tions in the quadrilateral to the binary
join Di-En. The effect of A1203 was,
however, ignored. Although the effect of
A1203 on the temperature estimate seems
to be negligible at 30 kb as inferred from
the data by Boyd (1970), 113 it is totally
unknown at pressures in the stability
field of spinel instead of garnet. Pres-
sure was estimated from the A1203 con-
tent of enstatite, using the data by Mac-
Fig. 65. Compositional relations of olivines,
orthopyroxenes and clinopyroxenes in lherzo-
lites. Temperatures ( 1000-1300° C) are from
Davis and Boyd (1966).112
Gregor (1973) 114 on the system MgO-
Al203-Si02. According to Boyd (1970),
at 30 kb and 1200°C there is about 1%
difference in A1203 between two enstat-
ites, one at the binary join En-Al203 and
the other at the apex of the three-phase
triangle En-Di-Gar. At lower pressures
this difference is presumably greater.
Therefore, 1.5% A1203 was added rather
arbitrarily to the analyzed values, and
pressure was then estimated from the
A1203 isopleths. Figure 66 illustrates the
estimated P and T for lherzolites to-
gether with some relevant phase bound-
aries. As shown in the figure, the esti-
mated conditions are reasonably close to
the stability limit of plagioclase in lherz-
olitic composition. It thus seems possible
that plagioclase participated in the pri-
mary equilibrium for these lherzolites.
Another problem is whether or not this
also holds true for the olivine eclogites.
Since the major element compositions of
the olivine eclogites are neither lherz-
olitic nor basaltic, the phase boundaries
shown in Fig. 66 cannot be applied with-
out modification. As discussed by Green
and Ringwood (1967), important factors
controlling the positions of boundaries
would be a complicated combination of
the MgO/MgO + FeO of the rock, the
DEPARTMENT OF TERRESTRIAL MAGNETISM
267
T,
i 1 1
III"
°C
/
1400
/
'lherz. dry
1200
'< \i
-
-
'^K
LHERZ.WET
1000
1 1 /^^^LHERZ. Px
■ >''
-
A B /
800
C(FOg2+AN59
PYROLITE
A: GAR IN,
) OL THOL.
B : PL OUT
i i i
10
15
20
25
P, kb
Fig. 66. Estimated P-T conditions of reequi-
librium for lherzolites. The wet and dry solidi
of a lherzolite are from Kushiro et at. (1968).115
The phase boundaries A and B are from Green
and Ringwood (1967)107 and C from Green and
Hibberson (1970).109
degree of silica saturation, the normative
plagioclase composition, and the molecu-
lar ratio of normative olivine and hypers-
thene to plagioclase. It is impossible to
discuss this possibility in great detail
until the phase relations are studied ex-
perimentally on the specific composition
of olivine eclogite. However, it seems
important to note the relative position of
the boundaries B and C in Fig. 66. These
are remarkably close to each other in
spite of the difference in chemical com-
positions. The figure also indicates that
at 1000°C, for example, the eclogitic
assemblage is stable even at lower pres-
sures in basaltic composition than the
spinel lherzolite assemblage in lherzolitic
composition. Since the major element
compositions of the olivine eclogites are
of intermediate nature between lherzolite
and olivine tholeiite in many respects, it
may not be unreasonable to assume that
the relative position of boundaries B and
C still holds true for the olivine eclogite
and the lherzolite now in question. The
molecular ratio of normative olivine to
plagioclase is about 1 in olivine tholeiite
B (Green and Ringwood, 1967), while
those in olivine eclogites are definitely
higher. This would be a good indication
that these rocks contain excess olivine,
assuming that both breakdown of plagio-
clase and formation of garnet are essen-
tially reactions between equal numbers of
molecules of plagioclase and olivine. The
presence of excess olivine components in
olivine eclogites suggests that olivine is
stable at any pressures of present
interest.
From these arguments, it seems possi-
ble that the hypothesis based on the trace
element variation also holds true for the
olivine eclogites. Judging from the pres-
sure, the liquid in equilibrium with those
rocks may well have been tholeiitic
rather than alkalic.
In contrast to this, the olivine-free
eclogites (or garnet pyroxenites) may
be of a different origin. Green (1966) 116
noted the similarity in chemical composi-
tions between the liquidus clinopyroxene
crystallized from an alkali basalt at 13-
18 kb and a whole-rock eclogite (No. 18
of Kuno, 1969). The texture also sug-
gests that garnet and orthopyroxene are
exsolved from clinopyroxene. Green
(1966) then concluded that the eclogite
of this type was originally monomineralic
clinopyroxene crystallized from alkali
basalt at high pressures which later ex-
solved orthopyroxene and garnet during
cooling. The sample used in his discus-
sion was also studied here (No. 18 in Fig.
62). The concentrations of trace ele-
ments in a hypothetical liquid calculated
from the measured trace-element abund-
ances in clinopyroxene and partition
coefficients experimentally obtained at
high pressures are 7000 ppm K, 10 ppm
Rb, and 800 ppm Sr, which are reason-
able as alkalic basalt. This appears to
support Green's argument on the origin
of this type of rock.
268
CARNEGIE INSTITUTION
Differential Dissolution Technique
(DDT): Chemical Separation of
Crystals from Glass
N. Shimizu and S. R. Hart
One of the most serious problems in the
experimental studies on partitioning of
trace elements between silicate crystal
and liquid is the separation of phases
from the small amount (less than 20 mg)
of high-pressure run product. Even if the
grain size is large enough for separation
by the conventional heavy liquid tech-
niques, it is impossible to eliminate crys-
tals with a very thin lining of glass on the
surface or those with tiny glass inclu-
sions. Even 0.5% impurity would change
the results significantly, because the
trace elements now in question (K, Rb,
Cs, Sr, and Ba) are strongly partitioned
into liquid relative to crystal. In order to
remove this difficulty, a new technique
has been developed and is presented here.
This technique is based on the fact
that clinopyroxene and glass (quenched
liquid) are dissolved by dilute HF at
significantly different rates. Several tests
using natural clinopyroxenes and glass
showed that glass is dissolved much
faster than clinopyroxene. Therefore, in
principle, starting with a mixture of
clinopyroxene and glass, we can separate
them by putting the mixture in dilute HF
for an appropriate period of time to dis-
solve only the glass. By repeating the
treatment several times, only pure clino-
pyroxene is left behind. If the amounts
of trace elements in the solution are
measured, and if the weight loss of the
sample during the treatment is measured,
we can calculate the concentrations of
elements in the dissolved sample (glass).
Ideally, in successive treatments the
amounts of trace elements in the solution
should come down to the chemical blank
levels before any clinopyroxene is dis-
solved. If, however, clinopyroxene is dis-
solved simultaneously, or if any selective
leaching of trace elements from clino-
pyroxene occurs, the amounts of trace
elements in the solution would never
come down to the blank level.
In order to test this technique, we pre-
pared a mixture of clinopyroxene and
glass (clinopyroxene from a spinel lherz-
olite from Salt Lake Crater, Hawaii;
tholeiitic basalt glass from the Juan de
Fuca Ridge) in a proportion so that the
simultaneous dissolution or the selective
leaching of trace elements from clinopy-
roxene could be detected. In this mix-
ture, the amount of Sr in the clinopyrox-
ene is more than twice that in the glass,
and Ba in the clinopyroxene is about
equal to that in the glass. A detailed
experimental setting is described below:
A pulverized mixture of clinopyroxene
(6.5 mg) and glass (2.8 mg) was put in a
Teflon centrifuge tube and 0.5 ml dilute
HF (4.9%) and a few drops of 2.5 N HC1
were added. After leaving the mixture
for 20 minutes at room temperature, the
solid was centrifuged down and the solu-
tion was withdrawn by a Teflon pipette
and transferred to a Teflon beaker. The
residue was rinsed three times with 2.5 N
HC1 and then three times with H20 and
the rinses were added to the Teflon
beaker. Spike solution was weighed into
the beaker and the solution was evapo-
rated to dryness. It was then dissolved
with 2.5 N HC1 and the elements were
separated by a cation-exchange resin col-
umn and analyzed by a mass spectrom-
eter. The residual solid was dried in the
centrifuge tube and weighed to determine
the weight loss of the sample during this
treatment. The treatment was repeated
three times and the amounts of the trace
elements in the successive solutions came
down to levels close to the chemical
blanks.
The results are summarized in Table
18 and illustrated in Fig. 67. As shown
in the figure, the amounts of trace ele-
ments in the solution are close to the
blanks at the third step. The largest dif-
ference was observed for Sr and may be
explained by the persistence of about 5
fig of glass (0.2% of that originally pres-
ent) . The concentrations of these ele-
DEPARTMENT OF TERRESTRIAL MAGNETISM
269
TABLE 18. Experimental Results
D10F*
DIOt
DDT,
1st, 2nd, and 4th
steps
Element,
1
2
4
ppm
acetone
SRH
CPX
glass
glass
cpx
K
1873
1671
32.0
1900
1733
27.8
Rb
2.54
1.84
0.063
2.50
1.95
0.054
Cs
0.076
0.021
0.083
0.022
0.0033
Sr
119.7
111
134.1
125.7
109.7
121.8
Ba
28.3
19.45
7.44
31.0
18.8
7.41
*Basalt glass D10 from Juan de Fuca Ridge, pulverized in reagent grade acetone.
fSame glass not pulverized with acetone.
ments were calculated and given in the
table. It is interesting to note that glass
1 is enriched in these elements relative to
glass 2. The glass was initially pulver-
ized in reagent grade acetone, so it might
have been contaminated by the acetone.
Glass 1 is in excellent agreement with
D10F (pulverized in acetone) and glass
2 with that not pulverized in acetone. It
is likely that the contaminant went into
the first solution and the pure glass (free
from contamination) was dissolved in
the second step. In the fourth step, the
remaining sample was dissolved with the
normal HF and HC104. The concentra-
tions are in good agreement with the
clinopyroxene of the starting material.
These results indicate that clinopyrox-
ene is clearly separated from glass in this
technique and that selective leaching of
trace elements from the pyroxene, if any,
is insignificant.
In the course of the experiments, we
observed armoring of undissolved grains
10 r
K
p9
1.0 -
.01
1.0 r
pg
.01
.012 5
.001
Ba
.01
.001
.00089
Rb
ng
10
1.0
.0021
.026
Fig. 67. Variations of trace elements in solutions during successive HF treatments. Dashed
lines with B designate blank levels of these elements.
270
CARNEGIE INSTITUTION
with insoluble precipitate (presumably
CaF2) and adsorption of dissolved ions
on the surface of remaining grains, but
these were removed by adding a few
drops of 2.5 N HC1 to the HF and by
rinsing the sample with 2.5 N HC1 after
the HF was withdrawn.
This technique was originally designed
to study the partitioning of trace ele-
ments between clinopyroxene and liquid
and is found to be applicable to other
problems, such as those described by
Shimizu and Kushiro (this Report, be-
low). An advantage of this technique is
that one does not have to worry about
the grain size of the samples. During
high-pressure synthesis of clinopyroxene,
it is generally difficult to obtain a grain
size large enough to do fission track map-
ping or even microprobe analysis very
easily. A disadvantage is that it is not
applicable to systems containing more
than two phases. It is not certain at this
stage whether the technique is directly
applicable to a system such as amphibole-
liquid, or whether some other acid or
more diluted HF would be needed.
Trace Element Content of Liquid Formed
by Partial Melting of a Garnet
Lherzolite at High Pressures:
A Preliminary Report
N. Shimizu and I. Kushiro
Many experimental works have been
carried out on the genesis of basalt mag-
mas from major element chemistry, but
very little has been done on the trace
element behavior during melting proc-
esses at high pressures. This was due to
the difficulty of separating phases by
conventional magnetic and heavy liquid
techniques from fine-grained run prod-
ucts. An attempt has been made to over-
come this difficulty by developing a new
technique of separation. A detailed ac-
count of the technique (DDT, differen-
tial dissolution technique) is given else-
where in this annual report (Shimizu and
Hart, pp. 268-270). Applying this tech-
nique, we have analyzed trace element
concentrations in liquid formed by par-
tial melting of a garnet lherzolite at high
pressures. Although the present paper
reports only some preliminary results,
more detailed study along this line would
provide possible explanations of trace
element variations in basaltic magmas
and a quantitative treatment of basalt
genesis from the standpoint of trace ele-
ment geochemistry.
The material used in this study is a
sheared garnet lherzolite in kimberlite
from Thaba Putsoa Pipe, Lesotho. The
sample was supplied by Dr. F. R. Boyd
of the Geophysical Laboratory. Major
element chemistry of whole-rock and
constituent minerals was described by
Nixon and Boyd (in preparation) . High-
pressure melting experiments were car-
ried out with a solid-medium piston cylin-
der apparatus using a graphite crucible
as a sample container. Phase relations
and major element compositions of co-
existing phases were given by Kushiro
(1973) ,117
K, Rb, Cs, Sr, and Ba were analyzed
by isotope dilution on the glass (quenched
liquid) of two samples, one run at 15
kb, 1450 °C, and the other at 20 kb,
1500 °C (both anhydrous) by sepa-
rating the glass from crystals with the
differential dissolution technique. Al-
though the chemical blanks are negligi-
ble, a large error is involved in the
weighing of samples of 1 mg or less. The
overall uncertainty in concentrations is
estimated to be 10-20%.
Table 19 summarizes the trace element
contents of whole rock, clinopyroxene
separated from the same rock, and two
glasses formed at the conditions given
above. The rock appears very fresh as
an inclusion in kimberlite; nevertheless,
the whole-rock trace element system has
been affected by alteration judging from
the unusually high Rb/Sr ratio for a
mantle material. Cs could not be meas-
ured because of unfavorable spiking ratio
(133/135 > 100). From K/Cs ratio in
one of the glasses (liquid A) and from K
DEPARTMENT OF TERRESTRIAL MAGNETISM
271
TABLE 19. Trace Element Concentrations
Liquid A
Liquid B
PHN 1611
15 kb,
20 kb,
Element
Whole-Rock
CPX
1450°C
E.F.*
1500°C
E.F.
2.4
A/Bt
K, ppm
1442
433.6
5638|
3.9
3476 §
1.6
Rb, ppm
7.17
0.196
35.7
5.0
22.5
3.1
1.6
Cs, ppm
0.00328
2.07
Sr, ppm
38.4
95.28
164.3
4.3
100.3
2.6
1.6
Ba, ppm
39.8
8.29
262.4
6.6
131.5
3.3
2.0
K/Rb
201
2212
158
154
K/Cs
132195
2724
K/Ba
36.2
52.3
21.5
26.4
K/Sr
37.5
4.55
34.3
34.7
Rb/Sr
0.187
0.0021
0.217
0.224
Sr/Ba
0.966
11.49
0.626
0.763
Sample wt.
(mg) anal.
41.8
14.5
1.0
0.7
*Enrichment factor, glass/starting material.
jConcentration ratios of liquid A/liquid B.
tK20 = 0.68%; microprobe value = 0.70%.
§K20 = 0.42%; microprobe value = 0.47%.
in the whole rock, Cs in the whole rock
would be as high as 0.5 ppm.
Berg (1968) 118 noted that high Cs and
resultant low K/Cs in kimberlite inclu-
sions are associated with secondary zeo-
lite (probably analcime) . The presence
of zeolite also increases Rb and Rb/Sr.
The situation may be similar for the sam-
ple used in the present experiments, al-
though zeolites have not been observed.
In Table 19, K20 values determined in-
dependently by microprobe and DDT
are compared. They are in excellent
agreement considering the possible 10—
20% weighing error of the present experi-
ment.
If a liquid is in equilibrium with a
solid phase in which a particular trace
element is preferentially partitioned, or
if two trace elements are effectively frac-
tionated by the solid phase, trace element
abundances (absolute as well as relative)
in the liquid are to some extent con-
trolled by this phase. In the present case,
however, the major solid phases coexist-
ing are olivine, orthopyroxene, and
clinopyroxene for liquid A; and olivine
and orthopyroxene for liquid B. Con-
ceivably these trace elements may not be
fractionated by these phases. It is ex-
pected, therefore, that the concentrations
of trace elements in the liquid are de-
pendent only on the degree of partial
melting. In other words, the enrichment
factors for all elements are equal. This
appears to be approximately correct, al-
though there are some variations of en-
richment factors (Table 19) which re-
main unexplained at this stage. For
example, in both liquids A and B, Rb
and Ba seem to be more enriched than K
and Sr. Judging from the concentration
ratios of the two liquids (A/B in Table
19), enrichment of K, Rb, and Sr is
rather straightforward, whereas Ba does
not behave as systematically (which
might partly be due to inhomogeneity of
the starting material with respect to Ba) .
As shown by Kushiro (1973), the
liquids are tholeiitic in major element
composition (quartz tholeiite for A and
olivine tholeiite for B), while trace ele-
ments in these liquids are more compli-
cated. For example, although the concen-
tration levels of K, Sr, and Ba are in the
range shown by many tholeiites, the
ratios of the elements are anomalous (i.e.,
low K/Rb and Sr/Ba, high K/Sr) . These
peculiarities are undoubtedly due to the
anomalous trace element pattern of the
starting material.
272
CARNEGIE INSTITUTION
It may be pointed out, however, that
it is possible to obtain concentration
levels of trace elements in natural basalts
by direct partial melting of a lherzolite
with concentrations of these elements
similar to those in the garnet lherzolite
used in the present experiments. If this
is true, about 1% phlogopite may be
present as a primary host phase for the
alkali trace elements and Ba. If phlogo-
pite does not persist during partial melt-
ing of 20-40% at 15-20 kb pressures, the
liquid should contain about the same
amounts of trace elements as those meas-
ured in the liquids of the present experi-
ments. In this case there is no need to
consider wall-rock reaction (Green and
Ringwood, 1967) 119 or eclogite fractiona-
tion (O'Hara, 1968) .12° If, however, the
trace element content in the upper man-
tle is much lower, then less partial melt-
ing is required, or it is necessary to con-
sider that some process such as wall-rock
reaction or eclogite fractionation is oper-
ating.
Trace Element Contents of Clinopyrox-
enes from Garnet Lherzolites in
Kimberlites
N. Shimizu and F. R. Boyd
The garnet lherzolite is widely as-
sumed to be a major constituent of the
upper mantle. It is actually abundant
among ultramafic inclusions found in
kimberlites. The recent works by Boyd
and Nixon (1972) 121 and Boyd (1973) 122
revealed that there are two groups of
garnet lherzolites, the sheared nodules
and the granular ones. The sheared nod-
ules appear to be more primitive upper
mantle material, judging from the major
element compositions, and can be a
potential source of basalt magmas (Ku-
shiro, 1973), 123 while the granular nod-
ules are depleted in the basaltic compo-
nents (CaO, A1203, and FeO relative to
MgO).
The trace element characteristics of
garnet lherzolites are therefore impor-
tant for a better understanding of the
upper mantle chemistry and processes
operating in the upper mantle. There
have been only a few trace element data
on these nodules available (Griffin and
Murthy ,1969; 124 Allsop et al, 1969; 125
Hutchison and Dawson, 1970 ;126 Phil-
potts et al., 1972 127). The sheared nod-
ules, in particular, have not been studied
because they are rare compared with the
granular ones and their chemical charac-
teristics have only recently been recog-
nized. Two technical difficulties that
were observed by these authors are the
low-level concentrations of trace ele-
ments and the difficulty of obtaining pure
mineral separates.
In this article, we report K, Rb, Cs,
Sr, and Ba abundances and Sr87/Sr86
ratios of clinopyroxenes separated from
garnet lherzolites. Our samples are pre-
dominantly from the Thaba Putsoa kim-
berlite pipe, Lesotho, with one (JJG 352)
from the Bultfontein kimberlite pipe
(supplied by Dr. I. Kushiro of the Geo-
physical Laboratory). Two of them,
PHN 1611 and 1596, are sheared nodules,
and PHN 1569 and JJG 352 are typical
granular nodules. PHN 1600E4 is a
discrete nodule, a diopside single crystal
about 2 cm in its smallest dimension, and
we analyzed the central part and the edge
(C and A, respectively) to find if there
are any significant variations within a
grain. A garnet separated from PHN
1596 was also analyzed for comparison.
Clinopyroxenes were selected for inten-
sive study because trace elements are en-
riched in this mineral relative to other
primary phases except phlogopite. The
mineral separation was made only by
handpicking under a binocular micro-
scope. The separates were washed with
warm 2.5 TV HC1 and water prior to the
chemical processing. The chemical blanks
are : 7 ng K ; 0.01 ng Rb ; 0.3 pg Cs ; 0.2 ng
Sr; and 0.8 ng Ba (negligible except in
garnet) (see Table 20). These low-level
blanks made it possible to handle small
amounts (less than 20 mg) of extremely
pure samples. Sr87/Sr86 ratios were
DEPARTMENT OF TERRESTRIAL MAGNETISM
273
2
c
©
CM
pq
GO t>
(M
-H
CO
03
pq
h
CO
M
CO
Ph
CO
oj
pq
M
d
cc
o
o
O
-^
M
a
a>
s
a
3
M
03
pq
CO
O
Si
P4
M
o
TEL
S
ca
CO
rH
-H
cm
CM
oo co
CM rH
id oo
OS CO
CM CO
rH CO
o o
o o
© d
rH CD
COOIN
CM i—l i-H
OON
N CO iO
d co h
CO CO ^
TjH ^H t^H
-H
-H -H
(M
iO
CO
o
CM
rH
CO
o
00
CO
o
00
io
o
d
d
d
d
lO OS
CM
OS
CM
r-3 d
1— I
00
OS
CM
1—1
CM
OS
OS
i—(
CO
CM
i— i
Tj5
iO
CM
i— i
1— 1
o oo
O rH
1— 1
o
o
o
CO
O
o
CM
CO
o
o
o
CO
O
o
o
rH
l— H
o
o
00
o
o
o
d d
d
d
d
d
d
d
iO
io io
CM
00
rH
rH
p
O
CM
CM
CO
CO
CO
CO
i— 1
CM
CM
d
d
CM CO
CM
r^
t^
iO
t^
CM
rH
-t-3
03
id cm
d
CO
CO
o
iO
i—i
£
CO iO
CO
r— 1
rH
CM
I-H
T3
C
03
00 iO
00
O
i—i
rH
o
rH OS
1—1
o
O
IO
i— i
iO i— i
00
iO
o
CO
00
w
CM CM
1— 1
CM
iO
CM
o
O CO
o
1— 1
o
CM
1— 1
^
T— 1 1— 1
1— 1
CM
rH
<M
CO
iO
CM
rH CM
1— 1
1—1
CO
rH
O
OS
O
o
CO i— i
o
o
oo
CM
CO
iO
o
CO CM
CM
CO
CM
rH
o
1— 1
rH
C
i-h CM
CM
CM
CO
CO
CO
O
iO
of
os
iO OS
O
rH
CO
<M
CO
OS
PQ
CM CM
00
CO
O
rH
t^
1— 1
t^
—
CM 00
d
d
d
d
d
d
id
^
K
i—i
CO
rH
00
CO OS
d d
00
rH 00
i-H O
o o
o o
o o
d d
co
iO co
rH O
O O
d d
CM t>
CO
iO o
CO CO
rH O
OS OS
CO CO
rH 00
o o
o o
o o
d d
OS i— i
CM CO
o o
d d
co oq
d rH
oo oo
co
co
o
iO
CM
o
o
d
CO
OS
CO
CO
d
00
iO
CO
CO
rH
iO
00
CO
PhPh W
tH
CO O
OS . O
iO , +T CO
XPh -
OS
£oo^£oo^u£ooo
•-9
1-5
CO
o
CM
+3
CD
a
o
Ph
o
id*
o
o
s-
o
-13
Ph
HH —
o3Ph
.5 CM
0 £
& 08
T3
CD
HH
^^
M TJ
C CD
CD hh
§ a
&*
2 "°
r3 "S
03
iO
d
T3
CD
03
.a -a
W PQ <?
o
bC
CD
"oT
c
CD
X
o
M
&
o
c
274
CARNEGIE INSTITUTION
measured with a precision of ±0.00014
or better at 95% confidence level. The
results are summarized in Table 20 and
illustrated in Figs. 68 and 69.
The whole-rock trace element concen-
trations (PHN 1611) are not accounted
for by the clinopyroxene data, indicating
that greater quantities of trace elements
should be along the grain boundaries
and/or in the secondary minerals, al-
though the rock appears fresh as an in-
clusion in the kimberlite. Potassium Ka
radiation image of this rock showed that
potassium is concentrated in the grain
boundaries. The intergranular phase in
which the potassium is concentrated can-
not be resolved under the microprobe but
is intergrown with thin veinlets of ser-
pentine.
The results for diopsides show that K
ranges from 110 to 430 ppm, and Rb
ranges from 0.015 to 0.69 ppm, being var-
iable by more than a factor of 40. The
Sr contents of the clinopyroxenes show
a bimodal distribution, one concentration
at about 90 ppm (the sheared nodules)
and the other at about 500 ppm (the
granular nodules). K/Rb ratios in the
clinopyroxenes from the sheared nodules
and a discrete diopside crystal are un-
usually high, especially the value ob-
served for PHN 1596 clinopyroxene
(22301), which is higher than any other
values ever observed. There seems to be
no coherence between K and Rb in the
clinopyroxenes, which is quite anomalous.
However, the coherence between Rb and
Cs is relatively good judging from the
narrow-range variation of Rb/Cs ratios
(59-139). As shown in Fig. 69, the
clinopyroxenes from the sheared nodules
appear to be distinguished from those in
the granular nodules primarily by the
difference in their Sr contents. The Sr
contents of the clinopyroxenes from the
sheared nodules are very similar to those
observed in clinopyroxenes from the
Hawaiian nodules.
1000
K
ppm
100
T 1 » I I I I |
10000
1596
1611
1600E4
/
HAWAII
' ' i i '
0.1
t — r
TTTT
T 1 I I I
1000
oo
JJG352
>1569
i i il
1.0
Rb ppm
Fig. 68. K vs. Rb plot for clinopyroxenes from garnet lherzolite inclusions in kimberlite. Open
circles, data found in the literature (Griffin and Murthy, 1969; Allsop et al, 1969; Philpotts et
al, 1972). Solid circles, present study. Dots, clinopyroxenes from Hawaiian ultramafic inclusions
(Shimizu, this Report).
DEPARTMENT OF TERRESTRIAL MAGNETISM
275
i i i i i i
I " 1
T
i i i ! t j r j ■
1
-
_
Rb
ppm
o
o
o
o
-
1.0
0
o
/
/
/
/
/
/ #1569
*JJG352
-
.1
/
/
/
.00 3
C
•
A •
1596
•
-
.01
l i i i i l
i . i
• i i i i i i l
10
100
Sr , ppm
1000
Fig. 69. Rb vs. Sr plot, Symbols are same as those in Fig. 68.
Comparing the two groups, the clino-
pyroxenes from the sheared nodules con-
tain higher K but lower Rb, Cs, Sr, and
Ba concentrations than those from the
granular nodules. These concentrations
are not consistent with the hypothesis
that the granular nodules are residua
produced by partial melting and extrac-
tion of liquid from more primitive rocks
like the sheared nodules because, if this
were true, all these trace elements would
be depleted in the granular nodules be-
cause they are strongly partitioned into
silicate melt relative to clinopyroxene.
The high Sr contents of the granular
nodules cannot be explained by partition
of Sr between clinopyroxene and silicate
melt because the partition coefficient of
about 0.1 for wt Sr in cpx/wt Sr in liquid
(Shimizu and Akimoto, 1971) 128 requires
that the clinopyroxene should have been
equilibrated with a silicate melt con-
taining more than 5000 ppm Sr, and that
is unrealistic.
The peculiar trace element character-
istics of these clinopyroxenes cannot be
primary features; most represent altera-
tion by some unknown secondary proc-
esses operating in the mantle. The low
Sr87/Sr86 ratios of these clinopyroxenes
(Table 20) may rule out the possibility
of crustal contamination. An indication
of a secondary process may be the pres-
ence of phlogopite in some of the granu-
lar nodules. The phlogopite is sometimes
present in two generations with different
276
CARNEGIE INSTITUTION
major element compositions. These
phlogopites are unusually high in Ti02
(about 2-4 wt %) , and an enrichment of
Ti02 in clinopyroxene has been observed
in the immediate vicinity of second-
generation phlogopite which is especially
rich in Ti02. These facts suggest that the
trace element variations may also be due
to partial equilibration with phlogopite
or with the fluid phase from which the
phlogopite was formed.
The high Sr content of clinopyroxenes
from granular nodules may be another
indication of the effects of secondary
processes. Bell and Powell (1969) 129
have measured concentrations of 5000-
11,000 ppm of Sr in carbonatites and
hence it appears possible that the high
content of Sr in the diopsides from kim-
berlite nodules is due to secondary reac-
tions between the nodules and carbonate-
rich kimberlite magma.
The unusually high K/Rb and K/Cs
ratios in the clinopyroxenes from the
sheared nodules remain unexplained. If
these high ratios also resulted from sec-
ondary reaction with some fluid or
magma, such a magma or fluid should
have anomalously high K/Rb and K/Cs
but normal Rb/Cs, or some phase should
be present by which K, Rb, and Cs abun-
dances in clinopyroxene are controlled.
From the results presented it is obvious
that until a plausible mechanism is found
to explain their characteristics and varia-
tions, caution must be exercised in apply-
ing trace element data from kimberlite
inclusions to any geochemical model of
the upper mantle.
Metasomatic Theory; Some Examples of
Diffusion and Infiltration
R. C. Fletcher and A. W. Hojmann
The importance of metasomatism (i.e.,
the change in bulk composition of a
rock) in metamorphic processes has been
debated by petrologists for many decades.
Dehydration and decarbonation reac-
tions are familiar and well-established
examples; alkali metasomatism and
granitization are perhaps among the
more controversial subjects of petrology.
In recent years, new methods have been
employed to determine the extent to which
mass transport is involved in metamorphic
processes. Examples are the use of isotopic
tracers (e.g., Shieh and Taylor, 1969) 13°
and the experimental and field observa-
tions of Vidale (1969).131 The electron
microprobe makes possible the detailed
analysis of chemical gradients in meta-
morphic-metasomatic rocks; from these
gradients and a knowledge of the mineral
equilibria involved, it is in principle
possible to deduce the transport mecha-
nism responsible for these gradients.
In the last several decades, D. S.
Korzhinskii has established a physico-
chemical basis for the thermodynamics
and the modes of transport involved (for
a recent summary, see Korzhinskii,.
1970) .132 In order to derive simple trans-
port equations, he used the concepts of
diffusion and infiltration metasomatism,
where both mechanisms involve chemical
transport through a pore fluid that is in
local equilibrium with the solid phase.
In the case of infiltration metasomatism,
the fluid percolates through and reacts
continuously with the rock. The result-
ing chromatographic effects have been
reviewed by Hofmann (1972) 133 (see also
Korzhinskii, 1970,132 1973 134 and Hof-
mann, 1973 135). In diffusion metasoma-
tism, the mobile components diffuse
through a standing pore fluid. Korzhin-
skii (1970) has given some general trans-
port equations for this case.132 Our pur-
pose here is to show solutions for a few
specific cases of pure diffusion and for
one case of combined diffusion and in-
filtration. These solutions were derived
by analytical and numerical methods
given by Carslaw and Jaeger (1959) 136
and by Crank (1956). 137 These examples
illustrate important qualitative features,
namely, the effects of different liquid-
solid equilibria (referred to as "iso-
therms") on the concentration profiles in
metasomatic replacement fronts. The
isotherm shapes are qualitatively the
DEPARTMENT OF TERRESTRIAL MAGNETISM
277
Fig. 70. Diffusion profiles for three continuous isotherms A, B, and C. The fluid concentrations
Ca and Cc are shown as dashed lines; solid concentrations Sa and Sc are shown as solid lines. A
single solid line represents the concentration profile for Sb and Cb. The distance units on the z
axis are arbitrary. The isotherms C — f(S) are shown in the inset.
same as those discussed by Hofmann
(1972) ,133
The simple graphical representation of
an isotherm (as used in Figs. 70 through
73) implies an isothermal, quasi-isobaric
equilibrium relationship between a solid
and a fluid where only one component is
independently variable. This includes
idealized cases of two-component ion ex-
change where the total ion concentration
in the fluid cannot be fixed. Because the
variability in the total ion concentration
is generally small, our examples are qual-
itatively, and probably semiquantita-
tively, applicable to two-component ion
exchange, which is geologically more in-
teresting than the one-component system.
The general mass balance for the sys-
tem containing porous rock and fluid is
(If), +(£),=» •"
if the concentration Ci<r of the ith com-
ponent in the rock (solid plus fluid) and
the flux Ji are continuous and differenti-
ate functions of time t and distance z
30 40
Z
Fig. 71. Diffusion profiles for a continuous, S-shaped isotherm. See legend for Fig. 70.
278
CARNEGIE INSTITUTION
CO
o
o
<
O
o
u
Fig. 72. Diffusion profiles for a discontinuous isotherm. The solubility gap is represented by a
dotted line. See legend for Fig. 70.
and if the transport is one dimensional
in the z direction.
When the flux Ji is specified by the
diffusion and infiltration processes, the
mass balance becomes
= —fiUi
(2)
On the right-hand side of equation 2, the
first term accounts for infiltration, the
second for diffusion; Di is the diffusion
coefficient of the ith component in the
fluid ; Cij is its concentration in the fluid ;
Ui is its velocity due to infiltration; and
(3 is the porosity of the rock. The formu-
lation of equation 2 requires that all
transport occur in the fluid phase, that is,
longitudinal diffusion in the solid is
neglected.
Fig. 73. Combined infiltration and diffusion profiles (labeled a) and pure diffusion profiles
(labeled b) for an isotherm with a "pseudosolubility gap." The a profiles are distinguished by
nearly horizontal segments (concentration plateaus) at the origin (C = 1, S = 1) and at the con-
centrations C = 0.5 and S = 0.15. For further explanations, see text.
DEPARTMENT OF TERRESTRIAL MAGNETISM
279
We now impose the condition of local
equilibrium, represented by the simple
isotherm C«,r = / (C*,/) , and obtain
put
V dt ) z " dc^/dd,
fiDt (d2Cu\
,t V ^2 )
( dCu \
(3)
Cl^i,r/ d^i,)
Equation 3 illustrates the dependence of
both infiltration and diffusion on the
slope of the isotherm dCi>r/dCij. If the
isotherm contains a solubility gap, there
is a concentration discontinuity and
equation 3 is not valid at this boundary.
This means that the solid acts as an in-
stantaneous sink or source of the com-
ponent in the fluid.
The examples discussed below and il-
lustrated in Figs. 70 through 73 are com-
puted for diffusion (and infiltration) into
a semi-infinite medium in which the con-
centrations are initially equal to zero. A
constant finite concentration is main-
tained at the origin z = 0. For conveni-
ence, we have assumed that the porosity
(3 is sufficiently small so that the concen-
tration of i in the total rock, Cir, is ap-
proximately equal to the concentration in
the solid phase alone, Cis. For the pur-
pose of representing several isotherms
and diffusion profiles in one diagram, we
have relabeled the fluid concentration C
and the solid concentration S.
An important feature of all solutions
for diffusion with these boundary condi-
tions is that the distance from the origin
of any given concentration is propor-
tional to the square root of time. This
means that the shapes of the profiles
shown in Figs. 70 through 72 (which have
arbitrary distance units) are independent
of time. An equivalent number of time
steps has been used for these profiles in
order to permit comparison of the rela-
tive rates of movement.
Diffusion profiles for curved isotherms
are most easily obtained by finite-
difference methods. We have used a
modified Schmidt method (see Crank,
1956, chapter X) ,137 Figure 70 shows the
concentration profiles for three continu-
ous isotherms, A, B, and C. The straight-
line isotherm (line B in Fig. 70) yields a
profile, CB and SB, that can be tested by
an analytic solution (in the form of an
error function) of the diffusion equation.
When the isotherm is convex upward
(line A), the fluid concentration CA
moves faster than CB but the solid con-
centration SA is retarded compared to SB
over most of the concentration range.
The most interesting profile is obtained
for the concave-upward isotherm C. The
fluid concentrations Cc form a nearly
linear profile with appreciable curvature
only in the lower ten percent of the con-
centration range. The solid concentra-
tion Sc forms an S-shaped profile. This
shape is quite distinctive and may be
used to distinguish between different dif-
fusion mechanisms. If, for example, the
operating mechanism were simple diffu-
sion through the solid phase with a con-
stant diffusion coefficient and no trans-
port through a pore fluid, there could be
no reversal in curvature, at least not for
the kind of boundary conditions we have
used.
The isotherm shapes shown in Fig. 70
are typical of solid solutions such as the
Na-Ca exchange equilibrium of plagio-
clases with chloride solutions (Orville,
1972). 138 The shapes shown in Figs. 71
and 72 resemble the alkali exchange
equilibria for alkali feldspars at different
temperatures as determined by Orville
(1963) ,139 The fluid-concentration pro-
files for the latter two isotherms are simi-
lar, but the solid-concentration profiles
differ in that there is a discontinuity at
the solubility gap.
The diffusion profiles for a solubility
gap (Fig. 72) were computed by Neu-
mann's method (see Carslaw and Jaeger,
1959, chapter XI) ,136 In this method, the
concentration profiles on either side of
the concentration discontinuity are ob-
tained by two separate analytic solutions,
and the velocity of the moving boundary
(concentration discontinuity) is obtained
numerically. The solubility gap acts as a
280
CARNEGIE INSTITUTION
sink, and the concentration profile be-
tween the origin and this sink closely
resembles that of a steady-state gradient,
i.e., in this case a straight line.
For the purpose of showing the effect
of combined diffusion and infiltration, we
have used an isotherm (Fig. 73) which
exhibits a "pseudo-solubility gap." Anal-
ysis of a true concentration discontinuity
is not possible with our present finite-
difference procedure. Comparison of the
diffusion profiles in Figs. 72 and 73 shows
a close similarity : The concentration dis-
continuity of Fig. 72 is replaced by a
steep concentration gradient which rep-
resents the "pseudo-solubility gap" in
Fig. 73.
When infiltration is superimposed on
diffusion, both solid and fluid concentra-
tion profiles become distinctly different.
The pronounced concentration plateaus
are characteristic of the infiltration proc-
ess. The sharp replacement front that
would be produced by infiltration alone
(Hofmann, 1972) 133 is here damped by
diffusion.
These examples show that the shape of
a metasomatic concentration profile can
be diagnostic of both the shape of the
isotherm and the transport mechanism.
In general, it will be necessary to know
the approximate shape of the isotherm in
order to determine the metasomatic
mechanism from a given concentration
profile.
Ion Microprobe Analysis of an Isotonic
Diffusion Experiment on Biotite
A. W. Hofmann, B. J. Giletti, J. R. Hinthorne,
D. Comaford, and C. A. Andersen
The ion microprobe offers several po-
tential advantages in the analysis of dif-
fusion experiments on minerals. Among
these are in situ spot analysis of isotopic
composition to obtain diffusion profiles
on individual mineral grains and the
ability to study isotopic variations in
depth (with a resolution on the order of
tens of angstroms) as successive layers of
the mineral are eroded by ion sputtering.
The resulting high resolution makes the
technique particularly suitable for the
analysis of the very slow diffusion rates
common in geological processes. Experi-
ence with metamorphic rocks shows that
characteristic transport distances due to
solid-state diffusion are commonly on the
order of 0.1 to 10 mm. As this distance
is usually (depending on boundary con-
ditions) proportional to the square root
of time, it is reduced to the order of 0.01
to 1 /xm during an experiment lasting 0.1
year if the natural process required 10
million years. The ability to analyze
diffusion gradients with a resolution of
10 to 100 A would therefore enable us to
perform the diffusion experiments at geo-
logically realistic temperatures with no
need for extrapolation.
In order to test the general feasibility
of this approach we analyzed a biotite
flake that had previously been subj ected to
a tracer diffusion experiment. The biotite
was sealed in a gold tube with an aqueous
2.0 normal alkali-chloride solution that
contained sodium, potassium enriched
with K41 tracer, and rubidium. The
K39/K41 ratio of the solution was 0.95
(which contrasts with a natural potas-
sium value of about 13.5), the Na/K
mole ratio was 5.73, and the Rb/K ratio
was 480. The charge was subjected to a
temperature of 650 °C and a pressure of
2.0 kb for 41 days. The diffusion experi-
ment is described in greater detail by
Hofmann and Giletti (1970) .14°
The isotopic analyses reported here
were made on an ion microprobe mass
analyzer (Andersen and Hinthorne,
1972) 141 at Applied Research Labora-
tories at Sunland, California. Figure 74
shows a sketch of the biotite flake and
the spots analyzed. Each spot was bom-
barded by a beam of 0.4 nanoamps of
negatively charged oxygen-16 ions which
had been accelerated to 20 kV in the pri-
mary mass spectrometer. The K39/K41
ratios were recorded at successively
deeper levels as continued ion sputtering
removed material from the surface. The
DEPARTMENT OF TERRESTRIAL MAGNETISM
281
Fig. 74. Schematic sketch of biotite flake. The small rectangles show the sites of the spot analy
ses. The large rectangle represents the rastered area.
results of the spot measurements are
shown in Fig. 75. The contours were
drawn by hand to show the overall con-
sistency of the data. The K39/K41 ratios
increase from the edge toward the center
of the flake. From these contours a rough
estimate can be made of the relative
transport rates in the different crystallo-
graphic directions. Because each spot
has a diameter of about 6 fim, much of
the steep lateral concentration gradient
near the edge was not resolved. If we
take the measured isotopic ratio to rep-
resent the value at the center of a spot, a
ratio K39/K41 = 11.5 is found at a depth
of >2000 A when measured 5 pm from
DISTANCE FROM EDGE (/xm)
10 20 30 40
50
% 600 \-
60
3500
SPOT NUMBER
Fig. 75. Analytical results of the spot analyses. Each black dot represents the site of an analy-
sis, and the associated number is the K39/K41 ratio measured at that site. The approximate depth
is measured from the top surface of the flake and is estimated from the elapsed sputtering time.
The black bars show the diameter of each spot. The contours were drawn by hand. They repre-
sent the ratios K39/41 = 9, 11, 12, 13, and 13.5.
282
CARNEGIE INSTITUTION
the left edge of the flake (spot 1). This
isotope ratio is reached at a depth of 400
A (measured in the c direction) in the
center of the flake (spot 4). Thus, the
transport distance appears to be greater
in the a,b direction than in the c direc-
tion by a factor of 125. The same factor
calculated using spot 5 instead of spot 1
is only about 40. Because of the steep
gradient near the edge, a small error in
the estimate of the location of the spot
or a small amount of beam wobble will
cause considerable errors in the measured
isotopic ratio of this spot. At present, we
can only conclude that the transport rate
of potassium parallel to the mica cleav-
age is considerably faster than in the
direction parallel to the c axis. In addi-
tion to the spot analyses, we rastered an
area 25 X 30 /xm (see Fig. 74) by rapid
scanning of the primary ion beam. This
procedure erodes the analyzed region
slowly and yields more precise, though
horizontally averaged, analyses at suc-
cessively deeper levels from the surface
of the grain. The dashed line in Fig. 76
gives the combined results of the rastered
analyses and the analyses from spot 4
(open circles).
Because of the preliminary nature of
the experiment, a number of unverified
assumptions are made in order to derive
a diffusion coefficient from these results:
(1) The depth has been estimated cor-
rectly from the sputtering rate. (2) The
first 80 A may be ignored because rem-
nants of the carbon film deposited on the
surface of the grain were sputtered to-
gether with the first few unit cells of
mica. (3) The beam drills a flat-bot-
tomed hole and does not sputter from the
hole walls. (4) The degree of vertical
isotopic mixing in the crystal due to ion
bombardment is negligible. (5) The
measured peak ratio of masses 39 and 41
equals the K39/K41 ratio (a check on the
X/2N/DT
,0
0.4
0.8
1.2
1.6
2.0
2.4
2.8
4 -
2 -
0
I ' I
I ■ I ' I ' I ' I ' I i I i I 'J
t~t
J i i i i L
J L_L
0 200
J i L
600
1000
DEPTH (A)
1400
1800
Fig. 76. K39/K41 ratio as a function of depth from the surface of the flake. The dashed line rep-
resents the combination of the analytical data of the rastered area for depths less than 350 A, and
the results from spot 4 for depths greater than 400 A. The solid line represents an ideal diffusion
profile (for the specific assumptions used to calculate this profile, see text).
DEPARTMENT OF TERRESTRIAL MAGNETISM
283
validity of this assumption is given by
the fact that the measured ratio K39/K41
= 13.5 in the interior of the grain is
within the range of currently accepted
values for the isotopic abundances of
natural potassium). (6) The total potas-
sium concentration is constant over the
range of the measured concentration pro-
file (in fact, some potassium was lost
from the biotite because of K-Na ex-
change with the hydrothermal solution).
(7) The boundary conditions can be ap-
proximated by constant concentrations of
K39 and K41 at the edge of the grain.
The last assumption may be justified
from our experience with diffusion ex-
periments on biotite (Hofmann and Gi-
letti, 1970) .14° We have found that the
isotopic composition of the fluid changes
very rapidly at first, probably because of
surface exchange and a small amount of
sample dissolution. After this initial ex-
change, the fluid composition and there-
fore the boundary conditions for the
solids change much more slowly as a re-
sult of volume diffusion. The situation is
further complicated by the apparent re-
versal in the profile at an approximate
depth of 40 to 80 A. If this reversal was
caused by a dislocation or an incipient
splitting along a cleavage plane, the
effective boundary for diffusion at greater
depths may have been at 80 A.
Because of these uncertainties we used
a constant boundary value of K39/K41 =
9.0 to obtain a theoretical diffusion profile
for a semi-infinite solid. Given these
boundary conditions and the above as-
sumptions, the concentration C of a
tracer isotope, say K41, is given by the
equation (see Crank, 1956, chapter 3)
L 0 — C i
= erf
x
2\/Dt
where Cx is the concentration at the sur-
face, C0 is the concentration in the inte-
rior at time t = 0, x is the distance from
the surface, and D is the diffusion co-
efficient. The theoretical curve can be
fitted to the measured values by adjust-
ing the free parameter D. This procedure
yields a diffusion coefficient for transport
parallel to the c axis, D = 2.4 X 10"18
cm2 sec"1.
Clearly, these results are not definitive
measures of potassium diffusion in bio-
tite. They do show that extremely slow
diffusion can be measured with the sput-
tering technique and that diffusion in
anisotropic crystals can be evaluated by
three-dimensional mapping of isotopic
abundances in a single grain.
Direct Determination of Rock Rheology
from Wavelength Selection in Folding
Raymond C. Fletcher
An assumption about the rheology of
rock in slow tectonic flow (strain rate
— 10~14/sec) in the crust and mantle is a
necessary part of any tectonic model.
Data from steady-state creep tests at
strain rates >10"9/sec can be adequately
described (Goetze and Brace, 1972) 142
by a power law (strain rate proportional
to the deviatoric stress raised to a power
n) . These results show significant non-
linearity, n =3. The question remains as
to whether this nonlinearity persists
through a reduction in strain rate of five
orders of magnitude or whether rocks at
tectonic strain rates may be adequately
treated as linearly viscous fluids (n = 1) .
Most modelers of tectonic deformation
assume a linear viscous rheology.
In principle, a direct determination of
rock rheology at tectonic strain rates is
possible by interpreting rheology-sensi-
tive features of natural structure on the
basis of theoretical models which employ
a range of rheologies. An analysis of
buckling of a layer of power-law fluid
embedded in a power-law medium has
been carried out and used to fit data on
wavelength selection in a natural single-
layer fold.
An appropriate tensor generalization
of a single-component power law (Nye,
1953; 143 Calladine and Drucker,
284
CARNEGIE INSTITUTION
1962 144), written for plane flow in the
x, z plane, is
&xx — "ft €XX Pi
<*zz = 2ft ezz — p,
&xz — "ft *-xz
(1)
where p is the pressure, 2/x = \BJ2in 2)]_1,
Jz — (Vi) (vxx — <?zz)2 + vxz2, and B is a
constant. If x and z are the principal
axes of a stress vif, a small stress incre-
ment atj and the associated strain-rate
increment e# will approximately satisfy
the linearized constitutive relations
S%x = 2fi° (l/n)££, — p7
<?ZZ
2/x° (1/n)
P,
&XZ
— 2a°
ft exz
(2),
where, from (1), o-^/ — azz° = 4/x° c^0.
These relations are equivalent to those
for an anisotropic viscous fluid whose
viscosity in shortening or extension par-
allel to x and z is less than its viscosity
in shear by the factor (1/n).
In the initial stages of buckling of a
single embedded layer (Sherwin and
Chappie, 1968), 145 wavelength compo-
nents in the initial irregularity of the
layer surfaces are selectively amplified,
and the form of the layer tends, generally
with much imperfection, towards a regu-
lar train of folds with a preferred arc
wavelength to thickness ratio, Lp/h.
During this phase the layer accommo-
dates itself to bulk shortening by thick-
ening approximately uniformly. When
the limb dip of the folds reaches 10-15°,
wavelength selection ceases, a sharp
transition to buckle-shortening of the
layer takes place, and thereafter the arc
wavelengths of existing folds and the
layer thickness are preserved.
The analysis of this process shows that
the power-law exponent of the layer nL
and the parameter Q = (nL/nM)1/2
(ftu°/ftL°) , where the subscripts L and M
refer to quantities of the layer and em-
bedding medium, respectively, may be
determined from (1) Lp/h; (2) A/A0, the
amplification of this component at the
time wavelength selection ceases; and
(3) the amount of layer thickening, rep-
resented, for example, by Ts = (h/h0)2,
where h and h0 are the final and initial
thicknesses. Lp/h is readily observed,
and from consideration of the probable
initial amplitude of irregularities, A'/A0
can be estimated. Ts may be estimated in
the fortunate circumstance that some in-
dicator in the layer records its finite
strain or is bounded by the amount of
shortening of the embedding medium.
Only a single case is presently known for
which Ts can be reliably estimated (Gro-
shong, 1971 ; 146 Fletcher and Groshong,
1973 147).
For a limestone layer embedded in
shale in the unmetamorphosed Rochester-
McKenzie Formation (Silurian) of the
Valley and Ridge Province, central Penn-
sylvania, Lp/h = 9.4, A/A0 is estimated
to lie in the range 20-100, and 1.07 < Ts
< 1.27. The rheological interpretation of
these data is illustrated in Fig. 77. The
upper bound on the layer thickening, Ts
= 1.27, which is determined from the de-
formation of brachiopods in the shale far
from the layer, implies that the limestone
exhibits significant rheological nonline-
arity, nL > 3. This might be expected for
a rock deformed at tectonic strain rates
and low temperature. The value Ts =
1.07 is derived from a calcite-twin strain-
CONTOUR5 OF
AMPLIFICATION
Ld/h = 9.4
Fig. 77. Evaluation of rheological parameters
from data for a natural fold.
DEPARTMENT OF TERRESTRIAL MAGNETISM
285
gauge procedure (Groshong, 1971) 146 and
probably underestimates the amount of
thickening. It implies a value nL « 12.
Corresponding values of Q range from
about 1/40 for nL — 3 to 1/60 for nL =
12. The results are not sensitive to the
amount of amplification assumed.
This is the only example of a direct
determination of rock rheology at tec-
tonic strain rates known to the author.
A variety of finite strain gauges may be
found in deformed rocks: deformed peb-
bles, fossils, oolites, and twinned crystals,
for example. The theoretical results are
therefore applicable to a wide variety of
rocks, provided the appropriate detailed
observations are made.
U-Pb Isotonic Studies of Zircons from
the Baltimore Gneiss of the Towson
Dome, Maryland
B. Grauert
The Towson dome is the easternmost
protrusion at the end of a chain of man-
tled gneiss domes along the crest of the
Washington-Baltimore anticlinorium in
the Maryland Piedmont (see Hopson,
1964, Fig. 12) ,148 The complex formation
of crystalline rocks exposed in the domes
represents a Precambrian basement and
is generally referred to as Baltimore
Gneiss. The mantle of the domes is
formed by the younger metasediments
and metavolcanics of the Glenarm Series
(see Higgins, 1972). 149 The lowest unit
of the Glenarm Series, the Setters Forma-
tion, consists chiefly of quartz-rich
schists and mica quartzites (see Knopf
and Jonas, 1929; 15° Hopson, 1964 148).
Both the geologic evidence and previous
isotopic age determinations (Tilton et al.,
1958,151 1970 ; 152 Wetherill et al, 1966,153
1968 154) indicate that the Baltimore
Gneiss had a polymetamorphic history:
(1) A strong orogeny at about 1000-1100
m.y. transformed earlier rocks to gneisses
and migmatites (Hopson, 1964). (2)
During an early Paleozoic orogeny, prior
to 425 ± 20 m.y., the basement was re-
activated and rose as domes beneath its
mantle. The Gunpowder granite, a gra-
nitic "intrusion" into the gneisses of the
Towson dome and its southern mantle
(see also p. 288, this Report), is inter-
preted as rheomorphic Baltimore Gneiss
(Hopson, 1964, p. 45; 148 Steiger and
Hopson, 1964 155). (3) Temperatures re-
mained sufficiently high for argon diffu-
sive loss and strontium isotopic reequili-
bration in minerals until about 290-280
m.y. The northeastern part of the Tow-
son dome, where the zircon samples were
collected, consists of coarse augen gneiss
grading into or intercalated with biotite
and hornblende gneisses. The augen gneiss,
which locally contains large crystals of
microcline, is also known as Hartley
augen gneiss (Knopf and Jonas, 1929) .150
Knopf and Jonas thought the porphyritic
gneiss ". . . represents an igneous magma
that has in other places formed a banded
ribbon gneiss by its injection into the
Baltimore Gneiss sediments" and was
later crushed and recrystallized into
augen gneiss. Hopson (1964, p. 43), how-
ever, believes ". . . the augen gneiss is a
local porphyroblastic phase of the Balti-
more Gneiss, formed chiefly by potassium
metasomatism along movement zones be-
neath the Setters quartzite, in part dur-
ing post-Glenarm doming and meta-
morphism."
To test these interpretations, I have
separated and analyzed zircons from
three gneiss samples, all collected within
30 feet of a fresh, blasted outcrop at
Gunpowder Falls (Towson quadrangle,
Maryland, 369420/4364470, 1000-meter
grid). At the sample locality the pre-
dominant rock type is a coarse augen to
flaser gneiss (sample BAL-11) with few
migmatized intercalations of biotite
gneiss (BAL-12). Veins of a medium-
grained granite (BAL-13), showing more
or less sharp contacts, have been de-
formed along with the augen gneiss. All
rock samples are relatively rich in zir-
cons (about 2-5 g zircon in 10 kg of
gneiss).
286
CARNEGIE INSTITUTION
A microscopic study reveals great simi-
larities in the morphological and optical
properties of the zircons, reflecting three
stages of zircon growth: (1) Translucent
cores, sometimes clear but more often
dark, can be observed in many of the
larger crystals. They most likely repre-
sent the oldest presumably inherited zir-
con component. (2) The second stage is
characterized by euhedral zircon growth,
evident from well-displayed repeated
zoning (up to ten zones) in the majority
of the crystals. The zoned part of the
zircons, which generally makes up more
than 80% of the volume, is clear in the
uranium-poor zircons and brownish
translucent in those rich in uranium.
From the microscopic picture it appears
that the zircons went through a period
during which the development of per-
fectly euhedral and repeatedly zoned
crystals was possible, thus suggesting a
magmatic or migmatic environment of
crystallization. (3) During a third stage,
clear uranium-poor overgrowths modi-
fied the shape of many crystals, forming
all sorts of grains from nearly euhedral
individuals to very irregularly shaped
clods. The assumption that the clear
overgrowths are poor in uranium is based
on analyses of 17 zircon fractions from
various types of Baltimore Gneiss which
yielded the following results: clear frac-
tions, 400—600 ppm U; brownish to pink-
ish translucent fractions, 500-2500 ppm
U; metamict fractions, more than 2500
ppm U. The isotopic data of six zircon
fractions are plotted on the concordia
diagram in Fig. 78. The points fall
within analytical uncertainty on a chord
which intersects the concordia curve at
about 450 m.y. and at 1180 m.y. Like the
zircons from the Baltimore Gneiss of the
Phoenix dome (see Year Book 71, pp.
301—305) the pattern on the diagram does
not fit any simple continuous diffusion
model (Tilton, I960,156 Wasserburg,
1963 157) although the degree of discord-
pb206/u238
O.I5
0.10
i 1 1 r
ZIRCONS FROM THE
TOW SON DOME
IOOO
BAL-II 663ppm
BAL-II 660 ppm -
BAL-II 789 ppm U
BAL- 12 1393 ppm U
BAL-13 1672 ppm U
BAL- 12 2032 ppm U
400
1.0
I I I I L
2.0
Pb207/U235
Fig. 78. Concordia diagram showing the isotopic data of zircons from the Baltimore Gneiss of
the Towson dome.
DEPARTMENT OF TERRESTRIAL MAGNETISM
287
ance could be interpreted as a function
of the uranium concentration (Silver,
1963) 158 or radiation damage.
A straightforward explanation of the
linear pattern in Fig. 78 would be the
simple episodic loss of lead from 1180-
m.y.-old zircons. Considering, however,
the morphological properties mentioned
above, the question arises whether the
zircons represent mixtures of old (zoned)
cores with 450-m.y.-old overgrowths, or
whether they are entirely old, having
been subject to radiogenic lead loss or
uranium gain during an Early Paleozoic
metamorphism. To test these possibil-
ities, the uranium concentrations have
been plotted versus the apparent Pb206/
TJ238 ages m ;pjg 79 (i) if we assume
a 1180-m.y.-old zircon population with
about 500 ppm uranium as the starting
material, a model of episodic uranium
gain at 450 m.y. without new zircon
growth involved (curve A) would ap-
proximately fit the measured pattern;
however, it is indistinguishable from a
model of episodic lead loss from zircons
with primary differences in uranium con-
centration. (2) A model of uranium-rich
overgrowth, e.g., with zircon substance
containing 3000 ppm uranium, is shown
by curve B. This model does not fit all
the points and requires additional pri-
APPARENT Pb^U238 AGE, M.Y.
1200
1100
1000-
900
800
700
600
500
400
1000
2000
3000
ppm U
Fig. 79. Diagram to explain various mixing models of 1180-m.y.-old zircons with zircon over-
growths or uranium gain at 450 m.y.
288
CARNEGIE INSTITUTION
mary differences with respect to the
uranium concentration. However, since
the overgrown part is low in uranium,
this possibility can be excluded. (3) In
accordance with the relatively uranium-
poor overgrowth observed, mixing curves
have been drawn for sample BAL-12
(2023 ppm U) assuming 500 ppm and
1000 ppm uranium in the overgrown part
(curves C and D, respectively). Ade-
quate curves for the other zircon frac-
tions analyzed would be similar but
steeper. This model requires large differ-
ences in the uranium concentrations of
the older zircon components. In addition,
the concentrations in the cores of the
uranium-rich crystals must be extremely
high — more than 8000 ppm. Such values,
however, occur only in completely met-
amict zircons.
From these considerations the follow-
ing conclusions are drawn: (1) The iso-
topic pattern shown in Fig. 79 is pri-
marily the result of episodic lead loss
about 450 m.y. ago. Minor episodic ura-
nium gain may be possible. (2) The
clear overgrowth may be old or young.
If it is young it would have a minor mix-
ing effect in addition to the episodic dis-
turbance of the older U-Pb systems.
U-Pb Isotopic Studies of Zircons from the
Gunpowder Granite, Baltimore County,
Maryland
B. Grauert
Along Gunpowder Falls in Baltimore
County, the Baltimore Gneiss of the east-
ern Towson dome and its metasedimen-
tary mantle have been intruded by a
fine- to medium-grained gneissic quartz
monzonite, the Gunpowder granite
(Knopf and Jonas, 1929) .150 According
to Hopson (1964, p. 47) 148 the Gun-
powder granite is a rheomorphic offshoot
of the Baltimore Gneiss and was em-
placed during deformation and regional
metamorphism. Where the granite breaks
through the mantle rocks the material is
relatively uniform and shows intrusive
contacts, but traced back to its roots in
the core of the dome, the granite becomes
more migmatic and the contacts against
the Baltimore Gneiss are often grada-
tional.
From the field observations and from
similarities in mineralogy and bulk com-
position with the Baltimore Gneiss, Hop-
son (1964) 148 concludes that the Gun-
powder granite was formed by anatectic
conversion of Baltimore Gneiss. This
interpretation seems to be supported by
the findings of Steiger and Hopson
(1964) ,155 who analyzed two zircon-size
fractions from the granite (their data are
included in Fig. 80) . In the coarse frac-
tion (point C) they observed many short,
rounded grains, often cloudy and pinkish,
similar to zircons found in the Baltimore
Gneiss. The fine fraction (point F) ,
however, was composed of about 70%
clear, needlelike euhedral zircons. Steiger
and Hopson therefore interpreted the
coarser zircons as inherited from the
Baltimore Gneiss, whereas the zircons of
the fine fraction should have been crys-
tallized during migmatization ^conver-
sion) of the Baltimore Gneiss. To estab-
lish a minimum age for the migmatization
and intrusion, Steiger and Hopson use the
apparent Pb207/Pb206 ages. They add,
however, that by assumption of episodic
lead loss during Paleozoic metamorphism
or of continuous diffusive lead loss, the
true age of the zircons would be higher
(up to 700 m.y.).
The major regional deformation and
metamorphism was terminated 425 ±
m.y. ago. This age is given by a Rb-Sr
isochron on minerals from various coarse
pegmatites which occur in the surround-
ings of the gneiss domes and which cross-
cut the foliation of the metasediments
(Wetherill et al, 1966). 153 It is, how-
ever, to be noted that the mineral iso-
chrons in some pegmatites have been
reset during later thermal (and tectonic?)
events: Two pegmatites yielded ages of
about 390 m.y. and three others gave
DEPARTMENT OF TERRESTRIAL MAGNETISM
289
Pb206/U238
0.08 -
0.06 -
0.05 -
0.3
-
1
/
^ 4.0
500 ^/
<■/ □ C 600 ppm U
-
400///
125 - 150 Mm 364 ppm U
75 -I25^m 484 ppm U
-
350
Aj < AOfim
810 ppm U
-
n
F - 325 mesh
(<44/*m) 900 ppm U
300 >
1
1 1
0.4
0.5
0.6
0.7
0.8
Pb207u235
Fig. 80. Concordia diagram showing the isotopic data of zircons from the Gunpowder granite.
Data for points C and F after Steiger and Hopson (1964)155 and Davis et al. (1965). 159 The grain
size of the zircon fractions and the uranium concentrations are indicated on the diagram.
ages of 345-347 m.y. (A Rb87 = 1.39 X
lo-^r1).
As the age of the Gunpowder granite
and the problem of its origin are of par-
ticular interest, I have collected rock
samples from two different localities
along Gunpowder Falls in order to carry
out isotopic analyses of small, well-
defined zircon fractions. Because episodic
lead loss appears to be the most likely
interpretation for the discordant zircons
of the Baltimore Gneiss (see this Report,
p. 288), detailed analyses were initiated
to find out whether this interpretation is
also applicable to the zircons of the
Gunpowder granite.
So far, I have separated zircons from
two blocks of slightly foliated, homo-
geneous-looking granite (GUN-1, GUN-
10) . Compared with the zircon yield from
various Baltimore Gneiss samples (0.2-0.5
g zircon/kg of rock) , the crop from the
Gunpowder granite was about a factor of
10 smaller (ca. 0.03 g/kg of rock). The
zircons are clear, mostly euhedral, and
have high elongations. They are appar-
ently identical to those described by
Steiger and Hopson (1964).155 How-
ever, zircons that may be interpreted as
inherited from the Baltimore Gneiss were
found only in sample GUN-10 and
amount to only about 3% of the total
zircon crop. The two types of zircons are
clearly distinguishable, and no zircons
have been found that could be regarded
as partly resorbed or as inherited from
the Baltimore Gneiss with later euhedral
overgrowth in the Gunpowder granite.
The gross differences in zircon yield be-
tween the Baltimore Gneiss and the Gun-
powder granite and the occurrence of
only few and distinctly different zircons
of Baltimore Gneiss type do not support
Hopson's (1964) 148 conclusion that the
Gunpowder granite was formed in situ by
anatectic conversion of Baltimore Gneiss,
290
CARNEGIE INSTITUTION
at least not at the present-day exposed
level. I would rather interpret the rock
as a homogeneous, although palingenic,
intrusive which locally migmatized and
assimilated parts of the country rock.
So far, three sizes of fractions of zircon
sample GUN-1 have been analyzed, and
the results are plotted on the concordia
diagram in Fig. 80. The points fall well
on a straight line that intersects the con-
cordia curve at about 330 m.y. Although
the zircon fractions contain no grains
that could be clearly related to the Balti-
more Gneiss, they obviously contain a
component of older radiogenic lead. Such
observations are quite common for mag-
matic zircons of presumably palingenic
intrusives. This problem is the major
subject of another part of this year's Re-
port (see Grauert and Hofmann) .
The new data, particularly the point
of the smallest size fraction, do not agree
with the results of Steiger and Hopson
(1964) ,155 If Steiger and Hopson's
method of interpretation were applied to
the new data, the minimum age of crys-
tallization for the euhedral zircons would
be considerably lower, about 430 m.y.
instead of 500 m.y. It is, however, strik-
ing that the age of 330 m.y., given by
the lower intersection of the best-fit line
with the concordia, nearly coincides with
the ages of the Rb-Sr mineral isochrons
(biotite, muscovite, and feldspars) of
three pegmatites determined by Wetherill
et al. (1966) ,153 It is therefore possible
that a distinct thermal (and tectonic?)
event took place 320-350 m.y. ago. In
that case, the data pattern in Fig. 80
would be the result of mixing of old and
newly formed zircon components com-
bined with later episodic disturbances of
the U-Pb systems at about 300 m.y. This
model, however, gives no satisfactory ex-
planation for the highly discordant ages
of Steiger and Hopson's point F, which
suggests instead a continuous diffusive
lead loss or episodic disturbance some
time after 300 m.y.
U-Pb Isotopic Analyses of Zircons from
Granulite and Amphibolite Fades Rocks
of the West Chester Prong and the
Avondale Anticline, Southeastern
Pennsylvania
B. Grauert, M. L. Crawford* and
M.E. Wagner]
The Precambrian basement of south-
eastern Pennsylvania, collectively termed
Baltimore Gneiss, outcrops along an axis
of intense Paleozoic deformation and
metamorphism. As in the Piedmont of
the Baltimore-Washington area, the
basement has been uplifted and thrust
over metamorphosed sediments of the
Glenarm Series (a detailed discussion of
the stratigraphic age of the Glenarm
Series is given in Higgins, 1972). 149 The
parts of the basement known as the West
Chester Prong and the Avondale Anti-
cline, where the zircons were collected,
are largely bounded by faults and mylo-
nite zones which separate the Baltimore
Gneiss from mica schists and serpenti-
nites in the north and from serpentinites,
schists, and gneisses, reaching the silli-
manite-orthoclase subfacies of the am-
phibolite facies in the south.
The Baltimore Gneiss in the middle
part of the West Chester Prong has been
recently studied by Wagner (1972) 160
and Wagner and Crawford (1973) .161
From the mineral parageneses observed,
they distinguish three areas of different
metamorphic facies (see Fig. 81). (1)
In the western granulite facies area,
the PT conditions were the highest and
reached approximately 10-11 kb and
900-950 °C. The main lithologic types
are pyroxene granulites (presumably the
metamorphic derivatives of olivine gab-
bros), quartzofeldspathic granulites,
and quartzose rocks. (2) In the eastern
granulite facies area, the PT conditions
were slightly lower than in the western
part and reached about 9 kb and 850-
* Department of Geology, Bryn Mawr Col-
lege, Bryn Mawr, Pa. 19010.
t Department of Geology, University of Penn-
sylvania, Philadelphia, Pa. 19104.
DEPARTMENT OF TERRESTRIAL MAGNETISM
291
175°30'
"■ /
/ ~ /
l'
\
\
/
N
\"
/ >
1
1 - 7S
LEGEND
Ultramafic rocks
Cambro-Ordovician
sedimentary rocks
Glenarm Series
Wilmington
complex
Baltimore Gneiss
undivided
Western granulite
fades
Eastern granulite
fades
Amphibolite
facies
Fig. 81. Simplified geologic sketch map showing the basement complex (=Baltimore Gneiss)
of the middle part of the West Chester Prong and the eastern Avondale Anticline, southeastern
Pennsylvania, with the sample locations. The irregular heptagon outlines the area studied by
Wagner (1972).
900 °C. Pyroxene granulites and quartzo-
feldspathic granulites are the two main
rock types. Quartzose rocks have not
been found. (3) Migmatites, amphibol-
ites, and ultramafic rocks occur in the
area of later amphibolite facies meta-
morphism, which has relatively sharp
boundaries against the western granulite
facies but is gradational into the eastern
granulite facies. Some of the amphibolite
facies rocks are probably completely
retrograded granulites; others are prob-
ably younger than the granulite facies
metamorphism.
At the time of the amphibolite facies
metamorphism, garnet coronas formed
around hypersthene and other mafic min-
erals in all the granulite facies rocks
wherever anhydrous conditions prevailed.
Where water had access, recrystallization
was more complete and biotite-garnet
gneisses formed which retain some evi-
dence of having once been in the granu-
lite facies. PT conditions during the
amphibolite facies metamorphism were
probably about 7-8 kb and 650-700°C.
According to isotopic age determina-
tions of zircons (Tilton et al., I960)162
and from comparison with radiometric
ages of the basement in the Baltimore-
Washington area, the granulite facies
metamorphism is thought to be about
1100 m.y. old. The younger amphibolite
facies metamorphism is attributed to the
Taconic and possibly Acadian orogenies.
It is well known, and discussed in an-
other part of this Report (see Grauert
and Hofmann) , that inherited zircons in
high-grade metamorphic sediments, mig-
matites, and even intrusive igneous rocks
did not reset their U-Pb radiometric
clocks completely under metamorphic
and magmatic conditions. Koppel and
Sommerauer (1973), 1G3 who studied de-
trital zircon suites of regional meta-
morphic sediments, found that ". . . zir-
cons of amphibolite facies paragneisses
show apparent lead loss of 70-90%, and
zircons from granulite facies rocks of
95-99%." This means that zircons which
crystallized or recrystallized under gran-
ulite facies conditions are apparently
292
CARNEGIE INSTITUTION
more useful for dating ages of regional
metamorphisms than zircons of amphib-
olite facies metasediments because sys-
tematic errors due to incomplete reset-
ting of the inherited U-Pb systems are
probably much smaller.
We have separated zircons of rocks
from the three facies areas mentioned
above. The sample localities are shown
in Fig. 81, and the analytical results are
plotted on a concordia diagram in Fig.
82. The points on the diagram cluster in
two groups which can be clearly corre-
lated with the granulite and younger
amphibolite facies rocks: (1) The zircons
of the granulites have high apparent U-
Pb ages. If we assume a model of episodic
lead loss, as found to be most likely for
the Baltimore Gneiss zircons of the Tow-
son dome, Maryland (see this Report),
and if we assume 450 m.y. (Taconic
orogeny) as the approximate age of the
episodic event that caused the lead loss,
then we obtain upper intersections with
the concordia curve between 980 and
1050 m.y., with the zircons from the
western granulite area giving the lowest
intersections. These ages are lower than
the zircon ages of the Baltimore Gneiss
from the Phoenix and Towson domes in
Maryland (see this and last year's an-
nual report). The Baltimore Gneiss in
Maryland comprises only amphibolite
facies rocks, and no relicts of an earlier
granulite facies are known. The ages are
also distinctly younger than the extra-
polated age of a zircon from the eastern
West Chester Prong analyzed by Tilton
et al. (I960).162 We therefore conclude
that the age of about 980 m.y. dates the
end of the Precambrian granulite facies
metamorphism which has reset the U-Pb
clocks in presumably much older in-
herited zircons (detrital zircons of sam-
ple WCP-316) as well as in zircons that
may have partly formed during earlier
stages of the same Precambrian orogeny
(sample WCP-1 and WCP-322). It
should be remarked that the Rb-Sr
whole-rock isochron of 1050 m.y. ob-
pb206/LJ238
0.15
O.IO
ZIRCONS OF BALTIMORE GNEISS
FROM SOUTHEASTERN PENNSYLVANIA
WESTERN GRANULITE FACIES
O WCP-316
D WCP - 322
EASTERN GRANULITE FACIES
O wcp -i
AMPHIBOLITE FACIES
A WCP -234
V AVA - 6
AFTER TILTON
ET AL. (I960)
600
400
pb207/Li235
Fig. 82. Concordia diagram showing the isotopic data of zircons from the Baltimore Gneiss of
southeastern Pennsylvania. The numbers at the points are the uranium concentrations in ppm.
DEPARTMENT OF TERRESTRIAL MAGNETISM
293
tained by Wetherill et al. (1968)154 on
basement gneisses of the Baltimore area
would give an age of 993 m.y. if
recalculated with the decay constant
A Rb87 = 1.47 X KT^y-1. (2) The
analytical data of the zircons from the
amphibolite facies rocks cannot be inter-
preted unequivocally. Morphologically,
the mostly subhedral zircons can be
easily distinguished from those of the
granulites, which are rounded or oval
without any crystal facies developed. It
is therefore unlikely that these rocks
were ever granulites. The low apparent
U-Pb ages — the uranium concentrations
are lower than in most other Baltimore
Gneiss zircons — suggest a strong Paleo-
zoic metamorphism or special meta-
morphic conditions to cause exceptionally
high radiogenic lead loss. An alternative
possibility would be that the rocks rep-
resent the metamorphic derivatives of
sediments or volcanic rocks deposited
after 980 m.y. but still containing an
inherited older zircon component. Rankin
et al. (1969) ,164 who analyzed zircons
from upper Precambrian felsic volcanics
of the Blue Ridge, obtained five points
defining a chord which intersects the Con-
cordia curve at 820 and 240 m.y. As the
volcanic rocks rest nonconformably upon
billion-year-old plutonic rocks, Rankin
et al. suggest an original age of 820 m.y.
and an episodic lead loss at 240 m.y. The
latter age they interpret as the time of
movement of the Blue Ridge thrust sheet.
Our points plot close to the chord of
Rankin et al. ; however, the spread of the
points is too small to decide whether the
zircons have the same original age and
the same age of episodic lead loss as those
of the Blue Ridge. Further investigations
are under way to find a better answer.
Age and Origin of Zircons from Meta-
morphic Rocks in the Manhattan Prong,
White Plains Area, Southeastern
New York
B. Grauert and L. Hall
The metamorphic rocks in the Man-
hattan Prong have been subdivided into
five formations: Yonkers Gneiss, Ford-
ham Gneiss, Lowerre Quartzite, Inwood
Marble, and Manhattan Schist (see Hall,
1968) .1G5 A diagrammatic illustration of
the interpreted stratigraphic history is
shown in Fig. 83. In the White Plains
MIDDLE
ORDOVICIAN
EARLY
ORDOVICIAN
PRECAMBRIAN
MANHATTAN -c
AAA-AAAaA 4 A ft
• MANHATTAN -b-y-
«•; — ; — ; — r— —s <*r&
- «kMAC-1
"MANHATTAN-a
Fig. 83. Diagrammatic illustration of the interpreted stratigraphic history in the Manhattan
Prong. The points with sample numbers indicate rock units from which zircons have been ana-
lyzed.
294
CARNEGIE INSTITUTION
area all these rocks underwent at least
sillimanite-grade metamorphism and
were locally intensely migmatized.
The Fordham Gneiss, which represents
the basement, has been subdivided into
five members A to E, with A as the inter-
preted oldest unit (Hall, 1966) ,166 The
Fordham Gneiss consists of various types
of paragneisses, amphibolites (members
A, B, C, and E), and minor quartzites
and rusty schists (member D). The
Fordham Gneiss is thought to be Pre-
cambrian because it is unconformably
overlain by the Lowerre Quartzite, which
correlates with Lower Cambrian Quartz-
ites farther north.
The Yonkers Gneiss, a well-foliated
biotite and/or hornblende quartz-feld-
spar gneiss, is intimately associated with
the Fordham Gneiss but has sharp bound-
aries. The Yonkers has been variously
interpreted as metamorphosed arkose
(Merill, 1890), 167 as metamorphosed in-
trusive granite (Merill, 1896), 168 and as
metamorphosed rhyolite or arkose (Hall,
1968) ,165
The Manhattan Schist, which has been
subdivided into three members (A to C) ,
consists of schists, schistose gneisses, am-
phibolites, and some calcite marbles.
Member A of the Manhattan Schist is
interpreted to be Middle Ordovician on
the basis of regional correlation (Hall,
1968) ,165 This age is based largely on
the recognition of an unconformity at the
base of member A, which is believed to
represent the widespread Middle Ordo-
vician unconformity found also elsewhere
in New York and New England (Cady,
1945; 169 Thompson in Zen, 1959; 17°
Zen, 1967 171) and is strengthened by the
discovery of fossils (Ratcliffe and
Knowles, 1968). 172 Members B and C of
the Manhattan Schist are believed to be
Cambrian rocks which were thrust onto
member A (Hall, 1968).
Yonkers and Fordham gneiss were
subject to at least one intense deforma-
tion and metamorphism before the de-
position of the Lower Paleozoic cover.
Long (1969) 173 obtained a Rb-Sr whole-
rock isochron of 575 ± 30 m.y. (A. Rb87
= 1.39 X 10_11y_1) for the Yonkers
gneiss and concluded that the Fordham
and Yonkers gneisses were involved in
major tectonism around 575 m.y. ago.
An extremely intense deformation and
metamorphism and another, apparently
less intense, took place during the
Taconic and/or Acadian orogenies. From
K-Ar and Rb-Sr age determinations of
micas, Long and Kulp (1962) 174 con-
cluded that the last metamorphic event,
"a mild pervasive re-heating," in the
Manhattan Prong took place 360 m.y.
ago. The age of the initial metamor-
phism is thought to be about 460 m.y.
(Long 1961). 175 Ar40/Ar39 incremental
heating "ages" of muscovites and biotites
from rocks in Dutchess County, New
York, which have K-Ar and Rb-Sr ages
(Long, 1962) 176 that are similar to those
determined in the Manhattan Prong
(Long, 1961 ),175 have led Bence and
Rajamani (1972) 177 to interpret the ages
of 450 m.y. and younger as cooling ages.
Bence and Rajamani (1972) 177 believe
these cooling ages reflect continuous up-
lift after the Taconic orogeny and ac-
cordingly there is no need for a distinct
Acadian thermal event.
In order to learn more about the meta-
morphic history of the bedrock in the
Manhattan Prong and to find at least
some hints as to the age of the source
area(s) that provided detrital material
for the original sediments, we have ini-
tiated U-Pb isotopic studies on zircons
from the various formations. The rock
units from which zircons have been ana-
lyzed up to now are indicated on Fig. 83.
We intend to make more analyses, and
the following discussion is concerned
with the preliminary results from the
Manhattan Prong rocks and the relations
of these results to zircon data from other
geologically comparable areas of the
Appalachians. The basement complexes
DEPARTMENT OF TERRESTRIAL MAGNETISM
295
in Maryland,* Pennsylvania,f Tennes-
seej: and North Carolina J and zircons
from the Setters Formation, the prob-
able equivalent of the Lowerre Quartzite
in Maryland, constitute the comparable
materials and areas. From the similarity
of the apparent age pattern of the zircons
from these areas, a model for their com-
mon origin and metamorphic history is
derived.
The isotopic analytical data are plotted
on a concordia diagram in Fig. 84. The
zircon data for the other areas are indi-
cated by small symbols. All zircons from
the Manhattan Prong that were studied
microscopically have in common that
they contain cores that are interpreted
*Data from Tilton et al. (1958,178 I960,162
1970 152), Grauert (1972 "9 and this Report).
tData from Tilton et al. (I960),162 Grauert
et al. (this Report).
% Data from Davis et al. (1962).180
as an older inherited zircon component.
The overgrown portions of different
grains show considerable morphologic
differences. The overgrowths produced
rounded to subhedral grains in the meta-
sediments and subhedral to euhedral
grains in the Yonkers Gneiss and in the
neosomes of migmatites found in member
C of the Fordham Gneiss. A late-stage
overgrowth that produced irregularly
shaped and spheroidal grains is present
in most zircon populations. All these
properties have also been observed among
the zircon populations of the Baltimore
Gneiss in Maryland and Pennsylvania
(see Grauert, this Report) except in the
zircons from the granulites.
The zircons of the Yonkers Gneiss
(sample YON 1) have the lowest appar-
ent ages, Pb206/U238:511 m.y., Pb207/
U235:515 m.y.; and these are consider-
ably lower than those of all other zircons
Pb206/U
238
I
CO
^
I
1700^
I
1000 y
y^*
K °
V
x^5
s% o yy
PRECAMBRIAN
(GRENVILLE)
OROGENY
M ^/^
SOURCE
REGIONS
0.15
4(
DQ>
100C
*p^
aoo
TACONIC AND ACADIAN
OROGENIES
7
00 //
^y
MANHATTAN PRONG
+ YON1
'{J* V/
V MAC1
{> yS *
A LOW 2
0.10
_ 600 /A
500 //^/
400 >r
Oj^
□ FOD 1 _
O FOC 2b
O FOC 1a
O FOB 1
• o MARYLAND
V NEW YORK
* NORTH CAROLINA
TENNESSEE
■ a PENNSYLVANIA
0.05
300/
I
i
1
0.5
1.0
1.5
2.0
pb207/u235
Fig. 84. Concordia diagram showing the isotopic data of zircons from the Manhattan Prong and
from other geologically comparable areas of the Appalachians.
296
CARNEGIE INSTITUTION
from the Manhattan Prong. If we as-
sume some disturbances of the U-Pb sys-
tems of zircons from the Yonkers during
Taconic and Acadian events, the age is
compatible with the well-defined 575 ±
30 m.y. whole-rock isochron determined
by Long (1969). 173 If the 575-m.y. age
represents an extensive metamorphism
affecting both Fordham and Yonkers, it
would seem that the U-Pb systems in the
Fordham zircons would have been more
disturbed because uranium concentra-
tions in the zircons of some Fordham
rocks (400-1300 ppm U) are similar to
and even greater than concentrations in
the Yonkers Gneiss (800 ppm U). This
suggests the possibility that the Yonkers
Gneiss represents a former granitic intru-
sion into the Fordham gneisses about 550
to 600 m.y. ago, and detailed mapping is
now being done to further test this possi-
bility.
The data points of zircons from the
Fordham Gneiss and the Manhattan
schists show a large spread on the Con-
cordia diagram, with Pb207/Pb206 ages
ranging from 800 to 1300 m.y. The fol-
lowing are two possible alternative con-
clusions that may be reached: (1) The
zircons crystallized during different
events of a long-lasting petrogenetic pe-
riod covering the time interval from 980
to about 1400 m.y. as evidenced by the
upper intersections of the extrapolated
chords from 430 m.y. to the data points,
e.g., line T on Fig. 84. (2) The zircons
are mixtures with various amounts of
much older zircon components as evi-
denced by the inherited cores. The
youngest ages of about 980 m.y., ob-
tained by extrapolation using data on
samples FOD 1 and FOC 2b, are in good
agreement with those found for the gran-
ulite zircons in Pennsylvania (see Grau-
ert et al, this Report) . They most likely
indicate the approximate end of zircon
recrystallization during the Precambrian
orogeny.
Looking at the data pattern on Fig.
84, we can make an observation that may
help us to interpret the older apparent
zircon ages. Many of the data points
giving young extrapolated ages, FOD 1,
fall relatively close to the concordia
curve, whereas those of higher ages, FOB
1 and FOC la, plot far below the extra-
polated upper intersection with the con-
cordia between 1300 and 1400 m.y.
Microscopic studies and the uranium
concentrations of the latter zircons pro-
vide no clues as to why the U-Pb systems
of these zircons should have been more
affected than other ones. Consequently,
the following hypothetical model is pro-
posed: (1) The Precambrian orogeny as
a whole happened within a relatively
short time interval (maybe 100-200
m.y.) and ended about 980 m.y. ago.
Zircons dated from neosomes in member
C of the Fordham Gneiss FOC 2b support
this. (2) Most, if not all, zircon popula-
tions analyzed so far contain a certain
amount of inherited zircons having "pri-
mary" ages of 1600 to 1700 m.y. (Fig.
84). Ages in this order of magnitude
have been obtained by extrapolation of
an episodic lead-loss line (line S in Fig.
84) to the detrital zircons of the Setters
Formation in Maryland (see Grauert,
1972). 179 Needless to say, the real spec-
trum of "primary" ages may be much
larger. (3) At the end of the Precam-
brian orogeny, the zircon fractions
plotted along mixing lines similar to line
M in Fig. 84. (4) The present-day dis-
played pattern is the result of further
disturbances during Paleozoic events.
(5) The zircons in the cover rocks (sam-
ples LOW 2 and MAC 1) derive largely
from the eroded basement but may possi-
bly contain additional components de-
rived directly from those regions that
supplied the sediments of the Fordham
and Baltimore Gneisses.
This model shows some similarities
with the interpreted zircon history of the
Central European Variscides, where the
old source regions are supposed to be
2000 to 2300 m.y. old. During an early
Paleozoic metamorphism, detrital zircons
from these source regions were recrystal-
lized and mixed with newly grown mate-
DEPARTMENT OF TERRESTRIAL MAGNETISM
297
rial. The U-Pb systems of these mixtures
were then again affected during Variscan
(=Hercynian) and Alpine events (Grau-
ert and Arnold, 1968; 181 Pidgeon et al.,
1970; 182 Grauert et al, 1973 183).
Old Radiogenic Lead Components in
Zircons from the Idaho Batholith and
its M eta sedimentary Aureole
B. Grauert and A. Hofmann
Supplementary to Rb-Sr and K-Ar iso-
topic studies on rocks of the metamorphic
lower Belt Series in northern Idaho (see
Hofmann, 1971, 184 1972; also, this Re-
port) , we have initiated U-Pb isotopic
analyses of zircons in order to provide
additional information about the meta-
morphic history and to learn more about
crystallization and recrystallization of
zircons in the metamorphic environment.
In the Beltian Series of northern
Idaho, the metamorphic grade generally
increases towards the vast mass of plu-
tonic rocks known as the Idaho batho-
lith, which is thought to have sup-
plied the thermal energy for the
regional metamorphism of the sediments
(Hietanen, 1968).lsn The age of intru-
sion of the various rocks of the batholith
is not very well known but is thought to
be Mesozoic to Tertiary. Therefore, we
also collected samples for zircon separa-
tion from several different parts of the
batholith. To our surprise the zircon
yield was extremely uneven. Two rock
samples, a quartz monzonite and a mafic
inclusion (both from the Beaver Creek
pluton), were practically free of zircon.
Figure 85 shows a simplified map of the
area with the sample locations. The
metamorphic zones in the lower Beltian
metasediments have been drawn after
Hietanen (1968). 1Sf) So far, we have
analyzed the detrital zircons of a garnet-
muscovite-biotite schist (1-119) and of a
quartzite (1-100) and the euhedral zir-
cons of a migmatite (1-9) and of two
plutonic rocks (quartz monzonite 1-50,
quartz diorite 1-252). The analytical re-
116°00'
St JoeRiver_
°tif9
^g? +!:5_3_^er
S t a u r o 1 1 1 e Si
/ 7
+ Kyanite/ /A
/ \
/
/
/
| I
/ Kyanite . Andalusite '
♦ Sil limanite, + Kyanite
r-i^o,,,t n \v-i
\ Staurolite . \ ,
\ \ . I
\ * Kyanite \ I /.
\ \ i si
t Kyanite \ 0 H00
Staurolite -»Chlorite + Mu
Chloritoid ->Chlorite
IDAHO BATHOLITH
Granite
Quartz monzonite
Quartz diorite
Lower Beltian
metasediments
Sample locations
+ Rb-Sr sample
O Zircon sample
Fig. 85. Simplified map of the metamorphic zones in the lower Beltian metasediments of north-
ern Idaho (after Hietanen, 1968)185 with the sample locations: BC, Beaver Creek pluton; Bu.
Bungalow pluton; RT , Roundtop pluton.
298
CARNEGIE INSTITUTION
pb206/LJ238
0.3
ZIRCONS OF METASEDIMENTS
AND PLUTONIC ROCKS FROM
NORTHERN IDAHO
0.2
0.04
0.03
-0.02
-0.01
Pb207/U235
Fig. 86. Concordia diagram showing the isotopic data of zircons from lower Beltian metasedi-
ments and the Idaho batholith.
suits are plotted on the concordia dia-
gram in Fig. 86. The high apparent ages
O1400 m.y.) of the detrital zircons are
consistent with the whole-rock isochron
of about 1300 m.y. which Obradovich
and Peterman (1968)18G obtained on un-
metamorphosed equivalents of the Belt
Series in Montana. K-Ar ages of the
metamorphic basement that underlies the
Beltian sediments are on the order of
1600 m.y. (Giletti, 1966). 18T Our zircon
data are consistent with the interpreta-
tion that these basement rocks consti-
tuted the source of the Beltian sediments.
The euhedral zircons of the migmatite
and the plutonic rocks have far younger
U-Pb ages, although the Pb207/Pb206
ages are about as high as those of the de-
trital zircons in the sediments. From this
result, one may suppose assimilation of
some older metasediments or infer a
general palingenic origin of the batholith
rocks. From the microscopic study of the
zircons, however, there is no indication
that the zircon fractions analyzed are
mixtures of old, inherited grains with
newly formed crystals. It is therefore
concluded that a visually unrecognizable,
older zircon component must be some-
how preserved within the crystals. In
zircons of magmatic and migmatic rocks
of anatectic basement complexes, rounded
or irregularly shaped cores in euhedral or
subhedral crystals are a widely observed
phenomenon (see also page 286 of this
Report) . In most cases these cores have
optical properties that differ from those
of the host zircons: e.g., more or other
kinds of inclusions, darker or different
color, or relict zoning. The zircons of the
batholith, however, show no such cores.
Only zones showing accumulations of
small inclusions can be observed in the
center of some zircons. If these parts of
the crystals are older, inherited material,
the former crystallographic and optical
differences must have been obliterated
during crystallization or recrystallization
of the outer part because no internal
boundaries can be found.
In the foregoing discussion the as-
sumption was made that the formation
HALE OBSERVATORIES
299
of the migmatites and the emplacement
of the batholith rocks took place during
the Mesozoic to Tertiary. An alternative
interpretation compatible with the zircon
ages alone would regard these rocks as
at least 1300-m.y.-old intrusives that
underwent an intense Mesozoic to Terti-
ary metamorphism, resetting the U-Pb
clocks in the zircons to 5-20% of their
primary ages. However, such an inter-
pretation is difficult to reconcile with the
geological evidence. At least some of the
intrusions are clearly posttectonic. For
example, sample 1-50 was taken from the
Roundtop phi ton, which intrudes low-
to medium-grade metasediments and
which shows no evidence of subsequent
metamorphism. The U-Pb systematics
of the four zircon fractions of sample
1-252 (syntectonic quartz diorite) appear
to support the geological evidence. The
four data points form a straight line
which intersects the concordia curve at
about 70 m.y. Contrary to the pattern
generally observed for oogenetic zircon
populations that subsequently lost some
of their radiogenic lead, the uranium con-
tent in this population decreases with de-
creasing apparent U-Pb age. This un-
usual pattern is best explained as a mix-
ture of an old, uranium-rich zircon com-
ponent with a much larger amount of
young, uranium-poor overgrowth.
We conclude that the magmatic rocks
(1-50 and 1-252) and the migmatite
(1-9) were formed about 40 to 100 m.y.
ago and that all these rocks retained a
Precambrian memory which presumably
reflects the age of the premagmatic and
premigmatic parent rocks. Although old
lead components have been found in
magmatic zircons in many other places
(see for example Grauert and Arnold,
1968; 181 Gulson and Krogh, 1973; 188
also, this Report p. 288) , an effect of this
magnitude has not been reported so far.
If these zircons were analyzed 1000 m.y.
later with the aim of dating the time of
the intrusion, the measured age would be
up to 300 m.y. too high. Ages found by
extrapolation of best-fit lines or continu-
ous diffusion curves to the isotopic data
of zircons from old orthogneisses are
therefore doubtful unless the results are
confirmed by other methods.
Effect of Regional Metamorphism on
Whole-Rock Rb-Sr Systems in
Sediments
A. W. Hojmann and B. Grauert
One of the most challenging problems
of geochronology is to find a reliable
method to measure the age of recrystal-
lization in regionally metamorphosed
rocks. Mica ages (K-Ar and Rb-Sr) tend
to be reset during slow cooling or moder-
ate reheating of the terrane (Armstrong,
1966; 189 Jager et al, 1967 190) ; zircons
from metamorphic rocks frequently re-
cord a geologic history too complex to
resolve (Krogh and Davis, 1972; 191 see
also the reports by Krogh and Davis and
by Grauert in this Year Book) ; and Rb-
Sr isochron ages of whole-rock samples
may record premetamorphic ages (at
least in orthogneisses; see Lanphere et
al., 1964) 192 or the data may scatter in
a way that precludes age determination
(Hofmann, 1972). To be sure, all the
above methods have doubtless in some
cases yielded the age of metamorphism,
but this was possible because of fortunate
circumstances such as the growth of an
identifiable zircon fraction that has re-
mained concordant since its crystalliza-
tion (Krogh and Davis, this Year Book)
or the isotopic homogenization of stron-
tium throughout a large region due to
pervasive metasomatism. In most cases,
however, the interpretation of the appar-
ent ages measured on metamorphic rocks
remains uncertain.
One method which may be promising
in this context but which has been ap-
plied to only a few cases so far (Krogh
and Davis, 1969; 193 and this Year Book;
Grauert et al, 1973; 183 Hofmann, 1972) ,
is the application of the Rb-Sr method to
small, contiguous whole-rock volumes.
We have tested this method on five sam-
300
CARNEGIE INSTITUTION
pies of varying metamorphic grade from
the regionally metamorphic terrane of
northern Idaho. Sample locations and a
rough sketch of the geology of the area
are shown in Fig. 85 of the companion
report by Grauert and Hofmann. Two of
these samples were included in last year's
Report (Hofmann, 1972).
The rocks were cut into 1- to 2-cm-
thick slabs normal to the compositional
boundaries, and then into smaller pieces
ranging from 5 mm to 5 cm in width.
This method yielded samples that varied
sufficiently in their Rb-Sr ratios so that
isochron measurements could be at-
tempted.
The results (Fig. 87) are both disap-
pointing and encouraging. Two of the
samples (1-250 and 1-90) failed to yield
isochrons or even a positively correlated
array of data points. The other three
samples yielded fairly well defined iso-
chrons although a few points show real
scatter. The three isochron ages are: 64
m.y. for the sillimanite-grade mica schist
1-13, 340 m.y. for the garnet-grade
quartzitic schist 1-53, and 530 m.y. for
the biotite-grade schist with chlorite
0.830 -
«L 0.825
GO
0.820
0.815
1 1 1 1 1
SILLIMANITE
i i I
-MUSCOVITE
i
-
G
6^t
" D :
-
E .-
-X
A
:
i i
C
i i i
i
I-
■ i
13 :
0
4 6
Rb87/Sr86
10
0.714
*. 0.713
go
0.712
0.711
'i i i i i i i i i
SILLIMANITE -MUSCOVITE
B iE
■C
i i i i i i i i i i
0 0.2 0.6 1.0
Rb87/Sr86
1-250
1 1 ' 1 ' 1 I 1
STAUROLITE-KYANITE
0.774
\ 0.773
sT 0.772
- E
-
w 0.771
I C
0.770
0 7KQ
, i
1 D 1-90 "
i i i i i
Rb87/Sr
86
1.30
1.25
co
1.20
\
oo
1.15
1.10
0.84
-i | 1 1 r
L GARNET
-i — i — i — r
RbB7Srl
co
1.05-
1.00-
0.95
Rb87/Sr86
Fig. 87. Whole-rock Rb-Sr data for small slabs cut from five larger samples. The sample num-
bers are shown in the lower right-hand corner of each diagram. The metamorphic zone (silliman-
lte-muscovite, staurolite-kyanite, garnet, biotite) is indicated for each sample.
DEPARTMENT OF TERRESTRIAL MAGNETISM
301
porphyroblasts 1-209. This decrease in
apparent age with increasing meta-
morphic grade parallels, but does not
coincide with, the trend of the biotite Rb-
Sr ages reported by Hofmann (1972).
Because the biotite ages tend to be sig-
nificantly lower than the isochron ages,
the isochrons must involve at least three
Rb-Sr bearing phases and cannot be re-
garded as two-phase mixing lines.
Only additional analyses will show
whether these isochrons date discrete
metamorphic events or some other meta-
morphic evolution of an undefined na-
ture. If we are dating discrete episodes,
the metamorphic history of the region
must be considerably more complex than
is indicated by the field and petrographic
evidence (Hietanen, 1968). lsr> However,
we regard the presently available evi-
dence as insufficient for any detailed in-
terpretation of the metamorphic history.
The geological interpretation of most
published isochrons for metamorphic
rocks is hampered by the fact that it is
usually difficult or impossible to demon-
strate that a given isochron was really
produced by the metamorphism. In the
absence of independent stratigraphic in-
formation, this demonstration requires
evidence of either a rotation of a pre-
existing isochron or a convergence of iso-
topically heterogeneous samples toward
an isochron. In the case at hand, the
extremely high initial SrS7/Sr8G ratios
may be regarded as sufficient proof that
any preexisting isochron was indeed ro-
tated. To our knowledge, no sedimentary
isochron with an initial Sr87/Sr8G ratio
greater than 0.80 has ever been reported.
More specifically, the values of this ratio
reported by Obradovich and Peterman
(1968) 186 for nonmetamorphic rocks of
the Belt Series (i.e., the approximate
stratigraphic equivalents of our samples)
range from 0.707 to 0.733.
Further evidence that our small-slab
isochrons were produced by metamor-
phism is shown in Fig. 88, which contains
all whole-rock Rb-Sr data of samples of
'0
10 20
Rb87/Sr86
30
Fig. 88. Rb-Sr data for all whole-rock sam-
ples of low to intermediate metamorphic grade
(up to staurolite-kyanite zone). The data do
not define an isochron but cluster around an
1100-m.y. reference isochron.
low to intermediate metamorphic grade
up to and including the staurolite-kyanite
mineral zone. Samples 1-209 and 1-53
are represented in this plot by their
weighted averages of all the layers ana-
lyzed. In contrast with the plot shown in
last year's Year Book (Hofmann, 1972),
which included samples from the higher
metamorphic zones and which showed ex-
treme scatter with no apparent correla-
tion, these data are clearly correlated and
scatter about the reference isochron of
1100 m.y. It should be noted that several
of the samples (in particular, all those
that deviate significantly from the refer-
ence line) included in Fig. 88 are no
larger than fist size. It is clear that the
metamorphism did not effect any kind of
regional isotopic homogenization and
that the isotopic disturbances were suffi-
ciently small so as to preserve a some-
what deteriorated Precambrian isochron.
This Precambrian memory is largely de-
stroyed in the sillimanite zone where the
scatter on the isochron diagram increases
sharply even for samples from a single
outcrop (Hofmann, 1972).
The two slab samples (1-250 and 1-90)
which failed to yield isochrons were pre-
sumably affected by local disturbances
during or after metamorphic recrystal-
lization. Especially in the case of sample
1-250, a quartz-feldspar-biotite migma-
302
CARNEGIE INSTITUTION
tite from the sillimanite zone of meta-
morphism, the range of Rb/Sr and Sr87/
Sr86 ratios in the rock is so small that
even a small amount of selective meta-
somatic disturbance could destroy the
isochron relationship of the individual
layers. Sample 1-90 was collected from
an outcrop near the locality where excess
radiogenic argon of apparently meta-
somatic origin has previously been re-
ported (Hofmann, 1971 ).184 It may be
that this same metasomatism disturbed
the Rb-Sr system in this rock.
Additional measurements of small-slab
isochrons are needed to determine
whether these isochrons are locally repro-
ducible and whether they represent the
timing of discrete metamorphic events.
Although we are not yet in a position to
evaluate the exact age significance of
these isochrons, we have demonstrated
that they are not as easily reset by later
disturbances of the Rb-Sr system as are
biotite ages and that they do not "look
through" the metamorphism. Thus, ap-
plication of the Rb-Sr technique to small
slabs appears to be the most promising
way to measure metamorphic ages.
Uranium Gain of Detrital Zircons Studied
by Isotopic Analyses and Fission
Track Mapping
B. Grauert and M . G. Seitz
Grauert and Soptrajanova {Year Book
71, pp. 305-307) reported U-Pb isotopic
analyses of detrital zircons of meta-
quartzites from the Martell Valley,
Italian Alps. The quartzites form part of
a phyllite series and occur also as 1 to 5
cm thick inclusions in a large pegmatite
which intruded the metasediments.
Zircons from the inclusions were found
to be high in uranium (700-850 ppm)
compared to zircons of the country rock
(150-350 ppm), even though both zircon
populations were similar in appearance.
A close inspection of the coarser zircon
fractions revealed no additional over-
growths to account for the higher ura-
nium concentrations in the zircons of the
inclusions. Grauert and Soptrajanova
sil»l|
200 /jm
Fig. 89. Induced fission-track map of detrital zircons from a metaquartzite (sample MAR 3),
DEPARTMENT OF TERRESTRIAL MAGNETISM
303
therefore suggested that most of the ura-
nium had been added directly to the de-
trital zircons.
To test this possibility, we mapped the
uranium distributions in various zircon
fractions by means of induced fission
tracks. Muscovite was used to record
tracks from the neutron-induced fission
of uranium in polished grains (for de-
tails of the technique see Seitz, 1971 ).194
The fission tracks are revealed by chem-
ical etching and form a map of the ura-
nium in the samples. Figure 89 shows an
example of the uranium pattern in zir-
cons from the country rock; in this case,
the map is of red grains selected by hand
picking from the zircon separate. The
track map reveals large differences in
uranium distribution and concentration
levels. All sorts of patterns from homo-
geneous to patchy or zoned distributions
have been observed.
A fission track map of zircons from the
quartzite inclusions in the pegmatite is
shown in Fig. 90. This map reveals an
accumulation of uranium in the rims of
the grains and in some marginal hotspots.
These are general features of the zircons
from the inclusions. Similar patterns
have been found in a few grains of the
country rock too; but here the relative
marginal accumulation is much lower
and no complete rims were observed.
The uranium distribution shown in
Fig. 89 was certainly formed prior to
sedimentation of the zircons. The zoning
observed in some grains most likely re-
flects chemical variations during crystal
growth. We consider the patterns repre-
sented in Fig. 89, however, as evidence
for uranium gain after deposition of the
grains in the sediment. This gain oc-
curred primarily in the rims of the de-
trital grains without new crystal growth.
The present case clearly shows that ura-
nium gain can occur in zircons and is a
possible cause of discordant U-Pb ages.
3
k^***^
Fig. 90. Induced fission-track map of detrital zircons from metaquartzites found as inclusions
in a pegmatite (sample MAR 2).
304
CARNEGIE INSTITUTION
SPECIAL ACTIVITIES
As in previous years, we have con-
tinued our cooperative educational effort
with the University of Maryland's De-
partment of Astronomy. This year four
students, Gordon Powell, Cynthia
Cheung, Bill Guit, and Barbara Wil-
liams, working under Dr. W. C. Erickson
and our staff, were able to get first-hand
observational experience in two separate
projects. First, an attempt was made to
map the sun at 1.5 cm, using our new K-
band mixer. Because the half-power
beamwidth at this wavelength is only
four minutes of arc, it was hoped to see
the limb brightening and maybe some
active regions on the solar surface. Un-
fortunately, the poor figure of our dish
for this wavelength makes the sidelobe
response so bad that all of these fine fea-
tures were washed out. They decided to
switch to our 21 -cm hydrogen line re-
ceiver and made three series of observa-
tions as the moon occulted the galactic
plane. By carefully subtracting out the
lunar continuum as it swept through the
galactic plane, they hope to detect small
HI emission regions.
Jean Goad, a Ph.D. candidate in as-
tronomy at Harvard College, continued
her thesis work on internal motions in
M81. Observations were carried out at
Kitt Peak National Observatory. Her
work is partly under the direction of Dr.
Rubin.
Drs. Brown, James, Rice, and Rubin
taught a one-term course in "Contempo-
rary Physical Sciences" at a local public
high school, covering atomic physics,
geophysics, biophysics, and astronomy.
At the end of the term, the major com-
plaint was that the course was not con-
tinuing.
In Argentina during the past year
three students, M. Gil, M. Gordon, and
H. Pena completed studies at the Insti-
tute Argentino de Radioastronomia qual-
ifying them for the Licenciado degree of
the University of Buenos Aires. W.
Poppel of the Institute was awarded his
Ph.D. at the University of Buenos Aires
for a thesis entitled "Models of cosmic
clouds in gravitational contraction."
Graduate students from the University
of La Plata or the University of Buenos
Aires beginning or continuing their
studies at the I. A. R. now number ten.
Forty-two scientists working in diverse
fields ranging from ceramics and metal-
lurgy to geochemistry and field geology
met to discuss "Geochemical Transport
and Kinetics." The ClW-supported con-
ference was organized by Drs. Albrecht
W. Hofmann (Department of Terrestrial
Magnetism), Hatten S. Yoder, Jr. (Geo-
physical Laboratory), Bruno J. Giletti
(Brown University) , Richard A. Yund
(Brown University) . It was held at Airlie
House, in Warrenton, Virginia, from
June 4 to 6, 1973.
Rates of diffusion, vapor transport, ex-
solution, and irreversible chemical reac-
tions are of fundamental importance to
the understanding of geochemical proc-
esses in the earth's crust and mantle
Little is known at present about these
rates and the factors controlling them.
The need for systematic research on rate
processes is widely recognized, but until
recently, progress in this field has been
disappointingly slow. Modern instrumen-
tation has opened new opportunities for
the solution of many of the problems that
were difficult to study in the past. For
example, it is now possible to analyze
diffusion gradients on the scale of a few
micrometers with the electron micro-
probe, and the electron microscope re-
solves incipient exsolution lamellae only
150 angstroms wide.
The attending scientists presented a
wide range of views on the subject: one
participant discussed "vacancy wind" in
solids; another held up an object and ex-
plained that "this is a rock." Sixteen
contributions were concerned with solid-
state diffusion, with particular emphasis
on common, rock-forming minerals such
DEPARTMENT OF TERRESTRIAL MAGNETISM
305
as feldspar, olivine, quartz, mica, and
calcite. Introductory reviews on diffusion
mechanisms in simple crystals and on
diffusion in sulfides illustrated the diffi-
culties one encounters in measuring diffu-
sion in crystals even when their structure
is comparatively simple.
Other topics of discussion were diffu-
sion theory in multicomponent systems,
diffusion related to geochronology, ki-
netics of calcite recrystallization, kinetics
of exsolution and disordering in feldspars,
equilibrium and mass transfer among
minerals and aqueous solutions at high
temperatures and pressures, and diffusion
in granitic melts.
An evening session dealt with the use
of hydrogen, oxygen, strontium, and
argon isotopes as tracers in geological
processes. The final day of the confer-
ence was devoted to transport in sedi-
mentary and metamorphic rocks. The
contributions covered a number of theo-
retical aspects (steady-state diffusion in
metamorphism, recrystallization under
nonhydrostatic stress, models of infiltra-
tion and diffusion metasomatism, and
volatile flux during metamorphism) as
well as geological observations (diffu-
sional permeability of sediments, the
relationship of metamorphic differentia-
tion to vein mineral assemblages, and
metasomatic zoning in skarns).
The conference discussion made it ap-
parent that research on chemical trans-
port and kinetics has made a significant
beginning in the earth sciences. In order
to encourage others to join in this re-
search, the Carnegie Institution of Wash-
ington will publish a volume of the pro-
ceedings of the conference. The book is
being assembled and edited by the or-
ganizers and will contain both research
contributions and review papers pre-
sented at the conference.
Dr. Aldrich has completed a term as
a representative of the American Geo-
physical Union to the Division of Earth
Sciences of the National Research
Council.
Dr. Hart has been appointed a co-chief
scientist for leg 34 of the Deep Sea Drill-
ing Project. This cruise of the Glomar
Challenger is scheduled for December
1973 through January 1974, and will tra-
verse the Pacific Ocean from Tahiti to
Chile. On this leg, the drilling ship will
occupy only three sites, with the objec-
tive of penetrating as deeply as possible
(hundreds of meters) into the upper ba-
saltic layer of the oceanic crust. He has
also been appointed to the Panel on Ori-
entations for Geochemistry of the U.S.
National Committee for Geochemistry,
National Academy of Sciences-National
Research Council. Dr. Hart recently
finished a three-year term on the Ad-
visory Panel to the Earth Science Divi-
sion, National Science Foundation and a
three-year term as Associate Editor of
Reviews of Geophysics and Space
Physics.
Dr. Hover has been appointed Chair-
man of the Adult Development and
Aging Research and Training Committee
of the National Institute of Child Health
and Human Development.
Dr. James has been appointed a mem-
ber of the Statutes and By-Laws Com-
mittee and Chairman of the Committee
on Publications of the American Geo-
physical Union.
Dr. R. B. Roberts is serving as an
Associate of the Neurosciences Research
Program, Brookline, Massachusetts.
Dr. Vera C. Rubin has been elected to
the Board of Directors of the Association
of Universities for Research in Astron-
omy, Inc., as a Director-at-Large, to
serve for a term of three years, through
their Annual Meeting in 1976. AURA,
Inc. operates the Kitt Peak National
Observatory, Tucson, Arizona, and the
Cerro Tololo Inter-American Observa-
tory, La Serena, Chile, under contract
with the National Science Foundation.
She has also been elected as a member of
the Nominating Committee of the Amer-
ican Astronomical Society for three years
306
CARNEGIE INSTITUTION
and has been appointed a member of the Dr. Sacks has been appointed a
Space Astronomy Committee of the member of the Committee on Seismology
Space Science Board of the National of the Division of Earth Sciences, Na-
Academy of Sciences for the period Jan- tional Academy of Sciences-National
uary 1973 through March 1975. Research Council.
REFERENCES CITED
1. Rubin. Vera C., and W. Kent Ford, Jr.,
Astrophys. J., 150, 379. 1970.
2. Rubin, Vera C, and W. Kent Ford, Jr.,
Astrophys. J., 170, 25, 1971.
3. Rubin, Vera C. C. Krishna Kumar, and
W. Kent Ford, Jr., Astrophys. J ., 177, 31,
1972.
4. Rubin, Vera C, W. Kent Ford, Jr., and
C. Krishna Kumar, Astrophys. J., 181, 61,
1973.
5. Oke. J. B., Astrophys. J. Lett., 161, L17,
1970.
6. Lowrance, J. L., D. C. Morton, P. Zuc-
chino, J. B. Oke, and M. Schmidt, Astro-
phys. J., 171, 233, 1972.
7. Oke, J. B., G. Neugebauer, and E. E.
Becklin, Astrophys. J., 159, 341, 1970.
8. Bahcall, J. N., Ben-Zion Kozlovsky, and
E. E. Salpeter, Astrophys. J., 171, 467,
1972.
9. Angione, R. J., Ph.D. thesis, University of
Texas at Austin, 1970.
10. Gregory, P. C. P. P. Kronberg, E. R.
Seaquist, V. A. Hughes, A. Woodsworth,
M. R. Viner, and D. Retallack, Nature,
230, 443, 1972.
11. Hjellming, R. M., and B. Balick, Nature,
230, 443. 1972.
12. Lauque, R., J. Lequeu, N. Q. Rieu, Nature
(Physical Science), 230, No. 95, 119, 1972.
13. Minkowski, R., Rev. Mod. Phys., 30, 1048,
1958.
14. Carswell. R. F, and P. A. Strittmatter,
Nature, 242, 394, 1973.
15. Lande, S., A. Witter, and D. de Wied,
J. Biol. Chem., 246, 2058-2062, 1971.
16. Lande, S., J. B. Flexner, and L. B. Flexner,
Proc. Natl. Acad. Sci. USA, 69, 558-560,
1972.
17. Kit, S, J. Mol. Biol, 3, 711, 1961.
18. Waring, M., and R. J. Britten, Science,
154, 791, 1966.
19. Sutton, W. D., and M. McCallum, J. Mol.
Biol, 71, 633, 1972.
20. Rice, N. R., and N. Straus, Carnegie Inst.
Wash. Year Book 71, 264, 1972.
21. Lieberman, M. W., Fed. Proc, 32, Fed.
Amer. Soc. Exp. Biol, 580, 1972.
22. Roderick, T. H., Mutat. Res., 11, 59, 1971.
23. Ellerman, J. R., The Families and Genera
of Living Rodents, Vol. 1, Rodents other
than Muridae, Brit. Mus. Nat. Hist., Lon-
don, 1941.
24. Simpson, G. G., The principles of classi-
fication and a classification of mammals,
Bull. Amer. Mus. Natur. Hist., 85, 1945.
25. Arata, A. A., Muroid, Gliroid, and Dipo-
doid rodents, in Recent Mammals of the
World, A Synopsis of Families, Anderson
and Jones, eds., Ronald Press, N. Y., pp.
226-253, 1967.
26. Davis, D. H. S.. S. Af. J. Med. Sci., 28, 53,
1963.
27. Kohne, D. E., J. A. Chiscon, and B. H.
Hoyer, J . Human Evolution, 1, 627, 1972.
28. Hennig, W., and P. M. B. Walker, Nature,
225, 915, 1970.
29. Britten. R. J., and J. Smith, Carnegie Inst.
Wash. Year Book 68, 378-386, 1970.
30. Davidson, E., C. Amenson, B. Hough, and
R. J. Britten, Carnegie Inst. Wash. Year
Book 71, 273-276, 1972.
31. Davidson, E. H., B. R. Hough, C. S.
Amenson, and R. J. Britten, J. Mol. Biol,
77, 1-23, 1973.
32. Alberts. B. M., and P. Doty, J. Mol. Biol,
32, 379-403, 1968.
33. Mulder, C., and P. Doty, J. Mol Biol, 32,
423-435, 1968.
34. Cocks, G. T., and A. C. Wilson, J. Bac-
teriol, 110, 793-802, 1972.
35. Murphy, T. M., and S. E. Mills, J. Bac-
teriol, 97, 1310-1320, 1969.
36. Brenner, D. J., G. R. Fanning, K. E. John-
son, R. V. Citarella, and S. Falkow, J.
Bacteriol, 98, 637-650, 1969.
37. McCarthy, B. J., and E. T. Bolton, Proc.
Natl Acad. Sci. USA, 50, 156, 164, 1963.
38. Stanier, R. Y., D. Wachter, C. Gasser, and
A. C. Wilson, J. Bacteriol, 102, 351-362,
1970.
39. Cohen, G. N., R. Y. Stanier, and G. Le
Bras, /. Bacteriol, 99, 791-801, 1969.
40. Truffa-Bachi, P., G. Le Bras, and G. N.
Cohen, Biochim. Biophys. Acta, 128, 450,
1966.
DEPARTMENT OF TERRESTRIAL MAGNETISM
307
41. Cowie, D. B., P. Truffa-Bachi, Jean-Marc
Costrejean, Marie-Claire Py, and G. N.
Cohen, in press, 1973.
42. Monod, J., J. Wyman, and J. Changeux,
/. Mol. Biol., 12, 88, 1965.
43. Hill, L. R., J. Gen. Microbiol., 44, 419-
437, 1966.
44. Brenner, D. J., et al., J. Bacteriol., 98, 637-
650, 1969.
45. Hayes, F. N., E. H. Lilly, R. I. Ratcliff,
D. A. Smith, and D. L. Williams, Thermal
transitions in mixtures of polydeoxyribo-
nucleotides, Biopolymers, 9, 1105-1117,
1970.
46. Hoyer, B. H., and N. W. van de Velde,
Carnegie Inst. Wash. Year Book 69, p. 513,
1971.
47. Commerford, S. L., Biochemistry , 10, 1993,
1971.
48. Sutton, W. D.. Biochim. Biophys. Acta,
240, 522, 1971.
49. Anderson, D. L., Earth's viscosity, Science,
151, 321-322, 1966.
50. Sacks, I. S., Viscosity and Q, Carnegie
Inst. Wash. Year Book 71, 327-329. 1972.
51. Sacks, I. S., Q for P waves in the mantle,
Carnegie Inst. Wash. Year Book 66, 28-29,
1968.
52. Sacks, I. S., A broad-band large dynamic
range seismograph, Amer. Geophys. Union
Mono. 10, 543-553, 1966.
53. Sacks, I. S., Distribution of absorption of
shear waves in South America and its tec-
tonic significance, Carnegie Inst. Wash.
Year Book 67, 339-344, 1969.
54. Sacks, I. S., The Q structure of South
America, Carnegie Inst. Wash. Year Book
70, 340-343, 1971.
55. Alexander, S. S., Crust-mantle structure of
shields and their role in global tectonics,
EOS, Trans. Amer. Geophys. Union, 53,
No. 11, 1972.
56. Utsu, T., Anomalies in seismic wave veloc-
ity and attenuation associated with a deep
earthquake zone (1), J. Fac. of Sci., Hok-
kaido Univ., Ser. 7: Geophysics 3, No. 1,
1967.
57. Farberov, A. I., and V. I. Gorelchik,
Anomalous seismic effect under volcanoes
and some features of deep-seated struc-
ture of volcanic areas, Bull. Volcanol.,
Tome XXXV-1, 212-224, 1971.
58. Tsujiura, M., Spectra of body waves and
their dependence on source depth 1. Jap-
anese arc, /. Phys. Earth, 20, 251-266, 1972.
59. Oliver, J., and B. Isacks, Deep earthquake
zones, anomalous structure in the upper
mantle and the lithosphere, J. Geophys.
Res., 72, 4259-4275, 1967.
60. Barazangi, M., B. Isacks, and J. Oliver,
Propagation of seismic waves through and
beneath the lithosphere that descends
under the Tonga Island arc, J. Geophys.
Res., 77, 952-958, 1972.
61. Sacks, I. S., A path for high-frequency
shear (S) waves in the upper mantle.
Carnegie Inst. Wash. Year Book 66, 31-32
1968.
62. Isacks, B. L., and M. Barazangi, High-
frequency shear waves guided by a contin-
uous lithosphere descending beneath west-
ern South America. Cornell University,
Department of Geological Sciences, Con-
tribution No. 549, to appear in Geophys.
J. Roy. Astron. Soc, 1973.
63. Sacks, I. S., Distribution of absorption of
shear waves in South America and its
tectonic significance, Carnegie Inst. Wash.
Year Book 67, 339-344, 1969.
64. Okada, H., Forerunners of ScS wave from
nearby deep earthquakes and upper man-
tle structure in Hokkaido (in Japanese),
Jishin: J. Seismol. Soc. Jap., Ser. 2, 24,
228-239. 1971.
65. Isacks, I., and P. Molnar, Distribution of
stresses in the descending lithosphere from
a global survey of focal-mechanism solu-
tions of mantle earthquakes, Rev. Geo-
phys., 9, 103-174, 1971.
66. Fukao, Y., Source process of a large deep-
focus earthquake and its tectonic implica-
tions— the western Brazil earthquake of
1963, Phys. Earth Planet. Interiors, 5,
61-76, 1972.
67. Sykes, L. R., Seismicity as a guide to
global tectonics and earthquake predic-
tion. In The Upper Mantle, A. R. Ritsema.
ed., Elsevier, 393-414, 1972.
68. McCamy, K., R, P. Meyer, and T. J.
Smith, Generally applicable solutions of
Zoeppritz' amplitude equations. Bull. Seis-
mol. Soc. Amer., 52, 923-955, 1962.
69. Costain, J. K.. K. L. Cook, and S. T.
Algermissen, Amplitude, energy, and phase
angles of plane SV waves and their appli-
cation to the earth crustal studies. Bull.
Seismol. Soc. Amer., 53. 1039-1074, 1963.
70. Talwani, M., G. H. Sutton, and J. L.
Worzel, A crustal section across the Puerto
Rico trench, J. Geophys. Res., 64, 1545-
1555, 1959.
71. Archambeau, C. B., General theory of
elastodynamic source fields, Rev. Geo-
phys., 16, 241-288. 1968.
72. Archambeau, C. B., The theory of stress
wave radiation from explosions in pre-
308
CARNEGIE INSTITUTION
stressed media, Geophijs. J. Roy. Astron.
Soc, 29, 329-366, 1972; 31, 361-363, 1973.
73. Archambeau, C. B., and J. B. Minster,
Private communication, April, 1973.
74. Linde, A. T., Source-spectral relations
from elastodynamic theory, J. Geophys.
Res., 76, 8066-8068, 1971.
75. Randall, M. J.. Spectral peaks and earth-
quake source dimensions, J. Geophys.
Res., 78, 2609-2611, 1973.
76. Aki, K., Scaling law of seismic spectrum,
J. Geophys. Res., 72, 1217-1231, 1967.
77. Ben-Menahem, A., Radiation of seismic
body-waves from a finite moving source,
/. Geophys. Res., 67(1), 345-350, 1962.
78. Haskell, N., Total energy and energy spec-
tral density of elastic wave radiation from
propagating faults, Bull. Seismol. Soc.
Amer., 54, 1811-1842, 1964.
79. Linde, A. T., and I. S. Sacks, Dimensions,
energy, and stress release for South Amer-
ican deep earthquakes, J. Geophys. Res.,
77, 1439-1451, 1972.
80. Sacks, I. S, Shigeji Suyehiro, D. W. Evert-
son, and Yokichi Yamagishi, Pap. in
Meteorot. Geophys., 22, Nos. 3-4, 195-208,
1971.
81. Case, J. E., L. G. Duran S., A. Lopez R.,
and W. R. Moore, Tectonic investigations
in western Colombia and eastern Panama,
Geol. Soc. Amer. Bull, 82, 2685-2712, 1971.
82. Case, J. E., J. Barnes, G. Paris, H. Gon-
zalez, and A. Vina, Trans- Andean geo-
physical profile, southern Colombia, Geol.
Soc. Amer. Bull, 84, 2895-2904, 1973.
83. Zeil, W., and H. Pichler, Die kanozoische
rhyolith-Formation in mittleren Abschnitt
der Anden, Geol. Rundsch., 57, 48-81, 1967.
84. Pichler, H., and W. Zeil, Die quartare
"Andesif'-Formation in der Hochkordil-
lere Nord-Chiles, Geol. Rundsch., 5S, 866-
903, 1969.
85. Pichler, H., and W. Zeil, The Cenozoic
rhyolite-andesite association of the Chil-
ean Andes, Bull. Volcanol., 35-2, 424-452,
1972.
86. Dickinson, W. R., and Trevor Hatherton,
Andesitic volcanism and seismicity around
the Pacific, Science, 157, 801-803, 1967.
87. Hatherton, T., and W. R. Dickinson,
Andesitic volcanism and seismicity in New
Zealand, /. Geophys. Res., 73, 4615-4619,
1968.
88. Hatherton, T., and W. R. Dickinson, The
relationship between andesitic volcanism
and seismicity in Indonesia, the Lesser
Antilles, and other island arcs, /. Geophys.
Res., 74, 5301-5310, 1969.
89. Jenks, W. F., and S. S. Goldich, Rhyolitic
tuff flows in southern Peru, J. Geol., 64,
156-172, 1956.
90. Guest, J. E., Upper Tertiary ignimbrites in
the Andean cordillera of part of Antofa-
gasta province, northern Chile, Geol. Soc.
Amer. Bull, 80, 337-362, 1969.
91. Dickinson, W. R., Relations of andesites,
granites, and derivative sandstones to arc-
trench tectonics, Rev. Geophys. (Space
Phys.), 8, No. 4, 813-860, 1970.
92. James, D. E., Plate tectonic model for
the evolution of the central Andes, Geol.
Soc. Amer. Bull, 82, 3325-3346, 1971.
93. Schilling, J. G., Iceland mantle plume
existence, and influence along the Rey-
kjanes Ridge: Geochemical evidence,
Nature, 242, 565, 1973.
94. Morgan, W. J., Convection plumes in the
lower mantle, Nature, 230, 42-43, 1971.
95. Vogt, P. R., Asthenosphere motion re-
corded by the ocean floor south of Iceland,
Earth Planet. Sci. Lett., 13, 153-160, 1971.
96. Kuno, H., Mafic and ultramafic nodules
in basaltic rocks of Hawaii, Geol. Soc.
Amer. Mem., 115, 189-234, 1969.
97. Kuno, H., and K. Aoki, Chemistry of
ultramafic nodules and their bearing on
the origin of basaltic magmas, Phys. Earth
Planet. Interiors, 3, 273-301, 1970.
98. White, R. W., Ultramafic inclusions in
basaltic rocks from Hawaii, Contrib. Min-
eral. Petrol, 12, 245-312, 1966.
99. Jackson, E. D., and T. L. Wright, Xeno-
liths in the Honolulu volcanic series,
Hawaii, J. Petrology, 11, 405-430, 1970.
100. Griffin, W. L., and V. R. Murthy, Distri-
bution of K, Rb, Sr and Ba in some min-
erals relevant to basalt genesis, Geo-
chim. Cosmochim. Acta, 33, 1389-1414,
1969.
101. Philpotts, J. A., C. C. Schnetzler, and
H. H. Thomas, Petrogenetic implications
of some new geochemical data on eclogitic
and ultrabasic inclusions, Geochim. Cos-
mochim. Acta, 36, 1131-1166, 1972.
102. Philpotts, J. A. and C. C. Schnetzler,
Phenocryst-matrix partition coefficients
for K, Rb, Sr and Ba, with applications to
anorthosite and basalt genesis, Geochim.
Cosmochim. Acta, 34, 307-322, 1970.
103. Shimizu, N., and S. Akimoto, Partitioning
of strontium between clinopyroxene and
liquid at high pressures: preliminary ex-
periments, Earth Planet. Sci. Lett., 13,
134-138, 1971.
104. Berlin, R., and C. M. B. Henderson, A
reinterpretation of Sr and Ca fractionation
DEPARTMENT OF TERRESTRIAL MAGNETISM
309
121.
122.
123.
124.
125.
trends in plagioclase from basic rocks, 119.
Earth Planet. Sci. Lett., 4, 79-83, 1968.
105. Drake, M. J., and D. F. Weill, Distribu-
tion of Sr, Ba, Eu2+, Eu3+, and other REE 120.
between plagioclase feldspar and mag-
matic liquid: an experimental study (ab-
stract), Trans. Amer. Geophys. Union, 53,
553, 1972.
106. Kushiro, I., and H. S. Yoder, Jr., Anorth-
ite-forsterite and anorthite-enstatite re-
actions and their bearing on the basalt-
eclogite transformation, J. Petrology, 7,
337-362, 1966.
107. Green, D. H., and A. E. Ringwood, An
experimental investigation of the gabbro
to eclogite transformation and its peno-
logical applications, Geochim. Cosmochim.
Acta, 31, 767-833, 1967.
108. Ito, K., and G. C. Kennedy, An experi-
mental study of the basalt-garnet granu-
lite-eclogite transition, in The Structure
and Physical Properties of the Earth's
Crust, J. J. Heacock, ed., Geophys. Mono.,
14, 303-314, 1971.
109. Green, D. H., and W. Hibberson, The in-
stability of plagioclase in peridotite at
high pressures, Lithos, 3, 209-221, 1970.
110. Herzberg. C. T., Stability fields of plagio-
clase- and spinel-lherzolite, Progr. Exp.
Petrol, 2, 145-148, 1972.
111. Boyd, F. R., A pyroxene geotherm, Geo-
chim. Cosmochim. Acta, in press, 1973.
112. Davis, B. T.C., and F. R. Boyd, The join
MgsSioOo-CaMgSiaOo at 30 kilobars pres-
sure and its application to pyroxenes from
kimberlites, J. Geophys. Res., 71, 3567-
3576, 1966.
113. Boyd, F. R., The system CaSiO,-MgSiO,-
A120:!, Carnegie Inst. Wash. Year Book 68,
214-221, 1970.
114. MacGregor, I. D., The system MgO-ALO,-
Si02: solubility of ALO.s in enstatite for 129.
spinel and garnet peridotite composition,
Amer. Mineral., in press, 1973.
115. Kushiro, I.. Y. Syono, and S. Akimoto,
Melting of a peridotite nodule at high
pressures and high water pressures, J. Geo-
phys. Res., 73, 60-23-6029. 1968. 130.
116. Green, D. H., The origin of the 'eclogites'
from Salt Lake Crater, Hawaii, Earth
Planet. Sci. Lett., 1, 414-420, 1966.
117. Kushiro, I., Partial melting of garnet
lherzolites in kimberlite at high pressures. 131.
in Lesotho Kimberlites, P. H. Nixon, ed.,
294-299, 1973.
118. Berg, G. W., Secondary alteration in eclo- 132.
gites from Kimberlite pipes, Amer. Min-
eral, 53, 1336-1346. 1968.
126.
127.
128.
Green, D. H., and A. E. Ringwood, The
genesis of basaltic magmas, Conlrib. Min-
eral. Petrol, 15, 103-190, 1967.
O'Hara, M. J.. The bearing of phase equi-
libria studies in synthetic and natural sys-
tems on the origin and evolution of basic
and ultrabasic rocks, Earth-Sci. Rev., 4,
69-133, 1968.
Boyd, F. R., and P. H. Nixon, Ultramafic
nodules from the Thaba Putsoa Kimber-
lite pipe, Carnegie Inst. Wash. Year Book
71, 362-373, 1972.
Boyd. F. R., A pyroxene geotherm, Geo-
chim. Cosmochim. Acta, in press, 1973.
Kushiro, I., Partial melting of garnet
lherzolites in kimberlite at high pressures,
in Lesotho Kimberlites, P. H. Nixon, ed.,
294-299, 1973.
Griffin, W. L., and V. R. Murthy, Distri-
bution of K, Rb, Sr and Ba in some min-
erals relevant to basalt genesis, Geochim.
Cosmochim. Acta, 33, 1389-1414, 1969.
Allsop, H. L., L. O. Nicolaysen, and P.
Hahn-Weinheimer, Rb/K ratios and Sr-
isotopic compositions of minerals in eclo-
gitic and peridotitic rocks, Earth Planet.
Sci. Lett., 5, 231-244, 1969.
Hutchison, R., and J. B. Dawson, Rb, Sr
and Sr87Sr86 in ultrabasic xenoliths and
host-rocks, Lashaine volcano, Tanzania.
Earth Planet. Sci. Lett., 0, 87-92, 1970.
Philpotts, J. A., C. C. Schnetzler, and
H. H. Thomas, Petrogenetic implications
of some new geochemical data on eclo-
gitic and ultrabasic inclusions, Geochim.
Cosmochim. Acta, SG, 1131-1166, 1972.
Shimizu, N., and S. Akimoto, Partitioning
of strontium between clinopyroxene and
liquid at high pressures: preliminary ex-
periments. Earth Planet. Sci. Lett., 13,
134-138, 1971.
Bell, K., and J. L. Powell, Strontium iso-
topic studies of alkalic rocks: the potas-
sium-rich lavas from the Birunga and
Toro-Ankole regions, east and central
equatorial Africa, /. Petrology, 10, 536-
572, 1969.
Shieh, Y. N., and H. P. Taylor, Jr., Oxy-
gen and hydrogen isotopic studies of con-
tact metamorphism in the Santa Rosa
Range, Nevada and other areas: Contrib.
Mineral and Petrol. 20, 306-356, 1969.
Vidale, R., Metasomatism in a chemical
gradient and the formation of calc-silicate
bands, Amer. J. Sci., 267, 857-874. 1969.
Korzhinskii, D. S., Theory of Metasomatic
Zoning, translated by J. Agrell. Oxford
U. Press, London, 162* pp., 1970.
310
CARNEGIE INSTITUTION
133. Hofmann, A. W., Chromatographic theory
of infiltration metasomatism and its appli-
cation to feldspars, Amer. J. Sci., 272,
69-90, 1972.
134. Korzhinskii, D. S., Theory of metasomatic
zoning. A reply to Dr. A. Hofmann. Amer.
J . Sci., in press, 1973.
135. Hofmann, A. W., Theory of metasomatic
zoning. A reply to Dr. D. S. Korzhinskii,
Amer. J . Sci., in press, 1973.
136. Carslaw, H. S., and J. C. Jaeger, Conduc-
tion of Heat in Solids, 2nd edition, Oxford
U. Press, London, 510 pp., 1959.
137. Crank, J., The Mathematics of Diffusion,
Oxford U. Press, London, 347 pp., 1956.
138. Orville, P. M., Plagioclase cation exchange
equilibria with aqueous chloride solution;
results at 700 °C and 2000 bars in the pres-
ence of quartz, Amer. J. Sci., 272, 234-272,
1972.
139. Orville, P. M., Alkali ion exchange be-
tween vapor and feldspar phases, Amer. J.
Sci., 261, 201-237, 1963.
140. Hofmann, A. W., and B. J. Giletti, Diffu-
sion of geochronologically important nu-
clides in minerals under hydrothermal
conditions, Eclogae Geol. Helv., 63, 141-
150, 1970.
141. Andersen, C. A., and J. R. Hinthorne, Ion
microprobe mass analyzer, Science, 175,
853-860, 1972.
142. Goetze, C. and W. F. Brace, Laboratory
observations of high-temperature rheology
of rocks, Tectonophysics, 13, 583, 1972.
143. Nye, J. F., The flow law of ice from meas-
urements in glacier tunnels, laboratory ex-
periments, and the Jungfrau borehole
experiment, Proc. Roy. Soc, Ser. A, 219,
477-489, 1953.
144. Calladine, C. R., and D. C. Drucker,
Nesting surfaces of constant rate of energy
dissipation in creep, J. Appl. Math., 29,
79-84, 1962.
145. Sherwin, Jo-Ann, and W. M. Chappie,
Wavelengths of single-layer folds: a com-
parison between theory and observation,
Amer. J. Sci., 266, 167-179, 1968.
146. Groshong, R. H., Jr., Strain in minor folds,
Valley and Ridge Province, Pennsylvania,
Ph.D. thesis. Brown University, 220 pp.,
1971.
147. Fletcher, R. C, and R. H. Groshong, Jr.,
Rock rheology from wavelength selection
in folding (abstract), Trans. Amer. Geo-
phys. Union, 54, 456, 1973.
148. Hopson, C. A., The crystalline rocks of
Howard and Montgomery Counties, in
The Geology of Howard and Montgomery
Counties: Baltimore, Maryland, Geol.
Survey, 27-215, 1964.
149. Higgins, N. W., Age, origin, regional rela-
tions, and nomenclature of the Glenarm
series, central Appalachian Piedmont: A
reinterpretation, Geol. Soc. Amer. Bull.,
83, 989-1026, 1972.
150. Knopf, E. B., and A. I. Jonas, Geology of
the crystalline rocks, Baltimore County,
Baltimore, Maryland, Geol. Survey, 97-
199, 1929.
151. Tilton, G. R., G. L. Davis, G. W. Wether-
ill, and L. T. Aldrich, Isotopic ages of
zircon from granites and pegmatites,
Trans. Amer. Geophys. Union, 38, 360-
371, 1958.
152. Tilton, G. R., B. R. Doe, and C. A.
Hopson, Zircon age measurements in the
Maryland Piedmont, with special refer-
ence to Baltimore Gneiss problems, in
Studies of Appalachian Geology — Central
and Southern, pp. 429-434, G. W. Fisher
et al., eds., New York, Interscience, 1970.
153. Wetherill, G. W, G. R. Tilton, G. L.
Davis, S. R. Hart, and C. A. Hopson,
Age measurements in the Maryland Pied-
mont, J. Geophys. Res., 71, 2139-2155,
1966.
154. Wetherill, G. W., G. L. Davis, and C. Lee-
Hu, Rb-Sr measurements on whole rocks
and separated minerals from the Balti-
more Gneiss, Maryland, Geol. Soc. Amer.
Bull, 79, 757-762, 1968.
155. Steiger, R. H., and C. A. Hopson. Mini-
mum age of the Glenarm Series, Balti-
more, Maryland: Geol. Soc. Amer., Abstr.
for 1964, Spec. Pap. 82, 194, 1964.
156. Tilton, G. R., Volume diffusion as a
mechanism for discordant lead ages, J.
Geophys. Res., 65, 2933-2945, 1960.
157. Wasserburg, G. J., Diffusion processes in
lead-uranium systems, J. Geophys. Res.,
68, 4823-4846, 1963.
158. Silver, L. T., The use of cogenetic ura-
nium-lead isotope systems in zircons in
geochronology, in Radioactive Dating,
Intern. Atomic Energy Agency Symp.
Athens, 1962 Proc, 279-287, 1963.
159. Davis, G. L., G. R. Tilton, L. T. Aldrich,
S. R. Hart, and R. H. Steiger, Geochron-
ology and isotope geochemistry, Carnegie
Inst. Year Book 64, 165-177, 1965.
160. Wagner, M. E., Metamorphism of the
Precambrian Baltimore Gneiss in south-
eastern Pennsylvania, Ph.D. thesis, Bryn
Mawr College, 89 pp., 1972.
161. Wagner, M. E., and M. L. Crawford,
Metamorphism of the Precambrian Balti-
more Gneiss in southeastern Pennsylvania
DEPARTMENT OF TERRESTRIAL MAGNETISM
311
(abstract), Geol. Soc. Amer., Abstracts
with Programs, 5, 234, 1973.
162. Tilton, G. R., G. W. Wetherill, G. L.
Davis, and M. N. Bass, 1000-million-year
old minerals from the eastern United
States and Canada, ./. Geophys. Res., 65,
4137-4179, 1960.
163. Koppel, V.. and J. Sommerauer, Electron
microprobe study of trace elements in
zircons having different resistability to
lead loss, Fortsch. Mineral, 50, 27-28,
1973.
164. Rankin, D. W., T. W. Stern, J. C. Reed,
and M. F. Newell, Zircon ages of felsic
volcanic rocks in the upper Precambrian
of the Blue Ridge, Appalachian Mountains,
Science, 166, 741-744, 1969.
165. Hall, L. M., Times of origin and deforma-
tion of bedrock in the Manhattan Prong,
in Studies of Applachian Geology — north-
em and maritime, pp. 117—127, E-an Zen
et al., eds., New York, Interscience, 1968.
166. Hall, L. M., Some stratigraphic relation-
ships within the New York City group in
Westchester County, New York (ab-
stracts). Geol. Soc. Amer. Spec. Pap. 87,
69, 1966.
167. Merill, F. J., On the metamorphic strata
of southeastern New York, Amer. J. Sci.,
39, 383-392, 1890.
168. Merill, F. J., The geology of the crystal-
line rocks of southeastern New York,
New York State Museum Ann. Rept. No.
50, Appendix A, 21-31. 1896.
169. Cady, W. M., Stratigraphy and structure
of westcentral Vermont, Geol. Soc. Amer.
Bull., 58, 515-587, 1945.
170. Zen, E-an, ed., Stratigraphy and structure
of westcentral Vermont and adjacent New
York, Guidebook 51st Ann. Mtg. New
England Intercollegiate Geol. Conf., Rut-
land, Vt.. 87 pp., 1959.
171. Zen, E-an, Time and space relationships
of the Taconic allochthon and autochthon,
Geol. Soc. Amer. Spec. Pap. 97, 107 pp.,
1967.
172. Ratcliffe, N. M., and R. R. Knowles,
Fossil evidence from the Manhattan
Schist-Inwood marble sequence at Ver-
planck, New York (abstract), Geol. Soc.
Amer., Program for Northeastern Section
Meeting, Washington, D.C., 48-49, 1968.
173 Long, L. E., Whole-rock Rb-Sr age of the
Yonkers Gneiss, Manhattan Prong., Geol.
Soc. Amer. Bull, 80, 2087-2090, 1969.
174. Long, L. E., and J. L. Kulp, Isotopic age
study of the metamorphic history of the
Manhattan and Reading Prongs., Geol.
Soc. Amer. Bull, 73, 969-996, 1962.
175. Long, L. E., Isotopic ages from northern
New Jersey and southern New York, Aim.
N. Y. Acad. Sci., 91, 400-407, 1961.
176. Long, L. E., Isotopic age study, Dutchess
County, New York, Geol Soc. Amer.
Bull, 73, 997-1005. 1962.
177. Bence, A. E., and V. Rajamani, 40Ar/89Ar
incremental heating "ages" of muscovites
and biotites from progressive metamorphic
terrain (abstract), Geol Soc. Amer., Ab-
stracts with Programs, 4, 449, 1972.
178. Tilton, G. R„ G. W. Wetherill, G. L.
Davis, and C. A. Hopson. Ages of minerals
from the Baltimore Gneiss near Baltimore.
Maryland, Geol Soc. Amer. Bull, 60.
1469-1474, 1958.
179. Grauert, B., New U-Pb isotopic analyses
of zircons from the Baltimore Gneiss and
the Setters Formation, Carnegie Inst.
Year Book 71, 301-305, 1972.
180. Davis, G. L., G. R. Tilton, and G. W.
Wetherill. Mineral ages from the Appa-
lachian Province in North Carolina and
Tennessee, J. Geophys. Res., 67, 1987-
1996, 1962.
181. Grauert, B., and A. Arnold, Deutung dis-
kordanter Zirkonalter der Silvrettadecke
und des Gotthardmassivs (Schweizer Al-
pen), Contrib. Mineral. Petrol, 20, 34-56,
1968.
182. Pidgeon, R. T., V. Koppel, and M. Griin-
enfelder. U-Pb isotopic relationships in
zircon suites from a para- and ortho-
gneiss from the Ceneri Zone, southern
Switzerland, Contrib. Mineral. Petrol, 26,
1-11, 1970.
183. Grauert, B., R. Hiinny, and G. Soptraja-
nova, Age and origin of detrital zircons
from the pre-Permian basements of the
Bohemian massif and the Alps, Contrib.
Mineral. Petrol, //), 105-130, 1973.
184. Hofmann, A.. Effect of regional metamor-
phism on radiometric ages in pelitic rocks
— preliminary results, Carnegie Inst. Year
Book 70, 242-243, 1971.
185. Hietanen, A., Belt series in the region
around Snow Peak and Mallard Peak,
Idaho, Geol Survey Prof. Paper 344-E,
34 pp., 1968.
186. Obradovich, J. D., and Z. E. Peterman.
Geochronology of the Belt Series, Mon-
tana. Can. J. Earth Set.. 5, 737-747. 1968.
187. Giletti, B. J., Isotopic ages from south-
western Montana, J. Geophys. Res., 71,
4029-4036, 1966.
312
CARNEGIE INSTITUTION
188. Gulson, B. L., and T. E. Krogh, Old lead
components in a young Alpine granitic
massif, Fortschr. Mineral, 50, 78, 1973.
189. Armstrong, R. L.. K-Ar dating of plutonic
and volcanic rocks in orogenic belts, in
Potassium Argon Dating, 0. A. Schaeffer
and J. Zahringer, eds., Springer-Verlag,
New York, pp. 117-133, 1966.
190. Jager, E., E. Niggli, and E. Wenk, Rb-Sr
Altersbestimmungen an Glimmern der
Zentralalpen, Beitr. Geol. Karte Schweiz,
N. F. 134, 1967.
191. Krogh, T. E, and G. L. Davis, The effect
of regional metamorphism on U-Pb sys-
tems in zircons and a comparison with
Rb-Sr systems in the same whole rock,
Carnegie Inst. Year Book 71, 564-571,
1972.
192. Lanphere, M. A., G. J. F. Wasserburg,
A. L. Albee, and G. R. Tilton, Redistribu-
tion of strontium and rubidium isotopes
during metamorphism, World Beater Com-
plex. Panamint Range, California, in Iso-
topic and Cosmic Chemistry, H. Carig,
S. L. Miller, and G. J. Wasserburg, eds.,
North Holland Publishing Co., Amster-
dam, pp. 269-320, 1964.
193. Krogh, T. E., and G. L. Davis, Old iso-
topic ages in the northwestern Grenville
Province, Ontario, Geol. Ass. Can. Spec.
Pap., 5, pp. 189-192, 1969.
194. Seitz, M. G., Heavy ion radiation studies
in terrestrial and extraterrestrial materials,
Ph.D. thesis, Washington University, St.
Louis, Mo., 1971.
BIBLIOGRAPHY
Ahlfeld, C. E., G. E. Assousa, R. A. LaSalle,
W. J. Thompson, H. A. Van Rinsvelt, and
N. P. Heydenburg, Angular correlation
studies for a particle excitation of the 1.78
MeV 2+ state in 28Si, Nucl. Phys., A191, 137-
144, 1972.
Akimoto, S., see Shimizu, N.
Aldrich, L. T.. see Ocola, L. C.
Aparicio, P., see Sacks, I. S.
Arnold, L. G., see Brown, L.
Assousa, G. E., and W. H. Smith, Radiative
lifetimes for some resonance transitions of
Fe I and Fe II in the region between 2300 A
and 3050 A, and the application to iron abun-
dance determinations in the sun and in the
QSO PHL 938, Astrophys. J., 176, 259-264,
1972.
Assousa, G. E., see also Ahlfeld, C. E., Brown,
L., and Kumar, C. K.
Benson, W. E., see Wright, T. L.
Bonner, T. I., see Straus, N. A.
Brooks, C, and S. R. Hart, An extrusive ba-
saltic komatiite from a Canadian Archean
metavolcanic belt, Can. J . Earth Sci., 9, 1250-
1253, 1972.
Brooks, C, and S. R. Hart, Pyroxenes from
Archean mafic rocks: trace element and
Sr87Sr86 data (abstract), Trans. Amer. Geo-
phys. Union, 54, 500, 1973.
Brooks, C, see also James, D. E.
Brown, L., G. E. Assousa, and H. A. Van Rins-
velt, Excitation of the K lines of copper by
alkali ions in the energy range of 1 to 5 MeV
(abstract), Bull. Amer. Phys. Soc, IS, 103,
1973.
Brown, L., E. Steiner, L. G. Arnold, and R. G.
Seyler, Polarization and phase shifts in
7Li(p,p)7Li from 0.4 to 2.5 MeV and the
structure of 8Be, Nucl. Phys., A206, 353-373,
1973.
Brown, L., see also Kumar, C. K., and Rohrer,
U.
Chattopadhyay, S. K., D. E. Kohne, and S. K.
Dutta, Ribosomal RNA genes of Neuro-
spora : Isolation and characterization, Proc.
Nat. Acad. Sci., USA, 69, 3256-3259, 1972.
D'Odorico, S., Vera C. Rubin, and W. K. Ford,
Jr., Line intensities and radial velocities for
12 planetary nebulae, Astron. Astrophys., 22,
469-472, 1973.
Dutta, S. K., see Chattopadhyay, S. K.
Fletcher, R. C, Folding and necking of a layer
with power law rheology (abstract), Geol.
Soc. Amer. Abstracts with Programs, 85th
Ann. Meeting, Minneapolis, Minn., p. 506,
Nov. 13-15, 1972.
Fletcher, R. C, and R. H. Groshong, Jr., Rock
rheology from wavelength selection in folding
(abstract), Trans. Amer. Geophys. Union, 54,
456, 1973.
Ford, W. K., Jr., see D'Odorico, S., Kumar,
C. K., and Rubin, Vera C.
Goodman, M., see Hoyer, B. H.
Grauert, B., New LT-Pb isotopic analyses of zir-
cons from the Baltimore Gneiss and the Setters
Formation, Maryland Piedmont (abstract),
Geol. Soc. Amer. Abstracts with Programs,
85th Ann. Meeting, Minneapolis, Minn., p.
519, Nov. 13-15, 1972.
Grauert, B., and M. G. Seitz, Uranium gain of
detrital zircons studied by isotopic anatyses
and fission track mapping (abstract), Trans.
Amer. Geophys. Union, 54, 495, 1973.
DEPARTMENT OF TERRESTRIAL MAGNETISM
313
Grauert, B., R. Hanny, and G. Soptrajanova,
Age and origin of detrital zircons from the
Pre-Permian basements of the Bohemian
massif and the Alps, Contrib. Mineral.
Petrol, 40, 105-130, 1973.
Groshong, R. H., Jr., see Fletcher, R. C.
Hanny, R., see Grauert, B.
Hart, S. R., D. E. James, and T. E. Krogh,
eds., Plate tectonics and evolution of conti-
nents, Geotimes, 17, 18-19, 1972.
Hart, S. R., and J. G. Schilling, Basalts from
Iceland and south along the mid-Atlantic
ridge: trace element and Sr87/Sr86 data (ab-
stract), Trans. Amer. Geophys. Union,, 54,
484, 1973.
Hart, S. R., see also Brooks, C., Melson, W. G.,
and Wright T. L.
Heydenburg, N. P, see Ahlfeld, C. E.
Hoyer, B. H., Neltje W. van de Velde, M.
Goodman, and R. B. Roberts, Examination
of hominid evolution by DNA sequences
homology, J. Human Evolution, 1, 645-649,
1972.
James, D. E., Crustral and upper mantle struc-
ture beneath west-central South America, in
Symposium on the Results of Upper Mantle
Investigations with Emphasis on Latin Amer-
ica, Vol. II, pp. 193-204, Comite Argentino
del Manto Superior, Buenos Aires, 1972.
James, D. E., Plate tectonic model for the
evolution of the central Andes: Reply, Geol.
Soc. Amer. Bull., 84, 1497-1500, 1973.
James, D. E., and C. Brooks. Sr87/Sr86 ratios in
volcanic and plutonic rocks of the central
Andes (abstract), Trans. Amer. Geophys.
Union, 54, 501, 1973.
James, D. E., see also Hart, S. R.
Kohne, D. E., see Chattopadhyay, S. K.
Krogh, T. E.. see Hart, S. R.
Kumar, C. K., G. E. Assousa, L. Brown, and
W. K. Ford, Jr., Radiative mean lifetimes of
levels in K II and Rb II, Phys. Rev. A, 7,
112-117, 1973.
Kumar, C. K., see also Rubin, Vera C.
LaSalle, R. A, see Ahlfeld, C. E.
Linde, A. T., and I. S. Sacks, Dimensions, en-
ergy, and stress release for South American
deep earthquakes, in Symposium on the
Results oj Upper Mantle Investigations with
Emphasis on Latin America, Vol. II, pp. 211—
227, Comite Argentino del Manto Superior,
Buenos Aires, 1972.
Lundsager, S., see Tuve, M. A.
Melson, W. G., S. R. Hart, and G. Thompson,
St. Paul's Rocks, equatorial Atlantic: pedo-
genesis, radiometric ages, and implications
on sea-floor spreading, in Studies in Earth
and Space Sciences, edited by R. Shagam
et al., pp. 241-272, Geol. Soc. Amer., Mem.
132, Boulder, Col., 1972.
Melson, W. G., see also Wright, T. L.
Meyer, R. P., see Ocola, L. C.
Mirabel, I. F., and K. C. Turner, A search for
neutral hydrogen remnants of strong tidal
disruption of the Small Magellanic Cloud.
Astron. Astrophys., 22, 437-440, 1973.
Ocola, L. C, R. P. Meyer, and L. T. Aldrich,
Gross crustal structure under the Peru-
Bolivia altiplano from 1968 explosion studies,
in Symposium on the Results of Upper
Mantle Investigations with Emphasis on
Latin America, Vol. II, pp. 233-243, Comite
Argentino del Manto Superior, Buenos Aires,
1972.
Okada, H., Conversion from ScS to ScSp at the
inclined deep seismic zone (abstract), Trans.
Amer. Geophys. Union, 54, 365, 1973.
Okada, H., see also Sacks, I. S.
Roberts, R. B., see Hoyer, B. H.
Rohrer, U., and L. Brown, Scattering of polar-
ized protons by beryllium from 0.9 to 2.7
MeV (abstract), Bull. Amer. Phys. Soc, 17,
923, 1972.
Rubin, Vera C, The dynamics of the Andro-
meda Nebula, Science, 228, 30-36, 1973.
Rubin. Vera C, W. K. Ford, Jr., and C. K.
Kumar, Stellar motions near the nucleus of
M31, Astrophys, J., 181, 61-77, 1973.
Rubin, Vera C, C. K. Kumar, and W. K. Ford,
Jr., Variation of emission-line strengths across
M31, Astrophys. J., 177, 31-44. 1972.
Rubin, Vera C, see also D'Odorico, S.
Saa, G., S.J., see Sacks, I. S.
Sacks, I. S., Geographic distribution of absorp-
tion of shear waves in South America and its
tectonic significance, in Symposium on the
Results of Upper Mantle Investigations with
Emphasis on Latin America, Vol. II, pp. 245-
253, Comite Argentino del Manto Superior.
Buenos Aires, 1972.
Sacks, I. S., and H. Okada, A comparison of the
Q structure beneath South America and
Japan (abstract), I'rans. Amer. Geophys.
Union, 54, 364, 1973.
Sacks, I. S., G. Saa, S.J.. and P. Apancio. P
travel-time anomalies and their structural
implications in west-central South America,
in Symposium on the Results of Upper Man-
tle Investigations with Etnphasis on Latin
America, Vol. II, pp. 255-264, Comite Argen-
tino del Manto Superior, Buenos Aires. 1972.
Sacks, I. S., see also Linde, A. T.
Schilling, J. G., see Hart, S. R.
Seitz, M. G., see Grauert, B.
314
CARNEGIE INSTITUTION
Seyler, R. G., see Brown, L.
Shimizu, N., Geochemistry of spinel lherzolites
from Salt Lake Crater (abstract), Geol. Soc.
Amer. Abstracts with Programs, 85th Ann.
Meeting, Minneapolis, Minnesota, p. 663,
Nov. 13-15, 1972.
Shimizu, N., Trace element comparison of
lherzolite and "eclogite" inclusions from Salt
Lake Crater, Hawaii (abstract), Trans. Amer.
Geophys. Union, 54, 501, 1973.
Shimizu, N., and S. Akimoto, Partitioning of
strontium between clinopyroxene and liquids
at high pressures: preliminary experiments,
Earth Planet. Sci. Lett., 13, 134-138, 1971.
Sinha, A. K., U-Th-Pb systematics and the age
of the Onverwacht Series, South Africa,
Earth Planet. Sci. Lett., 16, 219-227, 1972.
Smith, W. H., see Assousa, G. E.
Soptrajanova, G., see Grauert, B.
Steiner, E., see Brown, L.
Straus, N. A., and T. I. Bonner, Temperature
dependence of RNA-DNA hybridization
kinetics, Biochim. Biophys. Acta, 277, 87-95,
1972.
Thompson, G., see Melson, W. G.
Thompson, W. J., see Ahlfeld, C. E.
Turner, K. C, see Mirabel, I. F.
Tuve, M. A., and S. Lundsager, Velocity struc-
tures in hydrogen profiles, Astron. J., 77, 652-
660, 1972. *
Van Rinsvelt, H. A., see Ahlfeld, C. E., and
Brown, L.
Velde, Neltje W. van de, see Hoyer, B. H.
Wright, T. L., W. E. Benson, W. G. Melson,
and S. R. Hart, Petrology of basaltic rocks
collected from Leg 14, in Initial Reports of
the Deep Sea Drilling Project, Vol. XIV,
767-772, D. E. Hayes, A. C. Pimm et al.,
Washington, D.C., U.S. Government Printing
Office, 1972.
PERSONNEL
Director
Ellis T. Bolton
Associate Director
L. Thomas Aldrich
Staff Members
George E. Assousa
Roy J. Britten 1
Louis Brown
Dean B. Cowie
W. Kent Ford, Jr.2
Stanley R. Hart
Albrecht W. Hofmann
Bill H. Hoyer
Kenneth C
David E. James
Alan T. Linde
Nancy R. Rice
Richard B. Roberts
Vera C. Rubin
I. Selwyn Sacks
Diglio Simoni
Norbert Thonnard
Turner 3
Distinguished Service Member of Carnegie Institution
M. A. Tuve
Astrophysics: Louis Brown4
W. K. Ford, Jr.5
Section Chairmen
Biophysics: Dean B. Cowie
Geophysics: L. Thomas Aldrich
1 Visiting Associate in Biology, California In-
stitute of Technology.
"Visiting Scientist, Kitt Peak National Ob-
servatory, from January 1, 1973.
3 On leave from DTM as Director, Instituto
Argentino de Radioastronomia, through June
30, 1973.
4 From January 1, 1973.
5 To December 31, 1972.
DEPARTMENT OF TERRESTRIAL MAGNETISM
315
Research Associates
John Bannister, Universidad de Chile, San-
tiago, Chile
Mateo Casaverde, Instituto Geofisico del
Peru, Lima, Peru
Ruben Dugatkin, Instituto Argentino de Ra-
dioastronomia, Villa Elisa, Argentina
(through September 30, 1972)
Emilio Filloy, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
Leonidas Ocola, Instituto Geofisico del Peru,
Lima, Peru
Reynaldo Salgueiro, Instituto Geofisico Bo-
liviano, La Paz, Bolivia
T. Jefferson Smith, Kalamazoo College, Kala-
mazoo, Michigan (from September 1, 1972)
Shigeji Suyehiro, Japan Meterological Agency,
Tokyo, Japan
Fellows
Jesus A. Berrocal, Instituto Geofisico del
Peru, Lima, Peru, from September 1, 1972
Tom I. Bonner, University of Texas, Austin,
Texas; Postdoctoral Research Associate
under Public Health Service grant from
January 1, 1973
Michael B. Davis, Princeton University,
Princeton, New Jersey, from November 1,
1972
Borwin Grauert, Institut fur Kristallographie
unci Petrographie, Eidgenossiche Technische
Hochschule, Zurich, Switzerland; Fellow of
North Atlantic Treaty Organization
through March 31, 1973 ; Fellow of Carnegie
Institution of Washington from April 1,
1973
Cidambi K. Kumar, University of Michigan,
Ann Arbor, Michigan, through January 31,
1973
Allan T. Leffler II, M.D., University of Mary-
land School of Medicine, Baltimore, Mary-
land; Special Associate under Public
Health Service grant
Hiromu Okada, Hokkaido University, Sap-
poro, Japan; Predoctoral Research Asso-
ciate under National Science Foundation
grant from January 1, 1973
Urs C. Rohrer, University of Basel, Basel,
Switzerland
Nobumichi Shimizu, University of Tokyo,
Tokyo, Japan; Postdoctoral Research As-
sociate under National Science Founda-
tion grant through August 31, 1972, Fel-
low of Carnegie Institution of Washington
from September 1, 1972
J. Arthur Snoke, Middle East Technical Uni-
versity, Ankara, Turkey, from September
1, 1972
Neil A. Straus, University of Toronto, To-
ronto, Canada; Fellow of the Canadian
Medical Research Council through Sep-
tember 15, 1972
Thangasamy Velusamy, University of Mary-
land, College Park, Maryland, from May
15, 1973
Donald G. Wallace, University of Durham,
Durham, England; Postdoctoral Research
Associate under National Science Founda-
tion grant from December 11, 1972
John W. Warner III, Ohio State University,
Columbus, Ohio, from October 1, 1972
Trainee Fellows
Arturo Cuyubamba, Servicio de Geologia y
Mineria, Lima, Peru
Douglas M. Haefele, George Washington Uni-
versity, Washington, D.C.
Godwin C. Igiri, Howard University, Wash-
ington, D.C.
Jack M. Klitzman, LTniversity of Michigan,
Ann Arbor, Michigan
Ricardo Quiroga, Cuyo University, San Juan,
Argentina
Enrique Triep, University of Buenos Aires,
Buenos Aires. Argentina
Collaborators and Visiting Investigators
S. Alexander, Pennsylvania State University,
University Park, Pennsylvania
C. A. Andersen, Applied Research Labora-
tories, Goleta, California
E. Arnal, Instituto Argentino de Radioastron-
omia, Villa Elisa, Argentina
L. G. Arnold, Ohio State University, Colum-
bus, Ohio
316
CARNEGIE INSTITUTION
B. Balick, National Radio Astronomy Ob-
servatory, Charlottesville, Virginia
F. R. Boyd, Geophysical Laboratory, Car-
negie Institution of Washington, Washing-
ton, D.C.
D. J. Brenner, Walter Reed Army Institute
of Research, Washington, D.C.
C. Brooks, Universite de Montreal, Montreal,
Canada
M. J. Byers, Purdue University, Lafayette,
Indiana
R. Cabre, S.J., Observatorio San Calixto,
La Paz, Bolivia
A. Cerna, Instituto Geofisico del Peru, Lima,
Peru
S. P. Champe, Rutgers University, New
Brunswick, New Jersey
J. A. Chiscon, Purdue Univeristy, Lafayette,
Indiana
G. N. Cohen, Institut Pasteur, Paris, France
R. Colomb, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
D. Comaford, Applied Research Laboratories,
Sunland, California
M. L. Crawford, Bryn Mawr, Pennsylvania
E. H. Davidson, California Institute of Tech-
nology, Pasadena, California
G. L. Davis, Geophysical Laboratory, Car-
negie Institution of Washington, Washing-
ton, D.C.
S. del Pozo, Instituto Geofisico Boliviano, La
Paz, Bolivia
W. C. Erickson, University of Maryland, Col-
lege Park, Maryland
J. W. Erkes, State University of New York,
Albany, New York
A. J. Erlank, University of Capetown, Ron-
debosch, South Africa
I. Ermanovics, Canadian Geological Survey,
Ottawa, Canada
D. W. Evertson, University of Texas at
Austin, Austin, Texas
R. C. Fletcher, City College of New York,
New York
J. B. Flexner, University of Pennsylvania,
Philadelphia, Pennsylvania
L. B. Flexner, University of Pennsylvania,
Philadelphia, Pennsylvania
A. Flores, Instituto Geofisico del Peru, Lima,
Peru
S. E. Forbush, Department of Terrestrial
Magnetism, Carnegie Institution of Wash-
ington, Washington, D.C. (retired)
S. Garzoli, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
J. Gettrust, Instituto Geofisico del Peru,
Lima, Peru
A. A. Giesecke, Instituto Geofisico del Peru,
Lima, Peru
M. Gil, Instituto Argentino de Radioastron-
omia, Villa Elisa, Argentina
B. J. Giletti, Brown University, Providence,
Rhode Island
Jean Goad, Harvard College Observatory,
Cambridge, Massachusetts
D. Goniadzki, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
M. Goodman, Wayne University, Detroit,
Michigan
M. Gordon, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
J. A. Graham, Cerro Tololo Inter- American
Observatory, La Serena, Chile
R. H. Groshong, Syracuse University, Syra-
cuse, New York
L. Hall, University of Massachusetts, Am-
herst, Massachusetts
J. R. Hinthorne, Applied Research Labora-
tories, Goleta, California
D. Huaco, Instituto Geofisico del Peru, Lima,
Peru
T. E. Krogh, Geophysical Laboratory, Car-
negie Institution of Washington, Washing-
ton, D.C.
I. Kushiro, Geophysical Laboratory, Carnegie
Institution of Washington, Washington,
D.C.
B. T. R. Lewis, University of Washington,
Seattle, Washington
C. B. Lister, University of Washington,
Seattle, Washington
J. W. Mahoney, Jr., Woods Hole Oceano-
graphic Institute, Woods Hole, Massachu-
setts
M. A. Martin, National Institutes of Health,
Bethesda, Maryland
J. M. Mattinson, Geophysical Laboratory,
Carnegie Institution of Washington, Wash-
ington, D.C.
I. Mirabel, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
R. Morras, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
Y. Motoya, Kamikineusu Seismological Ob-
servatory of Sapporo University, Hokkaido,
Japan
B. R. Neufeld, California Institute of Tech-
nology, Pasadena, California
G. Olafsson, Akureyri, Iceland
DEPARTMENT OF TERRESTRIAL MAGNETISM
317
R. B. Partridge, Haverford College, Haver-
ford, Pennsylvania
W. Poppel, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
R. Quiroga, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
A. Rake, Pennsylvania State University, Uni-
versity Park, Pennsylvania
Fr. J. E. Ramirez, Instituto Geofisico de los
Andes Colombianos, Bogota, Colombia
M. S. Roberts, National Radio Astronomy
Observatory, Charlottesville, Virginia
A. Rodriguez, Universidad Nacional de San
Agustin, Arequipa, Peru
S. Rubin, Radcliffe College, Cambridge,
Massachusetts
J. G. Schilling, Narragansett Marine Labora-
tory, University of Rhode Island, Kingston,
Rhode Island
M. G. Seitz, Geophysical Laboratory, Car-
negie Institution of Washington, Washing-
ton, D.C.
R. G. Sevier, Ohio State University, Colum-
bus, Ohio
H. Sigtryggsson, Reykjavik, Iceland
F. Strauss, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
Z. Strauss, Instituto Argentino de Radio-
astronomia, Villa Elisa, Argentina
L. Tamayo, Universidad Nacional de San
Agustin, Arequipa, Peru
J. Tomblin, University of West Indies, Trin-
idad, West Indies
P. Truffa-Bachi, Institut Pasteur, Paris,
France
H. A. Van Rinsvelt, University of Florida,
Gainesville, Florida
V. Volponi, Universidad Nacional de Cuyo,
San Juan, Argentina
M. E. Wagner, University of Pennsylvania,
Philadelphia, Pennsylvania
R. F. Wing, Ohio State University, Colum-
bus, Ohio
Y. Yamagishi, Matsushiro Seismological Ob-
servatory of Japan, Meteorological Agency,
Matsushiro, Japan
Design Engineer
Everett T. Ecklund
Charles A. Little
Electronic Research Specialists
John B. Doak
Glenn R. Poe
Computer Systems Engineer
Kenneth D. Burrhus
Design Engineer
Michael Seemann (from September 25, 1972)
Research Assistants
Liselotte Beach
Patricia P. Esposito
Laura A. Gazze (from October 2, 1972)
Mary L. Kelly (through July 15, 1972)
Neltje W. van de Velde
Office
Chief, Fiscal Section: Helen E. Russell (re-
tired June 30, 1973)
Office Manager: William N. Dove
Assistant Fiscal Officer: Niels M. Pedersen
Secretary: E. Kathleen Hill
Stenographer: Dorothy B. Dillin
Clerk-typist: Thyra R. Jackson (through
September 8, 1972), Grethe J. Smith (Sep-
tember 11, 1972-February 26, 1973)
318
CARNEGIE INSTITUTION
Shop
Shop Manager, Electronic Research Special- Machinist: Francis J. Caherty (through Jan-
ist: Paul A. Johnson uary 31, 1973)
Instrument Makers: Carl M. Rinehart, Mil-
ton T. Taylor
Buildings and Grounds
Carpenter and Maintenance Foreman: Leo J. Maintenance Assistant: Stanley Gawrys
Haber Caretakers: Bennie Harris, Willis Kilgore, Jr.
Assistant Maintenance Foreman: Elliott M.
Quade
Michael J. Cooley
Thomas A. Dorian
Peter J. Fiekowsky
Fariborz Ghahremani
Part-Time and Temporary Employees
James R. Kelly, Jr.
Jack M. Klitzman
Joseph Scuderi
Department of Plant Biology
Stanford, California
C. Stacy French
Director
Contents
Introduction (French) 321
Reversible inhibition of energy transfer between photosynthetic pigments by hydro-
static pressure (Schreiber) 327
Distribution of photoconvertible, water-soluble chlorophyll protein complex CP668
in plants related to Chenopodium album (Takamiya) 330
Resolution of the low temperature absorption spectra of chlorophyll protein com-
plexes into their component bands (French, Takamiya, and Murata) . . . 336
A comparative study of the forms of chlorophyll and photochemical activity of sys-
tem 1 and system 2 fractions from spinach and Dunaliella (Brown, Gasanov,
and French) 351
The effect of manganese on DCIP reduction (Brown) 359
Rate measuring circuit for improved action spectra (Hagar, Ford, and French) . 361
Effect of blue light on respiration. Action spectrum for the photoinhibition of blue
light (Ninnemann and French) 365
Some recent improvements of a high-speed difference spectrophotometer (Hiyama
and Fork) 368
Computer resolution of complex kinetics (Hiyama) 371
Light-induced shifts in the absorption spectrum of carotenoids and chlorophyll b in
the green alga Ulva (Fork) 374
The function of plastocyanin in electron transport of photosynthesis. The content
of plastocyanin in chloroplast particles prepared by different techniques
(Murata and Fork) 376
The photochemical reactions of photosynthesis in an alga exposed to extreme con-
ditions (Fork and Hiyama) 384
Probit theory and applications in the analysis of synchronized growth systems
(Hagar) 388
Physiological adaptations to diverse environments: Approaches and facilities to
study plant response to contrasting thermal and water regimes (Bjorkman,
Nobs, Berry, Mooney, Nicholson, and Catanzaro) 393
Removal of contaminant inorganic phosphate and phosphoglycolate from ribulose-
1,5-diphosphate (Berry and Bowes) 403
Oxygen uptake in vitro by RuDP carboxylase of Chlamydomonas reinhardtii
(Berry and Bowes) 405
Bibliographic information retrieval (Brown) 407
Staff activities 410
Bibliography 411
Speeches 412
Personnel 414
Carnegie Institution of Washington Year Book 72, 1972-1973
INTRODUCTION
This year the Department enters into
a very promising period of reconstruction
and expansion under Dr. Winslow Briggs,
who comes to the Institution from Har-
vard University. In this twenty-sixth
and final report of the retiring Director
it is appropriate to make a few comments
on the history of botanical research in
the Institution.
In 1905 Dr. Daniel T. MacDougal be-
came Director of the newly organized
Desert Laboratory in Tucson, Arizona.
Since at that time the means of plant
survival and growth under desert condi-
tions were major mysteries, not only the
extraordinary structures and physiology
of the desert plants were studied but also
much attention was given to their phys-
ical environment. The topography and
climatology of the Sonoran Desert were
then as little known as were its plant
inhabitants. An important part of the
work of the Desert Laboratory was to
follow the individual growth of plants
over long periods of years. In the fenced
area surrounding the laboratory, one of
the few places where desert vegetation
has not been greatly altered by cattle
grazing, individual cacti such as the spec-
tacular Sahuaro, Carnegia gigantea, have
been measured and photographed at in-
tervals since the early part of the cen-
tury. This work still continues, although
no longer within the Institution. A his-
tory of the Desert Laboratory is now
being prepared for the Institution by Dr.
William G. McGinnies of Tucson.
Later at the Coastal Laboratory in
Carmel, California, MacDougal and his
colleagues investigated other aspects of
plant physiology including the daily
variation in the trunk diameter of trees.
In 1910 Dr. Herman A. Spoehr, fresh
from postdoctoral study in Emil Fischer's
laboratory of carbohydrate chemistry in
Berlin, joined the Desert Laboratory to
study photosynthesis. It is a matter of
some historical interest that during
Spoehr's time with Emil Fischer his post-
doctoral colleagues were Herman 0. L.
Fischer and Otto Warburg, who devel-
oped the study of photosynthesis in
Europe as Spoehr did in this country.
One of Spoehr's concerns was the daily
variation in organic acid content of fleshy
desert plants and the corresponding dark
uptake of C02 and its release in light, an
effect opposite to that in normal leaves.
This phenomenon, "Crassulacean Acid
Metabolism," known since 1819, remains
a subject of much current interest. From
1910 until Robert Emerson returned from
Warburg's laboratory in Berlin to Har-
vard in 1927 Spoehr remained essentially
the only investigator in this country with
a lifetime commitment to studies on the
mechanism of photosynthesis. In the
1930's decade, research on photosynthesis
expanded rapidly with the influx of Euro-
pean trained investigators. Spoehr sum-
marized the existing knowledge of photo-
synthesis in a monograph published in
1926.
The known and fancied lore relating
to fixation of atmospheric C02 by plants
was again thoroughly summarized in
Rabinowitch's treatises. These monu-
mental volumes appearing in 1945, 1951,
and 1956 emphasized the physical aspects
of the subject and brought up to date
the monograph of Spoehr which had
served as the definitive text on the sub-
ject for twenty years following its pub-
lication.
In 1927 Dr. Spoehr became Chairman
of the Institution's Division of Plant
Biology and in 1929 established the pres-
ent laboratory on the campus of Stanford
University. This was the administrative
"Central Laboratory" for the Institu-
tion's botanical research in progress at
Tucson, Carmel, Santa Barbara, Pikes
Peak, and the transplant gardens at
Mather and at the Harvey Monroe Hall
321
322
CARNEGIE INSTITUTION
Natural Area in the Sierra Nevada —
both established in 1926 by Dr. Hall.
The laboratory has remained a focal
point for research on photosynthesis and
on the relationships of plants to each
other and to their environment. The
work of Dr. and Mrs. Clements and their
colleagues outlined the basic concepts of
plant ecology. They developed a com-
plex terminology which may have tempo-
rarily clouded the importance of their
discoveries about plant succession which
are so basic to present-day ecology.
Dr. Jens Clausen came to the new lab-
oratory in 1929 from Denmark and after
Dr. Hall's death led the Experimental
Taxonomy group through prolonged re-
search on the means by which plants
achieve their wide variation in size, form,
distribution, and physiological adapta-
tions. This work documented compre-
hensive studies of the cooperative influ-
ences of heredity and of environment on
the morphology and physiology of plants.
The concept of climatic races and of
genetic coherence emerged from these
experiments of Clausen, Keck, Hiesey,
and Nobs under Clausen and later under
Hiesey's guidance.
To fit more closely the nature of its
present work, that group of the De-
partment, led by Dr. Olle Bjorkman,
now defines its field as Physiological
Ecology. This year their major efforts
have been devoted to construction of new
facilities with a joint National Science
Foundation grant to CIW and Stanford
University. New experimental gardens
have been established in the contrasting
climates at Death Valley and at the Uni-
versity of California's Marine Biological
Station at Bodega Bay.
Since 1947, when Dr. Spoehr retired as
Chairman of the Department, there have
been drastic changes in the understand-
ing of the process of photosynthesis, and
the number of people in the field has
increased greatly throughout the world.
Progress has been made in many differ-
ent aspects of the subject. The bio-
chemical pathways of atmospheric car-
bon dioxide incorporation into organic
matter have been clarified largely
through the work of Melvin Calvin and
his collaborators at the University of
California at Berkeley. During the past
quarter of a century, the idea that the
mechanism of photosynthesis could be
deduced solely from kinetic measure-
ments of such effects as oxygen evolution,
fluorescence intensity changes, or rates of
carbon dioxide uptake has slowly been
abandoned. In place of that procedure,
which still is profitably applied to
simpler photochemical reactions, the
breakdown of the overall photosynthetic
process into its individual chemical steps
has been found to yield more definitive
information. Preparative isolation bio-
chemistry has shown the presence of a
number of highly reactive compounds
such as cytochromes, plastoquinones,
high-energy phosphates, pyridine nucleo-
tides, plastocyanin, and non-heme iron
compounds in the photosynthetic system.
Current biochemical work centers on the
interactions of such substances when
light is absorbed by the chlorophyll com-
plexes of plant chloroplasts. Kinetic
measurements now usually refer specific-
ally to changes in known intermediate
substances rather than to hypothetical
components. Many of these known inter-
mediate electron carriers of the photo-
synthetic system change color with their
oxidation state so their participation in
photosynthesis can be followed by meas-
uring their absorption changes at appro-
priate wavelengths. This year Drs. Fork
and Hiyama have improved their equip-
ment in order to measure such changes at
shorter time intervals.
In 1947 the laboratory was deeply
involved in various aspects of photo-
synthesis research. Dr. Harold Strain
was identifying plant carotenoids by
column chromatography, an art that
had largely lain dormant since its inven-
tion by Tswett. Dr. James H. C. Smith
was isolating the phototransformable
DEPARTMENT OF PLANT BIOLOGY
323
protochlorophyll protein complex which
he had named protochlorophyll holo-
chrome. Dr. Spoehr and Mr. Harold
Milner were growing Chlorella on a large
laboratory scale to study the practicality
of its use for food. That work for some
years thereafter remained a major enter-
prise of the Department and led to the
Institution's best seller, Publication 600
— Algal Culture from Laboratory to
Pilot Plant. Algal culture for food use
has been studied in many countries, par-
ticularly Japan by Professor Hiro-
shi Tamiya and his collaborators, with
whom the Department has continuously
maintained close relations. However, the
cost of algal culture and its conversion to
edible form remain a barrier to its wide-
spread use.
Experiments initiated in 1947 on the
separation of plant chloroplasts into their
functional components remain today a
major part of the scene in photosynthetic
research. At that time the newly devel-
oped dye reduction reaction (a type of
Hill reaction) was used as a test for
chloroplast activity. When the work of
Blinks, at the Hopkins Marine Station,
on chromatic adaptation and that on
enhancement of Emerson and Lewis, who
had previously worked at this Depart-
ment, suggested the existence of two sep-
arate photochemical systems in photo-
synthesis, the present concept of the
process began to develop.
Much of the attention of the photosyn-
thesis group for the past quarter of a
century has centered on the spectroscopy
of photosynthetic pigments, primarily
chlorophyll complexes. For these studies
various instruments and techniques have
been developed at the laboratory.
A few of the results obtained this year
may be summarized as follows:
The accessory pigments of red algae
— phycoerythrin and phycocyanin —
function in photosynthesis by transfer-
ring their absorbed light energy to
chlorophyll. This year Dr. Ulrich
Schreiber of Aachen, Germany, now a
Research Fellow with Professor William
Vidaver (a former Research Fellow of
the Department, now at Simon Fraser
University) , spent part of the summer at
the Department. He found that hydro-
static pressure greatly reduces the effi-
ciency of energy transfer from the phyco-
bilin pigments to chlorophyll in red algae.
Professor Atusi Takamiya, former
head of the Department of Biochemistry
and Biophysics at the University of
Tokyo, spent five months at the Depart-
ment on leave from Toho University. He
found that a water-soluble and photo-
transformable chlorophyll protein is far
more widely distributed in wild plants
than had previously been realized. Its
presence in many species of Atriplex was
discovered. Accurate absorption spectra
of several chlorophyll proteins were
measured both at 24° and — 196°C. The
spectra of chlorophyll protein from Che-
nopodium. album and Brassica nigra with
peaks at 669 and 673 nm could be fitted
with the components of ordinary green
plant chloroplasts at 650, 662, 670, and
677 nm. The preparation from Lepidium
virginicum peaking at 658 nm, however,
required unusual components with peak
wavelengths of 659 and 667 nm. Because
these preparations lack carotenoids it
was also possible to analyze the blue part
of the chlorophyll protein spectrum.
Dr. Ralphreed Gasanov, Assistant
Professor of Plant Physiology at Baku,
Azerbaijan, USSR, visited the laboratory
for eight months. He found that the
photosystem fraction 1 isolated from
chloroplast grana by digitonin differs in
its pigment composition from the corre-
sponding material in the chloroplast
stroma particles located between the
grana stacks.
Drs. Brown and Gasanov are attempt-
ing to understand the state of chlorophyll
in vivo and how this essential pigment
functions in photosynthesis with the hope
that eventually a more direct use of solar
energy may be possible. They have ap-
proached this problem by dividing
324
CARNEGIE INSTITUTION
chloroplasts into the smallest units that
still retain photochemical activity. They
applied successfully an improved and ex-
tended fractionation procedure to an alga
that has unusually distinct absorption
bands of chlorophyll a. They were able
to obtain fractions of Dunaliella that
had activity only for photosystem 1 or
only for photosystem 2, the reducing and
oxidizing light reactions of complete
photosynthesis.
Curve analyses of the absorption spec-
tra of these fractions revealed the rela-
tive proportions of the forms of chloro-
phyll present in each photosystem.
Chlorophyll a-694 is lacking in the
photosystem 2 fraction and present in all
of the fractions with photosystem 1 ac-
tivity. The longer wavelength forms of
chlorophyll (>700 nm) seen in spinach
chloroplast particles are not found in this
alga, suggesting that these forms are not
essential for photosynthesis but may
function as additional light harvesting
pigment.
Research Fellow Dr. William Hagar,
Glenn Ford, and C. S. French have de-
veloped and tested an electronic circuit
for measuring the rates of slow reactions
normally obtained with action spectrum
measurements. This circuit, when trig-
gered, continuously produces a signal
of the voltage change per unit of time
until it is switched off. Thus the signal-
to-noise ratio keeps increasing as the
sampled increment becomes larger and
larger with time. The output of this rate
measuring circuit is monitored until it
reaches a steady value signifying the
sampling of a large enough increment to
avoid noise errors.
Dr. Hagar together with Professor T.
Punnett of Temple University have pro-
posed the use of probit transformation of
cell number data to define the timing
and synchrony of cell cultures. Probit
transformation converts a normal sig-
moidal curve into a straight line whose
slope is the reciprocal of the standard
deviation of the cell age spread, and the
abscissa value for probit of 5 is the mean
time of cell division. This technique of
analysis should be useful for describing
the degree of synchrony in many fields
of biology where either naturally or arti-
ficially synchronized systems are com-
monly found.
During his return visit to the labora-
tory last summer, Dr. Murata from the
Department of Biochemistry and Bio-
physics of the University of Tokyo col-
laborated with Dr. Fork on a continua-
tion of their study on the function of the
copper protein plastocyanin in photo-
synthesis. Their previous work, reported
in Year Book 70, showed that little
plastocyanin remained in particles pre-
pared by extrusion through a needle
valve. These particles nevertheless
showed rapid electron transfer reactions
from cytochrome / to the reaction center
of system 1 (P700) — as fast as those seen
in chloroplasts in their native state.
These observations suggested that cyto-
chrome / functions as the "secondary
electron donor" to P700, not plasto-
cyanin as is currently envisaged. How-
ever, the lack of plastocyanin in these
particles was questioned by some work-
ers, particularly by those who used the
so-called bioassay method to determine
plastocyanin. Drs. Murata and Fork
found that while the bioassay method is
highly sensitive for plastocyanin, it is
not necessarily specific for this substance.
After doing extensive studies on the con-
tents of plastocyanin in chloroplasts and
in subchloroplast particles prepared by
a variety of methods to release plasto-
cyanin, they confirmed their earlier
conclusions that plastocyanin is absent
from particles derived by needle valve
extrusion and that it does not function
between cytochrome / and P700. It would
seem from these results that cytochrome
/ plays the role of secondary electron
donor to P700 as described above. It is
still possible, however, that cytochrome /
and plastocyanin may function in paral-
lel to reduce P700.
DEPARTMENT OF PLANT BIOLOGY 325
Dr. George Bowes, a postdoctoral fel- to the Department for periods of a few
low, and Dr. Berry have continued their months to several years. In addition to
work on photosynthesis and glycolate these Research Fellows, we have had the
formation by the green alga Chlamydo- added stimulation of frequent working
monas. This year they have been at- visits from well-established scientists
tempting to compare the rate of forma- from other organizations, who have par-
tion of glycolate by an enzymatic mecha- ticipated actively in the educational
nism in vitro to the rate of glycolate for- work of the Institution,
mation by the alga in vivo. For this pur- To the Staff, Visiting Investigators,
pose a great deal of effort has been di- Research Fellows, and some other scien-
rected toward developing an unequivocal tists who have participated in the work
chemical assay for glycolate formation, of the laboratory, and to the supporting
Measurement of oxygen uptake has staff who have made the execution of
proven the best system so far tested, but their ideas possible, belongs a significant
it is not ideal. Using this system, Dr. part of the credit for the development of
Bowes and Dr. Berry report rates of the present understanding of photosyn-
glycolate formation in vitro that are thesis in plants and of the relationship of
nearly sufficient to account for glycolate plant families to each other and to their
formation in vivo. This work is being environment.
continued by Dr. Berry, and it is anti- The Desert Laboratory opened in
cipated that final results will be pub- 1903 and became headquarters of the
lished during the coming year. Department of Botanical Research in
Dr. Helga Ninnemann, a Research 1905. There were three name changes in
Fellow from Tubingen, measured the ac- later years: Laboratory for Plant Physi-
tion spectrum for the inactivation of ology in 1923; Division of Plant Biology,
yeast respiration. While the peak posi- including ecology, in 1928; and Depart-
tions of the action spectrum corresponded ment of Plant Biology in 1951. Year
to the absorption of cytochrome oxidase, Book 61 (p. 76) gives a list of the Staff
the active spectral region was greater Members and some other scientists and
than would be expected if cytochrome scholars associated with the Department
oxidase was the only material inacti- up to July 1, 1962. The Department's
vated. CIW Research Fellows to that time are
The fellowship program initiated by listed in the same Year Book on p. 102.
Dr. Vannevar Bush has kept a continual The following list of scientific workers
flow of talented young scientists coming extends the record to June 30, 1973.
Directors
Daniel T. MacDougal, 1905-1927
Herman A. Spoehr, 1927-1947 (Chairman)1
C. Stacy French, 1947-1973
Winslow R. Briggs, 1973-
Staff Members Since 1962
Joseph A. Berry, 1970- David C. Fork, 1960-
Olle Bjorkman, 1964- William M. Hiesey, 1926-1969
Jeanette S. Brown, 1958- Emeritus, 1969-
Jens C. Clausen, 1931-1956 Harold W. Milner, 1927-1965
Malcolm A. Nobs, 1938-40, 1951-
Emeritus, 1956-1969
1 On leave while Harvey Monroe Hall served
as Acting Director, Sept. 1930 to Sept. 1931. James H. C. Smith, 1925-1961
Chairman Emeritus, 1947-1956. Emeritus, 1961-1970
326
CARNEGIE INSTITUTION
Other Scientists Associated with the Department
Yvonne Aitken, 1962-1963
(University of Melbourne)
Jan Amesz, 1966-1967
(University of Leiden)
Jan M. Anderson, 1965-1966
(Division of Plant Industry, CSIRO)
Marcel Andre, 1968
(Service de Radioagronomie Commission a
l'Energie Atomique C.E.N. , Cadarache)
Lars-Olof Bjorn, 1969
(University of Lund)
George Bowes, 1971-1972
(University of Florida)
John E. Boynton, 1969
(Duke University)
Cornelis Bril, 1964-1965
(University of Utrecht), deceased, 1973
Georgi Detchev, 1972
(Bulgarian Academy of Sciences, Sofia)
Charles A. Fewson, 1962
(Cornell University)
Ralphreed A. Gasanov, 1972-1973
(Academy of Sciences of Azerbaijan, Baku)
David M. Gates, 1962-1963
(Missouri Botanical Garden, later, Univer-
sity of Michigan)
Eckard W. Gauhl, 1967-1969
(University of Frankfurt)
Martin Gibbs, 1962
(Brandeis University)
Govindjee, 1963
(University of Illinois)
Rajni Govindjee, 1963
(University of Illinois)
William G. Hagar, III, 1971-
Anthony E. Hall, 1970-1971
(University of California at Davis, later at
Riverside)
Per Halldal, 1955-1957, 1970
(University of Oslo)
Ulrich W. Heber, 1967-1968
(University of Dusseldorf )
Tetsuo Hiyama, 1971-
Gtinter Jacobi, 1969
(University of Gottingen)
Yaroslav de Kouchkovsky, 1963-1964, 1968
(Laboratoire de Photosynthese, C.N.R.S.,
Gif-sur-Yvette)
Eckhard E. Loos, 1968-1969
(University of Regensburg)
Daniel McMahon, 1967
(University of Chicago, later at California
Institute of Technology)
Axel Madsen, 1962-1963
(Royal Veterinarian and Agricultural Uni-
versity, Copenhagen)
Kenneth E. Mantai, 1968-1969
(State University College, Fredonia, New
York)
William G. McGinnies, Research Associate,
1972-1973
(University of Arizona)
Ernesto Medina, 1970
(Universidad Central de Caracas)
Wilhelm Menke, 1963
(Max-Planck-Institut fur Ziichtungsforsch-
ung, Koln)
Jean-Marie Michel, 1967-1968
(Universite de Sart Tilman, Liege)
Marie-Rose Michel-Wolwertz, 1967-1968
(Universite de Sart Tilman, Liege)
Alexander Miiller, 1965
(Max-Planck-Institut, Gottingen)
Norio Murata, 1969-1970, 1972
(University of Tokyo)
Teruyo Murata, 1969-1970
(Toho University)
Helga Ninnemann, 1972-1973
(University of Tubingen)
Robert W. Pearcy, 1969-1971
(State University of New York at Albany)
James M. Pickett, 1965-1967
(Montana State University)
Lawrence P. Raymond, 1970-1971
(University of California at Santa Cruz)
August Ried, 1964-1965
(University of Frankfurt)
Jerome A. Schiff, 1962-1963
(Brandeis University)
Ulrich Schreiber, 1972
(University of Aachen and Simon Fraser
University)
Marvin D. Schulman, 1962
(Cornell University)
Zdenek Sestak, 1968
(Czechoslovak Academy of Science, Prague)
James H. Silsbury, 1967-1968
(University of Adelaide)
DEPARTMENT OF PLANT BIOLOGY
327
Carl J. Soeder, 1964
(Kohlenstoffbiologische Forschungsstation,
Dortmund)
Atusi Takamiya, 1972
(Toho University)
John H. Troughton, 1973-
( Department of Scientific and Industrial
Research, New Zealand)
Wolfgang Urbach, 1964-1965
(Wiirzburg University)
William E. Vidaver, 1963-1965
(Simon Fraser University, Vancouver,
B.C.)
Hemming I. Virgin, 1953-1954, 1972
(University of Goteborg)
Dale A. Webster, 1963
(Illinois Institute of Technology)
Wolfgang Wiessner, 1969
(University of Gottingen)
Colin A. Wraight, 1969
(Cornell University)
REVERSIBLE INHIBITION OF ENERGY TRANSFER
BETWEEN PHOTOSYNTHETIC PIGMENTS BY
HYDROSTATIC PRESSURE
Ulrich Schreiber
Hydrostatic pressure strongly affects
all types of chemical and biochemical re-
actions proceeding with volume changes
(Johnson et al, 1954; Zimmerman, 1970;
Hamann, 1957). Vidaver (1963, 1964;
Vidaver and Chandler, 1969) studied the
effects of hydrostatic pressure on photo-
synthetic processes. They showed a re-
versible blockage of light-induced oxy-
gen uptake at about 400 atm and a
reversible inhibition of steady-state oxy-
gen evolution at about 1000 atm. Re-
cent investigations of pressure influence
on chlorophyll fluorescence induction
(Kautsky effect) support these findings
(Schreiber and Vidaver, 1972 a, b). Pres-
sures exceeding 1000 atm induce a re-
versible decrease in the fluorescence rise
believed to reflect pure photochemical
reactions (Kautsky et al., 1960; Delosme,
1967). Photochemical reactions are
normally not affected by these relatively
low pressures, and the pressure effect
could be on the energy transfer efficiency
within the pigment system. The results
of some spectrofluorometric experiments
with the red marine alga Porphyra per-
forata, reported here, confirm that the
pressure effect is on the energy transfer
process. In red algae, light energy ab-
sorbed by the phycobilins (phycoery-
thrin and phycocyanin) is transferred
with high efficiency by inductive reson-
ance via chlorophyll a to the photosyn-
thetic reaction centers. Duysens (1952)
and French and Young (1952) deter-
mined fluorescence action spectra for the
various pigment components in red algae.
They found that light absorbed by
phycoerythrin gives rise to some phyco-
erythrin fluorescence and, most interest-
ingly, to phycocyanin and chlorophyll a
emission as well.
Previous attempts to influence trans-
fer efficiency were made by methods that
caused irreversible damage to the photo-
£ 100
co
550 600 650
Wavelength, nm
700
Fig. 1. Effect of hydrostatic pressure on the
spectrum of stationary fluorescence emission
(F00) in Porphyra perforata. 5 X 10"5 M DCMU
added ; wavelength of excitation, 500 nm ; light
intensity, 3.3 nanoeinsteins cm"2.sec_1
328
CARNEGIE INSTITUTION
synthetic apparatus. Using methods to
apply hydrostatic pressure similar to
those of Vidaver (1969), it was found
that the application of pressure reversi-
bly induces a remarkable increase in
phycobilin fluorescence and a parallel
chlorophyll a fluorescence decrease.
Measurements were carried out with a
spectrofluorimeter, which automatically
corrected for the spectral sensitivity of
the apparatus (French and Koerper,
Year Book 65, p. 492) . Excitation light
(<500 nm) absorbed primarily by phyco-
erythrin gave a high chlorophyll a fluo-
rescence emission peak at 688 nm and
only a relatively low phycoerythrin peak
at 580 nm (for example, see 1-atm curve,
Fig. 1). The inhibitor 3-(3',4'-dichloro-
phenyl)-l,r-dimethylurea (DCMU) was
added to avoid possible pressure-induced
changes in a hypothetical fluorescence
quencher (Duysens and Sweers, 1963)
and to increase the stationary fluores-
at 580 nm
>\
0)
\
l-
\^
<v
^^^^
o
c
<D
i
o
t
W)
0)
1
o
Pressure on
Pressure off
n
^
1
!_
O
c
o
_o
''/
CO
at 688 nm
i i
l l l
Time.min
Fig. 3. Time course of simultaneous changes
in Porphyra perforata phycoerythrin fluores-
cence (580 nm) and chlorophyll a fluorescence
(688 nm) with the application and release of
1200 atm pressure. Maximum pressure at-
tained within 5 sec following on arrow; release
of pressure occurred within 0.1 sec. Fluorescence
intensities at the two wavelengths are not di-
rectly comparable. Other conditions as in
Fig. 1.
+ 100
+75
Q
+50
^25
25
50
75
- 100
Pressure difference
spectrum
F°° - F
1200 atm 1 atm
550
700
_J
600 650
Wavelength, nm
Fig. 2. Difference spectrum for fluorescence
emission (F^^atm — F°°iatm) calculated from
curves shown in Fig. 1.
cence yield. The 660 nm emission peak of
phycocyanin appeared as a small shoul-
der on the chlorophyll a peak. This re-
flects a high energy transfer between the
phycobilins and chlorophyll a and a low
photosynthetic energy conversion at the
reaction centers due to the DCMU in-
hibition. After the application of hydro-
static pressure there was a dramatic
change in the emission spectrum as
shown in Fig. 1. The chlorophyll a peak
at 688 nm is drastically reduced, a
phycocyanin peak at 660 nm emerges,
and the phycoerythrin peak at 580 nm
becomes predominant. The difference
spectrum (Fig. 2) shows a clear stimu-
lation of phycobilin fluorescence and sup-
pression of chlorophyll a fluorescence.
This effect was fully reversible upon re-
turn to atmospheric pressure. The char-
acteristic kinetics of the changes in fluo-
rescence intensities, when either applying
or releasing hydrostatic pressure, are
shown in Fig. 3. The application of pres-
sure caused a rapid biphasic rise in phyco-
erythrin fluorescence; at the same time,
there was an opposite chlorophyll a fluo-
DEPARTMENT OF PLANT BIOLOGY
329
rescence decrease. Upon release of pres-
sure, both returned within about 2 min to
their original levels. Without DCMU the
same general changes in fluorescence
emission occurred under pressure. Since
photosynthetic quenching of chlorophyll
a fluorescence was continuous without the
inhibitor, the magnitude of the unpres-
surized chlorophyll a peak was much
smaller and its suppression by pressure
less obvious.
Results similar to but somewhat less
reversible than the above were obtained
with pressurization of several other spe-
cies of red algae (Iridaea flaccidum,
Smithora naiadum, Porphyra nereocystis,
and Porphyra occidentalis) .
These results indicate that hydrostatic
pressure somehow blocks energy transfer
between phycobilin pigments and chloro-
phyll a. Under pressure, energy ab-
sorbed by phycoerythrin is not trans-
ferred to chlorophyll a and is reemitted
instead by the phycobilins. It is unlikely
that this effect results from shifts in
either the absorption or fluorescence spec-
tra of the individual pigments. We found
that in the range used here pressure has
no influence on the fluorescence of ex-
tracted red algal pigments. Additionally
no pressure effect was detected on the
absorption spectra of red algal pigments
in vivo and in vitro (Vidaver, 1969).
The question remains as to how hydro-
static pressure affects the energy transfer.
Gantt (1966) has reported that the
phycobilins are located in granules, so-
called phycobilisomes, loosely attached
to the stroma side of the thylakoid
lamellae, in which chlorophyll a is im-
bedded. While the pressure-induced
changes in fluorescence are fairly rapid,
the shifts are far from instantaneous
(Fig. 3). Half time (t1/2) for the first
phase is about 15 sec and the maximum
change is not yet attained after 2 min.
These relatively slow kinetics suggest
that pressure does not directly influence
the distances between pigment molecules.
Rather, the pressure effect appears to be
on membrane components of the pigment
systems. Membrane structure depends
on more or less patterned organization
resulting from the interactions of protein
and lipid molecules in an aqueous milieu.
Well-known effects involving molecular
volume changes (Johnson et at., 1954)
would affect the membrane structures
with pressures of the magnitude used
here. Within limits (e.g. irreversible pro-
tein denaturation) such changes are re-
versible, as is the inhibition of resonance
energy transfer shown here. It is not un-
expected, then, that applied pressure
alters a phycobilisome-thylakoid attach-
ment sufficiently to interfere with energy
transfer. On release of pressure, the
membranes restabilize and resonance
transfer resumes. All of this is consistent
with our observations that pressure does
not affect fluorescence of the individual
pigments in vitro.
Further studies of this nature may help
to clarify structural relationships of the
light-harvesting photosynthetic pigment
systems.
A postdoctoral research grant by the
Deutsche Forschungsgemeinschaft is
gratefully acknowledged. Dr. C. S.
French, Dr. N. Murata, and Dr. W. Vi-
daver are thanked for their advice and
discussion.
References
Delosme, R., Biochim. Biophys. Acta, lJf.3,
108-128, 1967.
Duysens, L. N. M., Transfer of Excitation
Energy in Photosynthesis, Thesis, Utrecht,
1952.
Duysens, L. N. M., and H. E. Sweers, Studies
on Microalgae and Photosynthetic Bac-
teria, Univ. Tokyo Press, pp. 353-372,
1963.
French, C. S., and V. K. Young, J. Gen.
Physiol, 35, 873-890, 1952.
Gantt, E., and S. F. Conti, Brookhaven Symp.
Biol., 19, 393-405, 1966.
Hamann, S. D., Physico-Chemical Effects of
Pressure, Academic Press, New York, 1957.
330
CARNEGIE INSTITUTION
Johnson, F. H., H. Eyring, and M. Polissar,
The Kinetic Basis of Molecular Biology,
Wiley, New York, 1954.
Kautsky, H., W. Appel, and H. Amann,
Biochem. Zh., 332, 277-292, 1960.
Schreiber, XL, and W. Vidaver, Proceedings
of the Canadian Society of Plant Physiolo-
gists, pp. 46-47, Dalhousie Univ., Halifax,
N. S., 1972a.
Schreiber, XL, and W. Vidaver, Proceedings
of the American Society of Plant Physiolo-
gists, Minneapolis. Plant Physiol, 1+9, (sup-
plement), 225, 1972b.
Vidaver, W., Photosynthetic Mechanisms of
Green Plants, NAS-NRC Publ. 1141, 726-
732, Washington, D.C., 1963.
Vidaver, W., Hydrostatic Pressure Effects on
Photosynthesis, Thesis, Stanford Univer-
sity, 1964.
Vidaver, W. and T. Chandler, Progress in
Photosynthesis Research, H. Metzner, ed.,
Vol. I, 514-520, Tubingen, 1969.
Vidaver, W., Int. Revue gesamten. Hydro-
biol, 54, 697-747, 1969. '
Zimmerman, A. M. (ed)., High Pressure
Effects on Cellular Processes, Academic
Press, New York, 1970.
DISTRIBUTION OF PHOTOCONVERTIBLE,
WATER-SOLUBLE CHLOROPHYLL PROTEIN
COMPLEX CP668 IN PLANTS RELATED TO
Chenopodium album
Atusi Takamiya *
A chlorophyll protein complex was dis-
covered by Yakushiji et al. (1963) from
the leaves of Chenopodium album. The
substance was characterized by its water
solubility. It could be extracted with
ordinary buffer solution from the leaves
without any addition of detergent or pre-
treatment of the material with acetone or
other reagents. The substance had a
definable constitution, a molecular
weight of 78,000 (Kimimura and Taka-
miya, unpublished data), and was com-
posed of 6 molecules of chlorophyll a
and one molecule of chlorophyll b per
molecule of protein. The most prominent
feature of this substance was that it was
light sensitive, undergoing, on illumina-
tion, a change of its chlorophyll a com-
ponent from the initial 668 nm-absorbing
form to a 743 nm-absorbing form. In
spite of these interesting features, it re-
mained an odd and unique substance of
rather dubious significance because the
only plant species in which it had been
detected were Chenopodium album, C.
rubrum, and Achyranthes japonica.
* Department of Biology, Faculty of Science,
Toho University, Narashino City, Chiba 275,
Japan.
This year a survey of the distribution
of this chlorophyll protein complex in
plants related to Chenopodium was car-
ried out. The substance, as identified by
the positions of the characteristic absorp-
tion bands in the initial and the photo-
converted forms, was found in several
species of A triplex (Chenopodiaceae). It
was also successfully extracted from
Amaranthus plants, with yields some-
times much higher than those from
Chenopodium album examined at the
same time. Another extension in distri-
bution of the substance was discovered
in Polygonum coccineum (Polygonaceae).
Table 1 summarizes the results obtained.
Materials and Methods
Plant material was collected from var-
ious localities (mostly California) as in-
dicated in Table 1. Plants cultivated in
this laboratory were also used for the
study. The materials available in the
vicinity were examined on the day of col-
lection. Those from remote places were
examined within a few days after the
trip, after storage in a cold room.
Extractions were made in the dark
or with a very dim green light. Plant
DEPARTMENT OF PLANT BIOLOGY
331
TABLE 1. Distribution of Water-Soluble, Photoconvertible Chlorophyll Protein (CP668)
in Plants of Chenopodiaceae and Related Families
CHENOPODIACEAE
Allenrolfea occidentalis (Watson) Kunze (DV)
Atriplex argentea Nuttall (CV, cult)
A. calif ornica Moquin-Tandon (cult)
A. confertifolia (Torrey and Fremont) Watson (CV)
A . coronata Watson (C V)
A . hortensis L. (St, cult)
A. hymenelytra (Torrey) Watson (DV)
A. lentiformis (Torrey) Watson (DV, cult)
A. littoralis L. (cult)
(-)
I I * *
''...(-)
+
+
..+ + **
+ +
+, (-) *
+ *
A. patula L. ssp. hastata Hall and Clements (St, PA) + + **
A. patula L. ssp. patula Hall and Clements (YS) +
A. phyllostegia (Torrey) Watson (ML) + *
A. rosea L. (St, PA) + *
A . semibaccata R. Brown (PA) ( —
A. serenana A. Nelson (LL, SJV) ( —
A. spinifera Macbride (OL) -f-
A. spongiosa Mueller (cult) ( —
Bassia hyssopifolia (Pall) Kunze (CV) ( —
Beta vulgaris L. (St, PA) ( —
Chenopodium album L. (St, PA cult) +, (++***) *
C. botrys L. (ML) (-
C. murale L. (St) ( —
C. rubrum L. (SJV) + + *
Salicornia ambigua Michaux (PA) ( —
Salsola Kali L. var. tenuifolia Tausch (OL) ( —
Spinacea oleracea L. (cult) ( —
Suaeda torreyana Watson var. ramosissima (Standley) Munz (DV) ( —
AMARANTHACEAE
Amaranthus albus L. (St) + + **
A . graecizans L. (SJV) +
A . retrofiexus L. (St) _!__!_**
Tidestromia oblongifolia Watson (Standley) (DV) ( — )
AIZOACEAE
Sesuvium sessile Robinson (CV) ( — )
PORTULACACEAE
Portulaca oleracea L. (St) ( — )
POLYGONACEAE
Polygonum coccineum Muhl (St) ( — ) *
Abbreviations for localities: CV, Central Valley; DV, Death Valley District; LL, Little Lake; ML,
Mono Lake; OL, Olancha; PA, Palo Alto; St, Stanford; SJV, San Joaquin Valley (California); YS,
Yellow Springs (Ohio) ; cult, cultivated at Carnegie Institution . Stanford.
* Present in stem and absent or insignificant in leaves.
** Predominance in stem was confirmed.
*** Result of previous test carried out with plant material collected in Japan.
material was ground in a mortar
with addition of quartz sand, or homog-
enized either in a Virtis high-speed
homogenizer (for smaller amounts of test
material) or in a Waring blender. There
was no essential difference in yields (if
any) of the substance. Two to four times
fresh weight amounts of 0.01 M K2HP04
or 0.05 M Na-K-phosphate buffer (pH
7.6) solution were added as the extrac-
tion medium to obtain about the same
amounts of clear supernatant solution
(extract) after ultracentrifugation of the
homogenate for 30 min at 144,000 g (Ta-
kamiya, 1971). The clear supernatant
solution was measured in the recording
spectrophotometer constructed and used
by the photosynthesis group of the In-
stitution. The absorption spectra were
determined usually with the same sample
332
CARNEGIE INSTITUTION
in the same cuvette before and after
illumination. When cloudy matter, prob-
ably consisting of lipid granules, ap-
peared in the top layer of the supernatant
solution, the lower clear portion was
gently removed with a Komagome pi-
pette and used for the determination of
the absorption spectrum. To avoid heat
denaturation of the substance, illumina-
tion was performed with the sample in a
cuvette dipped in ice water. White light
from a 100 W tungsten lamp (10 cm from
the position of the cuvette) was used.
Some experiments were performed with
red light passed through an interference
filter 668 nm (or 743 nm for reversal
tests; see below). After two minutes of
illumination with white light, practically
a maximum conversion of the chlorophyll
protein was obtained (Takamiya et al.,
1963). Because of likely contamination
by chlorophyll-containing material of
chloroplasts, the CP668 content of an ex-
tract was measured by the magnitude of
absorbance at 743 nm determined after
the sample had been fully illuminated.
Corrections were made for the distortion
of the spectrum due to light scattering
in cloudy samples as shown in Fig. 4.
The positive signs in Table 1 indicate the
presence of CP668 in terms of the mag-
nitudes of the absorbance at 743 nm in
extracts obtained by the procedures de-
scribed above: (-] — \-) , 743 nm absorb-
ance higher, and (+), lower than 0.01.
In view of wide fluctuations in yields of
extraction of the substance — as described
later — the best result obtained with a
particular species was recorded. The
lower limit of detection of the 743 nm
band with the instrument and the prepa-
ration procedures used was 0.0005 in ab-
sorbance units: (±) sign in the table.
Negative signs indicate that no trace of
a band at 743 nm was detected in a fully
illuminated sample.
Results and Discussion
A typical example for the detection of
CP668 is shown in Fig. 4. The decrease
in absorbance at 668 nm and the increase
at 742-743 nm caused by illumination
were taken as indicative of the presence
of CP668 in the initial sample. The
identity of the substance with the chloro-
phyll protein complex under investiga-
tion was further confirmed by the occur-
rence of the 565 nm band in the illumi-
nated sample. The revertibility of the
650
800
700 750
Wavelength, nm
Fig. 4. Light-induced absorption changes in an extract of a young stem of Atriplex hortensis.
The extract, prepared in the dark, was illuminated for 3 min at 0°C with white light. Measure-
ment was performed at 24 °C. The short wavelength part below the asterisk is shown above the
main curve.
DEPARTMENT OF PLANT BIOLOGY
333
743 nra absorbing form to the initial 668
nm absorbing form by illumination with
743 nm light in the presence of dithionite
(Takamiya et al., 1963) was also tested
when sufficient material was available for
the test, which always gave confirmative
results. There were samples showing the
red band of chlorophyll at around 670 nm
in which, however, illumination only
caused a decrease in the red band but did
not induce any absorbance increase at
743 or 565 nm. The red band in such
cases of photobleaching was interpreted
to represent some other nontransformable
form(s) of chlorophyll (protein) or con-
tamination by finely disrupted lamellar
material. The presence of such forms in
crude extracts obscures the exact position
(and magnitude) of the red band for
CP668, which seems to be also true of the
example shown in Fig. 4.
It will be seen from Table 1 summariz-
ing the results of the survey that the sub-
stance under investigation is not of
unique occurrence in particular plants of
Chenopodium but has a rather wide dis-
tribution, at least among such plant
groups as Atriplex and Amaranthus. In
fact, among 16 taxa of Atriplex exam-
ined, 13 showed a definite content of the
photoconvertible chlorophyll protein. All
of the three Amaranthus species exam-
ined showed the occurrence of the chloro-
phyll protein. The yields of extraction in
A. albus and A. retro flexus were much
higher than those obtained with Cheno-
podium album during the season covered
by the present study (June to October).
Polygonaceae, although represented only
by a single instance of Polygonum coc-
cineum in the table, was a new addition
to the list of plant families possessing
CP668. It will be noted that this family
is taxonomically closely related to
Chenopodiaceae and Amaranthaceae.
The last-named two families are classed
under the same group of Centrospermae
(Engler, 1897), or Chenopodiales (Hutch-
inson, 1926), to which the group Poly-
gonales, including Polygonaceae, is
placed adjacent both in Hutchinson's and
Engler's systems.
The negative sign in the table requires
some comment. The absence of the sub-
stance in the extract obtained as de-
scribed above might have been due to the
presence of some substance in the plant
tissue which interferes with the extrac-
tion or exerts some deteriorative effect on
the chlorophyll protein. Actually, no
positive result has ever been obtained
with plants giving, on extraction, deep-
brown extracts. This was a circumstance
encountered in many instances with
Chenopodiaceae plants, such as Allen-
rolfea (iodine bush), Salicornia (pickle
weed), Salsola (Russian thistle), and
Suaeda, to recall the worst. The season
of harvest, or the stage of growth, of the
plant is another factor that greatly influ-
ences the results. In general, the chloro-
phyll protein cannot be extracted from
plants at too young or too old a stage of
growth, as previously experienced with
Chenopodium album. The yield differs
markedly with the part of the plant body
used for the test, as will be described.
Caution, therefore, must be taken before
a final conclusion is reached as to the
absence of the substance in a plant
species. This is the reason why the nega-
tive signs for CP668 in the table are
enclosed in parentheses.
No systematic correlation has thus far
been established between the occurrence
of CP668 and the C3 or C4 type of carbon
dioxide assimilation in the plants tested.
In the same genus, Atriplex y there are C3
as well as C4 plants which give high
yields of the chlorophyll protein; for
example, Atriplex hortensis, A. patula
(C3) ; A. argentea, A. hymenelytra (C\).
It is not that all the C4-type Atriplex
species are positive in this respect; for
example, Atriplex lentiformis and A.
spongiosa were negative (in the same re-
spect). A similar, negative contrast in
C3 plants is encountered in Atriplex lit-
toralis. Similar discrepancies will be seen
on comparing representatives of Amaran-
thus and Chenopodium species showing
334
CARNEGIE INSTITUTION
high yields of the chlorophyll protein.
With spinach, a typical C3 plant of the
Chenopodiaceae, repeated trials failed to
detect the presence of CP668. Trials
have been made to correlate the occur-
rence and absence of CP668 and the
phylogenetic relationship of the Atriplex
plants (Hall et al, 1923) but without
success. CP668, as well as C3 and C4
photosynthesis, was not a specific attrib-
ute of some particular branch or sub-
group of the genus. The genetic problem
of the occurrence of CP668 has to be in-
vestigated in a future study. In this con-
nection, it should be noted that the hy-
brids between Atriplex rosea and A.
patula ssp hastata (Fl, F2, and F3) so
far examined were all strongly positive
in CP668, whereas one of the parent
plants, A. rosea, was mostly negative in
this respect.
Experiments were performed to inves-
tigate the localization of CP668 in vari-
ous parts of the plant. It was discovered
that generally stems are superior to leaves
of the same plant in yield of the sub-
stance. Younger leaves and younger
(upper) stems have much higher levels
of CP668 than the older, fully grown
ones, as exemplified by the results ob-
tained with a young plant of Atriplex
hortensis grown in the greenhouse of the
Institution (Table 2) . The plant was
still in the vegetative stage of develop-
ment, bearing seven pairs of leaves. The
yields of CP668 are expressed for con-
venience in terms of a743-units" of
CP668 per gram fresh weight of the
leaves; one 743-unit is defined as the
amount of CP668 contained in one ml of
extract giving a 743 nm absorbance
change of unity (after illumination). It
TABLE 2. Localization of CP668 in Atriplex hortensis
Size of Leaves (blades)
CP668 Content (743-units)
or Stem Segments
Fresh Weight
Fresh Weight
Per Leaf Blade
Specimen
mm
X mm
g (av.)
g"1
or Part of Tissue
Plant A
L->8*
0.07
0.0158
0.0011
L-8
55
X 45
0.35
0.0137
0.0048
L-7
110
X 78
1.2
0.0087
0.0105
L-6
120
X no
3.0
0.0022
0.0067
L-5
124
X 120
3.2
(...)
(...)
L-4
113
X 103
3.0
(...)
(...)
L-3
88
X 75
2.0
(...)
(...)
Plant B
L-8
85
X 57
0.59
0.051
0.030
(tip half, ca. 50 mm)
(0.27)
(0.061)
(0.0165)
(basal half, ca. 35 mm)
(0.32)
(0.043)
(0.0138)
L-7
123
X 88
1.64
0.0125
0.020
(tip half, ca. 75 mm)
(0.690)
(0.0105)
(0.0072)
(basal half, ca. 50 mm)
(0.95)
(0.0138)
(0.0132)
Stem 7-8f
length 63 mm
1.17
0.43
0.44
Stem 6-7
length 90 mm
2.33
0.21
0.49
Stem 5-6
length 98 mm
3.83
0.100
0.38
(upper half, 44 mm)
(1.73)
(0.102)
(0.177)
(lower half, 44 mm)
(2.08)
(0.098)
(0.204)
Plant material: Young plants (Plant A, 40 cm high, Plant B, 50 cm high) grown in the Institution's
greenhouse under long day conditions; harvested, August 25-26, 1972/Ext. No. 24.
*Position of the leaf pair numbered from the base upwards, taking that for the cotyledons as zero;
e.g., L->8 = youngest leaves (about 10) at the top of the erect stem.
fSegment of stem between L-7 and L-8 (etc.).
DEPARTMENT OF PLANT BIOLOGY
335
TABLE 3. Localization of CP668 in Stems and Petioles
Ratio,*
Chlorophyll
(743-units)
CP668 Content
Content
Tissues
(743-units)g~1
(663-units)g_1
(663-units)
Atriplex hortensis
Stem (upper part)
peripheral portion
0.100
13.1
0.0077
core
0.143
4.10
0.035
Stem (middle part)
0.049
7.95
0.0062
Petioles
0.007
6.70
0.0011
Atriplex argentea
Stem 5 (upper part)
0.0089
2.22
0.0040
Petioles
0.0028
10.7
0.0003
Atriplex rosea
Stems
0.023
5.90
0.0039
Plant materials: Test plants were grown in the greenhouse as in Table 2.
A. hortensis, 50 cm high; upper stem above 35 cm, middle stem between 35-30
cm from the base; Ext. No. 27, September 1-2, 1972. A. argentea, ca. 60 cm
high; stems, 10 cm portion, from the top; Ext. No. 28, September 1-2, 1972.
A. rosea, 60 cm high; stems, 7 cm portion from the top; Ext. No. 32, September
4-5, 1972.
* 663- and 743-units: See explanations in text.
will be noted that CP668 disappears from
leaves long before any sign of senescence
appears.
In another experiment, CP668 contents
of leaves and the stem were compared
with another individual of the same
species at a similar stage of growth
(Table 2, lower half). High yields (on a
one-gram fresh weight basis of tissue) of
CP668 in the stem deserve mentioning.
The relative amounts of CP668 to total
chlorophyll were tentatively measured
by the ratios (743-units/663-units) for
the leaf tissues, where the 663-unit (for
chlorophyll) is defined as above for the
743-unit and calculated as the product of
the 663 nm absorbance times volume
(ml) of the 80% acetone extract of the
given mass of tissue. Considering that
the absorbance of fully photoconverted
chlorophyll protein at 743 nm is not
much different from that of the initial,
nonilluminated form at 668 nm (Taka-
miya, 1971 ; see also Fig. 4) and that the
extinction coefficient of chlorophyll in
CP668 will be of the same order of mag-
nitude as that of chlorophyll dissolved in
acetone, the ratios (743-units/663-units)
shown in Tables 2 and 3 may be taken as
a rough measure for the actual ratio of
CP668 to total chlorophyll in a given
tissue. From the values shown in the
table, it is inferred that CP668 may
occur in younger parts of the stem of this
plant in significant amounts to attain a
level of several percent of the total
chlorophyll in the tissue.
Petioles seem to be much lower in con-
tent of CP668 as well as in its ratio to
chlorophyll (Table 3). The relative con-
tent of CP668 in leaves is generally far
lower, in spite of their high content of
chlorophyll in the tissue.
It has to be mentioned that all the
above discussions on the occurrence and
amounts of the chlorophyll protein in the
plants are based on the results of extrac-
tion of the substance from the plants. No
attempt to detect its presence in vivo has
thus far been successful. The physiolog-
ical role of the substance remains open to
future study.
Acknowledgments
The author expresses his gratitude to
Dr. French and all other members of the
Carnegie Institution in Stanford for their
336
CARNEGIE INSTITUTION
support and hospitality, which made this
study possible. The plant material was
collected and identified by the kind co-
operation of Dr. Malcolm Nobs. Pre-
cious plant material in the growth cabi-
nets and the greenhouses was provided
for use by the courtesy of the physiolog-
ical ecology group. Chenopodiwn album
was cultivated in the field of the Institu-
tion by Mr. Harry Lawrence and Mr.
Frank Nicholson. Special thanks are due
to Dr. Joseph A. Berry for first drawing
the author's attention to the Atriplex
plants to initiate the present study.
References
Engler, A., Ubersicht iiber die Unterabteil-
ungen, Klassen, Reihen, Unterreihen und
Familien der Embryophyta Siphonogama.
Naturliche Pflanz en familien, Nachtr. II-
IV, 1897.
Hall, H. M., and F. E. Clements, The North
American Species of Artemisia, Chryso-
thamnus and Atriplex in the Phylogenetic
Method in Taxonomy, Carnegie Institution
of Washington Publication No. 326, 1923.
Hutchinson, J., The Families of Flowering
Plants. I. Dicotyledons, 1926.
Takamiya, A., H. Obata, and E. Yakushiji, in
Photosynthesis Mechanisms in Green
Plants, National Academy of Science Pub-
lication 1145, National Research Council,
p. 479, 1963.
Takamiya, A., in Methods in Enzymology ,
Vol. XXIII, A. San Pietro, ed., p. 603,
1971.
Yakushiji, E., K. Uchino, Y. Sugimura, I.
Shiratori, and F. Takamiya, Biochim. Bio-
phys. Acta, 75, 293, 1963.
RESOLUTION OF THE LOW TEMPERATURE
ABSORPTION SPECTRA OF CHLOROPHYLL-
PROTEIN COMPLEXES INTO THEIR
COMPONENT BANDS
C. Stacy French, Atusi Takamiya, and Teruyo Murata
Water-soluble chlorophyll protein
preparations have been isolated, without
using detergents, from a variety of differ-
ent plants and some of these substances
have been crystallized (Yakushiji et at.,
1963). They have been described in a
review (Takamiya, 1971). One particu-
larly interesting chlorophyll protein, iso-
lated from Chenopodiwn album, has
been shown to be photoconvertible, a re-
action that reduces the 668 nm peak
while bands appear at 743 and 565 nm.
Other studies of the absorption spectra
of chloroplast fractions, measured at
— 196°C from a variety of algae and
land plants, have shown that most of the
ordinary bulk chlorophyll is present in
four different forms (French, Brown, and
Lawrence, 1972). These common forms
of chlorophyll a have their red absorp-
tion maxima at 662, 670, 678, and 684 nm
at the temperature of liquid nitrogen.
Longer wavelength forms in the 690 to
710 nm region also occur and are more
abundant in system 1 preparations. The
blue parts of the spectra from chloroplast
particles have not previously been re-
solved because of the overlapping ab-
sorption of chlorophyll and carotenoids.
The water-soluble chlorophyll protein
is only a very small fraction of the total
plant chlorophyll. These soluble chloro-
phyll complexes are, however, of particu-
lar interest since they have small molecu-
lar weights of about 80,000 (Takamiya,
1971). They are clearly definable forms
of chlorophyll which have the great ad-
vantage of freedom from carotenoids and
other components that interfere with the
blue part of the spectra of the usual
chloroplast fractions.
The question of whether the water-
soluble chlorophyll protein preparations
are made up of the same components (as
identified by their spectra) as are the
bulk chlorophylls can now be approached
DEPARTMENT OF PLANT BIOLOGY
337
by comparing the previous curve analy-
ses of chloroplast fractions with those of
a variety of chlorophyll protein com-
plexes. The spectra will first be com-
pared with each other to show differences
between the preparations from several
species and also the differences between
those recorded at low or at room temper-
atures. These differences will be dis-
cussed in terms of the components into
which the curves have been resolved.
Materials and Procedures
The preparations were made by pre-
viously described methods (Takamiya,
1971). Water-soluble chlorophyll pro-
teins of Chenopodium album, L., Lepi-
dium virginicum, L., and Brassica nigra
(L.) Koch were investigated. Cheno-
podium chlorophyll protein was purified
according to the method of Yakushiji
from the crude material extracted from
plants growing on the campus of the
Toho University in Narashino, Japan. A
pure crystalline preparation of Lepidium
chlorophyll protein obtained from plants
of the same campus was donated by Dr.
Yakushiji (prep. b). Another crystalline
preparation with a slight contamination
of flocculent sediment (prep, a; courtesy
of Mrs. Egi-Itoh) was also examined.
The spectrum measured in 1970 by N.
and T. Murata, using material isolated
from the Lepidium plants cultivated at
the Institution (initially introduced from
the Mono Lake district, California;
courtesy of Dr. M. Nobs), was also
analyzed. Brassica nigra chlorophyll
protein had been prepared and measured
in 1969-1970 by the Muratas, using plant
material growing in the Institution's
garden (for methods for extraction and
purification, see Takamiya, 1971). The
method of Takamiya, Obata, and Yaku-
shiji (1963) was used with Chenopodium
because its chlorophyll protein is basic
while those of Lepidium and Brassica,
being acidic, require the preparation
method described by Murata, Toda,
Uchino, and Yakushiji (1971) and by T.
and N. Murata (Year Book 70, p. 504).
Absorption spectra were measured in thin
cuvettes, or for some of the room temper-
ature measurements, in 1 cm cuvettes.
Water or dilute buffer mixtures were
used. The spectra here discussed fulfill
the requirements outlined for spectra to
be resolved by curve analysis (Pickett
and French, 1968).
Curve analyses were made as described
by French, Brown, and Lawrence (1972)
using the minimum number of bands that
are evident by inspection of the plotted
spectra. If adequate fits were not ob-
tained with the apparent minimum num-
ber of components, more were added as
necessary. In some cases the standard
bands that have been previously found to
fit many spectra of bulk chlorophyll of
chloroplasts were also tried. These input
bands were modified by the RESOLV
program in peak position, half width, and
height to achieve an optimum fit with the
number of bands entered. The band
shape was restricted to the Gaussian
form in all the results here reported, and
frequency, rather than wavelength, scales
were used for the analyses. The stand-
ard errors of fit are given as a percentage
of the mean, and the so-called error fac-
tor is the multiplier needed to make the
error curve a standard size. That factor
is given on the plots.
Spectral Measurements
The absorption spectra of chlorophyll
protein preparations from three plants
are shown in Fig. 5. A question we wish
to answer by curve analysis is whether
these preparations contain common com-
ponents whose differing proportions could
account for the striking differences be-
tween these spectra. It could be that each
species has its own characteristic com-
ponents, each corresponding to a specific
form of chlorophyll. A second question is
whether the usual forms of chlorophyll a
338
CARNEGIE INSTITUTION
600
650
Wavelength, nm
700
Fig. 5. The low temperature ( — 196°C) ab-
sorption spectra of chlorophyll proteins pre-
pared from three species.
found in ordinary land plants are also
present in these preparations.
In Fig. 6 we see comparisons of spectra
measured at room temperature, 24 °C,
and at liquid nitrogen temperature,
— 196°C. Because the spectra are
sharper at low temperature, the compo-
nents present are more easily detectable,
so most of our attention has been paid to
those curves. An understanding of room
temperature spectra is, however, essential
for comparisons of absorption and action
spectra. We do not know how to calcu-
late the room temperature spectra from
those at low temperature or vice versa
because the wavelength shifts and com-
ponent broadening are not always the
same. Figure 6 shows that the effect of
low temperature is to sharpen the spectra
and that there may or may not also be a
shift in the peak wavelength. Effects like
those in Fig. 6 have been reported for
spectra of whole algae (French et al,
1969).
A further comparison of spectra of two
species at two temperatures is shown in
the blue part of the spectrum in Fig. 7
(which also compares the preparations
from two species at room temperature) .
The 468 nm and the 459 nm bands for
the Lepidium and the Chenopodium
chlorophyll proteins, respectively, are
presumed to be due to chlorophyll b
(Murata et al, 1968).
The following spectra will be dis-
cussed:
Chenopodium album
C181 CZ181: 24°C, before phototrans-
formation, A.T. Oct. 14, 1972, prep. A,
curve 1, 580-810 nm (700-810 at 10 X
scale).
C182: 24 °C, after phototransforma-
tion, A.T. Oct. 14, 1972, prep. A, curve
2, 575-810 nm.
C183: — 196°C, after phototransfor-
mation, A.T. Oct. 14, 1972, prep. A,
curve 3, 575—810 nm.
C184 CZ184: 24°C, before phototrans-
formation, A.T. Oct. 27-28, 1972, prep.
A, curve 5, 380-770 nm (470-650 at 5 X
scale).
C185: — 196°C, before phototransfor-
mation, A.T. Oct. 27-28, 1972, prep. A,
curve II, 390—750 nm.
Lepidium virginicum
C105D: T.M. July 16, 1972; C114:
— 196°C, N.M., T.M., Dec. 18, 1970,
580-710 nm. The seeds of the material
used for C105 and CI 14 were collected at,
Mono Lake, California, and grown in the
greenhouse.
C177: 24°C, A.T. Sept. 25, 1972, prep.
a, curve 2, 400-800 nm.
C178: 24°C, A.T. Sept. 29, 1972, prep.
b, curve 2C, 280-740 nm.
C179: — 196°C, A.T. Sept, 28, 1972.
prep, b, curve 3, 600-800 nm.
The material for C177, C178, and C179
was harvested in Japan. These prepara-
tions have a low chlorophyll b content-
Some properties of different preparations
of the chlorophyll protein from two
sources are shown in Table 4.
Brassica nigra
C99: — 196°C, T.M. July 10, 1969.
C112: — 196°C, N.M., T.M., Nov. 5,
1970.
DEPARTMENT OF PLANT BIOLOGY
339
Chenopodium
before transformation
~\ 1 1 i r-
668 669
"i 1 r
"i 1 1 r
24°C
CI78
600
650 700
Wavelength, nm
750
800
Fig. 6. A comparison of the low temperature and room temperature spectra of chlorophyll
proteins. The red band of this Lepidium which is low in chlorophyll b shifts to a shorter
wavelength at low temperature while that of Chenopodium remains fixed in position. Both bands
are sharpened at low temperature.
Curve Analyses tion of the main peak of the Chenopo-
dium spectrum indicates that it may be
identical with Ca670 of the usual bulk
Chenopodium. The red band of this chlorophyll.
phototransformable chlorophyll protein, The curve analyses using two, three, or
seen in Fig. 6, is sharper than that from four components for the main red peak
either Lepidium or Brassica. The posi- of Chenopodium chlorophyll protein pre-
340
CARNEGIE INSTITUTION
1
" A
1 i i i i 1
i i i
I | I I l l
430 „o,
••'71 33
| I I I i | I
Chenopodium
CI85 [ \
-I96°C / I
-
3-^y •
A 1
/ CI84 B
-
/ 23°C |
-
/ \
459
458\\
u
0)
o
c
o
_Q
o
CO
_Q
- B
Lepidium
437
-
<
- \
CI78
24°C /""
7
A X
/ CI77 1
24°C '
469
\
l\ / 468 \
i
1 i i i i 1
i i i
i 1 i i i i
1 i i i i 1 i
300
350 400
Wavelength, nm
450
500
Fig. 7. The absorption in blue for Chenopodium at two temperatures and for Lepidium at room
temperature.
pared and kept in the dark, then meas-
ured at — 196 °C, are shown in Fig. 8. In
Fig. 8A the resolution of the main peak
into the two components 660.6, 10.8 and
669.1, 10.6 but without a band in the red
tail results in a sum curve needing im-
provement. The fit near the peak is
poor and a component is obviously miss-
ing near 677 nm in the tail. To improve
the situation a long wavelength compo-
nent estimated as 677, 12.0 nm was
added. This was transformed by
RESOLV to a 669.4, 26.9 nm component
as seen in Fig. 8B. The resulting fit near
643 nm (presumably chlorophyll b in
Chenopodium chlorophyll protein, chloro-
phyll a/b = ca 6.) was worse and the fit
near 677 improved only slightly. The
TABLE 4. Comparison of Two Ledipium Chlorophyll Proteins
Sample 1
Sample 2
Source
Japan
California
Curve designations
C105, C114
C117, C178, C179
Ratio, chlorophyll a/b
ca. 2.0
ca. 1.2
Molecular weight
80,000
40,000
Absn. peak at room temperature
660 nm
662 nm
DEPARTMENT OF PLANT BIOLOGY
34
170 160 150
Wavenumber.crrf1 XIO"3
Fig. 8. Attempts to match the low temperature spectrum of dark-prepared Chenopodium chloro-
phyll protein with two, three, or four components in the main red band. The standard error as a
percentage of the mean decreases as more components are added. A three-component fit is also
shown for the room temperature spectrum. The smaller bands are shown for other, more concen-
trated preparations in two spectral regions. The characteristics of the component curves are given
in Table 5.
apparent increase in the standard error
of fit in this analysis (Table 5) is merely
due to the fact that the easily fitted short
wavelength bands of Fig. 8A were
omitted in Fig. 8B and 8C, thus lowering
the statistical average goodness of fit. In
Fig. 8C another band — 657.6, 9.4 — was
used and 677 was more limited in its
allowable change per iteration. The re-
sulting fit is moderately good. The pa-
rameters of the component curves of Fig.
8 are given in Table 5. We believe the
bands of Fig. 8B may give the most rea-
sonable fit because those of 8C have un-
usually narrow widths. Furthermore,
those of Fig. 8C are consistent with the
room temperature bands of Fig. 8E.
For the room temperature spectrum of
Chenopodium, two main bands and a
small one in the tail gave an adequate
fit to the main peak as shown in Fig. 8E.
The two larger bands at — 196 °C — 661.1,
10.9, and 669.2, 10.1 — appear in the
24°C spectrum as 659.2, 13.3 and 669.0,
13.7 nm. The long wavelength bands at
669-682 nm, fitting the tail, are too small
for comparisons of their parameters in
the different curve analyses to have much
significance. These components are very
similar to those of ordinary chloroplasts.
Enlarged curves for the low parts of the
room temperature spectra are shown in
Fig. 8D and 8F with the band parameters
listed in the lower part of Table 5.
342
CARNEGIE INSTITUTION
t3
O
o
00
CD -■ ~
"^ fa
*-* d
. , 02
offl
O £
<H-H CS
h to
o fa
£ °
° s
o O
5£ '■£
OS O
l-H 0)
I «
,3 d3
•a 2
O 03
xti <C
03 • -h
|Q
la -5
d ■ g
cd
O
<
^ OS CD -tf
8s- t^ CO 00
i-H t^
o
< %x
^ fa ™ ^
H°° CO
O 6
i—l ' Oi
CO ' i—i
< bfl
o
d • r_l
CO
CO
id
CO
CO 1> CO
CO CO ©
t- i i—l CM
CN ©
OS OS CM
lO CO 00
co cO co
&S
O
W CO
ioqS os
d
rd
o
15 "5
CN
CO
00
1—1
CO
CN iq OS CO
co d os cd
HMkC
*q rH CO 00 00
co ai t> oo tjh
CO CO CO !>• (N
tJh N CO OS N
tH lO CO O N
CO co co co CO
S.o COCO^
^ o^ cn oo as
<! CN co
O
an W »
°0 00 t-h
8 i
co co
o
CD "3
00
CN
CO
OS
1—1
CO
CN
id
to
CO
^
°° oo — i
O !
CN
CN ^
OS
« s
tJh 00 iq CO
CO CO CO 1>
" (N i— i i— i i— i
^ d
i> co lq cn co t-h
Sic^io^ ' d os
d Q rH CN ** . CO CO
co co co co co co
£
G
O
LO
Sh
zr.
O
A
in
c
3
rri
02
CD
T)
k)
Fh
d
Sh
n -
CD
crt
RJ
>
Sh
P-.-Q T3
3
CD
H
d
3 oc
O
c3
fa
00
d
00 00 fa CO
• H 00 CO
bX)CS3 CO
faO d
CD
bfi
d
fa
d
o
,d
NhhhMNiOO
3 S t — i — i i—i i — f
co i— i as
i-h CN CN
lO in N N CO iC (N »C
£ id CO 00 00 CO Tf5 CN t£
CJOSOt-hCNCO^COOO
COt^I>t^t^l>t^t^
o
o
CN
73
d
oJ
PQ
. as i—i as |__I
o o
T— 1 1— 1
CO
CN
o3
s
CD
r-H CN
"tf
^d
i—i as
as
CO CO
CO co
CO
CO
fa s
as i— i
cdO
i-H 00
CN t^
X
o
*H
00 CO
Oh
o o
Oh
<1
1— 1 1— 1
^3
5 N kd N CD 6 ■*' CO
cCNCNCNCNCOCNCN
Q
00
4^ S
N CD iO N CD CO h
cn id id i-h oo as cn
iO iO iO lO lO CO CO
O
o
Sh
fa
o3
d
CD CD
I S3
d .gf J
DEPARTMENT OF PLANT BIOLOGY
343
Figure 9 shows the curve analyses of gives the characteristics of the compo-
the red spectrum of Chenopodium chloro- nents found. In the transformed mate-
phyll protein at the two temperatures rial the 743 peak was 1.1 X the height
after phototransformation, and Table 6 of the 668 peak in the low temperature
600
Wavelength, nm
650 700
750
CI83C
o
c
o
_Q
_Q
<
J L
l.i I
CI82B
24°C
155 145 135
Wavenumber.cm"1 XIO"3
Fig. 9. Gaussian components that fit the spectra of Chenopodium chlorophyll protein after
transformation by light. The whole curve, shown in Fig. 6, is here separated into two parts on
slightly different vertical scales for independent curve analyses. The 686 and 699 nm components
in Fig. 9A are the short wavelength side bands of the 743 peak. The 663, 670, and 680 nm
bands seem to have essentially the same wavelengths and apparent widths as the corresponding
bands of Fig. 8C before the transformation, but their proportions found here are different. The
bands found for the low temperature spectra, A and B, were used as input bands for the analyses
of the room temperature spectra, C and D. The components are described in Table 6.
344
CARNEGIE INSTITUTION
OS
bib
a
43
O
o
• ^
73
O
ft,
o
OS
O
73
CD
o
43
Oh
<
cd
o
44
PQ °8 ^>
S fto o Q
2,?2no3
O 5 I o
CO
§3S£oo
I a£ko1
O co i °
CO '
o3 73
d cd a a,
o a 73 g
Ejd £ d d
^
43 ^-v
73 S
K! d
bC
^ a
c3
c3 ^-^
73 d
bfi
^ a
03
c3 ^-v
CD >sP
CN 00 o
»d cd o6
H CO TtH
^ OS CO iO © CM »0
N iO h h (N N to
IM H HH H ^| M
q i> oo ^ co co oo
i-H d oo cd tjh id <n'
OMCNM^ion
t^ t^ I> t^ t^ I> t^
©NN
cd »d t>^
HCOTt*
i-5 oo cm' os os oo' cm'
cq i— i i— i i— i cm
OS CO i— i o lO l> OS
d os o i> tjh cd id
o -h CO w ^ CO oo
l> l>- b- l>- 1>- t^ b-
CO CM TjH
^ t^ oo *
i— i CO rt<
43
73
CM OS 00 CN i-H tJH CN © CM O
CO r-5 o "* CM* i-3 r-5 r-5 t-5 r-5
CO CM CM
CM
^OHNOqcOM "* ho
S o »d O rJH rjn CO d oo" CO* OS
flOlrHCO^lOCONNOOCO
— 'lOCOCDcocOcOcOcOcocO
73
^
N 00 00 N
»d os id oo
COiOCM-^OSOSCOiOCMCO
HHffiMNOlOO
COIN H H H ,— I t-H
CM O
i-i CM
•p-t>-oscqcococococooso
S cd ■<* os cd t»5 cd d os id os
AOl-HlM^iOCONNOOOl
— ^lOcOcOcOcOcocOcOcOco
o
o
CD
43
73
CD
a
o
a
CD
CD
42
CD
o3
43
CD
43
73
C
cd
o3
44
o3
CD
a
43
bJD
G
pCD
>
o3
CD
bfi
d
O
CD
43
43
+3
73 <3
CD CD
+= Oh
o3
■a °°
°^ 43
02 +=>
d .5
^H 0
c3 CD
a d
o
Oh
a
o
O CD
CD 43
43 +="
CD ©
c3
CD CD
cp oo
43 CD
DEPARTMENT OF PLANT BIOLOGY 345
spectrum and 1.27 X the height at 24°C 699.0, 20.6 band shown in Fig. 9A. With-
(Fig. 6). Since the red band does not out the 719.3, 18.0 band a moderately
completely disappear on phototransfor- good fit (not shown) to this part of the
mation, the question arises as to whether spectrum (SE = 0.353, error factor =
the remaining band is the same as that of 8.89) was found with the following
the material before transformation. The bands: 701.5, 43.0 (not consistent with
half-width of the red peak and of the two the adjacent part of the spectrum) ; 729.5,
major components of the transformed 13.8; 736.7, 9.3; 744.3, 9.8; 763.7, 17.4;
material is greater and its height is lower and 783.2, 26.2. The simplest possible
in proportion to the shorter wavelength analysis for the 743 band at — 196 °C
side bands. The width at half-height of that matches the adjacent part of the
the Chenopodium material before trans- spectrum is therefore the one defined by
formation is 12.8 and afterwards is 17.0 Fig. 9B. The somewhat broader bands
nm. However, the corresponding widths fitting the room temperature spectra are
at half the height above the 643 peak are shown in Fig. 9, C and D, and the band
11.2 and 12.4, showing that the apparent parameters for all of these fits are given
increase in half-width may be due mainly in Table 6.
to the fact that the overlapping absorp- In Fig. 8F we see greatly enlarged
tion by other bands becomes greater the curves for the long wavelength part of
less the absorption is in the main band the Chenopodium spectrum. Here, in
itself. We find no clearly significant curve CZ181A, it is clear from the long-
difference in the components of the main wavelength bands that a small amount
red peak before and after transforma- of phototransformation has already
tion, although the proportions of the two taken place during the preparation,
larger components do vary as seen in Among these, the 743 nm band is a
Tables 5 and 6 and in Figures 8 and 9. normal product of photoconversion, while
Table 6 lists the parameters of the bands the 763 band represents a by-product
shown in Fig. 9. For the curve analyses appearing after prolonged storage, which
these two peaks were plotted with equal was the case with this particular prepara-
heights and fitted separately. A moder- tion. The 762 band is higher than that
ately good fit (not illustrated) from 610 near 740, while in the spectrum of the
to 700 nm (SE = 0.496, error factor = intentionally transformed material of
7.99) was obtained for the red peak at Fig. 9B, the corresponding band (764) is
— 196°C after transformation with the relatively very much smaller. This shows
following eight bands: 596.2, 31.3; 614.3, that the two bands, 740 and 764 nm, do
22.2; 628.1, 19.2; 646.0, 18.6; 662.7, 12.2; not both belong to the same substance.
669.9, 12.0; 682.7, 9.8; and 697.5, 32.2. The 763 band seems to be light insensi-
However, the width of the 697.5 peak is tive.
not consistent with the other half of the Lepidium virginicum. A frequent ques-
same band in the analysis of the 743 peak tion that must be repeatedly answered in
shown in Fig. 9B. Furthermore, the 646.0 curve fitting is whether a small addi-
peak was obviously at too long a wave- tional component that helps to give a
length and the fit near 668 was not ideal, better fit is really significant enough to
Therefore, two more bands were added to represent an actual component. This
give the results shown in Fig. 9A, which problem is illustrated in Fig. 10 and
is considered to be the simplest accept- Table 7 with two chlorophyll protein
able resolution of this curve. preparations from Lepidium. These
Similarly for the 743 peak the 719.3 spectra resolve into components that are
band was added to make the 700.9, 21.3 not the same as those usually found for
band approximately consistent with the the bulk chlorophyll of usual green
346
CARNEGIE INSTITUTION
Wavelength, nm
700 600
0)
o
c
O
_Q
<
145 165
Wavenumber, cm"1 XIO"3
Fig. 10. Curve analyses of Lepidium chlorophyll protein spectra with and without a "643" nm
component which, while not evident as a hump in the curve, does improve the fits obtained. CI 14
has a higher chlorophyll b component than does the other preparation which was measured at
two temperatures. Table 7 gives the parameters of the components.
plants. The curve analyses for two Lepi-
dium spectra are summarized in Fig. 10,
and Table 7 gives the parameters of the
Gaussian curves used to fit each of the
spectra with and without a "643" nm
component. CI 14 contains more of a
chlorophyll b component than does C179.
The exact position and width of the 590
to 600 nm component is not of much sig-
nificance, but some estimate of it has to
be included because its longer wavelength
partly overlaps the other bands. It is of
interest to note that there is no hump
indicating a 643 nm component in Lepi-
dium even though that component is so
clearly evident in the preparations from
DEPARTMENT OF PLANT BIOLOGY
347
TABLE 7. Curve Analyses of Chlorophyll Protein Preparations from Lepidium of Fig. 10
Showing the Improved Fit When a Small Hidden Component Is Included
Without the "643" Component
With the "643" Component
Wavelength
Width
Area
Wavelength
Width
Area
(nm)
(nm)
(%)
(nm)
(nm)
(%)
A Preparation
with High Chlorophyll b Content, -
196°C
C114D, SE = 0.260
C114B
, SE = 0.248
597.2
41.7
592.5
32.0
612.6
21.7
611.9
21.6
633.5
21.5
631.1
640.7
19.5
13.6
3.9
649.2
11.1
28.4
649.4
11.6
29.0
658.7
11.2
44.2
659.4
12.0
46.0
666.7
10.6
19.2
668.2
11.3
16.8
673.4
12.9
5.0
(680.4
13.4
2.6)
(688.7
20.8
3.1)
(693.9
15.0
1.6)
A Preparation
With Low Chlorophyll b Content, -
196°C
C179A, SE = 0.269
C179B
, SE =0.248
600.0
34.3
600.0
35.9
...
613.4
19.3
613.3
19.2
*
631.3
18.0
631.6
644.7
18.5
9.1
6.5
647.3
11.8
23.2
652.0
14.6
33.2
657.4
11.3
45.8
659.4
11.9
41.6
661.1
10.7
19.9
667.2
8.6
9.6
(672.4
60.9
11.1)
672.8
(686.9
(699.7
8.6
24.3
64.2
2.9
2.9)
3.3)
A Preparation with Low Chlorophyll b Content, !
24°C
C178A, SE = 0.485
C178B
, SE = 0.269
587.3
59.7
590.9
68.9
615.0
25.5
615.8
26.0
. . .
634.1
20.5
635.8
643.3
21.9
11.1
3.9
650.0
15.4
20.4
652.0
11.9
17.8
661.4
14.5
50.5
661.5
12.7
44.9
670.8
14.4
21.0
669.8
13.0
22.7
679.0
24.0
8.1
677.1
(685.6
13.8
29.4
6.4
4.2)
Chenopodium. The use of a 643 nm com-
ponent does, however, reduce the stand-
ard error, as shown in Table 7. The small
bands averaging 613.7 and 632.6 nm
which are 18.9 nm apart from each other
may be vibrational side bands associated
with the main peaks at 648.7, 658.4, and
666.6 nm. If so, there should be three,
rather than two, such small bands in the
610-640 nm part of the spectrum. How-
ever, because the widths of these small
bands are large in proportion to their
separations, it is not easy to establish the
existence of a third band in this region.
The band near 650 nm is presumably
due to chlorophyll b and it is larger in
these preparations than in the bulk
chlorophyll of normal green plants. A
640 nm band (free chlorophyll b) is
usually necessary to match the spectra of
chloroplast preparations of chlorophyll
6-containing plants and it may be about
20% of the height of that at 650 nm. In
these chlorophyll protein preparations,
348
CARNEGIE INSTITUTION
TABLE 8. Curve Analyses of Two Mustard (Brassica) Chlorophyll Protein Spectra
at -196°C (Not Illustrated)
C99E
C112B*
Peak at 670 nm
Peak at 672 nm
Wavelength
Width
Area
Wavelength
Width
Area
(nm)
(nm)
(%)
(nm)
(nm)
(%)
586.8
43.3
596.7
621.1
45.3
19.3
627.7
40.7
637.6
23.0
652.8
10.9
5.8
652.6
11.6
6.3
662.2
12.1
24.3
662.1
11.7
21.4
669.7
11.1
40.3
670.4
10.9
36.8
676.5
10.8
22.2
675.8
10.5
25.2
683.0
9.9
6.2
683.5
8.8
10.3
(688.9
11.6
1.0)
"From French, Brown, and Lawrence (1972), but areas are here calculated to include the 653 band
however, the "640" component seems to
be nearer 643 nm. Analysis of another
curve for Lepidium (C150D: T.M. July
16, 1970) gave the following components
that further confirm the wavelengths of
the unusual bands: 600.1, 27.0; 615.1,
22.1; 633.6, 20.0; 649.5, 11.4, 35.3%;
659.2, 10.3, 46.9%; 666.9, 8.7, 13.6%;
672.2. 11.4, 4.2%. The red tailing of ab-
sorption, which gives 680 and 694 bands
of C114B, is caused by a denaturation of
the protein due to a long storage. This
tailing is not seen in another spectrum,
C105, of the high chlorophyll b-Lepidium
protein.
In normal green plants the bulk of the
chlorophyll is in the following major
bands: C6640, C6650, Ca662, Ca670,
Ca677, and Ca684 (French, Brown, and
Lawrence, 1972). In Lepidium the only
band we find that may be one of these,
C6650, is at 648.7. The other major
bands in Lepidium are unexpectedly at
about 657-661 and at 666-669 nm.
Brassica nigra (L.) Koch (Wild mus-
tard). This chlorophyll protein has its
peak at longer wavelengths than either
the Chenopodium or the Lepidium chloro-
phyll proteins and can be fitted by bands
identical to those fitting the usual chloro-
phyll spectra of ordinary green plants.
The curve analysis of C112B with a peak
at 672 nm appears in Fig. 7 of French,
Brown, and Lawrence (1972). Another
preparation with a peak at 670 nm can
also be fitted with the common chloro-
phyll bands as shown in Table 8. The
different percentages of the components
as well as the band parameters of the
components themselves determine the
peak position of the measured spectrum.
BLUE REGION
Chenopodium. The components found
for the blue part of the spectrum of
Chenopodium, seen in Fig. 11 and Table
9, have the same wavelength peaks at
— 196° and 24°C. The low temperature
bands are consistently narrower. Be-
cause the RESOLV program cannot han-
dle a component whose peak lies outside
the range of data wavelengths, the short-
est wavelength components shown are
not correct. Thus, for instance, the 390
nm Chenopodium component may well be
actually nearer to 380 nm.
The 468 peak of Lepidium in Fig. 12
and Table 9, presumably due to chloro-
phyll b, is much higher in another prep-
aration, C177, than in C178 of Fig. 12,
suggesting that the a/b ratio is depend-
ent on preparative procedures and is not
a definite characteristic of the complex.
DEPARTMENT OF PLANT BIOLOGY
349
o
c
o
_Q
JQ
<
CI85A
Error X4.85
X8.89
"~* \/'V«*(vv-^-
250 235 220
Wavenumber, cm"
205
XIO"3
190
Fig. 11. Resolution of the blue part of the
spectrum of Chenopodium chlorophyll proteins
at two temperatures.
Furthermore, the corresponding band in
Chenopodium comes at 458 nm, not 468
nm, a fact so far without adequate inter-
pretation. The 439 band is, however,
similar in the preparations from the two
species. In CI 77 a better fit was achieved
by breaking the 439 component into two
at 436 and 443 nm. The dominant 433
band of the Chenopodium preparation
seems to correspond to the 437 nm com-
ponent of Lepidium.
Because the RESOLV program is
limited to 10 components, Fig. 12 had to
be prepared in two separate parts, using
for the shorter wavelength part, an ab-
breviated spectrum from which the com-
ponents longer than 350 nm had been
subtracted. This accounts for the small
break in the 363 nm component.
The gift of a preparation of beautiful
blue crystals of chlorophyll protein from
Professor E. Yakushiji and Mrs. Egi-
Itoh is acknowledged with gratitude. We
are indebted to Dr. D. D. Tunnicliff of
the Shell Development Laboratory for
the original RESOLV program and to the
National Science Foundation for grant
no. GB8630, which paid for some of the
computer time.
Summary
The red band of the Chenopodium
chlorophyll protein spectrum at low tem-
perature can be fitted reasonably well by
two chlorophyll a components at 661
and 669 nm that are only 1 nm shorter in
wavelength than common constituents of
green plant chloroplasts. The chloro-
phyll b band of Chenopodium chloro-
phyll protein, however, comes at about
643-645 instead of at 650 nm.
The chlorophyll protein spectrum from
wild mustard (Brassica nigra) can be
fitted by the usual components of ordi-
nary green plants at 662, 670, 677, and
683 nm with its chlorophyll b band at
about 652 nm.
The Lepidium spectra have normal
chlorophyll b components at 649-652 nm
but contain unusual chlorophyll a forms
with peaks at 659 and 667-668 nm.
The blue part of the spectrum of
Chenopodium chlorophyll protein was re-
solved into Gaussian components at two
temperatures and that from Lepidium
measured at room temperature was ana-
lyzed in the blue and the ultraviolet part
of the spectrum.
350
CARNEGIE INSTITUTION
TABLE 9. Components Fitting the Absorption
in the Blue and Ultraviolet
Spectra of Chlorophyll Protein Complexes
Parts of the Spectrum
Chenopodium,
Fig. 11.
Lepidium,
24°C, Fig. 12.
C185A
-196°C
C184B 24°C
C178A, C178D
C177 (not illustrated)
Wave-
Wave-
Wave-
Wavelength Width
length
Width
length
Width
length
Width
(nm)
(nm)
(nm)
(nm)
(nm)
(nm)
(nm)
(nm)
(280.2
22.2)
292.7
(4.7)
299.7
14.0
316.4
29.6
339.8
24.4
363.3
24.7
(390.0
21.8)
(390.0
23.9)
382.3
401.0
21.9
21.9
404.7
14.5
404.9
17.6
416.3
14.0
416.3
14.6
418.6
22.1
(414.4
27.9)
425.8
9.1
425.3
10.9
432.7
7.1
432.8
10.7
436.2
25.1
439.4
11.2
440.5
13.1
439.3
21.5
443.0
20.0
458.2
21.8
457.3
22.5
469.2
19.8
468.3
476.3
19.5
23.3
488.3
22.1
0)
u
o
c
o
u
o
(/>
_Q
<
CI78A
Error X6.08
X4.04
340
220
310 280 250
Wavenumber.cm-1 XIO-3
Fig. 12. Resolution of the ultraviolet and blue part of the 24 °C spectrum of Lepidium chloro-
phyll protein. The 469 nm band is presumed to be due to chlorophyll b.
DEPARTMENT OF PLANT BIOLOGY
351
References
French, C. S., J. S. Brown, and M. C. Law-
rence, Plant Physiol, 49, 421, 1972.
French, C. S., M. R. Michel-Wolwertz, J.
Michel, J. S. Brown, and L. Prager, pp.
147-162, in Biochem. Soc. Symp., 28,
Porphyrins and Related Compounds, T. W.
Goodwin, ed., London, 1969.
Murata, T., Y. Odaka, K. Uchino, and E.
Yakushiji, pp. 222-239, in Comparative
Biochemistry and Biophysics of Photosyn-
thesis, K. Shibata, A. Takamiya, A. T.
Jagendorf, and R. C. Fuller, eds., Univ.
Tokyo Press, 1968.
Murata, T., F. Toda, K. Uchino, and E.
Yakushiji, Biochim. Biophys. Acta, 245,
208, 1971.
Pickett, J. M., and C. S. French, Carnegie
Inst. Year Book 66, 171, 1968.
Takamiya, A., H. Obata, and E. Yakushiji,
pp. 479-485 in Photosynthetic Mechanisms
of Green Plants, B. Kok and A. T. Jagen-
dorf, eds. NAS-NRC Wash., D.C., 1963.
Takamiya, A., Methods Enzymol, 23, 603,
1971.
Yakushiji, E., K. Uchino, Y. Sugimura, I.
Shiratori, and F. Takamiya, Biochim. Bio-
phys. Acta, 75, 293, 1963.
A COMPARATIVE STUDY OF THE FORMS OF
CHLOROPHYLL AND PHOTOCHEMICAL
ACTIVITY OF SYSTEM 1 AND SYSTEM 2
FRACTIONS FROM SPINACH AND Dun all ell a
J . S. Brown, R. A. Gasanov* and C. S. French
In vivo chlorophyll a exists as several
forms with distinct absorption spectra
(Brown, 1972). It is generally believed
that two photosystems function in green
plant photosynthesis. One photoreaction
(photosystem 2) is closer to the 02 evolv-
ing step and mainly utilizes visible light
of wavelengths shorter than 680 nra ab-
sorbed by such active pigments as chloro-
phyll b and certain forms of chlorophyll
a. The other (photosystem 1) is closer to
NADP reduction and is mediated prefer-
entially by the longer wavelength forms
of chlorophyll a. When the photosystems
have been partially separated physically
from each other, absorption spectra that
show the relative amounts of the different
pigments in each fraction may be used
to indicate the degree of separation.
A method developed by Michel and
Michel-Wolwertz (Year Book 67, p. 508,
1968) for partially separating these two
photochemical systems depends upon the
disintegration of chloroplasts by extru-
sion through a needle valve (French
press) followed by centrifugation in a
* Institute of Botany, Academy of Sciences of
Azerbaijan SSR, Baku 370122, U.S.S.R.
linear sucrose gradient. This gives a
green band near the top, termed fraction
1, which contains the system 1 particles,
and a lower band, called fraction 2, which
contains primarily the larger system 2
particles or fragments which still have
some system 1 activity. In addition, it is
known from the work of Jacobi and Leh-
mann (1969) and Sane et al. (1970) that
stroma lamellae contain photosystem 1
with little or no photosystem 2 and that
grana lamellae contain both photosys-
tems. These authors showed that one
may readily separate the grana and
stroma membranes by physical disrup-
tion followed by differential centrifuga-
tion.
Recently, Arntzen et al. (1972) have
further fractionated press-treated par-
ticles with digitonin. They were able to
fractionate the grana membranes (frac-
tion 2) into 60% photosystem 2 and
40% photosystem 1 particles on a chloro-
phyll basis. The photosystem 1 particles
obtained from stroma and grana mem-
branes were similar in electron transport
activity, P700 content, ultrastructure ap-
pearance, and ultrafiltration character-
istics. However, the stroma photosystem
352
CARNEGIE INSTITUTION
1 fragments did not recombine with the
grana photosystem 2 fraction to recon-
stitute electron transport activity from
diphenylcarbohydrazide (DPC) — >
NADP+ as did the grana system 1 frac-
tion. Thus these fractions do differ from
each other in some characteristics.
Most of the fractionation work has
been carried out on leaf chloroplasts,
usually spinach, because of the ease with
which large amounts of material can be
obtained. On the other hand, analyses
of various algal spectra have given sup-
port to the "forms of chlorophyll" hy-
pothesis because, for some unknown rea-
son, these spectra contain sharper, more
discrete maxima and shoulders than do
spectra of leaf chloroplasts. Therefore,
in order to study possible relationships
between the forms of chlorophyll in vivo
and photochemical activity, it would be
desirable to work with algal material.
This has not been done extensively be-
cause of the difficulty in extracting ac-
tive chloroplasts from large volumes of
particular algae.
We have overcome this problem with
Dunaliella by adapting it to grow in a
high salt medium and then subjecting the
cells to osmotic shock. This effectively
releases whole chloroplasts which are
photoreactive and may be treated the
same as spinach chloroplasts. However, we
have no information about the relative
proportions of grana or stroma lamellae
in Dunaliella grown under our conditions.
Thus, although fractions 1 and 2 from
Dunaliella were prepared in the same
way as the stroma and grana fractions
from spinach, they may not be struc-
turally equivalent.
This paper reports studies of the
photochemical activity and absorption
spectra of different fractions from spin-
ach stroma and grana and from Duna-
liella. Curve analyses of these different
spectra show that the pigment composi-
tion of the two photosystem 1 fractions
differs slightly and that the lack of the
chlorophyll form Ca694 correlates with
a lack of photosystem 1 activity.
Materials and Methods
Algae. Dunaliella is a halophilic, bi-
flagellate alga belonging to the suborder
Chlamydomonadineae. Many people
have studied its halophilic characteris-
tics, and, in particular, Wegmann (1971)
found that the organism apparently pro-
duces glycerol to protect itself from an
osmotically high external environment.
It grows well up to approx. 2.8 M NaCl.
Ben-Amotz and Avron (1972) noted
that active chloroplasts were easily pre-
pared from D. parva by osmotic shock.
For our experiments D. salina, originally
obtained from the Culture Collection at
Indiana University, No. 200 (Starr,
1964) , was adapted to a high salt medium
by serial transfer of large inocula for
about a month. Thereafter, the cultures
were grown in one liter volumes of arti-
ficial sea water (Jones et al., 1963) plus
\.b M NaCl on a shaker over 2 warm-
white and 2 cool-white 40 W fluorescent
lamps for from 5 to 7 days. Air en-
riched with 3% C02 was passed through
the flask continuously.
The algae were collected by centrifuga-
tion of the culture, washed once with 90
ml of 50 ml Tricine pH 7.5, and resus-
pended in about 30 ml of 150 ml KC1-
Tricine. They were then fractionated in
the same way as the spinach chloroplasts
described below.
Fractionation procedure. Chloroplasts
were isolated from spinach leaves by
blending in phosphate buffer (pH = 7.8)
with 400 ml sucrose and 10 ml NaCl.
They were resuspended in the same KC1-
Tricine buffer used for the algae and
forced through the French press three
times at 12,500 psi (800 kg/cm2) as pre-
viously described (Michel and Michel-
Wolwertz, Year Book 67, p. 508, 1967).
The subsequent steps in the fractionation
procedure are shown in Fig. 13.
Digitonin fractionation. As indicated
in Fig. 13, Fr 2 (0.25-0.3 mg chl ml"1)
was made up to 1% digitonin and incu-
bated at 4°C for 30 min with gentle stir-
ring. After this period, 3 ml aliquots of
DEPARTMENT OF PLANT BIOLOGY
353
Chloroplasts
1 3 passes through needle valve
Centrifuge, 3,000g
5min
t ( Homogenate)
Sediment
Discarded, unb
roken chloroplasts
bupernatar
lO.OOOg
, 30min
ant
, 30min
r
Sedimen
Resuspended ir
t
3.000g,
buffer
5min
Superna'
J
60,000g
r
"1
r ~t
Sediment
Supernatant
Supernatant Sediment
Discarded
1
Discarded
I44,000g,60min
1
i
|Fr2|
I
Sediment
Digitonin incubation
(1% Dig.),30min,4°C
[777]
r
"1
Differential cen
Sucrose gradient cen
I0,000g,30min
60,000g,45m,n
1
Sediment
t
"1
Supernatant
60,000g,30mm
1
— — |Fr ig|
|Fr2d|
Si
d
55%
sue.
w
r
diment
scarded
"1
Supernatant
1
I65,000g,60mm
— |Fr 2g
1
1
Sediment
1
Supernatant
*
IFrldl
I65,000g,6hrs
r
1
Supernatant Sediment
(Colorless) 1
| Fr 6hr|
Fig. 13. Chloroplast fractionation flow diagram.
the mixture were fractionated by centri-
fugation at 60,000 g for 45 min on linear
sucrose density gradients (12.5-55%
W/V, 150 ml KC1, 50 ml Tricine,
pH = 7.8) . In some experiments the
detergent-treated Fr 2 was further frac-
tionated by differential rather than by
gradient centrifugation.
Measurements. Chlorophyll a/b ratios
were determined by the method described
by Arnon (1949). The light-saturated
reaction rates of DCIP (dichlorophenol-
indophenol) reduction (system 2 activity)
and of cytochrome c oxidation (system 1
activity) were measured as previously
described in Year Book 70 (p. 499, 1971).
Absorption spectra at — 196 °C were re-
corded with the spectrophotometer de-
scribed previously (Year Book 66, p.
175, 1967). Curve analysis was carried
out on digitized spectra as described in
Year Book 69 (p. 662, 1970) .
Results and Discussion
When the algae were transferred from
their culture medium to the KC1 -Tricine
buffer, they gradually swelled and rup-
tured, but the chloroplasts did not break.
In preliminary experiments, Dunaliella
fractions enriched in systems 1 and 2
were prepared by the treatment of osmot-
ically shocked cells with digitonin ac-
cording to the method of Ohki and Taka-
miya (1970) for spinach. Usually it is
impossible to fractionate algal cells with
354
CARNEGIE INSTITUTION
detergents without first breaking them
by mechanical procedures that also dis-
integrate their chloroplasts and may
cause some uncontrolled fractionation.
The needle-valve treatment of both
spinach and Dunaliella chloroplasts sus-
pended in high salt buffer followed by
differential centrifugation yielded the
usual system 1 (Fr 1) and system 2 (Fr
2) particles. On centrifugation of the
digitonin-treated grana membranes (Fr
2) in a sucrose density gradient, two dis-
tinct green bands formed. They will be
designated, from the top downwards, as
Fr lg and Fr 2g. The chlorophyll com-
positions by curve analysis and the ac-
tivities of photochemical systems 1 and
2 of these four (Fr 1, Fr 2, Fr lg, and
Fr 2g) fractions were investigated.
The photochemical activities and
chlorophyll a/b ratios of the various
membrane fractions are shown in Table
10. The ratio of chlorophyll a to b has
frequently been used as an indicator of
degree of separation of the photosystems.
The lower the ratio, the greater the
amount of photosystem 2. Mechanical
breakage in a high-salt buffer produced
a Fr 2 with more chlorophyll b and a
Fr 1 with less. The further treatment of
Fr 2 with digitonin produced a second
fraction 2 particle (Fr 2g) with a still
lower ratio of chlorophyll a to b.
System 2 activity decreased upon
breaking the chloroplasts, with treatment
by digitonin and with time after such
treatments. Therefore, an enrichment of
system 2 activity in Fr 2 has not been
observed. As discussed in another sec-
tion of this report, chloroplast particles
were considerably more active in initially
reducing DCIP with Mn++ as electron
donor than with H20. With some frac-
tions the reduction with Mn++ was more
resistant to the damaging effects of the
treatment procedure than with DPC
(diphenylcarbohydrazide) as electron
donor. Since Mn++ and DPC may react
TABLE 10. Photochemical Activities and Chlorophyll a/b Ratios of Spinach
and Dunaliella Chloroplast Fractions
Photosystem 2 Activity*
DPC-
^DCIP
DCMU-
Photosystem
Ratio,
-DCMU
sensitive
1 Activityf
Fraction
Chi a/b
Cyt c oxid.
(jumoles(mg chl)-1
•hr"1)
Spinach homogenate
3.1
32
29
Fr 2, grana
2.7
43
35
23
Fr 1, stroma
8.1
3
1
53
Fr 1 + digitoninj
0
104
Fr lg, grana
4.6
6
3
174
Fr 2g, grana
2.2
38
35
10
Dunaliella homogenate
4.5
Mn+WDCIP
23
70
Fr 2
3.3
11
23
57
Fr 2g
2.0
14
15
13
Frig
3.8
3
13
348
Fr 1
8.0
6
6
174
* The photosystem 2 assay medium contained: 20 fiM DCIP, 400 ixM DPC
(or 16 ml MnCl2), 12 fxM DCMU when present, and 8 ng chlorophyll ml"1
in 50 ml Tricine (pH 7.5) in a total volume of 2.5 ml.
t The photosystem 1 assay medium contained: 40 to 60 nM reduced horse-
heart cytochrome c, 8 juM DCMU, 40 /xM methyl viologen, 4 ml KCN,
approx. 2 nM plastocyanin, 50 ml Tricine (pH 7.5), and 8 jug chlorophyll
ml-1, in a total volume of 2.5 ml.
% Conditions are described in the text.
DEPARTMENT OF PLANT BIOLOGY
355
at different sites, further experiments
with these two donors and different frac-
tions might be worthwhile.
Fr 2g from spinach showed good photo-
system 2 activity with DPC as the donor.
DCIP reduction was inhibited about
95% by 12 fxM DCMU (3-(3,4-dichloro-
phenyl)-l,l-dimethylurea). The rate of
DCIP reduction by Fr 2g was compar-
able to that of the needle-valve homog-
enate as shown in Table 10. Fr lg
showed some photoreduction of DCIP
with DPC as donor but this reaction was
relatively less sensitive to DCMU. Fr 2g
from either spinach or Dunaliella showed
very little photosystem 1 activity as
measured by cytochrome c oxidation.
The activity of Dunaliella prepara-
tions varied considerably from one ex-
periment to another. The particles in the
600
650
Wavelength, nm
700 600
650
700
I , i u i i
u
<U
O
c
o
_Q
u
O
</>
_Q
<
50 145 140 170 165 160
Wavenumber.cm"1 XIO"3
140
Fig. 14. Curve analyses of the absorption spectra of fragments from spinach stroma and grana
membranes measured at — 196°C. The error of fit at each point is shown on a scale below each
curve with the designated magnification; the higher the magnification, the better the fit. The
data from these analyses are given in Table 11.
356
CARNEGIE INSTITUTION
particular experiment shown in Table 10
happened to have about half the activity
for DCIP reduction that was seen in
some other preparations, but the relative
values were similar to fractions from
spinach. Why the rate of reduction of
DCIP with Mn++ should be the same in
Fr lg and Fr 2g from Dunaliella is not
apparent to us, but it was noted in sev-
eral experiments. This particular reac-
tion was not measured with spinach frac-
tions.
The oxidation of cytochrome c by
photosystem 1 is stimulated by deter-
gents. Probably this stimulation results
from an action which opens up the par-
ticles to allow exogenous plastocyanin to
act more efficiently as an electron carrier
between cytochrome c and P700, the re-
action center for system 1 (Year Book
70, p. 499, 1971 ; Fork and Murata, 1971) .
Incubation of digitonin with either Fr 2
or Fr 1 increased the rate of cytochrome
c oxidation by 2 to 2.5 times. This
digitonin-stimulated increase in system 1
activity must be distinguished from the
real enrichment in system 1 due to the
fractionation itself. The system 1 activ-
600
650
Wavelength, nm
700 600
650
700
170 165 160 155 150 145 170 165 160 155 150 145
Wavenumber.cnrf1 XIO"3
Fig. 15. Curve analysis of Dunaliella fractions. Same conditions as Fig. 14.
DEPARTMENT OF PLANT BIOLOGY 357
ity of Fr 1 is about the same as that of chlorophyll forms as Fr 1 but in different
Fr lg when the effect of digitonin is taken proportions in both spinach and Duna-
into consideration. liella. Perhaps these small differences in
The low-temperature absorption spec- absorption follow from differences in
tra of the spinach chloroplast fractions structure that are sufficient to explain
are shown in Fig. 14 and of the Duna- why Arntzen et al. (1972) could not re-
liella fractions in Fig. 15. The spectra combine stroma Fr 1 with grana Fr 2g.
have been matched with component The most unlike fractions in both ac-
curves. The standard error for each curve tivity and absorption were Fr 2g and Fr
resolution, the percentage of total area 1. Table 12 shows that Fr 2g had about
of each Gaussian component curve, and 30% more total absorption at shorter
the bandwidth (in parentheses) are given wavelengths than Fr 1. If only chloro-
in Table 11. phyll a is considered, this difference was
In general, the spectral components in about 20%. It also can be calculated
Frs 1 and 2 are similar to those seen from Table 11 that Fr 2g had about 20%
previously in spectra of fractionated leaf less of the long wavelength chlorophyll a
chloroplasts and algae (French et al., components.
Year Books 67 to 72) . The spectra of the Spinach Fr 1 showed very little system
digitonin subfractions have not been ana- 2 activity whereas Fr 2g gave 38 ^moles
lyzed before (curves C and D of Figs. 14 (mg chl)_1-hr_1. In Dunaliella, Fr 1 had
and 15) although some of the results a system 2 activity of 6 /xmoles (mg
with spinach are in Gasanov and French chl)_1,hr_1 and Fr 2g, of 14. We can con-
(1973). Fr 1 (Fig. 14B) from spinach elude that greater system 2 activity is
absorbs more at longer wavelengths than correlated with more absorption by
the corresponding Fr 1 (curve 15B) from chlorophyll b and the shorter wavelength
Dunaliella. Spinach Fr lg (Fig. 14D) forms of chlorophyll a, but an exact
also has a 700 nm component not seen in quantitative relationship is not apparent.
Dunaliella Fr Id (Fig. 15D) . The striking difference between these
The longest wavelength component in two fractions with regard to system 1 is
each spectrum is always the broadest the lack of Ca692 in Fr 2g. This lack is
(Table 11). Thus Ca684 in Fr 2g has a correlated with almost no activity for
much greater bandwidth than in the other cyt c oxidation. Therefore, the chloro-
fractions that have bands beyond 690 phyll component Ca692 may be essential
nm. Perhaps this broadening indicates for photosystem 1.
that Ca684 in system 2 is in fact more An interesting feature of the general
aggregated or polymerized than Ca684 problem of comparing spectra by curve
in system 1. On the other hand, it may fitting can be illustrated by a comparison
be a trivial result of the curve fitting of Figs. 14C and 15C. The standard
program that is attempting to match a errors (Table 11) of resolution of these
long wavelength part of the curve with two curves are higher than for the other
Gaussian-shaped components that may curves in spite of attempts to reduce this
not fit because of a distortion by the error by varying several parameters (e.g.,
overlap of the tail of a large, but not adding a 692 nm component) . The two
strictly Gaussian, band whose center is at curves are very different from each other,
shorter wavelengths. especially near the absorption maximum,
The spectrum of the system 1 fraction and yet they were best resolved with
(Fr lg or d) from Fr 2 is not shifted to- nearly the same components. The most
ward longer wavelengths compared to Fr significant difference between them is
2 or Fr 2g, as is usually observed in frac- that in spinach Ca677 had a half-band-
tions that have high system 1 activity, width of 11 nm and in Dunaliella Ca679
However, Fr lg has nearly the same had a width of 8 nm. It appears that this
358
CARNEGIE INSTITUTION
c3
CD
<
CO
b/0
o3
+3
a
CD
p
CD
Ph
o3 _
^ X
co ft
3 ,2
c3
w
a
CO
O
ft
a
o
u
T3 o
d e3
+3 "
CO *+h
o °
X
H
pq
<
o
^
+
o
o
©
CD
I
©
00
CO
oo
CO
I
CO
00
CO
©
CO
I
l>-
t^
CO
CO
I
©
CO
CO
<N
CO
CO
I
o
CO
CO
X
o
^
w
CO
I
©
CO
CO
I
©
CO
CO
a
ft
^
CD
>
o
CO <N
CN CO
OS h OO
CN i— I i— I
(NOMN
HCOOO
T-H 1 — I i—(
QiOOIN
© i— i O t-h
© t-h CO CO
CO "tf CO CO
© © T-H <N
© T-H T— I ©
CO CO CO CN
CN
t-H CO t-H TH
CN <N CN t-h
hMN^
CN CN t-h t-h
Tji
NCOhn
t^ I> 1> ©
(N
CO iO
GO t^ © CO
CN CN (N CO
mO t-H -hh T*
CO iO CO CN
©©'©'©
1> © © ©
T-H t-H T-H i— I
T*H © |> OS i>
CO 00 CN CN CN
tJH 00 b- o ©
t-H t-H (N t-H
z~s^-^ © t-H t-H
© 00 t-H t-H t-H
t* 00 © CO I>-
CN CN CN <N CN
© t-H © t-H t-H
© CO iO rfi iO
CN CO (N CN CN
(N <N CN CN CN
© t-H © t-H (N
CN CO CN CN CN
00 CN tH <N t-h
T— I CN T-H T-H T-H
t-h CN CO CO
iO O iO00 ii5
oo oo oo oo oo
© CN t-h CN t-h
iO O iO <N iO
T— I CN T-H T-H T-H
-hh CO © CN ©
CO ^ CO CO CN
o © © © ©
*0
CN
<JOQffl
CJ CN CN t-H t-H
03
,-H f_| ^ *H S-l
.9 Ph Pl, Ph. Ph
CO
KS +
."» bfit3
r-2 CN CN T-H T— I t— i
2 S_ >H tH Sh Sh
§ Ph Ph fe Ph Ph
o
iO
1— I
c
b£
H-3
c
CD
ft
>
c3
Q
>
a
CD
CD
S3
o
Oh
T3
a
o3
X
ft
o
X
a
CD
a
o
ft
£
o
o
CD
X
<=> erg
°o "si T3
CD *£ k
tIBtS
03 s a
O C ^
g § 2
(H CO
CD H CD
§ i «
T! 2 ft
CD G _,
+3 O M
^.J —
o o3 jj
° CD s
co -£ <;
ft
DEPARTMENT OF PLANT BIOLOGY
359
TABLE 12. A Comparison Between the
Amounts of Chlorophyll b and the
Shorter Wavelength Forms of
Chlorophyll a in Fr 2g and Fr 1
Percentage Difference between Fr 2g and Fr 1
Fr 2g had more Spinach Dunaliella
Chlorophyll b by 9 10
Ca662 by 9 10
Ca670 by 2 9
Co678 by 8 2
Total 28 31
narrowing in width and shift of 2 nm in
peak position makes possible the fitting
of the very sharp Dunaliella spectrum.
As mentioned above, spectra of algae
usually have sharper peaks and shoulders
than spectra of leaf chloroplast particles.
Also, it has been noted before that sev-
eral of the chlorophyll forms are broader
in system 1 fractions than in system 2
fractions. Perhaps the chlorophyll forms
can be modified in some small way that
results in their having different band-
widths in different chloroplast environ-
ments.
References
Arnon, D. I., Plant Physiol, 24, 1-15, 1949.
Arntzen, C. J., R. A. Dilley, G. A. Peters, and
E. R. Shaw, Biochim. Biophys. Acta, 256,
85-107, 1972.
Ben-Amotz, A., and M. Avron, Plant Physiol.,
49, 240-243, 1972.
Brown, J. S., Ann. Rev. Plant Physiol., 23,
73-86, 1972.
Fork, D. C, and N. Murata, Photochem.
Photobiol, 13, 33-44, 1971.
Gasanov, R., and C. S. French, Proc. Nat.
Acad. ScL, 70, 2082-2085, 1973.
Jacobi, G., and G. Lehmann, In Progress in
Photosynthesis Research, H. Metzner, ed.,
Vol. 1, 162-178, Munich, Goldman-Verlag,
1969.
Jones, R. F., H. L. Speer, and W. Kury,
Physiol. Plantarum, 16, 636-643, 1963.
Ohki, R., and A. Takamiya, Biochim. Bio-
phys. Acta, 197, 240-249, 1970.
Sane, P. V., D. J. Goodchild, and R. B. Park,
Biochim. Biophys. Acta, 216, 162-178, 1970.
Starr, R. C, Amer. J. Bot., 51, 1013-1044,
1964.
Wegmann, K., Biochim Biophys. Acta, 234,
317-323, 1971.
THE
EFFECT
DCIP
OF MANGANESE
REDUCTION
/. S. Brown
ON
Considerable work has been done on
the role of manganese in photosynthesis.
Papers up to 1969 were reviewed by
Cheniae (1970). He distinguished be-
tween the role of Mn++ as an exogenous
electron donor to system 2 or as a bound
chloroplast enzyme that specifically cata-
lyzes high rates of oxygen evolution.
In this work some observations were
made on the role that manganese plays
in the system 2 reactions of chloroplast
particles. It was found that a consider-
able increase in the initial rate of DCIP
(dichlorophenolindophenol) photoreduc-
tion occurred when millimolar quantities
of Mn++ were added to the reaction mix-
ture. This stimulation was apparently
not related to manganese bound as an
enzyme because the algae used wrere not
deficient in manganese during their
growth. Moreover, about three orders of
magnitude higher concentrations of man-
ganese were required to bring about the
stimulation of the DCIP-Hill reaction
than to supply the manganese needed for
synthesis of the enzyme.
DCIP photoreduction was measured
(Year Book 70, p. 500) as the increase in
transmission at 600 nm, using an extinc-
tion coefficient for DCIP of 20,000 M"1
cm-1. See Brown, Gasanov, and French,
this volume, for details about the chloro-
plast particles and the reaction proce-
dures.
360
CARNEGIE INSTITUTION
Fig. 16. The time course of DCIP reduction
by Fr 2 from Dunaliella with no additions
(HaO), 400 fiM DPC, 16 ml MnCl2, or both
DPC and MnCh. The cuvette also contained
20 fiM DCIP, 150 ml KC1, 50 mM Tricine
(pH 7.5), and 8 fig chl ml-1 in 2.5 ml. Each
sample was exposed to saturating red light for
22 sec. Arrows indicate light on and light off.
The numbers on the curves are the rates of
reduction in /mioles mg-1 chl hr"1.
Figure 16 shows the time course of
DCIP reduction without additions, with
DCP (diphenylcarbohydrazide), with
MnCl2, or with both DPC and MnCl2
in the reaction mixture. The increased
rate with Mn++ lasted for about 10 sec,
whereupon it gradually decreased to a
value lower than without Mn++. DCIP
was not immediately reoxidized in the
dark when H20 or DPC served as elec-
tron donors, whereas with Mn++ the back-
reaction began in the light and continued
when the light was turned off.
The initial rates of DCIP reduction
with homogenates of both Dunaliella and
spinach chloroplasts showed a linear de-
pendence on light at low intensities with
or without 20 ml MnCl2, but the slope
with Mn++ was about double that without
Mn++.
A small increase in DCIP reduction
could be observed with 2 ml MnCl2,
and the effect saturated at 16 ml MnCl2.
The highest concentration tested (40
mM) had no further effect.
Magnesium chloride added under the
same conditions as MnCl2 did not have
an effect in these preparations, indicating
that the stimulation was not produced by
a cation effect (Murata, 1969). Ribo-
flavin or NaCN had no effect on the
stimulation of DCIP reduction by
MnCl2, but 8 fiM DCMU inhibited the
reduction completely.
The effect of Mn++ on 02 evolution
with DCIP as electron acceptor by a
spinach chloroplast homogenate was de-
termined during the first 20 sec of illumi-
nation with an 02 electrode according to
Walker et al. (1968). The results pre-
sented in Table 13 show that Mn++ in-
hibited 02 evolution by about the same
amount as it stimulated the initial rate
of DCIP reduction.
These results are similar to those of
Itoh et al. (1969), Lozier et al. (1971),
and Ben-Hayyim and Avron (1970) . The
Japanese workers observed the reactiva-
tion of DCIP reduction and also of
NADP+ reduction (Yamashita et al.,
1969) by 0.1 mM Mn++ in chloroplasts
which had been partially inactivated by
Tris-washing. Since Mn++ is removed
from chloroplasts by Tris-washing, they
concluded that added Mn++ returns to its
original site and functions as an electron
carrier between H20 and the reducing
side of system 2.
The results reported here can be ex-
plained if it is assumed that Mn++ func-
tions in competition with or as a replace-
ment for water as an electron donor to
system 2 in the French-press treated
chloroplast particles. The curves in Fig.
16 suggest that after about 10 sec of
illumination, the Mn+++ formed begins to
TABLE 13. Light-Saturated Initial Rates of
DCIP Reduction and 02 Evolution by Spinach
Chloroplasts and Homogenate
DCIP
reduction 02 evolution
/xmoles (mg chl)_1hr-1
Chloroplasts
80 59
Homogenate
37 21
Homogenate +
20mM MnCl2
48 11
DEPARTMENT OF PLANT BIOLOGY
361
oxidize the reduced DCIP in a cyclic re-
action. The data in Table 13 show that
this cyclic reaction may compete with
the water-splitting reaction resulting in a
decreased rate of 02 evolution.
Usually the rate of DCIP reduction
was greater with both DPC and Mn++
than with saturating amounts of either
donor alone. In some experiments DPC
did not give a higher rate of reduction
than H20, whereas Mn++ was always a
better initial donor than H20. That these
two donors may react at different sites on
the oxidizing side of system 2 was also
recently suggested by Shneyour (1973).
He found that DPC reacted at a different
site than that inhibited by Tris-washing
in a mutant of Euglena.
References
Ben-Hayyim, G., and M. Avron, Biochim.
Biophys. Acta, 205, 86-94, 1970.
Cheniae, G. M., Ann. Rev. Plant Physiol.,
21, 467-498, 1970.
Itoh, M., K. Yamashita, T. Nishi, K. Konishi,
and K. Shibata, Biochim. Biophys. Acta.
180, 509-519, 1969.
Lozier, R., M. Baginsky, and W. L. Butler,
Photochem. Photobiol., 14, 323-328, 1971.
Murata, N., Biochim. Biophys. Acta, 189,
171-181, 1969.
Shneyour, A., Biochem. Biophys. Res. Com-
mun., 51, 391-398, 1973.
Walker, D. A., C. W. Baldry, and W. Cock-
burn, Plant Physiol, 43. 1419-1422, 1968.
Yamashita, K., M. Itoh, and K. Shibata, Bio-
chim. Biophys. Acta, 189, 133-135, 1969.
RATE MEASURING CIRCUIT FOR
IMPROVED ACTION SPECTRA
W. G. Hagar, G. A. Ford, and C. S. French
We have continued to improve the
methods for measuring the activities of
the two photosystems based on the oxida-
tion of cytochrome c for photosystem 1
and the reduction of dichlorophenolindo-
phenol (DCIP) for photosystem 2.
Modifications of the action spectrum sys-
tem and measurement procedures de-
scribed previously (Year Book 69, p.
670) , as well as new rate measuring elec-
tronic circuitry, are presently being
tested with a phototransformable chloro-
phyll protein (CP 668) (Yakushiji et al.,
1963).
The basic action spectrum instrumen-
tation and electronics are similar to the
previously reported system. In addition
to the rate measuring circuit to be de-
scribed, two minor modifications also
warrant reporting. A Bausch and Lomb
monochromator capable of higher mono-
chromatic intensities is now employed as
the actinic source, and the samples are
kept at a constant temperature by a
thermoelectric cooler.
We have returned to a point by point
procedure that is expected to minimize
any corrections based on interpolated
values, such as changes in activity due
to pretreatment and back reactions. The
method is based on a single sample in
which a repeated comparison of rates is
made between the measured wavelength
and a reference wavelength. Usually a
reference measurement comes immedi-
ately before and after each spectral
measurement. In this way we are not
dependent upon an external reference,
and any changes in activity and in back
reactions effectively account for them-
selves. The need to have the reference
measurements as close in time as possible
to the spectral measurement prompted
the development of electronics to provide
us with an immediate rate measurement.
To determine action spectra we must
measure the rate of change in concentra-
tion of a chloroplast component or of
added dyes. The rate of change in con-
centration is directly proportional to the
362
CARNEGIE INSTITUTION
change in transmission divided by the
total transmission at that time. Thus,
the rate is normally obtained by deter-
mining the transmission and its rate of
change from a recorder tracing of the
transmission signal. Analyzing recorder
tracing is a tedious, time consuming job.
The length of time required to obtain a
precise rate measurement depends on the
rate of the signal change with time which
is not the same for all wavelengths of
light. If the output signal is noisy, a line
may be drawn through the recorder trac-
ing and the slope of this "averaged line"
used as a measure of the rate. Our ap-
plication required a circuit which would
produce precise rate measurements di-
rectly. Standard analog differentiation
circuits, even when filtered, still produce
low signal to noise ratios when the sig-
nals are small and changing slowly, so
they were not satisfactory with our sys-
tem.
Consequently, to measure the small,
slow changes precisely, we developed an
electronic circuit which, when triggered,
continuously produces a signal of the
voltage change per unit time. The sam-
pled increment becomes larger and larger
with time until the required precision of
rate determination is obtained. Initially
the signal to noise ratio is so low that
there are variations of the output signal
about the true rate value; but as the
sampled increment becomes larger, the
noise is averaged out and a precise value
for the change in input signal with time
(rate) is obtained. The time required for
the measurement is dependent entirely
upon the signal change to noise ratio.
Rate Circuit Description
The rate measuring system can be
broken down into three separate circuits
(Fig. 17): (1) An initial sampling and
holding of the signal at time zero and a
continuous measurement of the differ-
ence between the held value and the
present value. (2) An internal timing
circuit which generates a continuous time
signal. (3) An analog divider which pro-
vides a ratio of these two signals.
Sample and hold
Voltage
follower
Input
signal S2
Calibration
Integrator
Recorder
IM =
Timer Integrator
Fig. 17. Diagram of rate measuring circuit. An operational amplifier manifold, Analog Devices
194, is used as the power supply and frame. Amplifiers 1, 2, and 3 are FET operational amplifiers,
Analog Devices (AD) 42 J. Numbers 4, 5. and 7 are chopper stabilized operational amplifiers,
AD233J. Amplifier 6 is also a chopper stabilized operational amplifier, AD233K. Donner pre-
cision polystyrene capacitors are used to prevent leakage. For measurements in the 80 microvolt
per second range the resistor values are as follows: Rl is 5 M; R2, R4, and R5 are 1 M; and R3
is 2 M. The divider (multiplier) is Analog Devices 427K.
DEPARTMENT OF PLANT BIOLOGY
363
Referring to the circuit diagram, the
voltage follower is used to offer high
input impedance to the measured circuit.
The sample and hold circuit, when trig-
gered, retains the initial signal for up to
several minutes. The held signal and the
continuously changing signal are con-
nected in a difference circuit which sub-
tracts one from the other. An integrator
is added to average out the random sig-
nal noise of the continuous voltage meas-
urement. The final output from this inte-
grator is the area under a curve of the
signal change versus relative time and is
constantly increasing.
The timer circuit is shown below the
voltage measuring circuit. An opera-
tional amplifier is used to integrate a
fixed voltage, thereby producing a volt-
age increasing linearly with time. An-
other integrator is added after the timer
to produce an identical area measurement
versus time like that from the signal
monitoring circuit. A single multipolar
switch controls the start of the timer and
the holding of the initial signal. The in-
tegrators after the continuous signal and
timer circuits reduce the error in the con-
tinually changing signal and shorten the
time required for the rate measurement.
However, an initial error in sampling of
the sample and hold circuit will still
affect the rate results until the sampled
increment is large enough to minimize
the importance of the initial error.
The ratio of the signal change integral
to the internal timer integral is obtained
by an analog divider. These integrated
signals are appropriately amplified be-
fore the divider to keep them in the opti-
mum division range of 2 to 10 volts.
Calibration
Inherent in the system is the capability
of internal calibration. A switch is used
to disconnect the input signal and to
connect the internal timer to the voltage
follower. This allows precise internal
calibration of the entire system, using
various changes in the amplification of
the voltage follower, integrator, and final
amplifier to produce a linear calibration
of various full-scale sensitivities. This
permits amplification changes in the mid-
dle of the experiment to increase the pre-
cision of measurements. The absolute
value of voltage change per time period
can be easily obtained from the relative
measurements.
Precision of the rate measuring circuit
was determined using the internal cali-
bration. This determination of a known
linear voltage change with time of 150
microvolts per second was reproducible
within less than 0.3% standard devia-
tion. A test of the entire system was
made by monitoring the phototransfor-
mation of chlorophyll protein 668 ex-
tracted from Chenop odium album. These
measurements have been in the 80 micro-
volts per second range.
System Test Measurement
Chlorophyll protein 668 is an ideal
substance to use in determining the pre-
cision of an action spectrum measuring
system. We are indebted to Professor
Atusi Takamiya of Toho University who
gave us some purified material and
showed us how to work with it. In the
presence of oxygen and light of the
proper wavelength, it is transformed into
a form which has an absorption maxi-
mum at 743 nm. Consequently, the rate
of transformation can be measured by
monitoring the increase in absorption at
743 nm. The rate of the transformation
at low light intensities is dependent on
the wavelength of light, on light inten-
sity, and on the concentration of the
chlorophyll protein. When the wave-
length and intensity of the exciting light-
are held constant, the rate is dependent
on the concentration of the chlorophyll
protein 668 remaining and follows first
order kinetics.
d [CP«68]
dt
— — A.'i [CPg68J
364
CARNEGIE INSTITUTION
.30
u
C
O
"5
£
o
.22
Fig. 18. A comparison of two methods of rate determination. The open circles are from the
rate circuit measurements, and the closed circles are from the eye-averaging method of measure-
ment. The data from the eye-averaging method were adjusted to give the same slope and inter-
cept as the data from least-squares analysis of the rate circuit measurements. The slope and inter-
cept of the line is —2.27 and 0.471. The correlation coefficient of the least-squares analysis was
-.994.
By plotting the rate of transformation
versus concentration, a linear relation-
ship is obtained. An analysis by the
method of least squares was used to
determine the slope and intercept of the
line, and also as a means of comparing
the new and the old methods of rate de-
termination. Figure 18 is a comparison
plot of rates of reaction obtained from
both the manual "eye averaging" tech-
nique and the rate measuring circuit
versus the increase in concentration of
the transformed chlorophyll protein 743.
Both methods gave approximately the
same slope and intercept and had almost
the same degree of precision. The rate
measuring circuit had standard devia-
tions for slope and intercept determina-
tions of 2.85% and 2.10%, respectively.
The manual "eye averaging" technique
had slightly higher values of 3.20% and
2.35%. Either method could have been
used for the slope analysis ; however, the
length of time used for each measure-
ment was decided on the basis of data
from the rate measurement circuit. If
the rate measuring circuit had not been
used, the length of time required to sam-
ple enough of the signal change to reduce
errors in measurement would not have
been known. Future benefits of the rate
circuit are foreseen in action spectra
measurements made by altering the light
intensity to produce equal rates of reac-
tions at the reference wavelength and the
monitored wavelength, since it gives rate
measurements immediately rather than
by later analysis.
Reference
Yakushiji, E., K. Uchino, Y. Sugimura, I.
Shiratori, and F. Takamiya, Biochim. Bio-
phys. Acta, 75, 293-298, 1963.
DEPARTMENT OF PLANT BIOLOGY
365
EFFECT OF BLUE LIGHT ON RESPIRATION
AN ACTION SPECTRUM FOR THE
PHOTOINHIBITION BY BLUE LIGHT
Helga Ninnemann and C. Stacy French
Photoinhibition by blue light of exo-
genous respiration in colorless organisms
is due to a destruction of cytochrome
oxidase: i.e., cytochromes a and a-A in
yeast (Ninnemann, Butler, and Epel,
1970a), and cytochrome a3 in Prototheca
(Epel and Butler, 1969, 1970) and beef
heart mitochondria (Ninnemann, Butler,
and Epel, 1970b) . These workers have
also shown that the Soret band of the
oxidized form of cytochrome oxidase is
the light-absorbing pigment.
In addition to the pronounced light
effect on cytochrome oxidase, other cyto-
chromes of the respiratory chain seem to
be affected by blue light, though to a
lesser extent. Absorption of blue light
causes a cytochrome b decrease in yeast,
beef heart mitochondria, and Prototheca
(Ninnemann, Butler, and Epel, 1970a
and b; Epel and Butler, 1970) , and cyto-
chrome c-551 is destroyed in Prototheca
(Epel and Butler, 1970).
An earlier paper presented a coarse
action spectrum and showed only blue
light to be effective in photoinhibition of
respiration (Ninnemann, Butler, and
Epel, 1970a), but for technical reasons no
detailed spectrum in the blue wavelength
range could be accomplished. It was the
purpose of the experiments presented in
this report to measure an action spectrum
for the photoinhibition of exogenous res-
piration in yeast in the range of 395-470
nm.
Materials and Methods
Commercial baker's yeast, Saccharo-
myces cerevisiae, was starved in dark-
ness for at least 18 hr in 50 ml KH2P04
with air bubbled through the culture.
The stock suspension with 2 X 109 cells
ml-1 was diluted 1:100 (density reading
at 587 nm 0.36 O.D.) . Starved cells were
irradiated at room temperature with
simultaneous aeration in a 1.2 ml Lucite
cuvette (8 mm diameter) , the backside of
which was covered by aluminum foil to
have as little loss of incident light as
possible. They were transferred, after
addition of ethanol (0.2% final cone),
into a second (dark) cuvette in a water
bath at 31 °C for measurement of 02-
uptake with a Clark-type 02 electrode
(Yellow Springs Instruments, Model
4004).
High energy light sources were a water-
cooled high pressure mercury lamp (type
A-l-B, Illumination Industries, Sunny-
vale, Cal.) and a 2500 W xenon lamp.
Both lamps were used in combination
with two glass lenses, a CuS04 bath
(15%, 38 mm thick), Balzer's interfer-
ence filters (8-14 nm half-bandwidth),
Schott KV-cutoff filters No. 370-418 for
the uv and near uv region and a Balzer
Bl/Kl Calflex infrared filter. Occasion-
ally a broad-band blue Corning filter no.
5562 was used.
Light intensities were measured with a
Kettering YSI radiometer and with a
calibrated thermopile in combination
with a Hewlett-Packard voltmeter. The
cells were irradiated with equal numbers
of quanta of blue light of 395-471 nm,
corresponding to 6 X 105 ergs cm"2 sec-1
at 433 nm.
Absorption spectra of highly scattering
samples were recorded in Dr. W. L. But-
ler's laboratory at the University of
California, San Diego, with a Cary 14
spectrophotometer on line with a PDP8/I
computer.
Exogenous Respiration
Relatively short irradiations (5-60
min) of intense broad-band blue light
(10G ergs cm-2 sec-1) as well as long-term
irradiation (several hours) with moder-
366
CARNEGIE INSTITUTION
ate intensities of white (1.5 X 104 ergs
cm-2 sec-1) or blue light (3 X 103 ergs
cm-2 sec-1) result in an inhibition of exo-
genous respiration and a decrease of ab-
sorption in the a-band of cytochrome
a/as in yeast (Ninnemann, Butler, and
Epel, 1970a). The viability of the cells
is not affected by these treatments. A
comparison of the effect of relatively
high intensity broad-band blue light
(Corning filter 5562, 106 ergs cm-2 sec-1)
with intense monochromatic blue light at
435 nm (Balzer's interference filter 435,
3 X 105 ergs cm-2 sec"1) showed that
even with monochromatic blue light, high
435nm int. filter
3XI0^ergs cnT^sec'
0
0
T
30
ime , min
60
Fig. 19. Oxygen consumption of yeast irradi-
ated for 0, 30, and 60 min, 2 X 107 cells ml"1
Room temperature. Substrate after irradiation:
0.2% ethanol.
intensities are required to inhibit the
respiratory capacity (Fig. 19) . This find-
ing pointed to a broad-banded action
spectrum in the blue wavelength range.
The requirement for relatively high in-
tensities of monochromatic light excluded
the use of a monochromator in the fol-
lowing experiments. Thus, yeast had to
be irradiated aerobically for 30 min with
monochromatic blue light obtained
through interference filters. At one time
a more light-sensitive strain was found
for which only 10 min irradiation with
403 nm was required for a 50% inhibi-
tion of respiration.
Fig. 20A and B show the result of irra-
diations of 10 min and 30 min, respec-
tively, with wavelengths between 395 and
471 nm. Each recorded point is the mean
value from 4 to 10 measurements; bars
indicate the standard error computed
according to
S.E. = l/SpE-m)2
(/ n (AT— 1)
where S.E. is standard error; iV" is num-
ber of measurements; m = mean value
of x.
The broad-peaked action spectrum
with a maximum near 427 nm follows
essentially the Soret band of oxidized
cytochrome oxidase. The Soret peak of
oxidized cytochrome oxidase of animals,
higher plants, yeast, and Neurospora is
generally reported in the literature at
425 nm (Mahler and Cordes, 1966).
There is some indication in the action
spectra that wavelengths from 395 to
411 nm are just as inhibitory as 427 nm
and more so than light at 416-419 nm.
This observation might indicate the par-
ticipation of 6-type cytochromes in the
absorption of destructive light. An ab-
sorption spectrum of whole yeast in an
oxidized state is shown in Fig. 21. The
peak at 413 nm is due to oxidized cyto-
chromes 6.
Measurements of activity and absorp-
tion of isolated complexes II and III of
heavy beef heart mitochondria, both of
DEPARTMENT OF PLANT BIOLOGY
1 1 1
367
50
40
30
20
_ a
c
0)
u
u
<v
a
10
c
o
■4-
J3
0
In
c
1
40
o
-♦—
o
_c
a
30
20
10
0
lOmin exposure
30min exposure
400
450
Wavelength.nm
Fig. 20. Action spectra for photoinhibition of exogenous respiration in yeast. 2 X 107 cells ml"1.
Irradiated with equal numbers of quanta, corresponding to 6 X 105 ergs cm"2 sec"1 at 433 nm.
02 measurements at 31 °C. Bars indicate standard error of 4-10 measurements.
which contain 6-cytochromes, show a
decrease in both parameters after irra-
diation with blue light (detailed results
to be published) .
0.04
0.02
400
500
Wavelength, nm
600
Fig. 21. Absorption spectrum of yeast aerated
in the dark without substrate, 0.3 ml of suspen-
sion with 2 X 109 cells ml"1. Room temperature.
Endogenous Respiration
Endogenous respiration in some strains
of yeast is stimulated by blue light of
high intensity. A 15 min irradiation with
monochromatic blue light under aerobic
conditions gave maximal stimulation.
However, the presence of 02 during the
irradiation resulted in a secondary respi-
ration decrease. Under anaerobiosis the
maximally stimulated respiratory rate
persisted even after prolonged illumina-
tion (Ninnemann, Butler, and Epel,
1970a). An action spectrum was re-
corded from 401 to 440 nm. Then we
unfortunately lost the yeast which could
be appreciably stimulated by blue light.
368
CARNEGIE INSTITUTION
The results obtained so far show an
increasing effectiveness of the longer blue
wavelengths with 440 nm light resulting
in a stimulation up to about 10-fold.
This observation agrees with action spec-
tra measurements on a yellow mutant of
Chlorella (Kowallik and Gaffron, 1967;
Kowallik, 1967) which showed increasing
stimulation of endogenous respiration
with longer wavelength blue light and a
maximal effect of 460 nm light.
Endogenous respiration of yeast was
not inhibited by most inhibitors of exo-
genous respiration like atebrin, amytal,
or antimycin A. Amytal (3.4 X 10"2 M)
increased the endogenous respiration 4-
fold. High concentrations of cyanide (3
X 10~2 M) and azide (3 X 10"3 M) ,
however, inhibited endogenous respira-
tion completely. Moderate concentra-
tions of thenoyltrifluoroacetone (4 X 10~4
M), which blocks electron transport be-
tween succinate dehydrogenase and co-
enzyme Q, stimulated, but high concen-
trations (10~2 M) partially inhibited,
endogenous respiration. These results
suggest that electron transport from
endogenous substrates to 02 does not
proceed via the respiratory chain from
NADH to cytochrome c; however, com-
plex II (SDH-cytochrome c reductase)
might participate in endogenous respira-
tion.
References
Epel, B., and W. L. Butler, Science, 166, 621-
622, 1969.
Epel, B., and W. L. Butler, Plant Physiol.,
45, 728-734, 1970.
Kowallik, W., Plant Physiol, 42, 672-676,
1967.
Kowallik, W., and H. Gaffron, Nature, 215,
1038-1040, 1967.
Mahler, H. R., and E. H. Cordes, Biological
Chemistry, p. 594, 1966.
Ninnemann, H., W. L. Butler, and B. Epel,
Biochim. Biophys. Acta, 205, 499-506,
1970a.
Ninnemann, H., W. L. Butler, and B. Epel,
Biochim. Biophys. Acta, 205, 507-512,
1970b.
SOME RECENT IMPROVEMENTS OF A
HIGH-SPEED DIFFERENCE
SPECTROPHOTOMETER
T. Hiyama and D. C. Fork
Last year, some modifications and ad-
ditions to the apparatus for the quantita-
tive study of rapid and small absorption
changes were described (Year Book 71,
p. 182) . Supported by a special fund
provided by the Institution, several fea-
tures have been added during the past
year in order to improve still further its
time resolution. Most of the construction
has been completed. Some interesting
results, which will be summarized in later
sections, have already come out.
We are now constructing another dif-
ference spectrophotometer also capable
of measuring the kinetics of very rapid
light-induced spectral changes, since we
anticipate increased use of both the pres-
ent apparatus and the proposed new one
by visiting investigators and students as
well as by ourselves. The new apparatus
will be like the present one in principle
but will be particularly well suited for
use with samples of subchloroplast par-
ticles. To define better the measuring
beam, a double grating monochromator
will be used and another monochromator
or appropriate filters will be used in front
of the photomultiplier.
A schematic diagram of how these set-
ups function is shown in Fig. 22. The
optics for measuring absorption changes
are essentially the same as reported last
year. The electronics, on the other hand,
have been substantially modified with
the introduction of a high-speed digital
data acquisition system (transient re-
DEPARTMENT OF PLANT BIOLOGY
•^Tungsten
amp
369
. >5
Measuring' r
light i Actinic
t / ngh+
s ,->
bample
High-speed difference spectrophotometer
and data acquisition
system
Osc. display Osc. display
Digital
printer
Preamplifier
Transient
recorder
X-Y plotter
Fig. 22. A schematic diagram of the spectrophotometer and high-speed data acquisition sys-
tem. Power supplies for a tungsten lamp, flash sources, and a phototube have been omitted. The
preamplifier includes an offset voltage circuit and a balance monitor.
corder, model 610B. Biomation Corp.,
Cupertino, Cal.). The photocurrent from
the photomultiplier is amplified by a
high-speed amplifier (instrumentation
amplifier, model FA 601, Intronics, Inc.,
Newton, Mass.), which has replaced the
existing preamplifier. The output of the
amplifier is recorded by the memory of
the Biomation at the rate of up to 0.1
/xsec per point with 6-bit resolution.
The recorded data are then either auto-
matically or manually transferred to a
signal averager (model 1010, Nicolet,
Madison, Wise.) at much lower rate
(typically 1 msec/point). The overall
time resolution has thus been improved
20 times over the previous set-up.
At present, the data are recorded on
chart paper and digitized manually. We
plan to feed the digital output from the
signal averager directly into a computer
for greater speed and accuracy (discussed
in a later section) .
To complement the improved time
resolution of the measuring system, we
have added a xenon flash lamp with an
appropriately short pulse duration. A
compact-arc xenon flash lamp (model
2CP-4, USSI, Watertown, Mass.) is fired
by a circuit with a high voltage pulser
and a storage capacitor. The maximum
energy input is 6 joules. The light en-
ergy output with a Balzers infrared cut-
off filter and a red glass filter is around
10-3 joules cm-2 sec"1, which is enough
to saturate the light reactions of most of
our samples. As shown in Fig. 23A, the
intensity reaches its maximum within 1
yusec and decays to almost 30% after 2
jtxsec.
For faster reactions, a recently devel-
oped organic dye laser has been added as
the excitation source. Substantial addi-
tions and modifications were made on the
original triggering circuit of a commer-
cial laser (Chromabeam 1050, Syner-
getics, Princeton, N. J.). The dye is
pumped by a coaxial xenon lamp. For
our purposes a solution of Rhodamine B
in ethanol was found to be the most
suitable among dyes presently available
because the output of the laser can be
370
CARNEGIE INSTITUTION
2/i/sec
Fig. 23. Output profile of flash lamp sources.
A, output of an ultra-short xenon flash; B, out-
put of a dye laser with Rhodamine B (10~4 M
in ethanol). The output was measured by a
photomultiplier (EMI 9558B) with filters and
recorded on a transient recorder (Biomation
610B).
tuned from 590 to 650 nm by changing
the angle of a rear reflecting grating. The
shape of the light pulse is shown in Fig.
23B. The rise time was less than 100
nsec with a decay of several hundred
nanoseconds.
The output energy ranged from 1 to
2 X 105 erg cm"2 at 647 nm, which in
most cases was enough to saturate photo-
reactions in green plant photosystems.
The laser beam was dispersed by a lens
to obtain a large enough area of illumi-
nation to cover the sample.
By proper shielding and grounding, the
electrical noise due to high voltage spikes
has been reduced to a tolerable level,
both with the xenon flash and the dye
laser.
The necessity and usefulness of signal
averaging are demonstrated in Fig. 24,
(D-144 of Anderson and Boardman,
1966) where small absorption changes at
443 nm in spinach photosystem 1 par-
ticles induced by a 3 /xsec flash are shown.
These extremely small and rapid absorp-
tion changes have been assigned to the
photoreduction and reoxidation of the
primary electron acceptor of photosys-
tem 1 (Hiyama, 1971; Year Book 71, p.
Absorption changes at 443nm induced by
3jusec flash in spinach subchloroplast particles
Control
I measurement
it-
AA
0.0001
Average
of 32 I
measurements \ ,w
f
I measurement
'T'".l El
i
M r »■»«
ili. .!.. Li] E
I measurement
irmm
iltliffii
JilJJlii^jj.
1 I
44
E
i M
it
I I
*- 50msec
iE -ir-+|E
ll'i"'!
m
i 4 n
-L-4-E,-
ill1!
Methyl
viologen
lOpM
Average
n»siuy« ▼ i - ■ j ■- - - - f ■ i -
of 16 J.'M ' *J^AjJm<^Mktfw thk-ukitk
Methyl
viologen
100/jM
Fig. 24. Absorption changes induced by a 3 ^sec flash at 443 nm in spinach photosystem 1
particles. The particles (D-144, Anderson and Boardman, 1966) were suspended in 0.2 M KC1,
0.05 M Tricine buffer (pH 7.8), with 100 /xM N JSl ,N\N1-tetramethyl-p-phenylenediamine, 2 ml
sodium ascorbate and methyl viologen of the given concentration. A short-pulse xenon flash
(see the text) with a red filter (Schott RG5, 3 mm) was used as an actinic light source.
DEPARTMENT OF PLANT BIOLOGY
371
182) . By comparing traces obtained
after only one light exposure to those
averaged for several exposures it is ob-
vious that a quantitative study of the
nature of the pigment becomes possible
only when a number of measurements
have been averaged. The situation was
dramatically demonstrated in the case
with a higher concentration of methyl
viologen that accelerates the recovery
(reoxidation) of the acceptor (Fig. 24,
right column). With one light exposure
the absorbance change at 443 nm pro-
vided virtually no information because of
the very high noise level. By contrast,
the signal averaged for 16 exposures
clearly reveals the fast reduction and re-
oxidation of the primary acceptor in the
presence of 100 jxM methyl viologen.
References
Anderson, J. M., and N. K. Boardman, Bio-
chim. Biophys. Acta, 112, 403, 1966.
Hiyama, T., Biophys. Soc. Abstr. 11, 31a,
1971.
A COMPUTER RESOLUTION OF
COMPLEX KINETICS
Tetsuo Hiyama
The measurement of time courses of
absorption changes both in vivo and in
vitro has been widely used as one of the
most powerful and convenient methods
to study biochemical reactions. Its non-
destructive nature as well as its high
sensitivity and fast time response were
found to be particularly useful and it has
been used almost exclusively for the
study of redox reactions in respiratory
and photosynthetic electron transport
systems.
Usually, certain wavelengths are
chosen so that only one component may
be observed without interference from
others. In a system where only two com-
ponents are expected in a certain wave-
length region, isosbestic points, if avail-
able, can be used for this purpose. In
many cases, it is often useful to construct
difference spectra based on kinetical cri-
teria. Often, however, one encounters a
situation where several overlapping com-
ponents may change their absorptions
with different kinetics and make it diffi-
cult to study individual components.
In principle, an absorption change at a
certain wavelength, Ai4A, could be repre-
sented by the sum of the individual ab-
sorption changes due to each component
at this wavelength:
&Ax = %AAKii= 1,2, '-,n. (1)
n
AAx = ZAEx,iX AC, (2)
where AE\t t is the difference extinction
coefficient of a component (i) at wave-
length, A;, and AC,- is the concentration
difference of the component (i) . Thus,
the following equation could be derived:
AAKi = X &EXji , X AC,,
i = l
j=zl,2,--,n. (3)
Thus, the following simultaneous equa-
tions, which consist of n equations at n
different wavelengths, would provide con-
centration differences of n different com-
ponents when all AE's are known:
A^Xl = X A#Xl. , X AC,
A^Xo = X A#x2. ,- X AC*
AAxn = X A£x„. , X AC,
(4)
?=i
372
CARNEGIE INSTITUTION
The present method is based on the above
equations. A digital computer has
greatly reduced the otherwise tedious
task of solving hundreds of sets of simul-
taneous equations for the analysis.
A program was written in STAN-
FORD/BASIC language for an IBM
360/67 computer at the Stanford Uni-
versity Computation Center to solve
these equations and plot AC's against
time on a typewriter print-out. Both
data input and print-out were done at a
terminal in the laboratory (Year Book
67, p. 534). Presently, two programs are
in use: one for a two-component system;
the other for a three-component system.
Necessary parameters such as amplifier
gain and offset voltage for the conversion
of voltage readings to absorption changes
as well as the extinction coefficients of
the components at each wavelength were
put into the program in the beginning.
The data and the readings of time and
voltage changes from a signal averager
(see the preceding section) were put into
the program at the end. For the moment,
the voltage readings are estimated by
measuring traces on a chart paper of an
X-Y recorder, either by hand or by a
curve digitizer [Year Book 67, p. 535).
The results, which represent the concen-
tration differences of the components at
a particular time, were either printed out
in a table or plotted against time.
In order to test the feasibility of the
present method, a reaction system was
chosen as a model. The model was the
flash-induced absorption changes in a
three-component system, where P700,
cytochrome c, and P430 were expected to
show absorption changes in the blue band
(400 nm-500 nm) . According to a pre-
vious study (Hiyama, 1972) with spin-
ach subchloroplast particles enriched in
the reaction center, P700, and the pri-
mary acceptor, P430 of photosystem 1,
the electron transport of the following
sequence would take place in the presence
of methyl viologen (MV) and plasto-
cyanin (PC) in addition to mammalian
cytochrome c (cyt c) :
Light
cyt c -» PC -» P700 -> P430 -> MV -> 02
Due to their much lower absorption in
the blue band, neither plastocyanin nor
methyl viologen would significantly con-
tribute to the absorption changes of the
reaction mixture. It has also been shown
that the oxidation of cytochrome c was
kinetically almost identical to the re-
reduction of P700, despite the presence
of the mediator, plastocyanin, due to its
extremely high turnover (Hiyama, 1972) .
As the reoxidation of P430 by 67 [xM
methyl viologen was faster than the time
resolution of the instrument (less than 2
msec), the contribution of the absorption
changes due to P430 was expected to be
small, i.e., the concentration changes
would stay zero throughout the time
course. Those results had been obtained
by observing each component at a se-
lected wavelength where the interference
from other components was insignificant
or negligible, i.e., P700 at 700 nm, cyto-
chrome c at 550 nm, and P430 at 443 nm.
Three wavelengths were arbitrarily
chosen for the program: 410, 420, and 430
nm. The time course of the absorption
changes at each wavelength are shown
in Fig. 25. It is obvious that kinetics
of the absorption changes of the com-
ponents induced by the flash could hardly
be determined either at 420 nm, the ab-
sorption maximum of cytochrome c, or at
430 nm, that of P700. Figure 26 is an
example of a computer print-out, proc-
essed by the program with the data de-
picted from the traces in Fig. 25. As ex-
pected, P430, plotted as A, showed little
change throughout the time course of the
reaction. On the other hand, P700 (D)
and cytochrome c (C) showed almost
identical reduction and oxidation
kinetics.
Although the present method may be a
powerful tool for the analysis of com-
DEPARTMENT OF PLANT BIOLOGY
373
4IOnm
420nm
430nm
100msec
AA
-IO-*J
0 -i
-10
4J
Fig. 25. Flash-induced absorption changes in
spinach photosystem 1 particles with plasto-
cyanin and mammalian cytochrome c.
plex absorption change kinetics, particu-
larly when appropriate wavelengths for
certain components cannot be found to
measure them separately, some precau-
tions must be taken for reliable results.
The wavelength must be chosen so that
the different extinction coefficients of
each component are reasonably large to
minimize the error. It is not only be-
cause of the inherent inaccuracy of the
coefficient of a smaller value, but also
because of smaller signals which include
more error and noise. Equation 4 treats
all A#'s and AA's equally. Thus, the
error and noise could be amplified par-
ticularly in the case of small changes. To
minimize the error as well as the tedious
task of data reading and input, the direct
input of the data from the signal aver-
ager into the program by the ACME sys-
tem (Year Book 67, p. 534) is being
planned.
It has also been encountered that some
of the extinction coefficients are either
unknown or different from those deter-
mined with purified samples, possibly
due to the flattening effect of particle
suspensions (Duysens, 1956) or for other
reasons. It would, however, be possible
to calculate the extinction coefficients
from the time courses if the latter could
be determined independently by some
other means. Such a program has been
written in BASIC and is presently being
tested for several reaction systems.
References
Duysens, L. N. M., Biochim. Biophys. Acta,
19, 1, 1956.
Hiyama, T., VI Inter. Congr. PhotobioL,
Bochum, Germany, 248, 1972.
Time.
C D
C
C
D C
C
c
c
5
6
7
8
9
10
II
12
13
14
15
16
17
18
Fig. 26. Computer resolution of traces in Fig. 25. Oxidized forms of P700 and cytochrome c and
reduced forms of P430 were denned to have positive absorption changes. The time scale was
arbitrary; 200 msec corresponds to 10 units.
374
CARNEGIE INSTITUTION
LIGHT-INDUCED SHIFTS IN THE ABSORPTION
SPECTRUM OF CAROTENOIDS AND
CHLOROPHYLL b IN THE GREEN ALGA Viva
David C. Fork
Since their discovery in 1954 by Duy-
sens {Year Book 53, p. 166; Duysens,
1954) the absorbance changes seen at
475 and 515 nra have been a challenge
to explain. It appears now that a light-
induced shift in both the absorption spec-
tra of chlorophyll b and carotenoids
contributes to absorption changes in this
region (Amesz and Visser, 1971 ; for a
review, see Fork and Amesz, 1970) . Witt
and his collaborators (Witt et al., 1969)
have sought to explain these changes on
the basis of the formation of a light-
induced electrical potential across the
thylakoid membrane. This potential, in
turn, is considered sufficient to produce
red shifts (Stark effect) in the absorption
maxima of the bulk pigments embedded
in the membranes such as chlorophyll b
and certain forms of chlorophyll a and
carotenoids. Certain types of antibiotics
such as Gramicidin D can cause these
potentials to be dissipated, giving rise to
an acceleration of the dark decay and a
return of the absorption spectrum to its
original state (blue shift).
In purple bacteria, light-induced shifts
in the absorption spectra of carotenoids
have been known for a number of years
(Smith and Ramirez, 1960; Nishimura
and Chance, 1963) . However, such shifts
were clearly observed in higher plants
only recently (Year Book 67, p. 496,
Fork and Amesz, 1967). So far, such
changes have been detected in red algae,
brown algae, and yellow-green algae
(Xanthophyceae) and in a barley mutant
lacking chlorophyll b.
In plants containing chlorophyll 6,
however, it is difficult to distinguish
changes produced by carotenoids from
those produced by chlorophyll b. The
marine green alga Viva is useful in this
regard because, as shown by French and
Elliot (Year Book 57, p. 278) , it contains
a large amount of chlorophyll b. Thus
this plant might be expected to have the
changes produced by chlorophyll b ac-
centuated allowing a distinction between
these changes and those produced by
other compounds.
When this alga was illuminated with
red actinic light, a difference spectrum
was observed (Fig. 27) that had peaks at
420, 431, 441, 460, 477.5, 497 (inflection) ,
510, 555, 648, 660, and 667 nm (inflec-
tion). The minima at 420 and 555 nm
can be attributed to the light-induced
oxidation of cytochrome /. With the ex-
ception of the peaks at 420 and 555 nm,
the difference spectrum of Viva was simi-
lar to that measured by Emrich et al.
(1969) in spinach chloroplasts for ab-
sorbance changes whose dark decay times
could be accelerated by the antibiotic
Gramicidin D.
In Viva symmetric negative 648 and
positive 660 nm peaks occur in the red
region. Fluorescence did not contribute
to or distort these measurements because
its contribution was biased out by a sec-
ond photomultiplier that monitored fluo-
rescence but was out of the path of the
measuring beam. The fluorescence signal
was likewise found not to be influenced
by the presence of the measuring beam at
the intensities used. These symmetric
negative 648 and positive 660 nm changes
can be attributed to a red shift of the red
band of chlorophyll b. Negative and
positive peaks near 650 and 660 nm were
first observed in the green algae Scene-
desmus and Chlorella by Kok (1957),
who interpreted them to be shifts in the
absorption of chlorophyll b. This conclu-
sion was also reached by Rumberg
(1964) who saw similar changes in
Chlorella. Recently, Amesz and Visser
(1971) reported alternate negative and
positive peaks in the red region of the
DEPARTMENT OF PLANT BIOLOGY
375
4.0
3.0
2.0
1.0
<
< -1.0
-2.0
■3.0
-4.0
-
i i i i 1 i i
I I 1 A I
510 3\
i i i i i i i i
- i i
i I i i i
i
-
497/ \
Ulva
-
-
460
660
-
431 3 \
\n rf^K^vi^
*oom:
r-
1 / L667
555
-
-
\ / 441
-
420 T
648
-
-
-
-
-
1 6 477-5
-
-
i i i i > i
a, i i i i
i
-
400
450 500 550
Wavelength, nm
600
650
700
Fig. 27. Light-minus-dark difference spectrum for Ulva obtained with 1 sec light exposures. The
spectrum from 400 to 620 nm was measured with red actinic light having wavelengths extending
from 650 to 750 nm and an intensity of 4 X 105 ergs cm"2 sec"1. The spectrum was measured from
600 to 690 nm, using blue actinic light having wavelengths from 360 to 500 nm and an intensity
of 4.6 X 104 ergs cm"2 sec"1. A second photomultiplier that measured only the fluorescence signal
was used with a Hewlett-Packard differential amplifier (Model 2470A) to cancel signals caused
by fluorescence. The half-bandwidth of the measuring beam was 2 nm. Gas phase, air; 20° C.
difference spectrum of Scenedesmus that
were attributed to red shifts of chloro-
phyll b and to the 670 nm band of chloro-
phyll a. These maxima and minima were
also seen in spinach chloroplasts by
Emrich et al. (1969) and were interpreted
in a similar way.
The minimum at 477.5 and maximum
at 510 nm would appear to be the same
as the well-known 475—515 nm change
discussed earlier. Emrich et al. (1969)
suggested that these peaks are produced
by a red shift of the blue maximum of
chlorophyll b and that the difference be-
tween the negative and positive peaks of
the difference spectrum in the blue and
red (37 versus 12 nm) may be caused by
different angles between transition mo-
ments of these bands with respect to the
direction of the electric field. Compared
to chloroplasts, the inflection at 497 nm
is particularly pronounced in Ulva that
is rich in chlorophyll b. It thus may be
more reasonable to suppose that the
minimum at 477.5 nm and the inflection
at 497 nm represent the red shift of the
Soret band of chlorophyll b rather than
the 475-515 nm changes. If this is the
case, then the difference between the
shifts of the Soret and red bands would
be 19.5 and 12 nm, respectively.
It would appear that the remainder of
the peaks in the blue region (with the
exception of the 420-nm change produced
by cytochrome /) are produced by a
light-induced red shift in the absorption
of a carotenoid. Since the absorption
spectrum of a typical carotenoid has
three maxima separated by about 30 nm,
a shift of this type of spectrum a few nm
to longer wavelengths yields a difference
spectrum having alternately three nega-
tive and three positive peaks. The posi-
tive maxima will be separated from each
other by about 30 nm as will the nega-
tive peaks. The last positive peak in the
difference spectrum that is produced by
a red shift of carotenoid absorption char-
acteristically decreases gradually toward
long wavelengths.
In Ulva the 510 nm peak is likely to
be the farthest red maximum in the dif-
ference spectrum of a carotenoid that is
shifted as discussed above. One difficulty
376
CARNEGIE INSTITUTION
with this idea is that its peak location
seems too far removed (50 nm) from the
next positive peak at 460 nm to be a
carotenoid. It may be that the peak loca-
tions at 460 and 510 nm are somewhat
distorted by the large neighboring 477.5
and 497 nm changes. Other evidence that
the 510 nm peak is likely to belong to a
carotenoid change is the result that the
difference spectrum of a mutant barley
that lacks chlorophyll b also exhibits a
peak near this wavelength (512 nm). In
this mutant a distance of 37 nm sepa-
rates the 512 nm peak from the next
positive peak at 475 nm (Fork, 1969) .
Emrich et al. (1969) have noted other
maxima and minima in their difference
spectrum for spinach chloroplasts treated
with Gramicidin D that were attributed
to shifts of various forms of bulk chloro-
phyll a. Such changes undoubtedly also
complicate the difference spectrum of
Viva.
In any case, some of these changes pro-
vide a convenient way to study rapid
membrane effects by means of spectral
changes.
References
Amesz, J., and J. W. M. Visser, Biochim. Bio-
phys.Acta, 234, 62, 1971.
Duysens, L. N. M., Science, 120, 353, 1954.
Emrich, H. M., W. Junge, and H. T. Witt,
Z. Naturforsch., 24b, 1144, 1969.
Fork, D. C, in Progress in Photosynthesis
Research, H. Metzner, ed., International
Union of Biological Sciences, Tubingen, 2,
800, 1969.
Fork, D. C, and J. Amesz, Photochem. Photo-
biol, 6, 913, 1967.
Fork, D. C, and J. Amesz, in Photophysiol-
ogy, A. C. Giese, ed., Academic Press, New
York, 5, 97, 1970.
Kok, B., Acta Botan. Neerland., 6, 316, 1957.
Nishimura, M., and B. Chance, Biochim. Bio-
phys. Acta, 66, 1, 1963.
Rumberg, B., Nature, 204, 860, 1964.
Smith, L., and J. Ramirez, J. Biol. Chem.,
235, 219, 1960.
Witt, H. T., B. Rumberg, W. Junge, G.
Doring, H. H. Stiehl, J. Weikard, and C.
Wolff, in Progress in Photosynthesis Re-
search, H. Metzner, ed., International
Union of Biological Sciences, Tubingen, 3,
1361, 1969.
THE FUNCTION OF PLASTOCYANIN IN
ELECTRON TRANSPORT OF PHOTOSYNTHESIS
The Content of Plastocyanin in Spinach Chloroplasts and
subchloroplast particles prepared by different techniques
Norio Murata and David C. Fork
We have been investigating the photo-
chemical activities and electron transport
reactions in particles prepared with the
French pressure cell (Fork, 1972; Fork
and Murata, 1971, 1972; Murata and
Brown, 1970; Murata and Fork, 1971;
Year Book 70, p. 468; Year Book 71, p.
171). This method, employing the
French pressure cell, yields two types of
particles: small particles (fraction 1)
having only system 1 activity and large
particles (fraction 2) having system 2
and weak system 1 activity (Michel and
Michel-Wolwertz, 1969; Murata and
Brown, 1970; Sane et al., 1970).
The fraction 1 particles, like untreated
chloroplasts, can reduce NADP in a sys-
tem 1 reaction in the presence of DCMU
(3-(3,4-dichlorophenyl)-l,l-dimethylu-
rea) and an artificial electron donor such
as DCIP (2,6-dichlorophenol-indophe-
nol). The addition of exogenous plasto-
cyanin does not stimulate this reaction
(Murata and Brown, 1970) . This lack of
a plastocyanin requirement in the system
1 reaction in the French press— treated
particles has been observed in some other
laboratories (Baszynski et al., 1971 ;
Knaff and Arnon, 1970).
DEPARTMENT OF PLANT BIOLOGY
377
In addition, we found that added
plastocyanin had little effect on the re-
duction of P700 in the French press par-
ticles but artificial electron carriers such
as DCIP could donate electrons to P700
as efficiently as in untreated chloroplasts.
Both the fraction 1 and 2 particles ob-
tained from the French press had rapid
light-induced oxidation reactions of cyto-
chrome /, as fast as that measured by
Hildreth (1968) for this cytochrome in
untreated chloroplasts illuminated with
a ruby laser. Detergent treatment of
French press particles abolished the rapid
light-induced oxidation of cytochrome /.
After detergent treatment the reduction
of P700 was mediated efficiently by
plastocyanin but not by artificial car-
riers. The above observations suggested
that the use of the French press is a mild
procedure that yields particles retaining
their electron transport system in its
native state.
In our previous experiments (Year
Book 70, p. 468) we found that only a
very small amount of plastocyanin re-
mained in either the fraction 1 or fraction
2 particles. For fraction 1 particles the
ratio of chlorophyll to plastocyanin was
3500 and in fraction 2 particles it was
7000. In contrast to this observation,
Baszynski et al. (1971) reported a ratio
of 300 for the fraction 1 particles and 500
for the fraction 2 particles; and Arntzen
et al. (1972) found intermediate values.
We have therefore done extensive
studies on the content of plastocyanin in
chloroplasts and in subchloroplast par-
ticles prepared by a variety of methods
to release plastocyanin, since our earlier
results suggested that plastocyanin is
absent from particles derived from
French pressure treatment, and that it
does not function between cytochrome /
and P700 or between artificial electron
donors and P700.
Methods
Class 2 chloroplasts were prepared by
disrupting market spinach leaves with a
Waring blender in 0.4 M sucrose, 0.01 M
NaCl, and 0.05 M phosphate buffer (pH
7.8) and by collecting the chloroplast
fraction, using differential centrifugation.
The chloroplasts were washed in the same
medium and resuspended in a solution
containing 0.15 M KC1 and 0.05 M Tri-
cine-KOH buffer (pH 7.5) .
Fraction 1 and 2 particles were pre-
pared from chloroplasts according to the
method developed by Michel and Michel-
Wolwertz (1969; Year Book 67, p. 508)
by using the French pressure cell made
in this laboratory and slightly modified
from that described by Milner et al.
(1950; Year Book 1±8, p. 88). To prepare
large quantities of fraction 1 particles we
sometimes used a step-wise sucrose gradi-
ent having densities of 12.5 to 22.5%
(W/V) instead of the linear gradient of
12.5 to 50%. After centrifugation the
particles were resuspended in 0.15 M KC1
and 0.05 M Tricine (pH 7.5). The yield
of fraction 1 particles on the basis of the
chlorophyll content was 10 to 15% com-
pared to chloroplasts as the starting ma-
terial. In the fraction 1 particles the
chlorophyll a/b ratios ranged from 6.5 to
8 and from 2.5 to 2.7 in fraction 2 par-
ticles.
A French pressure cell manufactured
by Ohtake Ltd. of Japan was also used,
but we were not able to get a good sepa-
ration of the two photosystems with this
cell.
Four kinds of procedures were used to
release and analyze the plastocyanin
contents in chloroplasts and chloroplast
particles: (1) sonication, (2) French
pressure cell treatment, (3) Triton X-100
treatment, and (4) acetone treatment
with buffer extraction.
In the first method, the chloroplasts or
particles were suspended in 0.15 M KC1
and 0.05 M Tricine-KOH buffer (pH
7.5) and treated for 5 min by sonic oscil-
lation at 20 Kc, 100 W with a MSE
Ultrasonic Disintegrator. The proteins
released were separated from the result-
ing particles by centrifugation at 144,000
378
CARNEGIE INSTITUTION
X g for 30 min. The content of plasto-
cyanin was determined in the superna-
tant.
In the second method, French pressure
treatment was substituted for sonication,
and the other procedures described above
were the same.
In the third method, Triton X-100 was
added to a suspension of chloroplasts or
particles until the ratio of Triton to
chlorophyll was 10 (W/W). The sus-
pension was then centrifuged at 144,000
X g for 60 min. The supernatant, which
still contained a large amount of chloro-
phyll, was put through a DEAE-cellulose
column equilibrated with 0.015 M phos-
phate buffer (pH 7.2). The solution
eluted with 0.2 M phosphate buffer con-
tained only a very small amount of green
material and was used for the determina-
tion of plastocyanin.
In the fourth method, the chloroplasts
or particles were treated and then washed
with 85% acetone until the acetone solu-
tion became colorless. Plastocyanin was
then extracted with a small amount of
0.015 M phosphate buffer (pH 7.2) . After
extraction for 2 hours the suspension was
centrifuged at 144,000 X g for 60 min.
The plastocyanin content in the super-
natant was determined.
The content of plastocyanin in a solu-
tion was determined by a spectroscopic
method. A base line was first meas-
ured using the solution to which
ferricyanide had been added to a final
concentration of 1 ml in both the refer-
ence and sample cuvettes. In order to
measure the absorbance of plastocyanin
in its oxidized form, a small amount of
sodium ascorbate was then added to re-
duce the plastocyanin in the reference
cuvette. The concentration of plasto-
cyanin was estimated from the difference
in absorption at 597 nm. In most cases,
the difference in absorption at 700 nm
was also measured to correct for the base
line shift after adding ascorbate. The
path lengths of the cuvettes were 10 and
5 cm when a Beckman (DK 2) spectro-
photometer was used and 1 cm when a
Shimadzu (MPS— 50) spectrophotometer
was used. The molar extinction coeffi-
cients used for the calculations of the
plastocyanin concentrations were 9.8 X
103 cm"1 M'1 at 597 nm and 3.2 X 103
cm"1 M1 at 700 nm (Katoh et al, 1962) .
The minimum detectable amount of
plastocyanin was 3 nmoles.
We also used a modification of the
"bioassay" method of Plesnicar and Ben-
dall (1970) to estimate the plastocyanin
content in a solution. This method takes
advantage of the high activity of plasto-
cyanin in reducing P700 in detergent-
treated particles. After a 50 msec light
exposure, which was long enough to oxi-
dize all of the P700, the initial rate of
dark reduction of P700 was measured.
Figure 28 shows a linear relationship be-
tween the amount of plastocyanin added
and the initial rate of P700 reduction. As
shown by the dashed line in Fig. 28 the
amount of plastocyanin in an unknown
O.I 0.2
Plastocyanin added , n mole
0.3
Fig. 28. The relationship between the amount
of plastocyanin added and the initial rate of
P700 reduction in system 1 particles prepared
with digitonin. The reaction mixture (in a
0.5 cm cuvette) contained 150 ml KC1, 50
ml Tricine-KOH buffer (pH 7.8), 1 ml
sodium ascorbate, 0.1 ml methyl viologen, and
particles amounting to 16 /mM chlorophyll. The
absorbance changes of P700 were measured at
430 nm (half-band, 2 nm). Actinic light con-
sisted of a broad band of red light extending
from 650 to 750 nm and an intensity of 4 X 105
ergs cm-2 sec-1. Each point is an average of 4
to 5 measurements.
DEPARTMENT OF PLANT BIOLOGY
379
solution could be found from this curve
by determining the effect of an aliquot of
the solution on the initial rate of P700
reduction.
Still another method was used to esti-
mate the plastocyanin content in fraction
1 particles. Triton X-100 was added to a
sufficient amount of fraction 1 particles
to make the ratio of Triton X-100 to
chlorophyll approximately 20 (W/W).
This procedure converted the French
press particles into detergent-treated
particles, thus changing the characteris-
tics of P700 with respect to the interac-
tion with an electron donating system;
the other effect was to release any plasto-
cyanin from the fraction 1 particles. We
measured the initial rates of P700 reduc-
tion in the absence and the presence of
known amounts of added plastocyanin.
Figure 29 shows an example of this type
of experiment. Figure 29 shows that 0.07
nmole was found in the fraction 1 par-
-0.07
0 0.1 0.2
Plastocyanin, n mole
0.3
Fig. 29. The relationship between the amount
of plastocyanin added and the initial rate of
P700 reduction in the presence of Triton X-100
in fraction 1 (system 1) particles prepared by
fragmenting chloroplasts in a French pressure
cell according to the procedure developed by
Michel and Michel-Wolwertz (1969). The re-
action mixture (in a 0.5 cm cuvette) contained
150 ml KC1, 50 ml Tricine-KOH buffer (pH
7.8), 0.1 mM sodium ascorbate, 0.1 ml methyl
viologen, 0.05% Triton X-100, and fraction 1
particles amounting to 26 \xM chlorophyll. Each
point is an average of 4 to 5 measurements.
The P700 changes were measured as described
for Fig. 28.
tides. In these experiments, ascorbate
(10"4 M) was always added to keep P700
in the reduced form in the dark. This
concentration of ascorbate had almost no
effect on the initial dark reduction of
P700. In both of the above-described
bioassay methods the minimum detect-
able amount of plastocyanin was 0.02
nmole.
These bioassay methods depend upon
the assumption that only plastocyanin
reduced P700 efficiently in the detergent-
treated particles. This is not always true
because we have found (Fork and Mu-
rata, 1971) that P700 reduction was also
accelerated by algal cytochromes of the c
type and that these cytochromes were
almost as efficient as plastocyanin. If
there are unknown substances that are
active in reducing P700, it is possible to
overestimate the plastocyanin content,
using the bioassay method.
Results
Table 14 shows the amounts of plasto-
cyanin released from chloroplasts by em-
ploying various disintegration techniques.
It can be seen that sonication released
plastocyanin most effectively and that
the ratio of plastocyanin released to
chlorophyll was 0.0021 (reciprocal, 470 ) .
By comparison the pressure cell made
and routinely used in this laboratory re-
leased about 81% as much plastocyanin
from chloroplast membranes. In this case
the plastocyanin/chlorophyll ratio was
0.0017. An important difference between
the Ohtake and the Carnegie pressure
cells is that the Ohtake cell produced
particles like those obtained after sonica-
tion or detergent treatment. With these
particles artificial electron carriers were
no longer active donors to P700 but
plastocyanin was. Katoh (1972) re-
ported that particles from the Ohtake
cell were like detergent or sonication
particles because the plastocyanin/
ascorbate couple donated electrons to
photosystem 1 more efficiently than did
380
CARNEGIE INSTITUTION
TABLE 14. Amount of Plastocyanin Liberated from Chloroplasts by Various Treatments *
Treatment
Chi (Mmole) PC (/xmole)
PC/Chl
Chl/PC
Sonication
81
0.198
0.0024
(410)
52
0.115
0.0022
(450)
67
0.128
0.0019
(520)
36
0.071
0.0020
(510)
average
0.0021
(dbO.OOOll)t
(470)
French press, Ohtake
54
0.112
0.0021
(480)
184
0.370
0.0020
(500)
78
0.167
0.0021
(480)
19
0.038
0.0020
(500)
average
0.0020
( ±0.00003) t
(490)
French press, Carnegie
30
0.047
0.0016
(640)
14
0.026
0.0018
(540)
87
0.159
0.0018
(550)
37
0.052
0.0014
(710)
75
0.149
0.0020
(500)
28
0.047
0.0017
(600)
60
0.109
0.0018
(550)
191
0.31
0.0018
(560)
average
0.0017
( ±0.00006) f
(570)
Triton X-100
37
0.065
0.0018
(560)
Acetone
81
0.130
0.0016
(620)
*Plastocyanin was assayed by the spectroscopic method described in the text.
fStandard error of the mean.
DCIP and ascorbate. We have found
(Fork, 1972; Fork and Murata, 1972)
that with the Carnegie pressure cell the
electron transport system was not dam-
aged. In these particles artificial electron
carriers donated electrons effectively to
P700 but plastocyanin did not.
In order to check whether any plasto-
cyanin remained in particles after soni-
cation or French pressure cell treatment,
the particles were further treated by
other methods to release plastocyanin.
Table 15 shows that if particles obtained
after sonication were given a further
acetone treatment and buffer extraction
no more plastocyanin was released. The
lack of plastocyanin in these particles
agrees with the results of Knaff and
TABLE 15.- The Amount of Plastocyanin Liberated from Chloroplast Particles
by Various Additional Treatments *
Subsequent
Initial Treatment
Treatment
Chl(/zmoles)
PC(/imoles)
PC/Chl
Chl/PC
Sonication
acetone
74
0.000
0.0000
French press, Ohtake
sonication
54
0.006
0.0001
78
0.008
0.0001
French press, Ohtake
triton X-100
162
0.008
0.0001
165
0.012
0.0001
French press, Ohtake
acetone
184
0.012
0.0001
French press, Carnegie
sonication
78
0.024
0.0003
(3,300)
32
0.010
0.0003
(3,200)
65
0.020
0.0003
(3,300)
French press, Carnegie
Triton X-100
60
0.024
0.0004
(2,500)
* Plastocyanin was assay
ed by the spectro
scopic method described in
the text.
DEPARTMENT OF PLANT BIOLOGY 381
Arnon (1970) that sonication liberates agreement with that found for plasto-
all of the plastocyanin from the chloro- cyanin originally by Katoh et al. (1962)
plast membranes. An estimation of the and also agrees with the contents of P700
amount of bound plastocyanin based and cytochrome / in chloroplasts (Board-
upon the in situ electron paramagnetic man and Anderson, 1967; Sane et al.,
resonance signal of this substance has 1970).
been made recently by Malkin and Because two different fractions can be
Bearden (1973). Their values for the obtained after treatment of chloroplasts
plastocyanin content in chloroplasts also with the Carnegie French pressure cell,
agree with the amount of plastocyanin we wanted to see if plastocyanin was
determined by the spectroscopic assay preferentially bound to one of them,
after it is released by sonication treat- Table 16 shows that when fraction 1
ment. particles were treated by sonication or
Particles obtained from the Ohtake acetone, the plastocyanin to chlorophyll
French pressure cell were found to con- ratio obtained was 0.0003— the same as
tain only 5% of the original amount of that found in the total Carnegie French
plastocyanin when they were further press-treated particles. Since the yield
treated with acetone, Triton X-100, or of fraction 1 particles is 10 to 15% of the
sonication. Those particles obtained original chloroplast preparation on a
from the Carnegie pressure cell released chlorophyll content basis, it can be con-
15 to 20% of the original amount of eluded that the ratio of plastocyanin to
plastocyanin when they were further chlorophyll does not vary significantly
treated with sonication or Triton X-100. in the fraction 1 or fraction 2 particles.
It may be noted that the sum of the Our results using the bioassay method
amounts of plastocyanin released from described earlier for the estimation of the
the chloroplasts by the Carnegie French plastocyanin to chlorophyll ratios are
pressure cell and by further treatment of summarized in Table 17. Chloroplasts
these particles with sonication or Triton treated with the Carnegie French press
X-100 gave the total amount of plasto- gave a plastocyanin to chlorophyll ratio
cyanin found originally in the chloro- of 0.0019. Further treatment of particles
plasts. A second treatment of French made with the Carnegie pressure cell
press particles with the Carnegie pressure gave a plastocyanin to chlorophyll ratio
cell did not release any more plasto- of 0.0006. Likewise in fraction 1 particles
cyanin. this ratio was 0.0006. These values are
Results shown in Tables 14 and 15 in- in good agreement with those obtained
dicate that sonication liberates plasto- by Arntzen et al. (1972) and were about
cyanin most efficiently from chloroplast twice the amounts determined by the
membranes and that the original ratio of spectroscopic assay method described
chlorophyll to plastocyanin in chloro- above. In this case the ratios were 0.0003
plasts is 450 to 500. This value is in for both particles.
TABLE 16.
Particles
The
by
Amount of Plastocyanin Liberated from
Subsequent Sonication or Acetone Trea
Fraction 1
tment *
Treatment
Chl(Mmole) PC(jumole)
PC/Chl
Chl/PC
Sonication
Acetone
13
18
0.003
0.005
0.0002
0.0003
(4,000)
(4,000)
* Plastocyanin was assayed by the spectroscopic method described in the
text.
382
CARNEGIE INSTITUTION
TABLE 17. Estimation of the Plastocyanin Content in Chloroplasts
and Subchloroplast Particles *
Material
Treatment
PC/Chl
(Chl/PC)
Chloroplasts
French press, Carnegie
0.0021
0.0019
0.0016
480
520
600
Average
0.0019
530
Particles from French press, Carnegie
Sonication
0.0005
0.0006
Average
0.0006
1,800
Particles from French press, Carnegie
Triton X-100
0.0007
0.0004
Average
0.0006
1,800
Fraction 1 particles
Sonication
0.0006
0.0005
Average
0.0006
1,800
Fraction 1 particles
Triton X-lOOf
0.0006
0.0009
Average
0.0007
1,300
*The bioassay method with digitonin system 1 particles was used. The values obtained were based
on the assumption that the activity of P700 reduction was due only to plastocyanin.
fTriton X-100 was added to the fraction 1 particles, and the change in redox state of P700 in the
particles was measured. See "Methods" for details.
Discussion
These results demonstrate that differ-
ent values for the plastocyanin content
of a sample can be obtained depending
upon the assay method used. We assume
that the spectroscopic method yields the
true values, since with this method the
plastocyanin concentration is determined
directly by the difference in absorption
between the oxidized and reduced forms.
The ratio of plastocyanin to chlorophyll
in the French press particles is 0.0003,
i.e., only 15% of the original amount of
this substance remains in the French
press-treated particles and in the frac-
tion 1 particles. This result confirms our
previous work (Year Book 70, p. 468;
Fork and Murata, 1972) that reported a
chlorophyll to plastocyanin ratio in frac-
tion 1 particles of about 4000.
The bioassay method yields values for
plastocyanin content that are about twice
that determined using the spectroscopic
method. It is possible that stimulation of
system 1 activity as measured by the
bioassay method is not specific to plasto-
cyanin and that there is a component
other than plastocyanin liberated by
sonication or detergent that enhances
dark reduction of P700. It is possible
that an unknown component stimulates
the reaction between P700 and plasto-
cyanin.
Sane et al. (1970) found that the ratio
of chlorophyll to P700 in fraction 1 par-
ticles was 105 and that of chlorophyll to
cytochrome / was 530. By comparing
these values to that for the ratio of
chlorophyll to plastocyanin obtained in
this study it can be concluded that
plastocyanin does not play an important
role in the electron transport system in
the fraction 1 particles. We observed
previously (Year Book 69, p. 682; Fork
and Murata, 1971) that there was prac-
tically no difference between untreated
chloroplasts and fraction 1 particles with
respect to the reaction between P700 and
cytochrome / and between P700 and the
artificial electron donor DAD (2,3,5,6-
tetramethyl-l,4-phenylenediamine) . Cy-
tochrome / was oxidized within 1 msec
by strong flash illumination of the par-
ticles. DAD was an efficient electron
donor to P700, but plastocyanin could
not donate electrons effectively to P700.
DEPARTMENT OF PLANT BIOLOGY
383
These characteristics contrast clearly to
those of detergent-treated system 1 par-
ticles in which the light-induced cyto-
chrome / change is slow and plastocyanin
is an efficient electron donor to P700, but
artificial electron donors are not (Fork
and Murata, 1971) . A very similar situa-
tion was observed with respect to the
reactions of P700 and cytochrome / in
the fraction 2 particles {Year Book 70,
p. 472; Fork, 1972).
Since plastocyanin is almost lacking
in the fraction 2 as well as in the frac-
tion 1 particles, the appearance of a
plastocyanin requirement for system 1
activity after detergent addition is not
due to the release of plastocyanin from
the chloroplast membrane but to an un-
known change in the electron transport
system. This conclusion is supported by
our previous experiments on P700 and
cytochrome / changes in fraction 1 par-
ticles that were treated with various con-
centrations of the detergent Triton X-100
(Fig. 9 in Year Book 70, p. 470) . We ob-
served that the dark reduction of P700,
mediated by an artificial electron donor
DAD, decreased at a concentration of
Triton X-100 of about 0.03%. But the
appearance of P700 reduction by reduced
plastocyanin occurred at a higher deter-
gent concentration of about 0.1%. Per-
haps low concentrations of Triton X-100
decouple the reaction between P700 and
cytochrome / and high concentrations
disrupt the reaction centers of system 1
allowing access of plastocyanin for the
reduction of P700.
Hind (1968) and Nelson and Racker
(1972) observed in detergent-treated
chloroplast particles that added plasto-
cyanin stimulated cytochrome / oxida-
tion. This suggested a sequence whereby
plastocyanin mediated electron transport
between cytochrome / and P700. We
have obtained similar results after add-
ing Triton X-100 to French press par-
ticles. However, this plastocyanin-
stimulated cytochrome / oxidation was
always considerably slower than that ob-
served originally in vivo even if sufficient
plastocyanin were added. These results
indicate that the reconstituted electron
transport system may not be equal to
that of the native state.
In view of the results reported here,
the more likely sequence of electron
transport may be plastocyanin — » cyto-
chrome / -» P700. This sequence was
also suggested from studies made with
Viva on absorbance changes produced
in vivo by plastocyanin (Fork and Ur-
bach, 1965). It is also possible that
plastocyanin and cytochrome / function
in parallel to reduce P700. Additionally,
these results would be compatible with
the idea that an unknown component of
electron transport is functioning between
cytochrome / and P700 and that this
substance would be released by Triton
X-100 treatment or sonic oscillation but
not by passage through the French pres-
sure cell.
References
Arntzen, C. J., R. A. Dilley, G. A. Peters, and
E. R. Shaw, Biochim. Biophys. Acta, 256,
85-107, 1972.
Baszynski, T., J. Brand, D. W. Krogmann,
and F. L. Crane, Biochim. Biophys. Acta,
234, 537-540, 1971.
Boardman, N. K., and J. M. Anderson, Bio-
chim. Biophys. Acta, 143, 187-203, 1967.
Fork, D. C, Biophys. J., 12, 909-921, 1972.
Fork, D. C, and N. Murata, Photochem.
Photobiol., 13, 33-44, 1971.
Fork, D. C, and N. Murata, in Proceedings
of the Ilnd International Congress on
Photosynthesis Research, Vol. I., G. Forti,
M. Avron and A. Melandri, eds., 847-857,
1972.
Fork, D. C., and W. Urbach, Proc. Nat.
Acad. Sci., 53, 1307-1315, 1965.
Hildreth, W. W., Biochim. Biophys. Acta.
153, 197-202, 1968.
Hind, G., Biochim. Biophys. Acta, 153, 235-
240, 1968.
Katoh, S., Biochim. Biophys. Acta, 283, 293-
301, 1972.
Katoh, S., I. Shiratori, and A. Takamiya, J.
Biochem., 51, 32-40, 1962.
384
CARNEGIE INSTITUTION
Knaff, D. B., and D. I. Arnon, Biochim. Bio-
phys. Acta, 223, 201-204, 1970.
Malkin, R., and A. J. Bearden, Biochim. Bio-
phys. Acta, 292, 169-185, 1973.
Michel, J. M., and M. R. Michel-Wolwertz,
in Progress in Photosynthesis Research,
Vol. I, H. Metzner, ed., Tubingen, 115-127,
1969.
Milner, H. W., N. S. Lawrence, and C. S.
French, Science, 111, 633-634, 1950.
Murata, N., and J. S. Brown, Plant Physiol.,
45, 360-361, 1970.
Murata, N., and D. C. Fork, Biochim. Bio-
phys. Acta, 245, 356-364, 1971.
Nelson, N., and E. Racker, J. Biol. Chem.,
247, 3848-3853, 1972.
Plesnicar, M., and D. S. Bendall, Biochim.
Biophys. Acta, 216, 192-199, 1970.
Sane, P. V., D. J. Goodchild, and R. B. Park,
Biochim. Biophys. Acta, 216, 162-178, 1970.
THE PHOTOCHEMICAL REACTIONS OF
PHOTOSYNTHESIS IN AN ALGA EXPOSED
TO EXTREME CONDITIONS
David C. Fork and Tetsuo Hiyama
Some plants must cope with recurrent
periods of rigorous conditions during
their growth cycle. Marine algae living
in the upper littoral zone are periodically
left by the receding tide to dry on the
rocks where they grow. It is not uncom-
mon to find the red alga Porphyra per-
forata drying during the summer months
in full sunlight until it becomes brittle.
This alga is ideally suited and is used for
many types of biophysical measurements,
since it has a broad, flat thallus only one
cell thick.
We have examined the effect of cer-
tain extreme conditions on some of the
photochemical reactions of photosynthe-
sis in Porphyra. For this purpose we
followed in vivo the light-induced oxi-
dation of P700 and of the c-type cyto-
chrome (cytochrome /) as a measure of
photosystem 1 activity. System 2 activ-
ity of photosynthesis was followed by
observing the dark reduction of this cyto-
chrome after exposure to light that ex-
cited not only system 1 but system 2 as
well.
Figure 30 shows the difference spec-
trum measured in the Soret region from
400 to 460 nm for absorbance changes
produced within 50 msec by red actinic
light having a duration of about 2 sec.
This red light was absorbed mainly by
the chlorophyll and to some extent by
the phycocyanin in this alga and served
to excite mostly system 1 reactions and
some system 2 reactions. The difference
spectrum of Fig. 30 shows that it is
mostly the cytochrome that accumulates
in the oxidized form within 50 msec.
P700 appears to contribute only a little
to this difference spectrum, presumably
3.0
-4.0
~4I8
400
450
Wavelength, nm
Fig. 30. Light-minus-dark difference spectrum
for the red alga Porphyra perforata obtained
using 2 sec exposures to a broad band of red
wavelengths extending from 650 to 750 nm and
having an intensity of 4 X 105 ergs cm-2 sec-1.
The half-bandwidth of the measuring beam was
2 nm. Seawater, 20°C; gas phase, air.
DEPARTMENT OF PLANT BIOLOGY
385
because it is kept reduced by electrons
derived from the oxidation of the cyto-
chrome. A spectrum with very similar
shape to that produced by continuous
illumination was also seen when 3 /xsec
flash illumination was used.
The sample used to measure the differ-
ence spectrum for Fig. 30 was then
blotted dry and left in the cuvette to dry
further in direct sunlight for about two
hours. After this treatment it was still
possible to observe in the dried plant
small, reversible, light-induced absorb-
ance changes. The difference spectrum
(Fig. 31, labeled continuous illumina-
tion) showed that only P700 produced
these changes because it had a negative
peak near 430 nm and positive peaks
.o -
o
X
<
<
.0 -
-2.0
1
i i 1 |
450
-
Por
P
hyra ( Dry) ...
-
Continuous / / _
i llumination / /
-
X P
-
\ /-*-Flash
Q^^P illumination
433
1
1 1 1 1
400
450
Wavelength, nm
Fig. 31. Light-minus-dark difference spectra
for the dried thallus of P. perforata exposed
to full sunlight for 2 hours after removing sur-
face water with tissue paper. Continuous illumi-
nation was provided by 1 sec exposures to red
light as described for Fig. 30. Saturating 3 /usee
flashes having a wavelength distribution like
that for the continuous beam were also used.
430
.v^HW^H^vW
420
40msec
Fig. 32. Light-induced absorbance changes
produced by the sun-dried thallus of P. per-
forata to 3 yttsec red flashes as described for
Fig. 31. The traces are the average of 8
measurements.
near 400 and 450 nm characteristic for
this substance (Hiyama and Ke, 1972).
A comparison of Fig. 31 with Fig. 30
shows that the cytochrome was not re-
versibly oxidized in the dried material.
Another sample of Porphyra was sun
dried in a similar way and illuminated
with 3 /xsec flashes of red light. The
spectrum labeled "flash illumination" in
Fig. 31 plots the changes produced 4
msec after the flash and shows that P700
is the substance responsible for these
changes. There may be a very small con-
tribution to this spectrum by cytochrome
/ because of the negative absorption
changes seen at 410 nm. The kinetics of
the changes at 430 and 420 nm are shown
in Fig. 32. It can be seen that even in the
dry Porphyra, P700 is oxidized more
rapidly than could be measured with the
3 //.sec flash.
Seawater was then added to dried
Porphyra used for the continuous differ-
ence spectrum of Fig. 31 and another
difference spectrum was measured im-
mediately afterwards. The result, shown
in Fig. 33, indicates that cytochrome /
regained its ability to undergo reversible
oxidation-reduction reactions as fast as
it could be measured. Figure 34 shows
the kinetics of the absorbance changes at
552.5 nm produced by cytochrome /. The
386
2.0
1.0
f0
O
X
<
<
■2.0
■3.0
■4.0
i r
i r
i — r
CARNEGIE INSTITUTION
1 1 1 1 I T"
Porphyra
(Seawaier added
to dried sample)
552.5
418
400
450 500
Waveleng+h,nm
550
600
Fig. 33. The light-minus-dark difference spectrum measured immediately after adding seawater
to the sun-dried thallus of P. perforata used for Fig. 31. Two-sec exposures to the red actinic
light (described in Fig. 30) were used. The changes occurring 50 msec after the onset of illumina-
tion were plotted. Half-bandwidth of the measuring beam, 4 nm.
rapid light-induced oxidation of this
cytochrome, mediated by system 1 in
the dried plant reconstituted with sea
water, was not affected by 100 ^M
DCMU. However, the rapid reduction
seen after darkening was markedly in-
hibited. It would thus appear that the
Off
Fig. 34. Light-induced absorbance changes at
552.5 nm (cytochrome /) measured as described
for Fig. 33 after adding seawater to the sun-dried
thallus of P. perforata. When used, the concen-
tration of DCMU (3-(3,4-dichlorophenyl-l, 1-
dimethylurea) was 10"4 M.
functioning of system 2 is rapidly re-
gained when the dried plant is given
water.
We next sought to determine whether
any photochemical activity remains in
dried Porphyra tenera. This alga, called
"nori" in Japan where it is extensively
cultivated for use as a food supplement,
was obtained from a local market as
sheets of the sun-dried plants. A number
of small pieces of nori were soaked in
seawater for a few minutes and arranged
over the cuvette surface.
Figure 35 shows examples of kinetics
of absorbance changes seen at 430 and
420 nm when this alga was illuminated
with 1 sec flashes of red light as used for
Fig. 30. It can be seen at both wave-
lengths that the absorbance change was
complex and consisted of an initial rapid
phase followed by a much slower phase
that was not completed even after 1 sec
of illumination. We measured the differ-
ence spectrum for the initial change (oc-
curring within 50 msec) and the slow
DEPARTMENT OF PLANT BIOLOGY
On
430
Off
On I-
-iO
I sec
420
AA
5X10-4
Off
Fig. 35. Light-induced absorbance changes at
430 and 420 nm in Porphyra tenera obtained
(as nori) from a local market importing Japa-
nese food. Small pieces of the dried plants were
soaked in seawater and layered over the surface
of a horizontal cuvette. One-sec exposures to
red actinic light (described in Fig. 30) were
used.
change that was produced after 1 sec of
illumination. Figure 36 shows the result.
The difference spectrum for the rapid
change seems to be produced only by
P700 since it has peaks at 433 and 448
nm and inflection points at 410 and 445
nm (Hiyama and Ke, 1972). By con-
trast, the spectrum for the slow change
is apparently caused by oxidation of
cytochrome / because it has a minimum
at 423 rather than at 433 nm. The shoul-
der around 435 nm may be produced by
a slow oxidation of P700.
These results show that the photosyn-
thetic apparatus is able to withstand
rigorous conditions such as those experi-
enced during prolonged periods of dry-
ness or during exposure to full sun in the
dry state. Dry Porphyra no longer
showed a light-induced oxidation and
dark reduction of cytochrome. When
dry, no electrons can be derived from the
oxidation of water, perhaps causing the
cytochrome to accumulate in its oxidized
form. It is also possible that the removal
of a considerable amount of the water
from the thylakoid membrane during
drying of Porphyra in the sun physically
and/or chemically uncoupled cytochrome
/ from P700.
387
The recovery kinetics of P700 oxidized
by a flash (Fig. 32) resembled those
observed in spinach subchloroplast par-
ticles treated with digitonin that had no
coupled cytochrome / (Hiyama and Ke,
1971). Similar kinetics have generally
been observed in other types of photo-
system 1 particles that have been treated
by sonication, aging, or certain inhibitors
and consequently no longer have photo-
oxidizable cytochrome /. From this re-
sult it appears that a cyclic type of elec-
tron flow may operate in system 1
involving the direct reoxidation of P430
by P700 in the dark.
It appears that only a little system 2
activity survives in nori. This is perhaps
not unexpected since the reactions of sys-
tem 2, especially those leading to oxygen
evolution, are known to be more labile
-2.0
nitial
change
423
433
400
450
Wavelength, nm
Fig. 36. Light-minus-dark difference spectra
of nori (P. tenera) soaked in seawater and
illuminated with 1 sec exposures to red actinic
light as described for Fig. 30 The changes oc-
curring 50 msec after onset of illumination were
plotted as the initial changes. Those occurring
1 sec after the initial change were used to plot
the slow change.
388
CARNEGIE INSTITUTION
than are system 1 reactions. It was
nevertheless surprising to see that P700
changes could still be seen in nori after
water was added.
These results show that at least in
some plants the primary photochemical
reactions of photosynthesis are stable
and can withstand rigorous conditions
such as high levels of UV and visible
irradiation, heat, and dryness. It is
likely that other marine algae growing
with Porphyra are similarly tolerant of
such conditions. These plants thus pro-
vide material not only for the study of
photosynthetic reactions themselves but
also for the study of mechanisms that
enable plants to tolerate these conditions.
References
Hiyama, T., and B. Ke, Arch. Biochem. Bio-
phys., llf.7, 99, 1971.
Hiyama, T., and B. Ke, Biochim. Biophys.
Acta, 267, 160, 1972.
PROBIT THEORY AND APPLICATIONS IN THE
ANALYSIS OF SYNCHRONIZED
GROWTH SYSTEMS
William G. Hagar
The use of synchronous cultures has
been of value in many measurements
with algae, bacteria, amoebae, mammal-
ian cells, and other organisms (Tamiya,
1966; Pirson and Lorenzen, 1966;
Schmidt, 1966; Burns, 1964; Padilla
et al., 1955; Rao and Engelberg, 1966).
By monitoring the cells at various stages
of the life cycle, it has been possible to
make comparisons of development
changes that would have been obscured
in measurements with nonsynchronous
populations. The synchronization of en-
tire populations of cells amplifies the
small changes occurring during develop-
ment into measurable quantities.
A problem has been how to make
meaningful comparisons between the de-
gree of synchrony of cultures in different
laboratories. Selecting the fewest possi-
ble parameters to express a biological
system is a difficult task. The problem is
compounded when there are several
methods for producing a given state, as
there are with inducing synchrony. Most
publications dealing with synchronized
cells give the synchronizing technique
and the growth conditions used in the
experiments. However, even when similar
procedures are used the synchrony often
varies between experiments and an un-
ambiguous comparison cannot be made
from one publication to another.
One of the easiest major parameters
that can be used for comparison is the
normal probability relationships between
elapsed time and rate of cell division.
The change in the rate of cell division
follows a normal distribution function
with the maximum rate at the midpoint
of the curve. This normal relationship is
usually overlooked when the total cell
number — or the percentage of cells that
have divided — is plotted against time.
The time curve for the total cell number is
the integral of the cell division rates up
to that time and has the same shape as
the integral of a normal probability
curve.
A practical way to compare the sig-
moid growth curves from various syn-
chronization experiments is to convert
cell numbers to probits and to use probit
parameters to define the synchrony (Ha-
gar and Punnett, 1973) . Probit transfor-
mation converts a normal sigmoid curve
to a straight line (Finney, 1953, p. 21).
The resulting probits of cell number
when plotted versus time produce a
straight line, having a midpoint corre-
sponding to the peak of the cell division
rate (probit of 5.0), and with the recip-
DEPARTMENT OF PLANT BIOLOGY
389
rocal of the slope equalling the standard
deviation of the cell release rate. These
two parameters, the time of maximum
cell division and the spread of the cell
release rates around the midpoint, are
all that are needed to compare the syn-
chrony from one experiment to another.
The standard deviation for cell division
is a critical parameter for describing the
synchrony of the culture. If the cell
division spread is too large, events which
reversed themselves every few hours
would be averaged out and not seen.
The time required for 95% of the cells
in synchronous cultures to complete divi-
sion is a span of ±2 standard deviations,
or 4 times the standard deviation.
Theory
Probit analysis has been used in the
past by pharmacologists to measure the
effectiveness of various drugs. Bliss in
1934 introduced the probit transforma-
tion that is a simplification of Gaddum's
method of "Normal Equivalent Deviate"
(NED) (Gaddum, 1933) and provided
an extremely useful tool for describing
dosage-response effects of various chem-
icals by determining the dosage which
would affect half of the organisms tested,
and also the effect that different dosages
had on the organisms tested (spread of
effectiveness). For drug dosage curves
the logarithm of the drug concentration
is used to produce the normal response
curve. With cell division data the dosage
value of time is best left linear rather
than logarithmic.
The normal distribution curve can be
expressed as follows (Treloar, 1939) :
of the curve and can range from — oo to
+ oo, but is usually measured within
limits of ±3 standard deviations. The
relative deviate is called the normal
equivalent deviate (NED) in probability
calculations.
The proportion of objects responding
in a large population is equal to the in-
tegral of the normal distribution curve
up to that dosage limit.
P< =
/-:
X, 1
Sy/%
_!/
exp
Xi — X0
dXt
(2)
Figure 37 shows the relationship be-
tween the normal distribution, its inte-
gral, and probit values for a normal
process plotted as percentage of the ob-
jects responding on the ordinate. Tables
relating the normal equivalent deviate
and probability of occurrence are found
in the back of most statistics textbooks.
Probit values are obtained by adding
+5.0 to the normal equivalent deviate
value. This conversion makes all values
of the abscissa positive. Y, the probit
value, corresponds to a given probability
in a normal distribution curve with a
variance of 1.0 and a mean of 5.0. As the
probability approaches 0 or 1.0 the probit
values decrease or increase without limit.
Methods of Calculation
To convert cell number (Nt) data to
probits, the proportion of objects re-
sponding (Pi) must be initially calcu-
lated as follows:
W —
1
exp
-Vo
xt - X
s
p< = lv
N- — V
N0
3)
X0 = mean of curve
S = standard deviation
W = ordinate value at Xi
The quantity (Xi — X0) /S is called the
relative deviate, K, of X,- from the mean
where P, = proportion responding; .V,- =
cells/ml at time (i) ; N0 = Cells/ml be-
fore division has started ; Ne = Cells/ml
after all division is completed.
To convert the proportion of cells that
have divided to a probit value, Fig. 37
390
I.O,
0.8
0.6
> 0.4
u
O
Eo.2
o
0:
CARNEGIE INSTITUTION
1 1 1 3-
Normal curve
integrated from -
\ Normal curve
\
\
NED -5-4-3-2-1 0 I 2 3 4 5
Probi + s 2 3 4 5 6 7 8
Fig. 37. Dashed line is the normal curve with a total frequency distribution of 1.0. Solid line is
the integrated normal curve. Data were obtained from Handbook of Mathematical Functions
(1964).
could be used and the probit value on the
abscissa for a given proportion could be
read off directly. Alternatively, the pro-
bit values can be obtained by adding 5.0
to the normal equivalent deviate ob-
tained from standard statistical tables
of probability or by using probit tables
directly (Fisher and Yates, 1953).
Finney has provided a computer pro-
gram for calculating probit values, stand-
ard deviations, and chi-square reliability
values (Finney, 1971). This program is
part of the standard IBM scientific sub-
routine package and is included in the
library subroutines at most computer
centers. By modifying this program
(Hagar and Punnett, 1973), the user can
analyze cell number data versus time
and produce an accurate statistical anal-
ysis of the cell number data. The print-
out of the program lists the y-intercept,
a, the slope, b, chi-squares and degrees
of freedom for a given set of data. The
time of maximum cell division rate, X0
(the mean of the normal curve), is ob-
tained by using the following formula :
x0 =
5 — a
(4)
Analysis of Synchrony
An analysis was made of synchronous
cell measurements reported in the litera-
ture. The cell number data were con-
verted into probits by using probit tables
(Fisher and Yates, 1953). Cell number
and time values were obtained directly
from a listing in the paper by Spencer
et al. (1961). However, in the rest of
the synchronous systems, analysis was
accomplished by interpolation of the
sigmoid growth curves listed in the pub-
lication, after enlargement, by using a
curve digitizer (Year Book 67, p. 535) to
determine the points and the scale.
Figure 38 is a comparison between
three different synchronized algal sys-
tems of Spencer et al. (1961), Senger
DEPARTMENT OF PLANT BIOLOGY
391
5 v
-§4
u
CL
3
2
1 1
i i i i 1 i i i i 1
/-^Spencer Senger-
/ e+al.
o/
/\
1
(. y
/Wang"
'< /
-CJ
•c /
00 /
1 <oF
00 1
"?*
O'A
" /
" /
Ay
3?/
-W
**/
-? /
0
Q./»
1 O /
Ml
|
09 /
J
/ i i
1 1 1 / 1
111/
' '
1— —
10 10.8
15
Time.hrs
18.9 2020.8
Fig. 38. Probit plot of three synchronized cul-
tures of algae. Probit values were obtained
from cell division data listed in or interpolated
from the publication. The slope and intercepts
of the probit plots were determined using the
modified IBM probit subroutine.
(1970), and Wang (1969). Notice how
the normally sigmoid shape of cell num-
ber versus time curve has been changed
by the probit transformation into straight
lines. The mean time of cell division
rate differs between the separate systems
due to the organism used and to the syn-
chronization techniques. The mean time
of cell division rate is obtained simply by
finding the abscissa value for a probit of
5.0. This parameter is useful for com-
paring the timing of cellular events from
one culture to another. The standard
deviation, the reciprocal of the slope, is
a critical parameter describing the dis-
tribution of cell ages within a population.
The larger the standard deviation, the
more likely that events which reversed
themselves every few hours would be
averaged and not seen.
Figure 39 is a probit plot of synchro-
nized mammalian cells (HeLa cells)
(Rao and Engelberg, 1966) and a ciliated
protozoan (Padilla et al., 1966). The
mammalian cells were synchronized by
giving the cells a large amount of thymi-
dine, which is used to inhibit DNA syn-
thesis, and then removing the thymidine
after all the cells had arrived at the
same state. The protozoan was synchro-
nized by repeated heat shocks. Both
synchronous systems followed a normal
curve generally, but the HeLa cells
started division too early and the proto-
zoa extended their division time. The
8 9.04
Time.hrs
10 io.94
Fig. 39. Probit plot of synchronized mammalian and protozoan cells. The protozoa, Tctrahy-
mena pyriformis HSM, was synchronized by repeated heat shocks. Mammalian (HeLa) cells were
synchronized by first blocking cell division with an excess of thymidine, and then — after waiting
16 hours — removing the block.
392
CARNEGIE INSTITUTION
early division probit value for the HeLa
cells represented only 0.5% of the total
cell division. These early and late probit
values are subject to the greatest errors
and should not be weighed very much in
the determination of probit slope and
midpoint (Finney, 1953, p. 24). If the
same type of misfit occurred in every
measurement of that synchronized sys-
tem, a skewed Gaussian curve could be
used as the basic equation (equation 1)
and its integrated form used similar to
the integrated normal curve as with the
probits. The amount of skewness re-
quired (negative or positive) could be
analyzed by taking the derivative of the
sigmoid cell number curve and then ana-
lyzing the resulting curve by a computer
curve fitting program.
The parameters of mean time of cell
division rate and the standard deviation
of the rate of cell division are all that are
needed to compare the synchrony of cells.
However, these values alone do not allow
the comparison of synchronized systems
with different life cycle times. To com-
pare the synchrony of these systems the
ratio of the standard deviation to the
total mean time of cell division is used.
This ratio of the standard deviation to
the mean time of cell division was first
used by Spencer et al. (1961) to compare
the synchrony of algal systems and is
called the coefficient of variation. Table
18 is a listing of the mean times of cell
division, standard deviations, and coeffi-
cients of variations for synchronized sys-
tems analyzed by the modified computer
program. The synchronized Chlorella
culture of Wang had the lowest coefficient
of variation, and hence had the greatest
total synchrony. These cells were grown
at high light intensities, 2200 footcandles,
and perhaps this is why their synchrony
is the highest.
Probit analysis of synchronized cells
offers a quick, convenient method of com-
paring the degree of synchrony of cells.
When listed with growth and synchroniz-
ing conditions a complete comparison can
be made with other synchronized cells.
These parameters can be thought of as
analogous to the wavelength of absorp-
tion maximum and the half-bandwidth
commonly reported with absorption spec-
tra.
TABLE 18. Computer Analysis of Synchronized Growth Systems
Cell
Division Standard Coefficient
Midpoint Deviation of Variation
Material Synchronization Methods Xo, hours S, hours S/Xo
*Chlorella pyrenoidosa
High temperature strain
(Spencer et al.)
*Scenedesmus obliquvs
(Senger)
*Chlorella pyrenoidosa
(Wang)
Scenedesmus obliquus
(Pfau et al.)
Chlorella pyrenoidosa
(Pi&uetal.)
HeLa cells (Rao et al.)
*HeLa cells (Rao et al.)
*protozoa cells
(Padilla et al.)
9 hours light-9 hours
dark
10.84
0.63
14 hours light-10 hours
dark
18.92
1.37
14 hours light-10 hours
dark
20.80
0.54
14 hours light-10 hours
dark
18.65
0.97
16 hours light-10 hours
dark
21.20
1.06
single thymidine block
and release
9.08
1.76
two separate thymidine
blocks and releases
9.04
1.14
temperature change,
12°-27°C
10.94
0.58
.058
.072
.026
.052
.050
.194
.126
.053
* Illustrated in Figure 38 or 39.
DEPARTMENT OF PLANT BIOLOGY
393
References
Bliss, C. I., Science, 79, 38-39, 409-410, 1934.
Burns, V. W., Synchrony in Cell Division
and Growth, E. Zeuthen, ed., Interscience
Publishers, New York, 433-439, 1964.
Finney, D. J., Probit Analysis: A Statistical
Treatment of Sigmoid Response Curve.
Cambridge University Press, London, 1953.
Finney, D. J., Probit Analysis, Cambridge
University Press, London, 81-99, 1971.
Fisher, R. A., and F. Yates, Statistical Tables
for Biological, Agricultural and Medical
Research, Oliver and Boyd, London, 60-62,
1953.
Gaddum, J. H., Spec. Rep. Ser. Med. Res.
Coun., Lond. 183, 1933.
Hagar, W. G., and T. R. Punnett, Probit
Transformation: An Improved Method for
Defining the Synchrony of Cell Cultures,
in press, Science.
Handbook of Mathematical Functions. Mil-
ton Abramowitz and Irene A. Stegun, eds.
National Bureau of Standards Applied
Mathematics Series, 55, 966-972, 1964.
IBM Application Program Gh20-0205~4 Sys-
tem/360 Scientific Subroutine Package Ver-
sion III, 44-46, Program Number 360-A-
CM-03X.
Padilla, G. M., I. L. Cameron, and L. H.
Elrod, in Cell Synchrony — Studies in Bio-
synthetic Regulation, I. L. Cameron and
G. M. Padilla, eds. Academic Press, New
York, pp. 269-288, 1966.
Pfau, J., K. Werthmuller, and H. Senger,
Arch. Mikrobiol, 75, 338-345, 1971.
Pirson, A., and H. Lorenzen, Ann. Rev. Plant
Physiol, 17, 439-458, 1966.
Rao, P. N., and J. Engelberg, in Cell Syn-
chrony— Studies in Biosynthetic Regula-
tion. I. L. Cameron and G. M. Padilla, eds.,
Academic Press, New York, 332-351, 1966.
Schmidt, R. R., in Cell Synchrony — Studies
in Biosynthetic Regulation. I. L. Cameron
and G. M. Padilla, eds., Academic Press,
New York, 180-235, 1966.
Senger, H., Planta, 90, 243-266, 1970.
Spencer, H. T., R. R. Schmidt, C. Y. Kramer,
W. B. C. Moore, and K. W. King, Exper.
Cell Res., 25, 485-497, 1961.
Tamiya, H., Ann. Rev. Plant Physiol., 17,
1-26, 1966.
Treloar, A. E., Elements of Statistical Rea-
soning, John Wiley and Sons, Inc., New
York, 250, 1939.
Wang, R., Ph.D. Thesis, University of Ro-
chester, 1969.
PHYSIOLOGICAL ADAPTATION TO DIVERSE
ENVIRONMENTS: APPROACHES AND
FACILITIES TO STUDY PLANT RESPONSES TO
CONTRASTING THERMAL AND
WATER REGIMES
Olle Bjorkman, Malcolm Nobs, Joseph Berry, Harold Mooney*
Frank Nicholson, and Benny Catanzaro *
The mechanisms underlying the re-
sponses and adaptation of plants to con-
trasting environments continue to be in
the focus of the physiological ecology in-
vestigations. An expanded research pro-
gram on a long-term comparative study
of adaptations of the photosynthetic
apparatus to ecologically diverse habi-
tats has been initiated this year. This is
in part a joint program with the Depart-
ment of Biological Sciences at Stanford
* Department of Biological Sciences, Stanford
University.
University. We believe that this program
is the most direct approach to an under-
standing of the evolutionary limitations
of the photosynthetic process. Knowledge
of the physiological and biochemical
mechanisms that enable plants to photo-
synthesize efficiently in the great divers-
ity of environments that exist on earth is
also fundamental to our understanding
of primary productivity in different eco-
systems and of plant evolution and dis-
tribution.
Such information is also basic to ap-
plied research concerned with improving
394
CARNEGIE INSTITUTION
the productivity and extending the range
of cultivation of economically important
plants. Inevitably, knowledge of the
evolutionary limits of the adaptation of
the photosynthetic process, derived from
experiments on wild plants occupying
extreme environments, is critical to any
assessment of the potential possibilities
of breeding crop plants that can be suc-
cessfully cultivated in areas now con-
sidered unsuitable. For example, the
incorporation of a photosynthetic char-
acteristic such as the C4 dicarboxylic
acid pathway, which may increase the
ratio of the amount of C02 fixed to the
amount of water lost through transpira-
tion, may make it economically feasible
to extend the cultivation of a crop into
areas where the cost of water available
for irrigation is high. At present there
are active programs that are screening
crops lacking the C4 pathway for strains
or mutants possessing at least some of the
desirable photosynthetic characteristics
associated with this pathway. Such pro-
grams are dependent on basic informa-
tion on the evolutionary adaptive mecha-
nisms of this pathway, derived from
comparative studies of related species
possessing and lacking the pathway, as
well as studies of its inheritance. Un-
doubtedly, there exist many other mecha-
nisms of photosynthetic adaptation to
extreme environments which yet remain
to be discovered.
In our expanded program special em-
phasis will be placed on the photosyn-
thetic performance and primary produc-
tivity of several species and ecotypes of
the same species that occupy habitats
extreme in their thermal load and water
supply. Thus, the studies should gener-
ally circumscribe the limits of adapta-
tion of the photosynthetic process to ex-
tremes of temperature and water rela-
tions and should provide important in-
formation both on how plants can adapt
their photosynthetic apparatus to the
primary stresses of the environment and
what the evolutionary restrictions are
that determine the limits of such adapta-
tion. An integral part of this general
problem is the extent to which the capac-
ity of a given genotype to photosynthe-
size efficiently in one particular stress
environment limits its capacity to do so
in an environment where the stresses are
of a different kind. For example, our
previous studies have shown that Tide-
stromia oblongifolia is remarkably well
adapted to photosynthesize at the very
high summer temperaures that prevail in
its native Death Valley habitat, but it is
unable to perform efficiently at cool and
moderate temperatures, which are near
the optimum for photosynthesis in plants
from cool-temperature environments. An
important question is whether the capac-
ity for efficient photosynthesis at high
temperatures in itself restricts the capac-
ity for efficient photosynthesis at low
temperatures, and if so — why?
It is of utmost importance that experi-
ments attempting to gain insight into
such problems be carried out at widely
different levels of organizational com-
plexity, from the whole plant to the cellu-
lar and the molecular levels. It is further
necessary that they include both a de-
scription of the performance and the
response of photosynthesis and growth to
the natural variation in the main external
factors in the respective habitats at one
end of the spectrum, as well as analyses
of chloroplast constituents and compo-
nent reactions of the overall photosyn-
thetic process at the other end.
Special efforts are being made to
bridge the usual gap between field and
laboratory investigations. For this pur-
pose each single genotype under intensive
investigation will be grown and studied
under a series of conditions: native habi-
tats with the natural differences in
climatic, edaphic, and biotic factors;
transplant gardens climatically similar
to the native habitats but where reason-
able control of water and nutrient supply
can be maintained; controlled growth
facilities using natural light but with
close control of thermal load, evaporative
demand, carbon dioxide, water, and nu-
DEPARTMENT OF PLANT BIOLOGY
395
trients; and finally, a series of artificial
environmental growth chambers where
all of the main external factors, including
the oxygen concentration of the atmos-
phere, can be strictly controlled.
A specially constructed, self-contained
mobile laboratory unit will permit meas-
urements of photosynthetic gas-exchange
characteristics on the whole plant and
single leaf levels to be performed with the
same instrumentation in the field, in the
transplant gardens, and at the controlled
growth facilities at Stanford.
Most of the group's efforts this year
have been devoted to the establishment
of new field transplant gardens, design
and construction of new measuring equip-
ment, controlled growth facilities, and
extensive remodeling and modification of
already existing mobile and stationary
laboratory equipment and facilities. This
was necessary to realize the expanded
program.
Transplant Gardens
For our comparative studies of adap-
tive differentiation of the photosynthetic
apparatus to extremes of thermal load we
have chosen dry and wet habitats on the
floor of Death Valley, California, and ex-
posed sites at Bodega Head on the cool
California coast north of the Point Reyes
Peninsula. During the year transplant
gardens have been established near the
natural habitats in both of these loca-
tions.
The Death Valley garden is situated
33 meters north of the U.S. Weather
Bureau Station, near the National Park
Service Headquarters at Furnace Creek
in the Death Valley National Monument,
36° 27' N. lat., 116° 52' W. long., 60 m
below sea level on essentially level
ground at the base of an alluvial fan.
The natural perennial vegetation in the
immediate proximity of the garden is
dominated by the xerophytic species
Atriplex hymenelytra (C4 photosynthe-
sis), Tidestromia oblongifolia (C4),
Euphorbia sp. (C4), and Larrea divari-
cata (C3). Within a radius of 1 km there
are also stands of Distichlis spicata (C4)
and the phreatophytic species, Prosopis
julifiora (C3), Typha domingensis (C3),
Phragmites communis (C3), Scirpus ol-
neyi (C3), and Pluchea sericea (C3).
The Bodega Head garden is about 73
m south of the Bodega Marine Biology
Laboratory and 200 m south of the local
weather station, in the Natural Lands
and Water Reserve of the University of
California, 38° 19' N. lat., 123° 4' W.
long, on an exposed coastal bluff, 8 m
above sea level and about 30 m from the
Pacific Ocean. The natural vegetation
in the immediate area is composed of
some 34 species in 19 families, all of a low
growth habit (<30 cm). No C4 species
grows on the bluff, but Distichlis spicata
(C4) is abundant on the strand just be-
low the bluff within 1 m of the high tide
mark. Other plants of particular inter-
est to the program, such as Atriplex pa-
tula ssp. hastata (C3), Atriplex californ-
ica (C3), Typha domingensis (C3), and
T. latifolia (C3) grow within a 2 km
radius of the garden site.
Each of the two gardens occupies 350
m2 of fenced-in area with access for the
mobile laboratory. They have identical
dimensions, orientations, and lay-outs.
Line power and piped fresh water have
been brought to the gardens. The water-
ing systems are automatic and the water
supply to each plant can be individually
controlled. Special precautions are taken
to minimize the effects of irrigation on
temperature and humidity of the sur-
rounding air and on root temperature.
Eight 125-liter capacity plastic tanks
have been installed in each garden for
growth of phreatophytes such as Typha
sp. and Phragmites communis. The salt
and nutrient levels in these tanks are
checked periodically and are kept within
desired limits. The meteorological data
collected at the weather stations are wind
speed and direction, rainfall, maximum
and minimum air temperatures (at both
stations) and irradiance, humidity, evap-
396
CARNEGIE INSTITUTION
oration, and continuous recordings of air
and ground surface temperature (at the
Death Valley Station). In addition, in-
struments for continuous recordings of
quantum flux in the photosynthetically
active region of the spectrum, air tem-
perature at different levels above ground,
and soil temperatures at different depths
are being installed in the transplant gar-
dens. Probes for periodical measurements
of soil— water potential will also be in-
stalled.
The two field sites contrast sharply in
thermal load and evaporative demand.
The floor of Death Valley provides the
upper extreme of thermal load in the
western hemisphere and probably on
earth. In summer, air temperatures of
50 °C are common, and plant tissue tem-
peratures may often exceed the tempera-
ture of the surrounding air. The long-
term average daily maximum air tem-
perature, recorded at the U. S. Weather
Bureau Station near the garden site for
July is 47 °C and the average daily mean
temperature is 39 °C. Garden soil sur-
face temperatures often reach 90 °C, and
even at depths of 10-40 cm soil tempera-
ture is in the range of 47°-43°C. The
average rainfall is 42 mm and 75% of
this falls during the winter months. Air
humidity is generally extremely low in
summer and relative humidities of 3% or
below have often been recorded. The po-
tential evapotranspiration is 4064 mm
per year. It thus exceeds the rainfall by
almost 100 times.
In contrast, Bodega Head has a cool-
temperature maritime climate with little
seasonal variation. Summer temperatures
are kept low, since the air passes across
coastal water of 12-13 °C surface tem-
perature. Long-term temperature rec-
ords for the Bodega Head site are not
available, but long-term records from the
U. S. Weather Bureau Stations at Point
Reyes, 34 km south and at Fort Ross,
27 km north of Bodega Head show aver-
age daily maximum temperatures for
July of 14.4 and 19.5 °C, respectively.
Mean daily temperatures for July are
13.2 and 14.4°C, respectively. Available
records for the Bodega Head indicate
that temperatures at this site are in be-
tween those recorded at the Point Reyes
and the Fort Ross Stations. The research
site is unprotected and the characteris-
tically strong northwesterly winds (aver-
age wind speed is 15 to 20 km per hour)
tend to couple the plant temperature to
that of the air, so it can be expected that
leaf temperatures will generally remain
within a few degrees of the air at this
site. The frequent occurrence of summer
fog also contributes to a very low thermal
load on the plants.
Because of the generally low tempera-
ture and high relative humidity the
evaporative demand on the plants is low.
Estimates indicate that the annual poten-
tial evapotranspiration is approximately
equal to or perhaps even somewhat less
than the annual rainfall of 750 to 1000
mm.
In this first season of operation, seed-
ling and cloned materials of a number of
species and ecotypes from habitats with
contrasting thermal loads and water re-
lations have been propagated at Stan-
ford and planted in the two transplant
gardens. The original plantings at the
two sites were identical. Additional
propagules of the same plants are main-
tained at Stanford and will be grown
under a series of controlled conditions
at the laboratory. The species and their
original habitats are listed in Table 19.
The initial transplantings were made
in late March and early April 1973. All
plants were irrigated during the first six-
week establishment period after trans-
planting. During the rest of the active
growing season only one-half of the indi-
viduals will receive continued irrigation.
Routine garden maintenance and note-
taking on plant survival and general de-
velopment are carried out daily by resi-
dent personnel at the Death Valley Na-
tional Monument and the Bodega Marine
Biology Laboratory. Detailed records of
DEPARTMENT OF PLANT BIOLOGY
397
TABLE 19. Original Habitats and C02 Fixation Pathways of Species Planted
in the Death Valley and Bodega Head Gardens
C02
Species
Original Habitat
Fixation Pathway
Atriplex patula ssp. glabriuscula
coastal strand, Swansea, British Isles
C3
Atriplex patula ssp. hastata
coastal marsh, Pescadero, California
c3
Atriplex patula ssp. spicata
interior valley, Los Banos, California
c3
Atriplex rosea (naturalized)
interior valley, Los Banos, California
c4
Atriplex halimus
near Dead Sea, Israel
c4
Atriplex hymenelytra
Death Valley floor, California
c4
A triplex lentiformis
coastal marsh, Point Magu, California
c4
Atriplex lentiformis
Death Valley floor, California
c4
Distichlis spicata
coastal strand, Point Reyes, California
c4
Distichlis spicata
Death Valley floor, California
c4
Tidestromia oblongifolia
Death Valley floor, California
c4
Typha domingensis
coastal pond, Bodega Head, California
c3
Typha domingensis
spring area, Death Valley floor, California
c3
Typha latifolia
coastal marsh, Point Reyes, California
c3
plant growth and biomass production
are taken at monthly intervals by the
main investigators. Intensive measure-
ments of microclimatic conditions and of
physiological responses and photosyn-
thetic characteristics will be made in situ
on selected key plant individuals using
the mobile laboratory unit.
Controlled Growth Facilities
A key item in our experimental ap-
proach is the growth of plants in different
controlled environments. It is of utmost
importance in programs such as the pres-
ent one that the main physical factors of
the contrasting natural environments be
isolated so that the effects of each varia-
ble can be studied separately. For ex-
ample, comparative investigations of
intrinsic responses and adaptation to con-
trasting temperatures require that the
transpirational demand during growth
be kept reasonably constant when the
thermal regime is varied, a condition
which cannot be obtained in natural en-
vironments and rarely even in controlled
growth chambers. It is also desirable
that, at the same time, the plant tem-
perature be closely coupled to that of the
ambient air even when solar radiation in-
tensities are high.
Two new environmentally controlled
growth chambers designed to meet these
requirements are now under construc-
tion. Since very high peak irradiance
levels are required in many of our experi-
ments, natural sunlight is used as the
main light source. Fortunately, the local
climate at our laboratory is characterized
by clear skies and high sunlight intensity
for at least 250 days per year. The cham-
ber design minimizes light attenuation by
the enclosure and should provide a
reasonably uniform light distribution.
With these units it should be possible to
obtain widely diverse temperature re-
gimes, such as those simulating Death
Valley and Bodega Head, while at the
same time maintaining similar fluxes of
photosynthetically active radiation and
water vapor pressure gradients between
the plant and the ambient atmosphere.
To achieve this a design incorporating
a horizontal air flow of high velocity
across the plant canopy is used (Fig. 40) .
The system is closed, so there is no gas
exchange with the air external to the
growth chambers. Heat exchangers with
a large surface area, extending across the
full length of the growth chamber units,
are used to minimize temperature differ-
ences between the cooling coils and the
air. In this way undesirable dehumidi-
398
CARNEGIE INSTITUTION
Aluminum frame
greenhouse
Electric heating
element
Double wall,
insulated
W13
Fig. 40. Diagramatic drawing of the new controlled-growth facilities.
fication of the air is avoided, and tem-
perature gradients across the plant
growth area can be kept small.
Air circulation in each unit is pro-
vided by three large propeller fans with
a total maximum air-handling capacity
of 500 m3 min-1, vertically mounted in a
partition under the floor of the plant
growth area. This floor is supported by
airtight, insulated 1.2 m high walls
which rest on a concrete slab. The space
enclosed by the concrete slab, the walls,
and the floor serves as an air circulation
duct. The finned heat exchange coils are
mounted horizontally in the air opening
in the floor on one side of the chamber.
The air exit opening is located in the floor
on the opposite side of the chamber. The
total volume of the air circulation sys-
tem is about 50 m3. At maximum fan
speed the total air volume should thus
be passed through the heat exchanger at
a rate of up to ten times per minute. The
total floor area is about 20 m2 and the
useful area for plant growth is about
14 m2 for each chamber. Airtight glass-
house units (Orlyt Imperial, Lord &
Burnham, Inc.) enclose the plant growth
area. A curved-eave design with large
glass panes and slim aluminum frames
minimizes reflection and shading of the
incident solar radiation.
The growth chambers are designed to
maintain temperature control in the range
from 10° to 55 °C with full solar radia-
tion and from 5° to 40 °C at night. The
very high heat load at full solar radia-
tion during the summer necessitates a
cooling system of large capacity. In the
present system, a three-stage, dual-com-
pressor water chiller (Westinghouse
Corp.) with a heat removing capacity of
66,000 kcal h"1 (21 tons, or 260,000 Btu
h_1) is used. Water from this unit is
passed through the heat exchanger coils
in the growth chambers whenever the
temperature control system calls for cool-
ing. Heating is provided by electrically
heated elements located immediately
above the cooling coils. Temperature is
continuously monitored and controlled
using thermocouple sensors and two
Honeywell model 111 temperature con-
trol systems in conjunction with two
Data-Track Program Followers, model
5300. This makes it possible to produce
any desired diurnal course of tempera-
ture in the growth chambers and thus
permits simulation of contrasting natural
temperature regimes.
DEPARTMENT OF PLANT BIOLOGY
399
C02 concentration will be precisely
controlled at levels from lower than that
of normal air to about 2000 ppm, using
Beckman infrared gas analyzers equipped
with specially designed solid-state elec-
tronic controllers which regulate the
amount of C02 fed to the growth cham-
bers from C02 cylinders. Radiant energy
and quantum fluxes incident on the
plants, and air, leaf, and root tempera-
tures as well as water vapor and C02
concentrations will be continuously re-
corded. The supply of water and nutri-
ents to the roots will be automatically
controlled.
In addition to the naturally lit growth
chambers, a growth room with artificial
illumination has been constructed during
the year and is now in operation. The
total area of the growth room is 12 m2
with a useful plant growth area of 6 m2.
Light is provided by two banks of Syl-
vania VHO fluorescent tubes supple-
mented with incandescent lamps giving
a maximum useful energy flux of 20 mW
cm-2 (in the waveband 400-700 nm).
The corresponding quantum flux is ap-
proximately equal to 92 nanoeinsteins
cm"2 sec"1. This compares with a peak
noon quantum flux of 200 to 250 nanoein-
steins cm"2 sec"1 reached in the naturally
lit growth chambers on clear summer
days. Thus the maximum light levels ob-
tainable in the artificially lit growth
chamber are only one-third to one-half
of peak solar intensities ; however, with a
light period of 16 hours per day, it is pos-
sible to attain a daily quantum flux ap-
proaching that received by plants grow-
ing under natural sunlight.
The artifically lit growth room uses a
downward vertical air flow. The air
circulation system can be either open or
closed to the external atmosphere. In
closed operation the C02 and water vapor
concentrations can be controlled and
maintained at levels from 200 to 2000
ppm of C02 and from 5 to 20 mm mercury
of water vapor (at 25° air temperature).
Air temperature can be controlled from
17 °C to 50 °C with full light load and
from 10 to 40°C in the dark. The growth
room utilizes the same water and nutrient
supply and C02 control system as the
naturally lit growth chambers. The same
physical parameters are also continu-
ously recorded using two multipoint po-
tentiometric recorders common to all
controlled-growth facilities.
The previously existing artificially lit
growth cabinets, each with a useful plant
growth area of 0.5 m2, have been im-
proved, and all electromechanical control
devices have been replaced with more
precise and dependable electronic solid-
state controllers. Two of these six units
provide simultaneous control of both
carbon dioxide and oxygen concentra-
tion. In future experiments it will also
be possible to grow plants under low
oxygen concentration during the light
period and normal oxygen concentration
during the dark period. This should be
helpful in experiments designed to re-
solve complex effects of low oxygen con-
centration on the overall growth rate in
plants. These effects include an enhance-
ment of photosynthetic productivity dur-
ing the light period and inhibition of non-
photosynthetic processes such as mito-
chondrial respiration during the dark
period.
The main controlling and recording
devices used with all of the controlled
growth facilities are conveniently located
in a central place within the physiologi-
cal ecology complex. Most of the various
sensors and gas sampling streams can
also be readily connected to the analyzers
and computerized data acquisition sys-
tem contained in the mobile laboratory
unit for cross-checking and calibration.
Mobile Laboratory Unit
Precise measurements in the native
habitats and in the transplant gardens of
the response of physiological processes
such as photosynthesis, respiration, and
transpiration require the use of highly
sophisticated instrumentation. A well-
400
CARNEGIE INSTITUTION
equipped mobile laboratory best meets
these requirements. During the year, the
mobile laboratory owned by the Depart-
ment of Biological Sciences at Stanford
University and in the past used jointly
by members of the physiological ecology
groups at Stanford University and the
Carnegie Institution, has been com-
pletely remodeled and equipped to meet
the demands of the new expanded pro-
gram. The vehicle is a converted Dodge-
Travco Motor Home with a total floor
area of 12.5 m2. As now equipped, the
unit houses all of the measuring systems
needed for continuous measurements of
the exchange of carbon dioxide, oxygen,
and water vapor and conductance to the
diffusive gas transport by whole plants
and single leaves, as well as for measure-
ment of external parameters such as
radiant energy and quantum fluxes; air
flow rates and pressures; air, tissue, and
soil temperatures; and water vapor C02,
and oxygen concentrations.
This integrated system also incorpo-
rates an extremely compact computerized
data acquisition, data processing, and
data control system (Compac, NLS, Inc.,
Del Mar, Cal.). This should greatly im-
prove the speed of data acquisition, the
flexibility and convenience of operation,
and the range and accuracy of the meas-
uring system. An especially important
consideration is that on-line data process-
ing should enable us to observe important
plant responses to a change in a certain
factor or combination of factors with
sufficient speed to take appropriate
action. A main limiting factor in our past
work has been that in spite of our best
efforts, we have been behind in data
processing and have therefore often
missed the opportunity to act in time.
As now equipped, the Compac will
monitor and record 200 analog signals
and 32 digital input lines. It also has 32
digital outputs. A large variety of
transducers and sensors producing a
variable voltage or resistance representa-
tion of a physical stimulus such as tem-
perature, light, pressure, and flow can be
directly connected to the analog inputs
for monitoring and entry into the system
processor. The analog-to-digital con-
verter is a 5-digit floating digital volt-
meter with four ranges. It has sufficiently
high sensitivity to accept signals from
very low voltage-generating sensors such
as thermocouples, pyranometers, and
quantum sensors without preamplifica-
tion. The multiplexer consists of a cross-
bar scanner with a scanning rate of 10 to
30 channels per second. The selection of
individual channels is completely ran-
dom so that all, any group, or any one
channel may be monitored and recorded
in predefined sequences under program
control. Data acquisition capacity or
speed, therefore, impose no limitation in
our applications.
In a typical field measurement of pho-
tosynthetic gas exchange, using an open
gas flow system, the Compac serves 10
to 20 thermocouple inputs for tempera-
ture measurements and control as well as
a number of other sensors of various
kinds. These sensors include two Cam-
bridge model 880 dew point hygrometers
for measurement of the water vapor con-
tent of the air entering and air leaving
the leaf chamber; two Beckman model
315A infrared C02 analyzers, one for
monitoring the absolute C02 concentra-
tion in the leaf chamber and the other for
measuring the difference in C02 between
the air entering and air leaving the leaf
chamber; a Westinghouse model 209
zirconium oxide oxygen analyzer for
monitoring the 02 concentration in the
leaf chamber, and in special applications,
for measuring the rate of photosynthetic
02 evolution or respiratory 02 uptake by
the leaf; a Validyne model DP 45 very
low range differential pressure transducer
for monitoring the pressure in and the
flow through the leaf chamber; an Ep-
pley pyranometer and a Lambda
quantum sensor for measurement of en-
ergy and quantum fluxes incident on the
leaf chamber; and a Lambda pyranome-
DEPARTMENT OF PLANT BIOLOGY
401
ter and a quantum sensor for energy and
quantum fluxes received by the leaf in-
side the chamber. The digital inputs
provide a direct interface between the
data processor and digital transducers
such as anemometers and various kinds
of counters as well as switch closures and
limit indicators in the control system.
The central processor consists of a
Data General Nova 1200 minicomputer.
It is equipped with 8191 16-bit words of
core memory, expandable to 32,000 16-
bit words; a real-time clock; and teletype
interface. All communication between
the experimenter and the computer is by
a teletype terminal in the mobile labora-
tory. Program instructions are entered
either directly on the key board or via a
tape reader. Output data are printed on
the teletype and punched on paper tape.
In addition, any four of the analog sig-
nals entering the Compac system can be
simultaneously recorded on two dual-
channel Honeywell Electronic 194 poten-
tiometric recorders. This provides an
independent and continuous recording of
critical variables and is also valuable in
trouble-shooting.
The program language used with the
Compac system, Simpac (Simplified Pro-
gramming for Acquisition and Control) ,
is especially designed for writing real-
time process acquisition and control pro-
grams. It is an extension of the BASIC
language which is widely used as a gen-
eral purpose language in interactive time-
sharing computer systems.
The central processor is used for pro-
gramming scan frequencies, scan inter-
vals, and analog-to-digital converter
ranges; for linearizing or curve fitting —
thus compensating for nonlinearities of
individual transducers or sensors or dif-
ferences between them; for averaging
signals from several sensors of the same
kind and over specified time intervals;
for conversion of input data to absolute
units of measurement; for on-line compu-
tations of actual rates of photosynthesis
and transpiration and conductance of the
stomata to the diffusive transport of
water vapor and C02 which require that
data derived from many diverse inde-
pendent transducers be combined in dif-
ferent equations.
The Compac is also capable of control-
ling discrete events such as opening and
closing relay controls and solenoid valves,
using the 32 digital outputs. By combin-
ing these outputs with inexpensive digi-
tal-to-analog converters, proportional
and integral control of a wide range of
devices using external analog or program-
generated set points can be obtained. In
our application, the digital outputs will
be used for control of leaf temperature
and air flow rate through the leaf
chamber.
The mobile laboratory also houses in-
strumentation for testing, servicing, and
calibration of the measuring systems as
well as refrigerated temperature circu-
lators, gas flow controllers, and com-
pressed gases. Calibration of the C02
analyzers is made with a Wosthoff model
G27/3-F gas mixing pump. This pump
provides a flow rate of over 3 1 min"1
and can also be used for generating gas
mixtures of different C02 concentrations
for photosynthesis measurements. For
calibration of the dew point hygrometers,
air streams of known water vapor con-
tent are procured by passing dry air
through water which is thermostatted at
a temperature of 2-4 °C above the desired
dew point and then through a 2 m long,
double-helix jacketed condenser coil
thermostatted at precisely the desired
dew point temperature, and finally
through the dew point sensors. A thermo-
couple placed in the condenser coil is
connected to the Compac system and the
dew point temperature is automatically
compared with the signals from the dew
point hygrometers. The same device can
also be used to control the water vapor
content of the air entering the photo-
synthesis cuvettes as long as the flow
402
CARNEGIE INSTITUTION
rate does not exceed 3.5 1 min"1. It is
also used for determining the signals pro-
duced by the C02 analyzers due to over-
lapping of the infrared absorption band
of water vapor with that of C02. In this
way, error signals caused by the presence
of water vapor in the gas stream can
automatically be compensated for, per-
mitting entry of the gas stream into the
C02 analyzers without first drying the
gas.
All gas lines connecting the gas moni-
toring system in the mobile laboratory
with the photosynthesis cuvette are of
stainless steel, and they are heated to a
temperature exceeding the highest dew
point temperature in the system. This
minimizes reversible absorption and con-
densation of water vapor and permits
determinations of temperature depend-
ence of photosynthesis without imposing
undesirable differences in the water vapor
pressure gradient between the leaf and
the surrounding air. Plant and leaf
cuvettes specifically designed for the
different plant materials are under con-
struction.
In field experiments where controlled
light fluxes are necessary, natural light
will be excluded and the photosynthesis
cuvette will be illuminated with light
from a Sylvania 1.5 kW Metal Arc lamp.
Neutral optical density screens will be
used for control of irradiance. The cu-
vette control system is designed to permit
precise temperature control in the range
from 5° to 60 °C, oxygen concentration
from 0.2 to 25%, C02 from 10 to 1500
ppm, and water vapor from 10 to 90% of
saturation concentration. In addition to
instrumentation and equipment that are
integral parts of the photosynthesis meas-
uring and data acquisition system, the
mobile laboratory is equipped with in-
strumentation for soil and plant water
potential measurements, including a
Scholander bomb, soil and water con-
ductivity bridges, radioactivity counting
instrumentation for translocation studies.
a light microscope for examination of
leaf anatomy, an analytical balance, de-
vices for determinations of leaf area, a
small centrifuge, a spectrometer, and
laboratory utensils for determination of
leaf chlorophyll and protein contents in
the field. It also provides a limited
amount of laboratory bench space and
cold storage. Other bulky accessory items
such as an oven for drying samples,
additional gas cylinders, and gasoline for
operating the power generator are con-
tained in the mobile laboratory support
vehicle, a Dodge % ton van which also
serves to haul the trailer-mounted power
generator and to transport personnel and
supplies when the mobile laboratory is
in operation.
In order to maintain the instruments
and the experimenters within tolerable
limits when operating in Death Valley,
the mobile laboratory is temperature
controlled. Two independent air-condi-
tioning units with a combined heat-re-
moving capacity of 5600 kcal h"1 (22,000
Btu h"1) have been installed. In the field,
electric power needed to operate the air-
conditioning units and all other equip-
ment that can tolerate voltage and fre-
quency variations comes from a 5 kW
gasoline-powered, water-cooled, AC gen-
erator. In the transplant gardens, line
power will be used for operating other
instrumentation, including the Compac
system. A line corrector (California In-
struments Corp., San Diego, Cal., model
LC-1000 B), mounted in the mobile
laboratory, eliminates voltage variations,
spikes, and other distortions.
In order to avoid duplication of the
high cost of the gas analysis and data
acquisition systems and of time-consum-
ing calibration and maintenance of the
various instruments, the mobile labora-
tory will also be used for kinetic studies
of photosynthesis at the Carnegie labora-
tory at Stanford. In this way the Compac
system can also serve other functions. A
new laboratory extension of the building
DEPARTMENT OF PLANT BIOLOGY
403
that houses the controlled growth facili-
ties has therefore been constructed this
year. It is designed to enable the mobile
unit to "dock" there and become an in-
tegral part of the stationary laboratory.
The new laboratory extension houses a
2.5 kW short-arc, high-pressure xenon arc
lamp (Christie Electric Company, Los
Angeles) and a 1.5 kW Metal Arc lamp
(Sylvania) for illumination of the plant
and leaf photosynthesis cuvettes. Addi-
tional temperature control and other
monitoring and recording equipment for
photosynthesis measurements are also
housed in this extension, as are appa-
ratuses for fluorescence measurements,
determinations of the capacities of com-
ponent reactions of photosynthesis in
isolated chloroplasts, and assays of pho-
tosynthetic enzymes. The signals from
several of these instruments and sensors,
including a liquid scintillation counter,
can readily be fed into the Compac sys-
tem when the mobile laboratory is in
position.
The new transplant gardens, the con-
trolled growth facilities, the mobile
laboratory, and other instrumentation
will certainly provide an excellent and in
several respects a unique tool in the com-
parative study of plant adaptation to
ecologically diverse environments and of
the underlying physiological and bio-
chemical mechanisms.
Acknowledgments. We gratefully ac-
knowledge the cooperation and support
of the Superintendent, Mr. James
Thompson, the Chief Naturalist, Mr.
Pete Sanchez, and the Chief Ranger, Mr.
Harold Thompson of the Death Valley
National Monument, and the Director of
the Bodega Marine Laboratory, Dr.
Cadet Hand, in establishing the new
transplant gardens. We also wish to
thank the National Park Service person-
nel in Death Valley and administrative
and technical personnel at Bodega for
their help with site preparation and con-
nection of water and electricity to the
gardens and for making lodging and other
facilities available to us. The excellent
assistance of Mr. Michael Prather and
Mrs. Nancy Prather in Death Valley
and Mr. Malcolm Erskain at Bodega,
all biologists, in maintaining the trans-
plant gardens during the early critical
stages, is also gratefully acknowledged.
This research program is in part fi-
nancially supported by National Science
Foundation Grants nos. 6B-35854-X and
BB-35855-X to 0. Bjorkman and H.
Mooney.
REMOVAL OF CONTAMINANT INORGANIC
PHOSPHATE AND P H 0 S P H 0 G L Y C 0 L A T E
FROM RIBULOSE-l,5-DIPHOSPHATE
Joseph A. Berry and George Bowes
Commercial preparations of ribulose-
1,5-diphosphate contain appreciable in-
organic phosphate and in one case trace
contamination with phosphoglycolate
was detected. These compounds interfere
with methods for the assay of RuDP oxi-
dation which we have been testing. A
procedure was devised to remove these
contaminants by enzymatic hydrolysis
followed by gel filtration chromatog-
raphy. Since this method results in good
recovery of RuDP and since the chroma-
tographic separation may be of more
general value, the procedure is described
here.
RuDP was assayed by reacting 10 p]
aliquots to completion with NaH14CO:^
(0.1 ^ci/yuinole) and RuDP carboxylase
in the reaction mixture of Bjorkman and
Gauhl (1969). Phosphate was assayed
by the procedure of Fiske and Subbarow
(1929). Phosphoglycolate was assayed
404
CARNEGIE INSTITUTION
by conversion to glycolic acid with phos-
phoglycolate phosphatase. Glycolic acid
was separated from other compounds on
a 0.8 X 2 cm column of Dowex-1-formate
and eluted with 10 ml of 0.2 N formic
acid. This was taken to dryness in a
gentle air stream and glycolate was de-
termined by the method of Calkins
(1943). Phosphoglycolate phosphatase
was prepared from spinach chloroplasts
according to Yu et al. (1964). Only a
trace of phosphatase activity was de-
tected with 3-phosphoglyceric acid or
RuDP in place of phosphoglycolate.
RuDP (7.8 /miole) in a volume of 0.5
ml was incubated for 30 min at 30 °C
with sufficient phosphatase to hydrolyze
0.1 /xmole of phosphoglycolate/min. The
solution was then placed on a column of
Sephadex G-25 fine and eluted with 5
ml NaCl, 5 ml Tris-HCl (pH 8.4).
Fractions of 2 ml each were collected and
assayed. Figure 41 shows the separation
obtained. Blue dextran 2000 (Pharmacia
Fine Chemicals) and presumably other
large molecules such as proteins eluted
ahead of RuDP which was also sepa-
rated from inorganic phosphate. Glyco-
late eluted together with inorganic phos-
phate in preliminary experiments in
which enough glycolate was added to
allow convenient assay. The fractions
containing RuDP were pooled and freeze
dried. Recovery was greater than 80%
of starting material. The major loss
probably occurred during enzymatic
hydrolysis. The recovered RuDP was
essentially free of inorganic phosphate
and phosphoglycolate; however, it did
contain NaCl and Tris-HCl from the
buffer. A reasonably high ionic strength
is necessary for successful separation
with this technique. Other salts or a
volatile buffer could be used for other
purposes. Treatment with phosphoglyco-
late phosphatase was necessary because
phosphoglycolate was not separated from
RuDP on the column.
In this separation RuDP (MW = 398)
behaves as if it were a larger molecule
(MW > 1000) . In preliminary experi-
ments with Sephadex G-10, which on the
basis of molecular weights of RuDP,
phosphate, and glycolate seemed the
more reasonable choice, we found very
1.2
1.0
J 0.8
"i — i — i — i — i — i — i — r
RuDP
Blue dextran
2000
j l i_i i i I i i i i l i i i i
o
0.06 O
0.04 co
0.02 | 8
n CD
40 50
Fraction number
Fig. 41. Elution of "Blue Dextran 2000," ribulose-1 ^-diphosphate (RuDP), and inorganic phos-
phate (Pi) from a column of Sephadex G-25 Fine. 0.5 ml containing 0.2 mg of blue dextran, 7.8
Mmoles of RuDP, approximately 0.2 fimoles of hydrolyzed phosphoglycolate, and 4.1 /mioles of Pi
was added to the column. Fractions (2 ml) of the eluant (5 ml NaCl, 5 ml Tris-HCl, pH 8.4)
were collected.
DEPARTMENT OF PLANT BIOLOGY
405
little separation. Both RuDP and phos-
phate eluted in the void volume. Appar-
ently these anions are excluded from gels
which are permeable to uncharged mole-
cules of the same molecular weight. This
is probably related to the presence of
some negative charges on the dextran
polymer that forms the gel.
References
Bjorkman, 0. E., and E. Gauhl, Planta, 88,
197, 1969.
Calkins, V. P., Anal. Chem., 15, 767, 1943.
Fiske, C. H., and Y. Subbarow, J. Biol.
Chem., 81, 629, 1929.
Yu, Y. L., N. E. Tolbert, and G. M. Orth,
Plant Physiol, 39, 643, 1964.
OXYGEN UPTAKE IN VITRO BY RuDP
CARBOXYLASE OF C hi amy d omonas reinhardtii
Joseph A. Berry and George Bowes
Last year studies of the effect of 02 on
the photosynthesis of the alga Chlamy-
domonas reinhardtii were reported ( Year
Book 71, 148-158). The presence of 02
stimulated excretion of glycolic acid to
the medium. Since this compound is
formed at the expense of photosynthetic
metabolism, it should result in a net
inhibition of the rate of photosynthesis.
However, 02 inhibited photosynthesis of
Chlamydomonas to a much greater ex-
tent than could be accounted for by the
quantity of glycolic acid recovered, sug-
gesting that 02 affects photosynthesis in
another way in addition to causing glyco-
late formation. On the basis of kinetic
studies it was proposed that 02 causes a
direct inhibition in vivo of the enzyme
ribulose-1 ,5-diphosphate carboxylase.
Such an inhibition has been well docu-
mented by studies of RuDP carboxylase
in vitro, isolated from higher plants and
Chlamydomonas (Bowes and Ogren,
1972; Year Book 71, 151). A competitive
interaction between 02 and C02 occurs,
both in photosynthesis in vivo and with
the isolated enzyme in vitro. There is
abundant evidence that the activity of
RuDP carboxylase is one factor limiting
the maximum photosynthetic rate of
higher plants and Chlamydomonas.
Competition between 02 and C02 for
binding to the enzyme in vivo could ex-
plain the kinetics of 02 inhibition of
photosynthesis. This postulate does not
require that 02 react further with the
RuDP-enzyme complex in order to in-
hibit photosynthesis. Thus, 02 could
inhibit photosynthesis without actively
participating in photosynthetic carbon
metabolism. In this respect it is funda-
mentally different from inhibition due to
photorespiration or to glycolate forma-
tion, as the latter type involves 02 as a
reactant. The results obtained last year
indicate that a passive inhibition could
explain most of the effect due to 02 in
this alga. There is no doubt, however,
that glycolate was formed by Chlamy-
domonas and that glycolate formed by
higher plants is metabolized (photo-
respiration) , resulting in C02 production.
The mechanisms which lead to these
processes also contribute to the inhibition
of net photosynthesis by 02. The mecha-
nism which leads to glycolate formation
in vivo is an essential step in photorespi-
ration, but it is not clearly understood.
Bowes et al. (1971) demonstrated that
RuDP carboxylase in vitro catalyzes an
oxidation of RuDP by 02 to yield phos-
phoglycolic acid, a precurser of glycolic
acid, and 3-phosphoglyceric acid. An-
drews et al. (1973) and Lorimer et al.
(1973) have extended these studies to
demonstrate the incorporation of 1802
into phosphoglycolate and the formation
of 14C labeled phosphoglycolate and 3-
phosphogly cerate from labeled RuDP.
They also studied some aspects of the
kinetics and specificity of the reaction.
They report that the pH optimum for
406
CARNEGIE INSTITUTION
oxidative activity is 9.4, while that for
carboxylase activity is 7.8. At pH 9.4
the specific activity of the enzyme for
phosphoglycolate formation was 23% of
that for C02 fixation at pH 7.8. There is
little doubt that oxidation of RuDP does
occur in vitro, and that it is catalyzed by
RuDP carboxylase. It has been sug-
gested that this reaction is responsible
for glycolate formation in vivo. How-
ever, no quantitative comparison be-
tween glycolate formation in vivo and
the activity for RuDP oxidation in vitro
has been reported.
Accurate measurements of glycolate
formation relative to photosynthesis of
Chlamydomonas were reported last year.
At 50% 02 and 2 mM bicarbonate (refer
to Fig. 40, p. 153, Year Book 71) the
rate of glycolate formation was 4.8
^moles of glycolate (mg chl)_1hr_1. These
algae were able to sustain a maximum
rate of photosynthesis of 371 jumoles
(mg chl)~1hr~1. If it is assumed that both
rates are limited by the activity of RuDP
carboxylase, then the oxidation activity
of the enzyme in vitro should be 1.3% of
the maximum carboxylation activity un-
der comparable conditions in order to
account for the rate of glycolate forma-
tion measured in vivo. The ratio of spe-
cific activities reported by Andrews et al.
(1973) is not directly comparable be-
cause these activities were measured at
different pH's and with enzyme from
another organism, spinach. Clearly the
ratio of the two activities must be meas-
ured at the same pH. However, the rela-
tive ratio of the two activities can be
expected to be a function of the pH
chosen, as the pH optima of the two
activities differ. The pH at which RuDP
carboxylase functions in vivo is not
known; however, it seems reasonable to
assume that it is near the pH optimum
for C02 fixation. Thus, we have made
our comparison at 50% 02 and pH 7.8,
the optimum for RuDP carboxylase.
RuDP carboxylase was extracted from
Chlamydomonas reinhardtii cells (Table
20) as described previously {Year Book
71 j 140), and the fraction precipitating
between 35% and 55% (NH4)2S04 was
redissolved in 0.1 M phosphate buffer
(pH 7.6) containing 5 mM mercapto-
ethanol and 0.1 mM EDTA. Aliquots of
this solution were stored frozen. Assays
for RuDP oxidation were performed in
50 mM HEPES buffer (pH 7.8), 10 mM
MgCl2, and 1 mM EDTA. This buffer
was prepared with carbonate-free NaOH
and stored under a C02-free atmosphere.
The enzyme preparation also contained
contaminating bicarbonate which se-
verely interfered with assays of RuDP
oxidation. The enzyme was therefore
transferred to C02-free buffer by treat-
ing it on a small Sephadex G-25 column
equilibrated with the assay buffer. As-
says were performed in 1 ml total volume
containing 1 to 2 mg protein with an
oxygen electrode (Rank Brothers, Bot-
tisham, Cambridge, England) . Temper-
ature was controlled at 30 °C. Sensitivity
TABLE 20. Comparison of the Ratio of C02 Fixation, to Glycolate Forma-
tion by Chalamydomas reinhardtii in vivo to the Ratio of C02 Fixation to
Oxidative Activity of RuDP Carboxylase Isolated from the Same Alga*
Glycolate
Formation in
vivo or O2
Uptake in
CO 2 Fixation vitro Ratio
Condition
Alga in vivo
Enzyme in vitro
371*
453f
4.8*
5.1f
1.3%
1.1%
* units = /xmole (mg chl)-1 hr._1
t units = nmole (mg protein)-1 min.-1
* Measurements were made at 50% 02 in the gas phase.
DEPARTMENT OF PLANT BIOLOGY
407
of the electrode to the small changes in
02 concentration caused by O2 uptake
was achieved by applying a bias poten-
tial appropriate to yield a zero signal at
the starting 02 concentration. This al-
lowed recording at expanded sensitivity.
Rates as low as 1 nmole 02 min-1 could
be detected. The background rate of 02
uptake by the assay system (before addi-
tion of RuDP) was typically 4-5 nmole
02 min"1. When RuDP (0.25 /xmole) was
added, there was a rapid increase of 02
uptake to 12 nmoles 02 min"1 or greater.
This rate was linear for the first 2-3
minutes, but gradually declined with
time. Both RuDP and enzyme were re-
quired for stimulation of 02 uptake. The
stimulation of 02 uptake resulting from
RuDP addition ceased when 10 /xmole of
NaHCOs was added, presumably due to
rapid removal of RuDP by carboxyla-
tion. With the enzyme preparation from
Chlamydomonas, no stimulation of 02
uptake was detected when glycolate or
phosphoglycolate was added in place of
RuDP, indicating that the oxidation does
not continue to glyoxylate formation.
The rate of carboxylation was meas-
ured as described previously ( Year Book
71, 151). Boynton et al. (1972) report
that these conditions yield maximal rates
of carboxylation with RuDP carboxylase
from Chlamydomonas. Comparison of
rates of 02 uptake at 0.55 ml 02 to C02
fixation, measured on the same prepara-
tion of RuDP carboxylase, yielded a rate
of oxidative activity relative to carboxyl-
ating activity of 1.1-1.2%. This is simi-
lar to the relative rate of glycolate for-
mation compared to total photosynthetic
capacity obtained in vivo. It is appar-
ent that the enzyme is not stable under
the assay conditions. Thus, the rates
obtained are surely an underestimate of
oxidative activity. The results obtained
in vitro suggest that RuDP oxidation in
vivo could account for the glycolate
formed. However, since the pH optima
for carboxylase activity and oxygenase
activity differ, the relative rates will be
a function of the pH chosen. The pH at
which this reaction proceeds in vivo is
not known; hence the choice of pH 7.8 is
conservative but nevertheless arbitrary.
References
Andrews, T. J., G. H. Lorimer, and N. E.
Tolbert, Biochemistry, 12, 11-18, 1973.
Bowes, G., and W. L. Ogren, J . Biol. Chem.,
257, 2171, 1972.
Bowes, G., W. L. Ogren, and R. H. Hageman,
Biochem. Biophys. Res. Comm., 1+5, 716—
722, 1971.
Boynton, J. E., N. W. Gillham, and J. F.
Chabot, «/. Cell. Sci., 10, 267-305, 1972.
Lorimer, G. H., T. J. Andrews, and N. E.
Tolbert, Biochemistry, 1, 18-23, 1973.
BIBLIOGRAPHIC INFORMATION RETRIEVAL
J. S. Brown
The problem of storing and retrieving
information from the vast volume of sci-
entific literature is well known. This
problem was mentioned previously ( Year
Book 68, p. 565) as it affects photo-
synthesis workers in particular. Our card
filing system has long since gone out of
date because at the rate of about 500
card entries a year, it soon became un-
manageable to sort and entries are now
filed alphabetically by first author. This,
of course, makes impossible the retrieval
of papers by other than first author.
We have taken advantage of an infor-
mation retrieval system (Spires) devel-
oped at Stanford University in October,
1972. This system is ideally suited for
our use and, in addition, can absorb our
previous card file system. General infor-
mation about how to use the system may
be found in the Spires User's Manual for
sale by the Systems Documentation
408
CARNEGIE INSTITUTION
Library, Campus Computing Center,
Stanford, California 94305.
Description of "Plantbio" File
Our file, called "Plantbio," is stored in
one of the interactive systems on the
IBM 360/67 computer of the Campus
Computing Center. Bibliographic data
are added to this file from a terminal in
our laboratory. The actual time involved
per reference is shortened because no
holes need be punched as required in a
manual system and because certain ab-
breviations may be used. A typical input
format is as follows :
IDEN = FREC 72.1 ;
AUTHOR = C. S. FRENCH/J. S.
BROWN/M. C. LAWRENCE;
AFFILIATION = CARNEGIE INST.,
STANFORD, CA. 94305;
TITLE = FOUR UNIVERSAL FORMS
OF CHLOROPHYLL A.;
CITATION = PLANT PHYSIOL. 49,
421-29 (1972) ;
CATEGORY = E-5, F-l;
NOTES = Anything may be entered
here.
The first word in each line is called an
element. In our file, only the first (key)
element is required. The key element is
unique to each reference, and the system
will not accept another reference with the
same key. All of the elements may be
abbreviated to one or two letters, and the
equals sign may be omitted on input.
The semicolon separates elements so that
when several lines are needed for any ele-
ment, the semicolon is placed at the end
of the last line before the next element.
The categories come from the thesaurus
shown in Table 21. This thesaurus was
begun for our card file in 1965 but has
nearly doubled in size since then. There
is virtually no limit to the size it can
become, so that fields of plant biology
other than photosynthesis can be easily
added and subdivided.
The references may be entered into
Spires one at a time, or any number of
them may be gathered into a data-set
and batched into Spires. At the moment
there is little difference in cost because
the system only updates during the over-
night service block.
Searching
The Spires System will accommodate
searching by any or all of the words in
all of the elements entered. However, the
major cost of the whole system comes
from building records which allow search-
ing by many elements. Therefore, we
have elected to search only by author
and category. One may "find" any au-
thor or category or combination of au-
thors and categories. The computer will
answer with the number of papers that fit
the "find" command. One then gives the
command "type" and these references
will be typed out in full as they were
entered. If the number of found refer-
ences is large, one may choose to "type"
only one or two elements (e.g., title)
from each reference and then choose the
more pertinent ones for listing in full.
We can omit the ability to search by title
word and still be efficient because of the
effort expended by a scientist who as-
signed the thesaurus categories to a
reference before it was entered.
In addition to on-line searching, we
plan to have all the records printed out
by author and category once or twice a
year. At this writing we have a listing
of about 1500 records alphabetically by
first author. With similar listings as the
file grows it will be possible to do some
searching when a computer terminal is
unavailable.
Computer Cost
The cost of this program is not exces-
sive when weighed against a scientist's
time and the great usefulness of the
system. Currently it costs about 10 cents
of computer time to add each record into
the Plantbio file. Since we are collecting
records at the rate of about 600 a year,
the file building cost will be about $60.
DEPARTMENT OF PLANT BIOLOGY
409
03
03
«p
o3
a 8 8
o-£;S
go 2
M
02
O
cu
3
©S O
-2 .22
CO M
5ft
.2
*C
c 73
2 fl
■3 2
03 43
'3,3
02 . T3
<N
so
©
-*2
e
« S
<u O
03
§42
1 s
6°
3
o
•—i
O
02
-P
>>.22
02
CD
'a
43
03
02
43
03
o
p
'oa
O
>j
X 43
ft a
■p 3
22 S
s.S
03 a
^3 _
o .2
"3 M
o33
bfi O
O 42
73
O
43
-P
sit
3 a
2 8
"-P b£
o3 o
53 ^
o3 o3
CO
73
©
•*i
OQ
3
O
02
3
° -I
o3
a ^
9 1
"oS
a «
a
0)
t^
43
ft
o3 .
bC
O
iH <N CO •<* iO CO
'■** oo M
Si
a
o
g -p -p
"-p ° S?
ft43 ^
O ft 03
02
02
-P
33
ft
C
o
73
3
o3
e3
a
03
43
"t3
5
>
03
■a -a °
8 c3
•3 33
^ Q O
IB U K
3 I a
03 CO
■P 8
3 £
02 S>
3^
<M CO Tfi ic CO jg O
03
o3
4h k g
o3 '43 »
©s
©>
©
O 02
's? 3
"° 5
=0-2
II
8 ©
© 02
O
43
ft
03
P 02
- O Q
03 +3 -P -P „
„ >> O >. c K
l_L rj 43 43 *. "S
WS a ft^ £
03
bf)
2 Oj
■S a
>< 43
3
O
03
c
c3
03
U
3 ft
a c
5 o
CO
03
43
-P
3
>>
<y&
m
o
43
Oh
o
-t-3
02
03
3
03
Sh
0
03
P3
O
03
"o
42
-^>
3
o3
02
3
s-
3
o3
02
03
43
H
>5
-t-3
'>Ph
^?^
•^ 3 feS
::3 g-O
-O O 3
® n *
-2 og
03 a n
73 ftW
03
T!
3
'-3
3
c3
03
a
02
0)
Fi
>i
03
N
43
3
C3
0J
02
3
o3
43
o
02 03
.§ a
03
3
o3
t-
42
a
03
a
a s
N00OJO h(N
02
O *3 -P
° 3 g
3 §3
■s a »
03 w Q3
3 ^
03 "2
■P .«
o ^>
03
CO
o
43
C3
02 3
a &
02 3
(ii .3
3
o3
>j
o
O
■P
02
o3
3 02
03 >< C
3 O -3
3 93
ft
M C «
03
Ph
- o3 ft i—,
ftq=| 43 K
xn
3
02
02
03
IB
03
02 >
o ^
2 ^
O o3
O 02
03 q
C3 -2
oi.a
02
Ih C3
03 02
43 to
O 03
3 „
£o
■p
03
43
+j
3
02
03
3
e3 .
bC 2
3 g
03 OS
O 44
So'hNM^' iO«ON00050h
W
8
c
43
ft
3 ^
O .3
3 43
03
03
bC
"33
03
03 3
bC 03
03
o3 ~
s-i b£
2 a T
43 03 03
bC 02 3
bC42
hSg
o
3
03 *
o3 C
bC S
73 <w
03 f«-i>
73 Is
b O
03
03
bC
"S 2
03 •-
i- 02
02 P
43 S
-i— o3
O 42
02
I
O
ft
>>
43
o3 v
JSP 8
03 o
5^
3
C
33
3
i— < «» _S »^j
w
3 43
o3 -fj
33 o3
03 "
02
02 Tfl O
(NCC*iOON
oo ci e d ^ ^ c
w
C 02
fi o
o ft
£ o
S C3
O 3
t-h (M
H
8
«W 02
C35 ^
I 03
.2 3
s* 3
co o3
-< 02
8 °
§ ^-3
« p § a
1'So ,
03
03
ft
03
>
g.s
o3 -p
ft 3
So
o >
02 02
02
3
O
• |-H
-4^
o3
fi
.3° 43 +s O
O 02 03
M tj »H -4J
« 5 o S p
| o o
C35
b/D «h
ft C3
42 »
^3
ft 3
5-
o
©
02 -si
go «
3^ ©
« 8 03
>> e £
n co 42
>.
o
o3
"oj
03
h (N ^ O h (N M
1 8
8 fi
tjh io g d
02 P
3^
? °
is 43
02
03
X
^03
'E,
fi
o
o
43
ft
O
1 2-
03 02 '+= 43
02 «
O 02
3-^
3 -P
O o3
02 ft 02 3
(N CO TfH IO
P 3
3 g
03 §
£ 3
ft S>42
O P. 33
— < © 03
CD C C
> CO 02
02 © '3
o3
fi £
£ C2
5 02
03 ft
ft ^
.2 S
-t-= 02
Sh CD
02 O
-° 4^
03 C3
o3
t-i
-4-3
03
03
ft
02
02
bC
3
fi ^
+3 02
° ft
02 B
ft.2
03
-4J
02
02
ft
02
02
bC
3
03
43
02
03 '
02
3
02
0.
■+3
03 ,_
3 o3
5.S
o
43
ft
O
C3 C2
3 03 a)
o i^ ft
43 cd
o3 jj
ft
o
,2 43
ft 3
. bC P
^0^^43 03
© o ° '
02 O
-° 3
03 C3
„ 3
a. 2
CO E C
o S3 .2
f-i ft -p
43 02 02
03 ^^
-4^
5 s O ^
C o r » oo
03
J2 73
CO ^d
(M CO rfi iO
CO t>^ CO
8
42 -3
^3 CD
ao
02
o
03
02 to
b£ 42
o3 £
•g 3
>< ^
03 &?
P I
a -^
o. .w
•^3 03
02 3
.££
-^ w
02 bC
b£ 3
■J.
3 03
03 -3
~z
43 C3
u 02
K^>
3 2
a
S 3
.2 03
3
—
-P 02
— 02
3
c t-
o3
K C
r
42 3
o* cn
O -h <N CO -7- if* CO
<!PQO
W
O
410
CARNEGIE INSTITUTION
The storage cost will vary with the time
of on-line accessibility that is desired.
If we wish to access 3000 records during
two days a week, the storage cost would
be about $800 a year. A 20 min search-
ing session currently costs about $5. This
is probably longer than the average
search request in which only one or two
authors and/or categories are wanted.
A listing of all the records in the file is
currently about $30 per 1000 records.
Inter department Users
Currently the Plantbio file is public
and can be searched by anyone at a
nonprofit institution with a valid Stan-
ford account number and an acoustic
coupler connection by telephone to the
Campus Computing Center. We hope
that other researchers in photosynthesis
and eventually in broader aspects of
plant biology will make use of this bib-
liographic file. Eventually, means will be
found to allow others to share in the
costs and building of the file as well as
in the searching that can be done now.
We may be able to send print-outs of
the file by author or by selected cate-
gories for a nominal charge if the number
of such requests is not too large. Of
course the value of the file depends on
the total number of references in it. As
of November 1, 1973, there are 4100 ref-
erences in Plantbio, and we plan to add
about 50 to 60 new records each month.
Acknowledgment
We are very grateful to Mr. John R.
Schroeder of the Stanford Computation
Center for creating the file definition for
Plantbio and patiently helping us to
learn how to build and use it.
STAFF ACTIVITIES
The Department is particularly fortu-
nate in that Dr. Winslow Briggs, Pro-
fessor of Biology at Harvard University,
will become the Director on July 1, 1973.
Professor Briggs is a specialist in plant
development and tropisms. He has done
much biochemical work on the plant pig-
ment phytochrome, which functions in
plants to control many reactions such as
flowering and photoperiodic responses.
In addition to maintaining strong groups
in photosynthesis and in physiological
ecology, Dr. Briggs plans to add a plant
development group to the Department.
This expansion will involve moderniza-
tion of the present building and the addi-
tion of new space. Plans for this work
are being prepared by the architectural
firm of Spencer, Lee, Busse, and Stypula,
and construction is expected to start in
the late fall of 1973.
During this report year a grant was
received from the National Science Foun-
dation for a joint study by Dr. Bjorkman
and Dr. Harold Mooney of Stanford Uni-
versity on the adaptation of plants to
contrasting environments. This enter-
prise has necessitated the addition of a
laboratory room, two naturally lit con-
trolled-growth units, and one artificially
lit growth room to the building housing
the plant growth chambers. Further-
more, new experimental gardens have
been established at Bodega Bay and in
Death Valley. The mobile laboratory,
operated jointly with Dr. Harold
Mooney of Stanford University, has been
completely remodeled and a computer-
ized data acquisition system has been
installed in it. Dr. John Troughton, De-
partment of Scientific and Industrial Re-
search, New Zealand, has joined the
group as a Research Fellow to partici-
pate in the field and laboratory work.
Dr. Troughton, Dr. Berry, and Dr.
Mooney of Stanford University took a
field trip to the Pacific coast of Baja
California, Mexico, in May. This cool
coastal desert abounds in cacti and other
succulent vegetation. Field measure-
ments were conducted and material was
collected for carbon isotope studies.
DEPARTMENT OF PLANT BIOLOGY
411
The use of the facilities of the Mather
and Timberline experimental stations by
various university classes has increased
during the past year. The evolution class
from the State University of California
at San Jose, led by Dr. Clifford Schmidt,
visited the stations during October. Dr.
John Thomas and his class in Plant
Taxonomy from Stanford University
conducted field studies in May at the
Mather Station. The Plant Taxonomy
class from the State University of Cali-
fornia at Sacramento, under Dr. Mike
Baad, used the station at Mather June
9-15 during their study of the flora of
the mid-Sierra. Mr. Richard Papp, a
graduate student from the University of
California at Berkeley, continued his
studies on the ecology of the insects
found at high elevations at the Timber-
line station. Dr. Craig Heller of Stan-
ford University and his graduate stu-
dents made a preliminary survey of the
rodents of the Timberline area and a
transect on the eastern escarpment dur-
ing the late summer and fall in prepara-
tion for an intensified study to be car-
ried out at Timberline during the
summer of 1973. Periodically during the
year Professor Theodosius Dobzhansky,
Dr. A. Ayala, and associates from the
University of California at Davis have
continued their long-term evolutionary
studies on the fruit fly Drosophila.
Dr. Olle Bjorkman, at the invitation of
the Israeli Government, presented a lec-
ture and was chairman of a session of a
meeting on Plant Response and Adapta-
tion to Environmental Stress in late De-
cember 1972. The conference, sponsored
by the National Council for Research
and Development for Israel, was held at
the Dead Sea. Dr. Bjorkman also pre-
sented a paper at the Brookhaven Sym-
posium on Basic Mechanisms in Plant
Morphogenesis in early June 1973.
Drs. Brown and Hiyama attended the
Biophysical Congress in Moscow and the
Photobiological Congress at Bochum,
Germany, in August 1972. They had the
opportunity to renew and extend contacts
with colleagues at both meetings. After
the meetings Dr. Hiyama visited labora-
tories in Leiden, Goteborg, Gottingen,
and Hannover.
Drs. French, Berry, and Bowes at-
tended the Society of American Plant
Physiologists' annual meeting in con-
junction with the AIBS at Minneapolis
in late August.
Dr. French was given the Award of
Merit at the annual meeting of the Bo-
tanical Society of America, at Amherst,
Massachusetts, on June 20, 1973.
BIBLIOGRAPHY
492* Bjorkman. Olle. Comparative studies on
photosynthesis in higher plants. In Cur-
rent Topics in Photophysiology , S, A.
Giese, ed., Academic Press, New York,
pp. 1-63, 1973.
488 Brown, J. S., Action spectra for partial re-
actions of photosynthesis. In Chloroplast
Fragments, G. Jacobi, ed., Gottingen, pp.
3-16, 1972. Proceedings of Workshop on
Chloroplast Fragmentation, University of
Gottingen, June 1971.
490 Brown, J. S., Separation of photosynthetic
systems 1 and 2. In Current Topics in
* CIW Department of Plant Biology publica-
tion number.
Photophysiology, S, A. C. Giese, ed.,
Academic Press, New York, 97-112. 1973.
461 Fork, David C, Oxygen electrode. In
Methods in Enzymol., 24, A. San Pietro.
ed., Academic Press, New York, 113-122,
1972.
489 Fork, David C, Light-induced absorbance
changes in subchloroplast particles. In
Chloroplast Fragments, G. Jacobi. ed.,
Gottingen, pp. 79-107. 1972. Proceedings
of Workshop on Chloroplast Fragmenta-
tion, University of Gottingen, June. 1971.
French, C. Stacy, see Virgin, Hemming I.
French, C. Stacy, see Gasanov, R. A.
412
CARNEGIE INSTITUTION
493 Gasanov, R. A., and C. S. French. Chloro-
phyll composition and photochemical ac-
tivity of photosystems detached from
chloroplast grana and stroma lamellae.
Proc. Nat. Acad. Sci., 70, 2082-2085, 1973.
487 Virgin, Hemming I., and C. Stacy French.
The light induced protochlorophyll-
chlorophyll a-transformation and the suc-
ceeding interconversions of the different
forms of chlorophyll. Physiol. Plant., 28,
350-357, 1973.
SPEECHES
Berry, Joseph A., and George Bowes, Effects of
Light Intensity on the Production of Glyco-
late during Photosynthesis by Chlamydo-
monas. American Society of Plant Physiolo-
gists, University of Minnesota, Minneapolis,
Minnesota, August 31, 1972.
Berry, Joseph A., Concepts of the Mechanism
of 02 Evolution in Photosynthesis. Depart-
ment of Biological Sciences, Stanford Uni-
versity, Stanford, California, October 25, 1972.
Berry, Joseph A., The Reductive Pentose Phos-
phate Pathway of Photosynthetic C02 Fixa-
tion. Department of Biological Sciences,
Stanford University, Stanford, California, Oc-
tober 27, 1972.
Berry, Joseph A., The d Pathway of Photo-
synthetic CO2 Fixation. Department of Bio-
logical Sciences, Stanford University, Stan-
ford, California, October 30, 1972.
Berry, Joseph A., Adaptive Differentiations of
the Photosynthetic Apparatus. Department
of Biological Sciences, Stanford University,
Stanford, California, November 1, 1972.
Berry, Joseph A., Oxygen Inhibition of Photo-
synthesis. Botany Tea, University of Cali-
fornia, Santa Cruz, California, February 1,
1973.
Berry, Joseph A., Physiological Ecology and
Biochemical Research at the Carnegie In-
stitution. First Year Graduate Students'
Seminar, Stanford University, Stanford, Cali-
fornia, May 1, 1973.
Berry, Joseph A., see also Bowes, George.
Bjorkman, Olle, Photosynthetic Adaptation to
Ecologically Diverse Environments. Seminar,
Department of Botany, University of Cali-
fornia, Riverside, California, November 8,
1972.
Bjorkman, Olle, Environmental and Evolution-
ary Aspects of C4 Photosynthesis. Lecture,
Reed College, Portland, Oregon, November
10, 1972.
Bjorkman, Olle, Photosynthetic Adaptations of
Rain Forest Plants. Stanford-Carnegie Plant
Physiology Graduate Seminar, Department of
Plant Biology, Carnegie Institution, Stanford,
California, November 15, 1972.
Bjorkman, Olle, Irradiance as a Stress Factor in
Photosynthesis. Israel Scientific Conference
on Plant Response and Adaptation to En-
vironmental Stresses. Ein Bokek, Israel,
December 20, 1972.
Bjorkman, Olle, d Photosynthesis and Its
Functional Significance. Charles F. Kettering
Research Laboratory, Yellow Springs, Ohio,
January 17, 1973.
Bjorkman, Olle, Photosynthetic Characteristics
of Plants from Different Light Climates.
Botany Tea, University of California, Santa
Cruz, California, February 22, 1973.
Bjorkman, Olle, Photosynthetic Function in
Relation to Leaf Structure. Brookhaven
Symposium on Basic Mechanisms in Plant
Morphogenesis. Brookhaven National Lab-
oratory, Upton, New York, June 5, 1973.
Bowes, George, and Joseph A. Berry, Effects of
02 and CO2 on Photosynthesis and Glycolate
Production by Chlamydomonas. American
Society of Plant Physiologists, University of
Minnesota, Minneapolis, Minnesota, August
31, 1972.
Bowes, George, Characteristics of the C3 and C4
Photosynthetic Pathways. California State
University, San Jose, California, November
21, 1972.
Bowes, George, see also Berry, Joseph A.
Brown, Jeanette S., Analyses of the Chloro-
phyll Absorption Spectra of Phaeodactylum
and Dunaliella. IV International Biophysics
Congress, Moscow, U.S.S.R., August 9, 1972.
Brown, Jeanette S., Photosystems 1 and 2 in
Fractions of Synchronously Grown Chlamy-
domonas. VI International Congress on
Photobiology, Bochum, Northrhine-West-
phalia, Germany, August 22, 1972.
Brown, Jeanette S., Separation of Photosyn-
thetic Systems One and Two. California
State University, San Jose, California, Sep-
tember 27, 1972.
Brown, Jeanette S., Biological Forms of Chloro-
phyll a and Separation of the Photosystems.
Photobiology Seminar, Department of Bio-
logical Sciences, Stanford University, Stan-
ford, California, June 6, 1973.
Fork, David C, Light Reactions and Photo-
chemistry of Photosynthesis. Department of
Biological Sciences, Stanford University,
Stanford, California, October 18, 1972.
DEPARTMENT OF PLANT BIOLOGY
413
Fork, David C, Electron Transport in Photo-
synthesis. Department of Biological Sciences,
Stanford University, Stanford, California,
October 20, 1972.
Fork, David C, Photosynthetic Phosphoryla-
tion. Department of Biological Sciences,
Stanford University, Stanford, California,
October 23, 1972.
Fork, David C, Observations on the Function-
ing of Cytochrome b 559 in Grana Prepara-
tions from Spinach. Current Topics in Photo-
synthesis Seminar, Laboratory of Chemical
Biodynamics, University of California, Berke-
ley, California, March 14, 1973.
Fork, David C, Photosynthesis Research at
the Carnegie Institution. First Year Graduate
Students' Seminar, Stanford University, Stan-
ford, California, May 15, 1973.
Fork, David C, The Photosynthetically Active
Pigments, Action Spectra and Energy Trans-
fer in Photosynthesis. Photobiology Seminar,
Department of Biological Sciences, Stanford
University, Stanford, California, May 30,
1973.
French, C. Stacy, and Hemming I. Virgin, Ab-
sorption Spectra of Individual Forms of
Protochlorophyll and of Chlorophyll a as
Derived from Measurements of Barley in
Early Stages of Greening. American So-
ciety of Plant Physiologists, University of
Minnesota, Minneapolis, Minnesota, August
31, 1972.
French, C. Stacy, Trends, Fads and Progress in
Photosynthesis Research, 1900-1973. Univer-
sity of California, Berkeley, California, April
18, 1973.
Hiyama, Tetsuo, Quantum Yield and Action
Spectrum of P700. IV International Bio-
physics Congress, Moscow, U.S.S.R., August
8, 1972.
Hiyama, Tetsuo, Secondary Reactions of
Photosystem 1 : Interactions of P700 and
P430 with Plastocyanin and** Cytochromes.
VI International Congress on Photobiology,
Bochum, Northrhine-Westphalia, Germany,
August 22, 1972.
Hiyama, Tetsuo, A Comment on Primary Ac-
ceptor of Photosystem 1. Postcongress Sem-
inar on Photosynthesis, sponsored by Bayer-
Leverkusen, Haus Hammerstein, Wuppertal,
Germany, August 26, 1972.
Hiyama, Tetsuo, Primary Reactions of Photo-
synthesis. Botanical Institute, University of
Goteborg, Goteborg, Sweden, August 29, 1972.
Hiyama, Tetsuo, P430: The Primary Acceptor
of Photosystem 1. Biophysical Laboratory,
The State University, Leiden, The Nether-
lands, August 31, 1972.
Hiyama, Tetsuo, Photosynthetic Electron
Transport. Department of Biology Seminar,
Toho University, Chiba, Japan, December 26,
1972.
Hiyama, Tetsuo, Green Plant Photosynthesis.
Institute of Applied Microbiology Seminar,
University of Tokyo, Tokyo, Japan, January
8, 1973.
Hiyama, Tetsuo, Techniques and Methods for
the Study of Photosynthetic Electron Trans-
port. Photobiology Seminar, Department of
Biological Sciences, Stanford University,
Stanford, California, June 1, 1973.
Ninnemann, Helga, Photoinhibition of Mito-
chondrial Respiration. Stanford-Carnegie
Plant Physiology Graduate Seminar, Depart-
ment of Plant Biology, Carnegie Institution,
Stanford, California, January 17, 1973.
Ninnemann, Helga, Light Effects on Respira-
tory Cytochromes. University of California
at San Diego, La Jolla, California, April 3,
1973.
Takamiya, Atusi, Occurrence, Distribution and
Properties of a Water-Soluble, Photoconver-
tible Chlorophyll Protein Complex; CP668-
743 in Leaves of Plants Related to Chenopo-
dium album. Department of Plant Sciences
and Microbiology Class Seminar, Indiana
University, Bloomington, Indiana, August 2,
1972.
Takamiya, Atusi, Autotrophic Form of Strepto-
myces Capable of Growing by Utilizing
Energy from the Knallgas Reaction. Depart-
ment of Plant Sciences and Microbiology
Class Seminar, Indiana University, Blooming-
ton, Indiana, xA.ugust 2, 1972.
Takamiya, Atusi, Water-Soluble Chlorophyll-
Protein Complex. Stanford-Carnegie Plant
Physiology Graduate Seminar, Department of
Plant Biology, Carnegie Institution, Stanford,
California, October 18, 1972.
Takamiya, Atusi, Researches on Water-Soluble
Chlorophyll-Protein Complexes. Department
of Biology, Revelle College. University of
California at San Diego, La Jolla, California,
October 29, 1972.
Takamiya, Atusi, On the Distribution and
Properties of a Water-Soluble Chlorophyll
Protein CP668. Japanese Society of Plant
Physiologists, Tokyo, Japan, March 31, 1973.
Virgin, Hemming I., see French, C. Stacy.
414
CARNEGIE INSTITUTION
PERSONNEL
Photosynthesis Group
Staff: C. Stacy French,1 Director; Jeanette
S. Brown, David C. Fork
Visiting Investigators: Georgi Detchev,2
Ralphreed A. Gasanov,3 Ulrich Schreiber,4
Atusi Takamiya 5
Carnegie Institution Research Fellows: Wil-
liam G. Hagar III, Tetsuo Hiyama, Norio
Murata,6 Helga I. Ninnemann 7
Technical Assistant : Glenn A. Ford
Physiological Ecology Group
Staff: Joseph A. Berry, Olle Bjorkman, Mal-
colm A. Nobs, William M. Hiesey, Emeritus
Carnegie Institution Research Fellows:
George Bowes,8 John H. Troughton 9
1 Retired June 30, 1973.
2 From November 21, 1972, through Decem-
ber 12, 1972. From Bulgarian Academy of
Sciences, Sofia, Bulgaria.
3 From September 25, 1972, through April 20,
1973. From Institute of Botany, Academy of
Sciences of Azerbaijan SSR, Baku, USSR.
4 From June 1, 1972, through August 30, 1972.
From Simon Fraser University, Burnaby 2, B.C.,
Canada.
5 From June 5, 1972, through October 31, 1972.
From Toho University, Chiba, Japan.
6 From June 20, 1972, through September 8,
1972. From University of Tokyo, Japan.
7 From November 16, 1972, through April 19,
1973. From University of Tubingen, Germany.
8 To December 31, 1972.
9 From April 18, 1973. From Department of
Scientific and Industrial Research, New Zea-
land.
Technical Assistants: Lis D. Berry,10 Ed-
ward G. Gausden,11 Frank Nicholson,
William E. Ward, Jr.12
Gardener: Archibald H. Lawrence13
Summer Research Assistant: John L. Horn 14
Research Associate: William G. McGinnies,
at Tucson, Arizona
Administrative and Operating Staff
Administrative Secretary -Accountant : Ruth
Fischer
Secretaries: Donna C. Moore,15 Christine
Shank i«
Clerical Assistant: Karen Franklin17
Mechanical Engineer: Richard W. Hart
Electrical Engineer: Kai Lanz
Computer Programmer : Thomas H. Mathie-
son
Custodian: Jan Kowalik
Grounds Maintenance Assistants: Ronald W.
Atteberry,18 Keith R. Weed 19
10 From February 21, 1973.
11 From April 30, 1973.
12 From June 15, 1973.
13 Retired June 30, 1973.
14 To September 16, 1972.
15 From July 17, 1972.
16 To July 31, 1972.
17 From April 30, 1973.
18 To September 16, 1972.
19 From February 26, 1973, through June 12,
1973.
Geophysical Laboratory
Washington, District of Columbia
Hatten S. Yoder, Jr.
Director
Contents
Director's Commentary and Review . . 418
Experimental and Field Studies in Igne-
ous Petrology 431
Structure of the upper mantle beneath
Lesotho (Boyd and Nixon) . . 431
Ultramafic nodules from Colorado-
Wyoming kimberlite pipes (Eggler
and McCallum) 446
Akermanite-C02 : relationship of meli-
lite-bearing rocks to kimberlite
(Yoder) 449
Role of C02 in melting processes in the
mantle (Eggler) 457
Melting in a hydrous mantle: phase
relations of mantle peridotite with
controlled water and oxygen fugac-
ities (Mysen) 467
Bridget Cove Volcanics, Juneau area,
Alaska: possible parental magma
of Alaskan-type ultramafic com-
plexes (Irvine) 478
Principles of melting of hydrous phases
in silicate melt (Eggler) . . . 491
Solubility of H20 in forsterite melt at
20 kbar (Hodges) 495
Regularities in the shift of liquidus
boundaries in silicate systems and
their significance in magma genesis
(Kushiro) 497
The system diopside-anorthite-albite :
determination of compositions of
coexisting phases (Kushiro) . . 502
Metamorphic Petrology 507
Infiltration metasomatism in the sys-
tem K20 - A1203 - Si02 - H20 - HC1
(Frantz and Weisbrod) .... 507
Cordierite-garnet equilibrium in the
system Fe-Mn-Al-Si-O-H (Weis-
brod) 515
Refinements of the equilibrium condi-
tions of the reaction Fe cordierite
<=^ almandine + quartz + silli-
manite ( + H20) (Weisbrod) . . 518
The problem of water in cordierite
(Weisbrod) 521
Effect of pressure on the composition
of coexisting pyroxenes and garnet
in the system CaSi03-MgSi03-
FeSi03-CaAlTi206 (Akella and
Boyd) 523
Pyroxenes and garnets as geothermom-
eters and barometers (Hensen) . 527
Physical and Experimental Mineralogy . 535
Pyroxenes from Mull andesites (Virgo
and Ross) 535
The "Fe Mossbauer spectra of syn-
thetic pyroxenes across part of the
join Fe85Eni5-wollastonite (Virgo) 540
A possible high-low temperature tran-
sition in orthopyroxenes and or-
thoamphiboles (Ohashi and Fin-
ger) 544
Thermal vibration ellipsoids and equi-
potential surfaces at the cation
sites in olivine and clinopyroxenes
(Ohashi and Finger) 547
Electrical and optical properties of the
olivine series at high pressure
(Mao) 552
Observations of optical absorption and
electrical conductivity in magnesio-
wiistite at high pressures (Mao) . 554
Thermal and electrical properties of
the earth's mantle (Mao) . . . 557
Polarized absorption spectra of gil-
lespite at high pressure (Abu-Eid,
Mao, and Burns) 564
The Fe3+ site preference in the solid
solution series MgCr204-MgFe204
(Virgo and Ulmer) 567
Lattice deformations in feldspars
(Ohashi and Finger) 569
Measurements of the polarized crystal-
field spectra of ferrous and ferric
iron in seven terrestrial plagio-
clases (Bell and Mao) .... 574
Electric-field-gradient calculations in
rare-earth iron garnets (Raymond) 576
Solid solution in aenigmatite (Hodges
and Barker) 578
Isotope Geochemistry 581
Uranium and thorium partitioning in
diopside-melt and whitlockite-melt
systems (Seitz) 581
Uranium and thorium diffusion in
diopside and fluorapatite (Seitz) . 586
Boron mapping and partitioning in
synthetic and natural systems:
crystal-melt assemblages, garnet
lherzolite, chondrites (Seitz) . . 588
The oxygen isotope geochemistry of
siliceous volcanic rocks from Ice-
land (Muehlenbachs) .... 593
The oxygen isotope composition of the
1973 Heimaey lava (Muehlenbachs
and Jakobsson) 597
Oxygen isotope compositions of some
basaltic lavas from the Snake
River Plain (Muehlenbachs and
Stone) 598
Radiometric Age Determinations . . . 601
The effect of regional metamorphism
on U-Pb systems in zircon and a
comparison with Rb-Sr systems in
the same whole rock and its con-
stituent minerals (Krogh and
Davis) 601
The significance of inherited zircons on
the age and origin of igneous rocks
— an investigation of the ages of
the Labrador adamellites (Krogh
and Davis) 610
Anomalous isotopic composition of
lead in young zircons (Mattinson) 613
Age and origin of ophiolitic rocks on
La Desirade Island, Lesser Antilles
Island Arc (Mattinson, Fink, and
Hopson) 615
Lunar Petrology 623
Partitioning of Ti and Al and its bear-
ing on the origin of mare basalts
(Akella and Boyd) 623
A study of charge-transfer and crystal-
field spectra of iron and titanium
in synthetic "basalt" glass as a
function of Po2 (Mao and Bell) . 629
Analytical study of the orange lunar
soil returned by the Apollo 17
astronauts (Mao, Virgo, and Bell) 631
Rust alteration of the Apollo 16 rocks
(Taylor, Mao, and Bell) ... 638
An analytical study of iron in plagio-
clase from Apollo 16 soils 64501,
64502, and 64802; Apollo 16 rock
66095; and Apollo 15 rock 15475
(Bell and Mao) 643
Petrology of Apollo 16 lunar highland
rocks (Hodges and Kushiro) . . 645
Liquidus phase relations of Apollo 15
mare basalt 15016 (Hodges and
Kushiro) 646
Crystallization of pyroxenes in Apollo
15 mare basalts (Kushiro) . . . 647
Crystallization and subsolidus cooling
history of Apollo 15 basalts 15076
and 15476 (Virgo) 650
Optical absorption studies of the Rus-
sian Luna 20 soil (Bell and Mao) . 656
Luna 20 plagioclase : crystal-field effects
and chemical analysis of iron
(Mao and Bell) 662
Volatilization of iron-bearing silicates
in the presence of carbon (Seitz) . 666
Systematic Petrography 668
Development and modification of the
information system RKNFSYS
(Chayes) 668
A relation between silica saturation and
oxidation state in class II, (ol +
hy) normative basalts (Chayes) . 671
A further note on the distribution of
silica-saturation components in
Cenozoic basalt (Chayes) . . . 673
Multivariate frequency information on
rock chemical data (Irvine) . . 675
Determining null values of ratio corre-
lation by simulation (Chayes) . 681
Biogeochemistry 682
A comparison of melanoidin and humic
acid (Hoering) 682
Separation of amino acid optical iso-
mers by gas chromatography
(Hare and Hoering) 690
New Techniques, Equipment, and Cali-
brations 694
A computer-automated, single-crystal
x-ray diffractometer (Finger, Hadi-
diacos, and Ohashi) 694
Modal analysis on the automated elec-
tron microprobe (Chayes) . . . 699
Fluorescent method for the analysis of
amino acids and peptides (Hare) . 701
The analysis of silica at submicrogram
levels using a flow-cell colorimetric
technique (Frantz and Hare) . . 704
Albite-jadeite-quartz equilibrium: a
hydrostatic determination (Hays
and Bell) 706
Incongruent melting of pure diopside
(Kushiro) 708
Staff Activities 710
Conference on Geochemical Transport
and Kinetics 710
Washington Crystal Colloquium . . 712
Seminar series 712
Field studies 713
Exhibits 714
Lectures 714
Penologists' Club 716
Bibliography 716
References Cited 717
Personnel 733
Carnegie Institution of Washington Year Book 72, 1972-1973
DIRECTOR'S COMMENTARY AND REVIEW
"To the Improvement of Mankind"
The present "bear market" in support
of science has not dampened the enthu-
siasm for basic research at the Geophys-
ical Laboratory. In fact, the public de-
mand for relevancy has stimulated the
staff, for they recognize that their efforts
will ultimately have a role in meeting
future societal problems. With great ma-
terials and energy shortages looming, the
need for research laboratories in earth
sciences is greater than ever.
The small group of scientists assem-
bled at the Geophysical Laboratory is
representative of the most talented and
creative in the scientific community.
They are supported, within financial lim-
itations, to do the research they think
best, because efforts of their own choice
are more likely to be useful to society
than those chosen by others remote from
the forefront of science. They are moti-
vated by curiosity, the desire to know
and understand without regard to imme-
diate practical applications. They are
fully aware, however, that this freedom
carries with it the responsibility to pro-
vide not only scientific leadership in de-
fining the scientific problems and finding
possible solutions but also insight into
societal needs.
What are the criteria, then, to deter-
mine that these exceptional people are
creating knowledge of benefit to man-
kind? The record of achievements result-
ing in practical applications of re-
searches at the Geophysical Laboratory
is reassuring. Industries (e.g., glass, ce-
ramics, steel making, electronics) have
received major contributions from the
work described in these reports. Greater
understanding of the geologic hazards
(such as volcanic eruptions and earth-
quakes) facing man has been acquired
from the studies presented here. Man's
continuing search for ore deposits and
sources of energy has been aided by the
researches of this Laboratory.
But who can forecast the needs of
tomorrow? Who can tell us what prac-
tical benefits may come from studies of
the origin of life, the condensation of the
solar nebula, the generation of magma,
or the formation of the lunar surface? Is
it not the wiser course to provide a broad
base of knowledge, as inefficient as it
may seem at the time, for the unpredict-
able array of new problems rather than
be committed to a few fixed goals of un-
certain long-range value? This long de-
bated question was clearly answered
many years ago for a young Staff Mem-
ber at the Laboratory by Dr. Vannevar
Bush: "If what you are doing today can
be used by industry now, you are not far
enough ahead in either theory or experi-
ment!" Little did they realize that the
very problem the young scientist was
studying would not only be of value in
mineralogy but would also lead to a new
transistor, a gem stone, and a laser mate-
rial.
The seeds of future benefits have been
planted in the research reported in the
pages that follow. It was an unusually
productive year.
A Brief Review of Accomplishments
Experimental and Field Studies in
Igneous Petrology
The exhilaration of developing a major
new idea was experienced by Boyd and
Nixon as their reconstruction of a fossil
geothermal gradient unfolded. Ultra-
mafic nodules from a number of kimber-
lites in northern Lesotho provided tem-
perature-depth points on an ancient geo-
thermal gradient extending over a depth
range of 100-200 km. Experimental data
on synthetic systems were used to esti-
mate equilibration temperatures and
418
GEOPHYSICAL LABORATORY 419
pressures between the coexisting and dis- wt % C02 could be dissolved in a liquid
aggregated phases in the nodules. The of diopside composition at 20 kbar. Much
gradient appears to increase at depths less C02 dissolved in a melt of albite
below 150 km, and the peridotite nodules composition, yet H20 enhanced the solu-
of deepest origin were found to be in- tion of C02 in all melts studied. The
tensely sheared. They speculate that dominant effect of C02 in partial melting
these rocks were sheared during the appears to be the shift of liquid corn-
breakup of Gondwanaland about 200 position from silica-saturated to silica-
m.y. ago and that stress-heating has in- undersaturated. His results provide a
creased the geothermal gradient. Mega- demonstration of the close relationship of
crysts from these kimberlites are inter- C02-rich fluids to silica-undersaturated
preted as having crystallized in crystal- melts.
mush magmas in the upper part of the Volatiles play a key role in Mysen's
low-velocity zone (150-200 km). The new hypothesis for the formation of kim-
work of Boyd and Nixon yields the first berlite and associated igneous rocks from
direct quantitative estimate of an an- peridotite. He melted four natural peri-
cient geothermal gradient. Additional dotite nodules under conditions of water
studies on nodules from a newly dis- fugacity ranging from anhydrous to
covered group of kimberlite pipes in the water-undersaturated and simultane-
northern Front Range of Colorado and ously controlled the oxygen fugacity with
Wyoming were carried out by Eggler and a new technique. The peridotite solidus
McCallum. They support the view that was found to increase isobarically with
the variety of nodules reflects the wide increasing Mg/(Mg + Fe2+) of the rock,
spectrum of origin as accidental xeno- increasing Ca/Al, and decreasing water
liths, liquidus phases in a melt, accumu- fugacity. Accordingly, he concludes that
lates, quenched liquid, or depleted the depth in the oceanic mantle corre-
residua. sponding to the first appearance of garnet
Yoder turned his attention to the ma- is sensitive to the Mg/(Mg + Fe2+) of
trix in which the nodules are found, the the rock and that the continental mantle
kimberlite itself. He demonstrated that is probably garnet-bearing immediately
kimberlite magma could not be derived beneath the Mohorovicic discontinuity,
from melilite-bearing rocks, as proposed The melts formed by partial melting of
by some investigators, but may be ge- peridotite under high water fugacity are
netically related to a magma with suit- andesitic, whereas the melts become
able volatiles having potential melilite. silica-undersaturated at water fugacities
His argument is based on the instability corresponding to a mole fraction of water
of the main component of melilite, between 0.5 and 0.6 in the coexisting
akermanite, at high pressures of H20 or vapor. Such melts resemble the volcanic
C02 and the conversion of melilite-bear- rocks associated with kimberlite.
ing rocks in the absence of H20 and C02 Many of the above experiments relat-
to pyroxenites at the high pressures ing to the genesis of kimberlite and the
where kimberlites are believed to origi- nodules it contains were stimulated by
nate. A particularly significant result of the forthcoming International Confer-
Yoder's study was the formation of ence on Kimberlite to be held in the
homogeneous silicate liquids containing a Republic of South Africa and sponsored
large amount of carbonate. in part by the Geophysical Laboratory.
The essential role of C02 in the forma- Irvine has been examining another prob-
tion of kimberlites, carbonatites, and lem concerning ultramafic rocks: the
highly alkalic rock suites was examined source of the magma for the concentric-
by Eggler. He found that as much as 9.5 ally zoned ultramafic complexes of
420
CARNEGIE INSTITUTION
southeastern Alaska. Irvine found mate-
rial that might be representative of the
magma in a volcanic belt associated with,
and roughly equivalent in age to, these
complexes. The volcanic rocks resemble
those from the ultramafic bodies in being
extremely rich in augite and appear to
show enough correlation in detailed min-
eralogical and chemical features to place
some definite compositional limits on the
parental magma. They are unusual alka-
line rocks with high CaO and low A1203
contents and are particularly distinctive
in being low in Ti02 and P2O5 and high
in K20. Irvine rules out the formation of
the ultramafic rocks either from ultra-
mafic magmas or from a tholeiitic pre-
cursor of calc-alkaline andesite and sug-
gests that a more likely parent is a water-
rich picritic ankaramite.
Eggler has worked out in detail the
principles of melting where hydrous
phases are involved. In addition, Hodges
measured the solubility of water in a
melt of forsterite composition and relates
the solubility of water to the number of
oxygen atoms in the melt. A dramatic
shift in the phase boundaries produced
by water is not unique to H20, and Ku-
shiro found that other oxides produce
similar effects, depending on the cationic
valency. Liquidus boundaries shift to-
ward silica-rich compositions by solution
of monovalent elements, whereas the
boundaries shift toward silica-poor com-
positions by solution of oxides of poly-
valent elements such as Cr203, Ti02 and
P205. For example, the addition of as
little as 3 wt % P205 results in a great
reduction of the forsterite field relative
to the enstatite field. Aluminum and
divalent elements have small effects. Ku-
shiro explains these observations on the
basis of the structural effects in the melt
imposed by the cations. The degree of
shift of the boundaries varies inversely
with the extent of polymerization of the
(Si04)-4 tetrahedra, the monovalent ions
preventing polymerization whereas the
polyvalent ions polymerize (Si04)"4
tetrahedra. The regularities in the shift
of liquidus boundaries found in Kushiro's
study can be applied to many problems
in igneous pedogenesis. In general, sig-
nificant amounts of polyvalent ions are
present in silica-undersaturated rocks,
whereas large amounts of monovalent
ions occur in silica-rich late differentiates
of magma. He emphasizes the significant
role of P205 in the generation of silica-
undersaturated magmas.
Almost every student of igneous pe-
trology has learned to appreciate the
concept of magma fractionation by
studying the classic work of Bowen on
the system diopside-albite-anorthite. The
first major revision of that system has
been made by Kushiro, using the electron
microprobe. He showed that the liquidus
boundary between diopside and plagio-
clase is concave toward diopside through-
out its entire range and directly deter-
mined the composition of the coexisting
phases at a series of temperatures. The
pantelleritic trend of liquids in this
pseudoternary system was confirmed.
Metamorphic Petrology
The evolution of metasomatic rocks in-
volves changes in their chemical com-
position. The requisite transport of ma-
terial through rocks, particularly over
long distances, to accomplish these
changes could not be accounted for by
diffusion through a static solvent, and a
theory dependent on the flow of a solvent
in response to gradients in fluid pressure
was devised by Korzhinskii (1936). No
quantitative application of this theory
had been made, and Frantz and Weis-
brod attempted to fulfill that need. They
calculated the evolution of an Si-Al bear-
ing rock infiltrated by a potassium-rich
solution and discovered that the applica-
bility of Korzhinskii's theory is restricted
if the reactions involved yield a signifi-
cant increase in volume. Frantz and
Weisbrod generalized the theory to re-
move the constraint of Lindgren's "law"
of constant volume replacement.
GEOPHYSICAL LABORATORY
421
Two critical minerals in the metamor-
phism of pelitic schists are cordierite
and garnet. The presence of manganese
in these minerals may greatly influence
their stability. For this reason Weisbrod
studied experimentally the system Mg-
Fe-Mn-Al-S-O-H to evaluate the influ-
ence of Mn. He specifically chose the
cordierite-garnet-sillimanite-quartz equi-
librium at 750 °C for quantitative meas-
urement. A small amount of Mn did in-
deed drastically reduce the field of cor-
dierite, even at very low pressures. Using
his experimental results he calculated
that the minimum pressure of metamor-
phism for a pelitic rock bearing 2% gar-
net and no cordierite is decreased about
30% if the normal MnO content of 0.2
wt % is taken into account. In the course
of the work he redetermined the high-
pressure stability limit of Fe cordierite
at low-oxygen fugacity and noted that
changes in oxidation significantly af-
fected the stability field of this mineral.
Weisbrod also tried a new approach for
calculating the water content of cordier-
ite, using the experimentally determined
partition coefficients of two exchangeable
metals (Fe and Mg or Fe and Mn) be-
tween coexisting cordierite and garnet.
This method enabled him to calculate
the water content of Fe cordierite at
750 °C, between 0 and 3 kbar. His results
are similar to those previously measured
for Mg cordierite under the same condi-
tions and indicate that, at least at high
temperatures and low pressures, the
water content of cordierite is a function
of pressure (and temperature) but prob-
ably not of the composition of the min-
eral. The amount of water in cordierite
(0.6 mole of H20 per mole of cordierite)
is significant and cannot be neglected.
In the highest grades of metamor-
phism, garnet coexists with pyroxene in
rocks such as eclogite. Akella and Boyd
determined the effect of pressure on the
composition of garnets and pyroxenes in
the system (Ca,Mg,Fe)Si03-CaAl2Ti06
at 1000° and 1100°C. In the garnet sta-
bility field, the solubility of A1208 in
synthetic orthopyroxene decreases as the
Mg/(Mg + Fe2+) of the rock decreases.
They point out that a direct application
of data on the enstatite-pyrope join to
natural systems will result in pressure
estimates that are too high. Using na-
tural pyroxene pairs, Hensen also ob-
tained new data on the distribution of
Fe2+ and Mg between pyroxenes and gar-
nets in the pyroxene two-phase region
and on the solubility of A1203 in pyrox-
enes as a function of temperature and
pressure. He concludes that the partition
coefficient of Mg/Fe2+ between garnet
and clinopyroxene is a potentially useful
geothermometer.
Physical and Experimental Mineralogy
The pyroxenes are the subject of in-
tensive study at the Laboratory. The
focus has been on the iron-rich pyroxenes
in the lunar samples ; however, the excep-
tional size and complexity of zoned au-
gite and pigeonite phenocrysts from the
lunar rocks has provided an impetus to
reexamine the interrelations of the vari-
ous pyroxene structures in the entire
CaSKVMgSiOa-FeSiOs system. Virgo
and Ross started with the reference
pigeonite from the Mull andesite because
of its unique properties — those proper-
ties that are found in the lunar samples!
Other specimens of pigeonite from Mull
revealed a two-phase assemblage of pi-
geonite (Wo4Fs54En42) and an ortho-
pyroxene (\y02Fs53En45) . This assem-
blage probably occurs relatively early in
the crystallization of magmas and has
been overlooked, no doubt, because of the
small difference in composition and the
lack of detailed structural characteriza-
tion of pyroxene phases. The observa-
tion, documented with two other natural
examples, is interpreted in terms of the
orthopyroxene — > pigeonite reaction first
defined by Bowen and Schairer (1935)
and in the light of the two-phase field
close in composition to the enstatite-
ferrosilite join recently described by
Ross, Huebner, and Hickling (1973).
422 CARNEGIE INSTITUTION
Supplementing the studies on natural in order to describe the vibrational be-
pyroxene assemblages, Virgo has charac- havior of atoms. The method, based on
terized single-phase synthetic pyroxenes the electrostatic-repulsive potential
across part of the join Fs85Eni5-Wo model, was applied to the study of the
using the 57Fe Mossbauer technique, thermal vibrations of the cation sites in
The C2/c pyroxenes containing 10, 15, olivine and clinopyroxene.
25, and 35 mole % wollastonite give In addition to observing the effect of
anomalous resonant absorption spectra high temperatures at temperature, it is
that suggest a significant excess of ca- necessary to make measurements of
tions in the M2 crystal site. These re- crystal properties at high pressure. Be-
sults are in contrast to those on the end- cause olivine is believed to be the major
member compositions, orthopyroxene and phase in the mantle, Mao measured the
a clinopyroxene ocntaining 50 mole % electrical and optical properties of the
wollastonite, which can be interpreted in olivine series. He found that, as the pres-
terms of their inferred stoichiometric sure is raised above 100 kbar, the ap-
compositions. The anomalous spectra of parent absorption edge of the olivine
clinopyroxene are consistent with a shifts to lower energy simultaneously
model in which there are multiple Ml with an exponential increase in electrical
sites in the C2/c clinopyroxene structure conductivity. The intensity of these re-
brought about by positional disorder of versible effects depends on the fayalite
the M2 site. content and is not accompanied by a
These anomalies may not exist at the first-order phase change. These strong
temperatures and pressures of formation pressure effects are used by Mao to inter-
of the pyroxene, and a new array of pret and evaluate the mechanisms of
structural features may govern its be- heat transfer in the mantle,
havior. Ohashi and Finger, therefore, in- In a similar way, Mao found that the
vestigated at various high temperatures optical absorption of magnesiowustite,
a possible phase transition in orthopy- another phase believed to be important in
roxene and orthoamphibole. It was pre- the lower mantle, increases rapidly with
sumed that a high-temperature ortho- pressure. In fact, the optical absorption
pyroxene polymorph could be related to of magnesiowustite is high enough to
the high-temperature clinopyroxene poly- block all thermal radiation ! However,
morph found earlier. The orthopyroxene simultaneous measurements of electrical
polymorph proposed presents a new fea- conductivity indicate that at first it in-
ture in which a regional symmetry ele- creases with pressure by a factor of 10
ment imposes nonspace group extinctions per 50 kbar, and then the rate reduces to
on the x-ray diffraction pattern. Ohashi a factor of 1.2 per 50 kbar when the ma-
and Finger conclude that one of the im- terial becomes opaque. Thermal conduc-
portant factors governing the high— low tivity is related to electrical conductiv-
temperature transition is the silicate ity, and therefore, Mao's results suggest
chain rotations, which are different in the that effective mechanisms for both con-
clinopyroxene and orthopyroxene poly- ductive processes are operative at high
morphs. pressures.
As the temperature of the crystal is On the basis of these new optical and
raised, the thermal vibrations of atoms electrical data on important minerals in
become more significant and play an the mantle, Mao has examined the cur-
important role not only in transitions but rent thermal and electrical models of the
in melting as well. Ohashi and Finger mantle. For example, thermal conduc-
have developed a means of estimating an tivity by phonons is important in the
anisotropic thermal vibration ellipsoid lower mantle and is probably one to two
GEOPHYSICAL LABORATORY
423
orders of magnitude greater than previ-
ously thought. The radiative heat trans-
fer is severely blocked in the upper man-
tle and the transition zone but may be
effective in the lower mantle by an ab-
sorption-reemission process. Exciton
transfer of heat is uncertain, but high
pressure could enhance localized exci-
tons. It is evident that a drastically dif-
ferent model of heat transfer from the
mantle is evolving.
The conclusion that high pressure also
tends to change the Fe2+ in silicates from
a high-spin state into a low-spin state is
based primarily on the behavior of the
rare mineral gillespite, BaFeSi4Oi0. Ac-
companying this change is a significant
volume reduction; hence Fe2+ in mantle
silicates is believed to be in a low-spin
state. On reexamination of gillespite it
was discovered by Abu-Eid, Mao, and
Burns using polarized optical spectra
that the observed behavior is probably
due to a second-order transition or a con-
version of only some of the iron to a low-
spin state. Unfortunately it is clearly
not a simple high-spin to low-spin con-
version; therefore, all previous calcula-
tions involving such transitions in the
mantle will have to be modified.
Highlights of other mineralogical stud-
ies include the discovery by Virgo and
Ulmer of an inversion of the normal
spinel structure in the composition range
of 60-70 wt % MgFe204 in the MgCr204-
MgFe204 system using the iron-57 Moss-
bauer technique. Ohashi and Finger de-
veloped a new technique of strain tensor
analysis, especially useful for triclinic
crystals where conventional description
of fractional changes in unit-cell param-
eters may be misleading. They used the
new technique in correlating the defor-
mation in feldspars with crystal struc-
ture. Bell and Mao obtained the first
polarized optical spectra on feldspars and
found ferric iron in plagioclase having as
little as 0.01 wt % Fe. The sensitivity of
the method in determining Fe3+/Fe2+ was
demonstrated, and the data will be help-
ful in the interpretation of diffuse reflec-
tance spectra of celestial bodies obtained
by telescope.
The bonding properties in crystals,
such as cohesive energies and site-prefer-
ence energies, govern the overall response
of a crystal to changes in pressure-tem-
perature conditions. Raymond has been
working with a model of the electrostatic
potential distributions in crystals in an
attempt to understand bonding properties.
Most recently he has included the effects
of overlapping deformable ions in the
calculation of electric-field gradients in
rare-earth iron garnet. The effect was
confirmed by a comparison with experi-
mental data derived from Mossbauer
studies.
Hodges and Barker describe a new
solid solution in aenigmatite, a mineral
common in peralkaline rocks. In addi-
tion to finding in nature a major solid
solution in the series Na2Fe2+5TiSi602o-
Na2Fe2+4Fe3+2Si602o, they also report
solid solution of the type Fe3+Al3+ ->
Fe2+Si4+. These solid solutions are more
strongly dependent on the abundance of
Ti in the melt than on oxygen fugacity.
Strangely enough, the Fe2+-rich end
member is stable at higher oxygen fugac-
ities than the Fe3+-rich end member.
Isotope Geochemistry
The very sensitive new analytical tech-
nique, particle-track mapping, was em-
ployed by Seitz to determine the parti-
tioning of uranium and thorium between
minerals and coexisting melt. Their par-
titioning between coexisting diopside and
melt is not strongly dependent on melt
composition or temperature. Silicate so-
lutions were found to be dilute with re-
spect to uranium at concentrations from
0.1 to 4 ppm U similar to those found in
natural systems. Most important, Seitz
found that the partitioning of these ele-
ments is unaffected by pressure up to 25
kbar. These properties considerably sim-
plify the application of the experiment-
424 CARNEGIE INSTITUTION
ally determined partition coefficients to position of lavas from the new eruption
studies of geologic processes. on Heimaey is similar to those of basalts
Using the same technique, Seitz was found elsewhere and, significantly, is
able to measure the diffusion rates of identical with that of basalt from the
uranium and thorium in some silicate and only other region of alkali olivine basalt
phosphate minerals. The experimental volcanism on Iceland. These regions
procedure approached natural conditions have a thicker crust than do areas of low-
by employing coexisting mineral and 180 tholeiitic basalt. If the low-180 con-
melt phases low in uranium and thorium, tent of the basalts is due to oxygen ex-
Uranium diffusion in diopside and tho- change between meteoric water and
rium diffusion in fluorapatite were found magma that occurs only in a thin crust,
to be slower than the diffusion of alu- then the mugearite of Heimaey might
minum and magnesium, respectively, in also be depleted in 180 because it prob-
these phases. ably differentiated in a shallow magma
A technique for mapping boron on a chamber. According to Muehlenbachs
microscopic scale is also described by and Jakobsson no such interaction was
Seitz. He used the method for determin- found in the rocks from Heimaey.
ing boron partitioning between olivine, Oxygen-isotope studies of Muehlen-
orthopyroxene, clinopyroxene, spinel, and bachs and Stone indicate that olivine
coexisting melt. Boron concentrations in tholeiites from the Snake River Plain,
the minerals of garnet lherzolite indicate Idaho, are not contaminated by crustal
that the rock was in equilibrium with a material. On the other hand, some asso-
melt containing 8 ppm boron — a value ciated, extremely iron-enriched lavas
typical of oceanic basalts. In meteorites might be. Furthermore, at least one
of chrondritic composition boron is con- olivine tholeiite is lower in 180 content
centrated in fine-grain interstitial than other unaltered continental basalts
areas, possibly in the layer-lattice sili- and in this way resembles basalts of Ice-
cate prevalent in carbonaceous chon- land. Muehlenbachs and Stone point out
drites. The high concentration of boron the tectonic, geochemical, and hydrologic
found in a melilite-fassaite chondrule of similarities of these two provinces,
the Allende meteorite is particularly sur-
prising because such chondrules are be- Radiometric Age Determinations
lieved to be high-temperature conden- m, ,, ,. . . .
- I he age ol a rock can now be de-
Tvyr i i u u u j. j- i iu termined with great analytical precision
Muehlenbachs has studied the oxygen , ,. f „ J , r , ,,
, , . , „ ... , irom exceedingly small amounts of the
isotope geochemistry of siliceous rocks ... , • , mu
- T i i t i <• , n , „ constituent minerals. The major ques-
from Iceland and has found that they tion ^ what eyent ig being measured?
like the basalts of Iceland, are depleted The datmg systemg are often disturbed
in 180. Although the data from one lo- by events postdating the origin of the
cahty are consistent with the hypothesis rock and in some cases by events taking
that the siliceous rocks are derived from piace even before the mineral was in-
basic magma by crystal fractionation, corporated into the present rock in which
Muehlenbachs argues that the anomalous it is found. The apparent chaos of ages
isotopic ratios are best explained by is in fact a record accurately dating sig-
models in which silicic magmas either nificant steps in complex processes,
exchange oxygen with hydrothermally Krogh and Davis have carefully dis-
altered basalts or are derived by partial sected the events that occurred in a lay-
melting of hydrothermally altered basalt, ered paragneiss that has undergone meta-
Not all volcanic rocks of Iceland are morphism to the amphibolite facies in the
low in 180 content. The isotopic com- Grenville province. A direct U-Pb age
GEOPHYSICAL LABORATORY
425
determination of the overgrowth material
from 200 /mi grains demonstrated that
each zircon from the paragneiss contains
material grown at two greatly different
times. The addition of uranium, with or
without visible overgrowths, is shown to
be a major factor that masks the primary
age of zircons from metamorphosed
rocks. The age of the overgrowths
(^1000 m.y.) is the same as the Rb-Sr
whole-rock age of an obvious reaction
zone between layers of contrasting com-
position in the paragneiss. Other layers
in the same outcrop remained closed to
the migration of Rb and Sr during the
reaction event and give an isochron age
of 1845 m.y. They show that isotopic
equilibrium for Sr had not been reached
on a scale of a few centimeters during the
regional metamorphism of felsic para-
gneiss to amphibolite facies! It is evi-
dent that great care must be taken in
sampling an outcrop if meaningful ages
of geologic events are to be measured.
Although inherited zircons may yield
spurious ages of igneous rocks, they can
be used in detecting crustal contamina-
tion of a magma. Inherited zircons were
recognized by Krogh in an adamellite
from Nain, Labrador, and analyzed sep-
arately from clear zircons also found in
the rock. The inherited zircons gave an
apparent 207Pb-206Pb age about 40 m.y.
older than the time of crystallization of
the clear zircons. Duplicate analyses of
zircons from duplicate samples of several
adamellite intrusions indicated a repro-
ducibility of about 1 m.y. for lead-
uranium and lead-lead ages.
Mattinson studied the relationships of
U and Pb isotopes in zircons from rocks
as young as 14 m.y. He notes systematic
deviations in the 207Pb/206Pb ratios that
are higher than would be expected for
ideal behavior. Younger zircons show
greater deviations from ideality than
older zircons. Mattinson concludes that
incorporation of intermediate daughter
products in other than secular equilib-
rium amounts during crystallization and
loss of intermediate daughter products
immediately after crystallization are the
main factors in the nonideal behavior.
Mattinson, Fink, and Hopson studied
a suite of ophiolitic rocks on La Desirade
Island, Lesser Antilles. On the basis of
isotopic and age relationships and seismic
reflection profiles, they conclude that the
ophiolitic rocks represent either western
Atlantic or eastern Caribbean oceanic
crust that developed during the Late
Jurassic period and served as the base-
ment on which the Eocene and younger
calc-alkaline rocks of the Lesser Antilles
Island Arc developed.
Lunar Petrology
The termination of the Apollo program
brings to an end a period of courageous
physical exploration of the moon. The
international group of scientists who
have been studying the returned samples
have been remarkably effective, and they
should continue to be supported. The
astronauts collected an invaluable suite
of samples. To realize the full value of
these treasures a vigorous program of
study of the returned lunar samples must
be maintained.
The exceptionally high titanium con-
tent of some mare basalts, especially
those collected during Apollo 17, has
turned the attention of Akella and Boyd
to pertinent systems that contain coexist-
ing Ti-rich oxides and silicates. They
equilibrated ilmenite and silicates over a
range of temperatures and pressures and
analyzed the products with the electron
microprobe. In subsolidus runs clino-
pyroxenes in equilibrium with ilmenite
contain 1-5 wt % Ti02, the most Ti-rich
pyroxenes crystallizing in the intermedi-
ate pressure range (5-20 kbar). In the
melting interval, clinopyroxenes contain
up to 6.5 wt % Ti02 and coexist with
Ti-rich liquids. An increase in silica
activity of the liquid causes a large de-
crease in the Ti and Al contents of the
clinopyroxenes. Moderate variations in
Mg/(Mg +Fe) 2+ do not affect the Ti and
Al of these clinopyroxenes. Ilmenite
426 CARNEGIE INSTITUTION
melts completely at the solidus from a synthetic equivalents provide a rough
wide range of compositions, and in runs approximation of Po2 in the range 10"8-
at atmospheric pressure, the liquids just 10"9 atm for conditions of crystallization
above the solidus have compositions simi- for the lunar glass. Although it is hi-
lar to those of the high-Ti mare basalts, ferred that these results are due to oxi-
The large variation in Ti02 content of dizing conditions greater than those
the mare basalts, according to Akella and determined from returned lunar rocks,
Boyd, is probably dependent on the ab- this is not to be construed as a highly
sence or presence of ilmenite in the source oxidizing condition because the 57Fe ab-
rocks. sorption spectra show ferric iron present
Mao and Bell measured the oxidation in an amount less than 1 wt %. It is be-
states of titanium and iron in the lunar lieved that the degree of oxidation in the
"basaltic" glasses and interpreted the ob- orange beads has not been altered since
served absorption, particularly that in they were originally formed 3.7 billion
the orange glass beads in the soils col- years ago. A P02 of 10"8-10~9 atm is
lected by Apollo 17. They prepared syn- close to that required for the stability
thetic glasses under oxidation conditions of magnetite in the temperature range
ranging from air to pure hydrogen and 1300°-1400°C, and this factor may
measured the spectral absorption due to prove to be important in explaining a
charge-transfer and crystal-field transi- source for the magnetization on the sur-
tions. It was found that Fe3+, Fe2+, Ti3+, face of the moon.
and Ti4+ (when Fe2+ is present) could be The orange soil of Apollo 17 initially
identified. The transitions observed due caused great excitement because it was
to charge transfer were Ti3+-ligand, Fe3+- thought to represent a large accumula-
ligand, and Ti4+-Fe2+. The crystal-field tion of rust. The rocks of Apollo 16 were
bands were caused by Fe2+. By compar- composed mostly of anorthite with minor
ing the spectra of synthetic glasses at olivine or pyroxene, contained free metal
various Po2 it was noted that Ti3+ formed alloys, had a high content of volatiles
under more oxidizing conditions (Po2 > (i.e., CI, Zn), and contained the H20-
10-6 atm) in glass than in crystals (esti- bearing mineral goethite, a principal
mated P02 < 10"9 atm). A grid of Po2 phase of rust. Taylor, Mao, and Bell
versus spectral change was prepared in deduce a history of anorthosite forma-
order to estimate the oxidation conditions tion, a period in which free metals were
under which lunar and terrestrial basalts introduced, and finally a stage of intro-
form. Attempts will be made to interpret duction of volatiles leading to alteration,
optical telescopic spectra from various Bell and Mao, using optical spectra,
parts of the lunar surface where Fe- and found no detectable Fe3+ in the plagio-
Ti-rich basalts outcrop. clase of the Apollo 16 rocks. However,
With these data at hand, Mao, Virgo, they suspect a very minor amount of Fe3+
and Bell were able to explain the orange on the basis of the hi£n total iron ob"
color of the Apollo 17 soil as due to the served with the electron microprobe rela-
minute orange-colored glass beads found tlve to the Fe'+ iron determined by op-
in smaller quantities at all the other tical spectra. The implication is that the
i«„j: ■* r^i • .-, i t , -, rocks were altered surncially to produce
landing sites. Their widespread distnbu- ,,, , - . ,, , J , ± ±i
,. . . . . . ., rust but not sufficiently to penetrate the
tion and chemical homogeneity suggest constituent crygtals Although their con.
that the beads may have been formed in cern ig that the mgt may be a product of
a single major event (e.g., impact melt- interaction with the earth's atmosphere,
ing, fumarolic activity). The iron and the presence of other volatiles is evidence
titanium absorption characteristics of the of a previously unknown process on the
orange glass matched against those of moon's surface.
GEOPHYSICAL LABORATORY 427
Using samples of the highland rocks of growth of the domains and then was very
Apollo 16, Hodges and Kushiro made a slowly lowered.
detailed study by electron microprobe of The study of the Russian Luna 20
cataclastic anorthosite, recrystallized sample was a feat of microanalytical
breccia, feldspathic basalt, and a spinel technique. Much was learned by the
troctolite. These small but critical sam- examination of single grains as well as an
pies of the highlands reveal a complex aggregate of grains weighing about 50
history of magmatic differentiation, im- mg. Bell and Mao compared the polar-
pact, and impact melting. High-pressure ized crystal-field spectra and the non-
melting experiments on the spinel trocto- polarized diffuse reflectance spectra with
lite 62295 as well as Apollo 15 basalt the scanning reflectance spectra obtained
15016 indicate that these rock types are through earth-based telescopes. The
the result of melting within the moon's crystal-field bands of titanium and iron
interior. Hodges and Kushiro conclude were documented in pyroxene, olivine,
that the lunar interior must be very plagioclase, and glass and identified in
heterogeneous. telescopic spectra of the Luna 20 site.
The pyroxenes from Apollo 15 mare On the basis of the results of this study
basalts were analyzed by Kushiro. The it was possible to determine the local
unusually large pigeonite phenocrysts mineralogy of various parts of the lunar
with rims of augite indicate a period of surface through calculation of reflectance
slow crystallization under subsurface "ratios." The technique will be useful in
conditions followed by rapid crystalliza- the remote sensing of the surface of other
tion on the lunar surface at very low oxy- planetary objects (e.g., Venus, Mars, and
gen fugacities. Kushiro outlined the the asteroid belt) .
range of composition of pigeonite from A detailed study of the Luna 20 plagio-
the detailed analyses. Virgo also exam- clases by Mao and Bell revealed features
ined the pyroxenes from two Apollo 15 never observed before. The zoning of
basalts but deduced a different cooling iron was "patchy" and unsystematic and
history. He characterized the first stage was interpreted as evidence of diffusion
of growth of the large pigeonites as being (perhaps due to shock effects of meteorite
of short duration followed by a rapid impact) rather than chemical zoning. In
heterogeneous crystallization of pheno- addition, fine (0.14 /mi diameter) needles
cryst rims and groundmass. The last of Fe-Ni alloy were discovered as ori-
stage of crystallization took place under ented inclusions. The data suggest
significantly lower /02, as suggested by that plagioclase had been strongly re-
the occurrence of Ti3+ in the pyroxene duced chemically after crystallization. It
structure. The pigeonite phenocrysts in will be important to assess the effects of
both rocks are characterized by the gen- postcrystallization reduction in all lunar
eral absence of exsolution lamellae, by samples in order to judge correctly the
diffuse 6 reflections (suggesting small crystallization path from the moon's
domains), and by low Fe2+-Mg equilibra- interior.
tion temperature (520°-560°C). It is Volatile fractionation apparently
suggested by Virgo that these diffusion played a major role in establishing abun-
properties are incompatible with a single dance patterns in the solar system. In
subsolidus cooling cycle. One possible order to evaluate its role, Seitz studied
explanation is the occurrence of a sub- the loss of elements from basalt and
sequent heating event, such as contact meteoritic material heated under vac-
with a superimposed lava flow, where the uum. Volatile fractionation was en-
temperature was raised but was still be- hanced under reducing conditions where
low that required to initiate unmixing or elements are lost with increasing tern-
428
CARNEGIE INSTITUTION
perature in the sequence Na, K, Fe, Si,
Mg, and Ca. In the presence of graphite,
iron, as well as the alkalies, may be
depleted from magmas on atmosphere-
free bodies.
Systematic Petrography
Pilot testing of the rock information
system devised by Chayes has resulted
in an updated version of the data base
of the system containing over 11,000
analyses of Cenozoic volcanic rocks and
a bibliography of over 600 titles. Sev-
eral new reduction programs have been
added, but the most important modifica-
tion of the program repertoire has been
the development of a new and much
improved function-generating subroutine.
It is attached to the problem file genera-
tor, which selectively scans the data base
and prepares working tables to be proc-
essed by the various reduction programs.
The new subroutine makes it possible to
generate, at operation time, any linear
combination of system variables that can
be stated in 80 characters of FORTRAN
IV with single-level parentheses. The
current values of up to ten such combina-
tions may be used as rejection criteria or
stored with other variables in the work
table. The net result is a notable increase
in the power of the system.
Interrelations between ratios are
widely used in geochemistry. Much of
the covariance between such ratios is
implicit in the variances of the initial
variables, but common approximation
procedures that yield estimates of this
implicit covariance are subject to the
constraints that the coefficients of varia-
tion of the initial variables are all small
and their covariances negligible. In
many practical situations neither condi-
tion is satisfied. Relaxation of these con-
straints is difficult to achieve with non-
linear approximation techniques, which
often involve further a priori assumptions
about distribution that cannot be evalu-
ated from small samples. Chayes has
developed a simulation procedure that
leads to experimental estimation of im-
plicit covariance between simple ratios
formed from initial variables of any as-
signed means and variances; the user of
the program can also make prior assign-
ments of correlations between initial
variables, provided that no such
variable is involved in more than one
correlation. Chayes finds that the rela-
tive variances of major constituents in
many suites of igneous rocks are large
enough to make first-order approxima-
tions of implicit covariance between their
ratios unreliable; accurate estimates of
such implicit covariances can now be
obtained experimentally.
Irvine has studied one aspect of the
problem of obtaining frequency informa-
tion from large files of rock chemical
data and has developed a ranking pro-
cedure for describing the joint distribu-
tion of a dozen more variables. This
method enables one to examine simul-
taneously all the constituents reported in
a conventional rock analysis, as opposed
to concentrating on only one or two con-
stituents or one or two related variables.
Concise summaries can be prepared of
the compositional characteristics of
major rock types or suites that should be
of considerable value both for general
petrological and geochemical purposes
and in selecting materials for experi-
mental study. Irvine has also devised a
method for projecting a multivariate
distribution onto simple graphs. The
projections in effect look through multi-
dimensional chemical space and show the
relationships of the graphed variables in
the most densely populated regions.
Biogeo chemistry
The bulk of the organic matter on
earth is contained in sedimentary rocks
as dark polymers. Although these ma-
terials are obviously produced from the
components of once-living organisms,
their origin, molecular structure, and
geological fate are not well understood.
GEOPHYSICAL LABORATORY
429
Hoering has produced new evidence
that shows how a portion of these poly-
mers, called humic acid, can arise from
the reaction of two of the most abun-
dant constituents of living cells. When
a pure sugar reacts with a pure amino
acid, a dark polymer called melanoidin
is formed. Hoering has compared a wide
range of chemical and physical proper-
ties of melanoidin and humic acid and
found them to agree. The results of the
comparison were combined with the
theory of the reactions leading to mel-
anoidin formation. Some insight into the
molecular constitution of humic acid was
gained. The melanoidin hypothesis is
valuable because it can account for the
rapid disappearance of common cellular
biochemicals and the rapid appearance
of a dark polymer. It gives a much-
needed focus for future research on a geo-
chemically significant material.
Studies on the geochemistry of amino
acids originated at the Geophysical Lab-
oratory nearly 20 years ago. Since then,
important contributions to the under-
standing of ancient life and to chemical
processes occurring in the earth have
been made in this subfield of organic geo-
chemistry. The success has been due, in
part, to great advances in the sophistica-
tion of amino acid analyses. Most of the
instrumentation used in organic geo-
chemical studies has been adapted from
biochemistry and organic chemistry. Be-
cause of the need to analyze amino acids
at unprecedentedly low levels, a rela-
tively simple instrument capable of de-
tecting picomoles (10~12 moles) of amino
acids has been developed by Hare. Com-
merical instrumentation now available
incorporates many of the features de-
veloped at the Geophysical Laboratory
for amino acid, peptide, and protein
analysis.
Most amino acids can exist in one of
two forms (optical isomers) , which are
called the d- and L-isomers. They differ
from each other only in the relative posi-
tions of parts of the molecule. Amino
acids in living organisms are exclusively
of one form, the L-isomer. During geo-
logic time, the L-form can be partially
converted to the D-form (racemization),
resulting in an equimolar mixture.
Hare and Hoering have applied a new
technique using gas chromatography to
determine the relative proportions of the
optical isomers in a number of amino
acids in fossils. They have shown that
the rate of racemization of different
amino acids varies by more than two
orders of magnitude, with aspartic acid
converting at the fastest rate and valine
at the slowest rate.
The amino acids in the clay fraction of
a sediment can retain their L-configura-
tion for periods of time significantly long
with respect to the geological time scale.
The rate of racemization of amino acids
appears to be sensitive primarily to time
and temperature. In principle, a com-
bined geochronometer and geothermom-
eter exists in the relative proportions of
d- and L-isomers. An extension of the
isoleucine-alloisoleucine dating method,
which was developed here, now seems
possible.
New Techniques, Equipment, and
Calibrations
Finger, Hadidiacos, and Ohashi have
designed and built an interface for a
computer-automated, single-crystal dif-
fractometer. Their design incorporates
several unique features, including an
integral furnace controller under com-
puter control. The system is programmed
in a high-level language, and the pro-
grams may be easily changed for special
applications. The programs for the de-
vice include a complete set of crystal ori-
entation functions and intensity collec-
tion routines. In addition, a special tech-
nique to measure intensities with con-
stant precision has been developed.
The computerized monitoring and con-
trol capability provided by Finger and
Hadidiacos {Year Book 7^ pp. 598-600)
for the electron microprobe gave Chaves
430
CARNEGIE INSTITUTION
the opportunity to explore the feasibility
of automating the routine operations of
modal analysis. Partial analysis of dis-
criminating elements was found to be
sufficient for mineral identification.
Present efforts are being devoted to re-
ducing errors resulting from polyminer-
alic points, sample imperfections, change
in focus, and alterations. Even if these
difficulties are minimized, Chayes does
not believe automated modal analyses
will be particularly rapid; however,
operator fatigue will obviously be elimi-
nated. Fatigue has been the greatest
obstacle to obtaining adequate modal
analyses.
Hare has developed a fluorescent
method for the analysis of amino acids
and peptides. The new reagent, fluores-
camine, forms highly fluorescent deriva-
tives of amino acids that can be detected
by sensitive fluorescence spectroscopy.
In many cases it is possible to distinguish
between free amino acids and peptides
by varying the pH of the reaction.
Frantz and Hare devised a microana-
lytical technique for the analysis of dis-
solved aqueous silica, using a flow-cell
calorimeter for the detection of silica
molybdate complexes. The method yields
precise results of concentrations from 1
to 10,000 ppm silica in 1 /xl of solution !
A new gas-media apparatus for gen-
erating hydrostatic pressures up to 30
kbar while heated internally up to
1400 °C was designed by Bell to study
geochemical reactions. The first reaction
studied was albite ^± jadeite + quartz
because it had been thoroughly investi-
gated, using solid-media (nonhydro-
static) pressure equipment, as a standard
for interlaboratory comparison. Fur-
thermore, because the reaction is geolog-
ically important, an accurate determina-
tion is required. The best value for the
reaction at 600°C is 16.4 ± 0.5 kbar,
which compares with the previously de-
termined value of 16.2 ± 2 kbar after
application of a friction correction of 2
kbar.
Kushiro determined the detailed melt-
ing relations of pure diopside, used in
thermocouple calibration, with the aid of
the electron microprobe. He showed that
pure diopside is not stable above at least
1375 °C and that for thermocouple cali-
bration within an error of ±0.5 °C a run
in which the amount of diopside (in solid
solution crystals) with enstatite is less
than 20% must be obtained.
GEOPHYSICAL LABORATORY
431
EXPERIMENTAL AND FIELD STUDIES
IN IGNEOUS PETROLOGY
Structure of the Upper Mantle
Beneath Lesotho
F. R. Boyd and P. H. Nixon?
It has long been understood that
diamond-bearing kimberlites have been
erupted from deep within the earth's in-
terior (e.g., Wagner, 1914) , and it is an ac-
cepted modern view that the fragments of
ultramafic rock that they contain are
samples of the mantle (e.g., Nixon, von
Knorring, and Rooke, 1963) . Neverthe-
less, the record provided by these sam-
ples has hitherto been difficult to read
because it has been scrambled in the
course of eruption. The stratigraphic
relations and the characteristic depths
from which came the different kinds of
mafic and ultramafic rock fragments
have been unknown.
Application of experimental studies
now suggests that kimberlites have been
erupted from depths in the mantle as
great as 200 km. In their explosive
ascent to the earth's surface they have
incorporated and transported upward a
remarkable variety of rocks that appear
to have come from the low-velocity zone
as well as the lithosphere. Estimates of
equilibration pressures for these rocks
provide the depths of origin, and thus
these fragments from the mantle can be
ordered into a geologic section.
Estimates of equilibration tempera-
tures and pressures are made through
consideration of experimentally deter-
mined phase studies and mineral com-
positions determined by electron micro-
probe analysis. It is necessary that a
sample have equilibrated with the assem-
blage Ca-rich pyroxene + Ca-poor py-
roxene + garnet for the equilibration
conditions to be estimated. At present, it
is also necessary that these phases be
* Department of Mines, Maseru, Lesotho.
poor in FeO. The diopside solvus (Davis
and Boyd, 1966) is used to estimate the
equilibration temperature. Once this
temperature is known, the equilibration
pressure can be estimated from the A1203
content of enstatite, using the data of
MacGregor (1973). This procedure is
discussed in more detail by Boyd (1973).
The Geotherm
Figure 1 (A-D) shows plots of esti-
mated equilibration temperatures and
depths for a group of lherzolite nodules
from northern Lesotho. These estimates
are subject to substantial uncertainties,
but it appears likely that the broad
stratigraphic relations will not be changed
by subsequent refinement. In Fig. 1A
the "raw" percentages of A1203 in the
enstatites were used to estimate the
equilibration pressures. These data ap-
pear to define an inflected geotherm in
which the points on the shallow limb plot
close to the predicted geotherm of Clark
and Ringwood (1964). In Fig. IB the
estimated equilibration pressures have
been corrected for the presence of minor
amounts of FeO in the natural enstatites
and garnets, using a method developed
by Wood and Banno (1973). This cor-
rection shifts the points to more shallow
depths. In Fig. 1C it has been assumed
that all the sodium in the natural en-
statites is present as jadeite (NaAlSioOu) ,
and an amount of Al equal to the atomic
proportion of Na has been subtracted
from the total Al content before calculat-
ing the equilibration pressure. This as-
sumption shifts the points to greater
depth. Figure ID shows a plot in which
the Al and Cr have been combined before
subtraction of Na, and this assumption
provides a plot intermediate between A
and C (Fig. 1). These operations on the
data illustrate but do not completely de-
432
CARNEGIE INSTITUTION
O
o
CD
^_
Z3
O
^_
CD
Q.
E
CD
1600
400
200
000
800
600
A
RAW AL203
SHIELD GEOTHERM^ ^^
PYROXENE GEOTHERM
50
200
250
Depth, kilometers
Fig. 1. Estimates of the temperatures and depths of equilibration of lherzolite nodules, pri-
marily from the kimberlites of northern Lesotho. Localities for these samples are: Thaba Putsoa,
Letseng-La-Terai, Monastery (Orange Free State), Matsoku, Lemphane, Kao 2, Mothae, and
Liqhobong. The anomalous sheared lherzolite is 1654 from Matsoku. Temperatures are estimated
from the diopside solvus of Davis and Boyd (1966) using the Ca/(Ca + Mg) ratio of the natural
diopsides. Pressures are estimated from the data of MacGregor (1973) on the system MgSi03-
MgaALSisOjn, using four different methods of reduction (A-D). See Boyd (1973) for a more de-
tailed description of these procedures. (A) Pressures estimated from the "raw" AI2O3 contents of
the natural enstatites. (B) Pressures estimated as in A but with corrections applied to compensate
for the effects of Fe in the natural enstatites and garnets using the calculation developed by Wood
and Banno (1973). (C) Atomic proportion of Na subtracted from Al prior to calculation of pres-
sure. (D) Atomic proportion of Na subtracted from combined Al and Cr prior to calculation of
pressure.
fine the uncertainties. Nevertheless, in
all these plots the overall form of the
geotherm and the order of points remain
the same.
The Lherzolites
The lherzolites from the kimberlites of
northern Lesotho form two groups that
differ markedly in texture, mineralogy,
bulk composition, and mineral chemistry.
A group whose equilibration points fall
on the shallow limb of the geotherm (Fig.
1) is characterized by a coarse-grained
(2-4 mm) , granular texture (Fig. 2A) . A
second group, whose equilibration points
fall on the inflected limb of the geotherm,
is intensely sheared. Lherzolites in the
second group contain coarse porphyro-
blasts of garnet and diopside in a fine-
grained, granulated groundmass of oli-
vine and enstatite (Fig. 2B-D). Some
of the granular lherzolites contain acces-
GEOPHYSICAL LABORATORY
433
1600
1400
o
1200
0
(D
v_
ZJ
|
O
1000
i_
<D
Q.
E
CD
h"
800
600
RAW AL203
BANNO AND WOOD
CORRECTION
100 150 200
Depth, kilometers
250
sory phlogopite, chromite, and rare
graphite, but these minerals have not
been found in the sheared lherzolites.
Bulk analyses (Table 1, Fig. 3) show
that the granular lherzolites are depleted
in Na, Ti, Fe, Ca, and Al relative to the
sheared lherzolites. These elements frac-
tionate into the liquid during partial fu-
sion, and escape of the liquid has evi-
dently caused the depletion. The two
groups of lherzolites can be distinguished
on the basis of FeO/(FeO + MgO),
which is less than 0.10 for all the granu-
lar lherzolites and greater than 0.12 for
all the sheared lherzolites (Fig. 3) . Most
of the lherzolites in both groups, how-
ever, are somewhat depleted relative to
pyrolite composition (Green, 19736).
Sample PHN 1611 from Thaba Putsoa
is closest to pyrolite, but it contains
substantially less Ti02 and Cr208 than
does the hypothetical pyrolite compo-
sition. Additional analyses of granular
lherzolite nodules from the Wesselton,
Bultfontein, and Jagersfontein pipes
have been provided by Chen (1971).
These differences in bulk composition
are reflected in the compositions of the
pyroxenes and garnets in the two groups
of lherzolites (Boyd and Nixon, Year
Book 71, pp. 362-373). Ti02 is strongly
depleted in the pyroxenes and garnets of
the granular lherzolites; in contrast,
Cr2Os is enriched in these minerals be-
cause depletion of Al by removal of
interstitial liquid causes Cr/Al to in-
crease. These variations can easily be
detected in hand specimens inasmuch as
Cr-rich garnets have a violet tint and
Cr-rich diopsides are a deep emerald
green.
Origin of the Inflection
It is unlikely that the geotherm illus-
trated in Fig. 1 describes a steady-state
434
CARNEGIE INSTITUTION
1600
1400-
o
0
1200
CD
i_
3
-t—
o
1000
CD
Q.
E
CD
H
800
600-
100 150 200
Depth, kilometers
250
situation. For a steady-state geotherm
to have such an inflection would require
that the rocks above and below the in-
flection have markedly different thermal
properties. There are compositional and
textural differences (described above)
between the sheared lherzolites on the
inflected limb of the geotherm and the
granular lherzolites on the shallower
limb, but it seems unlikely that these
differences would affect the thermal
properties to a significant degree. It ap-
pears more likely that the geotherm in
Fig. 1 was perturbed. In other words, it
appears that in a period prior to the
eruption of the kimberlites a steady-state
geotherm was established in the mantle
beneath Lesotho with a slope rather like
the "shield geotherm" of Clark and Ring-
wood (1964). This primitive geotherm
was then perturbed by an event that
caused major heating in the depth range
150-200 km. Possibly the postulated
heating was related to the intense shear-
ing of the lherzolites that define the in-
flected limb of the geotherm and possibly
that shearing resulted from the plate
movements that occurred during the
breakup and dispersal of Gondwanaland.
Briden (1967) has compiled paleomag-
netic data for the southern hemisphere
and found evidence for four episodes of
drift since the Cambrian separated by
quasi-static intervals. The principal epi-
sode of drift occurred when Gondwana-
land broke up in Late Triassic and Juras-
sic time after a quasi-static interval last-
ing for approximately 200 m.y. Smith
and Hallam (1970) concluded that the
initial rifting of Gondwanaland began
in the Late Jurassic and Early Cretace-
ous but that much of the dispersal oc-
curred in Late Cretaceous and Tertiary
times. Most of the kimberlites in Leso-
GEOPHYSICAL LABORATORY
435
50
100 150 200
Depth, kilometers
250
tho and South Africa are believed to have
been erupted in the Late Cretaceous (e.g.,
Wagner, 1914; Dempster and Richard,
1973), although there are as yet few ra-
diometric dates. Dr. T. E. Krogh (per-
sonal communication) has obtained an
age of 90-110 m.y. for a crystal of zircon
from one of the pipes in the Kimberley
area. Thus the African plate was in
motion after a long, quasi-static period
when the kimberlites were erupted.
The idea that friction may be an im-
portant source of heating in the mantle
has been extensively considered in recent
geophysical literature. McKenzie (1969)
suggested that stress heating at shallow
levels of the asthenosphere is the cause of
the anomalously high heat flow found in
regions behind island arcs, and Hasebe,
Fujii, and Uyeda (1970) have ascribed
an important role to frictional heat in
the thermal regime of a subduction zone.
The data reductions illustrated in Fig. 1
suggest that the sheared lherzolites of
deepest origin may have been stress-
heated by as much as 300°C above their
ambient preshearing temperature.
The Ilmenite Association and the
Discrete Nodules
Mg-rich ilmenite is a distinctive and
abundant mineral in kimberlites. Usually
the ilmenite occurs as rounded nodules,
which may be polygranular and which
commonly are over a centimeter in diam-
eter. Less commonly it forms granular
intergrowths with one or more of the
characteristic minerals of the garnet
lherzolite assemblage. Some megacrysts
of ilmenite contain inclusions of garnet,
pyroxene, or olivine; and megacrysts of
garnet and pyroxene may also include
ilmenite. Among the megacrysts at
Monastery and other pipes are extra-
ordinary lamellar intergrowths of ilmen-
436
CARNEGIE INSTITUTION
Fig. 2. Textures of typical granular and sheared lherzolites from northern Lesotho. Widths of
fields of view are approximately 0.8 cm. Photos of sheared nodules taken in transmitted light;
photo of granular nodule taken under crossed nicols. (A) Sheared lherzolite, 1611, Thaba Putsoa.
(B) Granular lherzolite, 1569, Thaba Putsoa. (C) and (D) Sheared lherzolite, 2273, Kao.
ite with diopside or enstatite in which the
lamellae of pyroxene and lamellae of
ilmenite each form single crystals (Boyd
and Nixon, 1973; Dawson and Reid,
1970; Ringwood and Lovering, 1970).
Compositions of garnets and pyroxenes
intergrown with ilmenite are remarkably
similar, regardless of the type of inter-
growth (Fig. 4). The garnets and en-
statites show a small range in Mg/(Mg
+ Fe), believed to be due to fractiona-
tion, and the diopsides show a small vari-
ation in Ca/(Ca + Mg), which prob-
ably reflects a variation in equilibration
temperature. Two of the nodules studied
in this investigation contain ilmenite to-
GEOPHYSICAL LABORATORY
437
gether with garnet, diopside, enstatite,
and olivine; and tie lines joining the com-
positions of coexisting pyroxenes and
garnets in these rocks are shown in Fig.
4. The compositional similarities be-
tween the minerals in these two rocks
and the other pyroxenes and garnets in
Fig. 4 suggest that the other ilmenite-
bearing nodules have also crystallized in
equilibrium with the phases enstatite +
diopside + garnet, although the original,
complete assemblages have evidently
been fragmented in the course of eruption.
Neither of the two ilmenite-bearing
rocks that contain the lherzolite mineral
assemblage resembles the granular or
sheared lherzolites described previously;
in fact, they appear to be cumulates.
Nevertheless, it is possible to estimate
their temperatures and pressures of
equilibration in the same manner as was
done with the lherzolites. Such estimates
are shown as points A and B in Fig. 5,
and they plot in a temperature range
438
CARNEGIE INSTITUTION
Ph
S-i
o3
a
o
U
o"
45
CD
h-5
•a;
+^
in
o
of
O
O
N
'—
cd
45
CD
55
Si
a
Is
$5
T3
CD
o
45
CO
53
C
a
o
J-H
Ph
CO
CO
CO
tO
CO
I
tO
OS
CO
OS
tO
OS
T— I
OS
o
tO
CM
o
CO
<M
OS
CO
COrHTtlCOTt<TtiI>00t^CO
f-H l> tO ■*+• © i— IT^OlOi— 1
CO CO
OS
I>
iddcodoooNMOO •
^ CO
'do
OS
OS
OiCWOOHMMCOCO'^iC'*
NNNMOOHNNMHOOl
eodiNHOodNcodddH
TjH CO
T3
00 CO 00 CO -^
iOh HNOl
r— I O r- 1 i— I 00
i-h t^ O (N CO O
H T)* O) H O H
o co
O O O O <M
i— i CM CO
O T-l <M
odd
V
iO iO N N iO (N
i— i i— i cn oq co t— i
O C<l O CO o
HOO^OMiOMH .
<Np<NOrHCOp<N#T3
OOcOOMNiO^iCMONHOM
OH00NNH00ONOHHONN
NOHHtfldHNoddNddd
TtH t|h y
OJOOhh^hcOMIMCONh
CO © O OS CO i— i »ONr i © © OS
*-<d'-<'-<dd'tfddd©c<i
MNN05-*NiO'<tHCOO
NHNOOOH(OC5(NOO
O T— I T— I CO O
<M t-h O O O CO
HtOfOHOOCONtONOH
MOiOMOOh^OJOOON
^doH^'diodddoN
CO^iMiOCDOHMNCOOiO
NOCNOONHCONOOOiO
MOOH^diOHoddd
<=> d
t- CO
<M CO
d d
r^ CO
d d
CO Tfl
d d
NINOlNCXJOWCNNtONiOHCO'*
WO«OOOtO'HN.iOOOHNqHN
^ ^ y
(NNMNOOH©CClNNN
lOpcqcoiOT-Hcq-^popco
t^ddoacodcoddddr-I
99S1
+
CO
d
o
CO
d
o
00
OS
OS
CO
o
d
o
00
d
os
CO
co
d
o
d
o
CO
d
o
o
to
d
o
co
os
d
os
o
H
OS
p
id
p
co
p
CO
oi
CO
d
d
<M
d
O
CD
X
o
o
O
+
o
<D
Pn
cd
Ph
CD
40
c3
O
GO
51
Ph
4D _&i
CJ3 'q.
45 ^
H 5
CD
s.i
CO
CO
OS
l-1 4)
-I
CO CD
to c3
--H 45
CO o
W^
CO~tO
OS C<l
tO OS
o ^
to OS
OS t-
CO t—l
tO OS
CD
4*
o
o
• Ph
^ 3
TO
CD
0.
O
-P
'55
O
a
o
CD
o
42 C
S 2
.S fl
CD
45
o
^ >>
bC45
0
o3
i-,
O
S
CO .
OS Ph
_, 42
^ Jh
flO 4^
O
X
^H (J}
03
J5
CD
45
o co et rs
c1Wh?0
T3
CD
a
CD
. ^
CD +3
55 (5
3 d
GEOPHYSICAL LABORATORY
439
4.0
GRANULAR
_ CaO
SHEARED
o
3.0 1-
L ai2o3
0
c
O
0.3 1-
22 n
0.4
CD
0
0.3r
0.
• •
• •
_ TJO
• %
m «:
Na20
Cr203 V
I I I I L
J I I I L
5 10 15
lOOx FeO/(FeO+MgO)
n4.0
0
13.0
0
0.3
0
-i0.4
0
0.3
0.1
I I 1 I
20
Fig. 3. Variations of minor elements among the rock analyses given in Table 1.
that is intermediate between the granular
lherzolites and the sheared lherzolites.
If the assumption is made that other
pyroxene-ilmenite nodules (Fig. 4) crys-
tallized in equilibrium with the assem-
blage enstatite + diopside + garnet, it
is also possible to make partial or crude
estimates of their equilibration condi-
tions. The temperature range of equili-
bration of a variety of diopside-ilmenite
nodules is indicated by a bracket in Fig.
5. It can be seen that this bracket spans
the temperature interval between the
granular lherzolites and the sheared
lherzolites and overlaps the range for the
sheared lherzolites. It is possible to esti-
mate the equilibration temperatures of
enstatite-ilmenite nodules from the data
shown in Fig. 6 and to estimate their
equilibration pressures from their AUO3
contents. These estimates are consider-
ably less precise than can be made for
lherzolites. Nevertheless, estimates for
all but one of the enstatite-ilmenite
nodules plot close to the sheared lherz-
olite limb of the geotherm, overlapping
its shallow end. The fact that these esti-
mates approximately fit the geotherm
440
CARNEGIE INSTITUTION
• LAMELLAR
O GRANULAR
□ DISCRETE
DIOPSIDES
1 GARNETS
ENSTATITES
V ^
_^_
Fe
Fig. 4. Compositions of diopsides, garnets, and enstatites that coexist with ilmenite in various
kinds of nodules. The solid triangle is for coexisting minerals in nodule 1680B (Matsoku), and the
dashed triangle is for nodule 1582 (Thaba Putsoa).
based on the lherzolites provides strong
support for the assumption that the vari-
ous ilmenite-bearing nodules were in fact
originally equilibrated with the assem-
blage enstatite + diopside + garnet.
The data in Figs. 1, 5, and 8 show that
at least some kimberlites (including Mon-
astery) have originated at depths greater
than the ilmenite-bearing nodules. It
thus appears that the ilmenite-bearing
nodules were not indigenous to the kim-
berlites but were picked up by erupting
kimberlites in the same manner as other
mantle rocks.
Some of the discrete nodules contain
ilmenite as a host, inclusion, or lamellar
intergrowth, but others consist solely of
garnet or pyroxene or intergrowths of
these phases. These ilmenite-free dis-
crete nodules are usually more magnesian
than the ilmenite-bearing varieties. The
diopsides are subcalcic (Fig. 7A) and
show a range of Ca/(Ca + Mg) corre-
sponding to a range in temperature over
100 °C. They are uniformly more Fe-rich
than the diopsides from the lherzolites.
Compositions of the garnet discrete nod-
ules especially contrast with those from
the lherzolites (Fig. 7B). They show a
substantial range in Mg/(Mg.+ Fe) with
the most Fe-rich varieties commonly
found in association with ilmenite. The
GEOPHYSICAL LABORATORY
441
O
o
o
1600
1400
1200
1000-
a>
Cl
E 800
<x>
600-
400
SHEARED
LHERZ0LITES
1/
B
A
/
/
/
/
/
/
/
/
GRANULAR
LHERZOLITES
TEMPERATURE
RANGE
_i_ DIOPSIDES
0
50
100
150
200
250
Depth, kilometers
Fig. 5. Temperature-depth relations for ilmenite-bearing nodules compared wtih the geotherm
based on the lherzolites. Points A and B are for ilmenite-bearing nodules 1680B and 1582, respec-
tively. The interval identified as "Temperature Range Diopsides" includes estimated equilibra-
tion temperatures for a variety of diopsides that coexist with ilmenite. The large square points are
estimates of the temperatures and depths of origin of. enstatites that coexist with ilmenite, based
in part on the data in Fig. 6. The version of the pyroxene geotherm shown in this figure was con-
structed using the "raw" weight percentages of A1203 in the enstatites (Fig. 1A).
discrete nodule garnets have a relatively
constant Ca content (Fig. 7B), whereas
those from the lherzolites have a range
of Ca content.
Estimates of the conditions of equili-
bration can be made for the pyroxene
discrete nodules in the same manner as
was done for the ilmenite-bearing varie-
ties (Fig. 8). The enstatites plot close to
the sheared lherzolite limb of the geo-
therm with a range in equilibration tem-
perature of 250 °C. The temperature
range of the diopsides overlaps the high-
est temperature sheared lherzolites, but
there is no overlap at lower temperatures
with the diopside-ilmenite lamellar inter-
growths (Fig. 8). This last relationship
is puzzling because one would expect that
some diopsides that coexisted with ilmen-
ite at depth became disaggregated on
eruption. Nevertheless, such specimens
have not yet been found.
Discussion
Thus the mantle section revealed by
this study of inclusions in the kimberlites
of northern Lesotho consists primarily of
depleted but phlogopite-bearing granular
lherzolites that are underlain by less de-
pleted but phlogopite-free sheared lherz-
olites. Most of the ilmenite-bearing
rocks and megacrysts appear to have
originated in a depth range extending
down from approximately the point of
inflection of the geotherm and overlap-
442
CARNEGIE INSTITUTION
1500
O
°^ 1400
^^_
Estimated temperature
o o o o o
o o o o o
—
Si ° '
I680BX.
Jv' ^-1582
OCX
78
j
800
(
/
l I 1 I i i i
)
5 10 15 20 25 30 35
Ca/(Ca + Mg) x |Q00
Fig. 6. An empirical "enstatite geothermometer" based on temperatures of equilibration esti-
mated for coexisting diopsides. Open points are for enstatites in the lherzolite nodules (Nixon
and Boyd, in preparation). Solid points are for two ilmenite-bearing nodules. See Mysen (this
Report) and Hensen (this Report) for experimental data bearing on the location of this solvus
boundary.
ping the more shallow sheared lherzolites
(Fig. 5) . More magnesian pyroxene and
garnet megacrysts are believed to have
originated at still greater depth along
with the deepest sheared lherzolites (Fig.
8).
The discrete nodules or megacrysts are
distinguished from their counterparts in
the lherzolites by an ultracoarse grain size
as well as by compositional differences.
Comparably coarse crystals of pyroxene
and olivine have been found in pegma-
titic zones and metasomatic rocks in
other kinds of ultramafic intrusions.
Irvine (1973) has described occurrences
of such rocks in the Duke Island ultra-
mafic complex and attributed their origin
to the action of volatiles emanating from
crystallizing interstitial liquids. The ac-
tion of volatiles alone would not appear
to explain the strong fractionation in
Mg/(Mg + Fe) shown by the discrete
nodules (e.g., Fig. 7B) . But it is possible
that crystallization in the presence of an
H20-rich melt over a long period of time
might produce ultracoarse crystals and
might also provide an environment in
which fractionation could occur.
The idea that the discrete nodules
might be phenocrysts in the magmas that
GEOPHYSICAL LABORATORY
443
CaSi03
MgSiO-
CaSi03
DIOPSIDES
-*r-^CaFeSi206
1400° — 30
MgSi03
Mole per cent
LHERZOLITE
o GRANULAR
© SHEARED
MEGACRYST
• NO ILMENITE
a WITH ILMENITE
Fig. 7. Compositions of diopside and garnet discrete nodules compared with the compositions of
diopsides and garnets from the granular and sheared lherzolites. (A) Diopside analyses plotted in
the northwest corner of the pyroxene quadrilateral. (B) Atomic proportions of divalent cations in
the garnet analyses.
initiated the kimberlite eruptions (e.g.,
Nixon, von Knorring, and Rooke, 1963)
now seems improbable because of the
large ranges in equilibration tempera-
ture exhibited by suites of discrete py-
roxene nodules from individual kimber-
lite pipes. For example, fourteen bronzite
nodules from Letseng-La-Terai have
equilibration temperatures ranging from
1170° to 1360°C and five diopside nod-
ules from Monastery (including diopside-
ilmenite varieties) have a range of equili-
bration temperature of 1200°-1365°C.
Nevertheless, the large range in equili-
bration temperature found for the dis-
crete nodules could be explained by a
model in which the discrete nodules are
taken to be megacrysts in large volumes
of crystal-mush magmas in the low-
velocity zone. Moreover, such a model
appears to be consistent with the occur-
rence of phlogopite in the lherzolites and
with experimental studies of the lherz-
olite solidus in the presence of H20.
Phlogopite is commonly found in the
granular lherzolites, and its occurrence
has also been observed in a few of the
ilmenite-bearing nodules. Nevertheless,
no phlogopite has been found in any of
the sheared lherzolites of deep origin;
sample 1654 from Matsoku is slightly
sheared and it contains phlogopite, but
its temperature-depth relations are equiv-
alent to those of the granular lherzolites
(Fig. 1). Sample 1582 has the highest
equilibration temperature of any of the
phlogopite-bearing nodules — it is esti-
mated to have crystallized at 1115°C.
444
CARNEGIE INSTITUTION
LHERZOLITE
o GRANULAR
e SHEARED
MEGACRYST
•NO ILMENITE
LWITH ILMENITE
This temperature is slightly above the
inflection in the geotherm (Fig. 5) . Lam-
bert and Wyllie (1968) have suggested
that H20 is largely stored in hydrous
minerals in the lithosphere, whereas in
the upper part of the asthenosphere it is
dissolved in small amounts of interstitial
silicate liquids. Possibly the occurrence
of phlogopite in these nodules reflects
such a relationship.
Experimental studies of the solidi of
peridotites at high pressures show that
H20 produces an extremely large de-
crease in the temperature range for the
beginning of melting (e.g., Kushiro,
Syono, and Akimoto, 1968; Mysen, this
Report). Mysen 's data suggest that
peridotite would begin to melt at ap-
proximately 900 °C at a depth of 150 km
if the H20 pressure were equal to the
total pressure. Moreover, the beginning
of melting remains below 1100°C at
this depth if C02 is added to the system
and the mole fraction of H20 in the vapor
phase is reduced to 0.25. Thus small
amounts of melting would be expected in
the depth range from which the sheared
lherzolites have come, provided small
amounts of H20 were present. Neverthe-
less, there would be no melting in the
absence of H20.
Thus various lines of evidence suggest
that the point of inflection in the geo-
therm (Fig. 1) might have been the top
of the low-velocity zone in late Cretace-
ous time when these kimberlites were
erupted. The megacrysts of garnet, py-
roxene, and ilmenite are interpreted as
having formed in large volumes of crys-
tal-mush magmas in the upper part of
the asthenosphere. The sheared lherz-
olites from this same depth range in the
GEOPHYSICAL LABORATORY
445
1600
1400
O
o
o>
CL
E
1200
1000
800
600
400
PYROXENE
DISCRETE
NODULES
LOW
VELOCITY
■►?
ZONE ?
/
O
SHEARED
LHERZOLITES
/
T
A
i.
o
GRANULAR
LHERZOLITES
CD T
B
1
/l
/
/
/
/
O ENSTATITE
ONLY
9 ENSTATITE
+ ILMENITE
-*■ UNCERTAINTY
50
100
150
200
250
Depth/^kilometers
Fig. 8. Temperature-depth relations for various pyroxene discrete nodules compared with the
geotherm based on the lherzolites. Procedures used in constructing this figure are analogous to
those used for Fig. 5. Temperature range A is for diopside discrete nodules that do not contain
ilmenite. Temperature range B is for diopside discrete nodules intergrown with ilmenite.
asthenosphere have probably been
erupted intact because they were dry.
Most of the sheared lherzolites show
some degree of depletion relative to
pyrolite (Table 1). Thus they may have
undergone small degrees of partial fu-
sion, and the liquids with dissolved H20
may have been kneaded out by the
shearing process. Weertman (1972) has
emphasized the importance of shearing
stresses in causing the coalescence of dis-
persed liquids in mantle rocks.
Green (1971) has suggested that mag-
mas in the low-velocity zone may ini-
tially contain less than about 5% liquid.
If such liquids gradually seep and ooze
upward through large volumes of crystal-
mush, following a P-T path along the
geotherm, some crystallization must oc-
cur as they cool. The effect of such a
process over a long period of time might
be to enrich the upper portion of the
asthenosphere in Fe and Ti and also in
certain trace elements (Green, 1971) . The
concentration of ilmenite-bearing rocks
and magmas (?) near the top of the low-
velocity zone might be caused by such
enrichment.
If the interpretation of the discrete
nodules as megacrysts in crystal-mush
magmas is correct, it appears inescapable
that erupting kimberlites would pick up
and include the interstitial liquids as
well as the crystalline grains. If so, the
liquid phase in a kimberlite eruption may
be hybrid. Various authors, including
Dawson and Hawthorne (1973), have
suggested a genetic relationship between
carbonatites and kimberlites. It is inter-
esting to speculate that kimberlite mag-
mas might be carbonatite-silicate hy-
brids.
446
CARNEGIE INSTITUTION
Ultramafic Nodules from Colorado-
Wyoming KlMBERLITE PlPES
David H. Eggler and M. E. McCallum*
Twenty-three kimberlite diatremes,
probably of Devonian age, penetrate
Precambrian crystalline rocks in the
northern Front Range of Colorado and
Wyoming. The pipes, which range in size
from a few tens of feet to 1800 feet in
longest dimension, occur in a north-south
trending zone between Tie Siding, Wyo-
ming, and Red Feather Lakes, Colorado.
Only the largest pipe, the Sloan diatreme,
has been described in detail (McCallum
and Eggler, 1971).
Kimberlite is predominantly an intru-
sive breccia of subrounded to angular
clasts in a finely crystalline matrix of
serpentine, calcite and dolomite, phlog-
opite, chlorite, and talc. Clasts include
individual grains of magnesian ilmenite,
olivine (rare), garnet, chromian diop-
side, phlogopite, chromite, picotite, pe-
rovskite, magnetite, and serpentine
pseudomorphs after olivine and enstatite ;
nodules of spinel and garnet lherzolite,
eclogite (rare), and sovite; and country
rocks. Although blocks of fossiliferous
Paleozoic limestone are present in the
pipes, oxygen and carbon isotopic analy-
ses indicate that carbonatite nodules and
most of the calcite in kimberlite are of
igneous origin (McCallum and Eggler,
1971).
Ultramafic nodules are abundant in
these pipes, but with few exceptions nod-
ules are altered, normally to hematite,
calcite, and serpentine, but also to talc
and quartz. Where possible, nodules,
both mono- and polymineralic, have been
analyzed by the electron microprobe tech-
nique described by Finger and Hadi-
diacos (Year Book 71). Nine polymin-
eralic nodules were analyzed, three
garnet lherzolites, three spinel lherzolites,
four eclogites, and two serpentinized
nodules of indeterminate origin. In the
* Colorado State University, Fort Collins,
Colorado.
garnet lherzolites all olivine and en-
statite have been replaced by serpentine
and quartz. Two spinel lherzolites are
quite fresh and preserve olivine (Fo9i)
and enstatite (En9i). These five nodules
have an equigranular (2-4 mm) , poly-
gonal texture and show no evidence of
shearing.
Because diopsides from the nodules are
closely similar in composition to mono-
mineralic nodules in the pipes (Fig. 9)
and to diopsides in South African kim-
berlite lherzolites (Boyd and Nixon,
Year Book 71) , it is inferred that all the
diopside monomineralic nodules are de-
rived from disaggregated lherzolite. If
this inference is true, they have been in
equilibrium with enstatite, and their
compositions may be compared with
compositions on the diopside solvus
(Davis and Boyd, 1966). Compositions
fall into two temperature groups, 950°-
1000°C and 1150°-1250°C. Calculation
of temperature of equilibration of three
garnet lherzolites from the latter group
using Fe-Mg partitioning between clino-
pyroxene and garnet (Hensen, this Re-
port) yields temperatures of 1120°—
1220°C. Diopsides from spinel lherzolite
indicate lower temperatures, 850°-
950 °C.
Magnesian garnets, which have high
Cr contents (Fig. 10), are also similar to
those in South African ultramafic nodules
(Boyd and Nixon, Year Book 71). t A
group of more iron-rich garnets includes
five megacrysts 2—13 cm in diameter.
This group has very low Cr content (Fig.
10) and higher Ti02 content (about 1.0
wt %). These data are consistent with
a model suggested by Boyd and Nixon
(Year Book 71 and this Report) . Accord-
ing to the model, chrome-rich garnets are
derived from accidental inclusions of
garnet lherzolite, which were subjected to
varying degrees of partial melting before
inclusion. Because Cr is strongly parti-
tioned into residual garnet rather than
liquid, Cr203 contents of garnet reflect
depletion of the lherzolite by melting. It
GEOPHYSICAL LABORATORY
447
Fig. 9. Compositions of clinopyroxenes (CPX) and garnets (GT) from Colorado- Wyoming kim-
berlite pipes. Temperatures on diopside solvus from Davis and Boyd (1966). Compositions are
in mole percent. Dashed lines connect coexisting phases.
is interesting that the most Cr-rich
garnets also have higher Ca/Mg ratios,
suggesting that they equilibrated under
a different temperature-pressure regime.
The lower Mg/Fe ratio and higher Ti
content of Cr-poor garnets indicate that
they are not disaggregated pieces of
lherzolite, but rather were once in equi-
librium with a phase very low in Cr and
high in Fe and Ti. Their large size sug-
gests that phase was a melt.
Megacrysts of ilmenite comparable in
size to garnet megacrysts are found in
many pipes, and ilmenite is believed to
have been a liquidus phase in addition to
garnet. Ilmenite grains show consider-
able variation in MgO and Fe203 con-
tent (Fig. 11) and in Cr203 content (Mg-
rich ilmenites contain over 3.0% Cr203;
Fe-rich ilmenites, less than 0.3%) . These
variations probably reflect a fractiona-
tion path of the liquid, an idea also ad-
vanced by Mitchell (1973). Geochemical
clues to the character of the liquid are
insufficient to distinguish whether it was
derived by partial melting of garnet
lherzolite at the same depth from which
lherzolite nodules were picked up or
448
CARNEGIE INSTITUTION
o
1
1 1 1
1 1
i i
1 1
CP
^ 0.30
X
+
oo
o
o
O 0.26
—
o
^ 0.22
o
o X
-
o _
O^
0*2°
o
o 0.18
-
O
* 1
CD
1 1 1
1 1
1 1
1 1
16
1
*
1 1 1
1 1
1 1
1 1
12
-
o
^ 8
6°
o ^
*°X o
o
oo
0 o X
o
o x
4
1
1 1 1
i i
1 1
i i
0 2 4 6 8 10 12
Cr203
Fig. 10. Compositions of garnets from Colorado-Wyoming kimberlite pipes. Symbols as in Fig. 9.
whether the megacrysts grew in a liquid the fourth is notably higher in Fe. The
at greater depth. eclogites are believed to be accidental
Four eclogite nodules have been found, inclusions from the upper mantle or deep
each from a separate pipe. Three are lower crust. Their low equilibrium tem-
quite similar in composition (Fig. 9) , but peratures (discussed below) and ex-
"* A A 7T
On
m
nU cP
cP
J. L U * X. x. u \/ v \£ y_
30 40 50 60 70 80
MgTi03
mole per cent FeT'103
FeTi03
Fig. 11. Compositions of ilmenite megacrysts from Colorado-Wyoming kimberlite pipes.
GEOPHYSICAL LABORATORY
449
tremely low K content indicate that they
are not related directly to melting epi-
sodes in the mantle. The high jadeite
content of the pyroxenes (about 35 mole
%) distinguishes the eclogites from gri-
quaites (Boyd and Nixon, Year Book 71,
p. 362). Chemical criteria of Coleman
et al. (1965) and Banno and Matsui
(1965), principally high pyrope and
grossular content of garnet and Fe-Mg
distribution (KD) , separate them from
eclogites from crustal glaucophane
schists, amphibolites, and granulites.
Phase compositions are similar, however,
to those of eclogite from other kimber-
lites (Kushiro and Aoki, 1968) , including
the Stockdale kimberlite in western
Kansas (Meyer and Brookins, 1971).
The stratigraphic position of eclogites
and other inclusions in the mantle before
kimberlite eruption can be estimated.
Diopsides from two spinel lherzolites in-
dicate temperatures of equilibration of
900° and 940°C, as discussed above. The
A1203 content of coexisting enstatite in-
dicates equilibration pressures of 21 and
22 kbar, calculated from data on the
system MgO-Al208-Si02 (I. D. Mac-
Gregor, in preparation). Because these
pressure estimates are uncorrected for
the presence of other elements, they are
probably high but can be checked by
another method. In Fig. 9 spinel lherz-
olite diopside compositions lie on the
low-temperature side of 940 °C and gar-
net lherzolite diopside compositions lie
on the high-temperature side. It appears
that 940°C is the temperature of the
phase transformation in the fossil upper
mantle where garnet first appeared and
that the position of that boundary can be
used as another estimate of equilibration
pressure. The position of the boundary,
which lies at 12 kbar and 940°C for CaO-
MgO-Al203-Si02 (MacGregor, Year
Book 64, p. 128) , shifts to higher pres-
sure with increasing Fe203 and Cr203
content and to lower pressure with in-
creasing FeO content (MacGregor, 1970) .
In these nodules Fe content is low, but
Cr203 contents could shift the boundary
by as much as 7 kbar. A corrected value
of 19 kbar (equivalent to 60 km depth)
and 940 °C lies on the pyrolite spinel-
garnet boundary (Green and Ringwood,
1967c) and occupies a P-T position be-
tween an oceanic and shield geotherm
(Ringwood, Boyd, and MacGregor, Year
Book 63). If other inclusions have equi-
librated at P and T on a geotherm drawn
through that point, garnet lherzolites,
with equilibration temperatures of 950°-
1250 °C, have been brought from depths
of 60-100 km. Values of KD from eclo-
gites indicate equilibration temperatures
of about 825 °C for the Mg-rich group
and 970 °C for the Fe-rich inclusion; the
latter estimate may be in error because
KD is not well known for Fe-rich phases.
Depths of equilibration on the geotherm
are 50 and 65 km, distances at and some-
what below the base of the crust in this
area (Pakiser and Zietz, 1965) . If crustal
thickness has not changed appreciably
since the Devonian, these data suggest
that eclogite is found in the deep lower
crust or as pockets in an upper mantle
that consists of a thin layer of spinel
peridotite underlain by garnet peridotite.
Akermanite-C02: Relationship of
Melilite-Bearing Rocks to
Kimberlite
H. S. Yoder, Jr.
The notion that kimberlite is genetic-
ally related to melilite "basalt" is attrib-
uted to H. C. Lewis (1897, p. 49) from
notes edited after his death by Professor
T. G. Bonney and later supported by
Wagner (1914, p. 106), Shand (1934),
Taljaard (1936), and Holmes (1937).
Reports of melilite in kimberlite have
been rare, however, and then the identi-
fication of the mineral was usually du-
bious or based on the interpretation of
alleged pseudomorphs. Recently the idea
was revived with the description by Uk-
hanov (1963) of the close association of
olivine melilitite and kimberlite breccia
450
CARNEGIE INSTITUTION
in the Bargydamalakh pipe, Anabar re-
gion, USSR. The olivine melilitite forms
the core of a kimberlite breccia that is
cemented with calcite and serpophite.
Although the mineralogy of the kimber-
lite portion was not described in detail,
to the best of the writer's knowledge this
is the first report of a direct connection
between these two rock types.
The principal minerals in the ground-
mass of kimberlite are forsterite, phlogo-
pite, and calcite; therefore, both C02 and
H20 play an essential role, not only in
the explosive character of most occur-
rences of kimberlite, but also in its
mineralogy. A study of phlogopite and
forsterite in the presence of C02 and
H20 ( Yoder, Year Book 68, pp. 236-240)
revealed that the stability field of phlogo-
pite can be expanded under certain con-
ditions in the presence of C02. In addi-
tion, the melting of phlogopite appeared
to be suppressed because of the relative
insolubility of C02 in the liquid in equi-
librium with phlogopite and forsterite.
On the other hand, akermanite, the prin-
cipal component of melilite, was found to
be stable only at pressures below 10.2
kbar in the presence of excess H20
(Yoder, Year Book 66, p. 473). If kim-
berlite was indeed formed at great depths
in the earth from melilite-bearing rocks
as proposed by Taljaard (1936, p. 314)
and others, then the stability field of
melilite would also have to expand with
excess C02. For this reason, later proved
erroneous, the stability field of akerma-
nite (Ak) was investigated in the pres-
ence of C02 in preparation for a study of
Ak-C02-H20.
Akermanite- CO ]2
Two mixtures (Fig. 12) having the
composition 2CaO:MgO:2Si02:C02 and
3CaO:MgO:2Si02:2C02 were studied up
Mole per cent
Fig. 12. The CaO-CaMgSbOe-C02 plane showing the location of diopside (Di), akermanite
(Ak), merwinite (Mer), and calcite (Ct). The two x's mark the compositions investigated.
GEOPHYSICAL LABORATORY
451
to 10 kbar total pressure at temperatures
from 950° to 1475 °C. The first composi-
tion was obtained by mixing purified
samples of natural diopside (Twin Lakes,
Cal.)* and calcite (Baker Chemical Co.;
grade: C.P.) ; the mixture is equivalent to
Ak + C02. The same mixture after com-
plete conversion to the assemblage
Akss -f- Di + G, was run simultaneously
to ensure reversibility. The second com-
position was obtained by mixing optical
grade natural dolomite (Thornwood,
N. Y.) with natural wollastonite (Wills-
boro, N. Y.) ; the mixture is equivalent to
akermanite + calcite + CO2. and also
merwinite + 2C02. The results for the
first composition run in gas-media, high-
pressure apparatus for times up to one
week are presented in Fig. 13.
The maximum stability limit of aker-
manite in the presence of excess C02
was found to be 6.1 kbar. At relatively
lower temperatures and below that pres-
sure, akermanite reacts with C02 to form
diopside + calcite. The reaction takes
place through a range of temperatures,
presumably because of possible solid so-
lutions of dolomite in calcite, akermanite
in diopside (Kushiro and Schairer, Year
Book 63, p. 133), and a variety of solid
solutions in akermanite (Schairer, Yoder,
and Tilley, Year Book 61, pp. 217-219).
The reaction curve, extended through the
measurements of Walter (1963), termi-
nates at its junction with the breakdown
curve of akermanite — » wollastonite -f-
monticellite (Yoder, Year Book 66, p.
473) near 685 °C at a pressure less than
100 bars.
The akermanite crystallizing in the
region marked Akss + Di + G is not
necessarily pure Ca2MgSi207, nor is the
gas phase (G) pure C02. The appear-
ance of diopside implies that the aker-
manite is a solid solution toward the
merwinite composition, akermanite itself
not being stable, or the gas contains ap-
* The analysis of this diopside, the closest to
the end-member composition on record, has been
previously published, with permission, by Smith
(1966).
preciable calcite in addition to C02
and negligible silicate components. The
amount of diopside increases noticeably
with increasing pressure in this region,
and so the extent of solution in Ak or G
is presumed to increase. The akermanite
was determined by electron microprobe
analyses, through the courtesy of Dr. I.
Kushiro, to be close to but not exactly
2CaO:MgO:2Si02 within the limits of
error. Unfortunately the akermanite
usually contains a large number of two-
phase fluid inclusions, and the summa-
tion of the analysis was generally low,
presumably owing to the volatile inclu-
sions. The diopside in equilibrium with
akermanite was stoichiometric within the
errors of analysis. It is believed, there-
fore, that the appearance of Di in the
Ak + C02 composition is due primarily
to the solution of additional components
in the gas phase.
The melting of akermanite in the pres-
ence of C02 appears to be congruent and
produces liquids that quench to a clear
glass. The lowering of the melting curve
relative to the dry melting curve of
Kushiro {Year Book 63, p. 85) is be-
lieved to be due to the solubility of C02
in the liquid. Considerable care was
taken to eliminate the presence of H20
in the starting materials. It is possible
that hydrogen from components of the
pressure vessel, from impurities in the
pressure medium (argon) , or from rubber
granules from the gasket diffused into
the sample container and formed H20.
Water so formed may be responsible for
the lowering of the melting temperature ;
however, the hydrogen pressure devel-
oped from interactions with the pre-
sumed iron-wiistite buffer of the vessel
wall or the nickel— nickel oxide buffer of
the furnace shell is believed to be most
inadequate to produce the observed effect.
The melting of the assemblage diop-
side + calcite is presumably more com-
plicated than the products appear to in-
dicate. If, in fact, both diopside (Di)
and calcite (Ct) are solid solutions, then
a region of Diss -f- Ctss -\- L -\- G should
452
CARNEGIE INSTITUTION
O
_Q
_*:
CD
(f)
CD
Cl.
Diopside +
Calcite
MOO 1200 1300
Temperature, °C
400 1500
1454 ±2
1600
Fig. 13. Pressure-temperature diagram illustrating the new results for akermanite-C02 using the
diopside + calcite (1:1 mole) composition. The short-dash curves represent an extrapolation to
low pressures of the melting data of Kushiro {Year Book 63, p. 85, Fig. 17) for akermanite.
exist immediately above the solidus. The
distinction between stable calcite and
quench calcite when only small amounts
of either are present is not sufficiently
clear where only textural evidence is
available. In addition to quench calcite,
a small amount of quench merwinite was
identified in some but not all of the runs
in the Di + L + G region. It is signifi-
cant that above the liquidus the liquids
quench to a clear, homogeneous glass
without crystalline quench products. The
relevance of this important observation is
given below.
GEOPHYSICAL LABORATORY
453
(G) (Ak)
(L) (Di) (Ct)
Fig. 14. Chemographic analysis of the uni-
variant curves about the invariant point gen-
erated by the phases akermanite, diopside, cal-
cite, liquid, and gas. Parentheses indicate
absent phase. Observed solid solution ignored.
The two x's mark the compositions studied.
The chemographic analysis of the uni-
variant curves that would be generated
in the absence of solid solution is useful
in ascertaining the general composition
of liquid. In Fig. 14 is displayed the
analysis for phases that lie wholly within
the system akermanite (Ak) -diopside
(Di)-calcite (Ct)-C02 (G). The results
on the diopside-calcite composition 1:2
were used to help fix the sequence of
curves and the composition of liquid.
The curve designated by the absent phase
(G) could not be determined experi-
mentally with the compositions em-
ployed. The sequence of univariant
curves observed suggests that the com-
position of liquid involved in the invari-
ant point lies in the area of composition
within the triangle Ak, Ct, and the point
of intersection of the joins Ak-G and Di-
Ct. The C02 content of the liquid pro-
duced on the congruent melting of aker-
manite in the presence of excess C02,
however, must be less — presumably by a
considerable amount — than 13.9 wt %,
the stoichiometric amount involved in
the reaction Di + Ct -» Ak -f- C02.
Relationship of Melilite- Bearing Rocks
and Magmas to Kimberlite
The restriction of akermanite to rela-
tively low pressures in the presence of
excess C02 is evident in Fig. 13. In the
absence of free C02, the assemblage Ak
+ Di + Ct, related to the alnoites, is
stable and would persist until the break-
down of Ak itself at about 15 kbar (Ku-
shiro, Year Book 63, p. 85, Fig. 17) . In-
clusion of the soda melilite component *
may extend this range to higher pres-
sures; however, the minerals in natural
melilite-bearing rocks were shown to re-
act at high pressures to form pyroxenites
in the absence of water and C02 (Tilley
and Yoder, Year Book 66, pp. 457-458).
Similar behavior would be expected in
the presence of the additional phase cal-
cite. It is not uncommon to see veins of
calcite crossing pyroxenites without the
formation of melilite. If kimberlite is
derived from the mantle, then melilite-
bearing rocks are not likely sources.
Melilite, however, is commonly asso-
ciated with leucite (Lc), kalsilite (Ks),
and phlogopite (Phlog), as illustrated in
Fig. 15 (A and B). Olivine melilite
leucitite (ugandite) is probably related
to olivine kalsilite pyroxenite (mafurite)
by the reaction :
2Ca2MgSi207 + 3KAlSi206 +
Ak Lc
(z+l)Mg2Si04->
Fo
4CaMgSi206 + 3KAlSi04 + zMg2Si04.
Di Ks Fo
*It is noted by Dr. I. Kushiro (personal
communication, 1973) that soda melilite may
undergo a reaction with C02 similar to that of
akermanite with C02 :
NaCaAlSi207 + C02^ jadeite + calcite.
Omphacite + aragonite is a well-known assem-
blage in low-temperature, high-pressure en-
vironments.
454
CARNEGIE INSTITUTION
Mol per cent
Fig. 15. (A) The kalsilite (Ks)-larnite (La)-forsterite (Fo)-quartz (Qz) tetrahedron showing
the relation of diopside (Di)-kalsilite (dashed join) to the akermanite (Ak)-leucite (Lc)-for-
sterite plane representing the olivine melilite leucitites. (B) The same tetrahedron as Fig. 15A
with H20. The sub tetrahedron Ak-Di-Fo-Phlog, representing alnoite for the most part, is illus-
trated with heavy lines. Other hydrous phases are neglected.
Diopside + kalsilite occurs with melilite
in the metamorphosed sedimentary
blocks of Brome Mountain, Quebec (Phil-
potts, Pattison, and Fox, 1967), and at
Hendricksplaats, Bushveld (Willemse
and Bensch, 1964, p. 20). The tie line
that sets in as a result of the reaction is
dashed in Fig. 15A. It is important to
note that the three phases-akermanite
(Kushiro, Year Book 63, p. 85, Fig. 17),
monticellite (Yoder, Year Book 66, p.
475, Fig. 75), and leucite (Scarfe, Luth,
and Tuttle, 1966, p. 728, Fig. 1), which
commonly occur together, all break down
at relatively low pressures.
Addition of C02 and H20 to different
proportions of the above products yields
the kimberlite groundmass assemblage:
CaMgSi206 + 3KAlSi04 +
Di Ks
(x + 4) Mg2Si04 + CQ2 + 3H20 -*
Fo
3KMg3AlSi3O10(OH)2 +
Phlog
CaC03 + zMg2Si04.
Ct Fo
It appears that there is some support for
the suggestion that a magma having
melilite affinities could be transformed
into kimberlite with the aid of suitable
volatiles. Loss of volatiles in transit or
crystallization at low pressure would
yield a melilite-bearing assemblage
within the upper crust.
From the relations exhibited in Fig. 13
and the equations listed immediately
GEOPHYSICAL LABORATORY
455
Mer
Mo
Mol per cent
Fo
above, it would seem that clinopyroxene
ought to be a common residual phase in
kimberlite. It is in fact a rare phase *
and then may be derived in part from
fragmented xenoliths incorporated in the
kimberlite. The large array of interme-
diate rocks that do contain clinopyrox-
ene, melitite, and the major phases in
the groundmass of kimberlite are usually
described as alnoite.
Immiscible Carbonatite
The observation that Di + Ct in the
molecular ratios of 1:1 and 1:2 melts to
form a clear, homogeneous glass does
not support the view that carbonatite is
derived through liquid immiscibility
from all ultrabasic compositions. The
concept was advanced by von Ecker-
* Wagner (1914, pp. 107, 112-115) described
a variety of micaceous kimberlite "of which
augite is an abundant microscopic constituent."
mann (1961), who described globular
masses rich in carbonate, which he be-
lieved ascended through liquified melilite
"basalt." "Possible liquid immiscibility"
in a limited sense was also proposed by
Dawson and Hawthorne (1973, pp. 76,
82) after a study of layered kimberlite
sills at Benfontein, South Africa. Ex-
perimental studies summarized by Wyl-
lie (1966) give strong support to the
immiscibility of carbonatitic and ultra-
basic magmas in some alkali-rich sys-
tems.
The high temperatures of melting of
calcite itself impose a limit on its inter-
action with silicates. A preliminary
melting curve based on experiments using
cleavage blocks is presented in Fig. 16,
incorporating the invariant point of
Baker (1962) and the 1-kbar bracket of
Wyllie and Tuttle (1960). Fair agree-
ment was obtained at 10 kbar (1460° ±
10 °C) with the value of Irving and Wyl-
456
CARNEGIE INSTITUTION
600
Temperature, °C
Fig. 16. Preliminary P-T diagram for the melting of calcite. Invariant point from Baker (1962)
and 1-kbar bracket from Wyllie and Tuttle (1960). The melting of calcite is compared with that
of diopside in the presence of an excess of C02 (Eggler, this Report). Calcite may melt at slightly
lower temperatures in the presence of excess C02.
GEOPHYSICAL LABORATORY
457
lie (1973), 1440° ± 10°C, determined in
solid-media, high-pressure apparatus.
Their value is based on 2-minute runs,
using powder, the piston-out method, and
no friction correction (P. J. Wyllie, per-
sonal communication).
Metamorphism of a Siliceous Dolomite
The results presented in Fig. 13 are
also important to the metamorphism of
carbonaceous rocks. The reaction
Diopside + calcite ^± akermanite + C02
is Bowen's (1940, pp. 250-251) step 8 in
the metamorphism of a siliceous dolo-
mite. Because of the exceptionally high
temperatures involved where the C02
pressure is the total pressure, it would
appear that the reaction runs in nature
at very low partial pressures of C02.
The lower limit of C02 pressure for the
above reaction is fixed by the pressure
of the invariant point (^685 °C) involv-
ing the decomposition of akermanite to
wollastonite + nionticellite mentioned
above. Solid solutions involving other
components may lower the temperature
further into a more reasonable range.
Step 11 of Bowen's decarbonation
series involves the reaction
Akermanite + calcite ^±
merwinite + C02.
Preliminary data suggest that this step
occurs at pressures below those investi-
gated. The results of Shmulovich (1969)
are shown in Fig. 13.
Role of C02 in Melting Processes
in the Mantle
David H . Eggler
Volatile species play important roles in
melting processes in the mantle. Al-
though H20 has received a great deal of
attention, there is abundant evidence
that C02 is an important constituent of
magmas, whose chemistry ranges from
ultrabasic to alkaline to tholeiitic, and of
their associated fluids. Examples include
the presence of carbonate associated with
kimberlite and magmatic carbonatites,
C02-rich fluid inclusions in minerals of
peridotite and in phenocrysts in some
basalts (Roedder, 1965), and C02-rich
gas associated with tholeiite eruptions at
Kilauea (Heald, Naughton, and Barnes,
1963).
A number of experimental studies have
been conducted on phase relations in-
volving C02 in systems modeling car-
bonatite and highly alkalic magmas
(Wyllie and Tuttle, 1960; Wyllie, 1965;
Wyllie and Haas, 1965; Koster van
Groos and Wyllie, 1968; Boettcher and
Wyllie, 1969a) . Previous experimental
studies concerned with basalt systems,
however, have assumed that C02 is es-
sentially an inert species (Holloway and
Burnham, 1972; Hill and Boettcher,
1970). Studies have therefore been ini-
tiated on the solubility of C02 in silicate
melts with compositions comprising some
of the components of basalts and on the
effect of C02 on phase relations in the
melting of peridotite.
Solubility of C02: CaMgSi206-C02-H20,
NaAlSi308-C02-H20, and
Mg2Si206-H20-C02
There have been almost no determina-
tions of COo solubility in silicate melts
at high pressure. Yoder found negligible
solubility in phlogopite melt at 10 kbar
{Year Book 68, p. 239) and a small
solubility in akermanite melt (Yoder,
this Report), and several workers (Hol-
loway and Burnham, 1972; Eggler,
1972a) have inferred negligible solubil-
ity in basaltic and andesitic melts at
pressures less than 10 kbar.
For this study, a variety of starting
materials has been used. For measuring
the effect of C02 on CaMgSi206, a mix
composed of MgO, Si02 (quartz), and
CaCOa, loaded alone, with H20, and with
crystalline diopside (CaMgSi206), was
458
CARNEGIE INSTITUTION
employed. For a few runs in which more
C02 was desired, diopside, H20, and
Ag2C204 were used. At run conditions
Ag2C204 produces C02 and a globule of
molten Ag. For investigating the effect
of C02 on NaAlSi308, albite (NaAlSi308)
glass, H20, and Ag2C204 were used. For
studies on Mg2Si206 (a double formula
is used for direct comparison with
CaMgSi206), a mix was prepared from
Si02 (cristobalite) and hydromagnesite.
The latter was supplied by Fisher as
5MgO:4C02:xH20; x was determined by
ignition to be 5.25. The mix was moni-
tored before each weighing for H20 gain
or loss. A portion was fired to crystal-
line enstatite, which was used with the
mix or with Ag2C204 for other composi-
tions. Approximately 10 mg of material
was loaded into Pt capsules for each run ;
the sizes of the run brackets in the
accompanying figures reflect estimated
errors in weighing the various materials
into the capsules.
The melting curve of CaMgSi206 with
excess C02 (Fig. 17) was determined
with dry carbonate mix. The mix was
loaded into capsules, which were then
dried several additional hours and im-
mediately welded. The apparent melting
temperatures are minima, as some H2
enters the capsule during the run and
forms about 1 mole % H20 (Eggler, this
Report). Plotting these runs in an iso-
baric section at 20 kbar (Fig. 21, below)
indicates, however, that the melting point
with pure C02 may be only 10 °C above
the determined curve. The amount of
C02 dissolved in the melt was investi-
gated in the system CaMgSi206-C02
30-
20 -
D
-Q
<D
o -
1 1 1
i i i i
1 / 1 1
1
-
V
/
- 1
o /
-
— ■ \
*Ay /
-
\
Q / /
-
" °\
-
- x w
-
* \
- o \
-
\
-
i i i
_/iii
1 1 1
1
1200
1400 1600
Temperature, °C
1800
Fig. 17. Melting curves of diopside (CaMgSi206) without volatiles (Boyd and England,
1963), with H20 (Yoder, Year Book 64, p. 87; this Report), and with C02.
GEOPHYSICAL LABORATORY
459
1700
□ 1600 -
500
CaMgSi206
3 10 15
Weight per cent C02
Fig. 18. Phase relations in the system
CaMgSi206-C02 at 20 kbar pressure. The sys-
tem is not binary, as all compositions contain
a small amount of H20. Run symbols for L
and L + V are not differentiated because
quench vapor could not be identified.
(Fig. 18). This system was the least
satisfactory of all those investigated,
both because it is not binary and because
graphite formed in charges containing
less than 10 wt % C02, in spite of run
times of only one minute. Graphite lined
most vapor bubbles and occurred only in
the center of the charge ; it may be either
a quench or a stable product. Run pro-
ducts are consistent with the geometry
shown, wherein the boundary between
Di + L and L passes between super-
solidus runs containing 3.0% H20 at
1580°C (all quench liquid) and at
1560°C (85% quench liquid + crystals).
Projection to the solidus indicates a solu-
bility of 4.0% C02 in the melt. Because
of difficulties associated with the system,
however, another method was also used.
Runs with excess C02 above the liquidus
quenched to clear glass containing large
C02 bubbles. Careful electron micro-
probe analysis of these glasses with a
CaMgSi206 glass standard yielded a
normalized oxide formula with diopside
stoichiometry but a total low by 5.8% at
20 kbar and by 7.0% at 30 kbar. These
differences are estimates of apparent C02
solubility. Because C02 may have ex-
solved from glass during the quench, the
estimates may be low. On the other
hand, as noted below, the runs may con-
tain a little H20, so that the estimates
could be high. The estimates cannot be
excessively high, since the value 4.0%
from the phase diagram at 20 kbar is a
minimum, the effect of a small amount of
H20 being to lower the melting curve.
The effect of C02 and H20 in combina-
tion was investigated with two isobaric,
isothermal sections in CaMgSi206-H20-
C02 at 20 kbar (Fig. 19A, B). These
sections are essentially ternary, as com-
positions contain only 1 mole % C02
(Eggler, this Report). One side, CaMg
Si206-H20, was determined (Fig. 20) to
define limits of the vapor-absent liquid
field. A pseudobinary section that forms
one trace across this plane was also ex-
amined (Fig. 21). In that section, the
univariant melting reaction of CaMg
Si206-H20 with excess H20 vapor (Fig.
20) becomes, with addition of C02,
divariant, and a narrow three-phase loop
appears; melting temperature is a func-
tion of vapor composition. The solution
of CaMgSi206 in vapor was determined
with a subsolidus run containing 97.5
mole % H20, 1.25% C02, and 1.25%
CaMgSi20G, which yielded half diopside
and half quench vapor. Because much of
the H20 in the run is not contained on
the quench within the quench vapor, the
vapor probably dissolved considerably
less than 1 mole % silicate material. Be-
cause the quench vapor of this run is
remarkably similar to quench diopside
liquid of other runs, no attempt was
made in subsequent runs to distinguish
those phases. For that reason, runs in
Fig. 19 were designed only to distinguish
boundaries between Di + L and L,
whether or not vapor was present.
The compositions of coexisting liquid
and vapor, in equilibrium with diopside,
presented in Figs. 18, 19, and 20, are
plotted as circles in Fig. 22, together with
two other tie lines deduced from the
geometry of Fig. 21. This diagram shows
that the 5.8 wt % (23 mole %) C02 in
a C02-bearing glass determined by elec-
460
CARNEGIE INSTITUTION
CaMgSi206
P = 20 kbar
T = I425°C
mole per cent
P=20 kbar
T = I5I0°C
mole per cent
HoO
CaMgSi206 H20
Fig. 19. Two isothermal, isobaric sections in the system CaMgSi206-C02-H20 at 20 kbar pres-
sure. Diamond symbols indicate a bracket determined in the isobaric section denoted by the
short-dashed line.
GEOPHYSICAL LABORATORY
461
1600
1500
o
o
° 1400
a>
o.
E
300
200
100
i 1 1 1 1 1 1 r
Di + L
L + V
.□_
Di +V
CaMgSi206 5
10
30
35
40
15 20 25
Weight per cent H20
Fig. 20. Phase relations in the system CaMgSi206-H20 at 20 kbar pressure.
tron microprobe must be a maximum
value because the glass also contains a
small amount of H20. Melts in equilib-
rium with vapors richer in H20 dissolve
more C02, to a maximum value of 34
mole % (9.5 wt %) at 1510°C in equilib-
rium with diopside and a vapor contain-
ing 87 mole % C02. Fluids still richer in
H20 are in equilibrium with liquids con-
taining less C02, so that at 1425 °C,
liquid containing 28 mole % (7.3 wt %)
C02 is in equilibrium with vapor con-
taining 71 mole % C02.
The effect of different melt composi-
tions on C02 solubility was investigated
with isothermal sections in Mg2Si20e-
H20-C02 and NaAlSi308-H20-C02. In
these systems quench vapor could be
recognized: in NaAlSi308-H20-C02 in
glass quenched from liquid, as bubbles
that were lined with quench material of
low refractive index; and in Mg2Si206-
C02-H20, as round, clear spheres of low
refractive index. A section in Mg2Si206-
C02-H20 appears in Fig. 23. The indi-
cated solubility of C02 in the melt at
1450°C is 15 mole % (3.7 wt %) in equi-
librium with vapor containing 62 mole %
C02, in contrast to 25 mole % C02 in the
CaMgSi20e melt in equilibrium with
vapor 'containing 62 mole % C02. An
isothermal, isobaric section in NaAlSi308-
462
CARNEGIE INSTITUTION
0.8 0.6
(C02/C02*H20)mole
0.4 02
1600
i r
1500-
1400
1300-
J L
CaMgSi206 :
I C02
0.1
— r
I — I L-V
^A Di + L + V
-*- bracket determined
in isothermal
section
H20
MOLE PERCENT
Fig. 21. Phase relations in the pseudobinary section 1 CaMgSi206:l C02-H20 at 20 kbar pressure.
CO;
P = 20 kbar
mole per cent
CaMgSi206
H?0
Fig. 22. Isobaric, polythermal diagram of CaMgSi206-C02-H20, showing the miscibility gap be-
tween the vaporous and vapor-saturated liquidus field boundaries and isothermal tie lines. Liquid
and vapor are in equilibrium with diopside.
GEOPHYSICAL LABORATORY
463
P = 20 kbar
T = I450°C
mole per cent
Mg2Si206
Fig. 23. Isobaric, isothermal section in the system Mg^Oe-ILO-CCX at 20 kbar pressure.
C02-H20 (Fig. 24) shows that the melt
dissolves even less C02, 5 mole % (0.9
wt %) in equilibrium with fluid contain-
ing 74 mole % C02.
It is evident from these data on C02
solubility in model silicate melts that the
solubility of C02 is strongly dependent
upon the composition of the melt, in con-
trast to H20, which, on a mole basis, has
nearly the same solubility in NaAlSiaOs,
CaMgSio06 and Mg2Si206 melts
(Eggler, unpublished data) . Morey
and Fleischer (1940) found from experi-
ments at pressures to 412 atm in the sys-
tem K20-Si02-C02-H20 that C02 solu-
bility is markedly dependent upon the
K20:Si02 ratio and that C02 appears to
enter the melt in fixed stoichiometry to
K20. These data indicate that much
more C02 dissolves in basic, less poly-
merized melt than in acidic, more poly-
merized melt. Low solubility of C02 in
relatively acid NaAlSi^Os melt supports
the argument, as does the observation
that more C02 dissolves in slightly hy-
drous melt than in anhydrous melt (Fig.
22). Because H20 dissolves as hydroxyl
ions, breaking Si-0 chains and producing
a less polymerized, more basic melt,
more C02 can dissolve in hydrous melt,
up to a limiting value when C02 and H20
compete for oxygens. The contrast be-
tween Mg2Si206 and CaMgSi206 melt
indicates that the cation may have some
effect on C02 solubility. In what form
C02 dissolves is not known, but the data
of Morey and Fleischer (1940) suggest
it forms a carbonate molecule, an idea
espoused by Wyllie and Tuttle (1959).
Quenched CaMgSi206 and Mg2Si206
liquids showed no evidence of carbonate.
Some NaAlSi308 glasses, however, con-
tained tiny rosettes of materials of high
birefringence.
The effect of pressure on C02 solubil-
ity has not been determined directly.
The melting curve of diopside (Fig. 17),
however, deviates considerably more
from the dry melting curve at high pres-
sure than at low pressure, implying
464
CARNEGIE INSTITUTION
P=20 kbar
T=950°C
mole per cent
NaAISi308 H20
Fig. 24. Isobaric, isothermal section in the system NaAlSi308-C02-H20 at 20 kbar pressure.
greater solubility at high pressure. Elec-
tron microprobe data on diopside glasses
suggest a moderate increase in solubility
from 5.8 to 7.0 wt % from 20 to 30 kbar.
These models show that basic melts in
the mantle may contain considerable
amounts of C02, whereas acidic melts,
such as NaAlSi3Os, contain very little.
Presumably, ultrabasic melts, such as
kimberlite (if there are kimberlite melts) ,
may have a large C02 solubility. If these
melts are in equilibrium with vapor, C02
is preferentially partitioned into vapor,
the extent of partition varying with the
composition of the fluid and with the
composition of the melt. Large differ-
ences in partitioning are evident from the
geometry of tie lines connecting liquid
and vapor compositions for albite (Fig.
24) and diopside (Fig. 19). The misci-
bility gap between vapor and liquid is
large for all compositions investigated.
If the melts are not in equilibrium with
vapor, C02 plays an interesting role. As
a simple model, consider the region en-
statite + liquid at 1450°C in Fig. 23,
and a composition containing 10 mole %
H20 (1.0 wt %) and 2 mole % C02 (0.4
wt % ) . No vapor is present. In fact, if
the composition is 60 mole % H20 (11.9
wt %) and 2 mole % C02, no vapor is
present. If the composition contains 10
mole % H20 and only 4 mole % C02
(0.9 wt %), however, the assemblage be-
comes enstatite + liquid + vapor. If
the composition is 10 mole % H20 and
16 mole % C02 (4.0 wt %), liquid is no
longer present. In summary, when C02
is present, vapor-absent conditions are
less likely, and melting does not occur
until higher temperatures are reached.
Phase Relations in the System
Mg2SiOrSi02-H20-C02
The effect of C02 on phase relations in
the system Mg2Si04-Si02-H20-C02 has
been determined at a pressure of 20 kbar.
In this system compositions of liquids in
equilibrium with forsterite (Mg2Si04)
GEOPHYSICAL LABORATORY
465
and enstatite (MgSi03) can be either
silica-oversaturated or silica-under-
saturated, depending on whether or not
enstatite melts to a liquid in equilibrium
with forsterite. Kushiro, Yoder, and
Nishikawa (1968) have demonstrated
these relations at a pressure of 10 kbar.
The critical join MgSi03-H20 has been
reexamined at 20 kbar (Fig. 25). The
subsolidus assemblage is enstatite (En)
+ forsterite (Fo) -f- vapor (V) because
the vapor dissolves Si02 in excess of a
stoichiometric MgSi03 composition. The
melting temperature of that assemblage
is 1305 °C. Liquids in equilibrium with
crystalline phases at temperatures below
1400 °C are silica-oversaturated, as the
liquidus assemblage is Fo + L. At tem-
peratures greater than 1400 °C, liquids
containing less than 13.0 wt % H20 are
exactly silica-saturated, as the liquidus
assemblage is En -f- L.
1700
1600
? 1500
1400
1300
i 1 1 1 r
P= 20 kbar
| L + V
■A Fo + L +V
J I I I I L
MgSi03 5 10 15 20 25 30
Weight per cent H20
Fig. 25. Phase relations on the join MgSiOs
H20 at 20 kbar pressure.
The effect of C02 was studied in the
join 5MgSi03:4C02-H20 (Fig. 26).
Starting materials are described in the
previous section. Three features of the
diagram are of interest. First, as in
CaMgSi206-C02-H20, the solidus tem-
perature varies continuously with vapor
composition, being higher at more C02-
rich compositions. Second, the subsolidus
assemblage changes, when the vapor con-
tains more than about 38.0 mole % C02,
from En + Fo + V to En + V, indicat-
ing that C02-rich vapor dissolves so
little excess Si02 that the presence of Fo
is not detectable optically. Third, the
temperature at which enstatite melts
congruently is 1400 °C, exactly the same
as in Mg2Si04-Si02-H20.
To clarify these phase relations, a
series of sections in Mg2Si04-Si02-C02-
H20 has been worked out. These sec-
tions, which are not shown but are avail-
able on request, are combined in Fig. 27.
In the three isothermal sections, the sub-
sections Si02-En-H20-C02 and the
vapor-rich regions are largely schematic,
but the configuration of the other parts
of the sections is reasonably accurate,
based on Figs. 23, 25, and 26 and data
of Hodges (this Report).
In the system Mg2Si04-Si02-H20
(Kushiro, Yoder, and Nishikawa, 1968;
this Report, Fig. 25) the assemblage Fo
+ En is in equilibrium with silica-over-
saturated, H20-rich melt at temperatures
below 1400°C, whereas above 1400°C
those phases are in equilibrium with
silica-undersaturated, less H_.0-rich melt.
Figure 27 shows that in the presence of
C02 the same situation holds. At 1375 °C,
the stippled tetrahedron is the four-phase
volume Fo + En + L -f V ; the vapor
contains about 45 wt % C02 and excess
Si02. The liquid, which is silica-over-
saturated, contains about 3.5% C02.
Back of that tetrahedron is the three-
phase volume Fo + En -f- L; these
vapor-absent liquids are also silica-over-
saturated. Compositions on the plane
En-H20-C02 or in the volume Fo-En-
466
CARNEGIE INSTITUTION
5MgSi03:
4C02
5MgSi03:
4C02:4H20
H,0
1600 -
P = 20 kbar
? 1500
Q-
E
1400 -
300-
1.0
En + V
Fo + L + V /j|
Fo + L +V //]"
0.8
0.6
_C05
C02-
0.4
0.2
0.0
:o2 \
i+H2°/mole fraction
Fig. 26. Phase relations on the join 5MgSi03:4C02-H20 at 20 kbar pressure.
S1O2
Si02
1450° C
1375°C
Fig. 27. Phase relations in the system Mg2Si04-Si02-C02-H20 at 20 kbar pressure. Compositions
are given as weight percentages. Stippled areas are the vaporous surface, in equilibrium with
liquid or solid phases, and the four-phase tetrahedron forsterite-enstatite-liquid-vapor.
GEOPHYSICAL LABORATORY 467
H20-COo lying to the right of the stippled (Fig. 26). At 1450°C, the subsolidus
tetrahedron do not melt at 1375°C. assemblage is En -f- V (Fig. 26), because
At 1400 °C, En and Fo are in equilib- the vapor is very C02-rich.
rium with a slightly more C02-rich vapor In summary, the model system Fo-En-
and with a liquid that now lies exactly H20-C02 indicates that in the presence
on the plane En-C02-H20 and is exactly of C02 or C02-H20 mixtures peridotite
silica-saturated. As noted above, H20- comprising the upper mantle melts at
undersaturated, vapor-absent liquid in higher temperature. Although the pres-
equilibrium with Fo + En lies on the ence of C02 has little effect on the silica-
join En-H20 at 1400°C and is exactly saturated or undersaturated nature of
silica-saturated. It is clear that the pres- liquid produced by partial melting
ence of C02, even in amounts sufficient at any particular temperature, liquids
to saturate the liquid and produce a free produced at higher temperatures
vapor phase, does not alter the silica- are less silica-saturated. The effect
saturated or undersaturated nature of the of C02 is therefore the production
melt. At 1450°C, En + Fo, both in of less silica-saturated melts. Inhomo-
equilibrium with a C02-rich vapor and in geneities in H20-C02 ratio could lead to
the vapor-absent region, are in equilib- production of basalts ranging in composi-
rium with silica-undersaturated liquids, tion from quartz tholeiitic to alkali ba-
At 1325°C, all liquids, even those in salt at the same depth in the mantle,
equilibrium with Fo or Fo -f- V, are provided appropriate temperatures were
silica-saturated. reached. Because the presence of C02
The assemblage Fo + En is a model also increases the probability of the pres-
peridotite. Figure 27 shows that the ence of a free vapor, silica-undersatu-
temperature at which that assemblage rated melts are more likely to be associ-
melts is a function of the H20-C02 con- ated with vapor, which will be C02-rich.
tent of the source region. The composi-
tion of the first-formed melt is a function Melting in a Hydrous Mantle : Phase
of temperature (and indirectly, therefore, Relations of Mantle Perid0Tite with
of H20-C02 ratio). Compositions con- Controlled Water and Oxygen
tainmg less than 55 mole % C02 will Fugacities
melt at temperatures below 1400 °C, and
first liquids are silica-oversaturated. Bj0m Mysen
Compositions containing more than 55
mole % C02 will melt above 1400°C, and Four natural peridotite nodules from
liquids are silica-undersaturated. These alkali basalt and kimberlite were melted
relationships hold whether or not the under conditions ranging from anhydrous
liquids are vapor-saturated. to water-saturated (0 < /H2o ^ /°h2o).
Figure 27 also illustrates the effect of Oxygen fugacity was controlled by the
C02 on vapor composition. At 1325°C, magnetite-hematite buffer in runs where
vapors more C02-rich than that forming /H2o < /°h2o, whereas the oxygen fugac-
one tip of the lined three-phase triangle ity was near that of the nickel-nickel
dissolve so little Si02 that tie lines con- oxide buffer in experiments with /H2o ^
necting En and V lie essentially on the /°h2o-
plane En-H20-C02. To the left of the The peridotite solidus temperature in-
triangle, however, the vapor dissolves creases isobarically with increasing
sufficient Si02 in excess of stoichiometric Mg/(Mg + Fe2+) and CaO/Al203 and
MgSi03 that the plane En-H20-C02 cuts with decreasing water fugacity in the bulk
the volume En -f- Fo + V, and the sub- rock. The depth in the oceanic mantle
solidus assemblage is En + Fo + V corresponding to the first appearance of
468
CARNEGIE INSTITUTION
garnet is sensitive to the Mg/(Mg +
Fe2+) of the rock. The continental man-
tle is probably garnet-bearing immedi-
ately beneath the Moho discontinuity.
Melts formed by partial melting of
mantle peridotite near 1000 °C under
conditions of high water fugacity are
andesitic at pressures ranging from 7.5
to at least 25 kbar. The melt becomes
silica-undersaturated at water fugacities
corresponding to a mole fraction of water
between 0.5 and 0.6 in the coexisting
vapor.
Phase Relations
The occurrence of hydrous minerals
and carbonate (Oxburgh, 1964; Dawson
and Powell, 1969; McGetchin and Silver,
1970; McGetchin and Besancon, 1973),
the composition of volcanic gases (Heald,
Naughton, and Barnes, 1963), and the
composition of fluid inclusions in olivine
from ultramafic nodules (Roedder, 1965)
strongly suggest the presence of H20 and
C02 in the mantle.
A selection of spinel and garnet lherz-
olites * (Table 2) have been partially
melted under conditions of controlled
water and oxygen fugacity. The study
has been carried out with solid-media,
high-pressure equipment at Pennsyl-
vania State University (Boettcher and
Wyllie, 1968) and at the Geophysical
Laboratory (Boyd and England, 1960)
using furnace assemblies of %-, %-, and
1-inch diameter.
The grain size of the starting material
was about 10 /mi. Run durations were
from 48 hours near the peridotite solidus
(<900°C) to 3 hours at temperatures
higher than 1200 °C. No pressure correc-
tion was applied to the emf of the
Pt/Pt90RhlO thermocouples used in the
experiments. The phases were identified
by optical and x-ray methods. Chemical
homogeneity in a given charge as well as
* The specimen numbers used throughout this
text are those used by the authors who orig-
inally described them (White, 1966; Jackson
and Wright, 1970).
TABLE 2. Chemical Composition of Starting Materials
Ga-p(l)
618-138b.l
66PAL-3
66SAL-1
Si02
Ti02
A1203
Fe203
FeO
MnO
MgO
CaO
Na20
K20
NiO
Cr203
P205
H20-
H20 +
C02
Totals
45.7
43.7
45.10
44.82
0.5
0.13
0.13
0.52
1.6
4.0
3.92
8.21
0.77
0.89
1.00
2.07
5.21
8.09
7.29
7.91
0.092
0.12
0.14
0.19
42.8
37.4
38.81
26.53
0.70
3.50
2.66
8.12
0.09
0.38
0.27
0.89
0.04
0.01
0.01
0.04
0.26
0.24
0.25
0.20
0.41
0.40
0.31
0.20
0.01
0.01
0.01
0.04
0.28
0.10
0.12
0.15
1.28
0.14
0.07
0.11
0.04
0.04
0.01
0.01
99.33
99.21
100.11
100.00
Ga-p(l): Garnet-peridotite, the Wesselton Mine. South Africa.
618-138b.l: Spinel lherzolite, Hawaii (White, 1966).
66PAL-3: Spinel lherzolite, Hawaii (Jackson and Wright, 1970).
66SAL-1: Garnet lherzolite, Hawaii (Jackson and Wright, 1970).
Analysts: Ga-p(l) and 618-138b.l, N. N. Suhr, Mineral Constitution Laboratory, The Pennsyl-
vania State University. 66PAL-3 and 66SAL-1, analyses from Jackson and Wright (1970).
All samples supplied to A. L. Boettcher and B. O. Mysen for this study.
GEOPHYSICAL LABORATORY
469
reversal of phase-field boundaries has
been used as an indicator of experimental
equilibrium. The garnet phase-field
boundaries are within ±1% kbar,
whereas other phase-field boundaries are
within ±i/2 kbar. The possibility of
overgrowth of minerals during the quench
(D. H. Green, 1973a) was checked using
the electron microprobe. Mineral grains
were traversed with a finely focused elec-
tron beam to within 1 /mi of mineral/glass
boundaries. No compositional variation
was observed.
Silver oxalate (Ag2C204) was used as
the source of gaseous C02 to dilute H20
in the vapor phase, thereby reducing the
/h2o below that of pure H20 (Boettcher,
Mysen, and Allen, 1973; Mysen, this
Report) . The phase relations of perido-
tite in the presence of essentially pure
water were explored to temperatures
above 1200°C (near the peridotite liqui-
30-
o
<d 20
en
to
0)
i-
Q_
0
ga-p(l)+ H20
'a
6I8-I38bl+H20
/a
/ /
, ^ 30
b o
_ *
V D'
^b
a> 20
•
Pressu
o
- E ♦.
\ 1
\ T
.c
*
i i
\ / j
700 900 1100
Temperature, °C
66SAL-I + H20
700 900 1100
Temperature,°C
66PAL-3+H20
i i i i n i
J I I I L
700 900 1100
Temperature, °C
700 900 1100
Temperature, °C
Fig. 28. Supersolidus phase relations of hydrous peridotite (X" — ' 1.0). a, Continental geo-
therm (Clark and Ringwood, 1964) ; b, oceanic geotherm (Ringwood, MacGregor and Boyd, Year
Book 63) ; A, ol + opx + cpx + amph + sp + vap + glass; A', ol + opx + cpx + amph + sp
+ ga + vap + glass; B, ol + opx + cpx + sp + vap + glass; C, ol + opx + sp + vap +
glass; D, ol + opx + cpx + ga + vap + glass; W, ol + opx + cpx + ga + sp + vap + glass;
E, subsolidus.
470
CARNEGIE INSTITUTION
dus) , while the study with reduced water
fugacity (H20 + C02-bearing fluids)
has been confined to conditions close to
the peridotite solidus. The preferential
solution of H20 over C02 in silicate
melts (see Eggler, this Report) becomes
an important uncertainty in the measure-
ments of the mole fraction of H20 in the
vapor (XvH2q) with larger degrees of
melting because of the limited amounts
of vapor present in the sample capsule
(about 30% of the total sample weight).
The phase relations of natural perido-
tites in the presence of essentially pure
H20 (XVH20 ^ 1.0) are shown in Fig. 28.
The phase relations are similar for all
peridotites; however, bulk rock chem-
istry and water fugacity determine the
detailed location of the phase-field
boundaries. Garnet appears at lower
pressures with lower Mg/(Mg + Fe2+)
of the starting material (Table 2, Fig.
29), because the Mg/(Mg + Fetotai)* of
garnet increases with increasing Mg/
* Fetotai : all iron as Fe2+.
(Mg + Fetotai) of the rock (Mysen and
Boettcher, 1972; Tables 2 and 3). Par-
gasite content of amphibole increases
with increasing pressure (Table 3), sug-
gesting that a decreased Ca/(Ca + Na
+ K) of the starting material may in-
crease the pressure stability of amphibole
in peridotitic mantle compositions (Table
2, Fig. 29). The generally similar P/T
coordinates of the lower pressure sta-
bility of garnet and the upper pressure
stability of amphibole (Fig. 29) indicate
a reaction relation between garnet and
amphibole at high pressures. Therefore,
the upper pressure stability limit of
amphibole may increase with increasing
Mg/(Mg + Fe2+) of the rock. (See
Table 2 and Fig. 29.)
The Mg/(Mg + Fe2+) and CaO/Al203
of the starting material appear important
for the P/T coordinates of the hydrous
(Xvn20 "^ 1.0) peridotite solidus. At con-
stant Ca/Al, the solidus temperature in-
creases with increasing Mg/(Mg -f- Fe2+)
(see samples ga-p(l), 618-138b.l, and
P-T stability
of garnet
2Px+OI + Amph
+ Sp+Vap+l_iq
100
75
50 a Q_
25
800 1000 1200 1400
Temperature, °C
P-T stability
of amphibole
-75
2Px+0l±Ga
±Sp+Vop+liq
100
E
a.
CD
50 a
25
800 1000 1200 1400
Temperature, °C
Fig. 29. Stability of garnet and amphibole as a function of pressure, temperature, and bulk-
rock chemistry, a, Continental geotherm (Clark and Ringwood, 1964) ; b, oceanic geotherm
(Ringwood, MacGregor, and Boyd, Year Book 63); A, Ga-p(l); B, 618-138b.l: C, 66PAL-3; D,
66SAL-1.
GEOPHYSICAL
LABORATORY
471
TABLE 3. Compositions of Amphiboles and Garnets *
Garnet
Garnet
Amphiboles 66SAL-1
618-138b.l
870°C
66SAL-1 -
840°C
950°C
1100°C
950°C
20 kbar
20 kbar
15 kbar
15 kbar
22 kbar
Si02
41.3
42.0
44.7
45.1
46.50
Ti02
0.93
0.26
1.27
1.35
1.09
A1203
19.46
22.8
13.4
14.1
11.2
FeOtotait
9.27
10.7
5.0
2.86
5.94
MnO
0.62
0.41
0.10
0.16
0.07
MgO
20.11
18.8
17.6
17.5
17.94
CaO
8.42
4.89
11.32
15.4
11.9
Na20
n.d.t
0.14
2.0
1.2
2.58
K20
n.d.
n.d.
0.13
0.04
0.14
NiO
0.13
0.07
0.13
n.d.
0.18
Cr203
0.56
0.36
0.27
95.92
0.14
97.85
0.26
Totals
100.83
100.43
97.78
Si
5.888
5.892
6.371
6.337
6.598
A1IV
0.112
0.108
1.629
1.663
1.402
A1VI
3.200
3.762
0.731
0.672
0.472
Ti
0.096
0.030
0.132
0.138
0.112
Fetotait
1.116
1.320
0.707
0.334
0.703
Mn
0.720
0.030
0.006
3.663
0.004
Mg
4.324
4.080
3.749
2.317
3.792
Ca
1.300
0.768
1.756
0.328
1.806
Na
n.d.
0.068
0.602
n.d.
0.707
K
n.d.
n.d.
0.017
n.d.
0.020
Ni
0.012
n.d.
0.012
0.012
0.016
Cr
0.064
0.048
0.029
0.029
Mg/(Mg+Fetotai)
0.790
0.76
0.86
0.92
0.89
Ca/(Ca+Na+K)
n.d.
n.d.
0.74
0.88
0.71
* The garnets show lower Mg/(Mg+Fe total) with lower Mg/(Mg+Fetotai) of starting material.
The amphiboles exhibit increasing Ca/(Ca+Na+K) and Mg/(Mg + Fetotai) with temperature at
constant pressure and decreasing Ca/(Ca+K+Na) with pressure at constant temperature. All
minerals were crystallized in the presence of essentially pure H20.
t Total iron as FeO (Fe2+).
X Not detected.
66SAL-1; Table 2, Fig. 30). With con-
stant Mg/(Mg -\- Fe2+), lowering of the
CaO/Al203 ratio shifts the peridotite soli-
dus to lower temperature at pressures in
excess of 10 kbar (see samples 618-138.1
and 66 PAL-3 ; Table 2, Fig. 30) . Mysen
and Kushiro (1973) suggested that the
melting of hydrous peridotite involves
the reactions :
Orthopyroxene + vapor ^±
olivine + liquid (1)
Amphibole + pyroxene (s) +
vapor <=± olivine -|- liquid. (2)
At higher temperatures, garnet may re-
place amphibole in reaction 2.
The peridotite solidi determined here
are 120° to 200 °C below those published
by Kushiro, Syono, and Akimoto (1968)
and D. H. Green (1972). The discrep-
ancy between the solidi of these authors
and those presented here (Fig. 30) may
be related to difficulties in obtaining ex-
perimental equilibrium.
Lowered water fugacity (/h2o) in-
creases the solidus temperature of perido-
tite. The temperature increases about
150°C between XvH2o ~ 1.0 and Xv H2o
~ ' 0.25. The rate of change of the solidus
temperature is ^20 °C per 0.1 XVH20 be-
tween 10 and 30 kbar for sample 618-
472
CARNEGIE INSTITUTION
o
_>£
if)
if)
CD
40
35
30
25
20
10
5-
a o} Vapor -Saturated (H2O)
-' /
7/rvjjf /
w5v T p-
co/co' 00 o
cp.co;
60'
125
100
75
Q.
0)
Q
200 1300
50
25
0
900 1000 1100
Temperature, °C
Fig. 30. Peridotite solidi (Xvn^o '—■' 1.0) as a function of bulk-rock chemistry, a, Continental
geotherm (Clark and Ringwood, 1964) ; b, oceanic geotherm (Ringwood, MacGregor, and Boyd,
Year Book 68).
138b.l (Fig. 31). The P/T coordinates
of melting reactions involving enstatite
and hydrous phases depend upon /H2o
(Eggler, this Report; Holloway, 1973),
indicating that the shift of the solidus
shown in Fig. 32 is probably related to
the effect of H20 on reactions similar to
1 and 2.
The depth in the oceanic mantle cor-
responding to the intercept between the
geotherm (Ringwood, MacGregor, and
Boyd, Year Book 63) and the peridotite
solidi determined here is essentially in-
dependent of the whole-rock chemistry of
the oceanic mantle (Fig. 30). If the low-
velocity zone (LVZ) is a zone of partial
melting, the depth to LVZ must be essen-
tially constant under oceanic regions
(with the exception of mid-oceanic ridges
and subduction zones) unless a param-
eter other than bulk-rock chemistry in-
fluences the position of the peridotite
solidus. Figure 31 shows that the perido-
tite solidus shifts with variation of /co-
variations of depth to LVZ beneath the
oceans could, therefore, be due to varia-
tion of /H2o in the oceanic upper mantle.
The depth to the first appearance of
garnet in the oceanic upper mantle may
shift substantially with bulk-rock chem-
istry because the oceanic geotherm
(Ringwood, MacGregor, and Boyd, Year
Book 63) is subparallel with and lies
inside the P/T zone that defines the in-
coming of garnet (Fig. 30). Garnet
peridotite is denser than garnet-free
peridotite; thus, small variations in
Mg/(Mg + Fe2+) of the rock can bring
about local lateral density variations in
the uppermost part (45-60 km depth) of
the oceanic mantle. Beneath continental
shields all peridotite compositions
studied would be garnetiferous at the
depth of the Moho. Thus, density varia-
GEOPHYSICAL LABORATORY
473
35 -
30-
25
- 20
10
1.0 '75/. 50/ .25i
' /' /
Solidi.6l8-l38b.H-
+ H20 + C02
900 1000 1100 1200
Temperature , °C
Fig. 31. Variation of peridotite solidi with
water fugacity expressed as Xvn2o).
tions similar to those described in an
oceanic upper mantle are unlikely in the
mantle beneath the continents.
Bulk chemical variations may play an
important role in determining the depth
at which melting occurs beneath the con-
tinents because of the steep slope of the
continental geotherm (Clark and Ring-
wood, 1964). The data obtained here
(Fig. 30) suggest a maximum difference
between the highest and the lowest pres-
sure of intercept between the continental
geotherm and the peridotite solidus of
about 15 kbar. A pressure of 15 kbar
corresponds to about 40 km. Boyd and
Nixon (this Report) suggest that the
depth corresponding to LVZ is about
160 ± 20 km beneath parts of the south-
east African continent (Lesotho). Gar-
net peridotite, ga-(l), is strongly de-
pleted in Ca, Al, Fe, and alkalies (Table
2) . Even this peridotite, however, will
melt at a depth of about 120 km beneath
a stable continental shield (Fig. 30) in
the presence of essentially pure H20.
i
1 '
1
1 1 '
1 i 1
0.9
A
-- /ON
(^AA
_£_
/ 9
/ v
0 o /
-
Li-
0.8
/
o o'
—
n-
/
/
en
/
/
^
/ o /
0.7
v9 '
o Glass
A Olivine
-
0.6
i
618
-I38b.l+H20
1 l 1 i
+ Amphibole
1 i I
—
800 900 1000 MOO
Temperature, °C
1200
Fig. 32. Compositional variations of glass, olivine, and amphibole with variations in tempera-
ture (degree of melting) .
474
CARNEGIE INSTITUTION
Lowered water fugacity may shift the
solidus to greater depth in the continental
mantle. Variations of depth to LVZ and
even its total absence, as observed under
some cratonic areas, are possibly due to
variations of the water fugacity in the
mantle. The data in Fig. 31 suggest that
^h2o < 0-5 may be sufficient to shift the
solidus temperatures above the conti-
nental shield geotherm (Fig. 31). It ap-
pears, therefore, that experiments with
controlled /H2o, in addition to whole-rock
chemical variations, can account for
some of the enigmatic features left un-
resolved by previous data.
Chemical Relations
Electron microprobe analyses were
made of the experimental products in the
peridotite systems reported above, using
the methods of Finger and Hadidiacos
(Year Book 71). Low beam-current
('^0.015 /xamp) and a defocused electron
beam (>10 /mi) have been employed to
reduce the loss of Na by volatilization.
The individual phases display continu-
ous chemical variations with temperature
or pressure, or both, suggesting that re-
actions occurring during melting of pe-
ridotite are spread out over extensive
temperatures and pressures (Fig. 28).
The Mg/(Mg + Fe^tai) of all phases
increases with temperature (Fig. 32), al-
though the ratio increases at different
rates in different phases. The Mg/(Mg
+ Fetotai) of liquid increases faster than
does the ratio of the coexisting minerals
(Fig. 32). The difference of rate would
be expected because a greater fraction
of the iron-rich end members of the min-
TABLE 4. Olivine Composition as a Function of Composition
of Starting Material *
Ga-p(l)
66PAL-3
66SAL-1
1000°C
1000°C
1000°C
15 kbar
15 kbar
15 kbar
Si02
40.31
40.9
41.71
Ti02
n.d.f
n.d.
n.d.
A1203
0.07
n.d.
0.07
FeOtotait
3.47
8.4
9.6
MnO
0.15
0.13
0.13
MgO
54.05
49.6
49.9
CaO
0.07
0.04
0.04
Na20
n.d.
n.d.
n.d.
K20
n.d.
n.d.
n.d.
NiO
0.06
0.34
0.36
Cr203
0.01
n.d.
n.d.
Totals
98.91
99.42
101.68
Si
0.980
1.002
1.002
Al
0.001
n.d.
0.001
Ti
n.d.
n.d.
n.d.
Fetotait
0.070
0.173
0.193
Mn
0.003
0.002
0.002
Mg
1.960
1.812
1.787
Ca
0.001
0.001
n.d.
Na
n.d.
n.d.
n.d.
K
n.d.
n.d.
n.d.
Ni
0.001
0.007
0.006
Cr
n.d.
n.d.
n.d.
Mg/(Mg + Fetotai)
0.93
0.91
0.90
* All olivines coexist with liquid and H20-rich vapor (Xvh2o
t Not detected.
% Total iron as FeO (Fe2+).
4.0).
GEOPHYSICAL LABORATORY
475
erals contributes to the melt at low tem-
peratures. There is also a positive cor-
relation between the Mg/ (Mg + Fetotai)
of the phases present and the Mg/(Mg
+ Fe2+) of the starting material (Tables
2 and 4). Therefore, the Mg/(Mg +
Fetotai) of liquid formed by partial melt-
ing of mantle peridotite can be critically
dependent on the chemical composition
of the parent of the liquid.
The Ca/(Ca + Mg + Fetotai) of
orthopyroxene coexisting with clinopy-
roxene also increases systematically with
temperature (Fig. 33), following the pat-
tern suggested by Boyd and Nixon (this
Report) . However, the geothermometer
suggested in Fig. 33 gives temperatures
100 °C lower than those suggested by
Boyd and Nixon (this Report) . From
studies of peridotitic compositions, Hen-
sen (this Report) obtained a temperature
scale that gives temperatures slightly
lower than those in Fig. 33.
The thermometers described by Boyd
and Nixon (this Report) and Hensen
(this Report) do not include iron,
whereas iron is included in Fig. 33. This
1300-
1200
¥ MOO
CD
| 1000
CD
Cl
£> 900
800
700
i — i — r
— i — i — i r
Orthopyroxene
j I L
i l i I i U
9 13 17 21 25 29
[Ca/(Ca+Mg + Fe)]xlOOO
Fig. 33. Orthopyroxene thermometry: Ca/
(Ca + Mg + Fetotai )oPx versus T. The pres-
sures are between 10 and 30 kbar. No clear
pressure dependence is observed, and all pres-
sures are included in the figure.
factor may explain the discrepancy be-
tween the curve presented here and that
of Hensen (this Report). Boyd and
Nixon (this Report) used the data of
Davis and Boyd (1966), obtained on the
system CaO-MgO-Si02. Both chemistry
and assemblage of the system studied by
Davis and Boyd (1966) are different
from those of the natural peridotite sys-
tems.
All liquids analyzed are silica- and
alumina-rich (at XVU20 > 0.6) and re-
semble andesite (see My sen and Kushiro,
1973; and Table 5) near 1000°C within
the pressure interval where analyses were
obtained (7.5-25 kbar). This tempera-
ture is comparable to that of an andesite
liquidus (Eggler, 19726; Allen et al.,
1972) in the presence of essentially pure
water. Mysen and Kushiro (1973) ob-
served that liquid formed by wet (XvH2o
"- ' 1.0) partial melting of sample 618-
138b.l at 940°C and 15 kbar (Table 5)
has its wet (XVU20 — 1.0) liquidus close
to 950 °C at 15 kbar, in further support
of a suggestion by O'Hara (1965) that
andesite can form by direct partial melt-
ing of mantle peridotite. The potassium
content of all liquids is low, probably
because of the low potassium content of
the starting materials (Table 2) . Experi-
ments on the melting behavior of phlogo-
pite (Yoder and Kushiro, 1969; Modre-
ski, 1972) , a possible carrier of potassium
in the mantle, indicate that potassium
could be bound in this mineral to tem-
peratures above those cited for the gen-
eration of andesitic liquids. Potassium
will partition to phlogopite rather than a
coexisting andesitic liquid. Thus, ande-
site formed by partial melting of mantle
peridotite would be restricted to low-K
andesite. Studies on the melting behavior
of phlogopite-bearing peridotite are ob-
viously needed. Nodule 66PAL-3
(Table 2) was subjected to a few experi-
ments with 3.6 and 10 wt % synthetic
phlogopite (Modreski, 1972) added. The
potassium content of the liquids formed
increases to ~0.6 wt % K20 when the
476
CARNEGIE INSTITUTION
TABLE 5. Composition of Liquids Formed by Partial Melting of Mantle Peridotite with
Variable Pressure, Temperature, Water Fugacity (Expressed at 1^)
and Composition of Starting Material
1
2
3
4
5
6
7
Si02
67.8
68.9
62.2
58.8
39.9
60.1
64.3
Ti02
0.6
0.2
0.8
0.6
1.0
0.6
0.1
AI2O3
16.1
19.6
20.1
18.0
18.9
18.1
22.3
FeOtotai*
1.1
1.1
1.7
4.8
6.2
1.7
0.6
MnO
0.1
n.d.
0.2
0.2
0.2
0.1
0.1
MgO
0.6
0.8
2.1
4.6
4.7
3.8
0.7
CaO
10.2
6.1
10.4
9.5
21.3
14.5
10.5
Na20
3.1t
1.8J
2.1t
3.5t
7.4t
Lit
0.7J
K20
0.3t
1.4J
0.2t
O.lt
0.2f
O.lt
0.6t
Totals
99.9
99.8
99.8
100.1
99.8
99.9
99.9
MgO/
(MgO+FeOtotai)
0.49
0.58
0.69
0.63
0.57
0.80
0.68
* Total iron as FeO.
t Analytical value of alkalies with no correction for volatilization loss of Na and K.
X Alkalies estimated from mass balance calculations.
1. Spinel lherzolite, 1190°C and 26 kbar (Kushiro et al, 1972).
2. Pyrolite, 1100°C and 20 kbar (D. H. Green, 1973a).
3. 618-138b.l, 940°C and 15 kbar, X"h2o~1.0.
4. 618-138b.l, 1100°C and 10 kbar, X"h2o ~ 0.6.
5. 618-138b.l, 1120°C and 15 kbar, X»h2o ~ 0.5.
6. 66PAL-3, 1120°C and 15 kbar, X»h2o ~ 1.0.
7. 66PAL-3 + 10 wt% phlogopite, 1025°C and 15 kbar, X"h2o ~ 1.0.
potassium content of the starting mate-
rial is increased up to 1 wt % K20.
Phlogopite is observed at 1150°C and 15
kbar. The liquids formed by wet partial
melting of nodule 66PAL-3 + phlogopite
display lowered MgO and FeO contents,
and Si02 and A1203 show some increase
compared with experiments without
phlogopite added to the starting material.
Optical examination of the run products
suggests that quench mineral (s) (amphi-
bole?) had crystallized. The composi-
tions of the glasses formed by wet (XvH2o
'— ' 1.0) partial melting of sample PAL-3
+ phlogopite resemble those published
by Kushiro et al. (1972) and D. H. Green
(1973a), supporting the suggestion by
D. H. Green (1973a) and Mysen and
Kushiro (1973) that the formation of
quench materials has altered the liquid
composition reported by them (Kushiro
et al., 1972; D. H. Green, 1973a). The
starting materials used by D. H. Green
(1973a) and Kushiro et al. (1972) were
rather potassic (0.13 and 0.15 wt %, re-
spectively) compared with the composi-
tion of the original starting materials
used here (Table 2). The discrepancies
in liquid composition between those
formed by partial melting of peridotite
with low potassium contents (Table 5)
and those of D. H. Green (1973a) and
Kushiro et al. (1972) may therefore be
explained by formation of quench min-
erals from their liquids.
The liquids formed by partial melting
of spinel lherzolite, 618-138b.l (Table 2),
with /H2o < /°h2o (XvH2o < 0.6) are
highly alkaline and silica-undersaturated
at pressures greater than 10 kbar (Table
5) . Similar chemical characters are ob-
served at XvH2o ^ 0.25. Eggler (this
Report) has shown that enstatite +
forsterite coexist with a quartz-norma-
tive liquid at XvH2o ^ 0.6 at 20 kbar
total pressure, whereas the liquid coexist-
ing with enstatite + forsterite is olivine-
normative at Xvh2o ^ 0.5. Liquids
formed by partial melting of mantle
peridotite are quartz -normative at XVU20
GEOPHYSICAL LABORATORY
477
> 0.60 and olivine-normative at XVH20
< 0.50 (Table 5). The remarkable coin-
cidence of melting behavior of mantle
peridotite and enstatite + forsterite
strongly suggests that the change in
liquid composition shown in Table 5 is
due to the effect discussed by Eggler (this
Report) .
The liquids formed with XVH20 < 0.50
have maintained or even increased their
aluminum, alkali, and calcium contents
(Table 5). The variation of these three
elements compared with experimental
results obtained with conditions of higher
water fugacity (XvH2o > 0.6) may be ex-
plained by the small degree of melting
(5-10%) at these relatively high tem-
peratures (1100°C). The lower Mg/
(Mg + Fetotai) may be related to the
same effect. The higher oxygen fugacity
used in experiments with H20 + C02
bearing vapors (/h2 = /h2 [MH]) com-
pared with /o2 in experiments with essen-
tially pure H20 (/o2 [NNO]) may also
explain this chemical feature (Mysen
and Kushiro, 1973).
Yoder (1969) and Wyllie (1971) dis-
cussed mechanisms of generating ande-
sitic liquids on and above subduction
zones beneath island arcs. The data dis-
cussed in this Report (Table 5) strongly
support the suggestion that andesite can
indeed form by partial melting of mantle
peridodite above subduction zones.
The data obtained on the melting be-
havior of mantle peridotite with con-
trolled water fugacity suggest an alterna-
tive hypothesis for the genesis of kimber-
lite and associated silica-undersaturated
alkaline rocks (see O'Hara and Yoder,
1967, and Dawson, 1971, for review and
discussion of previous models). The
points to be considered in this model are:
1. The peridotite solidus temperature
increases with decreasing water fugac-
ity (Fig. 31).
2. Although the liquids formed by par-
tial melting of peridotite are quartz-
normative at Xvh2o ^ 0.6, they become
strongly alkaline and silica-undersatu-
rated with XVH20 < 0.5. Under these con-
ditions the liquids resemble melilite
nephelinite, although the potassium con-
tent is too low and the calcium and
aluminum contents are high. The potas-
sium deficiency can be alleviated by
using phlogopite-bearing peridotite as
starting material. Further, the chemical
data in Table 5 were obtained on liquids
formed at rather low pressure (Table 5)
coexisting with Al-poor orthopyroxene,
olivine, and small amounts of Al-poor
clinopyroxene. Garnet and Tschermak's
component in clinopyroxene would be
stable at higher pressures and could take
up substantial amounts of Ca and Al,
thus eliminating the problem with these
two elements.
3. The presence of a separate multi-
component fluid in the mantle was sug-
gested by H. W. Green (1972). The
water fugacity of this fluid is defined by
its composition, temperature, and pres-
sure. The water fugacity is independent
of other reactions occurring in the sys-
tem.
The mechanism proposed for the gene-
sis of kimberlite and associated rocks,
based on the above points, is as follows.
Decreasing /H2o increases the peridotite
solidus temperature, which may, at some
intermediate water fugacity (O < XVH20
< 0.5) , exceed that required for the sta-
bilization of phlogopite, and phlogopite
breaks down at a temperature below that
of the peridotite solidus. When phlogo-
pite breaks down, water will enter the
mantle fluid without much change of
/h2o, but the peridotite does not melt.
The activity of potassium and to a
smaller extent, titanium, increases at this
stage. This activity determines whether
these elements dissolve in the mantle
fluid or enter other crystalline phases in
the peridotite, but no such phase is
known at present. When the solidus tem-
perature is reached, the melt composition
is similar to that shown in Table 5 and
strongly resembles the composition of
kimberlite and associated rocks. Re-
478
CARNEGIE INSTITUTION
gardless of whether K and Ti are located
in the fluid or a crystalline phase, the two
elements will partition strongly into the
silicate liquid formed when the solidus
temperature is reached. Although the
data obtained by partial melting of
mantle peridotite with water fugacities
below that of pure water display a trend
toward kimberlite and associated rocks,
the detailed water fugacity, temperature,
and pressure needed for the model pro-
posed have not yet been ascertained.
Bridget Cove Volcanics, Juneau Area,
Alaska: Possible Parental Magma of
Alaskan -Type Ultramafic Complexes
T.N. Irvine
Among ultramafic rocks, the concen-
trically zoned and lithologically similar
complexes of the Alaskan type present a
distinctive combination of features in
terms of setting, structure, petrography,
and geochemistry. It is well established
that they have formed essentially by the
accumulation of mafic minerals settled
from magma, but the nature of the
magma has been enigmatic. In this in-
vestigation an attempt has been made to
find lava that might be representative of
the magma in a volcanic belt associated
with the complexes in southeastern
Alaska. Although the results are not
definitive, the rocks studied show enough
of the features essential to the magma
to place certain constraints on its com-
position.
Approach
The characteristics of Alaskan-type
ultramafic bodies have been described
in detail by Ruckmick and Noble (1959) ,
Irvine (1959, 1967, 1973), Taylor and
Noble (1960, 1969) , Taylor (1967) , Find-
lay (1969), and others. The problem of
establishing the composition of their
parental magma arises because none of
the complexes shows a border facies that
can be convincingly interpreted as a
chilled margin. Postulated compositions
have been inferred from the gross com-
positions of the bodies and from struc-
tural and tectonic considerations, and
have variously ranged from a series of
ultramafic magmas to a tholeiitic ba-
saltic precursor of calc-alkaline andesite.
Several years ago, the writer noted
that the Duke Island complex in south-
eastern Alaska, manifesting a history of
repeated intrusion and disturbed crys-
tallization, might represent a subvolcanic
magma reservoir in which the magma
precipitated mafic minerals while under
temporary storage (Irvine, 1967) . In pur-
suing this notion it was found that the
main line of ultramafic bodies in south-
eastern Alaska is essentially coincident
with a discontinuous volcanic belt (Fig.
34), a feature also noted by Murray
(1972) and Berg, Jones, and Richter
(1972). The volcanic rocks had been
only tentatively assigned to the Late
Jurassic or Early Cretaceous, but pale-
ontological evidence now indicates that
they are, at least in part, late Early
Cretaceous (H. C. Berg, personal com-
munication, 1971) and thus roughly
equivalent in age to the ultramafic bodies,
which have been dated radiometrically
at 100-110 m.y. (Lanphere and Eberlain,
1966). It appeared worthwhile therefore
to investigate the possibility that the
parental magma of the ultramafic rocks
or its derivative might be represented
among the volcanic rocks.
The part of the belt that has been
studied is just northwest of Juneau. It
was selected because Knopf (1912) had
emphasized that the rocks are extremely
rich in augite; he called them augite
melaphyres and commented that they
must represent an unusually pyroxenic
magma. This observation was consid-
ered important because augite-rich cum-
ulates form sections 5000-7000 feet thick
in the Duke Island complex. Knopf also
described pillow structures, which guar-
anteed the opportunity of sampling lava
rather than just fragmental rocks such as
dominate the belt as a whole.
GEOPHYSICAL LABORATORY
479
Fig. 34. Map of southeastern Alaska showing the location of the Bridget Cove volcanics in re-
lation to other Jurassic-Cretaceous volcanics and to Alaskan-type ultramafic bodies.
Field Characteristics of the about 12 miles northwest of Juneau. The
Volcanic Rocks thickness of the unit is unknown, but it
is generally about 1V2 miles wide with
The rocks strike north-northwest along steeply dipping bedding. All observed
the east shore of Lynn Canal for a dis- pillow structures faced west, hence a sec-
tance of 24 miles, starting from a point tion of the order of 5000 feet is inferred.
480
CARNEGIE INSTITUTION
The rocks, which had previously been
unnamed, are conveniently termed the
Bridget Cove volcanics after a locality
where they are well exposed. Knopf
(1912) and others have noted that they
are probably correlative with a unit
called the Douglas Island volcanics oc-
curring southwest of Juneau (Lathram
et al., 1965), but as they are separate the
use of a different name was considered
advisable. Regionally, they are part of
a narrow strip of Middle (?) Jurassic to
late Early Cretaceous sedimentary and
volcanic rocks, called the Gravina-
Nutzotin belt, that extends for 700 miles
along southeastern Alaska into southern
Alaska. This belt has been interpreted as
a magmatic arc associated with an east-
erly dipping Mesozoic subduction zone
(Berg et al, 1972).
The Bridget Cove volcanics comprise
lavas, explosion breccias, tuffs, and mixed
volcaniclastic sediments with the pyro-
clastic and clastic rocks predominating.
They are basaltic in appearance and are
characterized by abundant augite, prin-
cipally as blocky, lustrous black pheno-
crysts. Feldspar phenocrysts, by con-
trast, are almost totally lacking ; in fact,
they were found in only one small out-
crop of breccia.
The rocks are everywhere chloritic
owing to metamorphism, but apart from
the filling of amygdules and local small
veins and patches of calcite, they show
little field evidence of chemical change
beyond the addition of H20 and C02.
Nor do they show any pronounced meta-
morphic fabric; most deformation has
evidently occurred by way of numerous
small faults.
Microscopic Characteristics
Almost all the rocks, whether lava or
pyroclastic, have phenocrysts of augite
and magnetite, and chlorite pseudo-
morphs after what were almost certainly
olivine phenocrysts. The augite pheno-
crysts generally amount to 25-40 modal
% ; the others are subordinate (Fig. 35) .
The matrix consists of chlorite and K-
feldspar variously combined with augite,
in w
*- 4
E 2
-
45A
41
^Specimen No.
22
I2C
I8E
I8D
5
9
I4B
I2A I8A
I2F
19
l
10 20 30 40 50
Augite phenocrysts, modal per cent
O Porphyritic rock
O Rock with skeletal
augite microphenocrysts
60
Typical modal range of
olivine clinopyroxenite
Olivine pseudomorphs
Typical modal range of
hornblende-magnetite
clinopyroxenite
80
90
Augite
Modal per cent
Fig. 35. Data on the modal abundances of phenocryst minerals in the Bridget Cove volcanics.
The typical ranges of the main pyroxenite units in Alaskan-type ultramafic bodies are shown for
comparison.
GEOPHYSICAL LABORATORY
481
**^^E':i ♦
i
*
■*.
mm
*%
4
Fig. 36. Photomicrograph of sample IB72-19, showing augite quench crystals. Reflected light.
magnetite, albite, tiny needles of apatite,
some epidote, and traces of chalcopyrite.
The amygdules are mainly filled with
calcite and chlorite, but some are tiny
pockets of K-feldspar or, less commonly,
albite. No unaltered intermediate to
calcic plagioclase has been found. The
feldspar phenocrysts in the one outcrop
in which they were observed are com-
pletely replaced by fine-grained albite (?)
and minor epidote.
The augite phenocrysts are generally
0.5^1 mm in length with the odd one as
long as 2 cm. They are characteristically
euhedral, and zoning is a principal fea-
ture. Most commonly they have colorless
or slightly yellowish cores bordered by
pale, grayish green rims. Less commonly
the zoning is oscillatory, and the occa-
sional crystal shows faint sector zoning.
Many of the phenocrysts enclose small
irregular pockets of chlorite that are ob-
viously altered from glass or crystalline
material formed from trapped magma.
In one exceptional rock from one of a
cluster of small pillows, most of the
augite occurs as skeletal micropheno-
crysts (Fig. 36) that are commonly in
cruciform growths and closely resemble
diopside crystals quenched from synthetic
melts. The rock is regarded as the best
sample obtained of lava that was rela-
tively free of suspended augite and chem-
ically is one of the more basic rocks.
The chlorite pseudomorphs that are
believed to be derived from olivine
phenocrysts are 0.5-1 mm in length.
Where best developed they are clustered
with augite phenocrysts and show
domatic forms typical of olivine. The
possibility that they represent Ca-poor
pyroxene seems ruled out by the com-
positions of the rocks and augite pheno-
crysts (see below) .
The magnetite phenocrysts are mostly
scattered through the matrix but can
generally also be found as inclusions in
the augite and olivine. It is apparent
that the three minerals overlapped con-
siderably in their crystallization periods.
Small euhedral chromite grains were
found in one sample of tuff.
Although the matrix of the lavas is
extensively altered, its fabric commonly
482
CARNEGIE INSTITUTION
reflects primary textures. Augite and
magnetite are least affected by the alter-
ation, and in some rocks the augite crys-
tallites show flow alignment around the
phenocrysts. The K-feldspar in places
shows lath-like forms and is even seen as
small phenocrysts. The chlorite appears
to have replaced glass in some rocks, and
a mafic mineral (again most probably
olivine) and plagioclase in others, with
albite, calcite, and epidote as by-prod-
ucts.
Mineral Chemistry
Electron microprobe analyses of the
augite in the volcanic rocks are sum-
marized in Figs. 37, 38, and 39. The min-
Si02
wt%
Al203
wt %
Mg
Mg+Fe
at %
Ca
Ca + Mg+Fe 47_
at %
Na20
wt %
Ti02
wt %
Cr2
°3
0.6
wt
%
0.4
0.2
0
1000 1500 2000 2500
Traverse distance, ju m
3000
Fig. 37. Compositional variations revealed by electron microprobe analyses along two traverses
across a large, zoned augite phenocryst in the Bridget Cove volcanics, sample IB72-14B.
GEOPHYSICAL LABORATORY
483
BRIDGET COVE VOLCANICS
-* -k CaFe
Shiant Isles trend
Skaergaard trend
Atomic per cent
ALASKAN -TYPE ULTRAMAFIC COMPLEXES
+
* 4- CaFe
Shiant Isles trend
^Skaerqaar
gaard trend
+ Chemical analysis • Electron microprobe analysis
Fig. 38. Ca-Mg-Fe ternary plots comparing augites of the Bridget Cove volcanics with those of
Alaskan-type ultramafic bodies. Chemical data from Ruckmick and Noble (1959), Irvine (1959
and unpublished data), and Findlay (1969). Skaergaard (tholeiitic) clinopyroxene trend from
Brown (1957); Shiant Isles (alkaline basaltic) clinopyroxene trend from Gibb (1971). Microprobe
data from 13 samples of volcanic rocks, four of ultramafic rocks.
eral generally ranges from Ca47Mg48Fe5
to Ca47Mg38Fei5. The high-Ca content
compares with that of clinopyroxenes
crystallized from alkaline basalt mag-
mas, as in the Shiant Isles sills (Fig. 38),
and is believed to preclude the possibility
that the chloritized phenocrysts were Ca-
poor pyroxene.
Along with the increase in Fe/Mg
ratio, A1203, Ti02, and Na20 increase
and Cr203 decreases. Even at their high-
est, however, Ti02 and Na20 are low
compared with their common levels in
alkaline basalt pyroxenes, and the augite
is far from being either titanaugite or ap-
preciably acmitic.
Considering briefly the origin of the
compositional variations, it is noted that
most of the variations are from core to
rim in the phenocrysts and from pheno-
cryst cores to matrix grains. There is
little difference between the ranges in
different samples. The change from core
to rim in the phenocryst in Fig. 37, for
example, spans almost the entire range
found in all samples. Moreover, the ratio
(Mg/Fe)corG/(Mg/Fe)rim is about 3.4,
roughly the value that one would expect
for a distribution coefficient, KD =
(Mg/Fe) au^ite/ (Mg/Fe) melt, describing
equilibrium between the cores and the
magmatic liquid. It is suggested there-
fore that the more iron-rich pyroxene was
essentially formed by quenching of the
magma on eruption. Presumably it grew
so rapidly that it took up much more
A1203, Ti02, and Na20 than it would
under equilibrium conditions, whereas
Cr203 decreased because it had already
been depleted in the melt through the
crystallization of the core augite and
spinel minerals.
In the magnetite phenocrysts, A1203
typically ranges from 0.5 to 8.5%, Ti02
from 0.5 to 7.5%, Cr203 from trace
amounts to 10%, and MgO from traces to
484
CARNEGIE INSTITUTION
10
8
Al203 6
wt % 4
2
n
i 1 1 1
BRIDGET COVE VOLCANICS
— • —
•
••• .
— •• • -
i i i i
Ti02 l0
Wt%
0.5
0
04
Na20
Wt %
0.2
o
•
••
^« •• • •
n i i i
•
•
1 1 1 1
0.6
Cr203
Wt % 0.4
0.2
0
•
?
- >:v.
l5Bk.ii :
I ' ' ' ' I
ALASKAN -TYPE ULTRAMAFIC COMPLEXES
100 90
80
70
+ +
+
+
• +
t.
+
.+ +
*v
+
• +
i
+ +
+
1 .,Ao^«-+-l_
60 50 100 90 80 70
60 50
100 Mg/(Mg-hFe). atomic %
Fig. 39. Comparison of the variations in A1203, Ti02, Na20, Cr203, and Mg/(Mg + Fe) in au-
gites of the Bridget Cove volcanics with those in clinopyroxenes from Alaskan-type ultramafic
bodies. Symbols and data sources as in Fig. 38.
5%. The phenocrysts showed some zon-
ing, but it is less pronounced and less
regular than that in augite. Compositional
correlations between the augite and mag-
netite are evident but difficult to define
because of the zoning. Analyses of the
only chromite grains that were found are
given in Table 6.
Rock Chemistry
Analyses of 17 rocks, including data
on all major constituents and 15 trace
elements, have been obtained through the
courtesy of the Geological Survey of
Canada. Data on the major oxide com-
positions of seven samples are given in
Table 7. Approximate analyses have
also been made of the matrix of ten
rocks with the electron microprobe, by
using a broad beam (60-90 /xm) and
analyzing 4 to 10 spots selected at ran-
dom over the area of a polished thin sec-
tion. Four of these analyses appear in
Table 7.
The whole-rock analyses show the vol-
canic rocks to be alkaline, and in view
of the abundance of clinopyroxene and
apparent presence of olivine, the rocks
are properly called ankaramites. All but
one of the 17 analyses show Ne in the
norm, and two even have Lc. Compared
GEOPHYSICAL LABORATORY
485
TABLE 6. Analyses of Chromites in the
Bridget Cove Volcanics and in Two
Alaskan-Type Ultramafic Com-
plexes, wt %
Bridget Cove
Alaskan-Type
Volcanics
Comp
3t
lexes
1*
2*
4t
Si02
0.3
0.4
0.61
0.38
Ti02
0.5
0.8
0.58
0.71
A1203
9.4
8.9
6.9
11.5
Cr203
46.5
42.1
43.9
46.5
Fe203
17.1
20.5
20.8
10.7
FeO
12.3
16.3
18.5
19.1
MgO
14.2
11.8
8.8
10.3
CaO
0.0
0.0
0.05
0.03
MnO
0.2
0.2
0.40
0.16
NiO
0.1
0.1
0.06
Totals
100.6
101.1
100.6
99.38
* Automated electron microprobe analyses,
Fe203 estimated from structural formula.
t Conventional chemical analyses, Analytical
Chemistry Section, Geological Survey of Canada.
1. Sample IB72-6, Bridget Cove volcanics,
one grain.
2. Sample IB72-6, Bridget Cove volcanics,
average of two grains.
3. Sample FJT60-55A, Tulameen complex,
British Columbia.
4. Sample IB65-30, Polaris ultramafic com-
plex, British Columbia.
with common basalts, A1203 is low and
CaO is high. Alkalies are also high, and
in 13 of the 17 analyses K20 exceeds
Na20. But Ti02 at less than 1% and
P205 at 0.2-0.5% are remarkably low
for alkaline rocks. As one would expect,
the norms are high in Cpx, showing 23—
40%, and even the rock with quench
crystals has 30%.
The matrix compositions are remark-
ably potassic, several of them showing
7-10% KoO. Their norms indicate that
the differentiation associated with frac-
tionation of the phenocrysts was toward
a trachytic residual liquid.
Comparison with Alaskan-Type
Ultramafic Rocks
There is, of course, no way of proving
that the volcanic rocks represent the
parental magma of the ultramafic com-
plexes— apart, perhaps, from tracing
them directly into one of the complexes.
The circumstantial evidence is consider-
able, however, and valid or not, it pro-
vides certain definite indications as to the
nature of the magma.
There are many requirements that the
magma must satisfy, but essentially it
must be able to crystallize the same min-
erals that occur in the ultramafic rocks,
in the same order and with the same com-
positions. The volcanic rocks will be
examined in this regard, mainly through
their phenocrysts, allowing for differences
in cooling rates and other probable fea-
tures of contrast between the volcanic
and plutonic environments. The general
order of crystallization in the ultramafic
rocks is olivine + minor chromite, oli-
vine + clinopyroxene, clinopyroxene +
magnetite, hornblende. Fractionation ac-
cording to this order has yielded the
lithologic association, dunite and wehrlite,
olivine clinopyroxenite, hornblende-mag-
netite clinopyroxenite, and hornblendite.
The rocks are virtually devoid of Ca-
poor pyroxene and plagioclase, and few
of the bodies are associated with gabbroic
rocks that can be assigned to the same
differentiation series. Hornblende is gen-
erally abundant, demonstrating the pres-
ence of abundant water in the magma.
Olivine-rich lava capable of precipitat-
ing dunite or wehrlite has not been found
among the volcanic rocks, so even if the
tack is right, a sample of the most primi-
tive magma necessary to form the ultra-
mafic bodies has not been recovered. The
closest approach may be represented by
the tuff containing the chromite grains,
as these are compositionally similar to
chromite from peridotite in the ultra-
mafic bodies (Table 6).
The volcanic rocks, however, contain
phenocrysts of all the cumulus minerals
in olivine clinopyroxenite and horn-
blende-magnetite clinopyroxenite — and
in about the right proportions (Fig. 35).
The olivine and magnetite have crystal-
lized concurrently rather than in succes-
sion as in the ultramafic rocks, but this
feature can reasonably be attributed to
486
CARNEGIE INSTITUTION
CU
nd
-p
«+H
o
in
#
w
■+=
>}
£
"3
^
d
«1
O
*d
0)
i?3
,£
O
O
"o
a>
o
OJ
o
>
§
o
O
d -^
K
d -^
oa a
W 5
Z? 02
C3 ^
d >
a3 ^,-1
o o
'a. *
<U Si
-d 12
04
o -I-3
O
o
b
o
O
d
CD
-d
PQ
H
so
CO
CM
M
o
o
CO
iO
TtH
CO
<N
q i>
CO CO
lO rH
00 iO
00 CO
00 CO
T* H
CO H
CO l>
io i— I
OCOCO-^^CqCTirHOp
swcoiNNodddd
tOM^NNWOOHOO
OOiO^'^'cOOOOOCJ
copLqcMpiopcMpp
oooM^coOHodd
q^H^tooo^oinqo
^Hfirlddddod
CO iO iO O O © O iO
q <q q « in ^ q q oo ■* q q q <o n
oioo^^^'ooco^ddddoNd
">* ,-H
iO iO 00 CO i— I CO "* CO
icoocoqHNwqoo^wqqHH
ffl^picoNoi^HoddddiNd
TjH ,-H
ONccqqNHNcoNiNqqNH
NNiNoddddc<idddddc6d
rJH t— i H
N CO OOr-lO^OO
■*qqi>ooqHiflNTH(Nqq^q
aJNNNffloiiNMddddoNd
i— i 00 ION H 1O00 H
pNrHrHpCMCMCON^T-HppNp
NHcooodHNddddddwd
TH T-H T-H
CO O
00 lOOHtO^iO
(NO00^NNTtH(Nqq>ON
CM* rH©ddddcM©
h tHNOh
HO! 00 OS CO © CO 00
oiioqqoiN^wNiNHqHiNTt;
NOCOOOHC>C<ii-HOOC>OOCOO
T^ 1— I T-H 1— I
■■■ ~> m9 3 O 6 o 9 O Q O jt
o 3 -
CO
CO T-H CO o •
LO iO "* iO
• CO 00
• CM 00
•NH05
■qqqqoo
id
OS
CD id CM <N
T* <N rH
'*£
© © rH OS lO
TjH CO
OS
<N
OS
CM
OS
iO
OS
00
©
o
CO
CO
©
o
c3
-t-3
o
H
CM CO © OS rH
n q w n q
cd os cd id d
IN NH
00 rH O OS H
iO ""* CM TtJ j-^
t-H* CM* "^ rH CO
CM CO rH
iO
CM
hHM
(, q oi n h
(-5 O rH cd id
to CO
to -co • iO
y-^ ■ t>- ■ CM "* CM
<*" ' d ' <n d id
CM iO CM
^h CO (M CO
IO TjH CM CO
d n d d
CO rH
rH CO
d J^
iq
d
00 OS CO
h'oio
CO CO
00 "HH CO 00 LO
HNOO00
Tji d cd ' tjh d
CM CM <N rH
(N O rt< O CO
l> CO rH O OS TtJ l>
d id i-^ d d <n" id
00 iO
CM
• o
CM
CO
CO
1— 1
• l>
N
CM
CO
T-H
1>
• iO
N
T)H
<N
p
• "tf
Tj<
T-H
o
p
T-H T-H
d
• d
d
d
l>
y—i
' "^
N
rH
d
T-H
d od
o
CM
rH
<N
T^
co co
i-H
go
a
CM
O
• n
rh
00
00
i— i
• o
00
TtH
00
CO
00
• ^
<N
iq
l>
00
• iq
p
T-H
o
LO
TjH O
d
" d
d
i>
1— 1
T-i
T-H
CO
T-H
d
d
id CM
o
o
^
i— i
i—^
(M
<M
l> iO
i— i
"d
#o
'•+3
o3
CO
o
■ y—i
CO
CO
CM
iO
• (N
^H
OS
tH
CM
00
■ OS
00
N
OS
CO
• O
T-H
OS
o
O
iO CO
d
' d
d
tH"
"*
id
' 00
CO
d
d
T-H
rjn 1>
o
CM
i— i
i— 1
(M
1— 1
1> iO
CO iO CO CO OS
p CO p lo p
cd -^ id cd d
t-H t-H t-H CO
IN "HO CON
^ CO O O p t-h CM
d cd rH d d cd cm'
t-H 1>- IO
00
CM
• tH O CO CM CO
• CO 00 p CM O
•CBNONN
• rJH rj^ O p p CM# CO
d
o
' d d TjH CO TjH
i— I i— 1 i— 1 CO
' TtH tJH rH d d od d
00
•■HH^NMN
• 00 ^ CO p CO
• OS rH OS t-H O
• t-H CM p rH p CO CM
d
o
' n d •* d d
CM rH CO
r-^i cd d d d i> y-t
CM I> TjH
tn -a d v
a Oh
tH Oh
[V, d
o o
Sh -" R * S-H Mr— I k-H ^ tH M<Q o
GEOPHYSICAL LABORATORY
487
co
43
Si
O
o
a>
o3
43
Si 03
o 2
s-s
©43
*£
03 (h
CO 03
ft£
-. o3
5-H
o
o
d
03
O
<g S o
•^ 73
he a 3
3 03
"" 03 02
."S Oh
o3
CO
co
>>
Si
o
o
d
03
43
o
> >
O "3
S-i Si
03 03
«4-l «+H
o3 o3
T3
d
o3
03"
'43
03
d
b£
c3
43
o
S
o
T2
03
CO
ft
CO*"
CO
>>
Si
a
O
d
03
43
ft
b/0
d
e3
03"
CO
o3
43 43
O
£
o
d
03
co
ft"o
* O
co c<3
b'E
§-d
fl 0>
03
o
03
CO
d
O
s*
03
■+3
O
a
o3
03
d
o
Si
«
03
ftft^
Ol 03 _
-P +3 "+H
■rt 'rt O
a
03 -f
03
O
-4-= -+^
co
03 03
d d -
bC bfi g
o3 o3 >
a a
o3
bfi
'co
Ph
S? «
. > 03
O Si CO
O P >>
d go 73
2 -a S
03 03
bC
a *
03 ►>
03 03
4* &
co o
. O
43 rd •="*
~ " ft
03 [S
if 8
'9 8
o3 ^3
- ft
$ 03
— « bC
d
o3
d d it,
o3 o3 -d .,
^ O
03 03
+= ■*= -
JH .1-1 +S O
be be d -2
d d 03 o
o3 c3 d 9
bC
o3 "*
I— I _» .S CO
o3
"S Oh
o3
is _o 0 ^
e8 H! '43 o
o o s ft
*-d ft o "-1
•s a o
©<^r - ft
^rQ^ -
Oi 00 CN O
H H H lO
-.'-.' ' •
CN (M CN CN
t^ t>. I> t^
PQ pq pq pq
^ ^ b
"d =3 03
d rd si
ft ft42
0
o>
Si
o
"■+3 "43 "43 *g
'"3 "C "s3 d
43 43 43 -f3
ft ft ftO
Si Si "
o o
ft ft
(1
Mh o
P0<f<f >>
-* lO T*H
i-H TJH CN
CN CN CM
t^ b- l^
PQ pq PQ
03
bC
o3
Si
03
►
<
CO*
03
03
42
O
U
ft
o
Si
o
42 S
o3
d
<
^03^03^03^03 03 03 03 Q
ft'BH'ft'ft'ft'ft'ftW t
a a a a a a a^t^
O3a3o3o3o3o3c32o3
wmmwmmm 5n
<! W
'— l N CO ^ W © N * ■+-
488
CARNEGIE INSTITUTION
more rapid cooling in the volcanic en-
vironment (see below) . The vesicular
(z=amygdaloidal) nature of the lava and
the abundant evidence of explosive erup-
tions in the form of breccias and tuffs
point to a high water content, which, if
it had been retained, would probably
have induced formation of hornblende.
Hornblende phenocrysts have been re-
ported in the Douglas Island volcanics
southwest of Juneau (Lathram et al.,
1965) .
The strongest correlation between the
lavas and the ultramafic rocks is in the
chemistry of their pyroxenes, which are
closely similar for all major constituents
(Figs. 37, 38). Some of the volcanic
augite shows higher MgO/FeO ratios and
Cr203 than its plutonic counterpart, but
this difference is believed to be a matter
of sampling. Electron microprobe work
on the ultramafic rocks has so far been
exclusively on the relatively iron-rich
hornblende - magnetite clinopyroxenite,
and it is anticipated that when peridotite
and olivine clinopyroxenite are similarly
investigated, more magnesian and chro-
miferous pyroxenes will be found.
The similarity of the volcanic and plu-
tonic pyroxenes is of course to some ex-
tent fortuitous. As noted above, the more
iron-rich volcanic pyroxene probably
crystallized during quenching of the lava,
whereas the analogous plutonic pyroxene
appears to have formed by equilibration
of relatively magnesian cumulus pyrox-
ene with a large proportion of intercumu-
lus liquid (Irvine, 1973). This difference,
however, may not be as crucial as it
seems. Consider, for example, crystal-
lization in the system forsterite-fayalite:
for a given melt, the olivine formed by a
perfect quench should be compositionally
identical with that produced by equilib-
rium crystallization. This relationship
obtains in many other systems with ex-
tensive solid solutions and must pertain
to some degree to magmas.
Some of the clinopyroxene in horn-
blende-magnetite clinopyroxenite is un-
usual in having a ratio, Ca/(Ca + Mg
+ Fe), greater than 0.5. In terms of end
members, this ratio reflects the combina-
tion of abundant Ca-Tschermak's mole-
cule (CaAl2Si06) with low enstatite
(MgSi03) and ferrosilite (FeSi03). In
terms of physical chemistry, it probably
relates to a low activity of silica, through
equilibria of the type
CaAl2Si08 *± CaAl2Si06 + Si02
2(Mg,Fe)Si03*± (Mg,Fe)2Si04 + Si02.
(With low Si02 activity, the first reaction
would tend to retard the formation of
plagioclase, and the second, to increase
the abundance of olivine, effects that are
consistent with the overall character of
the ultramafic rocks.) Or, considered in
yet another way, the high Ca content of
the plutonic pyroxene probably reflects
its association with hornblende and with
magnetite having exsolved spinel. This
association, induced in part by high H20
activity, probably presented more com-
petition for Si02 and created higher
A1203 activity than was experienced by
the volcanic pyroxene.
There is less similarity between the
magnetites in the volcanic and ultramafic
rocks, but the plutonic phase has ex-
solved large amounts of both ilmenite
and spinel. As seen in Fig. 40, the vol-
canic magnetite contains enough Ti02
and A1203 to undergo similar exsolution.
The volcanic magnetite also differs in
being richer in Cr203, but this too could
reflect a difference in cooling history.
Under plutonic conditions, much of the
Cr203 would probably be fractionated
into early formed chromite, and perhaps
then, in the absence of its stabilizing
effect on spinel structures, the magnetite
would precipitate later as compared with
augite and in proportionately larger
amounts, as in the ultramafic rocks (Fig.
35).
It would add considerable credibility
to the thesis that the volcanic rocks
represent the kind of magma that was
parental to the ultramafic bodies if they
were to show some unexpected feature
GEOPHYSICAL LABORATORY
489
2Ti
Ilmenite
O Bridget Cove volcanics
• Duke Island
hornblende -magnetite
clinopyroxenite
Atomic per cent
Fig. 40. Ternary plot of Ti, Fe3+, and Al, comparing magnetite phenocrysts in the Bridget Cove
volcanics with magnetite and exsolved spinel and ilmenite in hornblende-magnetite clinopyrox-
enite from Duke Island, Alaska. Ti has been doubled to show the end-member combination (Mg,
FeHTiO* in proper proportion to the combinations (Mg,Fe)Fe204 and (Mg,Fe)Al204. Data from
eight samples of volcanic rocks, four of ultramafic rocks.
that proved compatible with the nature
of the ultramafic rocks or, better still,
conducive to their formation. The high
content of alkalies, especially K20, is a
feature that possibly fits this category.
It was not anticipated, but in review the
ultramafic rocks show surprisingly abun-
dant evidence of the presence of consid-
erable K20. In Alaska, biotite is a con-
spicuous phase in magnetite clinopyrox-
enite and hornblendite at Snettisham,
Haines, and Klukwan, among other
places, and phlogopite occurs sporadic-
ally in olivine clinopyroxenite at Kane
Peak (Taylor, 1967; see Fig. 34 for local-
ities). Phlogopite is also a widespread
minor phase in the Polaris ultramafic
complex in northern British Columbia,
and this complex contains a few small
bodies of syenite (Roots, 1954; Irvine,
unpublished data). In southern British
Columbia, the Tulameen complex is ad-
joined by a large cogenetic body of sye-
nogabbro and syenodiorite with abun-
dant interstitial K-feldspar. The large
complexes at Duke Island and Union
Bay in Alaska do not contain much bio-
tite or phlogopite, but hornblende at
Duke Island has up to 1.5% K20.
Abundant alkalies should also be con-
ducive to the formation of ultramafic
cumulates. In the system diopside-
forsterite-anorthite-albite of Yoder and
Tilley (1962, p. 395), for example, melts
with a given total content of feldspar
components falling in the olivine and
pyroxene liquidus volumes should crys-
tallize more mafic material before pre-
cipitating plagioclase if they have high
albite/anorthite ratios. A similar rela-
tionship obtains in systems with K-
feldspar (see Fig. 41). The Bridget Cove
490
CARNEGIE INSTITUTION
Ab, Or
BRIDGET COVE
VOLCANICS
NORM DATA
Cation equivalents
Cpx + 01
3-»— Olivine
Augite
Whole-rock analyses
• General
(X Rock with augite
quench crystals
Matrix analyses
+ General
x Rock with augite
quench crystals
Ab + Or + Ne + Lc
Cation equivalents
Fig. 41. Diagrams comparing norm data on whole-rock and matrix compositions in the Bridget
Cove volcanics with liquidus boundaries in systems involving diopside (Di), forsterite (Fo), anor-
thite (An), albite (Ab), and K-feldspar (Or). The tetrahedral diagrams are modeled after the
system Di-Fo-Ab-An of Yoder and Tilley (1962, Fig. 10). The liquidus boundary for Di-An-Ab
is from Kushiro (this Report); that for Fo-An-Ab is from Schairer and Yoder (Year Book 65,
GEOPHYSICAL LABORATORY
491
volcanics have crystallized large amounts
of mafic material without precipitating
feldspar, and a comparison of the whole-
rock, phenocryst, and matrix composi-
tions with the liquidus boundaries of
various synthetic systems in Fig. 41 sug-
gests that the abundance of alkalies was
a significant contributing factor.
Figure 41 is also of interest in that it
suggests that to obtain materials balance
between the whole-rock and phenocryst
and matrix compositions, the matrix
must be largely paired off against the
more aluminous, iron-rich pyroxene that
is believed to have formed by quenching.
The implication is that the matrix com-
positions were largely determined by
quenching. Another point is that, even
though the high content of alkalies in the
lavas should increase their potential
yield of ultramafic differentiates, a high
content or activity of H20 would be even
more effective. It appears that at H2O
pressures of several kilobars, the lavas
would yield at least 30% and possibly
more than 40% of ultramafic material
before crystallizing feldspar, amounts
that are probably more than enough to
produce Alaskan-type ultramafic bodies.
In review, even if the Bridget Cove
volcanics have no genetic relationship to
the Alaskan-type complexes, they illus-
trate two points relevant to their origin.
First, ultramafic magmas are not essen-
tial to produce the large quantities of
mafic minerals contained in the com-
plexes: a reasonable alternative is avail-
able. And second, the kind of magma
that does produce these minerals is not
likely to be tholeiitic or to differentiate
to calc-alkaline andesite. The choice of
parental magma seems clearly to be lim-
ited to an alkaline type, and the most
probable composition appears to be a
water-rich picritic ankaramite.
Principles of Melting of Hydrous
Phases in Silicate Melt
David H. Eggler
Hydrous phases in the earth's mantle
contain H20, both in subsolidus phase
assemblages and in the presence of sili-
cate liquid produced by partial melting.
Stability of hydrous phases in such
liquids is one of the main factors affect-
ing the character of the melt (Yoder,
1969) . To clarify some of the principles
of hydrous phase stability, particularly
the stability of hydrous phases in rela-
tion to solidus temperatures in H20-
undersaturated melts, simple model sys-
tems have been analyzed geometrically.
Two models are presented here.
The first binary model A-H20 involves
a relatively refractory component A, the
volatile component H20, a hydrous phase
H, and liquid L. Relevant principles can
be developed with this simple system as
an aid in understanding a more compli-
cated ternary model A-B-H20, which
contains an additional refractory com-
ponent B. Phase relations of these sys-
tems in pressure-temperature (P-T) pro-
jection appear in Figs. 42, 43, and 44.
Phase relations have been derived by
Schreinemaker's rules from composi-
tional geometries that are characterized
by variable composition of the liquid
phase. Liquid composition is variable
because silicate melts can dissolve more
H20 at high pressure than at low pressure
and because vapor-absent liquids contain
less H20 than liquids in equilibrium with
vapor. In order to clarify reactions in the
system ^4-H20, liquids containing less H20
than H are called Lx and liquids contain-
ing more H20 than H, L2. Similarly, in
A-B-HoO, liquids containing various
amounts of H20 are called Li, L2, and L3.
The significance of these particular corn-
Fig. 2). Other boundaries are inferred from data of Schairer (1954), Morse (Year Book 67, Fig.
27), Yoder and Upton (Year Book 70, Fig. 4), and Yoder (1968, Figs. 2 and 8A). Complications
relating to the incongment melting of Or at 1 atm are realistically avoided by assuming an alkali
feldspar composition of about AbiOri.
492
CARNEGIE INSTITUTION
P=9.6
A+L\L L2+V
A + V
H + V
H20 0.
P=74
L2+V
.
P--
74
.
" \ A + L
-
\
•
/
•'"■
'•^M
He
/a+
H
■
■
P = 9.2
A
. +
1 1-
\ L+V .
L
-
\
- A
+
- H
H + V
A H
H20 0.
H^ 0. 0.5
aH20
P = 7 0
1.0
A H
P = 4.0
L2+V
M_o+H
H + V
'HoO
P=4.0
H20 0. 0.5 1.0 A H
aH20
J I L
H20 0. 0.5 1.0
aH20
Fig. 42. Phase relations in the model system .4-H20, including a P-T projection, isobaric T-
composition sections, and isobaric 7T-H20 activity diagrams. Solid lines refer to stable univariant
reactions and stable divariant assemblages ; dashed lines, to metastable reactions and assemblages.
P, T, and composition scales are arbitrary.
positions is that they define three sepa-
rate vapor-absent reactions, as the dia-
gram in the inset in Fig. 44 indicates. It
should be recognized that these composi-
tions are drawn as points only to- define
the labeling convention; the liquid field
exhibits continuous solution between
these compositions at appropriate tem-
peratures. Liquid compositions have
been labeled on diagrams where they
unambiguously conform to the labeling
convention.
From the P-T projection in Fig. 42
isobaric composition sections have been
drawn, and from the P-T projection in
Fig. 44 isobaric sections for one pseudo-
binary join have been drawn. At each
pressure an isobaric H20 activity (aH2o)
GEOPHYSICAL LABORATORY
493
P = 2.3
l_2+V
A H
P= 1.8
H20 0. 0.5 1.0 A H
aH20
P=l.8
A H
P = l.2
L| +V
H+V
H2O 0 0.5 1.0 A H
aH20
J I L
H20 0. 0.5 1.0
aH20
j 1 1 _j
H20 0. 0.5 1.0
aH20
Fig. 43. Additional isobaric sections for the model system A-H20.
diagram has also been constructed (aH2o
= /h2o//°h2o, where /°H2o = fugacity of
pure H20 at the T and P of interest) ;
curves in the aH2o diagrams show changes
of &h2o w^n T in H20-bearing phases in
divariant fields. The curves reflect the
fact that aH2o in liquid varies as a func-
tion of H20 content in liquid (Burnham
and Davis, 1971). The curves are de-
rived from aH2o contours on the P-T pro-
jection (not shown) drawn to conform to
the shape of such contours on diagrams
for actual rock or hydrate compositions.
An example is the system granite-H20,
incorporating the hydrate muscovite, of
Kerrick (1972).
These models are applicable to all
hydrous phases and to hydrate-bearing
rock compositions for which the dehydra-
tion curve intersects the solidus. Certain
portions of the diagram, however, will
not apply to certain hydrous phases. Two
cases in particular are of interest. First,
the models show a change in slope in the
metastable extension of the dehydration
curve (L), H — A + V in A-B.20 and
H — A + B + V in A-B-H20. Am-
phiboles exhibit this feature, generally
because breakdown products transform
with pressure to denser phases, a refine-
ment omitted from this simple model. A
mica such as phlogopite does not show
such a change in slope, however, at least
to a pressure of 40 kbar (Yoder and
Kushiro, 1969). For modeling of mica
behavior, therefore, that portion of Figs.
42 and 44 above P = 7.0 should be ig-
nored. Second, in some systems the de-
hydration curve may intersect the solidus
at a pressure higher than that of the
models. In Fig. 42, that intersection is at
a pressure sufficiently low that H20-
saturated liquid contains less H20 than
H (see isobaric section at P = 1.2, Fig.
43). The intersection may be at a pres-
sure sufficiently high, however, that H20-
saturated liquid contains more H20 than
H , and the singular point Si (or, analo-
gously, Si and S2 in Fig. 44) does not
occur on a stable melting curve.
Several principles of melting in the
vapor-absent region, which are of partic-
ular interest to petrologists, are well
explained by the binary model. Hydrous
phases may coexist with liquid at pres-
sures between invariant points I\ and I2
(Fig. 42) . At higher and lower pressures,
H breaks down by dehydration below
the solidus. Between I\ and I2, composi-
tions melt either by a vapor-saturated
melting reaction (A) or by vapor-absent
melting reaction (V) . Reference to an
494
CARNEGIE INSTITUTION
H20
Fig. 44. Phase relations in the model system A-B-H.2O, including a P-T projection, isobaric
pseudobinary joins, and isobaric H20 activity diagrams. P, T, and composition scales are arbi-
trary. Lined areas in sections are subsolidus regions, and stippled areas are regions where hy-
drous phase H coexists with liquid.
isobaric section, such as P = 7.4 (Fig.
42), reveals that the temperature of the
reaction (7) is located where aH2o curves
for A + H and for A + L intersect. This
situation is generated by the requirement
that the vapor-absent assemblage melts
when H supplies sufficient H20 by its
breakdown to satisfy the melt or, more
precisely, that aH2o in H equals aH20 in L.
Melt produced by reaction (V) always
contains less H20 than melt produced by
the vapor-saturated melting reaction. At
P = 7.4, H does not persist above the
temperature of the vapor-absent melting.
However, at P = 4.0, H may coexist with
liquid. This different topology (and,
concomitantly, change in the vapor-
absent melting reaction) results from the
decreased solubility of H20 in liquid in
equilibrium with A. The temperature
stability limit of H, which is a function
of &h2o or amount of H20 in liquid,
GEOPHYSICAL LABORATORY
495
reaches a maximum at an intermediate
value of aH2o- Such maxima have been
found in systems involving melting of
phlogopite (Yoder and Kushiro, 1969) ,
zoisite (Boettcher, 1970) , and pargasite
(Holloway, 1973) . The temperature
maximum for H occurs at a restricted
univariant reaction (Ricci, 1951, p. 29),
H = L, which is located at loci of inter-
section of aH2o contours for the divariant
assemblages H + L2 and H -f- L\. The
loci extend between singular points Si
and $2. For this simple system, the re-
stricted univariant reaction is a congru-
ent melting reaction, generated by com-
positional coincidence of H and L. In
systems with additional refractory com-
ponents, the temperature maxima will
occur at incongruent melting reactions
that have restricted univariancy by vir-
tue of phase composition colinearity.
Such reactions are noted for phlogopite
by Yoder and Kushiro (1969) and Mo-
dreski and Boettcher (1972).
The ternary system A-B-H20 (Fig.
44) is similar to A-H20, except that one
more reaction is present and two addi-
tional singular points are present. These
points exist because there are two re-
stricted univariant reactions, H + B =
L and H = A -\- L. Phase H melts in-
congruently in these reactions because
the liquid field does not extend to H until
very high temperatures are reached, a
feature incorporated into the model to
reproduce incongruent melting behavior
found for natural amphiboles and micas.
Pseudobinary joins of this system, one
of which is drawn for two pressures in
Fig. 44, closely model melting features of
rocks (e.g., Robertson and Wyllie, 1971).
The ternary model illustrates several
points of interest with regard to melting
of peridotite in the mantle. As empha-
sized above, at pressures between i\ and
I2, a hydrous phase disappears, as T is
raised, by melting rather than by dehy-
dration. For amphibole-bearing perido-
tite, an invariant point analogous to J2
occurs at 20-25 kbar (Kushiro, Year
Book 68, p. 246) . Therefore, in the pres-
sure range 1-25 kbar peridotite melts by
one of two reactions, depending on
whether or not the subsolidus assemblage
contains free vapor. The temperatures of
these reactions lie at the top of the lined
areas in the isobaric sections in Fig. 44.
Both the vapor-present and vapor-absent
reactions change with pressure, the for-
mer at singular point Si and the latter at
S3 and $4. At temperatures above these
reactions, the hydrous phase may or may
not be stable. At P = 5.0, for composi-
tions containing vapor in the subsolidus
region, H is stable in the presence of
liquid, whether or not vapor is also pres-
ent. In addition, for certain compositions
that are vapor-absent subsolidus, and
that melt at the vapor-absent solidus, H
is also stable above the solidus. At P =
8.4, however, the vapor-absent solidus
marks the upper temperature stability
limit of H, a situation produced because
the pressure is above the pressure of
singular point *S4. A small amount of
partial melting of the vapor-absent as-
semblage A + B + H at P — 8.4 will
produce a H20-rich liquid (and will melt
all H) , but further melting will reduce
the H20 content of the liquid.
It is evident that persistence of a hy-
drous phase into the melting range, above
the vapor-absent solidus, is a function
both of total pressure and of aH2o (or
amount of H20). Stability of hydrous
phases in melt has important implica-
tions for trace-element partitioning and
for fractionation trends.
Addition of a second volatile compo-
nent (C02) to these models does not
change principles of melting of the hy-
drous phase but introduces new phase
assemblages in equilibrium with H20-
poor fluids. These models will be dis-
cussed elsewhere.
Solubility of H20 in Forsterite Melt
at 20 KBAR
F. N. Hodges
The amount of water required to satu-
rate forsterite (Mg2Si04) melt has been
496
CARNEGIE INSTITUTION
determined at a pressure of 20 kbar.
Knowledge of the high-pressure proper-
ties of forsterite, a component of almost
all proposed models of the mantle, is
important for an understanding of deep-
seated processes within the earth. Davis
and England (1964) determined the an-
hydrous melting curve of forsterite to
50 kbar, and Kushiro and Yoder {Year
Book 67, pp. 153-154) determined the
water-saturated melting curve to 30 kbar.
Experimental runs were made in a
solid-media, high-pressure apparatus
similar to the one described by Boyd and
England (1960). Sealed platinum cap-
sules (1.8 mm OD) were surrounded by
finely powdered crushable alumina dur-
ing runs to prevent crumpling of the cap-
sule. Water contents in the runs ranged
from 17.5 to 23 wt % (Fig. 45). Atmos-
pheric water vapor, absorbed on finely
ground forsterite, was considered a pos-
sible source of error in these experiments.
A portion of dried forsterite powder,
kept in a capped bottle under normal
laboratory conditions (humidity un-
2000
10 15
Weight per cent H20
Fig. 45. Phase relations in the system Mg,
S1O4-H2O at 20 kbar.
known but high) for approximately one
month, was dried at 110°C for 16 hours.
Weight loss as a result of heating was
approximately 0.1 wt %, indicating that
gain of atmospheric water vapor is a
negligible source of error. The forsterite
used in this study was synthesized and
kindly provided by Dr. H. K. Mao.
Phase relations determined in this
study are presented in Fig. 45. Cloudy,
commonly nearly opaque grains of for-
sterite containing myriads of tiny inclu-
sions and bubbles arranged in a herring-
bone pattern are interpreted as quenched
liquid. In the field marked Fo + V, only
clear granular forsterite was found. In
the field marked Fo -f- L, clear granular
forsterite and quench liquid were found,
whereas in the fields marked L and L -f-
V, only quenched liquid was found. Dif-
ferent symbols for L and L + V have
not been used because of the difficulty of
detecting vapor in this system. The
boundary between L and L -f- V is drawn
as a dashed line on theoretical grounds
but not on the basis of experimental data.
The solidus of Mg2Si04-H20 at 20
kbar has been located at 1425 =± 10 °C,
in excellent agreement with the results of
Kushiro and Yoder {Year Book 67, pp.
153-154). The amount of water neces-
sary to saturate forsterite melt at 20 kbar
and 1425 °C, given by the water content
where Fo + L + V coexist, is 20.2 ± 1.0
wt °/o (66 mole %).
The solubility of water in forsterite
melt at 20 kbar (66 mole %) is equiva-
lent to approximately 0.5 mole of water
(H20) , per formula oxygen. Eggler (this
Report) has determined the solubilities
of water in enstatite, diopside, and albite
melts at 20 kbar. Within the limits of
experimental error, the values found by
Eggler are, when normalized to a per
formula oxygen basis, identical with the
value for forsterite. Solubility values at
10 kbar for enstatite (Kushiro and Yo-
der, Year Book 67, pp. 154-158) and
anorthite (Yoder, Year Book 64, pp.
82-89) are identical (approximately 0.24
GEOPHYSICAL LABORATORY
497
mole of water per formula oxygen) within
the limits of experimental accuracy.
The consistency of these few data is
extremely interesting, both for knowl-
edge of water solubility in complex sili-
cate melts at high pressures and for the
understanding of mechanisms of solution
in silicate melts under conditions of high
pressure. Hamilton, Burnham, and Os-
born (1964) found that at relatively low
pressure the solubility of water in sev-
eral molten silicates (Si02 is a notable
exception) varies linearly with the square
root of pressure and showed that this
phenomenon can be explained, at least to
a first approximation, by assuming that
water enters the melt by reacting with
oxygen in the silicate melt to form OH"
(H2Ogas + 02-meit ** 20H-meit). The
high-pressure solubility data presented
above deviate considerably from the low-
pressure linear relation. One possible ex-
planation for the deviation is the sugges-
tion offered by Shaw (1964) that at some
mole fraction of water (dependent upon
the number of formula oxygens), it may
enter silicate melts as discrete molecules
as well as by reaction with oxygen.
Regularities in the Shift of Liquidus
Boundaries in Silicate Systems and
Their Significance in Magma Genesis
/. Kushiro
Effects of various oxides on the shift
of the 1-atm liquidus boundaries between
forsterite (Mg2Si04) and enstatite
(MgSi03), enstatite and silica, pseudo-
wollastonite (CaSi03) and silica, and
fayalite (Fe2Si04) and silica have been
examined. It is found that these bound-
aries shift toward a silica-rich composi-
tion by solution of oxides of monovalent
elements (e.g., H20, K20, and Na20)
into the melt, but they shift toward a
silica-poor composition by solution of
oxides of polyvalent elements (e.g., Cr203,
Ti02, and P205). The oxides of divalent
elements have the smaller effect. These
observations can be explained by struc-
tural changes in the silicate melt; mono-
valent elements prevent polymerization
of (Si04)4" tetrahedra, so that olivine
(orthosilicate) would be favored over en-
statite (chain silicate) or silica minerals
(framework silicate) , in their crystalliza-
tion from the melt and enstatite or
pseudowollastonite (chain silicate) would
be favored over silica minerals. Poly-
valent elements, on the other hand, poly-
merize (Si04)4" tetrahedra and have the
opposite effect. At higher pressures the
boundaries shift toward a silica-poor
composition in the systems examined.
The regularity in shift of the liquidus
boundaries found in this study has a wide
application to the problems of genesis of
magmas.
Liquidus Boundaries between Olivine,
Pyroxene, CaSi03, and Silica
Polymorphs at 1 atm
The liquidus boundary between for-
sterite (Fo) and protoenstatite (Pr) at
1 atm in the systems MgO-Si02-X is
shown in Fig. 46, where X represents
MgSi03 60 70
SiC>2 Mole per cent
80
Fig. 46. Shift of the forsterite (Fo)-proto-
enstatite (Pr) liquidus boundary at 1 atm with
the addition of various oxides. Protoenstatite
forms appreciable amounts of solid solution
with the addition of CaO, FeO and A120;). The
ordinate is the number of moles of each oxide
(doubled for oxides containing two cations),
and abscissa, mole % silica on the join MgO-
Si02.
498
CARNEGIE INSTITUTION
K20, Na20, CaO, FeO, A1203, Cr203, and
Ti02. The boundaries are based on data
obtained by previous investigators. It is
found that the Fo-Pr liquidus boundary
shifts significantly toward silica with the
addition of K20 and Na20, while it shifts
away from silica with the addition of
Cr203 and Ti02. The boundary also
shifts toward silica with addition of CaO,
FeO, and A1203, but the effects are much
smaller than with Na20 and K20.
The order of shifting in Fig. 46 is
systematically related to the valency of
the element in the oxides added, in the
order monovalent - divalent - trivalent -
tetravalent. This effect obtains even
though forsterite and protoenstatite take
appreciable amounts of some of the
oxides (e.g., CaO, FeO, and A1203) in solid
solution. The effect of H20 cannot be
compared at 1 atm; a series of high-
pressure hydrothermal experiments have
shown, however, that the Fo-enstatite
(En) boundary shifts significantly to-
ward silica in the presence of excess
water (Kushiro, Yoder, and Nishikawa,
1968; Kushiro, 1969, 1972a). The vola-
tile C02 may have an effect similar to
that of Ti02. Eggler (this Report) has
found that several wt °/o C02 can be dis-
solved into diopside and enstatite melts
at 20 kbar, suggesting that the effect of
C02 may be significant. P205 apparently
shifts the Fo-Pr boundary strongly to-
ward silica-poor compositions, as will be
seen from the experiments described
below.
The liquidus boundary between proto-
enstatite and the silica polymorphs (cris-
tobalite and tridymite) at 1 atm in the
systems MgO-Si02-X is shown in Fig. 47,
where X represents K20, Na20, CaO,
FeO, A1203, and Cr203. As in the pre-
vious case, K20 and Na20 cause the
boundary to shift strongly toward silica,
whereas Cr203 shifts it to compositions
poorer in silica. In this case, however,
A1203 has a stronger effect than FeO. It
is also noted that the curvature of the
boundary changes with transition of the
E
C
o
0.3
1 / 1
FeO/
/ CaO
/ AI2°V
llr /
/Cr //
I
Na20
0.2
Pr
/ J/lr
/Cr X
1 / Cr
/Tr
K20
O.I
n
Cr203/y
i i
^
1
50
60
Si02
70
Mole
80
per cent
90
Fig. 47. Shift of the protoenstatite-silica poly-
morph liquidus boundary at 1 atm with the
addition of various oxides. Cr, cristobalite ; Tr,
tridymite. The abscissa is mole % silica on the
join MgO-Si02.
silica polymorph from cristobalite to
tridymite.
The liquidus boundary between pseu-
dowollastonite (Pwo) and the silica min-
eral at 1 atm in the systems CaO-Si02-X
is shown in Fig. 48, where X represents
K20, Na20, MgO, FeO, A1203, Ti02, and
P205. Pseudowollastonite is not a pyrox-
ene but has a chain structure related to
those of pyroxenes. The Pwo-silica poly-
morph liquidus boundary also shifts reg-
ularly in the way described above, except
with addition of A1203. Note in particu-
lar that P205, the oxide of a quinquiva-
lent element, causes an exceptionally
large shift away from silica.
Kracek (1930) determined the cristo-
balite liquidus in the systems Si02-X,
where X represents Li20, K20, Na20,
Rb20, and Cs20. He showed that the
order of effect of the oxide in lowering
the melting temperature of cristobalite is
Rb ~ Cs, K, Na, Li, Ba, Sr, Ca, and Mg.
The order K, Na, Ca (or Mg) is the same
as that found in shifting the Pr-silica and
Pwo-silica liquidus boundaries toward
silica.
For comparison, the liquidus boundary
between fayalite and the silica mineral
at 1 atm has been also examined in the
GEOPHYSICAL LABORATORY
499
CaSi02 60 70 80
Si02 Mole per cent
Shift of the pseudowollastonite (Pwo)-silica polymorph liquidus boundary at 1 atm
Fig. 48
with the addition of various oxides
The abscissa is mole % silica on the join CaO-Si02.
systems FeO-Si02-X, where X represents
KoO, Na20, CaO, MgO, A1203, and Fe203
(Fig. 49). The boundary is shifted
greatly toward silica by the addition of
KoO and Na20, whereas it shifts in the
opposite direction with the addition of
Fe203. As in the above systems, the
effects of the oxides of divalent elements
and A1203 are small, but the regularity of
shift is again confirmed.
The regularity of shift can be inter-
preted as relating to changes in the struc-
ture of the silicate melts. Monovalent
elements would tend to break Si-0 chains
and prevent the polymerization of
(Si04) 4~ tetrahedra so that crystallization
of silicate minerals in which there is less
linkage of (Si04)4~ tetrahedra should be
favored. Conversely, polyvalent elements
would tend to polymerize (Si04)4~ tetra-
hedra, so that crystallization of silicate
minerals in which there is more linkage
of (Si04) 4" tetrahedra should be favored.
Thus, the addition of oxides of monoval-
ent elements would cause the liquidus
field of forsterite to expand relative to
that of protoenstatite, and the liquidus
field of protoenstatite and pseudowollas-
tonite to expand relative to that of a
silica mineral, whereas the addition of
oxides of polyvalent elements should
cause the reverse effects.
At higher pressures the Fo-En bound-
ary shifts toward a silica-poor composi-
tion, as shown by Kushiro (1968, 1972a)
in the systems MgO-Si02-CaO, MgO-
Si02-Al203, Mg0-Si02-Na20, and MgO-
40 50 60 70 80
Si02 Mole percent (Join FeO-Si02)
Fig. 49. Shift of the fayalite (Fa)-tridymite
liquidus boundary at 1 atm with the addition of
various oxides. The abscissa is mole % silica
on the join FeO-Si02.
500
CARNEGIE INSTITUTION
Si02-H20 and by MacGregor (1969) in
the system MgO-Si02-Ti02. The expla-
nation in this case may be that pressure
causes a higher degree of polymerization
of (Si04)4" tetrahedra in the melt. If so,
one may also expect the liquidus bound-
aries between other silicate minerals to
shift toward silica-poor compositions
with increased pressure.
The regularity shown here should be
significant in the genesis of magmas.
Magmas formed by partial melting of
the upper mantle peridotite or by frac-
tional crystallization involving separa-
tion of olivine and orthopyroxene lie on
the olivine-pyroxene liquidus boundary.
Therefore, the compositions of such mag-
mas would be expected to become poorer
in silica in the presence of increased
amounts of polyvalent elements such as
Ti, Cr, P, and probably C. In fact, silica-
poor alkaline basaltic magmas are gen-
erally enriched in these elements. Many
of them are also enriched in alkalies.
There is no positive correlation between
alkalies and silica contents; very silica-
poor rocks (Si02, 35-45 wt %) are
not always rich in alkalies, suggesting
that alkalies are not the elements that
control silica content in the melt. Mag-
mas would also become silica-poor with
increasing pressure, as shown by several
investigators (e.g., Yoder and Tilley,
1962; Green and Ringwood, 1967a; Ku-
shiro, 1968). On the other hand, magmas
formed by partial melting of peridotite
or by fractional crystallization of basaltic
magma tend to be richer in silica in the
presence of H20 and alkalies. Recent
studies on the partial melting of perido-
tite compositions indicate that the liquids
are more silica-rich under H20-saturated
conditions (Kushiro, 1969, 1972a; Yoder,
Year Book 69, pp. 176-181; Mysen, this
Report) . The present study also suggests
that magmas formed by fractional crys-
tallization involving pyroxene, iron-rich
olivine, and a silica mineral should be
more silica-rich in the presence of greater
amounts of monovalent elements. Late-
stage differentiates such as dacitic, rhyo-
litic, and pegmatitic liquids are enriched
in silica and contain considerable
amounts of alkali elements and H20.
Crystallization of orthosilicate minerals
such as fayalite and almandine-rich gar-
net in silica-rich magmas (e.g., dacite
and rhyolite) may be explained by the
expansion of the liquidus fields of fayal-
ite (and probably garnet) in the presence
of appreciable amounts of alkalies and
H20 in these magmas.
Effect of P205 on the Forsterite-
Enstatite Liquidus Boundary
It was suggested in the previous sec-
tion that P2Os would greatly shift the
Fo-En liquidus boundary away from
silica. To test this possibility, the sys-
tems Fo-diopside (Di) -silica and Fo-
nepheline (Ne) -silica with the addition
of P205 have been studied. Because of
the difficulty of preparing anhydrous
starting materials containing P205, the
experiments were made under H20-satu-
rated conditions. The crystalline mix-
tures of each of seven different composi-
tions in the system Fo-Di-silica (Kushiro
and Schairer, Year Book 62, pp. 95-101)
and of four different compositions in the
system Fo-Ne-silica (Schairer and Yo-
der, Year Book 60, pp. 141-144) were
sealed in Pt capsules with H3P04 and
H20. The amounts of P205 and H20
range from 2.0 to 3.5 wt % and from
15 to 25 wt %, respectively. All the runs
were made at 20 kbar in solid-media,
high-pressure apparatus, using the
method described previously (Kushiro,
1969).
The results of the experiments are
shown in Fig. 50, in which the liquidus
boundaries under the same conditions but
with no P205 are shown for comparison.
In the system Fo-Di-silica-H20, the Fo-
Enss liquidus boundary is in the silica-
saturated region in the absence of P205,
but it is in the silica-undersaturated re-
gion in the presence of 2-3.5 wt % P205.
In the system Fo-Ne-silica, the liquidus
GEOPHYSICAL LABORATORY
501
CaMqSi206
20 kbar
Excess H20
Mg2Si04
Weight per cent
Fig. 50. (A) Forsterite-enstatite solid solution (En88) liquidus boundaries in the system forster-
ite-diopside-silica at 20 kbar under water-saturated conditions with 2.1-3.5 wt % P2O5. The liqui-
dus boundaries without P2O5 are from Kushiro (1969).
boundary is partly in the silica-saturated
region in the absence of P2O5, whereas it
shifts into the silica-undersaturated and
critically silica-undersaturated regions
by addition of 2.8-3.4 wt % P205. It is
evident that the Fo-Enss liquidus bound-
ary shifts toward silica-poor composi-
tions by addition of P2O5, confirming the
suggestion in the previous section.
The Fo-Enss liquidus boundaries in
these systems in the presence of 2—3.5 wt
% P2O5 are nearly the same as those
under anhydrous conditions with no PoO-,
at 20 kbar, indicating that 15-25 wt %
H20 and 2-3.5 wt % P205 have effects of
a similar degree on shifting the Fo-En
boundary but in opposite directions. It is
expected that the Fo-Enss liquidus bound-
502
CARNEGIE INSTITUTION
NaAlSiQ
NaAISi308
MgSiQ
MgSiO
50
Weight percent
Fig. 50. (B) Forsterite-enstatite solid solution liquidus boundaries in the system forsterite-
nepheline-silica at 20 kbar under water-saturated conditions with 2.8-3.4 wt % P2O5. The liquidus
boundary with no P205 (dashed line) is from Kushiro (1972a).
ary would shift to a compositional region
still more undersaturated in silica in the
presence of P2O5 under anhydrous con-
ditions. Magmas formed by partial melt-
ing of peridotite containing an appreci-
able amount of the jadeite component or
derived by crystallization of olivine and
orthopyroxene from basaltic magmas
should therefore be enriched in the neph-
eline component in the presence of P2O5
at high pressures.
Many alkali basalts contain more than
1 wt % P2O5, and some of them contain
up to 2.6 wt %, amounts comparable to
those used in the present experiments.
Thus the experimental results both sup-
port the regularity described in the pre-
vious section and suggest an important
role for P205 in the genesis of rocks such
as alkali basalts and nepheline-rich basic
rocks.
The System Diopside-Anorthite-
Albite: Determination of Com-
positions of Coexisting Phases
/. Kushiro
The compositions of coexisting diopside
solid solution, plagioclase, and liquid in
the system diopside (Di, CaMgSi206)-
anorthite (An, CaAl2Si208) -albite (Ab,
NaAlSisOg) at 1 atm have been deter-
mined with the electron microprobe. The
analyzed diopside solid solutions contain
up to 2.4 wt % A1203 and excess enstatite
(MgSi03), resulting in the generation
of excess wollastonite (CaSi03) and silica
components in the coexisting liquid. The
liquidus boundary between diopside solid
solution and plagioclase has been re-
vised; the boundary is concave only to-
ward diopside over its entire range.
Along the liquidus boundary, plagioclase
GEOPHYSICAL LABORATORY
503
changes its composition from An85Abi5
(wt %) at 1250 °C to An30Ab7o at
1150°C, comparable to the composition-
temperature relations of plagioclase in
natural magmas. A fractionation trend
toward pantelleritic compositions is re-
emphasized.
The system Di-An-Ab, denoted by
Bowen (1915) as a haplobasaltic and
haplodioritic system and studied in de-
tail by him, is a fundamental, simple
system for the understanding of crystal-
lization of basaltic and andesitic mag-
mas. The system had been considered a
ternary system; Osborn (1942), however,
found that the join Di-An is not binary
because of solid solution in diopside, and
Schairer and Yoder (1960) found that
the join Di-Ab is not binary either be-
cause of solid solution in plagioclase. It
is obvious from these findings that the
system Di-An-Ab is not ternary. In the
present experimental study, electron
microprobe analysis has been applied to
determine the compositions of coexisting
diopside solid solution, plagioclase, and
liquid and to clarify the crystallization
trend of the liquid.
Four different mixtures were selected
for the experiments: (1) Di7oAni5Abi5
(wt %), (2) Di5oAn25Ab25, (3) Di4oAn3o
Ab30, (4) Di2oAni5Ab65. Run duration
ranged from 3 days at 1230 °C to 30 days
at 1160°C. Electron microprobe analysis
has been made by the method described
by Finger and Hadidiacos {Year Book
70, pp. 269-275). For the analysis of
glass a 0.02 /xA specimen current and a
large electron beam (up to about 20 /xm)
were used.
with that determined by Bowen (1915),
although his boundary, drawn on the
basis of eleven points (shown by crosses
and plus signs in Fig. 51) and the minima
on the joins Di-An and Di-Ab, can be
moved by a small weight percentage in
terms of diopside content. In the albite-
rich portion of the system, analysis of
Na-rich glass is very difficult because of
the loss of Na, and since the analytical
data do not appear to be reliable, they
are not plotted. Examination of the pre-
existing data, however, indicates that the
previous boundary needs revision.
Schairer and Yoder (1960) have deter-
mined liquidus relations on the join
Di-Ab and found that the liquidus mini-
mum (piercing point) between diopside
solid solution and albite-rich plagioclase
is located at Dii0Ab9o (wt %) and that
the plagioclase liquidus has a maximum
owing to the effect of solid solution in
plagioclase. The minimum determined
by Schairer and Yoder (1960) , shown by
B in Fig. 51A, is about 7 wt % more
diopside-rich than that estimated by
Bowen (1915). In addition, by examin-
ing Bowen's (1915, p. 168) quenching
data it was found that the two critical
compositions (asterisks in Fig. 51) that
he had put in the diopside field could be
in the plagioclase field; these two mix-
tures contain both diopside and plagio-
clase at 1180° and 1150°C, respectively,
and they are all glass at 1183° and
1155°, respectively. The probable liqui-
dus boundary, drawn in Fig. 51A, is con-
cave only toward diopside, instead of
having double curvature as shown by
Bowen (1915).
Liquidus Boundary
The central portion of the liquidus
boundary has been determined at five
different temperatures from 1245° to
1225 °C, using mixtures 1, 2, and 3. The
results of analyses of five different
glasses, described below, are plotted in
Fig. 51A. The boundary determined in
this temperature range is nearly identical
Coexisting Phases
The diopside solid solution crystalliz-
ing along the liquidus boundary is found
to contain A1203 ranging from 0.37 to
2.4 wt % ; the A1203 content decreases
with lowering temperature. If all A1203
exists as CaAl2Si06 (Ca-Tschermak's
component) in the diopside solid solu-
tion, the amount of CaAl2Si06 ranges
504
CARNEGIE INSTITUTION
CaMgSipO,
2^6
NaAISi308
Weight per cent
CaAI2Si208
Fig. 51. (A) Revised liquidus boundary of the system diopside-anorthite-albite at 1 atm. Sym-
bols: open circles, electron microprobe analyses of glasses coexisting with diopside solid solu-
tion and plagioclase; plus signs and crosses, quenching runs by Bowen (1915) in the diopside
field and plagioclase field, respectively; asterisks, Bowen's quenching runs that he considered to be
in the diopside field (see text) ; A and B, liquidus minima on the joins diopside-anorthite
(Bowen, 1915; Osborn, 1942) and diopside-albite (Schairer and Yoder, 1960), respectively. Liquid
compositions are projected onto the plane diopside-anorthite-albite from Si02 and CaSi03.
from 0.8 to 5.1 wt %. The analyses also
show that diopside solid solution con-
tains excess Mg over Ca, indicating
that its composition lies on the plane
Di-En-CaAl2Si06 (or Di-En-Al203).
The pyroxene formula of one of the diop-
side solid solutions (A1203 2.41 wt %) is
Na0.02Ca0.89Mg0.94Al0.10Si2.00Oe, which
contains 5.5 mole % MgSi03 or 11 mole
% MgSi03 if CaAl2Si06 is subtracted.
Compositions of the glasses, which are
used to determine the liquidus boundary,
do not lie on the plane Di-An-Ab because
of solid solutions in diopside that do not
lie on this plane and for another reason,
mentioned below. Two examples of anal-
yses of glasses and their norms are given
in Table 8. The glass is usually very
homogeneous, and no quench crystals are
observed. The ratio of glass to crystals
is greater than 1 in most cases. The re-
sults indicate that these glasses contain
excess Si02 and wollastonite over the
composition on the plane Di-An-Ab.
This observation can be explained by the
presence of CaAl2Si06 and excess MgSi03
in diopside solid solution, mentioned
above. Another cause, which would be
GEOPHYSICAL LABORATORY
505
NaAISi308
1160° 50 /ISO" 1215° 1225" 1230'
Weight per cent
CaAI2Si208
Fig. 51. (B) Three-phase triangles at 1230° and 1160°C with compositions of plagioclase at dif-
ferent temperatures. Diopside solid solutions are projected onto the join diopside-anorthite from
Si02.
effective in the albite-rich portion of the
system, is suggested by Schairer and Yo-
der (1960) from their study on the Di-Ab
join. Plagioclase crystallizing near the
liquidus of this join is not pure albite but
contains appreciable amounts of anorth-
ite component. Consequently, the co-
existing liquid must contain sodium sili-
cate and enstatite components and excess
silica. In the light of suggestions by
Bailey and Schairer (1966) and the find-
ings of Schairer and Yoder (Year Book
69, pp. 160-163) on the presence of
3Na20-8Si03 at the quartz-bearing eu-
tectic in the system Na20-Si02, the prob-
able reaction is written as follows:
6NaAlSi308 + 3CaMgSi206
= 3CaAl2Si208 + 3Na20-8Si02 +
3MgSi03 + 7Si02
This reaction would also hold in the
albite-rich part of the system Di-An-Ab,
so that the liquid probably contains
3Na20-8Si02 component. Unfortunately
these relations could not be demonstrated
by the electron microprobe analysis be-
cause of the difficulty mentioned above.
Part of the MgSi08 component may be
combined with excess wollastonite com-
ponent to make diopside component in
the melt.
Plagioclase crystallizing along the
liquidus boundary (or coexisting with
506
CARNEGIE INSTITUTION
TABLE 8. Electron Microprobe Analyses and Norms of Glasses Formed
at 1245° and 1225° C in the System Diopside-Anorthite-Albite
T, °C
Si02
AI2O3 MgO
CaO
Na20
1245
1225
58.5
57.2
Q
Analyses, wt %
14.7 7.50
15.9 6.01
Norms
Ab An
16.9
13.8
Di
40.3
32.3
2.46
4.44
Wo
1245
1225
8.65
2.60
20.8 29.1
37.6 23.5
1.27
1.46
liquid and diopside solid solution)
changes its composition from An73Ab27
(wt %) at 1230°C to An35Ab65 at 1160°C
(Fig. 51B). The compositional change of
plagioclase with temperature along the
liquidus boundary is also shown in Fig.
52, onto which the compositions of the
coexisting liquids are projected from
diopside. The composition of plagioclase
changes extensively with a relatively
small change of temperature: from
An85Abi5 at 1250° to An30Ab70 at 1150°C.
It is noted that the composition-tempera-
ture relation of plagioclase along the
liquidus boundary in this system is com-
parable to that observed in natural ba-
saltic and andesitic magmas.
Two examples of three-phase triangles
determined partly with electron micro-
probe analysis are shown in Fig. 51B.
They are nearly the same as those drawn
by Bowen (1915), except for composi-
tions of diopside solid solutions that are
projected from silica onto the join Di-An.
The results of the present study do not
alter the crystallization paths in this
system shown by Bowen (1915) ; the
NaAISi308
50
Weight per cent
CaAI2Si208
Fig. 52. Temperature-composition relation of plagioclase crystallized along the liquidus bound-
ary (or coexisting with diopside solid solution and liquid) in the system diopside-anorthite-albite.
The liquidus curve is projected onto the join anorthite-albite from diopside.
GEOPHYSICAL LABORATORY
507
solidus fractionation trend during frac-
tional fusion shown by Presnall (1969),
however, needs significant change; soli-
dus fractionation lines no longer have
inflection points. The results suggest
that by crystallization of Ca-rich clino-
pyroxene, basaltic magmas whose com-
positions lie on or close to the plane of
silica saturation fractionate toward
silica-saturated compositions and that by
extreme fractionation such magmas
trend toward pantelleritic compositions,
as suggested by Bowen (1945) and Yoder
and Tilley (1962, p. 416).
METAMORPHIC PETROLOGY
Infiltration Metasomatism in the
System K20-Al203-Si02-H20-HCl
J . D. Frantz and A. Weisbrod
The use of quantitative mass transport
models can be of great assistance in un-
derstanding the important processes and
controls of the evolution of nonisochem-
ical rock systems. Models involving dif-
fusion of solutes through a static solvent
have successfully been used to explain
mineral textures that occur at meta-
morphic isograds (Carmichael, 1969).
Fisher (1970) used similar theories in ex-
plaining mineral segregations. A quite
different model, infiltration metasoma-
tism (Korzhinskii, 1936) , involves trans-
port of material by the flow of the solvent
in response to gradients in fluid pressure.
Such a process may provide a mechanism
explaining the transfer of material over
distances greater than one might expect
from diffusion. In spite of its potential
importance, however, the model has re-
ceived little attention in the western
literature, possibly because of some con-
fusion concerning its fundamental prin-
ciples and their application to rock
systems. This report is an attempt to
illustrate the use of infiltration meta-
somatism and to investigate quantita-
tively its potentials and weaknesses in
explaining long-range transport. The
character and sizes of the reaction zones
resulting from the alteration of a pyro-
phyllite-quartz rock by the infiltration of
a potassium-rich solution were calculated
at 500 °C and 1000 bars fluid pressure,
using data presented by Hemley (1959)
for the system K20-Si02-Al203-H20-
HC1.
Infiltration metasomatism, as ex-
plained by Korzhinskii (1965, 1970) and
Hofmann (1972), involves reactions be-
tween rock-forming minerals and aque-
ous solutes supplied by the movement of
an inflowing fluid. In many cases several
alteration zones of varying thickness and
mineralogy are formed. Assuming local
equilibrium and neglecting diffusion (i.e.,
dealing with pure infiltration as a theo-
retical end member of the transport proc-
esses), it can be demonstrated that (1)
the boundaries — or "fronts" — of the zones
are sharp (progressive variations of com-
position cannot occur in the rock) , and
(2) progressive variations of composition
cannot occur in the solution (gradients of
concentrations take place only at the
fronts and are infinite) .
It is possible to identify these assem-
blages of the zones and calculate the
relative rates of progression of their
fronts using the simplified expression:
dz/dv = -j (AC// AC/) =
j (AC// AC/) •
(1)
where dz/dv refers to the rate of pro-
gression of a front as a function of the
amount of fluid, by volume, passing
* The actual expression should also consider
changes in the amount of pore fluid and its
concentration.
508
CARNEGIE INSTITUTION
through a unit area of rock, A* The terms
AC/ = (CV - CJ) and AC/ = (CV -
C,/) refer to the changes in concentra-
tion of component i at the front between
the zones a and b in the fluid and in the
rock, respectively. Comparison of the
relative velocities of all possible fronts
indicates the expected sequence of reac-
tion zones and their relative thicknesses.
Concentration Units
The choice of units for the concentra-
tions in the fluid and in the rock is some-
what arbitrary because ratios of differ-
ences rather than absolute amounts are
important. The present calculations use
molality for fluid concentrations and the
number of moles per 1 cal/bar of rock
for rock concentrations. For example, a
rock formed only of muscovite with no
porosity would have silicon and potas-
sium concentrations calculated as fol-
lows:
Qsi = 3/ V mu
(2)
where FMu refers to the molar volume of
muscovite, KAl3Si3Oio(OH)2.
Rock Concentration Diagrams
The concentrations of the three com-
ponents in a rock composed of one or
more minerals can be represented on a
Cf—Cf—C/ concentration diagram of
three dimensions. A diagram for the
solid phases in the K^O-SiCVA^Os-H^O-
HC1 system is given in Fig. 53. The axes
CH2or and CHc/ are not present because
(1) H20 is an excess component and (2)
chloride is assumed not to be present as
a component of any of the solid phases.
Points Q, Pyr, Or, and Mu, representing
rocks composed entirely of quartz, pyro-
phyllite, orthoclase, and muscovite, re-
spectively, were calculated in a manner
similar to that shown in equation 2.
* The filtration coefficients (Korzhinskii,
1970), the meaning of which seems somewhat
questionable, are neglected here.
Fig. 53. Diagram showing concentrations of
silica, potassium, and aluminum in a rock com-
posed of one or more of the following minerals :
orthoclase (Or), quartz (Q), muscovite (Mu),
and pyrophyllite (Pyr). See the text for dis-
cussion of the units, p (origin) = 100% poros-
ity. A plane of constant Al concentration
(shown by the dashed lines on the C±\r-CvLr
and Cwr-C sir planes) cuts the polyhedron
along a polygon, shown by the crosshatched
area.
In the treatment of regional meta-
morphism, it is generally assumed that
volume constraints do not exist and that
pressure, rather than volume, is the im-
portant independent parameter. Thomp-
son (1970) stressed that although this
point of view is quite realistic, it is still
an assumption. In relatively small and
space-limited nonisochemical systems,
control of volume may be likely. In
order not to be limited by either assump-
tion, it is interesting, though rather un-
usual, to take the porosity of the rock as
one of the parameters describing the
state of the system. The origin p in Fig.
53 represents 100% porosity. The com-
position of a rock of any porosity com-
prising one or more of these minerals can
be represented by a point on or within
GEOPHYSICAL LABORATORY
509
the polyhedron Q-Pyr-Mu-Or-p. For
example, a porous rock containing quartz
and pyrophyllite lies on the p-Q-Pyr
plane; a nonporous rock of muscovite
and orthoclase lies on the Mu-Or line.
Because of the low solubility of alumi-
num in many supercritical aqueous fluids
(Morey and Hesselgesser, 1951), the alu-
minum content in the rock can be ex-
pected to remain approximately constant
during metasomatic processes. Consid-
ering a given rock volume including the
pores, the concentration of aluminum
(CA{) is also constant. Changes in CSir
and CKr can be represented on constant-
CA{ sections as shown in Fig. 53. Values
of CA{ can be calculated from the com-
position of the starting rock, using the
initial porosity and quartz/pyrophyllite.
Figure 54 shows two sections: (a) CA{
= 0.28; (b) CA{ = 0.52. The lines in
Fig. 54 represent the intersections of the
Fig. 54. CsZ-CV constant-aluminum sec-
tions. In section a, Ca\t = 0.28; section b, Cm*
= 0.52. See Fig. 53 for identification of mineral
labels. See text for explanation of units.
planes (Q-Pyr-p, Q-Mu-p, Q-Or-p, Q-
Mu-Or, Q-Pyr-Mu, Mu-Or-p, and Pyr-
Mu-p) in Fig. 53 with a plane perpendic-
ular to the CA{ axis at its calculated
value. In Fig. 54 (a) the aluminum con-
centration (CA{ = 0.28) is low enough
so that all planes are intersected. In Fig.
54 (b) (CMr = 0.52), however, the plane
Q-Or-p is not, indicating insufficient
silica in the original rock for a quartz-
orthoclase alteration zone to appear.
Therefore, using constant-aluminum sec-
tions, the original porosity and quartz/
pyrophyllite, one is able to follow changes
in the rock composition during infiltra-
tion in terms of the remaining param-
eters: CKr and CSir.
Solution Concentration Diagrams
Owing to the conservation of mass re-
striction, the relative changes in the CKr—
CSir sections must be identical with the
relative KC1 and SiC>2 changes in the
fluid at the fronts. This can be seen by
rewriting equation 1 :
AC
7ACsr
ACk
kciVaCW
(3)
Thus, information concerning the rela-
tive fluid concentrations is necessary to
predict reaction paths on the constant
aluminum section of Fig. 54. The silica
and potassium concentrations in equilib-
rium with the mineral pyrophyllite,
muscovite, quartz, and orthoclase are
represented in Fig. 55. A constant 1
molal chloride concentration is assumed.
Points A and C represent the following-
two sets of reactions:
3 Pyr + 2 KC1 = 2 Mu + 6 Q + 2 HC1
(4)
Quartz = Si02(aqueouS) (5)
Mu + 2 KC1 + 6 Q = 3 Or + 2 HC1
(6)
Quartz = Si02(aqueous) (7)
The curve C-n represents the equilibrium
Mu + 2KCl + 6Si02(aqueous)
= 3 Or + 2 HC1. (8)
510
CARNEGIE INSTITUTION
^Si
1 1 1
T=500°C
1 1 1 1 1
1 1
1 1
0.10
— Pf=l Kbar
—
-
b/
r
0.08
— Pyr + Q
Mu + Q
o7
0.06
Ay
6 +
i i
°l -
0.04
Quartz saturation Pr = Pf
DjK
Pyr
Mu
Dr
1 1 1
, I.I,
1 .
1 , n\
0.88
0.90
0.92
0.94
0.96
0.98
1.00 C
Fig. 55. mKci-mSio2(aq) solution-concentration diagram for 500°C, 1000 bars. See Fig. 53 for
identification of mineral labels. See text for identification of letter symbols.
The molality of silica (0.04) for quartz-
saturated solutions at 500 °C, 1000 bars
pressure (equations 5 and 7) has been
measured by Weill and Fyfe (1964) and
Anderson and Burnham (1965). The KC1/
HC1 ratios (8 and 148) for equations 4
and 6, respectively, are from Hemley
(1959). The equation of the curve C-n
(8) has been calculated assuming ideal
solution.
Predicted Reaction Paths
Using the data presented in Figs. 54
and 55, it is possible to predict the re-
action paths by which a porous quartz-
pyrophyllite rock may proceed if infil-
trated by a quartz-saturated, 0.998 molal
KC1 solution. For a rock with CMr =
0.28, four reaction paths are possible, de-
pending on quartz/pyrophyllite (Fig.
56).
Path 1 (p = 0.5; CSir = 0.7). The
path represents the alteration of a rock
with ample porosity but insufficient
quartz for conversion of the quartz-
pyrophyllite assemblage to quartz-ortho-
clase. The first front (Pyr-Q-p -» Mu-
Q-p) has quartz-bearing assemblages on
either side, indicating no change in Cs/
and thus CSir. It is represented by a
horizontal line (^4iCi, Fig. 56). Because
insufficient quartz exists to convert all the
muscovite in zone C\ to orthoclase, two
additional fronts occur: CiF1} the partial
conversion of the muscovite in zone C\ to
orthoclase; and F^Ei, conversion of the
remaining muscovite in this zone (F^ to
orthoclase, using silica from the infiltrat-
ing solution (as in equation 8) .
The values of CSir and CKr shown for
zone F\ were calculated as follows. In
accordance with equation 3, expressing
conservation of mass, the slope of line
F-E (Fig. 55) must equal that of F1E1 in
Fig. 56. The slope of F-lE-l and the initial
solution (point E) are known, and as
zone F1 contains the assemblage Mu +
Or, the solution composition of F± must
lie on curve C-n at point F (Fig. 55).
Again, owing to equation 3, the slope of
dFi (Fig. 56) must equal that of CF
(Fig. 55), thus defining the silica and
potassium concentrations in zone F\.
With this information (summarized in
Fig. 57(1) ), it is possible to calculate the
relative velocities of the three fronts
using equation 1 (see Fig. 58(1) ) :
GEOPHYSICAL LABORATORY
511
SsOr+Q
(4) Pyr + Q+p
Pyr+Mu+Q
Mu+Q
(3) Pyr+Q+p
Mu+Q+p
Mu+Q+Or
(2) Pyr+Q+p
Mu+Q+p
Q + p+Or
(1) Pyr+Q+p
Mu+Q+p
Mu+p+Or
Or+Q
Q + Or
Or+p
Fig. 56. Reaction paths depicted on constant-aluminum section a of Fig. 54 (Cai* = 0.28).
Dashed lines with arrows indicate four possible reaction paths; subscript letters indicate reaction
zones (see text). The inset is an enlargement of the upper part of the figure.
(dz/dv)AlCl = (C-A)/(C1 - A,)K
(9)
(dz/dv)ClFl =(F- C)/(F1 - C1)K
= (F-C)/(F1-C1)&i (10)
(dz/dv)PlEl= (E-F)/(E1-F1)K
= (E-*F)/(E1-F1)m (11)
thus providing the following metasomatic
sequence :
Pyr-Q-p | Mu-Q-p | Mu-Or-p | Or-p.
The relative concentrations of the min-
erals and the porosity can easily be cal-
culated for each zone, using the concen-
trations of elements in the rock.
Path 2 (p = 0.28; CSir = 1.10). This
path represents the alteration of a rock
with both ample porosity and quartz for
complete conversion of the porous pyro-
phyllite-quartz rock to a porous ortho-
clase-quartz rock. The fronts A2C2
(equation 4) and C2E2 (equation 6) were
calculated in a manner similar to that of
front A-lC-l because quartz is present in
512
CARNEGIE INSTITUTION
all assemblages. The concentration of
silica and potassium in the rock and fluid
are represented in Fig. 57(2). Calcula-
tion of the relative velocities (Fig. 58 (2) )
indicates the following sequence of re-
action zones :
Pyr-Q-p | Mu-Q-p | Or-Q-p.
Path 3 (p = 0.12; CSir = 1.40). In
this reaction path, the starting rock has
enough quartz and porosity for the
quartz -pyrophyllite assemblage to give
muscovite-quartz as in fronts AiCi and
A2C2. There is, however, insufficient po-
rosity to balance the volume increase as-
sociated with the alteration of muscovite
+ quartz to orthoclase + quartz. (Line
A^CS would intersect the Or-Mu-Q line
if extended, Fig. 56). Three possibilities
regarding further reaction exist.
The entire rock may expand. On these
grounds, the assumption of constant vol-
ume is no longer valid, and thus the sec-
tions in Figs. 54 and 56 are inappropriate.
This situation could be approximated by
considering a section with a larger initial
porosity but the same Cxf/Cgf ratio,
Si
0.6 0.7 0.8 0.9
0.1 0.2 0.3
c'
1.00
0.96 —
cf
0.92 —i
0.88
1.00 —
0.96 —
KCl
0.92 — i
0.88
1-3 1.4 1.5 1.6
1.00
cf
0.96 —
0.92 -
0.88
cr
1-3 1.4 1.5 1.6
(4)
c//
.12
'Si
, ty* _
E®-
.08
.04
0.1 0.2 0.3
cr
Fig. 57. Changes in silicon and potassium concentrations in the solution (Cko/,Csi0 and in the
rock (CKr,C8ir) at the fronts for the four reaction paths. Letters indicate the reaction zones.
Dashed lines, silicon; solid lines, potassium.
GEOPHYSICAL LABORATORY
513
0.3
0.2 -
cL
K
o.i y
Or+j>
Or + Mu+p
(1
Mu + Q+p
Pyr + Q+p
-1.0 -0.5
Log(dz/dv)
0.3 r-Or + Q+p
0.2
C[
K
0.1
(2)
Mu + Q+p
Pyr+Q+p
-1.5 -1.0 -0.5
Log(dz/dv)
0.3 i—
0.2
C[
0.1 -
Or + Q
(3)
Mu+Or+Q
Mu+Q+p
1.5 -1.0 -0.5
Log (dz/dv)
Pyr+Q+p
°-3 TOr + Q
0.2 -
C[
0.1
Mu + Q
Pyr + Mu+Q
Pyr+Q+p
-1.5 -1.0 -0.5 0
Log (dz/dv)
Fig. 58. Plot of the relative front velocities (log dz/dv) versus concentration of potassium in
the rock (CV). See Fig. 53 for mineral labels.
thereby giving a reaction path similar to
that of path 1 or path 2.
A second possibility is that the volume
remains constant and the pores are filled
during the formation of part of the ortho-
clase. Then the flow of the infiltrating
fluid would stop, thus prohibiting the
formation of any reaction zones (includ-
ing A3CS).
The third possibility assumes constant
volume during the formation of ortho-
clase with the maintenance of sufficient
porosity for continued infiltration. The
existence of two fronts is implied: C3D3,
the partial conversion of muscovite to
orthoclase filling up the major volume of
pores; D3ES, the equal volume replace-
ment of the remaining muscovite with a
change in the quartz concentration. But
zone D3 would have too many phases
(three instead of two), a situation that is
impossible unless the system contains
an additional independent constraint.
Thompson (1955) and Korzhinskii (1970)
suggested that situations may exist where
the pressure on the solids could be greater
than that on the fluid. Such an eventual-
ity provides the extra-state parameter
for the existence of assemblage D3 and
the equal volume transition of muscovite
+ orthoclase -j- quartz to orthoclase +
quartz as indicated by front DSES.
Quantitative calculations of this model
involve the determination of the equilib-
rium concentration of potassium and
silica in the solution for solid pressure
greater than the fluid pressure. Consider-
ing equations 6 and 7,
(3ln(
Qkci/ Qhci
)/dPs)Pf>T =&V8/2RT
(12)
^ln asi02(aq)/3^>s )PfT =Vq/RT
(13)
514
CARNEGIE INSTITUTION
where VQ is the molar volume of quartz
and AVS is the volume change for the
solids in reaction 6. Integrating and
combining expressions 12 and 13:
(flSi02(aq)/fl°Si02(aq)) —
[ (okoi/ohoi) / (a°KciA°HCi) ] 2FQ/AFs
(14)
where a °Si02( aqueous) and (a0Kci/«0Hci)
refer to activities of silica, KC1, and HC1
at Ps = Pf* This equation expresses the
relative fluid concentrations of silica and
KC1/HC1 in equilibrium with the Mu-Q-
Or assemblage as a function of the pres-
sure on the solids. Line CD (Fig. 55)
illustrates this relation assuming a 1 molal
chloride concentration and ideal mixing.
The relative fluid and solid concentra-
tions of silica and potassium are calcu-
lated as follows. Owing to equation 3 and
the constant volume restriction, line DE
in Fig. 55 must have a slope equal to that
of D%Ez. Because point E is fixed by the
infiltrating solution and zone D contains
the assemblage Q-Or-Mu, the solution
composition in zone D3 must lie at the
intersection of lines DE and C-n on the
solution concentration diagram. Simi-
larly, line C3D3 (Fig. 56) must have a
slope equal to that of CD (Fig. 55).
Thus the CSir and CKr for zone D3 are
determined and one now has the neces-
sary information (Fig. 57(3) ) for the cal-
culation of the relative front rates (Fig.
58(3)), giving the following sequence of
reaction zones:
Pyr-Q-p I Mu-Q-p | Mu-Or-Q | Or-Q.
Path 4 (V = 0.04; CSir = 1.54). In
this case, the porosity is not sufficient
to convert the pyrophyllite-quartz as-
semblage to muscovite-quartz (a hori-
zontal line from A± intersects the Pyr-
Q-Mu line in Fig. 56) and the three pos-
sibilities discussed for path 3 exist. Con-
sidering the possibility of a solid pres-
* Equations 12, 13, and 14 imply a hydro-
static-like pressure on the solids and do not
take into account the stress energy of the min-
erals. Therefore, these equations are used only
as approximations.
sure different from that on the fluid, cal-
culation of the relative reaction-front
velocities was attempted in a manner
similar to that for path 3. This entailed
determination of the equilibrium concen-
trations of the dissolved components
(Si02 and KC1) at various values of Ps
for the assemblages Pyr-Mu-Q and Mu-
Or-Q. These calculations were made using
equations similar to equation 14. As in
path 3, the volume constraints for the
transition of a Mu-Q rock to an Or-Q rock
(Fig. 56) require solution concentrations
of D and E (Fig. 55) in zones C4 and E4,
respectively (Fig. 56). Further, the tran-
sition of a Mu-Pyr-Q assemblage to
Mu-Q requires concentrations B and D
in zones B4 and D4 (slope ED = slope
E±D4] slope DB = slope D4£4). At the
front A4Z?4, the slope A4B4 in Fig. 56
must equal that of AB in Fig. 55, but
since the latter and A± are known, 2?4
may be located, thus providing the nec-
essary information to calculate the rela-
tive front velocities using equation 1
(Figs. 57(4) and 58 (4)). The following
is the predicted sequence of reaction
zones :
Pyr-Q-p I Pyr-Mu-Q | Mu-Q | Or-Q.
Conclusions
Diffusion metasomatism occurs as a re-
sponse to concentration gradients. Be-
cause these gradients decrease as the
thickness of the reaction zone (or zones)
increases, the process stops rapidly. Thus,
infiltration metasomatism seems to pro-
vide an adequate mechanism for long-
distance transport, for it depends on fluid
pressure gradients, which may remain un-
changed regardless of the thickness of the
reaction zones. However7 as has been
shown here, the occurrence during the
metasomatic evolution of reactions in-
volving positive volume changes puts
serious limitations on the applicability of
the infiltration model, unless one of the
following assumptions is accepted: (1)
The rock can expand. (2) The initial
GEOPHYSICAL LABORATORY
515
porosity is large enough to balance the
expansion of the solids (if this expansion
is important, the necessary minimum
value of the initial porosity would be
larger than expected in metamorphic
rocks). (3) The solid pressures can ex-
ceed the fluid pressure. It seems unlikely
that the situation involved in the third
assumption could exist without material
precipitation, thereby filling the pores,
releasing the nonhydrostatic pressure dif-
ferences, and stopping the flow of infil-
trating fluids. For instance, the solid
pressure on the zone B4 should be, at
500 °C and 1 kbar fluid pressure, around
7 kbar! This value is indeed not realistic
in rocks.
The example chosen may help others
to assess the applicability of the infiltra-
tion metasomatism model to the various
situations where metasomatic activity is
suspected to have occurred.
Cordierite-Garnet Equilibrium in the
System Fe-Mn-Al-Si-O-H
A. Weisbrod
Cordierite and garnet are among the
most common and interesting minerals of
metamorphic pelites. Several attempts
have been made over the last 10 years to
use one or both of them, in association
with other minerals, as a geothermometer
or geobarometer, or both. Until now,
none of these attempts actually yielded
results that could be used by field penol-
ogists. Almandine occurs mainly in high-
or intermediate-pressure types of meta-
morphism, even though it is stable at 1
atm, and cordierite in low-pressure (in-
cluding contact) types, though this min-
eral is known to be stable at pressures as
high as 10 kbar. Many investigators use
the presence or absence of these minerals
to characterize the type of metamorphism
when more reliable minerals (mainly the
Al2Si05 polymorphs) are not found. Sup-
port for using them for such a purpose
stems from recent experimental studies in
the Fe-Mg system: the lowest stability
pressure for the garnet + sillimanite +
quartz association at 650°-750°C is
about 3 to 4 kbar (Richardson, 1968;
Currie, 1971; Hensen and Green, 1971).
However, many occurrences of alman-
dine have been reported in low-pressure
gneisses, for instance, in the Hercynian
metamorphism of western Europe. It is
a very common mineral in the Cevennes
Medianes (southeastern part of the
French Massif Central), where the con-
ditions of the lower amphibolite facies
are about 650°-700°C and 2.5 to 3.5 kbar
(Weisbrod, 1968, 1970). In the latter
area, the garnet always contains a sig-
nificant amount of manganese (5 to 25
mole % of spessartine end member) and
almost no magnesium or calcium.
There is qualitative evidence that the
addition of manganese to the system en-
larges the stability field of garnet. Some
authors who used the mineral as a
rough geobarometer caution that only
garnets bearing less than 10% spessartite
can be used for this purpose (see for in-
stance Miyashiro, 1961; Winkler, 1967).
This value is arbitrary and cannot be
defended.
The only quantitative determination
of the "manganese effect" concerns the
low-temperature side of the stability field
of garnet and is useful for the greenschist
facies (chlorite-garnet equilibrium, Hsu,
1968). Because cordierite is one of the
most usual "counterparts" of garnet in
Si- and Al-rich rocks, the reaction
3 cordierite <=± 2 garnet +
4 sillimanite + 5 quartz (-f- H20)
has been experimentally studied in a
part of the Fe-Mn-Al-Si-O-H system at
750 °C and in the pressure range, 0.5-3.5
kbar.
Experimental Procedures
Reactants used in this study include
natural quartz (Lake Toxaway, 99.96%
Si02), natural sillimanite (Brandywine
Springs, traces of iron), and synthetic
garnets and cordierite. Garnets of vari-
516
CARNEGIE INSTITUTION
ous Mn contents (0, 10, 33, 50, 66, 100
mole % spessartite) were synthesized
with excess water in gold capsules in
cold-seal pressure vessels at various pres-
sures and temperatures ; the starting ma-
terial was a mix of Fe and Mn oxalates,
yAl203, and quartz. Grain size was large
(10-50 mm) for the end members but
always very small for the intermediate
members (1-8 fxm) . For these composi-
tions, products were rerun to increase
grain size. Nevertheless, the nucleation
rate of these garnets is so high that many
new minute garnets were formed, and the
seeds grew only up to 10-15 /mi. Crystal-
lization of Fe cordierite from a reduced
gel or glass in unbuffered runs always
yielded a large amount of hercynite.
Satisfactory results were obtained by
mixing Fe-cordierite glass, Fe oxalate,
7AI2O3, fired quartz, and a small amount
of sillimanite, with excess water. This mix
was run in gold capsules in a cold-seal
pressure vessel at 750°C and 1.1—1.3 kbar
for 5 days. The oxygen fugacity was
kept low by replacing a part of the filler-
rod by an open gold tube filled with iron
wire. Attempts to synthesize cordierite
bearing at least 10% of Mn end member
failed; these runs yielded either garnet
+ cordierite or fayalite + cordierite.
Experiments on the above reaction
were carried out in cold-seal pressure
vessels at 750° ± 5°C. Oxygen fugacity
was controlled by the quartz-iron-fayal-
ite external buffer. In fifteen runs, mixes
of previously synthesized Fe cordierite
and garnet, natural sillimanite, and
quartz were used. An estimation of the
equilibrium composition of the garnet
was calculated first, assuming no Mn in
the cordierite and a 3 kbar equilibrium
pressure (at 750 °C) for the pure Fe end
members (Richardson, 1968). Then, the
composition of the starting garnet was
chosen in order to reach the calculated
value from both sides. Because Fe-Mn
cordierite was not available, the same
approach was not possible for this min-
eral.
Five experiments were made using
mixes of Fe-cordierite glass, Fe and Mn
oxalates, sillimanite, and quartz, seeded
with Fe cordierite and Fe-Mn garnet.
Results of eight runs made for the pur-
pose of synthesizing Fe-Mn cordierite
also yielded interesting values. All the
experiments were carried out in the pres-
ence of an excess of water (approximately
5wt%).
Garnets of sufficient size for analysis
with the electron microprobe could be
synthesized only in one run, in which the
initial garnet composition was Alm90-
Spi0. Unfortunately, these crystals ap-
peared to be quite inhomogeneous. In all
other runs, grain size was too small
(<10 /mi) for electron microprobe analy-
sis. Therefore, garnet composition was
determined by means of powder x-ray
diffraction and refractive index. Neither
is considered very accurate: the former,
because many peaks of the garnet are
similar to those of quartz, sillimanite,
and cordierite; the latter, because the
range of refractive index in the Fe-Mn
series is very small (1.800-1.830, about
one-fourth of the range of the Fe-Mg
series, 1.715-1.830). Thus, the equilib-
rium composition of the garnet, Mn/(Mn
+ Fe) atomic ratio, cannot be given with
an uncertainty of less than 0.05.
Cordierite usually occurred with a
sponge-like habit. In only two runs were
grains suitable for analysis with the elec-
tron microprobe and reliable results ob-
tained. In the other runs, powder x-ray
diffraction and refraction index methods
were not suitable either.
Results
Compositions of coexisting garnet and
cordierite produced at 750 °C in the two
types of runs discussed above are plotted
in Fig. 59. The equilibrium pressure at
750 °C for the Fe end members (3.3 ±
0.1 kbar) is slightly higher than the
value found by Richardson (1968), 2.95
± 0.25 kbar. This small difference is
discussed elsewhere in this Report.
GEOPHYSICAL LABORATORY
517
P, bars
3500
3000
2500-
2000-4
1500-
iooohJ!
V
500 -^>
0
Fe
0.2 0.4 0.6
Mn/(Fe + Mn)
Mn —
Fig. 59. Composition of coexisting cordierite and garnet at 750 °C. Circles: starting composition
of garnet (filled) and cordierite (open) in runs using the association cordierite-garnet-sillimanite-
quartz as starting material. Squares: bulk composition of the runs using oxide mixes as starting
material; open, growth of cordierite; filled, growth of garnet; half-filled, growth of both cordierite
and garnet. Arrows and bars: resulting composition of coexisting garnet and cordierite, respec-
tively. Striped areas: range of uncertainty. Heavy dashed line: calculated curve, assuming no
H20 in cordierite.
The addition of Mn to the system pro-
duces garnet at much lower pressures.
Such garnet contains up to 65 mole %
spessartite. Coexisting cordierite always
contains a very small amount of Mn (less
than 5% at 750°C). This value agrees
with analyses of natural cordierites
(Leake, 1960).
The theoretical curve in Fig. 59 has
been derived from the following reaction,
in which water is assumed not to partici-
pate,
3 cordierite ^± 2 garnet +
4 sillimanite -j- 5 quartz.
Cordierite and garnet solid solutions are
assumed to be ideal. If PFe and PMn are
the equilibrium pressures at 750 °C for
the end members, X and Y are the molar
fractions of iron in garnet and cordierite,
respectively, and (AVFe) and (A7Mn) are
the volume changes in the reaction above
for Fe and Mn end members, then the
following relations are applicable:
X_ (AFFe)
Y- 6RT
tf To L
(1)
In
1— X
1 — Y
(A7Mn)
6RT
(Pm„ - P)
(2)
With the data (AFFe) = —3.95 cal/bar,
(AVMn) = —4.09 cal/bar (Weisbrod,
this Report); PFe = 3300 bars; Y =
518
CARNEGIE INSTITUTION
0.96 at P = 500 bars, then the values
PMn = —7636 bars and X — 0.39 at P =
500 bars (experimental value 0.45 ± 0.05)
are obtained. They can be substituted
into equations 1 and 2 which yield
lny =
-3.95
6 X 1.987 X 1023
(3300 — P) (3)
In
1-X
1-7
4.09
6 X 1.987 X 1023
(_7636-P). (4)
These equations are used to calculate X
and Y at 750 °C and various pressures.
Although the curve defined by these X
and Y values is close to that found
experimentally, it lies outside the uncer-
tainty range. It is believed that this dif-
ference can be explained by the occur-
rence of water in cordierite, rather than
by a nonideal behavior of garnet or
cordierite, or both, in the Fe-Mn system.
This important point will be discussed
further in this Report.
Application
The discovery of the stability of Mn-
bearing garnets at pressures much lower
than the stability range of Fe-Mg gar-
nets has important application to meta-
morphic petrology. As an example, con-
sider a pelitic rock containing 2% garnet,
excess quartz, and sillimanite and having
a bulk composition with 0.20 wt % MnO,
a typical value. Because Mn is domi-
nantly contained in garnet, this mineral
would contain about 25 mole % spessar-
tite. At 750 °C, the minimum stability
pressure for Mn-free garnet coexisting
with quartz and sillimanite is 3.3 kbar.
Garnet containing 25 mole % spessartite
could be stable, however, to 2.1 kbar.
Thus, the minimum estimated pressure
for the rock can be lowered by 30% by
addition of a small amount (0.20 wt %)
of MnO. Moreover, in the same rock, a
divariant assemblage containing cordier-
ite and garnet can occur at any pressure
below 3.3 kbar.
These preliminary results support the
view that garnet and cordierite assem-
blages must be used with caution as
quantitative indicators of conditions of
metamorphism.
Refinements of the Equilibrium Con-
ditions of the Reaction Fe Cordierite
^± Almandine -f- Quartz +
Sillimanite (+ H20)
A. Weisbrod
The system Fe-Al-Si-O-H has been
carefully studied by Richardson (1968).
In that system, the stability field of Fe
cordierite is limited at high pressure by
the reaction
3 Fe cordierite ±± 2 almandine +
4 sillimanite + 5 quartz (+ H20) .
During the course of a study of the sys-
tem Fe-Mn-Al-Si-O-H, the equilibrium
pressure at 750 °C for this reaction was
found to be about 400 bars higher than
Richardson's determination.
In order to understand this small dis-
crepancy, refine the previous results, and
calculate the thermodynamic properties
of cordierite, the reaction was studied
again over a larger temperature range,
575°-800°C.
Starting material was a mixture of
synthetic almandine, synthetic Fe cordi-
erite, natural sillimanite (Brandywine
Springs) , and natural quartz (Lake Toxa-
way) . The phases were mixed to give the
same bulk composition to each side of the
reaction, that is cordierite composition
plus a small excess (5 wt %) of quartz
and sillimanite to prevent metastable
formation of hercynite (see Richardson,
1968). The mixture was ground under
acetone in an agate mortar to a grain size
of 1-5 /xm. Each charge contained excess
water (approximately 5 wt %) .
Experiments were carried out in cold-
seal pressure vessels, and the oxygen
GEOPHYSICAL LABORATORY
519
fugacity was controlled by the quartz-
iron-fayalite (QIF) external buffer. It
was not possible to preserve the buffer for
more than 5 days at temperatures above
700 °C. Thus, these experiments were run
for 3-5 days, then quenched. The buffer
was replaced and the charge rerun for
another 3—4 days. In three experiments
the buffer did not contain iron at the end
of the run. Because no iron oxide was
detected, however, it is assumed that the
oxygen fugacity increased only slightly
above the QIF buffer.
Each whole charge was analyzed by
the powder x-ray diffraction method be-
fore and after the run. Changes in the
relative intensities of the cordierite and
garnet peaks were used to determine the
direction of reaction.
Results are presented in Fig. 60. The
equilibrium curve is essentially a straight
line, with a negative slope of — 3 ±1.5
bar/°C, passing through the point 750°
± 5°C and 3300 ± 100 bars. Derivation
from Richardson's experiments is rather
small, although the slope of his curve
(—8 to —9 bar/°C) is about three times
greater than the slope determined here
but with the same sign. Because the ex-
periments were carried out under the
same conditions, except for oxygen fu-
gacity (QMF versus QIF), the difference
may be related to the oxidation state.
The reaction, as written above, does not
involve oxygen. Moreover, according to
Hsu (1968), presence of Fe3+ in the gar-
net is unlikely under these condi-
tions. Analyses of natural iron-rich cor-
Fig. 60. Comparison of experiments defining the high-pressure stability limit of Fe cordierite
(quartz in excess). Richardson's experiments (1968) (QMF buffer): filled circles, cordierite; filled
squares, almandine + sillimanite; dashed lines, stability field boundaries. This work (QIF buf-
fer) : circles, cordierite growth; squares, garnet growth (open, large amount of growth; with cross,
slight or moderate amount of growth) ; triangles, no reaction. The low-temperature side of the
equilibrium curve (heavy line) is metastable, probably below about 675°C.
520
CARNEGIE INSTITUTION
dierites (Leake, 1960), however, show
significant amounts of this component.
Similarly, Iiyama's experiments (1960)
show that iron-bearing cordierite is easily
oxidized. Other workers (Rutherford,
1970; Keesmann et al., 1971) also believe
ferric iron may be important to the sta-
bility of cordierite. The amount of Fe3+
in iron-rich cordierites synthesized at
various oxygen fugacities could be ascer-
tained using the Mossbauer technique.
According to Richardson's results, the
curve obtained here should be metastable
below 675 °C, inasmuch as the reaction
occurs in the staurolite stability field.
Staurolite did not appear in the runs,
however, probably because of the low
nucleation rate of this mineral. Her-
cynite, which nucleates very easily, did
not appear in the runs carried out at
800 °C. Therefore, the high-temperature
side of the curve is probably stable, and
the stability field of hercynite + silli-
manite -j- quartz recedes toward the high
temperatures as the oxygen fugacity de-
creases.
The equilibrium curve in Fig. 60 may
be used to calculate the thermodynamic
parameters of the reaction
3 cordierite • nH20 — » 2 garnet +
4 sillimanite + 5 quartz + 3nH20,
where n is the number of moles of water
per mole of cordierite, under given P and
T conditions. The values of n have been
determined by Schreyer and Yoder
(1964) for Mg cordierite and seem to be
appropriate for Fe cordierite (Weisbrod,
this Report). Using a slope of —3 bar/
°C, data of Burnham, Holloway, and
Davis (1970) for thermodynamic prop-
erties of water, and a volume change of
solid phases of —3.95 cal/bar (Weisbrod,
this Report), the calculated entropy
change (ASS) for the solid phases in the
reaction above varies from —56.6 ± 3
cal/deg at 850 °C and 3 kbar, to 58.6 ±
3 cal/deg at 650°C and 3.6 kbar.
Denning (ASg) = 2£alm + 4Ssill + 5Sg,
and using the data from Zen (1973) for
almandine and of Robie and Waldbaum
(1968) for quartz and sillimanite, a value
of 278 cal/deg is obtained for
(AS,,)1-29815. Because of the following
relations,
(A&s) = {ASg) — 3£Cord-nH20,
(AS'S) = (AS,) - 3Scora,
it is possible to calculate the third law
molar entropy of low anhydrous cordie-
rite with the aid of one of the following
assumptions
$cordnH20 'T = $cordP'T -|- 7lkSu20P'T and
a(^.)=0tagw=a (1)
dT ' dP
Therefore
(AiS/.)1-298-l8= (AS'S)P<T =
(ASs)p'T-SnkSu2op'T)
a 1,298.15 —
'-'cord —
(as,)1-298-15— (AiSr.)1'298-15
The coefficient k represents the contri-
bution of 1 mole of water to the molar
entropy of hydrous cordierite. With
k = 0.85 (Newton, 1972) , the calculated
value is
Scord1'298'15 = 96 cal/deg -mole
C 1,298.15 C 1,298.15 _|_ ™CO
£Wd«H20 — £>cord T n*^ H20
and m± = o,^=o. (2)
dT
dP
Thus
(A&) 1,298.15 = (&Sa)P,T.
Q r ^1,298.15
«H20
(AS-)1'298-16— (AS,)1-298-16
and
,C 1,298.15 C 1,298.15 «,Qo
^cord — £>cord-nH20 — «*J H20-
The term £°H2o is the fictive entropy of
1 mole of water in cordierite at 25 °C and
1 bar, and its value of 17 cal/deg can be
GEOPHYSICAL LABORATORY
521
estimated (Zen, 1973) , assuming cordie-
rite has zeolitic water. The resulting
value for ^Scord1'29815 is 101 cal/deg-mole.
Both of those values (96 or 101 cal/
deg-mole) are lower than anticipated,
compared with the third-law entropy (97
cal/deg-mole) of Mg cordierite (Weller
and Kelley, 1963). Because the struc-
tural and hydration states of the sample
used by Weller and Kelley for their heat-
capacity measurements are not known,
no formal discussion of this discrepancy
can be undertaken.
The Problem of Water in Cordierite
A. Weisbrod
Metamorphic reactions involving cor-
dierite must now take into account the
water content of this mineral. The struc-
tural location of this water still remains
a problem: (H404)4" replacing (Si04)4~
in tetrahedra (Iiyama, 1960) or, more
likely, zeolite-like molecular water,
loosely bound in the channels of the min-
eral (Schreyer and Yoder, 1964; Newton,
1972) . Whatever the location, the impor-
tant question for the petrologist is not
"where?" but "how much?"
The most obvious way to answer this
question is to run cordierite with an
excess of water at a given temperature
and pressure, quench the run, and then
measure the amount of water in the cor-
dierite. This procedure has been suc-
cessfully used by the authors cited above.
Hydration and dehydration of cordierite
unfortunately seem to be fast processes,
and some reequilibration probably occurs
during the quench.
It is also possible to calculate the
amount of water in cordierite by the fol-
lowing method. Take the reaction, for
instance,
3 cordierite • nH20 +± 2 garnet +
4 sillimanite + 5 quartz -f- 3nH20.
The change in the Gibbs' free energy is
- (AGS)°T+P(&VS) + 3n^H2o,
where the subscript s refers to solid
phases.
For a divariant equilibrium involving
two exchangeable metals M (e.g., Mg)
and F (e.g., Fe)
6#Tln-^
(aG)tp
(2)
where X and Y are the molar fractions of
F in garnet and cordierite, and yy and yc
are the activity coefficients of F in garnet
and cordierite. At a given temperature
and pressure, yg and yc are functions of
X and Y.
So, at a given temperature T, it can be
written from equations 1 and 2
n = /(P,^H2o,X,Y)
(3)
Next the following assumptions have
to be made.
1. Water is the only constituent of the
fluid phase:
P,T
(aGV
P — PuoO ) /^H20 ' — Gh
2. Solid solutions are ideal between M
and F end members, in both garnet and
cordierite :
yc=yg — 1-
Even if this assumption is not completely
true, the departure from ideality must be
very small in these minerals as far as
Fe-Mg or Fe-Mn substitutions are con-
cerned.
3. The water content of cordierite is
the same for both end members and does
not depend on the M-F composition. This
assumption is important only for the
Mg-Fe system (in the Mn-Fe system,
only a very small amount of Mn enters
cordierite). The cell parameters of Mg
and Fe cordierites are very close to each
other, and neither analyses of natural
cordierites (Leake, 1960) nor experi-
mental work (Iiyama, 1960) shows any
systematic variation of the water con-
tent with the composition.
With these assumptions, equation 3
gives
(i)
n = j(P,Gn,oP'T,X/Y)
(4)
522
CARNEGIE INSTITUTION
At a given T and any P, GU2op'T can be
calculated and X/Y experimentally de-
termined, so it is possible to calculate n
and the relation n = /(P). If Pf is the
equilibrium pressure at T of the reaction
for the pure F end members, the relation is
n
QRT In (X/Y) - (AY.) (PF - P)
3(G^-G^')H2o
(5)
Equation 5 has been used to calculate
n at 750°C, between 0.5 and 2.8 kbar.
The mole fractions X and Y are taken
from experimental results obtained in the
Fe-Mn system (Weisbrod, this Report) ;
PF = 3300 bars (Weisbrod, this Report) ;
(GPf>t — Gp'T)H2o was found in tables
(Fisher and Zen, 1971) ; molar volumes
of quartz (22.7 cc) and sillimanite (69.9
cc) are from Robie, Bethke, and Beards-
ley (1967).
Skinner's (1956) and Hsu's (1968) de-
terminations of almandine cell param-
eters agree and yield a 115.3 cc molar
volume. Hsu's data (1968) for low Fe
cordierite (237 ± 1 cc), however, differ
from Eberhard's data (1962) (232 ± 0.4
cc) . It is noted that the latter is not con-
sistent with the values of high Fe cordie-
rite, and low and high Mg cordierite. For
this reason, the molar volume of low Fe
cordierite has been recalculated, taking
into account (1) the volume difference
between Mg and Fe cordierites in a given
structural state and (2) the volume dif-
ference between high and low cordierite
of the same composition. A value of
236.5 cc, which agrees with Hsu's data,
is obtained. Then, (AY,) for the reaction
mentioned above is about — 165 cc or
— 3.95 cal/bar. (The same calculation
for the Mn reaction gives a value of
—4.09 cal/bar, which compares favor-
ably with the value from Hsu's data of
—4.12 cal/bar.)
The results are shown in Fig. 61. No-
tice is taken of the very large uncer-
tainty, increasing when P approaches PF.
This uncertainty is expected, as the ratio
of two inaccurate values of X and Y is
the basis for the calculations.
Using smoothed values of X and Y
gives results that agree with Schreyer
and Yoder's measurements (1964), ex-
cept perhaps at very low pressures. A
part of this difference may result from
some partial rehydration of cordierite
during the quench of their low-pressure
runs. The calculated cruve can be
extrapolated down to n = 0 and is in
agreement with Iiyama's experiments
(1960) at low pressure.
(1) .— (2) -AH3) D(4) J^(5)
i
-a-
8
10
P.kbar
Fig. 61. Variation of the number of moles of water (n) in cordierite as a function of Ph2o =
Ptotai at 750°C. (1) Iiyama (1960), natural cordierites (direct). (2) Schreyer and Yoder (1964),
synthetic Mg cordierite (direct). (3) Currie (1971), synthetic Mg-Fe cordierites (calculated). (4)
Newton (1972), synthetic Mg cordierite (direct and calculated). (5) This work; synthetic Fe-
(Mn) cordierites (calculated).
GEOPHYSICAL LABORATORY
523
The same method has been used to
calculate n at 5, 6, and 7 kbar. The
values of X and Y are interpolated from
Currie's experiments (1971) in the Mg-
Fe system. The location of the points in
Fig. 61 is baffling. Although Currie's runs
were carried out with excess water, the
amount of this component in cordierite is
extremely low (less than 0.1 mole H20
per mole of cordierite) and decreases
with increasing PH2o.
The experiments of Hensen and Green
(1971) and Hensen {Year Book 71) were
carried out at much higher temperatures,
Ph2o < -Ptotai, and with a composition
more complex than the theoretical Fe-
Mg system (presence of Ca, Na, K) .
Therefore, their results cannot be extra-
polated and used here.
In conclusion, this study is consistent
with the direct experimental results of
Schreyer and Yoder (1964) in regard to
the water content of cordierite at various
pressures and temperatures. That is, the
amount of water is probably the same in
Mg and Fe cordierites, at least at 750°C
and between 0 and 3 kbar. The method
proposed here assumes that the experi-
mental results are not influenced by re-
actions such as dehydration or rehydra-
tion that may occur during the quench,
and represent actual equilibrium values.
Effect of Pressure on the Composition
of Coexisting Pyroxenes and
Garnet in the System
CaSiOa-MgSiOa-FeSiOa-CaAlTiaOe
Jagannadham Akella and F. R. Boyd
Diopsidic clinopyroxene and enstatitic
orthopyroxene coexist with pyrope-rich
garnet in some eclogites and garnet pe-
ridotites. In most petrological models of
the upper mantle, garnet peridotite is
considered as a major rock type. Rutile-
and ilmenite-bearing eclogites are com-
mon from many metamorphic terranes
and also as xenoliths in kimberlites. For
an understanding of upper mantle
mineralogy, it is essential to investigate
the compositional variations of pyrox-
enes, garnets, and Ti-oxide phases as a
function of pressure and temperature.
Boyd and England {Year Book 63, pp.
157-161) and MacGregor (1973) studied
the simple MgO-Al203-Si02 system as a
first approximation in determining the
conditions of formation of ultramanc
rocks. Subsolidus phase relations in the
system CaSi03-MgSi03-Al203 were in-
vestigated (O'Hara, Year Book 62, pp.
116-118; Boyd, Year Book 67, pp. 214-
221; Boyd, Year Book 68, pp. 214-221),
as well as those in a multicomponent
system using natural materials (Mac-
Gregor and Ringwood, Year Book 63, pp.
161-163; Green and Ringwood, 19676;
Hensen, this Report) . However, the
stability conditions and chemical com-
positions of the garnets coexisting with
ortho- and clinopyroxenes and also the
solubility of A1203 in these pyroxenes
may be affected by the presence of other
elements, such as Fe2+, Na, and Ti
(Banno, 1965; Kushiro, Syono, and
Akimoto, 1967).
For the present study two bulk com-
positions with the components MgO,
FeO, CaO, A1203, Si02, and Ti02 were
chosen (Table 9). These bulk composi-
tions have Mg/(Mg + Fe) ratios of
approximately 0.6 and 0.8. Details of
the analytical and experimental tech-
niques have been discussed by Akella
and Boyd {Year Book 71, pp. 378-384;
and 1973).
The subsolidus phase assemblage ob-
tained in the present experiments for the
Mg/(Mg + Fe) = 0.8 composition is
clinopyroxene + orthopyroxene + garnet
TABLE 9. Bulk Compositions
Mg/(Mg + Fe)
0.6 0.8
Si02
48.3 49.1
AI2O3
7.4 7.6
Ti02
5.7 5.7
CaO
12.9 13.4
MgO
11.9 16.3
FeO
13.7 7.9
524
CARNEGIE INSTITUTION
+ rutile + minor ilmenite; and for the
0.6 composition, clinopyroxene + ortho-
pyroxene + garnet + rutile. The com-
positions of garnets, clinopyroxenes, and
orthopyroxenes synthesized at 1100°C in
the range 30-45 kbar for both bulk
compositions are presented in Tables 10
and 11.
The compositions of Ca-rich and Ca-
poor pyroxenes synthesized in a number
of runs using the 0.6 and 0.8 starting
mixtures are plotted in the pyroxene
quadrilateral (Fig. 62) and are compared
with the Skaergaard trend (Wager and
Brown, 1968) and a pyroxene pair in a
garnet-pyroxene-olivine-ilmenite nodule
from a kimberlite pipe (Boyd and Nixon,
Year Book 71, pp. 362-373). The Ca
values used in these plots were not cor-
rected for CaAl2Ti06 and CaAl2Si06;
had this been done, the points in Fig. 62
would be shifted to a slightly lower Ca
value. Pyroxene pairs crystallized from
the 0.8 and 0.6 compositions plot close to
the Skaergaard trend. The synthetic
clinopyroxenes show a range of solid
solution toward orthopyroxene (Wo44-
Wo37) , which is due to a range in their
equilibration temperatures. The range
in Mg/(Mg + Fe) in the pyroxenes for
a given composition is caused by variable
reaction with the Fe capsules.
Orthopyroxenes crystallized from the
0.8 composition plot close to the Skaer-
gaard trend for orthorhombic enstatite.
On the other hand, the orthopyroxenes
crystallized from the 0.6 composition fall
on an extension of the Skaergaard trend
at a lower Ca concentration than the
Skaergaard pigeonites. This difference
can probably be attributed to the higher
pressure of equilibration (30-45 kbar) in
TABLE 10. Compositions of the Coexisting Pyroxenes and Garnets in the Subsolidus
Crystallized with H20 and Excess Si02 for Bulk Composition Having
Mg/(Mg + Fe) =0.8, T=1100°C
P,
30 kbar
P,
40 kbar
P,
45 kbar
Cpx
Opx
Cpx
Opx
Ga
Cpx
Opx
Ga
Si02
52.4
53.8
53.5
54.5
40.9
53.4
55.1
41.7
A1203
2.4
2.6
1.1
1.5
21.2
1.3
1.4
20.2
Ti02
0.5
0.5
0.2
0.3
1.7
0.4
0.5
1.6
CaO
18.9
1.3
19.1
1.4
7.6
19.9
1.8
8.8
MgO
17.9
29.8
18.2
29.1
14.9
17.6
29.1
14.4
FeO*
6.4
10.7
7.7
13.1
15.0
6.6
11.5
14.0
Totals
98.5 98.7 99.8 99.9 101.3
Number of Cations (X1000)
99.2
99.4 100.7
Si
1939
1927
1964
1948
2981
1968
1963
3051
Al
103
108
49
63
1821
56
61
1742
Ti
15
12
6
8
93
12
15
89
Ca
751
50
752
52
589
784
67
688
Mg
988
1590
997
1550
1620
964
1545
1564
Fe
198
319
237
391
912
203
341
855
Totals
3994
4007
4005
4012
8016
3987
3992
7989
Mg/(Mg + Fe)
0.833
0.833
0.808
0.798
0.640
0.826
0.819
0.647
Coexisting
phases
Cpx
+Opx+Gat+Hm
+Ru
Cpx+Opx+Ga+Ilir
i+Ru
Cpx+Opx+Ga+U
m+Ru
* Total Fe as FeO or Fe2+.
t Garnet is present in this run; however, analysis could not be obtained owing to inclusions.
Abbreviations used in tables: Cpx, clinopyroxene; Opx, orthopyroxene; Ga, garnet; Ilm, ilmenite;
Ru, rutile.
GEOPHYSICAL LABORATORY
525
TABLE 11. Compositions of the Coexisting Pyroxenes and Garnets in the Subsolidus
Crystallized with H20 and Excess Si02 for Bulk Composition Having
Mg/(Mg + Fe) =0.6, T= 1100° C
P, 30 kbar
P,
35 kbar
P, 40 kbar
Cpx
Opx
Ga
Cpx
Opx
Ga
Cpx
Opx
Ga
Si02
51.5
52.7
39.7
52.4
52.5
39.7
51.6
52.8
39.4
AI2O3
1.8
2.0
21.1
1.6
1.6
20.6
1.5
1.3
19.7
Ti02
0.5
0.4
0.95
0.4
0.2
1.6
0.5
0.4
1.5
CaO
17.5
2.0
6.6
16.4
1.7
6.8
17.6
1.7
7.4
MgO
14.6
22.4
10.8
14.6
20.2
8.9
14.7
21.6
9.3
FeO*
12.8
20.3
19.8
14.2
22.0
22.4
12.6
21.2
21.6
Totals
Si
Al
Ti
Ca
Mg
Fe
Totals
98.7 99.8 98.95 99.6 98.2 100.0
Number of Cations (X1000)
98.5
3992
3994
7982
3980
3971
7960
3992
99.0
3985
Coexisting
phases
98.9
1956
1951
3016
1975
1988
3024
1960
1975
3036
79
87
1894
70
70
1852
68
57
1793
13
12
54
10
6
90
15
12
85
711
79
534
660
69
557
715
70
611
827
1237
1223
818
1141
1009
834
1206
1067
407
628
1260
447
697
1429
401
665
1390
7982
Mg/(Mg + Fe) 0.670 0.663 0.493 0.646 0.621 0.414 0.676 0.645 0.430
Cpx+Opx+Ga+Ru+Ilm?? Cpx+Opx+Ga+Ru
Cpx+Opx+Ga+Ru
* Total Fe as FeO or Fe2+.
Abbreviations as in Table 10.
50
Atomic per cent
Fig. 62. Analysis of synthetic pyroxene pairs for both Mg/(Mg + Fe) = 0.8 and 0.6 bulk com-
positions, plotted in pyroxene quadrilateral. Skaergaard trend from Wager and Brown (1968)
Analyses of Matsoku pyroxenes from Boyd and Nixon (Year Book 71).
526
CARNEGIE INSTITUTION
ro
O
CVJ
c
o
Q.
CD
10
20 30
Pressure, kbar
Fig. 63. Weight percentage of A1203 in orthopyroxene coexisting with clinopyroxene and garnet
versus pressure. Open circles are orthopyroxene crystallized from a bulk composition with Mg/
(Mg + Fe) =0.6, and closed circles are orthopyroxenes crystallized from a bulk composition
with Mg/(Mg + Fe) = 0.8. Values in parentheses are temperature in degrees centigrade.
the present experiments (Akella and
Boyd, Year Book 71, pp. 378-384).
It has been shown by many authors
(e.g., Boyd and England, Year Book 59,
pp. 49-52; O'Hara and Mercy, 1963;
Banno, 1965) that orthopyroxene can
dissolve large quantities of A1203. Within
the pyrope stability field, increase of
pressure reduces the solubility limit of
A1203 in enstatite, because orthopyroxene
exsolves to form more dense garnet (Boyd,
1970). The A1203 content of the syn-
thetic orthopyroxene coexisting with
clinopyroxene and garnet for both the
0.8 and the 0.6 starting compositions is
plotted against pressure in Fig. 63. For
a given Mg/(Mg -j- Fe) ratio of the
original composition, the solubility of
A1203 decreases as pressure increases at
a constant temperature and also as tem-
perature decreases at a constant pressure.
The slopes of the A1203 isopleths in Fig.
63 are in general agreement with those
found by other workers for the system
MgO-Al203-Si02 (Boyd and England,
Year Book £3/ pp. 157-161; MacGregor,
1973) and for a multicomponent system
(Green and Ringwood, 19676).
In the present investigation the A1203
solubility in orthopyroxene at 30 kbar
and 1100°C is 2.6 wt % for the 0.8 com-
position and 2.0 wt % for the 0.6 com-
position, whereas under similar P-T
conditions for the pure Mg system, Mac-
Gregor (1973) reported a value of 3.3 wt
%. It is evident from Fig. 63 that with a
decrease in the Mg/(Mg + Fe) ratio,
the solubility of A1203 in orthopyroxene
decreases at constant temperature and
pressure. Thus a direct application of
data on the enstatite-pyrope join to a.
natural system will result in pressure
estimates that are too high.
GEOPHYSICAL LABORATORY
527
Pyroxenes and Garnets as
Geothermometers and
Barometers
B. J. Hensen
Most experimental work on pyroxenes
and garnets has been concentrated on the
study of simple three- or four-component
systems, modeling natural rock composi-
tions. Although the data obtained on
these simple systems can be usefully
applied to natural occurrences, more
information is needed on the influence of
additional components in order to con-
struct detailed and accurate petrogenetic
grids. The present study, using natural
starting materials, has been undertaken
to obtain data on (1) the orthopyroxene-
clinopyroxene two-phase region, (2) the
solubility of A1203 in enstatite, (3) the
partition of Mg and Fe2+ among garnet
and pyroxenes, and (4) the solvus be-
tween Ca-rich and Ca-poor (pigeonitic)
clinopyroxene.
Experimental Method
In view of the inherent uncertainties
associated with synthesis experiments in
these systems, the present experiments
were carried out on natural minerals that
were reequilibrated at high temperatures
in experiments of long duration. This
method also has the advantage that grain
size can be controlled such that the
sample is easily amenable to electron
microprobe analysis. Because reequili-
bration experiments require extended run
times (e.g., 2-3 weeks at 1100°C; run
times given in Fig. 64) , limiting the
number of runs that can be carried out,
a large sample holder simultaneously
containing six samples was used in the
experiments in order to improve effi-
ciency. The six-sample assembly can be
used in a %-inch, solid-media, high-
pressure vessel up to a pressure of 30
kbar.
The sample holder consists of two
graphite disks with three cylindrical
holes (diameter 0.0625 inch; depth, 0.125
inch) drilled in each disk. The holes are
closed by fitting lids. The disks are
0.250 inch in diameter and 0.156 inch
thick. The thermocouple is separated
from the graphite by a thin wall of
AlSiMag, a nonconducting, high-tem-
perature ceramic. The measured tem-
perature gradient across the sample is
less than 15°C (at 1400°C).
Experiments with two thermocouples
(Pt/Ptl0% Rh and W3%Re/W25%Re)
have been conducted at temperatures of
1100°, 1200°, 1300°, and 1400°C (13.5
kbar). Only at 1300° and 1400°C was a
relative drift between the two thermo-
couples observed. The relative down-
ward drift in temperature of approxi-
mately 2.5°C/hour, as read by the
Pt/Ptl0%Rh thermocouple, is believed
to be due to contamination of the couple
(c/. Mao and Bell, Year Book 70, p. 284) .
In the runs above 1200°C, where drift
occurred, the power input was adjusted
to maintain constant temperature, as
read by the tungsten-rhenium couple.
The experiments were carried out by the
piston-in technique, and the results in-
clude a friction correction of —10% (cf.
Johannes et al., 1971) .
Starting Materials
Five pairs of homogeneous natural
pyroxenes and the Salt Lake, Hawaii,
spinel lherzolite (Kushiro et al., 1972)
were used as starting materials for the
experiments. The pyroxene pairs are
Fe-rich, nonaluminous pyroxenes (nos.
207, 264, 277, and 278)^ described by
Butler (1969, Table 4) and a homogene-
ous aluminous orthopyroxene (6.6 wt %
A1203) and aluminous clinopyroxene (5.3
wt % A1203) from the Delegate breccia
pipe, Australia. The position of the pairs
in the pyroxene quadrilateral is shown in
Fig. 64A. Originally the minerals were
ground under acetone to an average grain
size of 20 yam. After it was found that
diffusion rates were a major problem, a
finer grain size of 5—10 /mi was used.
At 1100°C a small amount (approxi-
528
CARNEGIE INSTITUTION
En
A A. A A A A A A"
277
Fig. 64. (A-F) Analyses of experimentally produced pyroxenes at various pressures and tem-
peratures, plotted in pyroxene quadrilaterals. B, pyroxene composition of Brown (Year Book 66,
p. 349). G, extrapolated point of Grover, Lindsley, and Turnock (1973). R, extrapolated points
for ortho-clino inversion of Ross, Huebner, and Dowty (1973). D, diopside solvus after Davis
and Boyd (1966). Hypothetical phase boundaries stippled.
i
GEOPHYSICAL LABORATORY
529
N
K
•
••
1 \
•
•
n -
•
'/ ;
•
•
// /
•
r-
-*
%f
..B
R
_v V v v v*^^/
530
CARNEGIE INSTITUTION
mately 5 wt %) of oxalic acid was added
to promote reaction.
Results
The results of the experiments are
shown diagrammatically in Fig. 64 A-F.
Each point on these diagrams represents
one (or in some cases more than one)
spot analysis carried out with the elec-
tron microprobe analyzer (Finger and
Hadidiacos, Year Book 71, p. 598) . Only
analyses with structural formulae within
2% of the ideal values are shown. No
analyses have been excluded on other
criteria; the diagrams show all data
points obtained. The points from the
alumina-bearing compositions have been
projected through A1203.
Considerable compositional inhomo-
geneity is found in most runs (Fig. 64)
and overall equilibrium was not reached
in the experiments. It is common to find
unreacted material in the cores of grains
of more than 15 /xm. This lack of overall
equilibrium is discouraging, but as the
initial compositions of the minerals are
known, the direction of reaction can be
clearly established. The compositional
range shown for each mineral must lie
within its stability field. Those composi-
tions farthest removed from the original
composition are believed to be approach-
ing equilibrium most closely. The aver-
ages of three to eight analyses (per run)
of each mineral selected on this basis
show low standard deviations (Table 12) .
The alumina content of enstatite in
equilibrium with pyrope has been studied
experimentally by Boyd and England
(Year Book 63, p. 157) and more
recently by I. D. MacGregor (in prepara-
tion) . Data on natural ultramafic com-
positions have been reported by Mac-
Gregor and Ringwood (Year Book 63,
p. 161) and Green and Ringwood (1970).
TABLE 12. Analytical Results: Average Values with Standard Deviations
Mg/(Mg+ Ca/(Ca +
T, °C
P, kbar
Fe
2+), Mg),
AI2O3 wt%
Fobs, kbar Pcaic, kbar
mole % mole %
Orthopyroxene Analy
ses
1110
22.5 D*
0.80
(0.00) f 0.034 (0.00)
3.0 (0.23)
-8.5 -10.8
1110
40.5 D
0.83
(0.00) 0.024 (0.00)
1.3 (0.25)
-~10.0 -12.4
1410
27.0 S
0.86
(0.00) 0.047 (0.00)
5.9 (0.24)
-9.0 -6.5
1410
40.5 S
0.89
0.034
3.1
Clinopyroxene Analyses
1110
22.5
0.82
(0.00) 0.45 (0.00)
3.6 (0.11)
1110
40.5
0.84
(0.01) 0.44 (0.01)
3.0 (0.71)
1410
27.0
0.85
(0.00) 0.26 (0.01)
6.5 (0.23)
1410
40.5
0.88
0.30
Garnet Analyses
5.8(?)
Gross, mole %
1110
22.5
0.67
(0.01) 0.23 (0.01)
16.6
1110
40.5
0.68
(0.01) 0.21 (0.01)
15.2
1410
27.0
0.80
(0.01) 0.71 (0.02)
14.4
1410
40.5
0.85
0.08
KD Values
(using average compositions) x
KD (Ga-Cpx)
KD (Cpx-Opx)
1110
22.5
0.43
1.13
1110
40.5
0.41
1.11
1410
27
0.71
0.88
1410
40.5
0.65
0.87
* Starting material: D, Deltex (Fig. 64B); S, Salt Lake Crater, Hawaii, spinel lherzolite.
t Standard deviations in parentheses.
GEOPHYSICAL LABORATORY
531
The alumina content of orthopyroxene, in
equilibrium with garnet and clinopyrox-
ene, obtained in the present experiments
is considerably lower than that reported
by MacGregor for the same pressure and
temperature (Table 12). Recently
S. Banno and B. J. Wood (in prepara-
tion) have used a thermochemical calcu-
lation to evaluate the effect of Ca and
Fe2+ substitution in garnet and ortho-
pyroxene on the pyrope-enstatite equili-
brium. The calculated pressure difference
by the method of Banno and Wood is in
reasonable agreement with the observed
pressure shift (difference between the
present data and the data by MacGregor,
in preparation; Table 12).
The influence of chromium is expected
to be in the same direction as that of Ca
and Fe2+; that is, the stability field of
garnet is extended, and the stability of
aluminous enstatite is restricted. The
data of Akella and Boyd (this Report,
Table 10) on synthesis experiments at
1100°C in Ti-rich compositions show Al
contents in the orthopyroxene compar-
able to those obtained here.
The alumina content of clinopyroxene
is higher than that of the coexisting
orthopyroxene (Table 12). The clino-
pyroxenes, however, contain about 1 wt
% Na20, and when the A1203 content is
subtracted as jadeite component the
clinopyroxene contains slightly less
AI2O3 than the orthopyroxene.
The Orthopyroxene-Clinopyroxene
Two-Phase Region
In the Mg-rich portion of the two-
pyroxene field in the temperature range
1100°-1200°C and down to Mg/(Mg +
Fe2+) = 0.80, the clinopyroxene solvus
[Cpx(Opx)] projects towards the Fs
corner of the pyroxene quadrilateral. The
use of Ca/(Ca + Mg) in applying the
Di(En) solvus (Boyd and Schairer,
1964; Davis and Boyd, 1966) to deter-
mine temperature for Fe-bearing com-
positions is in order under these condi-
tions because the Fe-bearing pyroxene
composition is projected from Fs onto
the Di-En join.
At 1400°C and 13.5 kbar the solvus
under discussion is absent for composi-
tions with Mg/(Mg + Fe2+) < 0.90
because of the appearance of pigeonite
(Fig. 64F). The topology of the phase
diagram indicates that even minor
amounts of Fe cause errors in tempera-
tures estimated from the Di(En) solvus.
At 1400 °C and 27 kbar the temperature
of a pyroxene pair with Mg/(Mg +
Fe2+) = 0.85 would still be overesti-
mated by 50 °C or more. At 45 kbar only
one measurement could be made, and the
value obtained suggests that as the Ca-
poor clinopyroxene (pigeonite) field con-
tracts with increasing pressure (see later
section of this Report), the Cpx(Opx)
solvus boundary rotates, projecting more
and more in the direction of the Fs
corner of the quadrilateral.
The behavior of the Opx(Cpx) solvus
with varying pressure, temperature, and
Fe/Mg ratio is complicated. As shown in
Fig. 65 the Ca/(Ca + Mg) of ortho-
pyroxene increases with temperature and
decreases with pressure. The pyroxenes
with low Ca contents at high pressure,
however, contain the highest Mg/(Mg -f-
Fe), suggesting the possibility of control
by this chemical variable.
Comparison with the empirical curve
of Boyd and Nixon (this Report, Fig. 6)
and the experimental data of Boyd
(1970) on the system En-Di-Py and
Akella and Boyd (this Report) indicates
that the effects on Ca in orthopyroxene
of pressure and Mg/ (Mg + Fe2+) , on the
one hand, and temperature, on the other,
are of the same order of magnitude and
of opposite sign.
The Distribution of Fe and Mg between
Garnet and Pyroxenes
The experimentally determined KD
values for garnet-clinopyroxene and
clinopyroxene-orthopyroxene pairs are
532
CARNEGIE INSTITUTION
0.06 -
0.05
0.04
o
O
^ 0.03
o
o
0.02
0.0
I
[0.88] ■ (I)
[0.81] ■ (5)^ 13.5 kbar
I
[0.86] ((4)
[0.98]^"(D
>vBoyd and Nixon,
this Report
Boyd, 1970; 30 kbar
00
1200 1300
Temperature, °C
1400
Fig. 65. Variation of Ca/(Ca + Mg) ratio in orthopyroxene in equilibrium with clinopyroxene
as a function of temperature and pressure. Empirical curve for ultramafic nodules (Boyd and
Nixon, this Report) is shown for comparison. Figures in parentheses indicate number of analyses;
figures in brackets indicate the Mg/(Mg + Fe2+) ratio of the orthopyroxene.
given in Table 12. The data show a
clear-cut influence of temperature on the
partition coefficients for exchange equili-
bria between Ga-Cpx and Cpx-Opx.
Comparison with the experimental
data of Akella and Boyd (this Report,
Tables 10 and 11) at 1100°C indicates
that KD is approximately constant for
garnet (with Cpx) with a Mg/(Mg +
Fe2+) of 0.70 to 0.40.
The effect of pressure on KD (garnet-
clinopyroxene) has been variously esti-
mated at about 3% per kbar (Evans,
1965) to about 1% per kbar at 1100°C
(Banno, 1970). The observed decrease
in KD at 1100°C from 0.43 at 22.5 kbar
to 0.41 at 40.5 kbar is insignificant and
suggests that the pressure effect on KD
may be much smaller than theoretically
predicted.
The values obtained here have a spe-
cial interest because of the possible com-
parison with the large amount of data on
garnet lherzolites and related garnet-
clinopyroxene-orthopyroxene bearing
nodules from kimberlites collected by
Boyd and Nixon (this Report) . Figure 66
shows the KD values of garnet-clino-
pyroxene pairs plotted against tempera-
ture, estimated from the Di(En) solvus
(Boyd and Nixon, this Report). The
diagram shows a correlation of KD with
estimated temperature and with several
exceptions (7 out of 28) the points lie on
a straight line between 950° and 1420 °C.
The points obtained in this study plot
close to or on the line, as do the data of
Akella and Boyd (this Report). It is
concluded, therefore, that the KD (gar-
net-clinopyroxene) is a potentially useful
geothermometer and that the curve in
Fig. 66 can be used as a first approxi-
mation for high-temperature, high-pres-
sure ultrabasic rocks.
The grossular content of garnet in the
three-phase assemblage varies with (esti-
mated) temperature (Fig. 66). The
decrease of grossularss in the garnet must
GEOPHYSICAL LABORATORY
533
1600
1400
o
| 1200
Q.
E
1000
800
Grossular mole %
5 10 15 20
Kimberlite nodules,
Boyd and Nixon, this Report
* KD(Cpx-Opx)
• KD(Ga-Cpx)
■ Grossular mole %
(27) (40.5) (27) (40.5)
zv o m .
A A
A A
(40.5) \ (27)
B\ B
Open symbols are
experimental points
(22.5) (40.5)
Grossular-
mole %
a^Kd (Cpx-Opx)
KD(Ga-Cpx)
22.51 (40.5)
O
!•
(40.5) (22.5)
EH
.. \
• •
1.0
0.8
0.6
0.4
0.2
K,
Fig. 66. Variation of KD (Ga-Cpx), Kd (Cpx-Opx), and grossular content of garnet in ultra-
mafic nodules from kimberlites as a function of temperature (derived from Di(En) solvus; Boyd
and Nixon, this Report). Experimentally determined values are shown for comparison. Also
shown are the data of Boyd (1970) for the system MgSiOs-CaSiOs-AhOs. The pressures of the
experimental runs are given in parentheses.
be attributed mainly to an increase in
temperature, with increasing pressure as
a secondary factor. The influence of the
Mg/Fe ratio can probably be disregarded
here in view of the small compositional
range of the garnets. However, a possible
effect of the Mg/Fe on the grossular
content of the garnet cannot be dis-
counted. Note that the data of Boyd
(1970) on the system MgSi02-CaSi03-
A1203 at 1200°C fall on the curve in Fig.
66. The grossular contents of the garnets
in the present experiments compare rea-
sonably well with those from natural
specimens formed under similar physical
conditions (Fig. 66) . The data of
Kushiro, Syono, and Akimoto (1967) for
the Fe-free system are apparently in-
consistent with the temperature effect
observed here.
The distribution of Mg and Fe2+ be-
tween clinopyroxene and orthopyroxene
shows a nonlinear correlation with esti-
mated temperature (Fig. 66). At low
temperatures (1100°C and below) Mg
favors clinopyroxene over orthopyroxene,
but at high temperature this relationship
is reversed. This unusual behavior,
exemplified by the data of Boyd and
Nixon (this Report) , has been repro-
534
CARNEGIE INSTITUTION
duced experimentally (Table 12, Fig.
66). The inflection point lies close to
1100°C. Above 1100°C the value of KD
departs increasingly from unity, and a
near linear correlation with estimated
temperature is evident. The experimental
values are very close to those from na-
tural specimens at 1410 °C but plot
slightly off the empirical curve at 1100°C,
possibly suggesting that reequilibration
of the pyroxenes was not complete.
Extrapolation of the KD (Ga-Cpx)
curve to lower temperature is not pos-
sible at present. Theoretically the curves
of KD versus temperature should not pro-
duce straight lines even when mixing in
the phases is ideal (in the ideal case In K
versus 1/T should be linear). The value
of KD (Ga-Cpx) for eclogite from high-
grade metamorphic terranes ranges from
0.1 to 0.2 (Banno, 1970). Because these
eclogites probably formed in a tempera-
ture range of 550°-800°C, considerable
curvature of the KD line in Fig. 66 must
occur below 900 °C.
The Stability Field of Ca-Poor
Clinopyroxene
The results of the experiments relevant
to the stability of Ca-poor clinopyrox-
ene* are shown diagrammatically in Fig.
64 A-F. The results illustrate the expan-
sion of the Ca-poor clinopyroxene
stability field with increasing tempera-
ture and its contraction with increasing
pressure.
The wollastonite content of Ca-poor
clinopyroxene in the experiments shows a
wide spread, and the data do not indicate
the first composition to become stable
relative to orthopyroxene + clinopyrox-
* At room temperature most natural Ca-poor
clinopyroxenes have P2i/c symmetry (pigeon-
ite). Recent workers (Prewitt, Brown, and
Papike, 1971) have shown that these same
pyroxenes have C2/c symmetry at high tem-
perature. Therefore it is not deemed necessary
to distinguish between structural types in dis-
cussing clinopyroxene phase equilibria at high
temperatures.
ene with increasing temperature. Com-
paring the results at 15 kbar of Smith
(1972), Grover, Lindsley, and Turnock
(1973), and Kushiro and Yoder (Year
Book 68, p. 228), it appears that the
apex of the pigeonite field decreases in
Wo content from about Woi5 (mole %),
in the range Eni5Fs85 to En40Fs6o, to Wo8
on the diopside-enstatite join. Natural
pigeonites from volcanic rocks, probably
formed at very low pressure, typically
contain Woi0, whereas the Wo content of
pigeonite on the Di-En join at atmos-
pheric pressure is between 4 and 5 mole
% (Kushiro, 1972c). The foregoing
observations suggest that the wollaston-
ite content of the pyroxene forming the
apex of the Ca-poor clinopyroxene field,
i.e., the Ca-poor clinopyroxene in the
three-phase assemblage two clinopyrox-
enes + orthopyroxene (Fig. 64), is dis-
placed toward lower Ca content with
increasing Mg content (and tempera-
ture) and toward higher Ca content with
increasing pressure. The latter effect may
explain the phase relations at 1100°C
and 22.5 kbar. The topology of the dia-
gram suggests that at 22.5 kbar the com-
position of the Ca-poor clinopyroxene in
the three-phase assemblage may have
moved to Wo22- The data of Brown
{Year Book 66, Fig. 10) indicate that
the composition Wo7.6En4o.7Fs5i.7 is not
stable at 1100°C and 22.5 kbar. This
result is consistent with the phase bound-
aries of Fig. 64B. To obtain the dashed
hypothetical phase boundaries in Fig.
64 A-F, the data of Ross, Huebner, and
Dowty (1973) have been extrapolated to
higher pressures using the slope of the
orthopyroxene-to-clinopyroxene inver-
sion curve of Brown (Year Book 66, Fig.
10). This slope is almost identical with
that obtained by Kushiro and Yoder for
the lower stability of Ca-poor clino-
pyroxene on the Di-En join (Year Book
68, Fig. 18). The extrapolated results
are shown to be consistent with the
present data and provide a self-con-
sistent set of phase diagrams.
GEOPHYSICAL LABORATORY 535
PHYSICAL AND EXPERIMENTAL MINERALOGY
Pyroxenes from Mull Andesites
D. Virgo and M . Ross*
Mull Andesite, 15990
Monoclinic pigeonite from the Mull
andesite (Bailey et al., 1924, Table 3,
No. 15990) has provided the reference
standard for this type of pyroxene
(Morimoto and Giiven, 1970; Brown
et al., 1972). In addition, it serves as a
reference standard for the chemical com-
position of homogeneous pigeonites (the
Mull pigeonite referred to above contains
a minor percentage of augite exsolution
lamellae; Brown et al., 1972) in contrast
to inverted pigeonites from plutonic and
hypabyssal rocks (e.g., Hess, 1941 ;
Brown, 1957; Atkins, 1965). Some com-
ments on the rock texture of this sample
have been given by Virgo (this Report) ,
and caution should be taken in assuming
that the structural state of this pigeonite
is typical of volcanic environments. It is
reasonable to say that this sample is
extraordinary compared with other ter-
restrial pigeonites from a variety of rock
types because it is readily amenable to
mineral separation and final purification
techniques. It yields ideal samples for
further mineral studies. In fact, chemi-
cally and structurally similar pigeonites
are found only in some returned lunar
samples. However, the Mull pigeonite is
unfortunately also unique because of its
paucity. The samples used by Morimoto
and Giiven (1970) and Brown et al.
(1972) stem from the rapidly diminish-
ing supply curated by H. S. Yoder, Jr.
In an attempt to rectify this situation, a
sample of the Mull andesite collected
from the original locality of Hallimond's
sample was obtained from Dr. P. E.
Champness, University of Manchester
(Sample 05-441-1). Initial electron
microprobe analyses revealed significant
* U.S. Geological Survey, Washington, D.C.
20244.
differences in chemistry between this
sample and that of Hallimond, 1914 (cf.
Tables 13 and 14), and to avoid further
confusion, crystals of Mull pigeonite
(sample 15990) were obtained from
Yoder for new and detailed chemical
analysis.
The average chemical composition and
normalized En, Fs, Wo contents (mole
%) are given in Table 13 and compared
with Hallimond's analysis. It is note-
worthy in terms of the previous crystal-
structure refinements that Hallimond's
sample shows a range in chemical com-
position from Fs48.5 to Fs5i.5 and that the
average calcium content is somewhat
lower than previously assumed (Table
13).
Mull Andesite, 05-441-1
This sample consists of equigranular
euhedral phenocrysts of pyroxene and
plagioclase set in a glassy matrix con-
taining microlites of pyroxene. In grain
mounts, two structural types were recog-
nized: orthopyroxene that shows mod-
erate pleochroism and a clinopyroxene
that commonly shows straight extinction
and in some crystals is twinned. Both
pyroxene types appear to have a core in
some crystals, and in crystal 14-R this
core was found to be glass. Crystals of
both pyroxene types were selected for
electron microprobe analysis following
identification by single-crystal x-ray
precession techniques. Unit cell dimen-
sions for both pyroxene types are given
in Table 14. The clinopyroxene is identi-
fied as pigeonite with space group sym-
metry P21/c. Twinning was observed in
two of the six pigeonite crystals ex-
amined by x-ray techniques and was
identified as growth twinning. It should
be pointed out that this twinning is dis-
tinct from the polysynthetic twinning
observed in Mg-rich pigeonite ("clinoen-
statite") quenched from inferred "pro-
toenstatite" (Dallwitz, Green, and
536
CARNEGIE INSTITUTION
TABLE 13. Chemical Analyses of the Mull Pigeonite, 15990
Si02
Ti02
Cr203
A1203
Fe203
FeO
MnO
CaO
MgO
K20
Na20
Li20
H20
H20 +
Totals
Fs
En
Wo
Hallimond, 1914
Formula
This Report11
Formula
49.72
1.976
50.51 zb 0.66
1.985
0.85
0.025
0.42 ± 0.02
0.012
n.d.t
trace
0.90
0.042
0.72 dz 0.03
0.034
1.72
27.77
I 0.975J
28.97 § ±0.39
0.952
0.98
0.033
0.85 =b 0.03
0.028
3.80
0.162
3.29 zb 0.25
0.139
12.69
0.752
14.30 zb 0.33
0.837
0.12
0.006
n.d.
0.23
0.018
trace
trace
n.d.
0.08
1.27
100.13
99.07
Molecular Percentages
51.61
49.38
39.81
43.41
8.58
7.21
* Average of fourteen spot determinations on six distinct grains,
t n.d., not determined.
t Total Fe calculated as FeO in this table by the present authors.
§Total Fe measured as FeO.
TABLE 14. Electron Microprobe Analyses of Pigeonites from Mull Andesite, 05-441-1
Untwinned Cpx
Twinned Cpx Untwinned Cpx
Untwinned Cpx
X-9R-73
X-15R-73
X-13R-73
X-14R-73
Oxides, *wt %
Si02
48.55
49.33
48.11
48.92
Ti02
0.29
0.33
0.34
0.34
A1203
0.53
0.63
0.65
0.65
FeO
32.28
32.77
32.60
32.27
MgO
13.25
13.21
13.48
13.21
CaO
1.87
1.92
1.92
1.97
MnO
0.94
0.93
0.92
0.94
Totals
97.71
99.12
Molecular Percentages
98.02
98.30
Fs
55.39
55.75
55.28
54.14
En
40.51
40.08
40.74
40.39
Wo
3.99
4.17
Unit Cell Dimensions, A
4.14
4.28
a
9.683!
9.686
9.688
9.682
b
8.954
8.947
8.960
c
5.238
5.236
5.239
5.234
P
108.62°
108.67°
108.60°
108.58°
Space group
P2i/c
P2i/c
P21/c
P2,/c
* Each analysis is the average of five to ten spot determinations.
t Host crystal contains 2% augite on (100) with a = 9.805, c = 5.244 A, p
= 105.85'
GEOPHYSICAL LABORATORY
20
537
A A A A A
I
• ■ Mull Andesile
A x WeiselbergAndesite
X--X-] — xx-x x xx
— A-A A AA
r
■ »
\L
\L \L
45
50 55
ATOMIC PERCENT
60
65
70
Fig. 67. Ca-Mg-Fe plot of analyzed orthopyroxenes and pigeonites from the Mull andesite
(solid circles and squares) and the Weiselberg andesite (solid triangles and crosses; Nakamura
and Kushiro, 19706). Solid lines refer to the trend of crystallization. Arrows designate coexisting
augite composition.
Thompson, 1966) because the twin
lamellae are commonly coarse enough to
be observed optically. Also, the calcium
content of the twinned pigeonite is
greater than expected for a quenched
proto form. The twinning is also distinct
from the polysynthetic twinning ob-
served by heating natural orthopyroxenes
above the orthopyroxene-clinopyroxene
inversion curve (Ross, Huebner, and
Hickling, 1973). Both structural types
are free of exsolution features, with the
exception of one pigeonite crystal, which
contained 2% augite exsolved on (100).
No exsolution features could be seen
optically.
The electron microprobe analyses are
given in Table 14, and average grain
analyses are plotted in Fig. 67. The
orthopyroxene crystals show some zoning
from \V02Fs42En52 to Wo2Fs52En46 com-
pared with the pigeonite crystals, which
are essentially unzoned (average compo-
sition is Wo4Fs56En4o) . The difference in
chemical compositions is further distin-
guished by the Al-Ti relationship (Fig.
68) . It is noteworthy that the ortho-
pyroxene is richer in Al and Ti than the
So.020
0.015 —
•s 0.010
;o.oo5
0.01 0.02 0.03 0.04 0.05 0.06
Atomic ratio of Al on the basis of 6 0
0.07
Fig. 68. Ti-Al relations of analyzed orthopyroxenes (solid circles) and pigeonite (solid squares)
538
CARNEGIE INSTITUTION
coexisting pigeonite. Both pyroxene
types have Ti/Al ratios significantly less
than 1:2, and therefore some Al is
octahedrally coordinated (Fe203 and
Cr203 are present in trace amounts) .
Discussion
The range of composition for the
orthopyroxene crystals from Fs42-Fs52 is
indicative of progressive crystallization
from a fractionating liquid. Similar
fractionation trends have been found in
a single sample for other natural basaltic
and andesitic rocks. In this connection it
is noted that phenocrysts of both struc-
tural types occur with a slight rounding
of the crystal edges and that this effect is
more significant in the orthopyroxene
crystals. The calcium content of the
P2i/c pigeonite (Wo4) is significantly
less than that normally found for P2x/c
pigeonites from terrestrial basalts and
andesites (/WWoi0) . It is suggested that
the assemblage orthopyroxene (Wo2Fs53-
En45) -pigeonite (Wo4Fs54En42) repre-
sents a narrow two-phase field close to
the enstatite-ferrosilite join.
It is further suggested that these
results for a natural system can be ex-
plained in terms of the reaction of ortho-
pyroxene to pigeonite originally defined
by Bowen and Schairer (1935) and rede-
termined by Ross, Huebner, and Hickling
(1973). In these terms, the experimental
results of Ross and co-workers on the
orthopyroxene crystal No. 264 (W01.3-
En42.4Fs56.3) can be compared with the
two-phase assemblage from the Mull
andesite. This crystal commenced reac-
tion to pigeonite at 1038 °C after heating
for 17 hours. At this temperature the
host orthopyroxene grain contains about
2% pigeonite epitaxially oriented on
(100) of the host. Unit cell dimensions
for both phases are given in Table 16.
Further heating at this temperature did
not alter the ratio of pigeonite to augite
in the host crystal. From Tables 14, 15,
and 16 and the /^-nomogram of Papike
et al., (1971), there appears to be rea-
sonable agreement between the composi-
tions of the two-phase assemblage from
the Mull andesite and the experimental
result. A crystallization temperature
close to 1038 °C is suggested for the co-
TABLE 15. Electron Microprobe Analyses of Orthopyroxenes from Mull Andesite, 05-441-1
X-20R-73
X-18R-73
X-3R-73
X-8R-73
Oxides,* wt %
Si02
48.58
50.10
50.14
49.37
Ti02
0.50
0.47
0.47
0.48
AI2O3
1.56
1.44
1.27
1.18
FeO
24.81
25.54
27.47
30.76
MgO
19.01
19.37
17.98
15.43
CaO
0.99
1.00
1.11
0.98
MnO
0.63
0.66
0.68
0.79
Totals
96.08
98.58
Molecular Percentages
99.12
98.99
Fs
41.37
41.63
45.11
51 71
En
56.54
56.32
52.60
46.20
Wo
2.08
2.03
Unit Cell Dimensions, A
2.28
2.08
a
18.319
18.330
18.353
b
8.905
8.916
8.926
c
5.223
5.221
5.228
Space group
Pbca
Pbca
Pbca
Pbca
Each analysis is the average of five to ten spot determinations.
GEOPHYSICAL LABORATORY
539
TABLE 16. Unit Cell Dimensions of Ortho-
pyroxene Crystal No. 264 * (Wbi.aEn^Fsee.s)
and Coexisting Pigeonite t
Orthopyroxene
Pigeonite
(98%)
(2%)
a, A
18.303
9.676
6, A
8.941
8.941
c, A
5.241
5.224
P
90°
108.68°
Space group
Pbca
P2i/c
* Heated at 1038°C for 17 hours under con-
trolled /o2 (S. J. Huebner, personal communi-
cation).
t A synthetic pigeonite (\V05.0En47.5Fs47.5) syn-
thesized at 1200°C, atmospheric pressure, has
the following cell parameters: a = 9.662, b =
8.931, c = 5.218 A; p = 108.71° (Turnock,
Lindsley, and Grover, 1973).
existing orthopyroxene-pigeonite pair in
the Mull andesite, assuming that the
small amounts of Ti, Al, and Mn do not
greatly affect the temperatures. In fact,
the natural orthopyroxene used in this
heating experiment contained amounts
of these minor elements similar to those
found in the Mull pyroxenes.
It is noteworthy that other authors
(e.g., Brown, 1957; Kuno and Naga-
shima, 1952; Nakamura and Kushiro,
1970a) have also reported two-phase tie
lines, but between orthopyroxene and
pigeonite (^Woio). The tie line Opx
("- 'Wo3.5En65.5Fs3i.o)-pigeonite HWo85-
En61.5Fs3o) inferred from plutonic com-
plexes (e.g., the Skaergaard and Bushveld
intrusives) , however, has not actually
been observed in a single hand specimen.
It is also evident from this report that
crystallographic studies may be of ex-
treme importance in the identification of
low Ca-bearing pyroxene phases.
The Mull andesite and the Weiselberg
andesite (Nakamura and Kushiro,
19706) are comparable in bulk chemical
composition, and the constituent pyrox-
ene phases have similar Fe/Mg ratios,
but the pyroxene mineralogy is quite
distinct. An inquiry into some reasons
for this difference is therefore relevant.
The pyroxene compositions from both
andesites are compared in Fig. 67. It
should be noted that the authors are
aware that a qualitative comparison of
the crystallization history of these two
andesities must be undertaken because
the three-phase assemblage found by
Nakamura and Kushiro, orthopyroxene
(-— 'Wo4)-pigeonite (Wo85)-augite
(W025-37) (Fig. 67), may represent a
unique phase equilibrium. Their
orthopyroxene and pigeonite are satu-
rated with respect to Ca-rich clinopyrox-
ene compared with the two-phase
assemblage reported here, orthopyroxene
(^Wo2) -pigeonite (^Wo4). A preferred
explanation of the data in Fig. 67 is that
the two phases in the Mull andesite rep-
resent a quench across the narrow inver-
sion interval of orthopyroxene + low-Ca
pigeonite at a higher temperature com-
pared with the three-phase assemblage in
the Weiselberg andesite (Fig. 69). In
view of the slightly lower Fe/Mg ratio
of the Weiselberg andesite, an interpre-
tation that both andesites have crystal-
lized at the same temperature can
probably be ruled out because of the
significantly different Wo content (fac-
tor of 2) for orthopyroxenes from both
rocks. The above comparison of the Mull
O 10 20 30 40 50
(Mg,Fe)2Si206 CaSiOj (mote per cent) (CaMg,Fe)Si206
Fig. 69. Schematic diagram indicating the
subsolidus relationships of a pseudobinary join
across the pyroxene quadrilateral.
540
CARNEGIE INSTITUTION
and Weiselberg andesites leads to some
comments on the pigeonite stability field.
In natural systems, the frequently ob-
served phase assemblage is P2t/c pigeon-
ite (^Woio) + subcalcic augite/augite.
A compilation of pigeonite analyses
shows, however, an obvious gap in cal-
cium content for P21/c pigeonite between
Wo4 and W010 compositions. Alterna-
tively, Kushiro (this Report) and Naka-
mura (1971) suggested that the occur-
rence of lunar pigeonites of lower Fe/Mg
ratio but with cores of ~ Wo5 composition
is an argument for continuous extension
of the pigeonite field from W010 toward
the enstatite-ferrosilite join. Of major
importance in the lunar rocks, however,
is that the composition of the pigeonite
directly in contact with Ca-rich augite is
^Woi2. It is suggested that a rapid crys-
tallization sequence for lunar rocks may
preclude the equilibrium phase relation-
ships between cores and a second pyrox-
ene phase. A low-Ca, P2jc pigeonite-
Opx field distinct from a high-Ca, P2\/c
pigeonite-augite field during crystalliza-
tion from a melt, as shown in Fig. 69, is
suggested.
Finally, it is contended that the well-
documented pyroxene mineral assem-
blages in a fractionating natural melt are
coexisting two-phase assemblages: ortho-
pyroxene + augite; P2i/c pigeonite
(W010) + subcalcic augite/augite. The
field low-Ca, P2±/c pigeonite + ortho-
pyroxene is a necessary consequence
from the intersection of the essentially
colinear, two-phase fields augite +
orthopyroxene and augite -f- pigeon-
ite (Huebner, Ross, and Hickling,
1973) . The three-phase assemblage
orthopyroxene + pigeonite ( — Woi0) +
augite corresponds to a variable Fe/
Mg ratio at nearly singular tem-
perature conditions during crystallization
from a melt and on this basis alone
should not occur very commonly in
natural rocks.
The assemblage Opx (Wo2) + pigeon-
ite (Wo4) is probably also present rela-
tively early in the crystallization
sequence of a fractionating basaltic
liquid, but it may have been overlooked
because of the very small chemical
differences in the two phases and possibly
because of the absence of adequate struc-
tural identification of the phases. It is
also possible that low-Ca pyroxenes may
react out, leaving no indication of their
prior history. It is interesting to note
that Mg-rich pigeonite (■ — Wo5) was re-
cently reported in a diabase from
Connecticut (Philpotts, 1973). Also,
from analytical electron microscopy data
of Lorimer and Champness (1973) it can
be suggested that a bronzite from the
Stillwater complex may have resulted
from a low-temperature, subsolidus in-
version of pigeonite of composition
— Wo4. In the light of these results,
further detailed studies of low-Ca pyrox-
ene assemblages are suggested.
The 57Fe Mossbauer Spectra of Syn-
thetic Pyroxenes across Part of
the Join Fs8r>-Eni5-WoLLASTONiTE
D. Virgo
Crystal-structure data on C2/c clino-
pyroxene along the diopside-hedenbergite
join suggest that the M2 crystal site is
totally occupied by calcium, and in this
light the 57Fe Mossbauer spectra will be
characterized by a single absorption
doublet corresponding to ferrous iron in
the Ml crystal sites (Bancroft, Williams,
and Burns, 1971 ; Matsui, Syono, and
Maeda, 1971). In contrast, low-calcium
orthopyroxenes and pigeonites consist of
two well-resolved doublets corresponding
to ferrous iron in both M sites. In these
phases there is excellent agreement be-
tween the site-occupancy factors calcu-
lated from the intensity ratios of the
distinct doublets and x-ray structure
refinements (e.g., Burnham et al., 1971;
Ohashi and Finger, 1973). However, the
Mossbauer spectra of natural augite-fer-
roaugite compositions (Williams et al.,
1971), lunar subcalcic augites and ferro-
augites (Hafner, Virgo and Warburton,
GEOPHYSICAL LABORATORY
541
TABLE 17. Chemical Composition and Structural State of
Synthetic Pyroxenes across Part of the Join FsssEnis-Wollastonite
Chemical
Synthesis
Sample
Formula,
Structural
Temperature
mole %
State
at 15 kbar, °C
Woo
Fs85Eni5
Pbca
900
Wo 10*
WoioFs76.6Eni3.5
C2/c
950
WOl5
W015Fs72.oEn12.75
C2/c
925
WO 25
W025Fs63.75Enn.2
C2/c
950
Wo35
W035Fs55.25En9.75
C2/c
925
W050
Wo5oFs42.5En7.5
C2/c
925
* Contains a small amount of orthopyroxene (Smith, 1972).
19716; Dowty, Ross, and Cuttitta, 1972) ,
and synthetic Mg-free clinopyroxenes
with compositions between hedenbergite
and ferrosilite (Dowty and Lindsley,
1973) suggest that C2/c clinopyroxenes
are structurally more complex. Clino-
pyroxenes synthesized approximately
across part of the join Fs85Eni5-wol-
lastonite (Smith, 1972) provide the op-
portunity of examining the structural
characteristics of C2/c pyroxenes in the
pure Fe-Mg-Ca system.
The chemical compositions and struc-
tural states of the samples studied are
given in Table 17. The samples were
synthesized in the one-phase fields
delineated by Smith (1972) at 15 kbar,
900°-950°C, and were rapidly quenched.
The Mossbauer spectral measurements
were made at 77°K, using the technique
previously reported (Year Book 71,
p. 607) . The results of the least-squares
fits of Lorentzian lines to the data of
these samples are given in Tables 18 and
19, and the spectra of the Wo50 and Wo25
compositions are given in Figs. 70 and 71.
The hyperfine data, quadrupole splitting,
and isomer shift are plotted against the
mole percentage of Wo in Figs. 72 and 73.
No Fe3+ was detected in any of the
samples.
The end members of the join studied
are orthopyroxene and a clinopyroxene
containing 1.00 Ca ion per formula unit.
The spectra of both these samples are
characterized by narrow line widths and
can be interpreted in terms of the
inferred stoichiometry. Sample Wo0 is
considerably disordered, having a dis-
tribution coefficient, k, equal to 0.262 and
a standard Gibbs free energy of exchange
at 900°C of 3.12 kcal/mole. Both values
are consistent with the results on heat-
treated, natural orthopyroxenes (Virgo
TABLE 18. Quadrupole Splitting, Isomer Shifts, and Line Widths of Spectra * of Synthetic
Clinopyroxenes at 77 °K along Part of the Join FsssEnis-Wollastonite
Ml Parameters
M2 Parameters
Quadrupole
Isomer
Width
Quadrupole
Isomer
Width
Sample
Splitting,
Shift, f
Low
High
Splitting,
Shift, t
Low
High
mm/sec
mm/sec
Velocity
Velocity
mm/sec
mm /sec
Velocity
Velocity
Woo
2.74
1.21
0.29
0.28
1.98
1.17
0.26
0.27
Wo 10
2.95
1.27
0.31
0.30
1.97
1.22
0.32
0.39
W015
3.00
1.29
0.32
0.32
1.91
1.23
0.31
0.41
WO 25
2.98
1.29
0.36
0.37
1.84
1.22
0.31
0.35
WO35
2.90
1.30
0.39
0.32
1.90
1.23
0.34
0.40
WO50
2.70
1.29
0.30
0.30
* Data are the averages of duplicate experiments.
t Refined to metallic iron at 298°K.
542
CARNEGIE INSTITUTION
TABLE 19. Intensity Ratios and Fe2+,Mg Site Occupancy of Synthetic Clinopyroxenes
at 77 °K along Part of the Join FsssEnis-Wollastonite
Intensity Ratio
Peak
Area Height
Site Occupancy'
*
Distribu-
tion
Sample
A (Ml)
A(M1 +
I(M1)
I(M1 +
Ml
M2
Coefficient
Mg
Fe
Mg
Fe
Ca
fct
M2)
M2)
Woo
0.4676
0.4540
0.228
0.772
0.072
0.928
0.262
Wo io
0.4381
0.4668
0.286J
0.714
0.816
0.20
Wo15
0.4792
0.5033
0.275J
0.725
0.715
0.30
WO 25
0.5864
0.5624
0.282J
0.718
0.557
0.50
WO35
0.7114
0.6743
0.255J
0.745
0.360
0.75
WO50
0.150
0.850
1.00
* Calcium was assigned to the M2 site and the mole percentage of Fs was distributed according to
the measured intensity ratios (Virgo and Hafner, 1969). The peak-height values were used for these
calculations.
tfc = (Fe2+/Mg)Mi/(Fe2+/Mg)M2.
I These values are calculated so that the site occupancy of Ml is unity.
and Hafner, 1969). Sample Wo50 shows
a single Fe2+ doublet corresponding to
Fe2+ in Ml consistent with the complete
-4.0
-2.0 .0
l/ELDCITY
2.0
1 / s e c
4.0
Fig. 70. Iron-57 resonant absorption of syn-
thetic ferroaugite, Wo5oFs42.5En7.5. The solid line
is a least-squares fit assuming two Lorentzians
(six line variables, one background variable).
occupancy of M2 by calcium (Fig. 70).
The samples of intermediate composi-
tions, however, show broad line widths
(Table 18) and an intensity ratio for
Fe2+ in the M2 site that is too intense
.000
035
.070
Fig. 71. Iron-57 resonant absorption of clino-
pyroxene, WoasFsea.TJsEnu.a (12 line variables,
one background variable).
GEOPHYSICAL LABORATORY
543
15 20 25 30 35
Mole per cent wollastonite
Fig. 72. Quadrupole splitting of the Ml and
M2 sites at 77 °K against mole percentage of
Wo.
and leads to an excess number of cations
in the M2 site (Table 19). Crystal-
structure refinements do not support
calcium exchange between the M sites,
and electron microprobe analyses of some
crystals (Smith, 1972, Table 17) do not
indicate significant deviations from the
stoichiometric compositions. Dowty and
Lindsley (1973) suggested that this
anomaly in synthetic C2/c pyroxenes
along the Hd-Fs join resulted from the
partial overlap of the M2 doublet with
that due to multiple Ml doublets. In
terms of the crystal structure, it was sug-
gested that variations of the next-nearest
neighbor configurations of the Ml site,
due to random occupancy of M2 sites by
Ca and different combinations of Ca and
Fe2+, could lead to distinct local environ-
1.25
1 1 1
1 1
1
1
M2
-
1.20
1.15
^---~~ ""*"
-
1.10
-
J I I I
15 20 25 30 35
Mole per cent wollastonite
Fig. 73. Isomer shifts of the Ml and M2 sites
at 77 °K against mole percentage of Wo.
ments of the Ml sites. It was further
suggested that these distinct environ-
ments of the Ml sites would cause actual
splitting of the Ml absorption doublet
and consequently some degree of overlap
with the M2 doublet in the 57Fe spectra.
The spectra of the intermediate composi-
tions studied here, however, show well-
resolved absorption doublets due to Fe2+
(Fig. 71). An alternative interpretation
of the Mossbauer spectra of C2/c pyrox-
enes is that of an average absorption
envelope due to Fe2+ in the Ml sites
superimposed on but distinct from the
doublet due to Fe2+ in M2. In these
terms the anomaly in the proportions of
Fe2+ in the Ml and M2 sites stems from
the determinations of the intensities of
the Ml and M2 doublets as a ratio of the
total Fe2+ intensity. This explanation
can be verified using absolute determina-
tions of the Fe2+ site concentrations.
Nevertheless, the contrasting crystal
structures of C2/c pyroxenes suggested
from Mossbauer studies do find support
in recent structural refinements of
natural and synthetic clinopyroxenes
(Takeda 1972a,6; Ohashi, 1973). Anal-
yses of the atomic thermal ellipsoids
show large elongate ellipsoids of the M2
site, suggesting considerable positional
disorder of cations in the M2 site. This
effect is not common to the Ml site, but
oxygen atoms shared by both octahedral
sites show large displacements that do
suggest a significant variation in the
next-nearest configuration of oxygens
around Ml. Takeda (1972a,b) attrib-
uted these structural effects to rapid
quenching, and it will be of interest to
study these parameters by x-ray and
Mossbauer techniques in more slowly
equilibrated samples.
The hyperfine parameters, quadrupole
splittings, and isomer shifts are plotted
against atomic percentage of Ca in Figs.
72 and 73. A discussion of the M2 para-
meters is particularly relevant in view
of the preferred interpretation of the
absorption spectra. The data show sig-
544
CARNEGIE INSTITUTION
nificant changes in the quadrupole split-
ting with substitution of Ca on the M2
site. It is likely that these results are
primarily related to the marked changes
in the coordination of the M2 site across
the series, as suggested from crystal-
structure analyses (e.g., Takeda, 1972a;
Ohashi, 1973). These studies of natural
and synthetic crystals have shown that,
whereas the M2 coordination is eight for
compositions along the diopside-heden-
bergite join and six for the compositions
close to the enstatite-ferrosilite series,
the coordination may signflcantly de-
crease for compositions approaching the
two-phase region in the pyroxene quadri-
lateral. In fact, Takeda has suggested
that this decrease in the number of co-
ordinating oxygen atoms in the M2 site
might provide a crystal-structure basis
for clinopyroxene instability, resulting in
the two-phase region separating Ca-rich
and Ca-poor compositions within the
pyroxene quadrilateral. The quadrupole
splitting for M2 does show a systematic
trend for the compositions Woi5, Wo25,
and Wo35 (Fig. 72). On this basis, the
minimum in the curve ( — Wo25) may
present the most unstable configuration
and a greater proximity to the miscibility
gap. Smith (1972) in fact suggested that
the critical temperature for unmixing is
less well established in his experiments
and may be closer to 950 °C at 15 kbar.
The isomer-shift values for M2 are
uniform across the solid solution series
but are somewhat larger than for the
orthopyroxene composition. These re-
sults may reflect the smaller M-0 bond
distances in Ca-free orthopyroxenes.
In conclusion, it is evident that a large
portion of the pyroxene quadrilateral at
subsolidus temperatures is occupied by
structurally complicated pyroxenes,
which are characterized by severe de-
grees of positional disorder on the M2
site. For a given composition, this effect
will be continuous throughout the crystal
structure, since diffraction experiments
have not provided evidence for domain
structure based on distinct compositions
in C2/c pyroxenes. It is noteworthy that
the Mossbauer technique is sensitive to
detecting anomalies due to this effect.
A Possible High-Low Temperature
Transition in Orthopyroxenes
and orthoamphiboles
Y. Ohashi and L. W . Finger
Nonquenchable phase transitions have
been reported in clinopyroxenes (Smyth,
1970; Prewitt, Brown, and Papike, 1971;
Smyth and Burnham, 1972; Brown
et al., 1972) and clinoamphiboles
(Prewitt, Papike, and Ross, 1970; Sueno
et al., 1972). Structural changes asso-
ciated with the transition are (a) two
crystallographically distinct silicate
chains become identical in the high-
temperature phase; and (b) sites for
metal atoms, which are not on two-fold
axes in the low-temperature phase, are
on two-fold axes in the high-tempera-
ture phase. These changes add a
C-centering translation, which combined
with the two-fold axes, results in space
group changes from P21/c to C2/c
(clinopyroxene) and from P21/m to
C2/m (clinoamphibole) .
The crystal structural relationship be-
tween clino- and orthopolymorphs in
both pyroxenes and amphiboles can be
considered as polysynthetic twin opera-
tions on a unit-cell scale (Ito, 1935,
1950), and thus the major difference
between two polymorphs is the stacking
sequence of the unit "slabs" (100 layers
of primitive clino-unit cells; Burnham,
Year Book 65, pp. 285-290) along a*
(Fig. 74) . The atomic arrangement in
the unit slab is very similar in both poly-
morphs (e.g., Morimoto and Koto, 1969;
Burnham, Year Book 65) .
The existence of C-centered high mono-
clinic phases and the structural resem-
blance between orthopolymorphs and
clinopolymorphs has led to an investiga-
tion of possible high-temperature ortho-
GEOPHYSICAL LABORATORY
545
P21/c_and_C2/c
(C2/conly)
Solid symbols- P2)/c a C2/c
Open symbols- C2/c only
Fig. 74. Structure of orthopyroxene (Burn-
ham et at., 1971) and symmetry relationships
among space groups Pbca, P2i/c, and C2/c.
Note that the symmetry element relating two
adjacent chains within an A layer is a b-glide
plane in orthopyroxene and a two-fold screw
axis in clinopyroxene ; the symmetry in the B
layer, however, is the same in both polymorphs
(see also Fig. 76). In C2/c clinopyroxene the
A and B chains are related by two-fold axes,
and metal atoms occupy special positions on
these axes.
pyroxene and orthoamphibole. For this
hypothetical high-temperature orthopy-
roxene, there would be only one crystallo-
graphically distinct silicate chain. In
addition the nonspace-group hypersym-
metry would force the octahedral metal
atoms to occupy the position at (1/8, y,
3/8). Consequently this polymorph could
be considered as unit-cell scale twinning
of C-centered, rather than primitive, cli-
nopyroxene. This high-temperature ortho-
pyroxene would retain an 18 A a axis and
thus is different from protoenstatite,
which has space group P2xcn with an a
axis of 9 A (Smyth, 1971). These argu-
ments may also be extended to ortho-
amphibole.
On the basis of x-ray diffraction pat-
terns for orthorhombic polymorphs of
pyroxene and amphibole, Ito (1935,
1950) concluded that the unit slab of
orthorhombic polymorphs should be
primitive monoclinic and predicted the
possible existence of a P2x/c clinopyrox-
ene (later found by Morimoto, 1956) and
a P21/m amphibole (later found by
Bown, 1965). Thus this problem of high-
temperature orthopolymorphs is Ito's
approach in reverse; namely, the dis-
covery of a high-low transition in mono-
clinic polymorphs at high temperature
may suggest the possibility of an analo-
gous transition in orthopolymorphs.
As far as symmetry is concerned, a
transition to the high-temperature form
of monoclinic phases is the addition of
some symmetry elements (Fig. 74) , and
consequently the high-temperature sym-
metries C2/c and C2/m are supergroups
of the low-temperature forms P21/c and
P21/m, respectively. The orthorhombic
space groups, Pbca (orthopyroxene) and
Prima (orthoamphibole) , are not con-
sistent with the addition of two-fold
axes, however, and thus no new super-
groups are derived. Therefore, hypo-
thetical high-temperature orthorhombic
phases would be expected to have the
same space groups as the low-tempera-
ture phases, but there would be regional
symmetry that relates the two silicate
chains. This regional symmetry imposes
nonspace-group extinction conditions for
reflections that satisfy equations h -f-
2k + I = 4ra + 2 and h + 2k — I =
4n + 2 (m, n: integers) for both ortho-
pyroxene and orthoamphibole (Ohashi,
1973) . These reflections will be denoted
as c reflections.
Preliminary heating experiments have
been carried out on an orthopyroxene,
Eni3Fs85Wo2, using the precession
method. At 800 °C the c reflections are
too weak to be recorded on the photo-
graphs. This result does not necessarily
confirm a high-low transition in ortho-
pyroxene, however, because of the de-
tectability limit of the film. A better
criterion may be obtained by plotting
546
CARNEGIE INSTITUTION
c-reflection intensities versus tempera-
ture. If the curves for the c reflections
show a more rapid decrease to zero inten-
sity with increasing temperature than
those for other reflections, an intersection
of the c-reflection curves and the zero
intensity would give a transition tem-
perature. For this purpose, Dr. J. R.
Smyth kindly provided the high-tem-
perature intensity data of an orthopyrox-
ene, En3oFs68Wo2, prior to publication.
Examination of these data leads to the
conclusion that a proposed high-low
orthopyroxene transition does not occur
at least below 850 °C, which is higher
than the transition temperature (between
700 °C and 800 °C) of corresponding
clinopyroxene (Smyth and Burnham,
1972) . Possible phase relations are shown
schematically in Fig. 75. As for ortho-
amphiboles, precession photographs of
anthophyllite, ssMgsP^SigC^ (OH)2, at
500 °C still show c reflections.
It is possible that the transition is not
observed because the material either de-
Opx : Orthopyroxene
Cpx : Clinopyroxene
oy
t
r^
:>*
G<>*/
C7>
<U
^f
9)
-~*r~ /*- High Opx (?)
,<&'' /
\P* " /
U
<D
C
—
<**
Temperature
Fig. 75. Schematic internal energy diagram
for pyroxene polymorphs with an intermediate
composition on the enstatite-ferrosilite join.
Protopyroxene and forsterite + silica are as-
sumed not to be stable for the compositions
plotted. It is assumed that high-temperature
orthopyroxene is a stable phase and that the
high-low transition occurs at a higher tempera-
ture in orthopyroxene than in clinopyroxene.
High Clinopyroxene Low Clinopyroxene
Orthopyroxene
Fig. 76. Comparison of the silicate chain
layers in orthopyroxene (Burnham et al., 1971)
and high- and low-temperature clinopyroxene
(Smyth and Burnham, 1972). Two adjacent A
chains in orthopyroxene kink in the same direc-
tion, whereas A chains in clinopyroxene altern-
ate their kink directions.
composes or undergoes an unrelated
phase transition (e.g., ortho to clino)
before the temperature of the high-low
inversion is reached. The fact that ortho
phases either do not transform to high-
temperature phases or have a transition
temperature much higher than that for
the corresponding monoclinic phases
must be explained on the basis of the
crystal structure. In Fig. 76 silicate
chain layers in high and low clino-
pyroxene and orthopyroxene are com-
pared. There is a noticeable difference in
the A layer; two adjacent A chains in the
same layer are related by a two-fold
screw axis in clinopyroxene but by a
6-glide plane in orthopyroxene. As a
result, the silicate chain kink (lines con-
necting the bridging oxygen atoms) of
the A chain is parallel in orthopyroxene,
whereas in clinopyroxene two adjacent
A chains are oppositely kinked. Since
the high-low clinopyroxene transition
involves a reversal of the .4-chain rota-
tion direction (Smyth and Burnham,
1972; Brown et al, 1972; Ohashi, 1973),
the difference in the A-chain layer is im-
portant in the transition. In orthopyrox-
ene, the stacking of octahedral layers
below and above the A-chain layer is
also different from that in clinopyroxene.
These differences may require more
GEOPHYSICAL LABORATORY
547
energy for the parallelly aligned ortho-
pyroxene A chains to reverse their rota-
tion directions. The difference in the
A-chain layers would probably explain
the fact that a high-low transition in
orthopyroxene does not occur in the tem-
perature range where a high-low clino-
pyroxene transition occurs. Similar
arguments about the chain rotation can
be applied to orthoamphibole.
Thermal Vibration Ellipsoids and
Equipotential Surfaces at the
Cation Sites in Olivine and
Clinopyroxenes
Y. Ohashi and L. W. Finger
Equipotential surfaces of atomic
vibrations (Ohashi and Finger, in prepa-
ration) have been calculated for forster-
ite (Mg2Si04), diopside (CaMgSi206),
and hedenbergite (CaFeSi206) using re^
cently refined atomic coordinates. The
data on the crystal structures were
kindly provided prior to publication by
Drs. J. R. Smyth (forsterite), G. E.
Brown (hortonolite) , and S. Sueno
(diopside and hedenbergite) .
The second derivatives of the electro-
static potential were calculated by a
direct evaluation of the equations of
Ohashi (1973) using the techniques of
Ohashi and Finger (in preparation) and
also by the Ewald-method computer pro-
gram used for electric-field-gradient cal-
culation (Raymond, Year Book 70, pp.
227-229). In the repulsive potential
calculation, repulsive parameters are
those given by Ohashi and Burnham
(1972, Table 1). The repulsive para-
meters are assumed to be independent of
temperature; this assumption, however,
might be oversimplified at high tempera-
ture.
Forsterite
The cation sites in olivines exhibit a
remarkable contrast in site symmetry
and polyhedral shape. The Ml site is on
an inversion center and M2, on a mirror
plane. Both sites are octahedrally co-
ordinated by six oxygen atoms, but an
octahedral distortion is greater for the
Ml coordination polyhedron (Brown and
Prewitt, 1973). The differences make a
comparison of the two sites interesting in
terms of equipotential surfaces and
thermal vibration ellipsoids.
The results of calculations are listed in
Tables 20 and 21, and the orientations
of the principal axes are stereographi-
cally shown in Fig. 77. The orientations
of the observed and calculated ellipsoids
at room temperature agree in general
within 30°, although amplitudes of the
axes do not agree. The disagreement of
the results at high temperature is some-
times considerably greater (Table 21).
That the largest root-mean-square dis-
placement axis tends toward the open
direction of the polyhedron is physically
reasonable (see Fig. 77). The observed
thermal vibration ellipsoid for Ml is
somewhat anisotropic but not nearly so
distorted as the calculated ellipsoid.
Diopside and Hedenbergite
In diopside and hedenbergite the Ml
site is six-coordinated and M2 is eight-
coordinated. Both sites are on a two-
fold axis; thus one of the principal axes
of a thermal vibration ellipsoid must be
parallel to the b axis. The anisotropy of
the ellipsoid is more evident for the M2
site (Clark, Appleman, and Papike,
1969; Takeda, 1972a, b; M. Cameron
et al., 1973). Comparison of calculated
and observed thermal vibration ellipsoids
shows that the orientations are in fair
agreement but that the amplitudes of
vibrations are systematically larger for
the observed ellipsoids (Fig. 78, Table
21).
In addition to anisotropy of thermal
vibrations, change of vibrational ampli-
tudes as a function of temperature is an
548
CARNEGIE INSTITUTION
TABLE 20. Second Derivatives at the Cation Sites in Forsterite, Diopside, and
Hedenbergite from the Electrostatic and Repulsive Potential Model
Com-
ponent*
Electro-
static,
Repulsive X 105 dyn/cm
Total
Repulsive
Electro-
static,
X105 dyn/cm
Total
Forsterite {20°C) (l)f
Ml at (0,0,0)
f(aa) 2.42 0.55 2.97
f(bb) 1.43 -0.49 0.94
f(cc) 2.48 -0.06 2.42
f(ab) 0.05 0.54 0.58
f(ac) 0.60 0.28 0.89
/(6c) 0.75 0.07 0.82
Forsterite (900°C) (l)f
Ml at (0,0,0)
f(aa) 2.27 0.43 2.71
/(66) 1.22 -0.37 0.84
f(cc) 2.32 -0.06 2.26
flab) 0.00 0.55 0.55
f(ac) 0.59 0.24 0.83
/(6c) 0.72 0.02 0.73
Diopside (U°C) (2)f
Ml at (0,0.9082,0.25)
f(aa) 1.97 0.48 2.45
f(bb) 2.27 -0.26 2.00
f(cc) 2.33 -0.37 1.96
f(ab) 0 0 0
f(ac) -0.50 0.25 -0.25
/(6c) 0 0 0
Diopside (1000°C) (3)t
Ml at (0,0.9063,0.25)
fiaa) 1.97 0.54 2.52
/(66) 2.13 -0.25 1.88
f(cc) 2.23 -0.37 1.86
f(ab) 0 0 0
f(ac) -0.47 0.11 -0.36
/(6c) 0 0 0
Hedenbergite (2Jl°C) (2)f
Ml at (0,0.9075,0.25)
fiaa) 1.74 0.34 2.07
f(bb) 2.14 -0.43 1.71
f(cc) 1.71 -0.02 1.69
fiflb) 0 0 0
f(ac) -0.40 0.18 -0.22
/(6c) 0 0 0
Hedenbergite (1000°C) (2)f
Ml at (0,0.9049,0.25)
fiaa) 2.05 0.44 2.49
f(bb) 2.05 -0.36 1.69
f(cc) 1.76 -0.12 1.64
fiab) 0 0 0
fiac) -0.52 0.04 -0.48
/(6c) 0 0 0
M2 at (0.9915, 0.2774, 0.25)
1.45 0.33
1.77
2.25 0.36
2.62
2.15 -0.69
1.46
0.12 -0.54
-0.41
0 0
0
0 0
0
M2 at (0.9924, 0.2795, 0.25)
1.28 0.38
1.66
2.08 0.36
2.44
2.08 -0.73
1.35
0.13 -0.52
-0.39
0 0
0
0 0
0
M2 at (0,0.3015,0.25)
0.77 0.95
1.72
2.24 -0.62
1.62
1.84 -0.20
1.64
0 0
0
0.32 -0.34
-0.02
0 0
0
M2 at (0,0.3001,0.25)
0.66 0.91
1.57
2.07 -0.56
1.51
1.66 -0.18-
1.48
0 0
0
0.34 -0.39
-0.04
0 0
0
M2 at (0,0.3003,0.25)
0.83 0.85
1.68
2.09 -0.44
1.64
1.81 -0.27
1.54
0 0
0
0.35 -0.33
0.03
0 0
0
M2 at (0,0.2984,0.25)
0.72 0.82
1.54
2.01 -0.48
1.54
1.57 -0.17
1.40
0 0
0
0.33 -0.39
-0.06
0 0
0
* Components of second derivatives are based on crystallographic axes, i.e., f(aa) = d2U/(adx)2,
f(ab) = d2U/(adx) (bdy), etc.
t Source of data for atomic coordinates: (1) Smyth and Hazen, 1973; (2) M. Cameron et al., 1973;
(3) Clark, Appleman, and Papike, 1969.
GEOPHYSICAL LABORATORY
549
TABLE 21. A Comparison of Calculated Equipotential Ellipsoids and Observed Thermal
Vibration Ellipsoids for Forsterite, Diopside, and Hedenbergite
Calculated Ellipsoid
Observed Ellipsoid
Orientation of
Principal Axes
Orientation of
Principal Axes
Principal
Force r.m.s. r.m.s.
Site Axis Constant, Displace- angle (degrees) with Displace- angle (degrees) with
X105dyn/cm ment, A -\-a +6 +c ment, A +a +6 +c
Forsterite {20°C) (1):
Ml (Mg)
M2 (Mg)
3.94
1.84
0.55
2.79
1.60
1.46
Forsterite (900°C) (1):
Ml (Mg)
M2 (Mg)
Diopside (24°C) (2);
Ml (Mg)
M2 (Ca)
3.62
1.67
0.51
2.60
1.49
1.35
2.75
2.00
1.86
2.27
1.62
1.34
Diopside (1000°C) (3):
Ml (Mg)
M2 (Ca)
2.70
1.88
1.83
2.05
1.51
1.23
Hedenbergite (24°C) (3):
Ml (Fe)
M2 (Ca)
2.30
1.71
1.61
2.21
1.64
1.26
Hedenbergite (1000°C) (3):
Ml (Fe)
M2 (Ca)
2.52
1.69
1.64
1.92
1.54
1.21
0.032
42
72
53
0.037
66
73
30
0.047
132
77
45
0.051
156
80
68
0.086
95
22
112
0.077
92
160
70
0.039
112
22
90
0.044
4
86
90
0.051
22
68
90
0.056
86
176
90
0.053
90
90
0
0.058
90
90
0
0.067
43
72
52
0.117
58
75
37
0.098
132
79
44
0.140
32
99
121
0.177
96
21
110
0.185
89
162
72
0.079
112
22
90
0.140
53
143
90
0.104
22
68
90
0.149
144
126
90
0.110
90
90
0
0.151
90
90
0
0.039
36
90
70
0.052
56
90
49
0.045
90
0
90
0.055
90
180
90
0.047
125
90
20
0.065
146
90
41
0.043
51
90
55
0.066
66
90
39
0.050
90
0
90
0.068
90
180
90
0.055
141
90
35
0.103
156
90
50
0.081
27
90
79
0.141
77
90
29
0.097
90
0
90
0.160
90
180
90
0.098
117
90
11
0.161
168
90
62
0.093
49
90
57
0.154
71
90
35
0.108
90
0
90
0.162
90
180
90
0.120
139
90
33
0.240
161
90
55
0.042
35
90
70
0.070
90
0
90
0.049
90
0
90
0.075
120
90
135
0.051
55
90
160
0.078
150
90
45
0.043
48
90
57
0.079
69
90
35
0.050
90
0
90
0.084
90
180
90
0.057
42
90
147
0.108
159
90
55
0.083
12
90
93
0.145
73
90
32
0.102
90
0
90
0.160
90
180
90
0.102
102
90
3
0.167
163
90
58
0.096
47
90
59
0.153
70
90
35
0.107
90
0
90
0.168
90
180
90
0.121
137
90
31
0.240
160
90
55
* Source of observed thermal vibration ellipsoid data: (1) Smyth, personal communication;
(2) Clark, Appleman, and Papike, 1969; (3) Cameron et al., 1973.
550
CARNEGIE INSTITUTION
Fig. 77. (A) Stereographic projection of the
principal axes of the observed and calculated
thermal vibration ellipsoids for the Ml and M2
sites in forsterite. The data for olivines with
intermediate compositions (hortonolites) are
also shown for comparison. Diamonds, cal-
culated axes, forsterite; squares, observed axes,
forsterite (J. R. Smyth, personal communica-
tion) ; triangles, observed axes, Fo82Fai8 horton-
olite Bl (Finger, unpublished data) ; circles,
observed axes, lunar hortonolite 12018 (Brown
and Prewitt, 1973); L, largest axes; /, inter-
mediate axes; S, smallest axes. (B) Coordi-
nation polyhedra of the Ml and M2 sites in
forsterite. The height of atoms is given as per-
centage of the c axis. Note that the largest
vibration axis tends toward the relatively open
directions of the polyhedra.
interesting problem. The equivalent
isotropic temperature factors (Hamilton,
1959) were calculated and are plotted in
Fig. 79. In spite of its smaller polyhedral
volume, the Ml site of forsterite shows
an equivalent isotropic temperature
factor slightly larger than or comparable
to that for the M2 site. This relationship
results from a significant contribution of
the longest axis component of the Ml
site to the average value. The estimated
temperature factors at high temperature
are considerably lower than the observed
values in all three cases.
The isotropic temperature factor B is
related to the mean-square displacement
(m.s.d.) <r2> by the equation
and the m.s.d. is given by (Ohashi and
Finger, in preparation)
<r2> = kT/f
where k is the Boltzmann constant; T,
the absolute temperature; and /, a force
constant. Then
B = $7r2kT/f.
An isotropic force constant is a harmonic
mean of three principal force constants :
+
})
7_3 v/i ' h
and is on the order of 1 '— ' 2 X 105 dyn/
cm (Table 21) . If / is nearly independent
ofT,
dB
~dT
$7T2k
f
0.5 — 1 X 10~3 A2/deg.
Observed rates of increase of B are in the
range of 2 to 2.5 X 10"3 A2/deg, which
is larger by a factor of 2 to 4 than the
value calculated from the electrostatic
and repulsive potential model. A possible
cause of this discrepancy is the assump-
tion of constant repulsive parameters. If
Diopside
Hedenbergite
— Observed
— Calculated
B = 87r2<r2>,
Fig. 78. Comparison of observed and calcu-
lated thermal vibration ellipsoids for the Ml
and M2 sites in diopside and hedenbergite. In
all cases, the intermediate axis is parallel to the
b axis.
GEOPHYSICAL LABORATORY
551
CVJ
0
fc_
o
■■»»"
o
o
2
*^-
a>
k.
3
<^—
O
1
k_
<D
Q_
E
<v
^— •
o
0
Q.
O
• Observed
x Calculated
Forsterite
200 400 600
Temperature, °C
800 1000
Fig. 79. Comparison of observed and calculated equivalent isotropic temperature factors at var-
ious temperatures for the Ml and M2 sites in forsterite, diopside, and hedenbergite.
those parameters actually depend on
temperature, a change of force constants
with temperature change includes two
effects — slight changes of interatomic
distances and probably substantial
changes of the repulsive parameters.
The latter might be significant at high
temperature.
In conclusion, the orientations of
calculated and observed thermal vibra-
tion ellipsoids at room temperature are
in agreement. Therefore, the observed
ellipsoids derived from least-squares re-
finement of x-ray intensity data are con-
sistent with a physically reasonable
model of atomic potential. In addition,
the electrostatic-repulsive potential
model seems to be a valid starting point.
However, in view of the systematic
discrepancy in the magnitudes of vibra-
tions, especially at high temperature,
further study is required.
552
Electrical and Optical Properties
of the Olivine Series at
High Pressure
H. K. Mao
In a previous report synthetic and
natural fayalite and its high-pressure
spinel polymorph were observed to ex-
hibit unique electrical and optical prop-
erties at high pressure (Mao and Bell,
1972). The effects are characterized by
a strong shift in the optical absorption
to low energy and a rapid increase in
electrical conductivity as the pressure is
raised about 100 kbar.* These properties
are potentially important in geophysical
models of the earth's mantle because of
the assumed dependence on Boltzmann's
law of temperature and electrical con-
ductivity. Before the new effect can be
applied to geophysical problems, it is
necessary to have data on a wider
variety of olivine compositions. For that
reason the present study includes experi-
ments on synthetic and natural olivines
on the Fe2Si04-Mg2Si04 join.
The olivines used were analyzed with
the electron microprobe. Twenty-one
synthetic olivines, ranging from pure
fayalite to pure forsterite at roughly 5
mole % intervals were provided by H. S.
Yoder, Jr. The samples were prepared
by N. F. H. Bright, and their lattice
parameters were studied by Jambor and
Smith (1962). In addition, three natural
olivines were provided by D. Virgo.
Their compositions are Fa68Fo32 (hor-
tonolite A; Bush, Hafner, and Virgo,
1970) , Fa42.5Fo57.5 (L 10-273, Anderson,
1968), and Fai8Fo82 (Bl; Bush, Hafner,
and Virgo, 1970) . The crystallography
* Fayalite is stable at 25 °C below 20 kbar
and metastable above 20 kbar. The spinel form
of Fe2Si04 is metastable at 25 °C below 20 kbar,
stable between 20 and 220 kbar, and metastable
again above 220 kbar. The measurements of
optical and electrical properties of the meta-
stable phases of fayalite and spinel at 25 °C are
valid, since both x-ray data and instantaneous
reversibility of the measurements indicate that
they are truly metastable phases under the con-
ditions of the measurements.
CARNEGIE INSTITUTION
of the three natural olivines was studied
by x-ray and Mossbauer techniques
(Birle et al., 1968; Finger and Virgo,
Year Book 70, pp. 221-225) . All samples
were examined for the pressure-induced
absorption effect in the diamond-window
pressure cell. All natural olivines and
three synthetic olivines with the com-
positions Fa46Fo54, Fa3oFo7o, and Fa22-
Fo78 were placed with electrical leads in
the diamond cell to measure the electrical
conductivities. The technique, which in-
cludes the simultaneous measurement of
pressure, electrical conductivity, and
optical absorption, is described by Mao
and Bell (Year Book 71, pp. 520-524).
The samples studied show that the
high-pressure electrical and optical effect
occurs not only in fayalite but also in all
ferromagnesian olivines except pure
forsterite. The only silicates in which the
effect has been discovered to date are
olivine, spinel, and magnesiowustite
(Mao, elsewhere in this Report). Other
samples that have been examined are
orthoferrosilite, almandite, stishovite,
corundum, periclase, halite, and all
potassium halides.
The optical shift can be observed visu-
ally in the diamond pressure cell. As the
absorption edge sweeps through the
visible wavelength region, a strong dark-
ening is observed, in contrast to the
bright perimeter of the cell where the
pressure is lower. Figure 80 shows a
series of photomicrographs of this effect
taken through the diamond window of
the cell. The octagonal face of the gem-
cut diamond can be recognized, and the
darkening can be observed in all the
compositions. However, the intensity of
the absorption increases with increasing
fayalite content as well as with increas-
ing pressure. Pure fayalite becomes com-
pletely opaque at 150 kbar, but the
absorption in Fai8Fo82 is barely dis-
cernible at 30 kbar. The absorption effect
was not observed in Fa < 10 mole % at the
maximum pressures obtained (approxi-
mately 320 kbar) .
GEOPHYSICAL LABORATORY
553
Fig. 80. Photomicrographs of olivine taken through a diamond window of the pressure cell.
The octagonal outline is the shape of the anvil face of the diamond. The pressure distribution in-
side the octagon is nearly parabolic, with the maximum at the center and 1 bar at the edge.
The dimension of the octagon is 280 fim. The sample is compressed by the two opposite diamond
anvils into a thin foil, approximately 5 /tm thick, covering the whole octagonal area. The two
long black bands across the octagon are tungsten leads used to measure electrical resistance. The
samples and conditions are as follows. First row: Fa22Fo78; left to right, 57, 204, and 325 kbar.
Second row: Fa42.5F057.5i left to right, 110, 185, and 325 kbar. Third row: Fa6sFos2; left to right,
64, 130, and 310 kbar.
Electrical conductivity was measured
at various pressures, and like the optical
shifts the effects are completely rever-
sible. The conductivity was measured
stepwise as the pressure was cycled up
and down, and virtually no hysteresis
was observed. Figure 81 shows a plot of
electrical conductivity (10~12-10~2 (ohm
cm)"1) versus pressure (1 bar to 350
kbar) for the six olivines and includes
the data for fayalite and spinel. In
olivines containing less than 22 mole %
fayalite, the conductance was too low to
measure by the present technique until
the pressure increased the conductivity
to approximately 10"10-5 (ohm cm)-1. All
the olivines, however, show significant
increases in conductivity as a function of
pressure. The actual increment of con-
ductivity is systematically dependent
upon the fayalite content. In the 1 bar—
300 kbar interval, the conductivity of
fayalite increases six orders of magni-
tude, while that of the Fa42.5Fo57.5 olivine
increases only three orders of magnitude.
The results of this study suggest that
the optical and electrical effects are
related to the amount of iron present in
the olivines, regardless of whether they
are synthetic or natural. The fact that
these phenomena are not observed in
most other iron-bearing silicate and
oxide minerals suggests that the structure
site of iron plays an important role. The
mechanism is still not urderstood, but
the fact that the effect occurs through the
554
CARNEGIE INSTITUTION
200
Pressure, kbar
Fig. 81. Electrical conductivity of olivine series as a function of pressure at 25 °C. N, natural
olivine; S, synthetic olivine.
olivine series is established by the present
data. The significant change in conduc-
tivity with pressure must be considered
in models of the earth's mantle in which
olivine is a major constituent.
Observations of Optical Absorption and
Electrical Conductivity in Magnesio-
wtiSTiTE at High Pressures
H. K. Mao
Knowledge of the earth's mantle can
be derived from measurements of physi-
cal properties of phases believed to exist
in the mantle. If the chemical composi-
tion of the mantle is predominantly
(Mg,Fe)2Si04, as most models assume,
the stable phase is olivine in the upper
mantle, spinel in the transitional zone,
and probably magnesiowustite, (Mg,Fe)0,
and stishovite, Si02, in the lower mantle.
A major effect of pressure on the optical
and electrical properties of olivine and
spinel was discovered (Mao, preceding
section) , and it is important to examine
magnesiowustite as well.
GEOPHYSICAL LABORATORY
555
Behavior similar to that observed in
olivine and spinel was observed in
magnesiowlistite; that is, the optical ab-
sorption and electrical conductivity of
magnesiowiistite increase rapidly with
pressure. The compositions of the syn-
thetic magnesiowiistite studied were
Wu56Pe44, Wu39Pe6i, Wu22Pe78, and
Wu9Pe9i, where Wii stands for wlistite
(FeO) and Pe, for periclase (MgO). The
first two samples were provided through
the kindness of Dr. A. Muan. They were
synthesized from a mixture of Fe203 and
MgO by two successive 24-hour heatings
at 1100°C and P02 = 10"11 atm, with
intermediate grindings. The last two
samples were synthesized in an H-C02
gas-mixing furnace at 1150°C and P02 =
10"11-5 atm, and their compositions were
confirmed by electron microprobe anal-
yses. Observations of optical absorption
and electrical resistance were made with
the diamond-windowed pressure cell,
using the same techniques employed for
the olivines (Mao, preceding section) .
The effect of pressure on the optical
absorption properties of the magnesio-
wlistite series is large; however, for the
same Fe:Mg ratio, the magnesiowlistite
absorbs in the visible region at much
lower pressure than olivine and spinel.
Pure wlistite is nearly opaque at 1 bar
pressure. The compositions Wli56Pe44,
Wii39Pe61, and Wli22Pe78 begin to absorb
strongly at approximately 20, 80, and
150 kbar, respectively. The optical
absorption of Wli9Pe9i becomes intense
at high pressures, but the composition
does not turn completely opaque even at
300 kbar. Photomicrographs of the mag-
nesiowlistite series at high pressures are
shown in Fig. 82.
The optical absorption spectra of
Wli27Pe78 at high pressures were meas-
ured in a modified Cary 17-1 spectro-
photometer with dual microscope. The
techniques were described by Mao and
Bell (1973, and Year Book 71, pp. 608-
611). The area where the absorption
spectra were observed in the diamond cell
was limited by diaphragms and focusing
devices to a 50 /xm-diameter circle in the
center of the cell, where the pressure was
maximum and the pressure gradient was
minimum. The range of pressure within
the circular area is believed to be less
than 4% of the average pressure in the
area. The results are shown in Fig. 83
as a plot of absorption coefficient versus
wave number at several pressures. The
low-pressure spectrum shows an intense
absorption edge or shoulder in the near
ultraviolet region, and a crystal-field
band of the octahedrally coordinated
Fe2+ at approximately 10,000 cm-1 wave
numbers (Low and Weger, 1960). With
increasing pressure, the crystal-field
band shifts to higher energy, as expected,
but the absorption edge shifts rapidly
to lower energy and obscures the crystal-
field band. The absorption becomes ex-
tremely intense, and it is clear that
radiative transfer of heat would be
severely blocked at high pressure.
At relatively low pressures the pres-
sure-induced increase in electrical con-
ductivity of the magnesiowlistite is
Fig. 82. Photomicrographs of magnesiowlistite through a diamond window of the pressure cell.
For detailed explanation see Fig. 80. (A) WlioPeoi at 310 kbar. (B) WiissPew at 310 kbar. (C)
WlUPeei at 275 kbar.
556
CARNEGIE INSTITUTION
400
10,000
Wavelength, nm
500 600 800 1000
T
2000 3000 oo
i r
8,000 -
6,000
£- 4,000
o
v>
JQ
<
2,000 -
15 10
Wave number, I03cm-1
Fig. 83. Optical absorption spectra of Wu22Pe78 at high pressure and room temperature. Dashed
curves indicate region where the detector sensitivity is low.
similar to that of olivines; i.e., the
conductivity increases approximately
ten-fold at each 50 kbar increment of
pressure. At higher pressures, however,
the electrical conductivity of the mag-
nesiowlistite increases by only a factor of
approximately 1.2 per 50 kbar. Figure
84 shows a plot of conductivity versus
pressure for the four compositions. The
point where the rate of increase of con-
ductivity with pressure is sharply re-
duced corresponds to the pressure where
strong absorption occurs. A sharp change
in rate is not observed, however, for
Wu9Pe9i, which did not begin to absorb
at 310 kbar.
The temperature dependence of the
electrical conductivity of Wu22Pe78 was
studied at 150 kbar in the diamond cell
with an external heater, and the results
are plotted in Fig. 85. The measurements
were reproducible when the temperature
was cycled between 300° and 700 °K and
are linear in the range 400° to 700 °K,
suggesting a single mechanism of con-
ductivity. Above 700 °K, the conduc-
tivity started to drift, indicating possible
oxidation, reduction, or extrusion of the
sample. The pressure-induced absorption
was also examined in this temperature
range, but no change of absorption due
to temperature was observed. The
logarithm of the electrical conductivities
of Wu22Pe78 at 150 kbar fits the straight
line:
log a = log a0 — A/2.303 kT
where o- is the conductivity; o-0 and A,
constants; k, Boltzmann constant; and
T, temperature in degrees K. The pa-
rameters (T0 and A were found to be 101-34
ohm-1 cm"1 and 0.37 eV, respectively.
GEOPHYSICAL LABORATORY
557
200 300
Pressure, kbar
Fig. 84. Electrical conductivity of four magnesiowiistites as a function of pressure.
It is tempting to relate the strong
optical absorption and high electrical
conductivity at high pressures with non-
stoichiometry, especially because the
nonstoichiometry of wiistite is known
to affect its electrical conductivity
(Adler, 1968; Tannhauser, 1962). How-
ever, the olivines and spinels, which show
high-pressure behavior similar to that
of the magnesiowiistites, are stoichio-
metric, according to x-ray and Moss-
bauer studies. It would appear, therefore,
that the mechanism is unrelated to
stoichiometry.
Although the mechanism of the strong
pressure effect is not yet understood, the
effect may be important in the earth's
mantle. Pressure increases the optical
absorption and blocks radiative heat
transfer. At the same time, pressure
increases heat transfer by electronic and
possibly by excitonic processes. All the
major sources of thermal conduction
should be notably affected by the
phenomenon observed in this study.
Thermal and Electrical Properties
of the Earth's Mantle
H. K. Mao
Recent experiments on the phase
equilibria and the electrical, optical, and
elastic properties of materials thought to
occur in the earth's mantle suggest that
558
CARNEGIE INSTITUTION
800
600
Temperature, °K
500 400
300
1.0 2.0 3.0
1000/T, °K-1
Fig. 85. Electrical conductivity of Wu22Pe78 at 150 kbar as a function of temperature.
many of the physical constraints em-
ployed in currently accepted geophysical
models are inadequate. The purpose of
this study has been to reexamine the lines
of argument, to revise parameters for
geophysical models within the limitations
of the recent data, and to suggest new
experiments for constructing better
models.
The major chemical component of the
earth's mantle is thought to be magne-
sium-rich (Mg,Fe)2Si04. Recently, it
has been experimentally demonstrated
that Fe2Si04 spinel reacts to wiistite and
stishovite at 200 kbar (Bassett and
Takahashi, 1970; Mao and Bell, Year
Book 70, pp. 176-178; Bassett and Ming,
1972). Both shock-wave experiments
and thermodynamic calculations (Mao,
1967; Mao et al, 1969) indicate that the
magnesium-rich (Mg,Fe)2Si04 spinel is
also unstable relative to magnesiowustite
and stishovite under the conditions of the
lower mantle. Therefore, most physical
models of the earth's mantle calculated
on the basis of the properties of the low-
pressure phases such as olivine, spinel,
and pyroxene should be revised to ac-
count for the observed reactions.
Recent observations of the strong pres-
sure dependence of optical absorption
and electrical conductivity of olivine,
spinel, and magnesiowustite (Mao and
Bell, 1972; Mao, elsewhere in this Re-
port) also affect current views on the
electrical and thermal behavior of mantle
materials. With recent measurements of
parameters of equations of state (Ahrens,
Takahashi, and Davis, 1970; Mao et al.,
1969) and assignments of optical spectra
(Burns, 1970; Mao and Be\\,Year Book
71, pp. 524-527; Runciman, Sengupta,
GEOPHYSICAL LABORATORY
559
and Gourley, 1973) one can better com-
prehend thermal conductivity in the
mantle. In the present study, these new
data are used to update the state of
knowledge of electrical conductivity,
phonon thermal conductivity, radiative
heat transfer, exciton thermal conduc-
tivity, and electronic thermal conduc-
tivity of the earth. These parameters are
essential in constructing models of tem-
perature distribution and thermal history
of the earth.
Electrical Conductivity and Temperature
Distribution of the Mantle
The radial distribution of electrical
conductivity of the earth has been esti-
mated from magnetic variations (Fig.
86A). At a constant pressure, the elec-
trical conductivity, o-, is related to the
absolute temperature, T,
cr = S<70texp(-A4/fc!T), (1)
i
where i denotes a specific conducting
mechanism, k is the Boltzmann constant,
cr0l and Ai are preexponential term and
activation energy, respectively, and are
pressure dependent. Temperature dis-
tribution of the earth's mantle has been
calculated from electrical conductivity
based on extrapolated values of a0i and
At of olivine (Tozer, 1959). The calcula-
tions were revised using extrapolated
values of a0i and At of Fe2Si04 spinel
(McKenzie, 1967) ; previous experi-
mental measurements of conductivity, on
which these extrapolations are based,
however, were made at pressures below
60 kbar (e.g., Bradley, Jamil, and
Munro, 1964; Hamilton, 1965; Duba,
1972).
At pressures above 60 kbar, it was dis-
covered (Mao and Bell, 1972) that the
conductivities of the olivine and spinel
forms of Fe2Si04 are highly pressure
dependent. Under room conditions,
metastable Fe2Si04 spinel is a better
conductor than fayalite by a factor of
100. At 300 kbar and 25°C, however,
Depth, 10° km
Fig. 86. (A) Electrical conductivity of the
mantle: (1) Eckhart, Larsen, and Madden,
1963; (2) McDonald, 1957; (3) Banks, 1969;
(4) Rikitake, 1966. (B) Electrical conductivity
of olivine at 1 bar and magnesiowiistite at 150
kb: (1) Shankland, 1969; (2) Noritomi, 1961;
(3) Hughes, 1953; (4) Duba, 1972; (5) Kobaya-
shi and Maruyama, 1971 ; (6) Mizutani and
Kanamori, 1967; (7) Magnesiowiistite; Mao,
this Report. (C) Temperature within the man-
tle estimated from electrical conductivity by
Tozer (1959), Tu, and by McKenzie (1967), Tc.
(D) Thermal conductivity in the mantle esti-
mated by (1) Lawson and Jamieson (1958) and
(2) Lubimova (1967), due to various mecha-
nisms: ph, phonon; r, radiative; ex, exciton;
t, total.
metastable fayalite increases in conduc-
tivity by six orders of magnitude and
becomes a better conductor than stable
Fe2Si04 spinel, suggesting that extra-
polation of the olivine-spinel data was
not justified. The recent data indicate
that new electrical mechanisms occur at
pressures equivalent to those at the base
of the mantle's transition zone and there-
fore imply that a large uncertainty exists
in thermal models that relate tempera-
ture and electrical conductivity.
Olivine and spinel are unstable in the
lower mantle, and if one assumes the
560
CARNEGIE INSTITUTION
2000
Temperature, °K
1000 800 600 500
0.5
2.0
.0 1.5
1000/T, °K
Fig. 86B.
lower mantle to be made of magnesio-
wustite and stishovite, with the bulk
composition of olivine, magnesiowustite
will occupy 60% of the volume. Stisho-
4
1 ' 1
1
/^*
x.
0
to
2 3
— 1
<u
=J
o
~
i_
/ ^^""t„
Q>
/ j>r c
CL
E ^
5 2
— / s
—
h-
- f
-
1
1,1,1
C
-
0
Depth, I03 km
Fig. 86C.
Depth, I03 km
Fig. 86D.
vite is predicted to be a poor conductor,
so the conductivity of such a mixture
may be determined mainly by magnesio-
wustite (Tolland and Strens, 1972) . The
present analysis considers a magnesio-
wustite with Fe/(Fe + Mg) — 0.22, a
ratio suggested by Press (1968) for the
lower mantle.
The electrical conductivity of Wu22-
Pe78 magnesiowustite has three impor-
tant features (Mao, this Report) . First,
it increases rapidly with pressure. Sec-
ond, the absolute conductivity is very
high ; at 150 kbar, it is six orders of magni-
tude higher than that of olivine with the
same Fe:Mg ratio at the same pressure
(Fig. 86B). Third, the activation energy
is relatively low; with A = 0.37 eV, the
conductivity mechanism is most likely to
be extrinsic. A consequence of these find-
ings is that the argument in which the
mechanism of extrinsic conduction in the
mantle is ignored (Tozer, 1959; Mc-
Kenzie, 1967) is no longer valid. It was
believed that the extrinsic conductivity
of olivine or pyroxene would not be high
enough to account for the high conduc-
tivity of the mantle, but for pure
Wu22Pe78 magnesiowustite, the conduc-
GEOPHYSICAL LABORATORY
561
tivity extrapolated to 300 kbar and
1500 °K (using extrinsic activation
energy, A = 0.37 eV) is actually higher
than the calculated conductivity of the
mantle.
If extrinsic conduction cannot be ex-
cluded, the low activation energy will
lead to an exceedingly large uncertainty
in the estimation of the mantle tempera-
ture from conductivity. With A = 0.37
eV, an uncertainty of two orders of
magnitude in the estimated conductivity
distribution in the mantle is equivalent
to an infinite temperature uncertainty,
and temperature estimation is therefore
meaningless. Although the conductivity
becomes insensitive to temperature be-
cause of the low activation energy, it
becomes increasingly sensitive to the
composition and, conversely, can be used
to estimate the mantle's composition.
Ignoring the effects of pressure and com-
position on A, the Fe/(Fe + Mg)
ratio at 1000 km is approximately 0.15,
for temperatures between 1500° and
3000°C.
Heat Transfer in the Mantle
In most current models of heat trans-
fer (Lubimova, 1967; Lawson and Jamie-
son, 1958; Clark, 1957), the earth's
mantle is regarded as an electrical
insulator or semiconductor. Near the
earth's surface, heat is conducted by
lattice vibrations (phonons) in an elec-
trical insulator or semiconductor. At
temperatures above 800°-1500°C, a sig-
nificant amount of heat can be trans-
ferred through the mantle by radiation
(photons). At still higher temperatures,
heat transfer by excitons becomes ap-
preciable, but because of the mantle's
relatively low electrical conductivity, the
energy transferred by free electrons is
thought to be insignificant. The total
thermal conductivity is a summation of
conductivities of all mechanisms (Fig.
86D), and it is useful to reexamine the
various processes.
Phonon conductivity. The thermal
conductivity due to lattice vibrations
(phonon) , Kph, may be expressed by
Kphz= (aBT^)/(3y2P^T), (2)
where a is the lattice constant, BT the
isothermal bulk modulus, y the Griieisen
parameter, and p the density (Lawson,
1957) . It is apparent that a relatively
incompressible (large BT) mineral usu-
ally has high Kph. Under room condi-
tions, the Kph of dense incompressible
stishovite, estimated from equation 2,
and that of Wui5Pe85 magnesiowiistite
are 0.07 and 0.09 cal cm-1 sec-1 deg"1,
respectively. The estimations are based
on data on equations of state (Mao,
1967; Mao et al, 1969; Ahrens, Taka-
hashi, and Davis, 1970; Anderson and
Anderson, 1970) and thermal conduc-
tivity of quartz and periclase (Mc-
Quarrie, 1954). A mixture of stishovite
and magnesiowiistite has Kph four times
as high as that of the low-density com-
pressible phases, olivine and pyroxene,
which were used to calculate the phonon
conductivity of the mantle in previous
models.
As implied by equation 2, high pres-
sure increases Kph and high temperature
decreases Kp1l. From the parameters of
the equation of state of stishovite and
magnesiowiistite, the estimated pressure
effect far exceeds the temperature effect
under the conditions of the lower mantle.
At the bottom of the mantle (2900 km
deep), Kp1l of the mixture of stishovite
and magnesiowiistite increases ten-fold
and approaches 1 cal cm"1 sec"1 deg"1,
which is one or two orders of magnitude
higher than the phonon conductivity pre-
viously estimated on the basis of the data
of the surface rocks (Fig. 86D). This
value is comparable to the total thermal
conductivity estimated previously. The
present result indicates that the phonon
conductivity in the lower mantle could
be a major mechanism of heat transfer.
Radiative transfer. The energy trans-
fer by photons through a "gray body"
562
CARNEGIE INSTITUTION
(absorption of photon finite and non-
zero) is given by
Kr = 16n2>ST3/3e,
(3)
where n and e are the mean refractive
index and the mean opacity of the gray
body, and are temperature and pressure
dependent, and S is the Stef an-Boltzmann
constant. Because Kr increases propor-
tional to T73, it was thought to be the domi-
nant mechanism under the high-tem-
perature conditions of the mantle.
The strong, pressure-induced optical
absorption in olivine, spinel, and mag-
nesiowiistite (Mao and Bell, 1972; Mao,
accompanying reports) suggests that
radiative heat transfer is probably not
so important in the mantle as the T3 law
implies. At 300 kbar, a pressure corre-
sponding to the top of the lower mantle
(1000 km deep), the absorption coef-
ficient of Fe22Mg78 magnesiowiistite at a
wavelength of 1500 nm increases to 1500
cm-1, and at 700 nm increases to 10,000
cm"1 (25°C). The absorption is 1000
times stronger than the absorption of
olivine usually proposed for radiative
transfer in the mantle (Clark, 1957) and
will effectively block radiation at wave-
lengths shorter than 1500 nm. Calculated
from the function of black-body radia-
tion, 75% of the radiative energy is in
the wavelength range 0-1500 nm at
3000°K, and 87% is in this range at
4000°K. The observed absorption (25°C)
is sufficient to eliminate the major part of
radiative heat transfer. At temperatures
of the mantLe, high-pressure absorption
is likely to broaden and spread to longer
wavelengths, thus completely excluding
radiative transfer in magnesiowiistite.
Higher pressures or higher iron content
can cause the same effect.
Even though magnesiowiistite may
completely block direct thermal radia-
tion, radiative heat transfer through a
heterogeneous mixture of magnesio-
wiistite and stishovite in the lower
mantle may still be effective. Possibly
the radiative energy could pass through
contacting stishovite crystals by a series
of absorption-reemission steps. This
process strongly depends on the opacity
of stishovite, the amount of magnesio-
wiistite present, and the fabric of the
assemblage.
In the transition zone of the mantle,
ferromagnesian silicate spinel is presum-
ably the major phase. The Fe2Si04 spinel
becomes opaque at the pressures of the
transition zone (Mao and Bell, 1972). If
the same effect occurs in the magnesium-
rich (Fe,Mg)2Si04 spinel by analogy
to olivine and magnesiowiistite, radiative
heat transfer in the transition zone will
be reduced.
In the upper mantle, the pressure alone
is not high enough to induce a strong
optical absorption in a magnesium-rich
olivine. Radiative transfer is controlled
mainly by the broadening of the Fe2+
crystal-field band at high temperatures.
Simultaneous measurement of the lattice
and radiative conductivity at room pres-
sure (Schatz and Simmons, 1972) indi-
cates that the increase of Kr with tem-
perature is far less than the amount
predicted by the T3 law. At the pressures
prevailing in the lower part of the upper
mantle, the combined effect of tempera-
ture- and pressure-induced absorption
may further reduce radiative heat
transfer.
The observed increase of electrical
conductivity in olivine, spinel, and mag-
nesiowiistite also increases the absorp-
tion of photons by free carriers. The
absorption due to a free carrier is given
by
e/ctcm"1) = 120«tr(©)/n, (4)
where o-(w) in ohm-1 cm"1 is the electrical
conductivity at the same angular fre-
quency, (o, as the photon. Nevertheless,
the electrical conductivities of these
minerals are too low. Even if the electri-
cal conductivity increases by several
orders of magnitude at high pressure, the
resultant opacity is less than the order of
1 cm-1, which is negligible in comparison
GEOPHYSICAL LABORATORY
563
to the strong optical absorption at high
pressures. At high pressures and tem-
peratures, the electrical conductivity
increases by several more orders of mag-
nitude, and its effect could be significant
as the free carriers absorb across the
whole spectral range.
Transfer of energy by excitons. In
previous models of energy transfer by
excitons, six optical absorption bands of
olivine were attributed to exciton transi-
tions. The calculation based on assumed
exciton bands at 0.31 and 1.19 eV (Law-
son and Jamieson, 1958) indicates that
the exciton thermal transfer could be
important in the bottom of the mantle.
The calculation based on assumed ex-
citon bands at 0.40, 2.80, 3.12, and 3.60
eV (Lubimova, 1967) suggests that the
exciton thermal transfer exceeded all
other mechanisms at depths >100 km
and increases with depth more rapidly
than the other mechanisms (Fig. 86 D).
With the new data available, it be-
comes apparent now that probably none
of the six bands is an intrinsic exciton
band. The band at 1.19 eV is a crystal
field band of octahedrally coordinated
Fe2+ (Burns, 1970; Runciman, Sengupta,
and Gourley, 1973; Mao and Bell, Year
Book 71, pp. 524-527). The bands at
2.80 and 3.12 eV have intensities pro-
portional to the iron content in olivine
and therefore are more likely assigned
as charge-transfer bands of Fe3+ rather
than the type of exciton band related to
the intrinsic absorption edge of olivine at
3.60 eV (Mao and Bell, Year Book 71,
pp. 524-527). The bands at 0.31 and
0.40 eV were observed as a cluster of
four sharp, weak bands (White and
Keester, 1966), whereas the proposed
exciton bands (with the quantum num-
ber n — 1) should not form such a
multiplet.
In addition to the problems of band
assignment, exciton transitions in olivine
have application limited to the upper
mantle where olivine is stable. The ex-
citon mechanism may possibly become
important, however, at high pressures.
Recent observations of the pressure-
induced optical absorption in olivine,
spinel, and magnesiowiistite could be
caused by the pressure-intensified ab-
sorption of a localized exciton. The
exciton localized at the Fe2+ site could
move to another Fe2+ site by "hopping,"
thereby transferring energy.
Electronic transfer of energy. The
electronic thermal conductivity due to
electron-hole ambipolar-pair diffusion
of an intrinsic semiconductor is of the
form (Price, 1955)
Kel = *T(k/e)2 (n + a0)
0= (w/«r2) [(Eg/kT)2 +
a(Eg/kT) + /?]
(5)
where <rE and crH are the contributions of
electrons and holes to the electrical con-
ductivity, o- = aE + &H is the total elec-
trical conductivity, e is the charge of an
electron, Eg is the gap energy, a and fi
depend on the relaxation processes, and
O0 is a classical constant. The high elec-
trical conductivity of olivine, spinel, and
magnesiowiistite recently observed at
high pressure and high temperature sug-
gests that Kei in the lower mantle could
be appreciable. Taking aE = <th — 5
ohm-1 cm-1 for magnesiowiistite at
3000 °K and 800 kbar, equivalent to
lower mantle conditions, and the energy
gap of MgO at Eg = 8.7 eV (Reiling and
Hensley, 1958; Cohen et al, 1967) , KPl =
0.02 cal cm"1 sec-1 deg-1. This calculation
is at best only qualitative, but it shows
that the electronic thermal conductivity
may be significant at the pressures and
temperatures of the lower mantle. Fur-
thermore, phase transitions or electronic
transitions that increase the electrical
conductivity abruptly by several orders
of magnitude may occur at extremely
high pressures, as is the case for FeoOs,
Cr203, and Ti02 (Kawai and Mochizuki,
1971). The electrical properties are
nearly metallic, and electronic thermal
conductivity becomes predominant.
564
CARNEGIE INSTITUTION
Conclusions
At this time it can be concluded that
the electrical conductivity of the consti-
tuents thought to occur in the earth's
lower mantle is probably greater than
previously believed by four orders of
magnitude. The Fe/(Fe + Mg) ratio of
the mantle at a depth of 1000 km, esti-
mated from the present data on electrical
conductivity, is 0.15.
The value determined for the phonon
thermal conductivity at the bottom of
the mantle is approximately 1 cal cm"1
sec-1 deg"1, which is one to two orders of
magnitude higher than previous values.
The process of phonon transfer could,
therefore, be a more important mecha-
nism of conductivity in the lower mantle
than previously believed. The process of
radiative heat transfer, which had been
postulated to be the major mechanism in
the mantle, is severely blocked in the
upper mantle and the transition zone.
That process may become important,
however, if an absorption-reemission
process occurs in the lower mantle.
Exciton thermal conductivity, which
was considered as another major mecha-
nism in the lower mantle, may occur, but
its significance is uncertain. All pre-
viously assigned exciton bands are prob-
ably invalid, but the present results
indicate the possibility of localized ex-
citon enhancement caused by high pres-
sure. Electronic thermal conductivity,
which had been deemed negligible, could
be 0.02 cal cm"1 sec"1 deg"1 or higher at
a depth of 1800 km. The present analysis
points to the need for revision of all
previous models of the electrical and
thermal properties in the mantle. It may
be possible to construct a much more
quantitative model when sufficient data
become available. The desirability of
extending the high-pressure, high-tem-
perature experiments is immediately
apparent, and it will be advantageous to
experiment with a judicious set of com-
positional mixtures now that the pressure
effects are known to be important in heat
transfer.
Polarized Absorption Spectra of
GlLLESPITE AT HlGH PRESSURE
R. M. Abu-Eid* H. K. Mao, and R. G. Burns*
The rare mineral gillespite, BaFeSi4Oi0
(Pl/ncc; a0 = 7.495 A; c0 = 16.05
A; z = 4), is of considerable interest, not
only because it contains high-spin Fe2+
ions in square-planar coordination
(Fe2+-0 = 1.97 A), but also because the
crystal-field splitting of the ferrous 3d
orbitals is larger in the gillespite struc-
ture than in any other ferromagnesian
silicate. Since crystal-field splittings in-
crease with decreasing metal-oxygen
distances and increasing pressure, spin-
pairing is predicted to take place in Fe2+
ions in gillespite at elevated pressures.
Strens (1966) reported a featureless
spectrum and the disappearance of color
in gillespite above 26 kbar, which he
attributed to a high-spin to low-spin
transition in Fe2+. These observations do
not conform with properties of com-
pounds containing isoelectronic low-spin
ions (e.g., Co3+) which are colored and
give spectra with intense absorption
bands. Such inconsistencies have led to
further high-pressure spectral studies of
gillespite and an alternative interpreta-
tion of the observed transition at 26
kbar.
Measurements were made in a dia-
mond-anvil cell containing a single
crystal of gillespite mounted in oil with
the basal cleavage parallel to the pres-
sure axis. With this orientation both the
E 1 1 c and E 1 c spectra could be meas-
ured in polarized light. At atmospheric
pressure, gillespite has the pleochroic
scheme: 2£||c, bright red; E±c, pale
pink. The intensities of these colors
correlate with the polarized absorption
spectra illustrated in Fig. 87, which
shows the absorption band centered
around 500 nm to be most intense in
the E \\c spectrum. Note the polariza-
tion dependence of the 1200-nm band
* Department of Earth and Planetary Sci-
ences, Massachusetts Institute of Technology.
GEOPHYSICAL LABORATORY
565
1.5
.0
0.5
0.0
E-Spectrum
I bar
above 26 kbar
O-Spectrum
I bar
above 26 kbar
500
1000
Wavelength, nm
500
Fig. 87. The polarized absorption spectra of gillespite at 1 atm and above 26 kbar.
that occurs only in the E ±c spectrum,
defining the Fe2+ ground state as 5Alg
and the 1200-nm and 500-nm bands as the
5Alg -> 5B2g and 5Alg -» 5Blg transitions,
respectively (Burns, Clark, and Stone,
1966). As the pressure is increased, ab-
sorption maxima of all bands first move to
higher energies (shorter wavelengths).
Above 26 kbar, however, both bands
broaden, intensify, and move to lower en-
ergies with rising pressures, whereas the
band around 500 nm becomes a doublet
(Fig. 87). There is a corresponding change
in color, with E \\ c bright blue and E 1 c
pale blue. These results and observations
differ from those reported by Strens
(1966). The discrepancies may be due
to his having used a (001) cleavage frag-
ment of gillespite mounted perpendicular
to the pressure axis. With this orienta-
tion only light polarized E 1 c could be
studied. The polarization dependence in
the present measurements of the band
around 1200 nm suggests that the ground
state may be 1Alg, corresponding to dia-
magnetic low-spin Fe2+ if the transition
observed at 26 kbar represents spin-
pairing. One apparent inconsistency,
however, is that transitions to the singlet
B2g and Blg states occur at lower energies
than those to the corresponding quintet
states for high-spin Fe2+; higher energy
bands might be expected since shorter
metal-oxygen distances are anticipated
in the low-spin state.
An alternative interpretation of the
changes at 26 kbar is a displacive phase
transition of the gillespite structure, with
a change in site symmetry of Fe2+ ions
from square planar to distorted tetra-
hedral coordination. Pressure directed
along the cleavage planes of gillespite
may lead to distortion of the four-
membered ring of linked Si04 tetrahedra
(Pabst, 1937) and puckering of the
square planar Fe2+ site, thereby produc-
ing an oxygen coordination polyhedron
intermediate between square planar and
tetrahedral symmetry (Fig. 88) . Support
for this interpretation stems from high-
pressure Mossbauer studies of synthetic
gillespite enriched in 57Fe (Tossell and
Vaughan, unpublished results). The
Mossbauer spectrum above 26 kbar con-
sists of two doublets with parameters
(based on the stainless steel standard) :
inner doublet, isomer shift 0.87 mm/sec,
quadrupole splitting 0.64 mm/sec; outer
doublet, isomer shift 1.13 mm/sec, quad-
rupole splitting 2.02 mm/sec. The pa-
rameters of the inner doublet correspond
to those expected for square planar Fe2+
ions; those for the outer doublet fall into
the range expected for tetrahedrally co-
566
CARNEGIE INSTITUTION
o
d 6 6 £
• © o •
VSZfr
#
Q
QJ
XI
-I-3
M-H
o
d
o
•rH
-1-3
«
a
• *—*
o
u
ft
0>
/»— s
4b*
1— 1
O
«J
O
^.
0
a3
k
02
■o
*
•
O
X
^3
0
CO
k
CU
*•
J-H
tt
3
h-
bD
"O
cu
0)
^5
k.
H
o
«/>
d
• _
o
G
°02
>»
f3
k.
03
0)
S-l
-4-3
>
03
c
-Q
Jbjl
a>
CD
u_
CM
CU
-t-i
><
cu
>
02
a
m
o
,£5
o
+3
O
03
03
•
T3
ti
«o
cp
CU
"«•
+=
bfi
ii
-t-3
o
O
-d
ft
a
T3
03
a
CU
c3
fe
a
73
-4-2
CU
c3
■+3
a
1— 1
cu
■+3
"cu
tt>
03
CO
(/>
T3
CU
O
k.
0
c
'a
0
'■+J
o
in
o
to
0) ft
o ©
-1-3
CO a3
■^ .-s
ft s
co I?
r-t CU
CO CU
00 gf
00 ^
S 2
GEOPHYSICAL LABORATORY
567
ordinated Fe2+ ions (Bancroft, Burns,
and Maddock, 1967). The Mossbauer
evidence for two coexisting spin-states of
Fe2+ of two coordination symmetries
correlates with the general trend in
high-pressure Mossbauer spectroscopy
for the presence of two iron species in
equilibrium over a wide pressure range
above the transition pressure (Drick-
amer and Frank, 1973).
The two alternative interpretations of
the transition in gillespite at 26 kbar
could be resolved from high-pressure
x-ray diffraction data. Appreciable
shortening of the Fe2+-0 distances in the
high-pressure phase would support the
Fe2+ spin-pairing hypothesis, whereas
atomic parameters showing pucker-
ing of the planar oxygen atoms would
indicate a displacive phase transforma-
tion and tetrahedrally coordinated Fe2+.
The Fe3+ Site Preference in the Solid
Solution Series MgCr204-MgFe204
D. Virgo and G. Ulmer
The crystal chemistry of the spinel
solid solution series MgCr204-MgFe204
is of interest because Mg[Cr2]04 is a
normal spinel and magnesioferrite,
Fe3+[MgFe3+]04, is an inverse spinel. A
preliminary inquiry into the transition
from normal to inverse crystal structure
across the series was made from measure-
ments of 57Fe Mossbauer effect at 298°K.
The samples studied were synthesized
at 1300 °C in air (Ulmer and Smothers,
1968) and have compositions at 10 wt %
intervals on the join MgCr204-MgFe204
in the system MgO-Cr203-Fe203. In
actual fact, this system is not strictly
ternary because of varying amounts of
divalent iron. In the end-member com-
position MgFe204 the Fe2+ content is
1.0 wt % FeO but decreases rapidly with
increasing amounts of Cr3+ (Ulmer and
Smothers, 1968).
The hyperfine parameters, quadrupole
splitting, and isomer shift are listed in
Table 22, and the spectrum of the sample
TABLE 22. Composition, Hyperfine Param-
eters, Quadrupole Splitting, and Isomer Shift
at 298 °K for the Series MgCr204-MgFe204
Quadrupole
Isomer
Sample
MgFe204,
Splitting,
Shift, t
No. *
wt %
mm/sec
mm/sec
XB
10
0.43
0.33
IXB
20
0.47
0.32
VIIIB
30
0.50
0.31
VIIB
40
0.52
0.31
VIB
50
0.57
0.31
VB
60
0.56
0.30
* Ulmer and Smothers (1968).
t Referred to metallic Fe at 298°K.
containing 60 wt % MgFe204 is plotted
in Fig. 89. The substitution of trivalent
iron for chromium in samples containing
10-60 wt % MgFe204 results in the
appearance of a single, well-resolved
quadrupole split doublet (Fig. 89). The
isomer shift (IS) and quadrupole split-
ting (A) , given in Table 22, are in
000
.!*•
..'.
•
'.-.
..i -X-I--
i*.l
•«
* </"■" a*w •
z
i.«
wPT
f*y. *
□
i
■
H
JjSm
"t
h-
"1
CL
\r
a
in
& 020-
I
1 ^
h
z
EC
z
a
in
LI
OH 040-
t>" *
m/ sec
Fig. 89. Mossbauer absorption spectrum of
spinel containing 60 wt % MgFe204 at 298° K.
Solid line is a least-squares fit to the data as-
suming two Lorentzian lines (six line variables,
one background variable).
568
CARNEGIE INSTITUTION
agreement with the values obtained for
octahedrally coordinated ferric iron, as
in Zn[Fe2]04 (IS = 0.39 mm/sec,
A = 0.33 mm/sec; Dobson, Linnett, and
Rahman, 1970) and Fe2+[Cr2-*Fe8+J04
(IS = -0.38 mm/sec, A = 0.46-0.42
mm/sec, for values of x from 0.25-0.75;
Robbins et al, 1971).
These values contrast with the few
examples of tetrahedrally coordinated
ferric iron, as in FeNiA104 (IS = 0.35
mm/sec, A = 1.06 mm/sec, 559°K) and
FeNiCr04 (IS = 0.32 mm/sec, A = 0.80
mm/sec at 746 °K). A slight asymmetry
in the shape and intensity of the peaks
in the spectra is probably due to the
presence of ferrous iron. In fact, ferrous
iron was detected in these spinels in both
the tetrahedral and octahedral sites from
diffuse reflectance spectra (Ulmer and
Smothers, 1968) . The absence of distinct
Fe2+ absorption peaks in the resonant
absorption spectra, however, attests to
the small concentrations of divalent iron.
The spectra of samples containing 80,
90, and 100 wt % MgFe204 exhibit a
broadened magnetic six-line pattern, but
resolution of this pattern for Fe3+ on both
the octahedral and tetrahedral sites is
not observed at 298 °K. Sawatzky, Van
der Woude, and Moorish (1969) have
shown that complete resolution of Fe3+
in the sublattices is obtained for spectra
of MgFe204 taken in an applied magnetic
field at low temperature. At this time,
no attempt has been made to further
analyze the spectra of those samples
exhibiting magnetic phenomena in terms
of their degree of inversion. It is note-
worthy, however, that the conclusions
reached by Sawatzky and co-workers
regarding the dependence of the degree
of inversion on the nature of the heat
treatment of the sample do not take into
account the presence of significant Fe2+,
which is suggested from the phase-equi-
librium studies of Speidel (1967) and
Ulmer and Smothers (1968). The spec-
tra of the sample containing 70 wt %
MgFe204 exhibit magnetic splitting but
with a quadrupole split doublet due to
iron in the paramagnetic state.
Interpretation of the Mossbauer effect
measurements indicates that spinels
along this join are characterized by iron
in the trivalent state and that for com-
positions in the range 10-60 wt %
MgFe204, iron occurs entirely in the
octahedral sites. The suggestion of two
Zeeman patterns for the compositions
70-100 wt % MgFe204 from this study
and for MgFe204 by Sawatzky, Van der
Woude, and Moorish (1969) argues for
the distribution of ferric iron on both the
tetrahedral and octahedral sites for
spinels with >70 wt % MgFe204. These
results attest to a transition to an in-
verted spinel structure in the composition
range 60-70% MgFe204 and are con-
sistent with the very strong octahedral
preference of Cr3+ (e.g., Miller, 1959).
Comparatively few other studies of
this binary join have been made. Ulmer
and Smothers (1968) argued for a
continuous replacement of [Fe3+] by
Cr3+ in MgFe204 to give a (Fe3+)IV
[Mg2+Cr3+a,Fe3+i.J YI04 solid solution.
They suggested that when x increases
to one, the addition of further Cr3+ would
normalize the solid solution structure to
(Mg2+)IV[Fe3+yCr3+2.JVI04 because Cr3+
will not readily substitute for (Fe3+)IV.
The value of x or y becomes unity at
50.98 wt % MgFe204. The observation
of a discontinuity in the cell-dimension
data at this composition supported their
discussion of the crystal chemistry of
spinels along this join. In the light of the
quantitative site occupancies for iron
determined in this investigation (at least
in the range 10-60 wt % MgFe204),
however, a more complicated variation
in the crystal chemistry must be inferred.
In an attempt to unravel this crystal
chemistry, the Curie-temperature mag-
netic behavior of this solid solution series
was also examined. The Neel tempera-
ture (paramagnetic to antiferromagnetic
transition) is reported as 15 °K for
MgCr204, whereas the Curie temperature
GEOPHYSICAL LABORATORY
569
(paramagnetic to ferromagnetic transi-
tion) is reported as 593 °K for MgFe204
(Schieber, 1967). An assumed linear plot
of these transition temperatures versus
spinel composition suggests a room-
temperature (296 °K) Curie point for
the composition containing 48 wt %
MgFe204. Simple magnetic examination
of the spinels along this join was per-
formed with a Bruker magnet (model
B-M6) with an electrically produced
field of 130 Oe. For the MgCr204-
MgFe204 solid solution series, these tests
at room temperature showed that com-
positions from 50 to 100% MgFe204
were magnetic, whereas compositions
from 0 to 40% MgFe204 were non-
magnetic.
High Curie temperatures are generally
attributable to strong tetrahedral-octa-
hedral (Fe3+-Fe3+) interactions. The 57Fe
Mossbauer spectra indicate that no
tetrahedral-octahedral interactions exist
in these spinels over the composition
range 50-60% MgFe204 but begin only
at compositions greater than 60%
MgFe204. The magnetic behavior of the
50% and 60% MgFe204 compositions
may be either the result of magnetic
ordering of the Fe3+ in the octahedral
sites or the result of oblique-spin align-
ment of Fe3+ on octahedral sites. Mag-
netic resonance spectra will be required
to decide which of the two models best
describes the crystal chemistry of the
spinels in the composition range 50-60%
MgFe204.
The ferrites (magnesioferrite and mag-
netite) as well as the titanates (geikielite
and ulvospinel) of magnesium and iron
are all the inverse type of spinel, whereas
the aluminates (spinel and hercynite)
and the chromites (picrochromite and
chromite) of magnesium and iron are all
the normal type of spinel. A further
understanding of the crystal chemistry
for solid solutions between normal and
inverse spinel types is important, there-
fore, to both terrestrial and lunar
petrologic studies.
Lattice Deformations in Feldspars
Y . Ohashi and L. W. Finger
The geometry of a crystal lattice
changes (a) when the temperature is
changed, (b) when the composition is
varied in a solid solution series, or (c) in
some minerals when the structural state
is changed. The data for the changes in
the unit cell are usually interpreted by
computing the fractional changes in the
axial lengths. This approach is satisfac-
tory when the lattice symmetry con-
strains the expansion. In the case of
monoclinic and triclinic crystals, how-
ever, these calculations may lead to an
erroneous view of the expansion because
the directions of the extrema of the
changes deviate significantly from the
axial directions. This difficulty does not
affect the calculation of the changes in
the unit cell volume. Ohashi and Burn-
ham (1973) have developed a technique
for interpreting these effects in terms
of strain tensors and the corresponding
ellipsoids. Their calculations are free of
Fig. 90. Stereographic projection of the prin-
cipal axes of the strain ellipsoids associated
with chemical substitution in the microcline-
low-albite series. Points in the upper and
lower hemispheres are represented by solid and
open circles, respectively, with L, I, and S in-
dicating the largest, intermediate, and smallest
axes.
570
CARNEGIE INSTITUTION
the restrictions caused by crystal sym-
metry and therefore may be used on
triclinic crystals. Accordingly, the tech-
nique has been applied to feldspars with
an attempt to separate the three possible
sources of lattice deformation.
A FORTRAN program, STRAIN
(available on request), was used for the
strain-tensor calculation described above.
If standard errors or a variance-covari-
ance matrix of unit-cell parameters are
known, the program also computes esti-
mated standard errors for principal
strain components and their orientations
according to the equations of error
propagation (equations 3a and 3b,
Finger, Year Book 71, pp. 600-603). The
first derivatives required in the equations
are numerically evaluated.
In this study, the effects of chemical
substitution on lattice deformation were
determined by calculating the strain
ellipsoids for the microcline (KAlSisOs)-
low albite (NaAlSi308) series. The
TABLE 23. Principal Strain Components: Microcline-Low Albite Series *
Compositional Range,
Mole % of Or
Principal Strain
Components per
l%of Ab-*Or,
X10~3
Orientation of Principal Axes
+a
Angle (degrees) with
+c
(1) 0.94 -+ 10.27
(2) 10.27 -> 19.71
(3) 19.71 -> 29.27
(4) 29.27 -> 38.96
(5) 38.96 -+ 48.76
(6) 48.76 -» 58.68
(7) 58.68 -> 68.71
(8) 68.71 -> 78.86
(9) 78.86 -+89.11
(10) 89.11 -> 99.48
0.74(ll)f
0.30(8)
-0.14(9)
1.16(11)
0.62(7)
-0.30(8)
0.78(5)
0.68(5)
-0.26(5)
1.37(10)
0.42(5)
-0.52(7)
0.68(9)
0.40(6)
-0.26(8)
0.54(6)
0.40(5)
-0.28(6)
0.68(3)
0.27(4)
■0.09(5)
0.50(3)
0.23(5)
-0.18(4)
0.44(3)
0.06(4)
■0.08(4)
0.49(5)
0.11(5)
0.00(5)
45(8)
134(9)
99(7)
67(6)
27(5)
102(4)
86(20)
6(15)
94(2)
59(2)
33(3)
99(3)
43(11)
131(11)
99(4)
39(17)
127(17)
100(3)
23(4)
111(4)
98(2)
17(2)
95(6)
107(2)
7(4)
94(5)
96(3)
9(6)
97(6)
96(4)
66(8)
58(8)
138(6)
52(4)
113(5)
133(2)
42(3)
93(16)
131(1)
58(2)
114(3)
138(1)
65(8)
55(7)
134(2)
69(12)
52(9)
134(2)
78(4)
46(4)
133(4)
96(5)
45(3)
134(3)
89(4)
39(8)
129(8)
87(5)
46(13)
135(13)
79(6)
47(7)
45(7)
65(3)
120(4)
41(3)
58(10)
116(12)
43(1)
68(2)
129(2)
47(1)
80(8)
47(3)
45(3)
82(12)
46(3)
45(2)
95(3)
46(4)
44(4)
100(4)
48(3)
44(3)
109(4)
54(8)
43(7)
107(6)
46(13)
48(11)
* Lattice parameters are after Orville (1967, Table 2A).
t Standard errors given in parentheses were calculated from variances of lattice parameters. Con-
tributions from covariances of lattice parameters and compositional errors were not included.
GEOPHYSICAL LABORATORY
571
TABLE 24. Principal Strain Components: Low- and High-Albite Thermal Expansion *
Principal Strain
Components
per 1°C
X10~5
Orientation of Principal Axe
s
Temperature Range, °C
+a
Angle (degrees) with
+b
+ c
Low Albite
(1) 26 -* 245
2.06(29)1
0.49(21)
-0.38(19)
32(6)
113(9)
111(5)
71(7)
27(9)
109(9)
90(4)
75(9)
15(9)
(2) 245 -* 488
2.28(27)
0.79(20)
-0.57(17)
53(6)
137(7)
108(5)
63(5)
53(6)
131(5)
72(4)
59(5)
37(5)
(3) 488 -* 759
2.45(23)
1.33(15)
-0.66(20)
52(7)
140(7)
100(3)
59(5)
59(6)
134(3)
75(5)
51(3)
42(3)
(4) 759 -> 1010
2.43(31)
1.51(22)
-0.45(27)
52(12)
140(13)
101(5)
63(8)
60(9)
137(4)
73(8)
50(5)
45(4)
Low Albite (after heating to 1010° C)
(5) 26^141 2.11(106)
0.47(59)
0.03(77)
41(22)
130(23)
83(40)
71(18)
82(60)
159(32)
80(20)
19(53)
73(62)
(6) 141 -> 364
2.37(38)
0.57(24)
-0.58(24)
42(7)
124(9)
111(6)
71(7)
43(9)
127(8)
79(5)
60(8)
32(8)
(7) 364 -» 628
2.18(26)
1.16(17)
-0.52(20)
63(8)
29(8)
99(4)
52(5)
113(6)
133(3)
69(5)
125(5)
42(4)
(8) 628 -* 883
2.23(27)
1.29(19)
-0.47(22)
40(10)
128(11)
102(4)
67(8)
51(8)
132(4)
83(7)
51(4)
40(4)
(9) 883 -> 1127
3.53(40)
1.69(33)
-1.83(33)
78(7)
18(5)
77(4)
53(3)
106(5)
42(2)
55(4)
115(5)
134(2)
High Albite
(10) 26 -> 154
2.11(67)
0.96(66)
-2.00(68)
98(23)
9(18)
95(9)
43(7)
88(21)
133(7)
53(10)
108(14)
43(7)
(11) 154 -^318
3.01(55)
0.49(31)
-0.99(48)
75(6)
15(6)
91(11)
49(4)
100(10)
43(5)
62(6)
127(9)
130(9)
(12) 318 -> 605
2.89(20)
1.51(16)
-1.21(25)
66(6)
24(6)
90(3)
53(3)
106(5)
42(3)
67(5)
132(4)
129(3)
(13) 605 -> 902
4.60(43)
1.23(24)
-2.49(38)
81(4)
11(4)
97(4)
48(2)
101(4)
136(2)
57(3)
117(3)
45(3)
* Lattice parameters are after Stewart and von Limbach (1967, Tables 2 and 5A).
t Standard errors given in parentheses were caclulated from variances of lattice parameters. Con-
tributions from co variances of lattice parameters were not included.
572
CARNEGIE INSTITUTION
ordering of Si-Al is considered to be high
and nearly constant for this series. Thus
the major effect in the strain ellipsoids
represents the effect of K-Na substitu-
tion. The unit-cell data are those re-
ported by Orville (1967). The results of
strain-ellipsoid calculations are given in
Table 23 and shown stereographically in
Fig. 90. For all results shown in stereo-
graphic projection, the c axis is selected
as the polar axis and the a* (reciprocal
a) axis is horizontal with an azimuth of
180°. Throughout the microcline-low
albite series the position of the shortest
axis of the strain ellipsoids is essentially
invariant, whereas the largest axis moves
approximately on a great circle as shown
in Fig. 90. With increasing potassium
content, the axis of largest expansion
moves from point 1 to point 3 and then
rotates in the opposite direction.
The ellipsoids for thermal expansion
(Table 24 and Fig. 91) were calculated
using the low- and high-albite data of
Stewart and von Limbach (1967). In all
cases, the smallest axes are tightly
grouped and show contraction with in-
creasing temperature. The largest axes
of expansion are closest to the a axis for
low temperature and rotate toward the
b-c plane with increasing temperature.
The high-albite data show the opposite
trend: a rotation generally away from
the b-c plane with increasing tempera-
ture. Although Stewart and von Limbach
(1967) apparently did not detect any
irreversible changes in the low-albite
samples, the direction of maximum
thermal expansion in the second series
of samples, which were quenched from
1010 °C and reheated, is rotated toward
the b-c plane to a greater extent than
that of the samples before quenching.
This rotation suggests a possible slight
degree of disorder during the first heating
experiment.
To show the effect of Si-Al ordering,
unit-cell parameters were compared for
feldspars with identical chemistry but
with a different structural state. The
lattice parameters of KAlSi303 (micro-
a Low albite
■ Low albite (after heating to 1010°C)
• High albite
Fig. 91. Stereographic projection of the prin-
cipal axes of the thermal expansion ellipsoids
for high and low albite. Conventions for orien-
tation and symbols as in Fig. 90. The numbers
correspond to the numerical data in Table 24.
° Microcline — sanidine
Q Low albite — high albite
Fig. 92. Stereographic projection of the prin-
cipal axes of the strain ellipsoids associated with
Si-Al disordering for microcline-sanidine (in-
dicated by circles) and low albite-high albite
(indicated by squares). Conventions for orien-
tation and symbols as in Fig. 90. For numerical
data, see Table 25.
GEOPHYSICAL LABORATORY
573
TABLE 25. Principal Strain Components: Si-Al Disordering *
Principal Strain
Components,
X 10~2
Orientatio
n of Principal Axes
Ordered Phase — >
Disordered Phase
+a
Angle
(degrees) with
+6 +c
Microcline — > sanidine
(KAlSi308)
Low albite — * high albite
(NaAlSi308)
2.54(2) t
-0.68(2)
-1.66(2)
2.73(3)
-0.58(2)
-1.64(3)
131.4(2)
102.3(5)
44.0(3)
130.1(3)
98.5(10)
41.3(3)
44.5(2) 79.1(2)
92.6(5) 13.9(5)
45.7(2) 98.4(7)
43.1(3) 81.7(4)
89.1(7) 18.6(11)
46.9(3) 106.5(11)
* Lattice parameters are after Waldbaum (1968, Table 7).
f Standard errors given in parentheses were calculated from variances of lattice parameters. Con-
tributions from covariances of lattice parameters were not included.
cline-sanidine) and NaAlSi308 (low
albite-high albite) reported by Wald-
baum (1968) were used for construction
Y
X
iA
Fig. 93. Alkali polyhedron in albite projected
along the smallest axis of the strain ellipsoid.
X, Y, and Z are the axes numbered 1 in Fig.
90, X being the largest axis and Z, the smallest.
The height of the oxygen atoms from an arbi-
trary plane is shown in angstroms. An enantio-
morphic polyhedron is also found in the same
unit cell. If the lengths of the edges shared
with the silicate tetrahedra (indicated by
double lines) are fixed, an expansion of the
alkali polyhedron must be accomplished by
expansion of the nonshared edges, resulting in
a slight contraction along the direction of pro-
jection and an expansion in the plane of the
diagram. The rotation of the largest expansion
axis with a change in chemical composition (see
Fig. 90) would be related to a rotation of the
shaded triangular face.
of the strain ellipsoids. In both cases the
orientation of the strain ellipsoids is
similar (Fig. 92 and Table 25) but shows
a marked contrast with the strain
ellipsoids associated with chemical sub-
stitution and increasing temperature. In
the strain ellipsoid due to disordering of
Si-Al, the intermediate axis is approxi-
mately parallel to the c axis. The largest
axis is in the second ( — a and -\-b)
quadrant, and the smallest axis is in the
first quadrant.
The directions of maximum expansion
may be interpreted by using Fig. 93, a
projection of the feldspar structure
adapted from Ribbe, Megaw, and Taylor
(1969). As the size of the atom in the
alkali position is increased (i.e., with
substitution of K for Na), it seems rea-
sonable to assume that the Si04 tetra-
hedral framework will be essentially
unaffected. In particular, the edges
shared between the tetrahedra and the
alkali polyhedron will not change in
length. A distortion of the alkali site
will therefore be required, with the rela-
tive distortions as shown. An analogous
effect would be expected for increasing
temperature since the tetrahedra will
undergo a much smaller expansion than
the alkali site (Brown et al.} 1972). On
the other hand, the structural changes
associated with Si-Al order-disorder af-
fect the framework, and a greatly differ-
ent mode of expansion would be expected.
574
CARNEGIE INSTITUTION
Measurements of the Polarized
Crystal-Field Spectra of Ferrous
and Ferric Iron in Seven
Terrestrial Plagioclases
P. M. Bell and H. K. Mao
The search for ferric iron in lunar
samples led to the first measurements of
polarized crystal-field spectra in plagio-
clase from Apollo 14 (Bell and Mao,
Year Book 71). Ferrous and ferric iron
should cause very different absorption,
and the bands should be detectable even
at the low concentrations of iron found
in plagioclase (0.05-1.0 wt % Fe). The
technique should therefore provide data
that can be used as a sensitive indicator
of the oxidation state, an important
parameter in the chemical processes of
crystallization on the moon.
Oxidation potentials are equally im-
portant parameters in the crystalliza-
tion history of terrestrial rocks. The par-
tial pressures of oxygen vary sufficiently
to stabilize significant proportions of
Fe3+:Fe2+. Hence, terrestrial plagioclases
have a good range in the ratio of ferric
to ferrous iron which could be used to
calibrate the crystal-field technique.
An excellent suite of seven plagioclases
ranging from albite to anorthite in com-
position and from 0.01 to 0.55 wt % iron
content was provided by Dr. Paul Ribbe,
Virginia Polytechnic Institute. Part of
the suite formed the basis of a crystallo-
graphic and analytical study (Ribbe and
Smith, 1966) and was well characterized.
The crystals are colorless, clear, and free
from alteration. Table 26 gives the
chemical analyses determined in this
study by electron microprobe and the
geographic locations.
The crystals were sufficiently large
(average diameter 2 mm) so that meas-
urements of the crystal-field spectra
could be made directly with the Cary
17-1 instrument. Crystal surfaces were
ground and polished parallel for each
orientation. The suite varied from all
ferric to all ferrous iron. Figure 94 shows
an example of absorption spectra for the
a and f3 polarizations of a plagioclase
from the suite with a ratio of ferric to
ferrous iron of approximately 10. The ab-
sorption bands are extremely sharp and
intense, even though the total iron con-
tent is only 0.32 wt %. Some absorption
bands are offscale in the figure and have
been replotted with the origin displaced
by 5 cm-1. A plot with a compressed scale
is shown in the inset.
TABLE 26. Average Chemical Composition of Plagioclase Single Crystals
Determined by Electron Microprobe
Anorthite,
Bytow-
Labra-
Labra-
Labra-
Miyakei
nite,
dorite,
dorite,
dorite,
Andesine,
Island,
Adel,
Sonora,
Lake Co.,
Hart Mt.,
Minsen,
Albite,
Japan
Ore.
Mexico
Ore.
Ore.
Korea
Crete
MnO
0.02
0.02
0.03
0.00
0.02
0.02
0.00
FeO*
0.55
0.43
0.38
0.41
0.41
0.29
0.01
NiO
0.04
0.01
0.01
0.01
0.00
0.00
0.01
Na20
0.47
3.87
3.91
3.75
3.82
6.27
11.93
MgO
0.07
0.16
0.11
0.14
0.15
0.03
0.00
AI2O3
35.3
29.9
30.1
30.1
30.5
26.5
19.5
Si02
43.8
50.9
51.0
50.6
51.6
56.8
68.6
K20
0.00
0.11
0.23
0.12
0.13
0.87
0.02
CaO
19.6
14.0
13.4
13.6
13.4
8.9
0.0
Ti02
0.00
0.05
0.07
0.04
0.03
0.06
0.00
Cr203
0.01
0.00
0.00
0.00
0.00
0.00
0.00
Totals
99.9
99.5
99.3
98.9
100.1
99.9
100.27
* Total Fe as FeO.
GEOPHYSICAL LABORATORY
575
o
o
c
o 2
o
CO
<
30 25
1 11/ 1 (li I | i I i i — | — i — i — I — i — | — r
300
Wave number , I03cm_l
20 15
10
I03cm-'
30 20 - 10
Plagioclase,Lake County.Ore.
An 67, Ab33
Total Fe 0.32 weight percent
400
500
Wavelength , n m
1000
2000 3000
Fig. 94. Polarized crystal-field spectra of iron (both Fe2+ and Fe3+) in plagioclase from Lake
County, Oregon. Inset shows same spectra in reduced scale.
Ribbe and Smith (1966) deduced that
iron in plagioclase occupies two struc-
tural sites. One contains Fe3+ and is
tetrahedral (the aluminum site) ; the
other contains Fe2+ and is octahedral
(the calcium site) . In lunar plagioclase,
only the polarized bands centered at
approximately 8000 and 4500 cm-1 wave
numbers are observed. These bands are
caused by transitions in ferrous iron
(Bell and Mao, Year Book 71). As
indicated by Mao and Bell (this Re-
port), the assignment of these bands
cannot be derived by simple calculation.
It is tempting to use the method of Burns
(1970) and draw analogy with the ab-
sorption bands in pyroxene, where Fe2+
is in octahedral coordination. However,
the site symmetries, bond lengths, and
energies of absorption are not close
enough for correspondence. In both
pyroxene and plagioclase structures, the
octahedral sites are distorted, but this is
where the similarity ends.
Professor W. B. White (personal com-
munication) has suggested that the band
at 8000 cm-1 may be caused by tetra-
hedral coordination of Fe2+ and that the
band at 4500 cm-1 may be caused by a
small amount of Fe2+ in the calcium site.
A method of comparing the crystal-
field splitting of a single quintet ground-
state transition of Fe2+ in MgO was used,
which was based on successful applica-
tions to ferrous-iron-bearing garnets
(Moore and White, 1972 ) . Models of the
plagioclase sites (ignoring the effects of
site and Jahn-Teller distortions) give a
moderately close fit to the observed
values. From the calculations it can be
concluded that these bands are caused
only by Fe2+, and the assignment of most
of the Fe2+ to tetrahedral (aluminum)
sites appears to be the best approxima-
576
CARNEGIE INSTITUTION
tion at this time. It may be possible to
calculate the distortion and its relation-
ship to the observed molar extinction
coefficient.
Higher energy bands in the spectra
shown in Fig. 94 are not detected in lunar
plagioclase. The weak bands at approxi-
mately 24,000 and 26,000 wave numbers
are consistent with spin-forbidden transi-
tions in tetrahedrally coordinated Fe3+.
The strong band at 32,000 wave numbers
is probably a charge-transfer band in
Fe3+, similar to the one documented in
pyroxene (Bell and Mao, Year Book 71,
p. 533). The sharpness and intensity of
the band make it usable for detecting
Fe3+.
The present measurements are the first
available on Fe3+ and Fe2+ in plagioclase,
and they are sufficiently accurate to be
used in distinguishing the oxidation
states of iron. The band assignments are
preliminary but nevertheless imply a
distinct partitioning of ferrous and
ferric iron, which may be related to the
thermal history of crystallization.
Electric-Field-Gradient Calculations
in Rare- Earth Iron Garnets
Michael Raymond
The electric-field gradients (EFG's)
at the iron sites in yttrium and lutetium
iron garnets have been calculated to
ascertain the nuclear quadrupole inter-
actions in these crystals and to test an
improvement in the point-multipole lat-
tice summation model of Raymond
(1971) for calculating nuclear quadru-
pole interactions in solids. The improve-
ment is the incorporation into the model
of overlap effects, which are the contri-
bution to the EFG caused by the overlap
distortion of the metal orbitals by the
surrounding ligand orbitals. This contri-
bution is neglected in the ionic point-
multipole model, which specifically as-
sumes nonoverlapping ions.
Rare-earth iron garnets were chosen
for study because of the availability of
crystal-structure data, Mossbauer meas-
urements of the quadrupole coupling
constants, and previous calculations of
the overlap contribution in these crystals.
The last-mentioned consideration is im-
portant because it was desired to test the
complicated overlap calculations for a
relatively simple crystal before attempt-
ing those for the aluminum silicates, with
their much greater number of EFG
components and nonequivalent sites.
The structural parameters used in
this calculation were taken from a re-
finement of the rare-earth aluminum,
iron, and gallium garnet structure by
Euler and Bruce (1965). In iron garnets,
#3Fe3+2(Fe3+04)3, where i2 is a rare-
earth element, there are two nonequiva-
lent Fe3+ sites, one tetrahedral with point
symmetry 4 and one octahedral with
point symmetry 3. This symmetry
reduces the number of nonzero compo-
nents of the EFG tensor from five to one
at both sites, so there are only two EFG
components per crystal to measure and
calculate.
Mossbauer determinations at room
temperature of the quadrupole coupling
constants have recently been made by
Housley and Grant (1972) and Belo-
zerskii et al. (1970) for the octahedral
and tetrahedral sites, respectively, in
yttrium iron garnet (YIG). The quad-
rupole coupling constants in lutetium
iron garnet (LuIG) were measured by
Nicholson and Burns (1964) at 610°K.
The rare-earth iron garnets are ferri-
magnetic below about 570 °K, making
room-temperature measurements more
difficult; however, room-temperature
data are more useful for making com-
parisons with EFG calculations because
the EFG's were calculated using the
crystal structure determined at room
temperature. The experimental results
are displayed in Table 27.
The EFG's in YIG and LuIG were
first calculated using the ionic point-
multipole model described by Raymond
(1971). Point-charge and dipole contri-
butions were calculated using a lattice
summation method discussed by Ray-
GEOPHYSICAL LABORATORY
577
TABLE 27. Calculated and Experimental Field-Gradient Components
at the Fe Sites in YIG and LuIG (units of 1014 esu)
YIG
LuIG
Tetrahedral
Octahedral
Tetrahedral
Octahedral
-0.622
-1.297
-0.752
-1.611
-0.475
+0.971
-0.437
+ 1.073
-1.097
-0.325
-1.189
-0.538
-11.123
-3.300
-12.059
-5.458
-15.148
-11.701
-16.037
-13.943
-10.301
-7.957
-10.905
-9.481
-21.424
-11.257
-22.964
-14.939
- 18.465 §
-8.234H
-16.3
-9.27
-16.9
-10.1
-14.2315**
-19.2001**
-16.00tt
-23.55tt
Point charge
Dipole*
Total lattice sum
Lattice sum X (1 — Toot)
Overlap
Overlap X (l-Rt)
Lattice sum + overlap (this calc.)
Experiment (300°K)
Experiment (610°K)1f
Lattice sum + overlap (previous calc.)
* aD (O2-) = 1.0 A3.
t To, (Fe3+) = -9.14; Sternheimer, 1963.
tR (Fe3+) = +0.32; Ingalls, 1964.
§ Belozerskii et al., 1970.
|| Housley and Grant, 1972.
T[ Nicholson and Burns, 1964.
** Sharma and Teng, 1971.
tt Sharma, 1972.
mond (1971). The dipole polarizability
of oxygen was treated as a variable
parameter, but only the results for a
value of 1 A3 are given here. An approxi-
mate average value for the oxygen
polarizability (Raymond, Year Book
71, pp. 500-504) was used because an
accurate value is not available at this
time.
Overlap contributions to EFG's have
been previously calculated by D. R.
Taylor (1968), Sawatzky and Hupkes
(1970), and Sharma (1970), and espe-
cially for rare-earth iron garnets by
Sharma and Teng (1971) and Sharma
(1972). Sawatzky and Hupkes (1970)
used a simplified formulation that
neglects EFG contributions from the
ligand valence orbitals, whereas Sharma
(1970, 1972) used a more complete
formulation. Sharma, however, did not
include induced dipoles in his lattice sum
calculations, and perhaps for this reason,
his calculations do not agree well with
experiment.
For the present calculation, the sim-
pler Sawatzky formulation was used,
where the overlap contribution to the
EFG is
\* zz i overlap —
^e<r-3>2>V (3 cos2 ®K
5 K
1). (1)
It is necessary to calculate the contribu-
tions from the overlaps of the iron 2p,
3p, and 3d orbitals with the oxygen 2s
and 2j> orbitals. The contribution to the
EFG due to overlap of the iron 3p orbi-
tals with the oxygen 2s and 2p orbitals
may be written
(Vzz') (3p) =
^e<r3>3p:>(3cos20A'-l) X
5 K (2)
[S2sP,2s(K) + ShPt2Pa(K) -
S2
3p,2p.
.(#)]
Here, the *S's are overlap integrals and
<r"3> is the expectation value of r~3 over
the iron-ion radial functions. The over-
lap integrals were calculated according
to the method of Corbato and Switendick
(1963), using dementi's (1965) wave
functions for Fe3+ and Watson's (1958)
O2" wave functions in a 2+ stabilizing
potential well. The total EFG is, then,
578
CARNEGIE INSTITUTION
(Vgz) = (1 — R) (Vzz) overlap +
(1 — 7oo) \Vzz) lattice sum- (3)
The shielding factors R and y^, were
determined by Ingalls (1964) and Stern-
heimer (1963), respectively; the values
are given in Table 27.
Results of the calculations are also
shown in Table 27. The lattice sum term
alone does not agree well with experi-
ment ; the agreement is much poorer than
for most silicates discussed by Raymond
{Year Book 71, pp. 500-504). Using a
different value for the dipole polariza-
bility of oxygen does not decrease the
discrepancy. The dipole contribution,
neglected by Sharma (1970, 1972), is
large and may account for the lack of
agreement obtained by him. The present
calculations agree with Sharma and Teng
(1971) and Sharma (1972) in the values
of the point-charge summation and the
local overlap contribution. The dis-
agreement of the present calculations
with experiment is about 23% and can
be reduced to half as much by choosing
different values for the polarizability of
oxygen and the Fe3+ antishielding factor.
These calculations will therefore be per-
formed for the aluminum silicate crys-
tals, where the large number of EFG
components makes it possible to mini-
mize the effects of uncertainties in these
parameters.
Because the overlap contribution to
the EFG is large, its contribution to
Madelung constants and potentials at
individual sites may also be large. For
this reason the overlap effect would affect
cohesive energy and site-preference en-
ergy calculations.
Solid Solution in Aenigmatite
F. N. Hodges and D. S. Barker*
Aenigmatite, once thought a mineral-
ogical rarity, has been found since the
advent of the electron microprobe to be
* Department of Geological Sciences, Univer-
sity of Texas at Austin.
a widespread constituent of peralkaline
rocks. Investigation of the northern
(Diablo Plateau) portion of the Tertiary
Trans Pecos igneous province, including
a detailed electron microprobe study
(Barker and Hodges, in preparation) ,
has revealed that aenigmatite, a common
accessory mineral, contains considerably
less titanium than the ideal formula
Na2Fe2+5TiSi602o (Kelsey and McKie,
1964; Cannillo et al, 1971). The Ti02
content of aenigmatite from the Sierra
Prieta nepheline-analcime syenite ranges
from 6.8 to 1.7 wt %, considerably lower
than values reported from other alkalic
provinces (Na2Fe2yriSi602o contains
9.27 wt % Ti02). Analysis of aenigma-
tites from other intrusions in the area
yields Ti02 contents within the same
range.
On the assumption that the aenigma-
tite is stoichiometric (cations = 14;
oxygen = 20) , ferric iron was calculated
from the microprobe analyses by con-
straining the total number of cations and
total number of oxygens (Finger, Year
Book 71, pp. 600-603). Plotting of Fe3+
values obtained in this manner against
Ti (Fig. 95) gives a clear indication
that these aenigmatites are a solid
solution between "ideal aenigmatite"
(Na2Fe2+5TiSi602o) and the titanium-
free aenigmatite (Na2Fe2+4Fe3+2Si602o)
of Ernst (1962). The principal substitu-
tion involved is Fe2+Ti4+^±2Fe3+. Other
substitutions of this type, involving Si
and Al in addition to Fe and Ti, almost
certainly occur in natural aenigmatites.
Although a large part of the scatter in
data points (Fig. 95) is due to statistical
uncertainties inherent in electron micro-
probe analysis and possibly also to un-
certainties in the method used to esti-
mate Fe3+, the general tendency of aenig-
matites, those reported both here and
elsewhere in the literature, to contain
Fe3+ in excess of that required to replace
Ti is of considerable interest. A few
analyses from the literature indicate the
possibility of Fe3+ in tetrahedral coordina-
GEOPHYSICAL LABORATORY
579
o
CVJ
X
a>
Ll_
O |
O -
0.4 0.6
[Ti] on
[o]
Fig. 95. Plot of calculated Fe3+ versus Ti. Circles, analyses from Sierra Prieta nepheline-anal-
cime syenite. Squares, analyses from other Diablo Plateau intrusives. Diamonds, analyses listed
in Kelsey and McKie (1964), Abbott (1967), and Zies (1966).
tion, a situation probably more likely in
undersaturated peralkaline rock, defi-
cient in both silicon and aluminum. The
possibility of Fe3+ in tetrahedral coordi-
nation is also indicated by detailed
structural analysis of natural aenigma-
tite (Cannillo et al, 1971). Most of the
excess Fe3+, however, is due to the substi-
tution Fe2+Si^±Fe3+Al, Fe3+ in octahedral
coordination balancing the charge of Al
in tetrahedral coordination. Figure 96 is
a plot of the same aenigmatite analyses,
but Fe3+ has been replaced as a coordi-
nate by Fe3+ - A1(IV) + A1(VI) to re-
move the effect of the substitution
Fe2+Si^±Fe3+Al. In this plot the con-
cordance between the data points and
the line connecting Fe3+ and Ti end mem-
bers is greatly improved. A better fit to
the line may be achieved by subtracting
Fe3+ necessary to fill tetrahedral sites
from Fe3+ - A1(IV) + A1(VI) (crosses in
Fig. 96 indicate analyses after subtrac-
tion of tetrahedral Fe3+ from ordinate).
The reason for the substitution of
Fe3+ for Ti in these aenigmatites is not
580
CARNEGIE INSTITUTION
Z°z
o
CO
X
>_
<
+
>
<
I
fO
x x
0
0.2
0.4
[Ti]
[0]
0.6
0.8
X20
Fig. 96. Plot of Fe3+ — A1^IV> + A1(VI> versus Ti. Symbols as in Fig. 95. Crosses at ends of
arrows represent analyses with Fe3+ necessary to fill tetrahedral site subtracted from the ordinate.
clear. Oxygen fugacity might conceiv-
ably be an important factor in this sub-
stitution; however, published experi-
mental results shed little light on the
matter. Ernst (1962) synthesized ti-
tanium-free aenigmatite at the oxygen
fugacity of the wustite-iron (WI) and
magnetite- wiistite (MW) buffers but re-
ported that aenigmatite is unstable at
the oxygen fugacity of the quartz-faya-
lite-magnetite (QFM) buffer. On the
basis of this experimental work, various
authors (Abbott, 1967; Kelsey and
McKie, 1964) have put forward the idea
that the presence of aenigmatite in a
rock indicates formation under condi-
tions of relatively low oxygen fugacity.
Lindsley {Year Book 69, pp. 188-190),
however, synthesized aenigmatite of
"ideal" composition at the oxygen fuga-
city of the QFM buffer and reported evi-
dence that it is unstable at the oxygen
fugacity of the nickel-nickel oxide
(NNO) buffer. Thus, the Fe3+-rich end
member apparently has an upper sta-
bility limit at lower oxygen fugacity
GEOPHYSICAL LABORATORY 5gl
than the Ti end member, which should rocks highly depleted in titanium
contain little or no Fe3+. Obviously (Ti02 = 0.04 to 0.11 wt %). In addi-
nothing meaningful can be said at the tion, as noted by Abbott (1967), there is
present time concerning the relationship a tendency for groundmass aenigmatite
between aenigmatite formation and to contain less Ti than aenigmatite
oxygen fugacity. phenocrysts. It appears that aenigma-
The activity of titanium in the magma tite forms in response to the peralkaline
may have more influence on aenigmatite condition, its composition depending
solid solution than oxygen fugacity. The largely upon the availability of titanium
aenigmatites described here occur in at the time of formation.
ISOTOPE GEOCHEMISTRY
Uranium and Thorium Partitioning weight from which diopside crystals were
in Diopside-Melt and Whitlockite- grown in liquid. The second consisted of
Melt Systems the same mix plus 25.5% Ca3(P04)2 and
M r „ . yielded large crystals of the tricalcium-
phosphate, whitlockite. Diopside, or
The partitioning of uranium and more generally clinopyroxene, is one m of
thorium in silicate and phosphate sys- the main repositories of uranium and
terns was studied at pressures to 25 kbar. thorium in common mafic rocks, and
The uranium studies differ from a whitlockite is an important host to
method previously described (Seitz and uranium and thorium in meteorites.
Shimizu, Year Book 71, p. 548) in that
pure uranium-235 was used instead of Experimental Method
uranium with a natural isotopic ratio. ^ n -, , 0 ,■> . , ,x
mi- i xt i fi • rrom 0.1 to 3 ppm (by weight) ura-
1ms change greatly enhances the sensi- • ooc in , «n ,, non
,. ., r x i • x i i -j inum-235 or 10 to 30 ppm thonum-230
tivity oi the experiments and makes it
possible to measure the distribution of
was added to the starting mixtures. The
charges used in atmospheric experiments
uranium at concentrations to 0.1 ppb in •■ . , ,. £ -, ,
. ,. . , , . -^ .... r^ were wrapped in platinum toil and run
individual grains. Partitioning expen- . ,. , . , , -^
&, .. i. in a conventional resistance turnace. ror
ments were done over the range of ura- , ■ u ■,-,• ,x , ,
„ . . nigh-pressure conditions they were sealed
mum concentrations found in common , ,• i ± ±\ -.i u .
, x . .... . . . in platinum capsules together with about
igneous rocks. In addition, a technique oc/ , ■>■, ■, , , ,, ,-,
, . . ,, . ' , o% water, added to promote the growth
lor microscopically mapping concentra- ^ i /-mn \ j. i rni i
» ,, : . , r & .,,. or larger (100 um) crystals. 1 he capsules
tions oi thorium in parts per million was rj •■• , • ,
, , , I,, , . . i were run m a solid-media, high-pressure
developed and has been used to study , e m j. oe uu t«i
,, . r .... . . , / apparatus from 10 to 25 kbar. Ihe
tnormm partitioning between crystals , , G . • , x ,,
, .\. lf & J temperature was first raised above the
and coexisting melt over a pressure range melti interyal of the ch and then
from 1 atm to 25 kbar. The ability to lowered ^ the yalues repofted in Tableg
study uranium and thorium, demon- 2g and 2g to promote crystallization,
strated here, and the implicit ability to Run times ranged from x to 200 hours
study plutonium partitioning are impor- after which time the charge was rapi(lly
tant in light of the central role these cooled to preServe the high-temperature
elements have played in determining the phase assemblage as quenched silicate
chronology of elements in the solar sys- mejt in the form of glass surrounding
tem. the crystals.
The first of two systems studied was After the runs, the samples were
a simple basalt system of 50% diopside, polished and covered with an appropri-
25% albite, and 25% anorthite by ate particle-track detector. Fission
582
CARNEGIE INSTITUTION
TABLE 28. Uranium Partitioning Between Diopside and Coexisting Melt
Uranium
Run U in
Uin
Equilibrium
Time, Melt,
Diopside,
Partition
P, kbar
T, °C
min ppm
ppb
Coefficient*
Diopside-Meltf
10
1150
50 4.3
6.6
0.0017
10
1150
120 3.5
5.0
0.0016
20
1150
360 4.4
5.3
0.0014
25
1150
120 4.4
Diopside-MeltJ
7.0
0.0020
10
1150
120 0.91
2.0
0.0025
20
1150
120 0.96
0.96
0.0012
25
1150
120 1.0
Diopside-Melt§
1.0
0.0012
20
1100
60 0.148
0.13
0.0014(4)
1100
120 0.148
0.11
0.0010(2)
1150
60 0.135
0.09
0.0008(3)
1150
120 0.139
0.30
0.0025(10)
1200
120 0.126
0.09
0.0008(2)
* The partition coefficient at equilibrium was calculated from equation 1. Standard deviations of
the final digits from track-counting statistics are given in parentheses. Standard deviations less than
8% are not given.
t Diopside 50 wt %, albite 25%, anorthite 25%; an average of 8.6 wt % water was added.
t 2100 ppm (by weight) potassium, 5 ppm rubidium, 1 ppm cesium, 100 ppm strontium, 240 ppm
barium, 0.83 ppm thorium, 2.5 ppm lead; 2.1% iron and 8.8% water were added to the system.
§ An average of 8.2 wt % water was added.
tracks from the neutron-induced fission
of uranium-235 were recorded in mus-
covite, and alpha particles from the
spontaneous decay of thorium-230 were
recorded in cellulose nitrate plastic.
If the trace element was strongly frac-
tionated into the melt, two track maps
were used — one of low density, to record
the element distribution in the glass,
and the other of higher density, to reveal
the distribution in the crystals. Uranium
concentrations were obtained by multi-
plying the ratio of the track density from
the unknown over that in a standard
TABLE 29. Thorium Partitioning Between Diopside and Melt and Whitlockite and Melt
P, kbar
T, °C
Run
Time,
hours
Thin
Melt,
ppm
Thin
Crystals,
ppm
Thorium
Equilibrium
Partition
Coefficient*
0
25
1240
1150
1240
1240
Diopside-Meltf
240 17
4 30
Diopside-MeltJ
120 30.1
Whitlockite-Melt§
24 34.1
0.10
0.15
<0.5
72.7
0.007(3)
0.006(1)
<0.017
2.13(9)
* The partition coefficient at equilibrium was calculated from equation 1 for diopside-melt systems.
Standard deviations of the final digits from counting statistics are given in parentheses.
t Diopside 50 wt %, albite 25%, anorthite 25%; 15 wt % water was added to the 25 kbar run.
t Diopside 51 wt %, whitlockite 17.5%, albite 15.75%, anorthite 15.75%.
§ Starting mixture: diopside 30 wt %, whitlockite 25%, albite 22.5%, anorthite 22.5%.
GEOPHYSICAL LABORATORY 583
glass by the uranium concentration in by 50% in MgO in borders about 10 /xin
the standard. Thorium concentrations wide. This effect is probably due to
were calculated from the measured track rapid growth of the crystals.
densities, track lengths, exposure time, Despite this effect, however, there were
and half-life of the thorium-230 isotope, no detectable variations in the uranium
Background corrections for uranium concentrations of the glass or crystals,
and thorium were less than 2%, even at The uniform concentration of uranium
the lowest concentrations measured, in glass showing major-element varia-
Partition coefficients were calculated tions may be interpreted in two ways,
from the ratio of the trace-element con- First, during the rapid crystal growth
centrations in the crystal and glass so that is proposed to account for the de-
that any systematic bias in the measure- pleted zone, the uranium may not have
ment of the trace-element concentrations been excluded from the diopside and thus
was largely eliminated. The statistical would not have accumulated in the zone
uncertainty from the number of tracks around the crystal. Attempts to see a
counted decreased with increasing con- uranium-laden diopside rim were made
centration, giving a standard deviation by superimposing photographs of the
of the partition coefficient of less than crystals and the uranium map. Although
3% in the system containing 3 ppm no such rim was observed, the existence
uranium. Standard deviations greater of a rim less than 4 /xm wide could not
than 8% from this source of error are be excluded. Alternatively, the uranium
given in Tables 28 and 29. Glass cores may have accumulated in the glass near
in the crystals were easily identified by the growing crystal but diffused away
the corresponding clusters of tracks they more rapidly than the major elements,
produced in the track map so that errors If this were the case, since uranium does
from cross-contamination of tracks from not mimic the major-element variations,
crystals or glass are believed to be small, its partitioning between diopside and
melt would appear not to be sensitive to
Uranium the actual melt composition.
The degree of chemical equilibration
Experiments were done to study ura- of uranium between the rims and the
nium partitioning between diopside and centers of the crystals depends in part
melt at various pressures to 25 kbar, on the diffusion rates of uranium in the
trace-element and minor-element concen- crystals. The measured diffusion rate of
trations, temperatures, and concentra- uranium in diopside (Seitz, this Report!
tions of uranium. The major-element ig ciearly not high enough to affect
chemistry of the phases was determined equilibrium in crystals greater than 10
by electron microprobe. The Na,0 and ^m in size In the experiments con-
A1203 concentrations in the diopside in- gidered, it can be shown that the parti-
creased with pressure from 0.13% and tion coefficient K', measured as the ratio
1.40% by weight, respectively, at 1 atm of the average uranium concentrations
pressure, to 0.25% and 1.73% at 10 kbar, of the crygtal and mdt> ig related tQ the
and to 0.73% and 5.52% at 20 kbar partition coefficientf K that would be
None of the diopside crystals showed obtained at equilibrium by the equation
variations in chemistry trom core to rim,
even those as large as 200 ^m. This was K'S
not true, in general, for the quenched -(i — «S) In ( 1 — S)
liquid, some of which was depleted in (i(
magnesium and calcium around the
crystals. In the most extreme case, the where S is the fraction of the material
melt was depleted by 30% in CaO and that has crystallized. This equation was
584
CARNEGIE INSTITUTION
derived assuming no trace-element dif-
fusion in the solid and rapid diffusion in
the melt, and assuming that the trace-
element concentration of the crystals is
small compared with that in the melt.
Values of S were calculated from the
final crystal and melt compositions and
are in the range from 0.20 to 0.36 in these
experiments. Equilibrium partition co-
efficients were computed on the basis of
equation 1 and are 12% to 24% higher
than the observed ratios of the concentra-
tion in the crystal over that in the glass.
The experimental conditions, uranium
concentration, and equilibrium partition
coefficients for runs showing large diop-
side crystals are given in Table 28.
The partition coefficients measured in
the first set of runs reported in Table 28
are seen to be similar for pressures from
10 to 25 kbar. Thus it appears that there
is no strong effect of pressure on uranium
partitioning. If so, the differentiation of
uranium by partial melting will be un-
affected by pressure to depths of 80 km.
In the second set of experiments iron
and seven trace elements were added to
the simple basalt mixture, the trace ele-
ments in concentrations common to na-
tural basalts. The concentrations of
uranium, potassium, and iron were low
before their addition (less than 10 ppb,
less than 0.02%, and less than 0.04%, re-
spectively) so that the addition of these
elements represents significant increases
in their concentrations. From Table 28
it can be seen that there is no systematic
change in the partitioning of uranium in
the presence of the added elements. This
observation suggests that the incorpora-
tion of uranium into diopside does not
require charge compensation by the trace
elements considered here.
For the diopside-plagioclase mixture
used in these experiments only a limited
temperature range could be explored be-
cause of the small temperature interval
over which the diopside and melt phases
coexist. Moreover, any temperature ef-
fect could be masked by the effects of
changes in the chemical composition of
the silicate melt that accompany the
temperature changes. The uranium con-
centration in the glass is about 0.1 ppm.
The low uranium concentration in the
crystals produced few tracks and resulted
in a large standard deviation of the
measured partition coefficients. The re-
sults of partitioning studies at 1100°,
1150°, and 1200°C are given in Table
28, from which it can be seen that there
is no systematic variation of the parti-
tion coefficient with temperature. This
behavior holds in spite of the fact that
the concentration of elements such as
magnesium in the melt changed by more
than 50% in this temperature interval.
It seems unlikely that the effects of tem-
perature and chemical variations in the
melt would exactly cancel one another.
The insensitivity of uranium partition-
ing to the composition of the liquid in-
dicates that uranium concentrations
should not be expected to follow major-
element inhomogeneities in the glass.
It is important to note that even
though the uranium concentration was
varied by more than a factor of 30, the
partition coefficient remained essentially
unchanged. This result suggests that the
activity coefficients of uranium in the
crystal and melt are independent of con-
centration and implies that uranium acts
as a dilute solute at concentrations be-
low 4 ppm.
Thorium
Thorium partitioning was studied in
systems containing diopside crystals plus
silicate melt, diopside crystals plus sili-
cate-phosphate melt, and whitlockite
crystals plus silicate-phosphate melt. In
all the runs, no inhomogeneities of
thorium in the crystals or glass were
detected. From the slow rate of diffusion
of thorium in crystals (Seitz, this Re-
port), however, only local equilibrium
was obtained. Therefore the equilibrium
GEOPHYSICAL LABORATORY
585
partition coefficients were determined
from the ratio of the concentrations in
the crystal and glass on the basis of
equation 1. The experimental run con-
ditions, thorium concentrations, and
equilibrium partition coefficients are
given in Table 29.
The thorium-partitioning coefficient
between diopside and melt is consider-
ably higher than that for uranium. The
difference is believed to reflect a differ-
ence in valence, thorium being always in
the +4 valence state, whereas uranium
is likely to be in the +6 valence state.
Like uranium partitioning, however, there
is no detectable difference in thorium
partitioning between atmospheric pres-
sure and 25 kbar. These measurements
establish that the fractionation of tho-
rium in diopside-melt systems will be
essentially constant to depths of 80 km in
the earth.
In experiments on the system com-
prised of diopside and silicate-phosphate
melt, the counting period (5 days) was
too short to determine the thorium con-
centration in the diopside, but an upper
limit was established for the partition
coefficient that is compatible with the
value measured in the system with the
pure silicate melt.
Crystals were grown from a melt con-
taining 25% Ca;H(P04)2 by weight and
were identified by x-ray diffraction as
the fi form of the tricalcium phosphate,
whitlockite. From electron microprobe
analyses the crystals were found to con-
tain 3.23% MgO, 0.39% Si02, and less
than 0.05% Na20. The melt that formed
in equilibrium with these crystals at
1240 °C contained 10.8% P305. The
crystals incorporated concentrations of
thorium 2.13 times higher than were in
the melt. The equilibrium partition co-
efficient is accurately given by the ratio
of the concentrations in the crystal and
melt because the extent of crystallization
was small (less than 5%) and because
thorium was only slightly fractionated in
this system.
Discussion
Several properties of uranium and
thorium geochemistry have been revealed
from these experiments. They appear to
be incorporated in diopside solid solu-
tions without the aid of other trace ele-
ments. Their partitioning between diop-
side and melt does not seem to be ap-
preciably dependent on the composition
of the melt or the temperature or pres-
sure of the system. The partitioning of
uranium was independent of its concen-
tration, indicating that it acts as a dilute
solute at concentrations found in natural
systems.
These properties considerably simplify
the application of the experimental data
to studies of geologic processes. For ex-
ample, trace-element concentrations
carry information concerning the genetic
relationships of various liquid and solid
phases. From the concentrations of ura-
nium and other trace elements, it appears
that ultramafic rocks and basalts from
oceanic ridge systems have been formed
in the same fractionation event (Seitz
and Hart, 1973). Because uranium par-
titioning is not sensitive to pressure, this
conclusion does not depend on assump-
tions concerning the depth at which frac-
tionation took place.
The methods developed to study tho-
rium and uranium partitioning between
silicate and phosphate phases are ap-
plicable to other geophysical problems.
The isotope plutonium-239 has a large
thermal neutron cross section for fission
and may be studied in much the same
manner as uranium. The radioactive
elements thorium, uranium, and plu-
tonium were formed with the other ele-
ments in a stellar environment by suc-
cessive neutron capture, and their half-
lives provide a mechanism for determin-
ing when this synthesis took place.
Models of nucleosynthesis depend, among
other things, on the 244Pu/238U ratio
measured in meteorites. This ratio is not
constant but varies by more than a
factor of 3 between different phases of
586
CARNEGIE INSTITUTION
the same meteorite and by an even larger
factor between whole meteorites of the
same apparent age (Podosek and Lewis,
1972). Thorium and uranium are known
to fractionate on a microscopic scale in
natural samples (Burnett et al., 1971),
and it is likely that plutonium and
uranium have also chemically fraction-
ated and caused the observed variability
in the 244Pu/238U ratio.
In regard to this application, it is in-
teresting to note that the whitlockite
grown in the present experiments con-
tains small amounts of Si02 and MgO as
did the uranium- and thorium-rich phos-
phate of the St. Severin meteorite (Seitz,
Year Book 71, p. 553). A mixture of
15% apatite, 29% diopside, 25.5%
albite and 25.5% anorthite was found to
have a liquidus temperature of 1340°C
and to nucleate large fluorapatite crystals
upon slow cooling. Partitioning studies
on these and other minerals found in
meteorites should lead to a better under-
standing of the nucleo-synthesis chrono-
logy.
Uranium and Thorium Diffusion in
Diopside and Fluorapatite
M. G. Seitz
As a crystal grows or dissolves in a
melt, equilibrium between the center of the
crystal and the melt is controlled by dif-
fusion through the crystal. The diffusion
rate of the migrating species may be
used to determine the extent to which a
particular system has attained equi-
librium.
Diffusion rates of uranium and tho-
rium in geologically important crystals
were estimated using the trace-element
mapping techniques developed for par-
titioning studies (Seitz, this Report).
Crystals of diopside or fluorapatite were
seeded in a uranium- or thorium-rich
glass and equilibrated at high tempera-
tures for periods of 1 to 300 hours. The
trace elements proceeded to diffuse from
the rims into the centers of the grains.
The development of the diffusion profiles
was recorded by mapping the trace ele-
ments in the resulting charges. No fila-
ments or other penetration features of
the trace element in the crystals were
observed, indicating that absorption
along cracks or crystal boundaries had
not occurred. Natural crystals of zircon,
having rims of high uranium concentra-
tion, were annealed to promote uranium
migration. The extent of diffusion was
then determined from a comparison of
the initial and final uranium profiles.
Diopside
Exceptionally pure diopside crystals
obtained from the Twin Lakes locality,
California (Yoder, this Report) , were
free of aluminum (less than 0.04%) and
contained less than 0.1 ppb uranium-235.
Crystals about 100 /xm in size were seeded
into a glass of composition 40% diopside,
30% albite, 30% anorthite, and 10 ppm
uranium-235 by weight. The uranium
concentration of diopside in equilibrium
with the melt was predicted from the
measured partition coefficients (Seitz,
this Report) to be 20 ppb and easily de-
tectable by fission-track mapping. Runs
were made in the diopside-melt field at
1240°C for 1, 4, 15, and 96 hours. The
uranium distribution in the quenched
material was mapped in a muscovite
detector, using a neutron dose of 2 X
1017 ft/cm2.
It was found that none of the charges
except the one run for 96 hours contained
a detectable diffusion profile. In the
96-hour run fission tracks extended to a
depth of 12 fxm into the crystal area.
Track density profiles were measured
from grain-melt surfaces that were per-
pendicular to the polished surface. Long
tracks originated near the surface of the
crystals, and their spatial position could
be accurately determined by viewing
them in reflected light (Seitz, Walker,
and Carpenter, 1973). Both the density
of all tracks and of those greater than 6
/xm in length were measured. The re-
GEOPHYSICAL LABORATORY
587
sultant track distribution was consistent
with a diffusion profile in the diopside
having a uranium concentration equal
to 50% of the final equilibrium concen-
tration at a depth of 5 /mn in the crystals.
Because of the short penetration of the
profile, the diffusion process could be
accurately described as proceeding in one
direction perpendicular to the crystal-
melt plane. Assuming that there was no
rate-controlling process at the crystal-
melt interface, the uranium diffusion rate
in diopside at 1240 °C was estimated
from Fick's law of diffusion to be 9.9 X
10~13 cm2 sec-1. It is interesting to note
that for the five crystal-melt interfaces
investigated, the diffusion profiles ap-
peared to be the same, indicating that
the diffusion rate is not strongly depend-
ent on crystal direction.
The resolution of the diffusion profile
was governed by the range of the fission
particles in diopside. When only long-
tracks were counted, the resolution was
lowered to about 2 ^m. Because of the
short penetration depth of the profile,
however, the accuracy of the diffusion
constant measurement is not high.
Because of the absence of aluminum
in the initial grains, the diffusion rate
of this element in diopside could be de-
termined as well. Aluminum analyses
were done by electron microprobe. No
diffusion profiles were observed for runs
shorter than 96 hours. An equilibration
time of 256 hours produced an aluminum
profile having 0.3% aluminum at a dis-
tance of about 20 /xm from the crystal-
melt interface. This value is half the
concentration level found at the edge
of the crystals, which is taken as
the equilibrium concentration (consistent
with the value of 1.4 wt % Al2Oa found
in diopside grown from a similar melt;
see Seitz, this Report). These measure-
ments resulted in a calculated diffusion
rate for aluminum in diopside of 6 X
10"12 cm2 sec-1 and is higher than the
diffusion rate estimated for uranium.
Fluorapatite
Fluorapatite from the Durango, Mex-
ico, locality was ground and sieved to
produce a crystal population about 200
fim in size. These crystals were added
to an equal proportion of glass, composed
of 40% diopside, 30% albite, 30% an-
orthite, and 30 ppm thorium-230 by
weight. Runs were made in the apatite-
melt field at 1240° and 1337°C for 271
hours. From microprobe analyses it
was found that the crystals partially dis-
solved to form a homogeneous melt hav-
ing 4.0% P205 at 1240° and 7.2%
P,05 at 1337°C.
Alpha particles from thorium-230 were
registered in cellulose nitrate plastic and
served as a map of the thorium in the
quenched samples. It was found that the
thorium had only slightly penetrated the
grains in spite of the great tendency of
apatite to incorporate thorium (Burnett
et al., 1971). Because of the longer range
of alpha particles compared with fission
fragments, the spatial resolution was not
so high as that obtained for uranium.
The equilibrium concentration of thorium
in apatite is not known and was as-
sumed to be one half the concentration
of the melt. This value is consistent with
thorium partitioning estimated from na-
tural systems (Burnett et al., 1971) using
the measured partition coefficient be-
tween whitlockite and melt (Seitz, this
Report) . The penetration depth of the
diffusion profile was measured in cylin-
drical grains that had been polished
parallel to their long axis. Track densi-
ties were measured from the center of
the polished surface. For the run at
1337 °C the grains contained one fifth of
their equilibrium concentration at an
average depth of 8.5 /xin, giving a dif-
fusion rate of 5.6 X 10"13 cm2 sec"1. The
existence of a diffusion profile for the
1240°C run could not be established.
From the limits of the measurements,
however, the diffusion rate was deter-
mined to be less than 2.0 X 10"13 cm2
sec-1.
588
CARNEGIE INSTITUTION
The final apatite grains contained
0.22% magnesium compared with less
than 0.05% in the initial grains. A con-
centration of 0.23% magnesium was
found in fluorapatite crystals grown from
a similar melt. The influx of magnesium
most likely occurred by diffusion, the
rate of which would have to be greater
than 2.8 X 10"10 cm2 sec"1 at 1240°C.
This high value is comparable to meas-
ured diffusion rates of magnesium in
olivine (Meisner, Year Book 71, p. 516).
Zircon
Single zircon crystals from quartzite
inclusions in a pegmatite from the Mar-
tell Valley, Italy, contain rims of high
uranium concentration (Grauert and
Seitz, this Year Book, Annual Report
of the Director, Department of Terres-
trial Magnetism) and are suitable for
uranium diffusion studies. Annealing
these grains for 51 hours at 1200 °C pro-
duced no perceptible change in the
uranium gradients. From this experi-
ment, the diffusion rate of uranium in
zircon was estimated to be less than
1.9 X 10"12 cm2 sec"1 at 1200°C.
Discussion
The particle-track technique of map-
ping the diffusion profile of the migrat-
ing species is applicable to a number of
other elements. Experiments have been
conducted that show the feasibility of
studying oxygen diffusion by this method.
Silicate glass fragments held at 700 °C,
5 kbar pressure, in water having 1% oxy-
gen-17 were permeated by oxygen from
the water to a depth of 50 /mi in 7 days.
The extent of oxygen penetration was re-
vealed by recording alpha particles in
plastic from the thermal neutron reac-
tion with oxygen-17. Similarly, alpha
particles from neutron reactions with
boron and lithium may be recorded in
cellulose nitrate plastic (Seitz, this Re-
port). Other stable isotopes, such as
chlorine-35 and sulfur-33, undergo neu-
tron-alpha reactions and are potentially
amenable to this method of diffusion
study.
The experimental method used in this
study results in intimate contact between
the solid and the melt that serves as the
element reservoir and thus has advan-
tages over solid-solid couples where sur-
face irregularities may control the rate
of diffusion. The diffusion profiles of the
trace elements are seen directly, and
diffusion rates are calculated from a
simple diffusion geometry.
Boron Mapping and Partitioning in
Synthetic and Natural Systems:
Crystal- Melt Assemblages, Garnet
Lherzolite, Chondrites
M. G. Seitz
Boron, being chemically similar to
aluminum, forms compounds having
aluminum analogues such as CaB2Si208
(danburite) , suggesting that boron may
substitute into common mineral phases.
Following the analogy with aluminum,
the concentration and distribution of
boron between mineral phases may be
a valuable recorder of the pressure and
thermal history of rocks.
Boron is rare in terrestrial material
and in meteorites, its concentration in
the crust of the earth being less than
10 ppm. Because boron is usually present
in trace quantities and its microdistribu-
tion is difficult to determine, the par-
titioning of boron in rocks has received
little attention. Boron undergoes an
(n, a) reaction upon thermal neutron
bombardment, and the reaction products
may be registered in cellulose nitrate
plastics. This behavior forms the basis
for a technique that reveals the micro-
distribution of boron in natural samples
at concentrations below 1.0 ppm. In spite
of its low abundance, boron is a major
contributor of neutron-induced particle
tracks from natural materials. In oceanic
ultramafic rocks, for example, induced
tracks from boron are about 50 times
GEOPHYSICAL LABORATORY
589
more abundant than induced tracks from
all other elements (Seitz and Hart, 1973) .
The technique described in the follow-
ing paragraphs was used to determine
boron partitioning between olivine, ortho-
pyroxene, clinopyroxene, spinel, and co-
existing melt at high pressures and
temperatures and to map boron in a
garnet lherzolite and two meteorites of
chondritic composition.
Samples
The boron distribution was studied in
two crystal-melt assemblages synthesized
in the laboratory. The first sample of
orthopyroxene, olivine, spinel, and melt
was equilibrated at 10 kbar and 1375°C
from a mixture of natural olivine 20 wt
%, orthopyroxene 20%, clinopyroxene
30%, garnet 30% (sample JJG-352,3 de-
scribed by Kushiro, 1973) . The second
sample consisted of diopside plus melt
formed at 25 kbar, 1150°C, from a syn-
thetic diopside 50 wt %, albite 25%, an-
orthite 25% mixture plus 8% water. In
an inital survey of boron distributions in
natural materials, the following samples
were studied: a garnet lherzolite inclusion
from the Bultfontein kimberlite mine,
Republic of South Africa (Kushiro,
1973) ; a sample of the St. Severin am-
photerite; and fragments of a gehlenite-
fassaite chondrule of the Allende mete-
orite (Clark et al, 1970) .
Analytical Method
The samples were mounted in epoxy,
ground to a fresh interior surface, and
polished. The mounts were then cleaned
with distilled water in an ultrasonic
cleaner, covered with a cellulose nitrate
detector, and clamped to a similar mount
of a standard glass containing 32 ppm
boron of natural isotopic composition.
Two cellulose nitrate plastics were used.
The plastic from the Nixon-Baldwin
Chemical Company, U.S.A., was polished
to remove scratches and artifacts from
the surface. The second plastic obtained
from the Kodak Pathe Company, France,
consisted of a cellulose nitrate layer on
an Estar base and was used without
further preparation. The assemblies were
irradiated with doses of 2 X 1018 to 2 X
1014 n/cm2 in the National Bureau of
Standards reactor facility RT-4 having
an unshielded to shielded (0.05-cm-thick
cadmium) activation ratio of 395 for
cobalt. After irradiation, the plastics
were removed and treated in 6.2 N NaOH
solution at 53 °C. The etching time varied
from 3 minutes for the Nixon-Baldwin
plastic to 6 minutes for the Kodak
Pathe plastic.
Results
After being etched, the plastics con-
tained an abundance of particle tracks.
The track lengths varied from zero to
a maximum length of 6.2 /xm and could
be viewed and counted in a manner
completely analogous to that used for
fission tracks. Both plastics had similar
track registration characteristics, and
the Kodak Pathe plastic was arbitrarily
chosen for detailed track analysis.
The large number of tracks from these
samples may be attributed only to neu-
tron reactions with lithium or boron. In
a previous study of oceanic ultra-mafic
rocks (Seitz and Hart, 1973) it had been
determined that 98.4% of the tracks
registered in cellulose nitrate were from
boron and 1.5% were from lithium. It
is reasonable to assume that the relative
proportions of tracks from boron and
lithium in the lherzolite are comparable
to those determined for the oceanic ultra-
mafic rocks. By the procedure outlined
by Seitz and Hart (1973), it was cal-
culated that 98.5% of the tracks from
the Allende chondrite and 95% of the
tracks from the St. Severin amphoterite
are from boron. Boron-to-lithium ratios
in meteorites given by Quijano-Rico and
Wanke (1969) were used in the calcula-
tion. Thus it is reasonable to assume, as
a first approximation, that in these
samples all tracks are from boron.
590
CARNEGIE INSTITUTION
Because neither boron nor lithium was
specifically added to the diopside-melt
assemblage, it is uncertain which of these
elements is responsible for the observed
particle tracks. Because boron is more
abundant in nature and produces particle
tracks nearly 15 times more efficiently
than lithium, however, it is assumed that
most of the tracks in this sample are
from boron.
From the above considerations, the
etched plastic detectors serve as maps
of the boron concentrations in the
samples. The boron map of the reequi-
librated garnet lherzolite is shown in
Fig. 97. The tracks are not resolvable at
this low magnification but blend to form
the shading in the photograph. The glass
that was quenched from the melt cor-
responds to the dark regions of high
track density and is rich in boron. The
large euhedral orthopyroxene and smaller
olivine and spinel grains, being compara-
tively poor in boron, form the sharply
defined areas of low track density.
The track densities from the glass and
large mineral phases were uniform, indi-
cating that the boron had attained chem-
ical equilibrium in the system. The
boron concentrations were calculated by
multiplying the ratio of the track den-
sities from the unknown and standard
by the boron concentration in the stand-
ard glass. The track density in plastics
that were sandwiched together during
irradiation was found to be attributable
solely to neutron-alpha reactions with
oxygen- 17 in the detectors. The calcu-
lated track density due to oxygen in the
sample and detector was subtracted from
Fig. 97. Boron map of a reequilibrated lherzolite. The dark regions of high track density cor-
respond to quenched melt (glass) having 61 ppm boron. The light areas of low track density cor-
respond to large orthopyroxene grains and smaller olivine and spinel gains. Scale bar is 100 /mi
long.
GEOPHYSICAL LABORATORY
591
the measured track densities. These cor-
rections were important only at the low-
est track densities, the largest correction
being 10%. The concentrations and
measured partition coefficients for various
phases grown in the laboratory are listed
in Table 30.
High track densities could not be ac-
curately measured owing to some overlap
of the tracks. About 30% of the tracks
overlapped at the highest concentration,
most of these tracks still being resolv-
able. Consequently, the resulting errors
in boron concentrations above 40 ppm
were estimated to be less than 20%.
Most spinel grains were too small to give
a measurable number of tracks. The
largest grain produced only four tracks,
causing a large statistical uncertainty
in the measured boron concentrations.
The uncertainty of the other values in
Table 30 from counting statistics is less
than 6%.
The tendency of the four minerals
listed in Table 30 to incorporate boron
in their structure can be seen to follow
their affinity for aluminum. Spinel in-
corporates boron more readily than the
other minerals, followed by clinopyrox-
ene and orthopyroxene. Of the four
TABLE 30. Concentrations and Partitioning
Coefficients of Boron in Laboratory Runs
Boron
B (crystal)
Concentra-
tion, ppm
B(melt)
Lherzolite*
Orthopyroxene
2.8
0.046
Olivine
1.7
0.028
Spinel
5 ± 3
0.08
Melt
61
Diopside- Albite- Anorthite t
Clinopyroxene 2.7 0.067
Melt ' 40
* Olivine 20 wt%, orthopyroxene 20%, clino-
pyroxene 30%, garnet 30%; heated to 1450°C
then equilibrated at 1340°C, 10 kbar, for 2 hours.
f Diopside 50 wt%, albite 25%, anorthite 25%,
plus 8% water equilibrated at 1150°C, 20 kbar,
for 2 hours. The alpha tracks from neutron
bombardment were assumed to be from boron.
minerals, olivine contains the lowest con-
centration of boron relative to the glass.
Quantitatively, however, the partition-
ing differs considerably from that of
aluminum. In the sample studied here
aluminum is enriched 3-fold in the spinel
relative to the glass as compared with
a 12-fold depletion of boron in the same
mineral. Similarly, diopside and ortho-
pyroxene are enriched in aluminum com-
pared with boron. Of the four minerals,
only in olivine is aluminum more de-
pleted than boron.
Natural Samples
The boron map of the garnet lherzo-
lite showed a distribution of tracks that
correlated with mineral and textural pat-
terns in the rock. The high track densi-
ties in the map were visible without a
microscope and correlated with large
cracks and filaments of serpentinite in
the rock. The primary minerals were
associated with regions of low track
density and thus have lower boron con-
centrations than the altered material. On
a microscopic scale, lines and patches of
high track densities, presumably from
altered material, were found to permeate
the mineral grains. These features ex-
tended to a scale of less than 100 /xm.
The boron concentrations were deter-
mined in various portions of the rocks
and are given in Table 31. In measuring
the track densities from mineral grains,
care was taken not to count tracks from
filaments or spots of altered material.
The boron distributions in the minerals
do not show reaction profiles and are
thus believed not to have been changed
during rock alteration. Because the boron
concentrations in the minerals are low
compared with the rock average, it is
likely that most of the boron in the rock
was introduced upon serpentinization, as
was the case for oceanic ultramafic rocks.
Kushiro (1973) suggested that this
lherzolite was partially melted in the
mantle and depleted of its basaltic com-
ponents. From the partition coefficients
592
CARNEGIE INSTITUTION
TABLE 31. Boron Concentrations in Portions
of a Garnet Lherzolite and Two Chondrites
Boron Concen-
Phase
tration, ppm
Garnet Lherzolite
Clinopyroxene
0.37
Orthopyroxene
0.45
Olivine
0.27
Garnet
0.47
Serpentinite
34 to 43
Rock average
12
St. Severin Meteorite
Olivine
0.05
Pyroxene
1.2
Troilite
0.3
Plagioclase
~2.5
Interstitial
5.0 to 10
Average
1.4
Allende Chondrule
Gehlenite
0.35
Fassaite
0.18
Fine-grained regions
4 to 6
Hot spots
~25
Average
0.8
given in Table 30, it is estimated that a
melt in equilibrium with this mineral
assemblage would contain about 8 ppm
boron. That concentration is typical of
oceanic basalts, which may have been
derived by partial melting of mantle
material such as lherzolite.
A portion of the boron map of the St.
Severin amphoterite is shown in Fig. 98.
The boron is seen to be extremely in-
homogeneous. Much of the boron occurs
in grain boundaries or patches of fine-
grained material, containing up to 10
ppm boron. Because of the small grain
size in these regions, no phases of high
boron concentration could be identified.
Boron concentrations in major phases of
the amphoterite are given in Table 31.
As was the case in the synthetic sys-
tems, olivine has the lowest boron con-
centration of the major minerals. The
*-.r ^.m^ ™:'1L- ;%r4*
-»
Fig. 98. Boron map of the St. Severin meteorite. The boron is inhomogeneous and occurs mostly
in grain boundaries or patches of fine-grained material. Boron concentrations in the major phases
are given in Table 31. Scale bar is 100 ^m long.
GEOPHYSICAL LABORATORY
593
average boron concentration of 1.4 ppm
for the sample is the same as that found
in the Parnallee amphoterite (Quijano-
Rico and Wanke, 1969) .
Like the St. Severin meteorite, the
gehlenite-fassaite chondrule of the Al-
lende chondrite contains boron inhomo-
geneously distributed and concentrated
in fine-grained regions. It is possible
that the boron-rich phase is the layer-
lattice silicate that is prevalent in car-
bonaceous chondrites. This hydrated
phase has tentatively been identified as
chlorite, which with its structural re-
semblance to serpentine, would be ex-
pected to accommodate high concentra-
tions of boron. Boron concentrations in
the major phases and the chondrule
average are given in Table 31.
Chondrules, such as the one studied
here, have a composition that has been
interpreted as arising from silicate ma-
terial condensed at high temperature.
They would thus be expected to be de-
pleted of volatile boron. On the con-
trary, the chondrule contains an unex-
pectedly high concentration of boron, the
value of 0.8 ppm being close to the value
of 1 ppm for the entire meteorite (Clarke
et al., 1970). In this regard, the presence
of boron is similar to the anomalous pres-
ence of sodium in these chondrules.
Boron microdistributions in natural
samples can be studied by particle-track
mapping. From the partitioning co-
efficients measured here, boron concen-
trations of a garnet lherzolite indicate
that the assemblage was in equilibrium
with a melt having 8 ppm boron. The
major mineral phases of chondritic mete-
orites, like those of terrestrial ultramafic
material, are low in boron. As yet un-
identified material of high boron concen-
tration occurs in intergranular and fine-
grained portions of the meteorites. The
high concentration of boron found in a
gehlenite-fassaite chondrule of the Al-
lende meteorite is unexpected consider-
ing the premise that such chondrules
represent high-temperature condensates.
The Oxygen Isotope Geochemistry of
Siliceous Volcanic Rocks
from Iceland
K. Muehlenbachs
An interesting feature of the petrology
of Iceland is the large abundance of vol-
canic rocks with a high silica content.
Studies in experimental petrology (Car-
michael, 1964) and radiogenic isotopes
(Moorbath and Walker, 1965) indicate
that the siliceous rocks could have
arisen by crystal fractionation from basic
magma. Their sheer volume (10%) has
led others to propose an independent
source of silicic material in the Icelandic
crust (Walker, 1966; Thorarinsson,
1967) or models for generating acidic
magmas by partial melting of mantle
material (Yoder, 1973) .
The majority of recent basalts on Ice-
land have uniquely low 180 contents
(Muehlenbachs, Anderson, and Sigvalda-
son, 1972). Several possibilities have
been considered to explain this observa-
tion: (1) exchange of oxygen between
magma and hydrothermally altered, 180-
depleted wall rocks; (2) near-surface
interaction between meteoric water and
magma; (3) unusual crystallization his-
tory; and (4) an exceptional property
of the mantle under Iceland. The follow-
ing observations also bear on the prob-
lem.
Empirical and theoretical studies show
that any silicic magma derived by crys-
tal fractionation would have nearly the
same isotopic composition as the basaltic
parent (Garlick, 1966; Anderson, Clay-
ton, and Mayeda, 1971). Alteration
processes can change the isotopic compo-
sition of volcanic rocks. At low temper-
atures, the groundmass or glassy parts of
volcanic rocks exchange oxygen with
water during weathering more readily
than do phenocrysts, causing an enrich-
ment in 180 in the weathering products
(Muehlenbachs and Clayton, 1972a t.
At temperatures above 200°-300°C,
594
CARNEGIE INSTITUTION
rocks and some minerals rapidly exchange
oxygen with water. Because meteoric
waters are low in 180, hydrothermally
altered rocks are depleted in 180 (Taylor
and Forester, 1971).
Oxygen isotope analysis and Si02 con-
tent of siliceous rocks and minerals from
Iceland are shown in Table 32. Their
locations are indicated in Fig. 99. The
samples of this study display an extra-
ordinary range of isotopic composition
for volcanic rocks. The analyses of
phenocrysts and fresh glasses show that
siliceous magmas of Iceland range from
2.8 to 5.6%e. Such a range is not known
in fresh volcanic rocks from any other
location. A number of processes have
caused the more extreme values observed.
Samples from the Tertiary central vol-
canoes of Kroksfjordur and Breiddalur
show unambiguous signs of low-tempera-
ture alteration or weathering, the ground-
mass and glasses being much richer in
180 than the coexisting phenocrysts.
Quartz-cordierite xenoliths that have
probably been previously altered hydro-
thermally and are exceptionally depleted
in 180 ( — 10 to — 5%0; Muehlenbachs,
Anderson, and Sigvaldason, 1972) are
common in the violently erupted pumice
of Askja, 1875. The low 180 content of
the pumice (0.4 to 1.0%o) might be due
to mixture of xenolithic material and
basalt.
If the acidic rocks are derived by
crystal fractionation from basaltic mag-
mas, their isotopic composition should
be nearly the same as that of basalts.
The andesites from Hekla are noteworthy
in the uniformity of their isotopic com-
positions. The young and unaltered ob-
sidians from Hrafntinnuhryggur and
Prestahnukur differ in 180 content from
nearby basalts. In those areas a genetic
relationship, if one exists, cannot be es-
tablished between the rhyolites and
basalts. The emulsion rock from Breid-
dalur has been weathered and its pri-
Fig. 99. Map of Iceland. Numbers refer to sample localities of Table 32.
GEOPHYSICAL LABORATORY
TABLE 32. Iso topic Composition of Oxygen in Siliceous Rocks from Iceland
595
&180, %o (SMOW)
Loca-
Sample*
Si02,
Feld- Pyrox- Magne-
tion
Refer-
wt %
Rock
spar ene tite
Glass
on Fig.
99
encet
Kroksfjordur
1
1
KM72-23, dacite
66.5
5.3
3.8 3.0
KM72-26, dacite
67.6
5.9
2.9 2.3 -1.9
Breiddalur
6.9
2
2
Emulsion rock:
Acidic
73.0
11.6
Basic
52.4
6.9
Feldspar
5.6
Hrafntinnuhryggur
3
3
Obsidian
75.0
2.8
Askja, 1875
4
4
Viti I, pumice
66.5
1.0
Viti IV, pumice
70.6
0.4
Viti VII, pumice
69.4
0.8
Kerlingarfjoll
5
Obsidian
4.4
Prestahnukur
4.8
6
5
Obsidian
4.4
Perlite
3.8
Landmannalaugar
7
6
KM72-47A, obsidian
69.6
4.1
3.9 3.0 0.7
4.3
KM72-47D, mixed:
Acidic
4.0
Basic
4.2
4.3
KM72-47E, mixed:
Basic
4.3
4.2
BMR, obsidian
70.7
2.8
HeklaJ
8
7
8
1845:
I
53.4
4.2
4.2
1-922
53.4
4.8
1947:
4
H47-2
59.1
4.8
1970:
9
H70-3
55.2
4.9
H'70-13
4.9
H'70-14
54.5
4.8
H'70-15
5.0
4.8
H'70-16
4.8
H70-17
4.8
H'70-70
53.6
4.8
* Samples donated by R. F. Martin, G. E. Sigvaldason, R. L. Smith, and H. S. Yoder, Jr.
t References: (1) Hald, Noe-Nygaard, and Pedersen, 1971. (2) Walker, 1963; Yoder, 1973. (3)
Wright, 1915. (4) G. E. Sigvaldason, personal communication, 1971. (5) Friedman and Smith, 1951.
(6) Sigurdsson, 1970. (7) R. F. Martin, personal communication, 1973. (8) Hoppe, 1938. (9) Thora-
rinsson and Sigvaldason, 1972.
t Analysis performed at the University of Chicago.
5180 (%0) -
(18Q/160) sample - (180/160) standard
X 1000
(18Q/16Q) 8tandard
The standard is Standard Mean Ocean Water (SMOW) (Craig, 1961). Experimental procedure
same as Clayton and Mayeda (1963). Standard deviation of replicate analysis, 0.13%o-
596
CARNEGIE INSTITUTION
mary isotopic composition obscured
(Table 32) .
It is particularly instructive to study
composite rocks, rocks that result from
mixing of acidic and basaltic magmas.
A young, postglacial composite flow,
Domadalshraun, is found at Landman-
nalaugar. The oxygen isotope composi-
tions of acidic and basic fractions are
listed in Table 32. Only very small dif-
ferences in 180 content are found between
the acidic and basic parts. The isotopic
fractionation between the plagioclase and
the basalt is typical of those found else-
where (Anderson, Clayton, and Mayeda,
1971). However, the plagioclase of the
basaltic part is not in isotopic equili-
brium with that of the obsidian. Pos-
sibly the basic parts were quenched
while the acidic magma was still fluid,
during which time the phenocrysts and
acidic melt exchanged oxygen at a lower
temperature. A reasonable isotopic tem-
perature of 930°C can be calculated
(Anderson, Clayton, and Mayeda, 1971)
from the isotopic fractionations among
phenocrysts in an obsidian hand speci-
men, KM 72-47A.
Two nearby young basalt flows are
just slightly lower in 180 than Domadal-
shraun. The 180 data for Domadalshraun
are consistent with the hypothesis that
acidic magmas on Iceland are derived
from basic ones by crystal fractionation.
However, an obsidian from a neighboring
hill of Domadalshraun, which is fairly
similar to it in bulk chemistry, has a
very different isotopic composition of
2.8%, (0.7 wt % water). Such differ-
ences in isotopic composition cannot be
attributed to crystal fractionation.
A similar situation is found at Kroks-
fjordur, where the data on feldspars and
pyroxenes indicate that the two dacite
magmas differed in S180 by about l%e.
Such a difference is large and surprising
for two magmas of similar chemical
composition and age.
Large differences in isotopic composi-
tion among these rocks would result if
surface waters were involved in their
petrogenesis. Meteoric waters on Iceland
are low in 180 and deuterium. Any
putative interaction between such water
and magma would result in the depletion
of 180 in the magma. Evidence of water-
magma interaction should also be given by
the deuterium-hydrogen ratio in obsidians.
Friedman and Smith (1951) found
a SD (SMOW) value of — 150%o in an
obsidian (0.17 wt % H20) from Prestah-
nukur. This value is much lower than
that estimated for "juvenile water"
(Sheppard and Epstein, 1970) and is also
lower than any value for water found
on Iceland. Water near Prestahnukur
now has SD values of —70 to — 80%o.
Conceivably the low deuterium water
in the obsidian might be some "relict
glacial water" that entered into the
magma. The D/H ratio in a coexisting
perlite (Friedman and Smith, 1951) in-
dicates that since the time of extrusion
of the obsidian, the SD of the local water
was not much lower than — 100%o. The
anhydrous nature of Icelandic rhyolites
is not compatible with massive direct
exchange of oxygen between water and
magma.
Alternatively, two-stage models can
be proposed that produce magmatic
water with a SD value of — 150%o and
also explain features of the oxygen and
strontium isotope geochemistry of sili-
ceous rocks of Iceland. Water in hydro-
thermal minerals is depleted roughly
70%0 in deuterium (Suzoki and Epstein,
1970) relative to the water in which
they were formed. The hydrogen isotope
composition of water bound in hydro-
thermally altered rocks on Iceland may
be expected to be near — 150%o. Such
rocks are also depleted of 180 (Taylor
and Forester, 1971 ; Muehlenbachs and
Clayton, 19726), and magmas could ac-
quire these same features by large-scale
isotopic exchange with them. If, on the
other hand, hydrothermally altered ba-
salts were buried in the Icelandic crust,
parts of which might be as hot as 1000 °C
GEOPHYSICAL LABORATORY
597
(Palmasson, 1971 ; Hermance and Grillot,
1970), the hydrated basalts might be
partially melted, giving rise to low-deu-
terium and low-180 siliceous magmas.
Iceland is not older than 20 m.y. (Moor-
bath, Sigurdsson, and Goodwin, 1968) ;
hence the acidic rocks produced would
have, as is observed (Moorbath and
Walker, 1965) , very nearly the same
strontium isotope ratios as the basalts.
The highly variable 180 content of the
acidic xenoliths (Muehlenbachs, Ander-
son and Sigvaldason, 1972) and acidic
rocks found within small geographic
areas would result from the vagaries of
the hydrothermal system before burial.
Sigvaldason (personal communication,
1973) has suggested a similar model for
the generation of acidic rocks on Iceland
by the partial melting of hydrothermally
altered basalts.
The Oxygen Isotope Composition
of the 1973 Heimaey Lava
K. Muehlenbachs and S. P. Jakobsson*
A new volcanic eruption in Iceland
occurred January 23, 1973, on Heimaey,
the largest of the Vestmann Islands
(Fig. 99). Postglacial volcanism on Ice-
land is confined to relatively narrow
zones that can be subdivided on the basis
of normative character of the erupted
basalts (Jakobsson, 1972). The recent
eruption of Heimaey occurs in the zone
of alkali olivine basalts, as did the
eruption between 1963 and 1967 on
neighboring Surtsey. The last eruption
on Heimaey, the Saefell-Helgafell erup-
tion, occurring 5500 years ago (Kjar-
tansson, 1967) , also produced alkali
olivine basalt (Jakobsson, 1968). The
new material, however, is differentiated
mugearite-hawaiite (Thorarinsson et al.,
1973; Jakobsson et al, 1973).
Nepheline normative basalts are rare
on Iceland (3% of postglacial volcanic
rocks, Jakobsson, 1972) and are found
* Museum of Natural History, Reykjavik,
Iceland.
only in the Vestmann Island region and
on the Snaefellsnes peninsula. Differen-
tiated alkali rocks have been reported
previously from both locations (Si-
gurdsson, 1970; Jakobsson, 1968). Be-
tween these areas, seismic profiles across
Iceland are symmetrical, in that the Ice-
landic crust is thickest under areas of
alkali olivine basalts and thinnest under
areas of tholeiitic volcanism (Palmason,
1971). Significantly, the 180 contents of
fresh basalts from Surtsey and Snae-
fellsnes, areas of thick crust, are close to
that of basalts from other oceanic areas,
whereas most of the fresh Icelandic
tholeiitic basalts, from areas of thin
crust, are appreciably lower in 180
(Muehlenbachs, Anderson, and Sig-
valdason, 1972).
No isotopic fractionations are to be
expected accompanying magmatic dif-
ferentiation (Garlick, 1966; Anderson,
Clayton, and Mayeda, 1971). A suite of
rocks from Reunion Island (Upton and
Wadsworth, 1966), southwest Indian
Ocean, illustrates (Table 33) that re-
lated rocks as diverse in major-element
composition as trachyte and olivine
tholeiite have essentially the same oxy-
gen isotopic composition.
If the low-180 content of the tholeiitic
basalt is due to the exchange of oxygen
between magma and low-180 meteoric
water, the apparent correlation of iso-
topic composition with crustal thickness
indicates that this postulated mechanism
operates only in a shallow crust. The
above exchange mechanism should also
deplete material of 180 in shallow magma
chambers lodged in a thick crust.
On this basis, the Heimaey magma of
1973 might be expected to be depleted in
180 because its differentiated nature
indicates that it may have resided in a
shallow magma chamber for a significant
time, possibly since the previous erup-
tion of alkali olivine basalt at Saefell-
Helgafell (5500 y.b.p.).
The isotopic compositions (Table 33)
of both the new (VE 110-112) and older
(VE 93) hawaiites from Heimaey are the
598
CARNEGIE INSTITUTION
TABLE 33. Oxygen Isotope Composition of Volcanic Rocks from Heimaey
Sample
8«0, * %.
Date of Eruption
Description
VE 72
VE 93
VE 67
VE 112
VE 111
VE 110
S-l
Plagioclase
1244
Re 315
Re 364
Re 114G
Re 168
5.6
5.5
5.6
5.6
5.5
5.7
5.5
5.7
5.5
5.5
5.9
5.5
5.5
Heimaey
Last glaciation(?)
Last glaciation(?)
5500 y.b.p.
A.D. 1973
A.D. 1973
A.D. 1973
Surtseyt
A.D. 1964
A.D. 1964
A.D. 1966
Reunion Island
A.D. 1961
Hain volcanic neck, alkali olivine basalt
Dalfjall volcanic neck, hawaiite
Helgafell alkali olivine basalt
Pumice, Jan. 23, first hours, mugearite
Lava flow, Jan. 23, mugearite
Volcanic bomb, Jan. 23, mugearite
First lava, alkali olivine basalt
First lava, alkali olivine basalt
Alkali olivine basalt, Aug. 22
Syenite
Trachyte
Oceanite, groundmass
Olivine tholeiite
* Definitions and experimental procedure same as Muehlenbachs, this Report.
t Data from Muehlenbachs, Anderson, and Sigvaldason, 1972.
same as that of alkali olivine basalts
from Heimaey and Surtsey. No inter-
action of meteoric or sea water with the
magma may be inferred. If this lack of
interaction is a result of insufficient time
or too great a depth of the magma
chamber for the interaction to take place,
then the Icelandic tholeiitic basalts must
have exchanged oxygen for times longer
than 5000 years or at depths shallower
than the Heimaey magma chamber.
Magma would be depleted of 180 by
0.5%e if it exchanged oxygen with 5 wt %
sea water. If no interaction with me-
teoric water was detected because the
process is not significant, then the origins
of the low 180 tholeiitic basalts of Ice-
land are even more enigmatic.
Oxygen Isotope Compositions of Some
Basaltic Lavas from the Snake
River Plain
K. Muehlenbachs and G. T . Stone
The Snake River Plain (SRP; Pliocene
to Holocene) is an elongate structural
and topographic depression that arcs
nearly 400 miles across southern Idaho
from the rhyolitic Yellowstone plateau
(Pleistocene) on the northeast to the
tholeiitic Columbia River (Miocene) and
high-alumina Malheur (Pliocene-Holo-
cene) plateaus on the northwest and
west. The Plain includes two or three
zones of late Cenozoic crustal extension
(Stone, 1969) , having been downfaulted
in the west (Malde, 1959) and down-
folded in the east (Kirkham, 1931). It
is characterized, at least in the west, by
anomalously thick intermediate crust
(Vp = 6.7 km/sec) and a deep Mohoro-
vicic discontinuity (40-45 km; Hill and
Pakiser, 1967).
The SRP is filled by dacitic to
rhyolitic volcanic rocks, fluvial and
lacustrine sediments, and basaltic lavas,
predominantly olivine tholeiites (nomen-
clature of Yoder and Tilley, 1962) . The
latter are distinctive in their compara-
tively high iron enrichment (F/(F +
M)* = 0.64), high Ti02 (2.8 wt %),
and high P205 (0.6), low Si02 (46.6),
and low to moderate total alkalies (3.1)
(Powers, 1960a; Stone, 19663 1967).
The oxygen isotope compositions of a
variety of basaltic lavas and constituent
phenocrysts from the SRP range from
*F/(F + M) = (0.9Fe2O3 + FeO)/(0.9Fe2O3
+ FeO + MgO)(wt%).
GEOPHYSICAL LABORATORY 599
TABLE 34. Oxygen Isotope Compositions of Volcanic Rocks from the Snake River Plain, Idaho
Sample No.
&180*, %0
Description
62-P-69
Whole rock
Plagioclase
Augite
55-P-275
Whole rock
59-P-13
Glass
59-P-160
54-P-65
Plagioclase
OB-32
Whole rock
59-P-276A
Glass
5.4
5.6
5.1
5.6
5.0
4.8
5.5
6.3
5.9
72-S-16A
Xenolith
Host lava
Plagioclase
9.1
8.6
7.2
72-S-15A
Groundmass
Plagioclase
Augite
8.9
7.1
6.3
Washington
71-104 Z
Whole rock
5.5
71-114 Y
Whole rock
5.6
Oregon
Pl-7-3f
5.9
California
MLf
CA5J
5.9-6
6.9
Olivine tholeiite, Snake River Group; National Reactor
Testing Station, well No. 80, 63 feet below surface;
uppermost flow.
Madson olivine tholeiite, Snake River Group: near Malad
Springs, Gooding County, Idaho (olivine-plagioclase
pillow vitrophyre).
McKinney olivine tholeiite, Snake River Group: near
mouth of Malad River, Gooding County, Idaho.
Cognate xenolith in McKinney olivine tholeiite (olivine-
plagioclase cumulate), Snake River Group: McKinney
Butte, Gooding County, Idaho.
Olivine tholeiite (olivine and plagioclase phenocrysts),
Snake River Group: near Devil's Washboard Falls,
Gooding County, Idaho.
King Hill lava — basic end of series, iron-enriched basalt
(olivine + plagioclase + magnetite + ilmenite micro-
phenocrysts) ; Bruneau Formation, Idaho Group: near
King Hill, Elmore County, Idaho.
Cold Springs Creek lava — basic end of series, iron-
enriched pillow basalt (olivine + plagioclase + mag-
netite + ilmenite -f- apatite microphenocrysts),
Bruneau Formation, Idaho Group: Ryegrass Creek,
Elmore County, Idaho.
Granitic gneiss xenolith (oligoclase + intermediate alkali
feldspar + quartz) in lava like 72-S-15A; same location.
Plagioclase-augite andesitic porphyry, east rim of Crystal
Butte, Fremont County, Idaho; approximately 10 feet
from xenolith location.
Upper Yakima, high-titanium basalt, plagioclase pheno-
crysts, Ice Harbor Dam, near Pasco, Washington.
Upper Yakima, high-titanium basalt, plagioclase, py-
roxene, olivine phenocryst, Ice Harbor Dam, near
Pasco, Washington.
Basalt, Newberry Crater.
Basalts, Medicine Lake Highland.
Subalkaline basalt, Mono Craters.
* Definitions and experimental techniques as in Muehlenbachs, this Report.
t Data from H. P. Taylor, 1968.
X Datum from Anderson, Clayton, and Mayeda, 1971.
600
CARNEGIE INSTITUTION
5.0 to 6.3%0 (Table 34) . This spread of
8180 values is broader than expected for
a basaltic suite that presumably origi-
nated by partial melting of upper mantle
material, and the 180 content of Mc-
Kinney basalt (samples 59P13 and
59P160) is lower than that of any other
fresh continental basalt.
With the notable exception of Iceland
(Muehlenbachs, Anderson, and Sig-
valdason, 1972), the oxygen-isotope
compositions of basalts from oceanic
regions fall in the range 5.5 to 6.0%o
(Garlick, 1966; H. P. Taylor, 1968; An-
derson, Clayton, and Mayeda, 1971 ;
Muehlenbachs and Clayton, 1972a) .
Most rocks of the continental crust are
richer in 180 (8180 of shales averages
19%r, Savin and Epstein, 1970; that of
granites, 8 to 10%0, H. P. Taylor, 1968).
Basaltic magma contaminated by wall
rock during its ascent through conti-
nental crust should therefore be en-
riched in 180. Indicative of this effect
are 180-rich quartz xenocrysts in basalts
that are themselves relatively 180-rich
(H. P. Taylor, 1968) . Basalts of Medi-
cine Lake Highlands and Mono Craters
(H. P. Taylor, 1968) are slightly en-
riched in 180 relative to oceanic basalt
(8"0 = 6.8 and 6.9%c; Table 34) ; this
amount is consistent with contamination
by crustal rocks. However, some con-
tinental basalts have oxygen isotope
compositions identical with those of
oceanic basalts, thereby indicating no
significant exchange of oxygen with 180-
rich crustal material. For example, two
samples of the youngest Columbia River
basalts (Upper Yakima basalt at Ice
Harbor Dam; Wright, Grolier, and
Swanson, 1973) have 8180 values of 5.5
and 5.6%0 (Table 34).
The Snake River olivine tholeiites, like
most other continental tholeiites, are
characterized by 87Sr/86Sr values (0.706
to 0.708; Leeman and Manton, 1971)
that are markedly higher than those of
most oceanic tholeiites (0.702 to 0.704,
Hart, Year Book 70, p. 354, and Year
Book 71, pp. 288-290). The possibility
that they may have been contaminated
by relatively radiogenic continental ma-
terial must therefore be considered.
Moreover, some highly evolved Snake
River lavas have bulk chemical and
strontium isotope compositions that are
difficult to explain by simple schemes of
igneous differentiation.
Leeman and Manton (1971) con-
cluded, however, that the uniformly high
87Sr/86Sr values of Snake River olivine
tholeiites, which are widely distributed
in space and time (30,000 sq mi, 10
m.y.), are more likely characteristic of
their parent magmas than due to ho-
mogenized crustal contamination. The
oxygen isotope data presented here are
consistent with their interpretation; the
Snake River olivine tholeiites are not
contaminated or only slightly contami-
nated by 180-rich crustal material.
Of the Snake River lava samples
analyzed thus far, those having the
highest 180 contents, 8180 = 6.3 and
5.9%e, are, respectively, the most basic
representatives of the highly evolved
King Hill (F/(F + M) = 0.78 to 0.88)
and Cold Springs Creek (F/(F + M) =
0.74 to 0.90) series (Powers, 19606;
Stone, 1970) . The isotopic data are
consistent with the hypothesis that the
highly evolved lavas may have been
slightly contaminated by crustal rocks.
The King Hill rocks were reported by
Leeman and Manton (1971) to have un-
usually high and variable strontium iso-
tope ratios (0.7107 to 0.7180), which
define a pseudo-isochron "with an ap-
parent age of 870 million years and an
initial 87Sr/86Sr ratio of 9.709." They
interpreted the pseudo-isochron as a
contamination mixing line.
Although xenoliths of crustal rocks are
not common in Snake River lavas, they
have been reported in some localities;
e.g., in the Mud Lake (Stearns, 1926),
Craters of the Moon (Stearns, Crandall,
and Steward, 1938) , and Bellevue areas
(Schmidt, 1961). A granitic gneiss
xenolith from Crystal Butte west of the
Island Park caldera (72S16A; oligoclase
GEOPHYSICAL LABORATORY
601
+ intermediate alkali feldspar +
quartz) has a 8180 of 9.1%e, a reasonable
value for crustal rocks. It manifests
incipient melting on its margins and in-
ternally parallel to its banding. The host
lava, an andesine-augite porphyry of
andesitic composition, is enriched in 180.
However, its groundmass oxygen com-
position is not in equilibrium with those
of its coexisting minerals. This non-
equilibrium might be due to some very
low-temperature, secondary alteration
of the groundmass that has not affected
the phenocrysts, or the preservation by
quenching of differences in the rate of
oxygen exchange between xenoliths, melt,
and phenocrysts.
The relatively low 8180 value of Mc-
Kinney olivine tholeiite (59P13 and
59P160) is comparable to that of some
Icelandic basalts (8180 = 1.8 to 5.7%0).
Other characteristics in common between
the Snake River and Icelandic provinces
are crustal extension and postulated hot
spots or mantle plumes (Morgan, 1971) ;
low-180 meteoric waters; and the pres-
ence of high-titanium, iron-enriched
basalts in a bimodal basalt-rhyolite
assemblage. In both provinces, it re-
mains to be demonstrated whether oxy-
gen isotope compositions reflect mantle
characteristics or processes or as yet ill
defined interactions between magma and
meteoric water.
RADIOMETRIC AGE DETERMINATIONS
The Effect of Regional Metamor-
phism on U-Pb Systems in Zircon and
a Comparison with Rb-Sr Systems in
the Same Whole Rock and its
Constituent Minerals
T. E. Krogh and G. L. Davis
The ultimate usefulness of geochrono-
logy and isotope geochemistry in eluci-
dating complex geological history is
limited, not by analytical procedures,
but by the understanding of the response
of each dating system to geological
processes. This study describes the ef-
fects of a regional metamorphism on the
migration of rubidium and strontium in
whole-rock and mineral systems and on
the uranium-lead system in zircons. The
rocks chosen for this test were meta-
morphosed at least 800 m.y. after they
were formed. During the long time be-
tween formation and metamorphism,
large differences in the 87Sr/86Sr ratio
developed in rocks with different Sr-Rb
ratios. These differences, which are
easily detectable, yield a precise deter-
mination of the limits of isotopic equili-
bration for strontium during metamor-
phism. For the first time, it is shown that
a single outcrop can yield two vastly
different whole-rock Rb-Sr isochron ages.
The rock is a paragneiss composed of
intimately interlayered black and pink
bands. The black layers are amphibole,
biotite, and plagioclase; and the pink
layers are of granitic composition —
quartz, plagioclase, and microcline, with
minor biotite. At the contact between
the two rock types, potassium and rubi-
dium have migrated from the granitic
gneiss to form biotite in a 1- to 5-cm
zone at the margins of the amphibolite
layers. The Rb-Sr isochron age of the
biotite-enriched zone is 1035 m.y., mark-
ing the time of metamorphism. Calcium
from the amphibolite has migrated in the
opposite direction, resulting in the for-
mation of a microcline-free, plagioclase-
rich layer in the granitic gneiss.
Analyses of granitic layers away from
the contact yield an isochron age of
1845 m.y., which is evidence that these
layers were closed to migration of rubi-
dium and strontium during the meta-
morphism. These data demonstrate
unequivocally that isotopic equilibrium
for strontium was not attained during
the metamorphism, even on a scale of a
few centimeters. The analysis of samples
from the zone between lithologic units
where rubidium and strontium have
diffused in response to concentration
602
CARNEGIE INSTITUTION
gradients and to the formation of new
minerals provides insight as to the effect
of metamorphism on any Rb-Sr dating
system.
Studies on the U-Pb systems in zir-
cons from the rocks reveal that the
renewed growth of zircon during meta-
morphism can mask the original age and
can produce a complex age pattern on a
concordia diagram. In this study the
age of the overgrowth material was
measured directly, showing that each
grain is made up of zircon of two differ-
ent ages.
The use of the two independent decay
systems gives results that are mutually
reinforcing. In both cases, means have
been developed for obtaining the age of
the products of metamorphism and for
estimating the original time of formation
of the rocks.
In the test locale, the French River
area of the Grenville province in Ontario,
a series of paragneisses and quartzites
has been folded into a southwest-plung-
ing syncline and metamorphosed on a
regional scale to the amphibolite facies
(Lumbers, 1970) . In a previous study it
was shown that a granitic sill in the para-
gneiss section has a whole-rock Rb-Sr
age of 1725 m.y. and that isotopic equi-
librium of strontium between the adja-
cent minerals in each rock had occurred
about 1000 m.y. ago {Year Books 66,
68, 71) . In contrast to this pervasive
isotopic equilibrium between adjacent
grains, it was shown that isotopic equi-
librium did not occur between adjacent
whole-rock layers only a few centimeters
thick. These data were interpreted as
indicating that the major metamorphism
of these rocks occurred long before a
relatively minor event of recrystalliza-
tion about 1000 m.y. ago. This conclu-
sion must now be modified in the light
of new data in this Report.
Zircons in the French River Paragneiss
and Granite
Krogh and Davis {Year Book 71, pp.
566-571) gave data for zircons from one
block of granite and for three samples
of paragneiss from the French River
area. These data have been recalculated
using the new set of decay constants for
uranium (Jaffey et al., 1971) and a re-
finement of the method of correcting for
the small amount of common lead found.
The results are shown in Fig. 100. By
analyzing small portions of the zircon
populations that represent the extreme
variations in size and magnetic prop-
erties, it can be shown that for each rock
the data points fall within an area rather
than on a line on the concordia diagram
shown in Fig. 100. For the least mag-
netic zircons in each sample, the smallest
grains have the youngest apparent age
and the highest uranium content. Within
each size fraction the most magnetic
zircons are richest in uranium. Their
data points lie farthest below the con-
cordia curve.
Because data points for paragneiss
66-88 plot closest to the time of meta-
morphism (1000 m.y.) and because it has
the greatest abundance of overgrowths,
it was suspected that rims of new zircon
may have formed during metamorphism.
The smaller grains then would be af-
fected the most because of their larger
surface-to-volume ratio. This hypothesis
is also supported by a fission-track in-
vestigation by Dr. M. G. Seitz (personal
communication, 1973), which shows that
in paragneiss 66-88 the overgrowths are
10 to 100 times richer in uranium than
the cores. Zircons from each of the other
rocks studied show similar but smaller
uranium additions. A direct determina-
tion of the U-Pb age of the overgrowths
was therefore essential if the factors
contributing to the complex array of data
on Fig. 100 were to be understood.
A sample weighing 0.6 mg was pre-
pared by breaking the overgrowth ends
from 200/mi grains in alcohol under a
binocular microscope and was analyzed
by the low-contamination method pre-
viously described (Krogh, Year Book 69,
pp. 341-344; Krogh, Year Book 70,
pp. 258-266; Krogh, 1973). The lead
GEOPHYSICAL LABORATORY
603
o
1
<3-
i
O
<cf
CO
o
co
00
IT)
*3"
O
cj>
M
lO
a>
UJ
>-
CD
CD
1
lO
CD
to
o
Q.
Li_
S
1—
to
a>
a>
<o
O
LU
M
CO
to
a>
m
C\J
O
o
1
o
o
CO
O
O
QC
CD
C
o
o
Q_
M
o
=>
O
CD
C
CD
o
E
a>
o_
1
1
o
+
o
CD
e
0 □ +
>
""•»
<a
<
,5s
K
N
-U
U
a
p
cr
to
d
o
•-H
-^
e3
O
bC
<D
^T
i-t
bC
CD
to
o
•-H
-4-3
'■+3
o3
S
fcX)
CD
^ to
if)
CD <->
ro
-5 O
o3 .ft
'a N
O o
CO --h
CO -M
o3 CD
. (3
from
mag
Tj CO
O
d c3
(73 <D
ro
^^
ID
CO QJ
*51 -M
m
CD a3
fO
d ,
CM
b£ ft
?3<S
-O
Q_
o3 CD
r-
c3
o
-—
CM
CD CO
in
Ph o
oJ
J a
CD S-
d CD
CD -d
S-, -t_
Ph O
CD *•>
-P «5
S CD
2^
q
c3
cJ
co c
d c
§1
N
to
ft d
0 o
*+H .Zi
_ -t-i
5 c3
SS ^
S-i <+H
bC
in
2 -2
■ —
"0 "£
03 ?
•2 bC
-C c3
c3^
CD
d
O CD
U-a
d
• 03
O ?-
O bC
q
i— i
be d
u^ c3
^ ^
^
'902
604
CARNEGIE INSTITUTION
extracted from this small sample was
highly radiogenic (206Pb/204Pb = 8800),
and a precise age could be calculated.
The isotopic ratio and concentrations
obtained confirm the inference that new
material was added to each zircon during
a regional metamorphism about 1000
m.y. ago. The data (Fig. 100) indicate
that the overgrowths probably formed at
the same time as the large zircons in the
local pegmatite pods. The displacement
of the data point for the overgrowths
from the isochron line through data for
the pegmatite zircons is probably the
result of a small amount of older zircon
in the material analyzed. A maximum
age of 1070 m.y. is deduced if a lead-loss
trajectory parallel to that found for the
zircons in the mobilizates* is assumed.
It is now clear that each grain is com-
posed of material of two different ages,
and that the addition of new zircon and
uranium during metamorphism is a
major factor resulting in the spread of
points shown on Fig. 100, even though in
most cases obvious overgrowths are not
present.
Several general characteristics of the
zircons studied and their geologic setting
can be used to differentiate between
populations with single and multiple
stages of zircon growth. These criteria
can be applied to limited data from other
areas to detect which zircon ages prob-
ably date a specific event and which are
apparent ages between the time of for-
mation of the system and a later meta-
morphism. The type of material analyzed
and the geological setting might suggest
that multiple stages of zircon growth
have occurred. In this study, and in
others in progress, it appears that zircons
from metamorphosed igneous rocks are
least affected by metamorphism ; in these
rocks the largest grain size fractions give
apparent diffusion ages (Tilton, 1960)
approaching the time of crystallization.
* Small pegmatitic pods occurring in the
shadow zones around boudin structures in the
deformed parts of the paragneiss.
Zircons in metasedimentary rocks are
probably more likely to undergo meta-
morphic addition of uranium because
detrital grains are not in equilibrium
with the rocks in which they occur. The
hypothesis that a metamorphosed sedi-
ment will be essentially immune to
further new zircon growth is currently
being tested. The occurrence of local
mobilizates containing zircons younger
than the host implies that more than one
stage of zircon growth in the host is
probable.
The relative amount of new to old
zircon should vary with the grain size of
the sample; in this study the smallest
grains were found to have the highest
abundance of new zircon, but this need
not always be the case. A uniform layer
of new zircon on all grains will result in
more new zircon for the smallest grains,
which have a greater surface-to-volume
ratio, but if large grains grow at the
expense of small grains, the largest grains
will contain more new material. Differ-
ences in the ratio of new to old zircon are
reflected in the lead isotope data in two
ways. First, the data for the smallest
zircons will not lie on a line defined by
magnetic fractions of the coarser zircons
(Fig. 100). Second, if the ratio of Th to
U is different in the overgrowths than in
the core, changes in the relative abun-
dance of new to old material will result
in differences in the observed amount of
lead-208 relative to lead-206 (produced
by radioactive decay of Th and U,
respectively) . In this study zircons from
the grain overgrowths and from the
mobilizates have a Th-U ratio, inferred
from the abundance of lead-208 and
lead-206, about ten times lower than is
commonly observed in zircons from
igneous rocks. In the least magnetic
fraction of paragneiss zircons, the finest
grains have 60% to 80% less lead-208
than the coarsest grains, a finding con-
sistent with a higher abundance of over-
growth-type zircon in the finest fractions.
The observed variation is the opposite
of that found in the current investiga-
GEOPHYSICAL LABORATORY 605
tions for grain-size fractions of zircons The degree of displacement is not large
from igneous rocks. in this case because the indicated inter-
The disposition of the data for mag- cepts of 1000 and 1850 m.y. agree quali-
netic fractions of zircons of the same tatively with the age of the zircon over-
size may be suggestive of more than one growths and the two Rb-Sr isochron ages
stage of zircon growth. The addition of found for these rocks. It is evident, there-
a fixed amount of uranium will have the fore, that whenever two periods of
greatest effect on the grains with the growth have occurred the upper and
lowest initial concentration of uranium, lower intercepts with concordia will in-
Lines joining data points for fractions dicate ages that are too old and too
with different magnetic properties and young, respectively. The sense of rota-
different uranium concentrations will in tion of such a line will be related to the
general be rotated so that in the extreme concentration of uranium in the old and
case a lower intercept, to the right of the new component. For example, the
the origin of a concordia diagram, would age of 3550 m.y. for the Morton gneiss,
result. Zircons from each rock studied, as deduced by Goldich, Hedge, and
except paragneiss 66-88, show this effect. Stern (1970), is probably too old, and
Paragneiss 66-88 probably contains so their interpretation that the age indi-
much new material that the magnetic cated by the lower intercept is too young
fractions are reflecting properties of the is probably valid,
overgrowths. In any case, large differ-
ences in the slope of the lead-loss tra- Migration of Rubidium and Strontium
jectories for zircons from several rocks between Layers in the French
from the same area may signify that River Paragneiss
more than one period of zircon growth
has occurred Krogh et al. (Year Book 66) and
Mixtures of zircon formed at two Kr°gh and Davis (Year Book 68> re"
different times will define a single line Ported that complete isotopic mixing of
intersecting the concordia curve at each strontium had taken place between adja-
period of zircon growth. The departure cent mmeral Srams durmS regional meta-
from a single line observed in Fig. 100 morphism about 1000 m.y. ago but that a
implies that a second phenomenon has similar mixm§ had not occurred between
affected those zircons. In each grain-size adJacent layers of plagioclase-quartz-
fraction those grains richest in uranium blotlte §neiss °^ a few centimeters
lie farthest below the concordia curve. thlck- 0ne 5'5"cm la^er of biotite-amphi-
This type of phenomenon, which is ob- bohte however gave results that im-
j • i in- i ,. plied that isotopic mixing may have
served in almost all zircon populations, , , .,, ,, ,.
., , . . . ' taken place with the adjacent granitic
even n no metamorphism has occurred, , t?-xi +i • i A-& .
.. 1,1, , layers. Either this layer was different
is usually attributed to lead loss due to from aR the otherg or essentially all the
radiation damage. In Fig. 100 the hand- rubidium in the biotite-amphibolite layer
picked, clear zircons are lowest in ura- had been added during the metamor-
nium and therefore plot closest to a hy- phism. To distinguish between the possi-
pothetical line joining the primary age bilities, suitable samples were collected
and the time of metamorphism. from the same outcrop used for the study
Because all the zircons represented in of zircons just described. These were
Fig. 100 have probably lost some lead, sawed into blocks 10 cm square and
it is fair to assume that the broken line sliced parallel to the layering as repre-
shown, drawn through data for the least sented in Fig. 101. The concentrations
magnetic zircons from each rock, has of rubidium and strontium were measured
been displaced downward and rotated, by x-ray fluorescence of the powdered
606
CARNEGIE INSTITUTION
700
600-
500-
400-
300-
200-
100
0 cm
ppm Sr
Section 70- 138
»K— Section 72 -22
ppm Rb
-200
-100
-800
500-1
, 5 cm j
400-
300-
ppm Sr
200-
100 -
*-
Section (2 ;•!
>
1
2
3
4
5
6
7
8
9
10
II
12
13
14
15-16
17
18
19
20
21
300-
200-
100-
PP
m
Rb
— |
qJ
Fig. 101. Cross sections of French River paragneiss, showing variations in the concentration of
strontium and rubidium in amphibole-plagioclase-biotite gneiss (indicated by black bar) and in
quartz-plagioclase-microcline gneiss (indicated by white bar).
GEOPHYSICAL LABORATORY
607
slices, agreeing within ±10% with later
isotopic dilution determinations. The
concentration profiles in Fig. 101 are re-
markably similar. The strontium content
in the granitic gneiss increases abruptly
adjacent to each layer of biotite-amphi-
bolite, and there is a corresponding de-
crease in the strontium in the margin
of each layer of amphibolite. These
gradients reflect variations in the stron-
tium content of the microcline and
plagioclase.
The variation in the concentration of
rubidium in these sections reflects the
abundance of microcline and biotite.
Microcline, which normally makes up
about 30% of the granitic gneiss, is
totally absent in a 2— 3-cm plagioclase-
rich zone adjacent to each amphibolite
layer. The low points on the rubidium
profile indicate these zones. A zone of
rubidium enrichment at the margin of
each layer of amphibolite corresponds to
an increase in the amount of biotite. In
most sections amphibole is completely
absent in a zone about 1 cm thick at the
edge of each dark layer. The components
that could have formed microcline and
amphibole have reacted to form biotite
and plagioclase, but the trace-element
strontium contained in the feldspar failed
to reach chemical and isotopic equi-
librium even within these reaction layers.
The extent of migration of the alkali ele-
ments into the amphibolite layers varies
from 7 cm in section 72-22 to 1 cm in
section 72-23.
Because of differences in the rubidium-
to-strontium ratio, each layer developed
a different isotopic composition of stron-
tium before regional metamorphism. The
limits of isotopic mixing during meta-
morphism then can be uniquely deter-
mined. On an isochron diagram, points
for minerals or related whole-rock
samples that had the same isotopic com-
position after metamorphism will lie on
a line with a slope indicating the time
since metamorphism. In this case, migra-
tion of strontium or rubidium between
such systems during metamorphism does
not disturb this line. Migration between
systems that do not have the same iso-
topic composition of strontium will, in
general, disturb any isochronal relation-
ship in a predictable way. In one special
case, rocks with extremely low Rb-Sr
ratios, such as the biotite-amphibo-
lite layers, will accumulate negligible
amounts of radiogenic strontium prior
to metamorphism. If rubidium is added
to such a system by later metamorphic
reactions — by the formation of biotite
at the expense of amphibolite, for ex-
ample— an isochron corresponding to the
time of addition of rubidium will result.
Isotopic data for three samples show-
ing different amounts of rubidium en-
richment in section 72-22 (layers 2, 3, 4)
and enriched edge samples from section
72-23 (layer 5) and 72-21 (layer 10)
plot close to an isochron with a slope
of 1035 m.y. on Fig. 102. However, the
data do not fit the isochron within ex-
perimental error, indicating, as might be
expected, slight variations in the isotopic
composition of strontium after the meta-
morphism.
In sharp contrast, data for samples
collected from the central parts of thick,
chemically homogeneous, metasedimen-
tary layers of granitic gneiss define a
line with a slope corresponding to 1845 ±
11 m.y., within experimental error (Figs.
102 and 103). A 5-cm slice from the
central portion of a 13-cm granitic gneiss
layer (87-lc) from a rock section de-
scribed by Krogh et at. (Year Book 66)
also lies on the 1845-m.y. isochron, as
though it were closed to migration of Rb
and Sr during the metamorphism 1000
m.y. ago. The rocks that give the two
different isochrons occur in the same
outcrop and are intimately interlayered
on a scale of a few centimeters to several
meters.
The zone between lithologic units
where chemical reactions have taken
place provides insight as to the effect of
metamorphism on any whole-rock Rb-
608
CARNEGIE INSTITUTION
0.800
0.780
CO
go
0.760
0.700
■ //
70-142
70-140
/
0.2
0.6 1.0
Rb /Sr
1.4
1.8
Fig. 102. Whole-rock Rb-Sr data for layers in the French River paragneiss. Samples B to F are
a series of adjacent slabs from section 70-138 shown in Fig. 101. Samples designated 2, 3, and 4
are from section 72-22 in Fig. 101. Samples 5 and 10 are from sections 72-23 and 72-21, respec-
tively. One sigma errors for the 87Sr/86Sr ratio are 0.04%; for the Rb/Sr ratio, ±1%. Lambda
equals 1.39 X 10"11 yr1.
Sr dating system. Migration of strontium
and rubidium in rocks will generally be
affected by concentration gradients and
the distribution of strontium- or rubi-
dium-bearing minerals. The abundance
of rubidium in rocks is strongly influ-
enced by reactions involving the pro-
duction or consumption of the two major
alkali-bearing mineral groups, the potas-
sium feldspars and the micas. Strontium,
for the most part, is dispersed in both
potassium feldspar and plagioclase. For
example, the data for layers B and C
(Figs. 101 and 102) (which lost Rb to
the amphibolite) show that a loss of rubi-
dium during metamorphism will cause a
data point to be displaced to the left of
the isochron. If an isochron were to be
drawn through a series of such points,
an excessively old age would be indicated.
On the other hand, the incorporation of
these data with those for other samples
of granitic gneiss would indicate too
young an age and too high a value for
the initial 87Sr-86Sr ratio. As described
earlier, layers 2, 3, and 4 from section
72-22 gained essentially all their rubi-
dium during metamorphism and give an
isochron age indicating the time of rubi-
dium addition. In most cases, however,
GEOPHYSICAL LABORATORY
609
1.100
1.000
0.900
0.800
0.700
,'jr 66-88
£,* 70-142
|f 70-140 1
1.0
2.0
Rb /Sr
3.0
4.0
5.0
Fig. 103. Whole-rock Rb-Sr data for layers C, D, E, F, and G of French River paragneiss from
section 70-138, Fig. 101. Samples 70-142 and 66-88 are from thick layers of granitic paragneiss.
Sample 1C is a 5-cm slice from the center of a granitic gneiss layer 13 cm thick, described by
Krogh et al. (Year Book 66).
only a part of the rubidium may have
been added by the metamorphism, and
the points will lie anywhere between the
two limiting isochrons in Fig. 102.
If samples having a high concentration
of strontium and a low initial ratio lose
strontium during metamorphism, the
data point on an isochron diagram will
be shifted slightly to tliu right and the
isotopic composition will be preserved.
The projected initial 87Sr-86Sr ratio will
be too low. When the rubidium concen-
tration is fixed, an addition of strontium
displaces the point from the isochron
toward the 87Sr-86Sr ratio of the added
strontium. In a strontium diffusion
gradient such as in section 70-138 (Fig.
101), the isotopic composition of the
strontium entering the successive layers
(B, C, D, E) should be progressively
more radiogenic, and the data for layers
closest to the source of the strontium
should be displaced the most from the
isochron. A curve similar to B, C, D,
(Fig. 102) should result, but in this case
the pattern is complicated by the effect
of a loss of rubidium from layers B and
C. Data from layers E and F plot on the
1845-m.y. isochron as if their concentra-
tions of rubidium and strontium have
remained fixed since that time. It is
probable, therefore, that some migration
of strontium occurred at the time of de-
position or diagenesis. If isotopic equilib-
rium of strontium between adjacent lay-
ers was achieved during metamorphism,
the data points would lie on a line par-
allel to the 1035-m.y. isochron in Fig.
102. No such mixing occurred. If the
reaction layer of biotite in the amphi-
bolite had formed during an earlier meta-
morphism and radiogenic strontium had
simply diffused from this zone during
the lOOO-m.y.-old metamorphism, then
the strontium in layer B would be more
radiogenic than that in layer C. It is not,
thus substantiating the hypothesis that
the chemical reactions took place 1035
m.y. ago.
A random pattern of sampling is some-
times used by geochronologists. If sam-
ples were taken from this outcrop in
610 CARNEGIE INSTITUTION
such a manner, a scatter of points mination of a time of intense meta-
bounded by B, C, and D and the 1035- morphism rather than a cooling age can
m.y. isochron would have resulted. If be made.
only the granitic gneiss had been sam- The analysis of related whole-rock
pled, any samples from next to the volumes provides a simple and direct
biotite-amphibolite layers would raise way of testing for local isotopic and
the apparent initial ratio and decrease chemical equilibrium. The lack of local
the apparent age. A suite of samples col- isotopic equilibrium precludes the possi-
lected at random from the biotite-amphi- bility of attainment of isotopic equi-
bolite layers would have yielded an iso- librium on a regional scale, sometimes
chron for the time of metamorphism if invoked by geochronologists. Where the
material from the margins of the layers same host mineral for strontium occurs
had been included. In this study, the in adjacent rock layers, isotopic equi-
lithologic units and reacting components librium will also be accompanied by
have been isolated and analyzed sepa- chemical equilibrium. A lack of stron-
rately; the long time between formation tium trace-element equilibrium between
and metamorphism in this case produced related rocks will, in general, preclude
differences of such magnitude as to be the attainment of isotopic equilibrium,
easily detected. Effects for rocks with a This study shows that isotopic and chem-
shorter premetamorphic history may be ical equilibrium was not reached over
so small as to go undetected and may add distances of even a few centimeters dur-
significantly to the error in the isochron ing a regional metamorphism reaching
age. the amphibolite facies.
Conclusions the Significance of Inherited Zircons
This study of diffusion during meta- °* THE Age atnd 0rigin of Igneous
morphism, utilizing natural isotopic Rocks-an Investigation of the
tracers, depended on the establishment Ages of the Labra™r Adamellites
of the original age (1845 m.y. or older) T.E.Krogh and G.L.Davis
and the time and extent of metamorphism
(1000 to 1070 m.y.) by radiometric Recent advances in the chemical pro-
methods. The grade of these rocks prior cessing of zircons (Krogh, 1973) coupled
to metamorphism is not certain, but the with the high-precision mass spectro-
lack of old overgrowths on the zircons metry available at the Department of
and old reaction layers suggest that they Terrestrial Magnetism make possible an
had not been intensely heated earlier analytical reproducibility for U-Pb ages
than 1000 m.y. ago. The results con- of zircons of 1 to 2 m.y. for Precambrian
tribute greatly to knowledge of the geo- samples. The improved precision can
logical history of the region, illustrate now be applied to evaluate the geological
several methods for dating a period of factors that limit the accuracy of age
intense metamorphism, and demonstrate determinations. In the preceding section
the limits of isotopic equilibrium for the meaning of the variations in appar-
strontium. The extent of the metamor- ent. age found in some metamorphosed
phism is indicated by the production of igneous and sedimentary rocks was ex-
zircons in the mobilizates, the over- plored. Of particular interest are the
growths on the zircons in the paragneiss, differences in age for grains of different
and the reaction zones between layers size, color, or shape from a single sample
of contrasting composition. Each of these of unmetamorphosed igneous rocks and
products is able to withstand minor re- the possibility of recognizing and making
heating (unlike micas) and the deter- an age determination of old inherited
GEOPHYSICAL LABORATORY
611
zircons as well as new zircons formed
during magmatic crystallization. The
occurrence of an older component in the
zircons in an igneous rock should provide
unequivocal evidence for crustal con-
tamination or remobilization. Such a
component, unlike the isotopic composi-
tion of trace elements, cannot be intro-
duced or seriously modified by circulat-
ing magmatic fluids or ground water.
Inherited zircons have been recognized
and are most easily detected in young
intrusive rocks in a region of older crust,
but they might occur in some Precam-
brian rocks more than 1200 m.y. old.
The adamellite intrusions associated
with the large anorthosites in Labrador,
Canada, were chosen for testing because
assimilation might have been an impor-
tant factor in their production and be-
cause they contain relatively large
amounts of low-uranium, almost con-
cordant zircon. The ages of these rocks
are not well known, and precise data are
required to establish their genetic rela-
tionship to the anorthosite. Moreover,
the results for zircons from these un-
metamorphosed intrusives will help in
evaluating the effects of regional meta-
morphism on the zircons from similar
rocks occurring in the Grenville province.
Samples of the Michikamau and Harp
Lake intrusions (Emslie, 1970) were
supplied by Dr. R. F. Emslie of the
Canadian Geological Survey. Dr. J. M.
Barton of the University of Montreal
provided a sample from an adamellite
intrusion situated near the eastern mar-
gin of the Nain anorthosite body.
On the assumption that crustal con-
tamination would probably vary in
amount from one sample to another,
zircons from two samples collected sev-
eral miles apart from both the Harp
Lake and Michikamau intrusives were
analyzed (samples 71-45,46 and 71-47,48,
respectively). The data shown in Table
35 and Fig. 104 show exceptional agree-
ment, equal to that expected if two parts
of the same sample had been analyzed.
Crustal contamination must be very
limited because the samples are con-
cordant within analytical uncertainty.
For this comparative study, only the
least magnetic fractions were analyzed.
From sample 71-45, the coarsest grains
from a fraction that was slightly more
magnetic were also analyzed. The re-
sults in Table 35 and Fig. 104 again
show excellent agreement. It appears
certain that in these rocks the zircons all
formed at the same time or that a small
TABLE 35. U-Pb Isotopic Ages for the Labrador Adamellites
Isotopic Ages, m.y.
Sample Location and Number
206pb/238TJ
207Pb/235U
207Pb/206pb
Michikamau
71-45 NO,* -65 mesh
71-45 MO, +100 mesh
71-45 M2.5, -65 mesh
71-46 NO, -65 mesh
1454.9
1455.7
1452.1
1455.9
1457.7
1457.6
1456.6
1457.5
1461.7
1460.5
1463.3
1459.8
Harp Lake
71-47 NO, -100 + 200 mesh
71-48 NO, -100 + 200 mesh
1432.5
1432.2
1438.7
1438.3
1447.8
1447.2
Nain
71-50 clear grains MO, +100 mesh
71-50 turbid grains MO, +100 mesh
71-50 NO, +100 mesh
1288.0
1308.0
1292.1
1290.5
1319.0
1294.7
1294.6
1336.3
1299.1
Decay constants: 238U = 0.15513 X 10~9 yr"1;
235U = 0.98485
X
IO"9 yr"1; 231
3TJ/235TJ = 137.88
NO, nonmagnetic at 0°; MO, magnetic; Frantz Isodynamic Separator.
612
CARNEGIE INSTITUTION
0.27
0.26
0.25
Q.
0 24
0.23
0.22
0.21
71-45 M2\
1450,
-71-46 NO
-71-45 MO
-71-45 NO
-71-47 NO
-71-48 NO
CONTINUOUS DIFFUSION
CURVE FOR 1450 M.Y.
,+ 71-50 MO TURBID
71-50 NO
71-50 MO CLEAR
2.5
2.7
2.9 3.1
207pb/235u
3.3
3.5
Fig. 104. Concordia diagram for zircons from adamellite intrusions at Harp Lake (71-47,48),
Michikamau Lake (71-45,46), and Nain (71-50), Labrador.
inherited component, if present, occurs
in all samples. Nonmagnetic zircons
from the two samples of each intrusive
body have an average difference in their
207Pb_206Pb ages of L1 my The aver.
age difference in the238U/206Pb ages
(less than 1 m.y.) is slightly better than
expected from the estimated analytical
uncertainty. Because the points are al-
most concordant, and hence lead loss is
essentially nonexistent, an age difference
between the two intrusives of about 10
m.y. is clearly resolvable. Data for a
third fraction, from sample 71-45, which
is magnetic and richer in uranium, are
slightly more discordant than those for
the other samples, as if it had lost more
lead since it formed. The point for this
sample, 71-45 M2.5° (Fig. 104), lies
close to a diffusion trajectory through
the other two points.
Three fractions of zircon taken from a
single sample of adamellite that crops
out along the eastern margin of the
Nain anorthosite give very different re-
sults. The data point for a hand-picked
fraction of perfectly clear, inclusion-free
zircons from sample 71-50 is almost con-
cordant (Fig. 104), whereas an older
apparent age is obtained from a fraction
composed of grains with cracked or tur-
bid cores, suggestive of an older zircon
component. A bulk sample composed of
the two types of material plots on a mix-
ing line between the two data points.
Most zircon grains in this population are
GEOPHYSICAL LABORATORY
613
clear, and where turbid cores are visible
they are surrounded by clear, well-
faceted overgrowths. The clear grains
apparently represent the overgrowth ma-
terial and give essentially the time of
magmatic crystallization. The turbid
grains contain an older component that
can be recognized, concentrated by hand-
picking, and verified by isotopic anal-
ysis. The data point for the core material
by itself, if it were measured, would lie
on an extension of the mixing line through
data for the clear and turbid grains. It
would plot only slightly to the right of
the point for turbid grains in Fig. 104
because the turbid grains contain 284
ppm U, whereas the clear grains contain
only 23 ppm U. The latter uranium
concentration is among the lowest values
observed for zircon from crustal rocks
and probably accounts for the fact that
little or no lead has been lost from these
zircons. The amount of lead lost by con-
tinuous diffusion is normally propor-
tional to the amount of uranium pres-
ent (Silver, 1963) so that if an inherited
component has a different uranium level
than the magmatic zircon, different de-
grees of lead loss would occur and false
diffusion trajectories and inferred ages
would result.
Inherited zircons were detected in only
one of the three adamellite bodies.
Crustal contamination, therefore, does
not appear to be essential in the produc-
tion of these magmas. Even in the rock
where inherited zircons were found, a
precise age for the time of crystallization
was determined by analyzing properly
selected zircon grains.
Anomalous Isotopic Composition of
Lead in Young Zircons
James M . Mattinson
In an isotopic study of plutonic rocks
from the Salinian block of California
(Mattinson, Hopson, and T. E. Davis,
Year Book 71 ) , small but consistent devi-
ations in the U-Pb systematics of zircons
were pointed out. Zircon fractions sepa-
rated from single populations had nearly
identical 206Pb/238U ages, indicating con-
cordant behavior. The 207Pb/206Pb ages
were also essentially identical but were
consistently higher than the 206Pb/238U
ages. It seemed likely that these dis-
crepancies resulted from an error in the
rather poorly known decay constant of
235JJ
The commonly used decay constant for
235U (9.72 X 10"10 yr"1; Fleming,
Ghiorso, and Cunningham, 1952; Stieff
et al., 1959) is based on direct alpha
counting of enriched 235U and has an
estimated uncertainty of 2.2%. Banks
and Silver (1966) used natural samples
to evaluate this decay constant. On the
basis of a study of U-Pb systematics of
zircons and uranothorites from the south-
ern California batholith, they concluded
that the constant was probably accurate
to within 1%, relative to the decay con-
stant of 238U. Their zircon samples, how-
ever, had high U concentrations and had
evidently lost some Pb. The resulting
uncertainties in estimating the true ages
of the slightly discordant samples pre-
cluded further refinement of the decay
constant.
Since the study of Banks and Silver
(1966) was published, a number of im-
provements in the techniques of zircon
analysis have been made: the develop-
ment of new methods for the digestion of
zircon and chemical isolation of Pb and
U (Krogh, Year Book 70 and 1973), the
development of the silica-gel method for
Pb mass spectrometry (Cameron, Smith,
and Walker, 1969) , and the availability
of absolute isotopic Pb standards (Na-
tional Bureau of Standards). Accord-
ingly, a geologic redetermination of the
decay constant of 235U seemed worth-
while and timely. The approach of Banks
and Silver (1966) was used but with the
newer methods and more suitable samples
(low-U contents and a range of ages, in-
cluding very young zircons).
614
CARNEGIE INSTITUTION
While this study was in progress, new
and highly precise direct measurements
(by alpha counting of enriched samples)
of the decay constants of 235U and 238U
became available (Jaffey et al., 1971).
These new decay constants are in excel-
lent agreement with the results of the
present study for samples 100 m.y. old or
older (Fig. 105). Therefore, adoption of
the new constants of Jaffey et al. (1971)
for U-Pb age calculations is strongly
recommended. Younger zircon samples,
however, show progressively greater de-
viations from ideal behavior (Fig. 105) .
These deviations are not related to the
U decay constants and are the subject of
the remainder of this report.
Young zircon samples have 207Pb*/
206Pb* ratios higher than those predicted
for ideal, closed system behavior (Fig.
105). These zircons have 207Pb excesses,
206Pb deficiencies, or both. The simple
pattern of decreasing deviation with age
(Fig. 105) suggests that the Pb anomalies
are produced at the time of crystalliza-
tion or shortly thereafter, then are pro-
gressively "swamped out" by normal,
closed system accumulation of Pb.
There are several possible mechanisms
for producing such Pb excesses or defi-
Q_
(£>
O
OJ
*
Q_
£-
O
OJ
<
+ 2%
+ 1%
O
1%
"i i i i r~TT
"i 1 — i — i i i
i (4)
\
\T
\
(5H
L \#(2)
(
»\
\
\
(I) •^(7)(
.(7)
(3)
(I)
J I i i i i I
\
i i i
lOm.y.
IOO m.y.
AGE
IOOO m.y.
Fig. 105. A(207Pb*/206Pb*) 0f concordant zircon samples plotted against age. A(207Pb*/206Pb*) is
expressed as the percentage difference between measured and theoretical 207Pb*/206Pb* ratios. Posi-
tive values indicate that measured 207Pb*/206Pb* is higher than the theoretical value. Zircon data
have all been normalized to "absolute" values by applying corrections based on replicate analyses
of the NBS-983 radiogenic Pb standard. Theoretical 207Pb*/206Pb* ratios are represented by the
heavy horizontal line and are based on the following constants: X238U and X235U = 1.5513 X 10"10
yr"1 and 9.8485 X 10"10 yr"\ respectively (Jaffey et al, 1971); 238TJ/235U (atomic ratio) = 137.88.
Lower dashed line is based on the older constants used in the age determination tables of Stieff
et al. (1959) \ \™V = 1.537 X lO^yr-1; \235U = 9.72X 10"10 yr"1; 238U/235U (atomic ratio) = 137.7.
Each dot represents the average value of the number of determinations shown in parentheses.
Vertical bars show the total range of values.
GEOPHYSICAL LABORATORY
615
ciencies, including incorporation of an
older zircon component, incorporation of
intermediate daughter products (IDP's)
in other than secular equilibrium amounts
during crystallization, and loss of IDP's
from the 238U-20GPb decay chain subse-
quent to crystallization. Perhaps the most
obvious mechanism is incorporation of
an older zircon component in the zircon
sample. This mechanism is thought to be
extremely unlikely in this case. The
youngest zircons in this study (14-26
m.y., Mattinson, 1973) were collected
from a young volcanic terrane that has
no old basement rocks. Moreover, rapidly
chilled phases of the complex from which
the samples were collected contain no
zircons. Rocks that cooled more slowly
contain amounts of zircon approximately
proportional to their general grain size,
indicating that the zircons all crystallized
from the magma after its emplacement in
the country rocks. Finally, great care
was taken to prevent contamination
during all phases of mineral separation
and chemical analysis.
Incorporation of IDP's in nonequilib-
rium amounts during crystallization
could have important effects on the U-Pb
systematics of young samples. The nor-
mal equations used to calculate U-Pb
ages are based on the assumption that
for each U atom that decays a Pb atom
will be formed. Uranium does not decay
directly to Pb, however; decay proceeds
through a chain of IDP's. If, at the time
a mineral becomes a closed system with
respect to U and Pb, U is in secular
equilibrium with these IDP's, the usual
equations will be valid. If, on the other
hand, IDP's are not incorporated in equi-
librium amounts, the equations will no
longer precisely describe the evolution of
the U-Pb system with time.
Incorporation of each element in the
decay chain will be controlled by different
partition coefficients. Therefore, in most
minerals, secular equilibrium at the time
of crystallization is likely to be the
exception rather than the rule. In prac-
tice, only IDP's with relatively long half-
lives will be present in sufficient amounts
to be of quantitative significance. Of the
IDP's in the 238U-206Pb chain? only 234U
and 230Th are significant in this respect.
The 234U cannot be fractionated from
238U by normal chemical processes. It
will always be in secular equilibrium
with 238U. On the other hand, 230Th will
be strongly depleted, relative to U, in a
crystallizing zircon. For example, Tilton
et al. (1955) reported Th/U in a granite
of ca. 15. Zircon separated from the
granite has Th/U of only ca. 0.8. In this
case, [Th/U (zircon)] /[Th/U (gran-
ite)] equals ca. 0.05. Thus, only a small
fraction of the equilibrium amount of
230Th would be incorporated into a crys-
tallizing zircon. The quantitative effect
of this partitioning on the U-Pb syste-
matics of zircon can be readily calculated.
A simple equation relates the amounts of
a radioactive daughter isotope (Nd) in
secular equilibrium with its parent iso-
tope (Np) with their decay constants (kd
and Ap) : Nd/Np = Xp/\d. In the case
where p is 238U and d is 230Th, Nd/Np
equals 1.7 X 10"5 at secular equilibrium.
Each atom of 230Th represents a poten-
tial atom of 20GPb. Therefore, exclusion
of 230Th from zircon during crystalliza-
tion will result in a deficiency of 206Pb
equivalent to the amount of 230Th ex-
cluded. Combining the U-Th partitioning
data and 230Th equilibrium data discussed
above, it is possible to predict that 206Pb
(deficiency) /238U equals ca. 1.6 X 10"5.
For a lOO-m.y.-old sample this amounts
to about 0.1% of the total 206Pb in the
sample. For the youngest sample in this
study (14 m.y., Fig. 105), however, the
20GPb deficiency is about 0.7%, and the
measured 207Pb*/20GPb* ratio will be
high by that amount. This value accounts
for about a third of the discrepancy
actually observed between measured and
theoretical 207pD*/206Pb* ratios (Fig.
105).
Incorporation of excess amounts of
IDP's in the 23HJ-207Pb chain would also
616
CARNEGIE INSTITUTION
raise the observed 207Pb*/206Pb* ratios.
However, the only relatively long-lived
IDP in this chain is 231Pa. The ionic
radius of Pa4+ is 0.91 A, intermediate to
that of U4+ (0.89 A) and Th4+ (0.95 A).
This fact suggests that rather than being
incorporated in greater than equilibrium
amounts, 231Pa will be depleted, relative
to uranium, though not to the extent that
Th is. The effect will be to reduce the
observed 207Pb*/206Pb* ratio. Depend-
ing on the exact partition coefficient, as
much as one-half of the effect of 230Th
depletion might be cancelled out by 231Pa
depletion. Thus, incorporation of IDP's
in nonequilibrium amounts, though an
important factor, accounts for only part
of the observed deviations in U-Pb syste-
matics of young zircons.
Loss of IDP's from the 238Tj_206Pb
chain subsequent to crystallization is
another possible mechanism for produc-
ing 206Pb deficiencies. Some IDP's (e.g.,
Th, Pa) will be chemically bound in the
zircon because of their similarities to Zr
in ionic radius and charge. Others such
as Rn (inert gas) will not be chemically
bound and, depending on diffusion rates
and half-lives, may escape from the crys-
tal. Both U-Pb decay chains include a
Rn isotope. Rn from the 238-[j_206pk
chain (222Rn, T V2 = 3.8 d), however, is
much longer lived than Rn from the
235TJ_207pb chain (219Rn? rp yg _ 39
sec). Thus, in any time-dependent dif-
fusion process, 222Rn should be prefer-
entially lost. That Rn is lost from zir-
cons under various conditions has been
demonstrated in a number of studies
(e.g., Giletti and Kulp, 1955; Adams,
Barretto, and Clark, 1972). Usually
only 222Rn is measured, however, owing
to the difficulty in measuring the short-
lived 219Rn. The problem of preferential
loss is unresolved. Moreover, the normal
patterns of discordance observed in most
zircon studies seem best explained in
terms of loss of Pb or gain of U rather
than preferential loss of 222Rn. Never-
theless, it seems likely that 222Rn could
be lost during a geologically short inter-
val while the zircon was subjected to
magmatic or near-magmatic tempera-
tures. If such a mechanism had a high
efficiency and at the same time retention
of 207Pb was high, the total observed
isotopic anomalies (Fig. 105) could be
generated in about 3 X 105 years. More
likely, neither the escape of 222Rn nor
the retention of 207Pb would be complete.
A substantially longer time would be
required.
In summary, the results of this study
confirm the accuracy of the U decay con-
stants determined by Jaffey et al. (1971) .
However, young zircon samples have
anomalously high 207pk*/2oepk» ratios.
These high ratios are evidently the re-
sult of partitioning of IDP's during crys-
tallization and possible loss of 222Rn in
response to high-temperature conditions
subsequent to crystallization. Other
mechanisms, as yet undetermined, may
also be involved. Isotopic studies of
other minerals and still younger zircons
and measurements of IDP's in zircon
from freshly erupted volcanic rocks
would help resolve the problem.
The deviations from ideal behavior
noted in this study are such that, for
samples >100 m.y. old, they can safely
be ignored. For younger samples, 206Pb/
238U ages should be approximately cor-
rect, but 207Pb/206Pb ages should be in-
terpreted with great caution.
Age and Origin of Ophiolitic Rocks on
La Desirade Island, Lesser Antilles
Island Arc
James M. Mattinson, L. Kenneth Fink, Jr.*
and Clifford A . Hopson t
Modern concepts of global tectonics
place major emphasis on the evolution of
ocean basins and the interactions be-
* Department of Oceanography, Ira C. Dar-
ling Center, University of Maine, Walpole,
Maine.
t Department of Geological Sciences, Univer-
sity of California, Santa Barbara, California.
GEOPHYSICAL LABORATORY
617
tween oceans and continents. Ophiolites,
generally agreed to be of oceanic prove-
nance but commonly found within conti-
nental margins and intracontinental oro-
genic belts, may provide evidence of
previous episodes of sea-floor spreading
and may delineate suture zones along
which ancient ocean basins have closed
and continents collided. Ophiolites, there-
fore, are the subject of much current
research interest.
Although the oceanic provenance of
ophiolites is generally accepted, the ex-
act mode of origin remains the subject
of debate. Are ophiolites the products of
(1) normal crustal generation at mid-
ocean ridges in major ocean basins, (2)
sea-floor spreading "false starts" in
which narrow marginal basins open
briefly then close, (3) spreading behind
island arcs, (4) the earliest stages of
island arc development, or (5) some other
mechanism? Unfortunately, ophiolites in
their usual tectonic setting — orogenic
belts — are allochthonous, precluding de-
termination of their original structural
and stratigraphic position in the oceanic
environment.
Recently, autochthonous ophiolites
have been recognized in a number of
island arcs, including Eua, Tongan
Islands (Ewart and Bryan, 1972) ; the
Yap Islands (Shiraki, 1971) ; and La
Desirade Island, Lesser Antilles (this
Report). These ophiolites evidently are
the basement rocks on which the arcs
have subsequently developed. Two ori-
gins seem most likely for ophiolites in
this setting: oceanic crust, probably
formed during the early stages of opening
of an ocean basin, and early-formed
island arc rocks. In this article, the iso-
topic and age relationships of ophiolitic
rocks, calc-alkaline island arc rocks, and
adjacent crust in the Lesser Antilles are
therefore considered.
La Desirade is a small island east of
Guadeloupe in the Lesser Antilles island
arc (Fig. 106). The geology and ba-
thymetry of the region have been studied
18°
1
Virgin Is.
Q
1 1
_ Barbuda
St.
Kitts%
u
_ Antigua
16°
Guadeloupe p-TV,, o *■ Desirade
\ J Dominica
\/*l Martinique
14°
c«
\/;
/\ St. Lucia —
Barbados
12'
10° -
64°
Fig. 106. Index map of the Lesser Antilles
island arc.
by Fink (1972). The island is capped by
Lower Miocene limestone, but exposed
along its rugged coastline is an older
igneous complex comprising massive
flows and pillow lavas of spilitic-kerato-
phyric affinities, interbedded and overly-
ing cherts, and subjacent trondhjemitic
rocks. The trondhjemite is cut by swarms
of mafic-to-silicic dikes. Dredging on the
Desirade scarp northeast of the island
reveals that abundant gabbroic rocks lie
stratigraphically beneath the trondhje-
mite. Thus, the Desirade Igneous Com-
plex closely resembles the upper part of
a typical ophiolite complex and is inter-
preted as such in this report.
Feldspar and zircon were separated
from two samples of trondhjemite from
the Desirade ophiolite. The trondhjemite
was chosen for study because of its zircon
content and its relatively fresh, unaltered
nature. Lead and Sr from the trondhje-
618
CARNEGIE INSTITUTION
mite are thought to be more representa-
tive of the original isotopic composition
of the ophiolite than are Pb and Sr from
the altered volcanic rocks. The feldspars
were analyzed for Rb and Sr concentra-
tions and Sr and Pb isotopic composi-
tions. Isotopic ages of the zircons were
determined by the U-Pb method.
The Sr data, corrected for the decay
of Rb, are presented in Fig. 107, along
with previously reported data from other
ophiolites, mid-ocean ridge basalts, and
calc-alkaline volcanic rocks from the
Lesser Antilles arc. The Desirade ophio-
lite contains Sr that is considerably less
radiogenic than that previously reported
for Mediterranean ophiolites but is simi-
lar to the least radiogenic Sr reported for
the young calc-alkaline volcanic rocks of
the arc and the most radiogenic Sr de-
termined for fresh mid-ocean ridge ba-
salts (Fig. 107).
Lead from the Desirade trondhjemite
(Fig. 108) is less radiogenic than Pb from
-PA
Trondhjemite , Desirade complex (I)
Mid-ocean ridge basalts
(2)
Ophiolites, Mediterranean
(3)
Calk-alkaline volcanics, Lesser Antilles
(4)
4
3
2-
Chemically Primitive" group, Virgin Is.
(5)
0.702
0.704 0.706
87Sr/86Sr
0.708
Fig. 107. Isotopic composition of Sr from the Desirade Igneous Complex and possibly related
rocks. All Sr values are normalized to 86Sr/88Sr = 0.1194 and to an Eimer and Amend 87Sr/86Sr
value of 0.7080. (1) Sr values corrected for the decay of Rb (corrections ca. 0.0001). (2) P, Pacific
Ocean; A, Atlantic Ocean. Both average values from Hart (Year Book 69). I, Indian Ocean;
average value from Subbarao and Hedge (1973). (3) Ten samples from the Troodos ophiolite,
Cyprus, corrected for the decay of Rb (Peterman, Coleman, and Hildreth, 1972) ; two average
values from the Pindes ophiolite, Greece, uncorrected for Rb decay (Montigny, Javoy, and Al-
legre, 1970) . (4) Eocene to Quaternary volcanic rocks from the Lesser Antilles island arc, uncor-
rected for Rb decay (Hedge and Lewis, 1971; Donnelly et al, 1971). (5) Pre-Late Cretaceous
spilite and keratophyre from the Virgin Islands (Faure, Hurley, and Powell, 1965).
GEOPHYSICAL LABORATORY
619
Q_
O
\
JD
0-
00
8
Q_
O
C\J
CL
39.0-
38.0-
37.0-
Pb/204Pb
Fig. 108. Isotopic composition of Pb from the Desirade Igneous Complex (solid circle) and
possibly related rocks. Open circles, Eocene to Quaternary volcanic rocks from the Lesser Antilles
island arc from Donnelly et al. (1971). Triangles, "chemically primitive" group, Virgin Islands,
from Donnelly et al. (1971). Plus signs, Pacific Ocean Rise basalts, data of M. Tatsumoto as
reported in Church (1970). Crosses, Mid-Atlantic Ridge basalts from Tatsumoto (1966). Data
from Donnelly et al. (1971) are absolute values based on the double-spike method. All other
data are normalized to "absolute" values using average fractionation corrections.
the younger arc volcanics and more
closely resembles Pb from mid-ocean
ridge basalts. However, Pb from vol-
canic rocks from other arcs (e.g., Tonga)
is similar in isotopic composition to Pb
from the ophiolite. Thus, the common Sr
and Pb data are not sufficient to clearly
distinguish between an island arc and a
mid-ocean ridge origin.
Another group of rocks shows close
isotopic affinities with the rocks of the
Desirade ophiolite, mid-ocean ridge ba-
salts, and some arc volcanics. This is the
"chemically primitive" group (Donnelly
et al., 1971) of spilites and keratophyres
of pre-Late Cretaceous age from the
Virgin Islands (Figs. 107 and 108). Don-
nelly et al. (1971) noted the lithologic
resemblance of these rocks to some rocks
of the Desirade ophiolite. They argued
that the "chemically primitive" group
may have been generated by early arc
magmatism, a conclusion evidently based
heavily on the abundance of siliceous
keratophyres in the group (Donnelly
et al., 1971, p. 215). By analogy with the
Desirade ophiolite, the abundant silice-
ous rocks may not be representative of
the "chemically primitive" group as a
whole. Much more abundant mafic rocks
may occur at deeper stratigraphic levels
beneath the Virgin Islands.
620
CARNEGIE INSTITUTION
Perhaps the most crucial evidence re-
lating to the origin of the Desirade ophio-
lite is its age. A previously reported K-Ar
date (142 m.y., Fink, 1970) and the zir-
con dates (Table 36) indicate a Late
Jurassic age. The zircons have moder-
ately high U contents (500-1200 ppm),
and small amounts of Pb may have been
lost from some fractions of zircon as a
result of structural damage from radia-
tion. This loss would account for the
spread in 206Pb/238U ages. Considering
all the data, a crystallization age of 145-
150 m.y. seems most likely for the
trondhjemite and presumably the as-
sociated spilites, keratophyres, and
gabbros.
The next oldest rocks in the Lesser
Antilles arc are Eocene (ca. 40-55 m.y.)
in age (Donnelly et al., 1971). Thus, if
the ophiolite represents an early phase of
island arc magmatism, it preceded the
initial development of the main arc by
some 100 m.y. This possibility seems
unlikely and casts doubt on this mode of
origin for the ophiolite. The alternative
possibility, that the Desirade ophiolite
represents oceanic crust formed during
the early stages of opening of the At-
lantic or the Caribbean Ocean basins, is
considered below.
The time of opening of the Atlantic at
the latitude of the Lesser Antilles is
somewhat problematic because of the
structural complexities. According to
current interpretations, major opening of
the North Atlantic began about 180 m.y.
ago (Pitman and Talwani, 1972) and
TABLE 36. Isotopic Ages of Zircons from
Trondhjemite, Desirade Igneous Complex
Age, million years
Sample
206
Pb*/238U
207Pb*/206Pb*
Des 1 A
143
147 ± 10
B
148
129 ± 15
C
140
160 ± 35
Des 2 A
145
149 ± 6
B
141
152 ± 5
* Radiogenic
lead.
that of the South Atlantic, about 110
m.y. ago (Larson and Pitman, 1972) .
Opening of the gap between North and
South America, now occupied by the
Caribbean, may have accompanied the
early stages of opening of the North
Atlantic (Dietz and Holden, 1970; Vogt,
Anderson, and Bracey, 1971, Fig. 15).
The age of the Desirade ophiolite sug-
gests that it may represent an oceanic
crustal remnant that corresponds with
either western Atlantic crust character-
ized by the Keathley sequence of mag-
netic anomalies (Vogt, Anderson, and
Bracey, 1971; Pitman and Talwani, 1972,
Fig. 2) or perhaps the eastern Caribbean
crust. Subsequent westward underthrust-
ing that led to the development of the
Lesser Antilles island arc was initiated
when the oceanic crust broke to the east
of the site of the Desirade ophiolite.
Therefore, rather than being subducted,
the Desirade ophiolite was preserved be-
hind the trench and served as the base-
ment on which the younger, calc-alkaline
volcanic rocks were deposited.
The development of the Desirade ophi-
olite and the possible significance of its
occurrence as basement in the frontal
zone of the Lesser Antilles arc are con-
sidered in two alternative models. In
model A (Fig. 109) the ophiolite corre-
sponds with western Atlantic crust. The
following constraints apply: (1) The
Caribbean and the Atlantic Ocean crust
on each side of the arc appear to be
younger than the ophiolitic basement of
the arc itself. A Late Cretaceous (Conia-
cian) age for the floor of the Venezuelan
Basin west of the arc is indicated by the
JOIDES drilling (leg 15), and Atlantic
crust immediately east of the arc also
appears to be no older than Late Cretace-
ous on the basis of magnetic lineations
(Pitman and Talwani, 1972, Fig. 2). (2)
The Aves Ridge, between the Lesser
Antilles and Venezuelan Basin, is partly
cored by Late Cretaceous and Paleocene
(?) epizonal granodiorite and mafic vol-
GEOPHYSICAL LABORATORY
621
Atlantic Ocean
Sea level
M.A.R.
(65) —
Caribbean
Sea Aves Arc
=185) ^/>T\ (145)
iii iin Mill U1D jV jrapn— — " — ■■■■■— —^^^m
Caribbean ft ^>'^ - American Plate
Plate
Upper Cretaceous
epizonal plutons a. 65 m.y. b.p.
Aves Arc
(extinct)
Lesser Antilles Arc
MMMMM.
M.A.R.
(65)-
Shallow carbonate^
sediments/'/'
Basalt, diabase I ^ ' Oceanic crustal remnant
Granodiorite Arc volcanism (uPPer Jurassic)
(Upper Cret. - Paleocene) B.40 m.y. b.p.
Lesser Antilles Arc
Aves Swell c/y** Guadeloupe
\^ ? Desirade
Miocene to Recent
volcanics
(Basse Terre)
Jurassic ocean crustal remnant,
capped by lower Miocene limestone
^Eocene-Oligo. volcanics
and Miocene limestones C. Present
(Grande Terre)
Fig. 109. Model A for development of the Desirade ophiolite and its incorporation into the
Lesser Antilles island arc. Three stages in the convergence of the American (western Atlantic)
and Caribbean plates at a section through La Desirade are shown. Numbers in parentheses indi-
cate approximate age of oceanic crust in million years before present (m.y .b.p.). (A) Latest Cre-
taceous (65 m.y .b.p.). Oceanic crust formed 145 m.y. ago at the Mid- Atlantic Ridge (M.A.R.)
has moved eastward in close proximity to the Aves island arc (considered to be a southward ex-
tension of the eastern Greater Antilles arc). (B) Late Eocene (40 m.y .b.p.). Eastward stepping
of subduction zone causes Aves volcanic arc to become inactive and Lesser Antilles arc to develop
atop isolated remnant of Late Jurassic (145 m.y.) Atlantic Ocean crust. (C) Present time. Upper
Jurassic Atlantic Ocean crustal remnant (145 m.y. old) is exposed at La Desirade, east of the
volcanic axis of the Lesser Antilles arc.
canic rocks, covered by sediments repre-
senting shallow carbonate shelf (mid-
Eocene to Lower Miocene) and then open-
ocean planktonic (mid-Miocene to Pleis-
tocene) environments (Fox, Schreiber,
and Heezen, 1971). The Aves Ridge may
represent a volcanic arc that became
inactive in latest Cretaceous or early
Tertiary time.
Figure 109 portrays the interaction of
converging Caribbean and American
(western Atlantic) plates at the latitude
of La Desirade, from mid-Cretaceous to
the present time. Volcanism along the
eastern Greater Antilles— Aves island arc
resulted from subduction of western At-
lantic Ocean floor beneath the Caribbean
plate, then moving relatively northeast-
ward (Fig. 109A). Between Paleocene
and Late Eocene time subduction ceased
in front of the eastern Greater Antilles
(Puerto Rico and Virgin Islands) and
stepped farther eastward from the front
of the Aves arc, trapping Upper Jurassic
Atlantic Ocean crust behind the new
break (Fig. 109B). Renewed subduction
622
CARNEGIE INSTITUTION
from Late Eocene to the present led to
growth of the Lesser Antilles volcanic arc
atop the older oceanic crustal remnant
(Fig. 109B-C). This model accords
closely with that of Malfait and Dinkel-
man (1972) as it applies to the Lesser
Antilles region.
However, new seismic reflection studies
(L. K. Fink, Jr., unpublished) suggest
that the Caribbean crust may be consid-
erably older than indicated by JOIDES
drilling (leg 15, discussed above). The
new reflection profiles suggest accumula-
tion of about 2100 m of sediments in the
Venezuelan Basin. In contrast, JOIDES
penetrated only about 1200 m of sedi-
ments in the Venezuelan Basin before
bottoming in basalt "basement." The age
of the sediments immediately overlying
the basalt is Upper Cretaceous (ca. 85
m.y.) . The large discrepancy in sediment
thickness suggests that JOIDES bot-
tomed in a sill rather than true volcanic
basement. A simple extrapolation of the
JOIDES age, based on thickness of sedi-
mentary section, suggests an age of ca.
148 m.y. for sediments immediately above
the deepest seismic reflector (possibly
true basement) in the Venezuelan Basin.
It should be stressed that this age is at
best a crude approximation. The signifi-
cant point is that the Caribbean crust
Caribbean Sea
Aves Rise
Caribbean A " " Plate
A. 65m.y. b.p.
Aves Rise Lesser Antilles Arc
(extinct) p^ ,45
-(100)-
M.A.R.
II /l I) i mm a x "" m
(65)-
B. 40m. y. b.p.
Aves Swell
Lesser Antilles Arc
Guadeloupe
Desirade
iiniimjl i
M.A.R.
-(40)-
Miocene to Recent i ,,
volcanics
(Basse Terre)
Late Jurassic ocean crustal remnant,
capped by lower Miocene limestone
Eocene- Oligo. volcanics
and Miocene limestones
(Grande Terre) C. Present
Fig. 110. Model B for development of the Desirade ophiolite and its incorporation into the
Lesser Antilles island arc. Three stages in the convergence of the American (western Atlantic)
and Caribbean plates, at a section through La Desirade, are shown. (A) Latest Cretaceous (65
m.y .b.p.). Western Atlantic Ocean crust is subducted beneath eastern Caribbean Sea crust, initi-
ating arc volcanism and incipient spreading behind the arc (Aves Rise). (B) Late Eocene (40
m.y .b.p.). Continued subduction leads to the development of the Lesser Antilles island arc. The
Aves Rise ceases activity. (C) Present time. Upper Jurassic Caribbean Sea plate margin (145
m.y. old) is exposed at La Desirade, east of the volcanic axis of the Lesser Antilles arc.
GEOPHYSICAL LABORATORY
623
appears to be considerably older than
indicated by JOIDES. An older (Juras-
sic) age for Caribbean crust is more com-
patible with recent interpretations of the
evolution of the Atlantic and Caribbean
(Dietz and Holden, 1970; Vogt, Ander-
son, and Bracey, 1971, Fig. 15) . A Juras-
sic Caribbean crust is still compatible
with model A (Fig. 109) for development
of the Desirade ophiolite but is also com-
patible with an alternative model (B) in
which the Desirade ophiolite represents
the eastern edge of the Caribbean plate.
Model B is portrayed in Fig. 110. Arc
volcanism began with subduction of the
western Atlantic Ocean floor beneath the
eastern margin of the Caribbean plate
(Fig. 110A). The Aves Rise may repre-
sent a short-lived spreading center simi-
lar to the interarc spreading center in the
Tonga-Kermadec arc system (Karig,
1970). Continued subduction resulted in
development of the Lesser Antilles arc
atop Jurassic Caribbean crust (Fig.
hob, o.
Choosing between the two models is
difficult. Clearly, much more extensive
data than are now available, particularly
regarding the nature of the Caribbean
crust and the Aves Rise, will be required
before the tectonic evolution of this com-
plex area can be understood.
LUNAR PETROLOGY
Partitioning of Ti and Al and Its
Bearing on the Origin of
Mare Basalts
Jagannadham Akella and F. R. Boyd
The basalts that have flooded the mare
basins on the moon vary widely in Ti02
content from about 2 wt % to as much as
13 wt %. Those collected on the Apollo
11 and Apollo 17 missions, where the
landing sites were on the western and
northern margins of Mare Tranquillitatis,
respectively, are Ti02-rich (Fig. 111).
In contrast, mare basalts collected on the
Apollo 15 mission at the eastern edge of
Mare Imbrium and on the Apollo 12 mis-
sion in Oceanus Procellarum have
markedly less Ti02 (Fig. 111).
Experimental melting of ultramafic
compositions in which ilmenite is the
solidus phase shows that Ti02 is
markedly enriched in the liquid at tem-
peratures immediately above the solidus.
These experiments suggest that the large
differences in Ti02 content exhibited by
these two groups of mare basalts (Fig.
Ill) might be due to the presence of
ilmenite in the source rock of the high-Ti
basalts and its absence in the source
rocks of the low-Ti basalts.
Results were reported last year (Akella
and Boyd, Year Book 71) for the par-
titioning of Ti and Al between pyroxenes
and garnets coexisting with ilmenite or
rutile in subsolidus runs at pressures of
15—40 kbar. For these experiments a bulk
composition with Mg/(Mg + Fe) near
0.6 was used (Table 37) . This year an
equivalent set of experiments has been
made using a bulk composition with Mg/
(Mg + Fe) near 0.8. Results have also
been extended to atmospheric pressure
and to an examination of partitioning be-
tween crystals and liquid in experiments
in the melting interval.
Subsolidus Relations
Three assemblages were obtained at
temperatures immediately below the
solidus in the pressure range from 1 atm
to 40 kbar. The low-pressure assemblage
includes anorthite with Ca-rich pyroxene,
ilmenite, and olivine. Anorthite and
olivine react to form aluminous pyrox-
enes above about 8 kbar at temperatures
immediately below the solidus, and these
Ca-rich and Ca-poor pyroxenes coexist
with ilmenite. A reaction to form garnet
intersects the solidus at about 20 kbar,
and above that pressure the assemblage
624
CARNEGIE INSTITUTION
14
12-
10-
O
<L>
Q_
8 -
.? 6
(V
2 4
2 -
1
1 1 1 1
□
▲
i i
▲
A* A
MARE BASALTS
—
A D
A
APOLLO II A
APOLLO 12 O
APOLLO 15 •
"
-
APOLLO 17 □
0
O
1
••• !•
•
i i i i
COq
1 1
10 20 30 40 50
100 Mg/(Mg + Fe)
60
70
80
Fig. 111. A plot of Ti02 content against Mg/(Mg + Fe) for the mare basalts collected on the
Apollo 11, 12, 15, and 17 missions.
is Ca-rich pyroxene -j- garnet + ilmenite
+ olivine. Ilmenite melts at the solidus,
and olivine is the liquidus phase for both
bulk compositions over the pressure
range investigated.
Electron microprobe analyses of co-
existing phases in the subsolidus assem-
blages (Table 38) show that the Ca-rich
pyroxene contains small amounts of Ti02
(1.3 wt %) and A1203 (1.4 wt %) at
atmospheric pressure. In the intermedi-
ate pressure range (5—20 kbar) , however,
TiOo in the Ca-rich pyroxene increases
nearly fourfold and A1203 increases
nearly sixfold (Table 38) . In the garneti-
ferous assemblage obtained above 20
kbar, Ti02 in the Ca-rich pyroxenes
drops to more moderate levels (2-3 wt
%) but A1203 remains moderately high
(— 5wt%).
Ilmenite is intimately intergrown with
silicates in these runs, and it is most diffi-
cult to obtain an analysis of pure ilme-
nite. The rather large amount of Si02
(2.8 wt %) shown in the two ilmenite
analyses in Table 38 is probably due to
TABLE 37. Primary Bulk Compositions
Mg/(Mg+Fe) =
0.6
0.8
Intended
Obtained
Intended
Obtained
Si02
42.85
42.70
44.34
43.40
A1203
8.08
8.05
8.36
8.24
Ti02
6.33
6.26
6.55
6.23
CaO
14.44
14.05
14.94
14.63
MgO
12.94
13.02
17.85
17.68
FeO
15.37
14.92
7.96
8.60
GEOPHYSICAL LABORATORY
625
CO
p
P
CO
CD
XI
IB
o3
-P
Ph
fafi
a
x
cd
o
O
00
CO
w
pq
Eh
OS
I
<y
Ph
-Q
M
i— i
CO
I
a*
Ph
3
CO
I
£
<
©
a
p
«
C
03
p
go
co
CD
HI
0H
a-
o
H
<&
g
^
8
5-
1> I-H t^ O i— I T*f
t->- i-h io o os CO
cm o d d d d
CO
iC
o
o
CO CO "<*(
o
CO CM
o
o
CO CN) rjn iO iO CM
00 H O H |> C35
CO CO r- 1
"5 "* M CN O iO
O d CO d CO ©
""* T-H l-H r-H
OS !>- CO oq N rr
00 iO CM |> cd CO
"** r- I .— I
CM iO iO 00 OS CO
00 00 cm' csj io pH
O CO TfH 05 CO i>-
"2 od Tt? o cd i>
"F ,-H —4
oo d d o © d
CO ,-p ^
<M t>. io tJH CO <N
2 S o n H" h
^r co ,— i
"3 "* OS Tfi i> CM
*2 N th" ]>.' d N
"'" ~4 -H
22 © co ^ !o io
oq co cm b- os ic
CM
O o d co od
>« CO
(N CO CO OS tJh
co°od
i-h CM
CM Tt<
•^NN -^ M N
co d d oo co co
^ CM T-H
N ** CO Ttj CO tH
O H H" N CO >C
"J .—I *-4 —
00
os
OS
00
OS
00
OS
CM
d
OS
OS
00
OS
p
l-H
o
CO
OS
o
d
OS
CM
OS
d
OS
o
o
CO
d
OS
OS
OS
6°~6q°*o
^5 FH <pH 03 l_7 flj
o
iO CO
co oo
CM IO
CM 00 00 CO CM O
O CO tt CO CM
co co co i> os t^
00 00 00 CM IC <*
OS CO i— I !>. |>. CD
(Nh ^h
CO i-h iO CM O O
CM iO CO t-h CO O
00 CM t^ OS CM
OS CM t^ 00 CO O
CO CO CO O OS t£
» CO i-h CO CO
C-00 O t)< CON O
O N b- CM t- O tJh
O CO CO i-h CO OS CM
X
oo 1> <M OS CM iO CO
S»h CM <M tP
© OS IQ^
O O i-h oo »o to co
^ b- O CO CM IO CSS
W CO CO lO ^t1 r— I
rOCON CO
3
CO IO N IO 00 tt
OS CM CO OS CM <M
CO CO i-h CO OS -M
O 00 iC O CO 00
l> i- 1 CO CM "* C5
OS T—< t^.
00NNO5O5H
OS t-h CO CM
OS OS o
3h OS CO N CM CO
OS tP O Tf O 00
CM CO i— i N OS io
00 CO CO
CO CO 00 »o 00 ^
CM CO CO i-h N 00
os r>- ^ ^
.„ _• .ph 03
bfie^
OS
CO
OS
CM
00
00
OS
00
OS
CO
Os
o
o
o
o
rr
oo
OS
OS
CM
OS
OS
iO
o
o
-Hi
00
OS
OS
OS
CM
00
CM
o
CM
o
p
p
P4
o
Ph
c
in
03
of
o
a
*>
3
a
o
•rt -p
CD t-i
s- P
03 03
fac ^
<u P
T3 <!
p ...
p
CP
T3
CD
p
CD
X
o
O
O
a
I
03
O ft
o
o
iO
o
o U
CD
co Ol,
co Q
ft^ .
o 13 92
CD
"•a
o -^
^8
J2 P
la o3
CD CD
'p *s
CD CD
s^
o o o
iOO
"H <^ ft
^ ^ o
CM
+ ^ &«,
^3
^ w w —
-t-2 +-J -^ +J .«
03
■ m •— I L»H 03 l_7 CD
c
#o
'■+3
03
U
o
o
3 Oh
g 8
CD CO
HH O
-^
* <!
00 00
d d
oa
6>6>
Ph Ph
•• > — <
OS CD r^
O^X^ CD
Ph <J-43
3
626
CARNEGIE INSTITUTION
admixture of silicates. The Fe203 con-
tents of ilmenite and other phases can be
estimated from the mineral formulas
(Finger, Year Book 71). Ilmenites crys-
tallized in dry runs from the primary
bulk compositions do not contain signifi-
cant Fe203, but ilmenites crystallized in
runs to which H20 was added do contain
a few percent hematite.
The Ti02 and A1203 contents of Ca-
rich pyroxenes in runs of differing bulk
compositions are contrasted in Table 39.
All three runs were made at 1350 °C and
30 kbar, which are subsolidus conditions
in the absence of H20. Change in the bulk
Mg/(Mg + Fe) ratio of the starting
composition from 0.6 to 0.8 does not
affect either the Ti02 or A1203 contents
of the pyroxene. Addition of an excess
of silica does not affect the Ti02 content
of the pyroxene, but it does significantly
reduce the A1203. It will be shown here-
after that the addition of silica to an as-
semblage that contains liquid has a
drastic effect on the titanium content of
the pyroxene, but this effect does not
occur in the subsolidus.
Melting Relations
Pyroxenes that crystallize in equilib-
rium with liquid at temperatures above
the solidus are richer in Ti02 and A1203
than are the subsolidus pyroxenes. When
ilmenite melts at the solidus, most of the
Ti02 enters the liquid but some dissolves
in the pyroxene. The sequence of runs in
the melting interval at atmospheric pres-
sure (Table 40) illustrates this relation-
ship and also shows that the Ti02 and
A1203 contents of the Ca-rich pyroxenes
are markedly reduced when anorthite
begins to crystallize. In the highest tem-
perature run shown in Table 40 (1160°)
the Ca-rich pyroxene coexists with oli-
vine and liquid and contains 4.5 wt %
A1203 and 3.3 wt % Ti02. Anorthite is
present along with Ca-rich pyroxene, oli-
vine, and liquid in the runs at 1110° and
1130°C; the Ti02 content of the pyrox-
enes drops to 2.3-2.5 wt %, and A1203
content drops to 2.7-3.0 wt %. The Al/Ti
ratios in these pyroxenes remain near 2
(Table 40) , however, and do not show the
fluctuations that have been observed in
TABLE 39. Compositions of Clinopyroxenes in Three Bulk
Compositions Run at 1350° C and 31 kbar
Composition
0.6
0.8 0.8 + 10 wt % Si02
Si02
48.2
48.9
49.4
A1203
6.1
5.7
4.6
Ti02
2.3
2.3
2.5
CaO
19.0
17.8
14.7
MgO
13.0
16.7
17.6
FeO*
11.6
6.4
9.3
Totals
100.2
97.8
98.1
Number
of Cations ( X 1000)
Si
1808
1826
1847
Al
268
251
203
Ti
65
65
70
Ca
762
712
589
Mg
725
930
981
Fe
364
200
291
Cation totals
3992
3984
3981
* Total Fe calculated as FeO.
Assemblages: Columns 1 and 2, Ga + Cpx + 01 + Ilm; Column 3,
Ga + Cpx + Pgt + Ilm.
GEOPHYSICAL LABORATORY
627
TABLE 40. Compositional Variations in Clinopyroxene and Glass in the Melting Interval
at Atmospheric Pressure *
1130°C
1110°C
1050°C
1160°C
Cpx + 01 +
Cpx + 01 +
Cpx + 01 +
Cpx + 01 + Gl
An + G1
An + Gl
An + Ilm
Cpx
Gl
Cpx
Gl
Cpx
Gl
Cpx
Si02
47.5
37.7 (0.5) f
49.3
39.1 (0.1)
50.4
39.4 (0.4)
50.2
A1203
4.5
9.4 (0.3)
2.7
8.7 (0.0)
3.0
9.8 (0.3)
1.4
Ti02
3.3
7.5 (0.5)
2.3
9.8 (0.1)
2.5
9.4 (0.3)
1.3
CaO
20.0
16.4 (0.4)
19.7
12.2 (0.2)
20.1
12.4 (0.6)
17.4
MgO
14.7
4.0 (0.6)
14.8
5.3 (0.1)
14.9
4.3 (0.4)
13.6
FeOJ
8.7
21.9(1.1)
9.2
23.8 (0.4)
9.1
24.1 (1.3)
15.1
Totals
98.7
96.9
98.0
98.9
100.0
99.4
99.0
Number of Cations (X1000)
Si
1800
1880
1880
1930
Al
200
120
130
70
Ti
90
70
70
40
Ca
810
800
800
720
Mg
830
840
830
780
Fe
280
290
280
480
Cation totals
4010
4000
3990
4020
* Bulk composition 0.6 + 2 wt % Si02.
t Values in parentheses are standard deviations for analyses of five to ten spots.
% Total Fe calculated as FeO.
Abbreviations as in Table 38.
natural lunar pyroxenes (e.g., Bence and
Papike, 1972) or in melting experiments
on lunar rocks (Grove et al., 1973).
Kushiro (1960) and Verhoogen (1962)
have suggested that the amount of ti-
tanium that a pyroxene crystallizing from
a silicate melt will dissolve should be a
function of the silica activity in the melt.
If the silica activity is relatively low,
there will be a greater tendency for the
pyroxene to accept aluminum in tetra-
hedral coordination compensated by
titanium in octahedral coordination.
The analyses presented in Table 41 ap-
pear to provide a remarkable confirma-
tion of this hypothesis. Two sets of runs
are given, one for the 0.8 composition at
1250 °C and one for the 0.6 composition
at 1200°C. In each set, one run was made
with the original composition and a sec-
ond run with a composition in which 10%
Si02 was added to the original composi-
tion. Addition of Si02 in both sets of runs
reduces the concentration of A1203 and
Ti02 in the Ca-rich pyroxenes by factors
of 2-3. Addition of Si02 also causes the
crystallization of pigeonite rather than
olivine, but this is probably not responsi-
ble for the major change in the Cpx-liquid
equilibria because the pigeonites contain
only 1.1-1.4 wt % Ti02 and 1.4-1.7 wt %
A1203, and are much less abundant than
the Ca-rich pyroxenes.
Liquid C ompositions
Interstitial patches of clear brown
glass are present in some runs quenched
from the melting interval. In a few cases
there has been a gravitational separation
of crystals and liquid with a concentra-
tion of liquid at the top of the capsule.
Runs made at atmospheric pressure in
evacuated silica-glass tubes were
quenched by rapidly immersing the tubes
in mercury, and in these runs it was
possible to obtain glass believed to be of
liquid composition. Devitrification is
more prevalent in high-pressure runs,
and accordingly there is less certainty
628
CARNEGIE INSTITUTION
TABLE 41. Compositional Variations in the Clinopyroxenes with Change of Silica Activity
in the Liquid in Runs at 5 kbar
1250°C
1200°C
Bulk
Composition :
0.8
0.8 + 10 wt % Si02
0.6
0.6 + 10 wt % Si02
Assemblage :
Cpx + 01 + Gl
Cpx + Pgt + Gl
Cpx + 01 + Gl
Cpx + Pgt + Gl
Cpx
Gl
Cpx
Gl
Cpx
Gl
Cpx
Gl
Si02
45.1
41.9
51.3
45.9
44.5
38.0
51.0
44.1
AI2O3
7.9
11.0
2.8
9.1
7.5
11.8
2.5
8.4
Ti02
6.0
8.1
2.3
6.7
5.4
8.5
1.9
6.6
CaO
19.4
15.4
11.8
12.8
17.9
15.1
12.3
12.2
MgO
14.7
7.5
23.2
11.0
12.4
7.5
20.6
9.9
FeO*
6.7
14.8
7.2
13.1
12.5
17.9
11.4
19.4
Totals
99.8
98.7
98.6
98.5
100.2
98.8
99.7
98.6
Number of Cations ( X 1000)
Si
1676
1873
1685
1881
Al
346
121
335
109
Ti
169
64
154
53
Ca
771
462
726
486
Mg
813
1263
700
1133
Fe
208
221
396
352
Cation totals
3983
4004
3996
4014
* Total Fe calculated as FeO.
Abbreviations as in Table 38.
that glass compositions obtained repre-
sent liquid compositions.
Interstitial patches of glass in some
runs were free of devitrification products
and were sufficiently large to be analyzed
with the electron microprobe. Some ex-
amples of such analyses are given in
Table 40. The reproducibility of glass
analyses for a given run is satisfactory,
as can be judged from the standard devia-
tions given in Table 40. In one run,
where there was a pronounced gravita-
tional separation of liquid and crystals,
the composition of the interstitial glass
in the crystal-rich portion of the run was
identical with that of the crystal-poor
glass at the top of the capsule.
It is of interest to note that some of
the glasses obtained from runs in which
the phase assemblage was olivine -f-
clinopyroxene + anorthite + liquid
(Table 40) are remarkably similar in
composition to high-titanum mare ba-
salts (Table 42). Ilmenite melts at the
solidus, and at temperatures immediately
above the solidus there is a strong con-
centration of TiC>2 in the liquid. The
marked difference in Ti02 content of the
Apollo 11 and 17 basalts and the mare
basalts collected at other landing sites
might be due to the presence of ilmenite
in the source rocks that provided the
high-titanium basalts. The amount of
ilmenite that would have to be present in
TABLE 42. Comparison of Compositions of
Rock 10047 with Synthetic Glass (A-4) Co-
existing with Anorthite, Olivine, and Clino-
pyroxene at Atmospheric Pressure and
1130°C
10047*
A-4
Si02
41.3
39.1
A1203
9.8
8.7
Ti02
10.2
9.8
CaO
12.2
12.2
MgO
6.1
5.2
FeOt
19.0
23.8
MnO
0.29
Na20
0.65
K20
0.11
Others
0.45
* Rose et al. (1970).
t Total Fe calculated
as FeO.
GEOPHYSICAL LABORATORY
629
the source rocks to provide liquids with
~40 wt % Ti02 depends on the degree of
partial melting. For a few percent of
partial melting, only a small amount of
ilmenite needs to be present.
A Study of Charge-Transfer and
Crystal-Field Spectra of Iron and
Titanium in Synthetic "Basalt"
Glass as a Function of P02*
H. K. Mao and P. M. Bell
Little is known of the relationship be-
tween the optical absorption spectra of
iron and titanium in glasses and the
oxygen fugacity at the instant of quench
(~500°-900°C). The relationship, if
known, can be applied to the study of
terrestrial and lunar basaltic glass as a
method of determining the oxidation
state. Controlled oxidation experiments
may provide this relationship and data
on the mechanisms of charge-transfer as
well.
Two synthetic glasses were prepared in
air, at partial pressures of oxygen (Po2)
of 1065, 10"7, 10"8-1, 10~91, 10 116, and
in pure hydrogen. The technique em-
ploys calibrated gas mixtures of hydro-
gen and carbon dioxide (Nafziger, Ul-
mer, and Woermann, 1971). Tempera-
tures from which the glasses were
quenched in mercury were 1300°, 1350°,
and 1400 °C. The actual temperature
during the formation of glass was con-
siderably below 1200 °C, judging from
experiments with similar compositions
(J. Akella, personal communication).
Starting compositions are listed in Table
43. The glasses are Ti-rich and Fe-rich
and contain CaO, MgO, and Si02.
All glasses studied were prepared with
flat, parallel, polished surfaces and
placed with optical masks in a Cary 17-1
crystal-field spectrometer. After absorp-
tion spectra were measured, the grains
were analyzed by electron microprobe to
* Work supported in part by National Science
Foundation grant GA 22707.
TABLE 43. Synthetic Compositions
Ti Glass
Fe Glass
FeO
MgO
A1203
Si02
CaO
Ti02
0
24.71
19.31
16.46
8.19
6.98
50.85
43.34
9.99
8.51
11.66
0
Totals
100.00
100.00
guard against using samples whose com-
positions had changed significantly dur-
ing melting.
Titanium- Rich Synthetic Glass
Quenched in air, the Ti-rich glass
showed only very minor absorption and
was visibly clear with a slightly yellow
cast. Titanium is oxidized to Ti4+, in
which transitions caused by photon ab-
sorption do not occur unless coupled with
another transition element in intrava-
lence charge transfer. However, at lower
Po2 than air, Ti3+ forms, as evidenced by
a characteristic crystal-field band at
about 500 nm and charge-transfer
shoulder (ligand-metal) absorption start-
ing at approximately 375 nm. Surpris-
ingly, Ti3+ absorption was observed at
Po2 as high as 10"6 and maximizes at
approximately 10"9. Figure 112 shows a
plot of the spectrum. These data can be
used to document the onset of formation
of trivalent titanium as Po2 is lowered.
Iron-Rich Synthetic Glass
Figure 113 shows the optical absorp-
tion spectrum of the synthetic Fe-rich
glass quenched in air, at Po2 = 10"9-1,
and at P02 = 10"11,e. The air-quenched
glass shows only a strong absorption
"edge" or "shoulder" sweeping from the
ultraviolet into the near infrared. This
absorption is caused by a transition in
Fe3+ (charge transfer) . In a reduced
glass, such as the one quenched at Po2 =
10"91 shown in Fig. 120 (Mao, Virgo and
Bell, this Report) , the intensity of the
630
CARNEGIE INSTITUTION
3 „_-l
Wave number, 10 cm
25
20
.£15
sio
5 —
Synthetic
Ti silicate glass
log Po2 = -9.l 1400 °C
350
400
500
600
Wavelength, nm
1000
1500 3000
2000
Fig. 112. Optical absorption spectrum of synthetic titanium-"basalt" glass. Part of the spectrum
is off scale and has been replotted to the scale of the figure.
Wove number, 10 cm
20 10
Synthetic
Fe silicate glass
1400 °C
400
500
600
Wavelength, nm
1000
1500 2000
Fig. 113. Optical absorption spectrum of syn-
thetic iron-"basalt" glass. The dashed curve
(air) starts at 2000 nm and goes off scale to-
ward higher energy. Two extensions or off-
scale segments are also plotted by lowering
them to the scale of the figure. Therefore, start-
ing from right to left, the first segment of the
dashed curve goes off scale at 200 cm-1, the
second at 400 cm"1, and the third at 600 cm-1.
Fe3+ shoulder is less, and three distinct
bands are observed. The first band, at
about 550 nm, is probably a charge-
transfer band in Fe3+, as indicated by its
disappearance at lower P02 and its known
behavior in crystalline silicates (Bell
and Mao, Year Book 71, pp. 531-534).
The strong bands at 1 and 2 ^m are
caused by crystal-field absorptions in
Fe2+, mostly in octahedral coordination.
Structural control in these glasses is pro-
vided by the 57Fe Mossbauer data de-
scribed by Mao, Virgo, and Bell (this
Report) .
At P02 = 10"11-6, most of the Fe3+ is
reduced and only Fe3+ absorption is ob-
served. Judging from the appearance of
metallic iron as a primary phase on the
liquidus at and below P02 = 10"11-5, it is
possible that iron metal is in solution in
these glasses at low oxygen fugacity. The
Fe2+ bands become less intense as the
partial pressure of oxygen is reduced
toward pure hydrogen, supporting the
GEOPHYSICAL LABORATORY
631
idea of reduction to Fe°. A glass of simi-
lar composition, but containing both
iron and titanium in proportions approxi-
mately the same as those in the lunar
orange glass, is described by Mao, Virgo,
and Bell (this Report) .
Conclusions
The results demonstrate that it is pos-
sible to recognize absorptions for Fe3+,
Fe2+, Ti4+, and Ti3+ in complex silicate
glass. Intermediate oxidation states can
be determined from the slopes and posi-
tions of the charge-transfer edge. Charge
transfer is known to occur between Fe2+
and Ti4+, Ti4+ and Ti3+, and Fe2+ and
Fe3+, but present data suggest that
charge transfer occurs between oxygen
and Fe3+ and between oxygen and Ti3+.
The data provide a basis for analyzing
the amounts and oxidation states of iron
and titanium in basaltic glasses. An
obvious application is in the interpreta-
tion of optical spectra obtained by tele-
scope (see Mao and Bell, this Report).
The spectra for glass are sufficiently dis-
similar from crystal spectra to distinguish
the properties of both — for example, in
the glass-rich regolith of parts of the
moon's surface. Although the absorption
bands of glasses shown in Figs. 112 and
113 have been assigned previously in
studies of ceramic materials (Smith and
Cohen, 1963) and crystals, the present
data have been obtained under better
controlled conditions of oxidation-re-
duction.
Analytical Study of the Orange Lunar
Soil Returned by the Apollo 17
Astronauts
H. K. Mao, D. Virgo, and P. M. Bell
An orange-colored soil was discovered
on the edge of Shorty Crater during the
northern traverse of the Apollo 17 land-
ing site. The soil was composed mostly
of minute glass beads (mean diameter
100 /mi), colored by an absorption proc-
ess that is caused by electronic charge-
transfer mechanisms in iron and tita-
nium. The interaction of iron and
titanium absorption bands shifts in wave-
length with the degree of oxidation, and
therefore the color can be used as a
crude indication of the lunar atmosphere
that was in equilibrium with the glass
at the time of its formation.
The orange color is not caused by oxi-
dation and should not be interpreted as a
sign of alteration by water or fumarolic
activity. In this study, however, there is
evidence that the orange soil is relatively
oxidized compared with other lunar rocks.
Meteoritic components may occur in the
soil, and the areal extent of the glass
suggests that it is the product of a major
event on the lunar surface. The present
investigation is an attempt to character-
ize individual glass beads chemically and
to determine their oxidation state experi-
mentally.
In this study chemical analysis was
obtained by electron microprobe on both
lunar glass from sample 74220 and syn-
thetic glasses quenched under controlled
and measured oxygen fugacity. Optical
absorption caused by crystal-field and
charge-transfer processes in iron and
titanium present as constituents in the
lunar and synthetic glasses was measured
in order to establish a scale of oxidation.
Measurements of 57Fe Mossbauer reso-
nance provided precise structural con-
trol of the coordination of iron on which
the interpretation of the optical absorp-
tion spectra was based.
Chemical Properties of the Orange Glass
The electron microprobe analyses
given in Table 44 include the bulk sam-
ple 74220, an average for several orange
glass fragments separated from the bulk,
and an errant green glass fragment also
found in the sample. Comparison of the
major-element compositions with those
of glasses from other lunar missions (e.g.,
Bell and Mao, 1972) indicates a remark-
able similarity. Originally from a single
632
CARNEGIE INSTITUTION
TABLE 44. Electron Microprobe Analyses : Chemical Composition
Apollo 17, 74220 Orange Soil
Synthetic
Bulk*
Orange Glass
Green Glass
Fe-Ti Glass
0.74
0.59
0
0.31
0.24
0.25
0
22.01
22.21
17.68
22.48
0
0.08
0
0.42
0.24
0
14.37
15.81
12.61
14.97
6.35
5.76
10.21
6.35
38.60
38.88
47.54
39.42
0.08
0.06
0.03
0
7.74
7.17
10.01
7.74
8.85
8.70
0.71
9.04
0.05
0
0.06
0
Cr203
MnO
FeO
NiO
Na20
MgO
A1203
Si02
K20
CaO
Ti02
P205
S
Totals
98.42
99.99
99.95
100.00
* X-ray fluorescence analysis of the orange soil kindly provided by the Preliminary Examination
Team, Johnson Spacecraft Center, Houston, Texas.
source or perhaps several sources, the
orange glasses are uniform in composition
and are distributed at all the landing
sites.
It is intriguing that the sample is
enriched in volatile elements such as
chlorine, zinc, and lead but is depleted in
KREEP (postassium, rare-earth elements,
and phosphorus) and uranium (G. Reed
and P. Eberhart, personal communica-
tion, 1973). It is also significant that the
orange glass is not simply equivalent in
bulk composition to any lunar rock
analyzed.
The 57Fe Mossbauer Resonance of the
Bulk Soil and of Orange Glass Separates
Soil sample 74220,61 was initially
sieved into —200 (separate A) and +200
sieve fractions. The latter fraction was
further separated into distinct heavy
(separate B) and light (separate C) frac-
tions by progressive dilution of Clerici
solution. Under the microscope, separate
C was comprised almost entirely of
orange-colored glass spheres and glass
fragments. This separate was handpicked
to a purity greater than 99% and subse-
quently crushed under acetone for the
57Fe resonant absorption experiment.
Separate B contained minor orange glass
composites, brownish clinopyroxene, and
olivine grains. The bulk of this fraction
was black, opaque fragments. Separates
A and B were likewise crushed under
acetone prior to the 57Fe experiments.
The 57Fe resonant absorption experi-
ments were measured using the technique
previously reported (Virgo, Year Book
71). A count rate of 1-2 X 106 counts
per channel was accumulated for sepa-
rate A in attempts to detect minor soil
constituents such as magnetite, troilite,
Fe metal, and spinel. The spectra were
fitted assuming a number, -N, of Lorent-
zian lines with arbitrary positions, and
N was determined on the basis of the
mineralogy and data for previously
measured mineral and glass standards.
The density of iron in the lucite absorb-
ers was in the range 2-5 mg Fe/sq cm.
The 57Fe spectra of separates A and B
are shown in Figs. 114 and 115, and the
hyperfine parameters, quadrupole split-
ting, isomer shift, and line widths at half-
peak height are given in Table 45. The
spectra consist of several superimposed
patterns of absorption doublets charac-
teristic of high-spin ferrous iron. The
GEOPHYSICAL LABORATORY
633
Fig. 114. The iron-57 resonant absorption in
lunar soil (separate A). The solid line is a
least-squares fit assuming six Lorentzians (eigh-
teen line variables, one background variable)
with unconstrained intensities, widths, and posi-
tions. The fit is not statistically acceptable.
solid line in Figs. 114 and 115 is a least-
squares fit, assuming Lorentzian lines, to
uncorrected analyzer counts without em-
ploying constraints. The quadrupole
l/ELDCITY mm/szc
Fig. 115. The iron-57 resonant absorption in
lunar soil separate B. The fit is almost statis-
tically acceptable. See Fig. 114 caption.
splitting and isomer shifts of the outer
doublet agree closely with those for Mg-
Fe olivines, and those of the inner
doublet (well resolved in Fig. 115) agree
closely with stoichiometric ilmenite from
the Apollo 11 soil (c/. Tables 45 and 46).
TABLE 45. Isomer Shifts, Quadrupole Splittings, and Line Widths
for Apollo 17 Soil Separates A and B at 298° K
Constituent
FWHH,*
mm/sec
Isomer Shift, f
mm /sec Fe2+
Quadrupole
Splitting,
mm/sec Fe2+
Separate A
Olivine
Glass
Ilmenite
0.28-0.30
0.55-0.67
0.44-0.45
Separate B
1.14
—1.08
—1.07
2.85
—1.98
—0.93
Olivine
Glass
Ilmenite
0.26-0.31
0.58-0.66
0.32-0.29
1.14
—1.07
1.07
2.90
—1.91
0.74
* Full width at half-peak height. First stated value refers to the negative
velocity component.
t Referred to metallic Fe at 298°K.
634
CARNEGIE INSTITUTION
TABLE 46. Isomer Shifts and Quadrupole Splittings of Some Reference Standards at 298° K
Isomer Shift,*
Quadrupole
Splitting,
Crystal Site
mm/sec
mm/
sec
Sample
Fe2+
Fe3+
Fe2+
Fe3+
Synthetic fayalite
Ml, M2
1.15
2.83
Modoc olivine f
Ml, M2
1.05
2.95
Synthetic ferridiopsidej
tetra.
0.18
1.49
oct.
0.42
1.01
Natural orthopyroxenes
Ml
1.17
2.48-2.35 §
M2
1.13-1.16§
1.91-2.11 §
Ilmenite from Apollo 11 soil
oct.
1.07
0.68-0.71
* Referred to metallic Fe at room temperature.
t Virgo and Hafner (1972).
t Hafner and Huckenholz (1971).
§ Range of values across the En-Fs join (Hafner, Virgo, and Warburton, 1971a).
The most intense doublet (Fig. 114) can
be attributed principally to glass and
minor pyroxene. No additional absorp-
tion peaks that might be attributed to
minor magnetic components, such as
magnetic minerals, were detected.
The computer fit to spectrum A is not
acceptable statistically. The residual
count rates, that is, the deviations of the
fits from the data shown below the reso-
nant absorption spectra (e.g., Fig. 114),
are anomalous in the velocity regions — 1,
+ 1, and +2.5 mm/sec. Such deviations
are also evident in the spectra of glasses
(Figs. 116, 117, and 118). It is inferred
that the anomalies in the total soil spec-
trum are due principally to the non-
Lorentzian line shape of the Fe2+ absorp-
tion in the glass component of the soil
(discussed further below). In this con-
nection, it is noted that the computer fit
to the data of separate B in which the
glass has been mostly removed is almost
statistically acceptable. In fact, the line
widths of the olivine and ilmenite com-
ponents (Table 45) are close to the ex-
pected line widths. The anomalies in the
total soil spectrum are dissimilar to the
excess absorption observed in the region
around zero velocity as found in Apollo
11 and 12 soils. The excess absorption
was attributed to minor concentrations of
superparamagnetic iron (Hafner, Janik,
and Virgo, 1971 ; Herzenberg, Moler, and
Riley, 1971; Housley et al, 1971).
Separate A can be taken as representa-
tive of the average soil, and the ratio of
the integrated absorption areas cor-
responds to the approximate relative
proportions of iron in the soil constitu-
ents. Using these parameters, the ap-
l/ELDCITY
Fig. 116. The iron-57 resonant absorption of
the orange glass separated from the lunar soil
(separate C). The solid line is a least-squares
fit assuming two Lorentzians (six line variables,
one background variable). The fit deviates
from the fitted line in the regions —0.85, +0.85,
and 1.55 mm/sec. The absorption is due prin-
cipally to Fe2+.
GEOPHYSICAL LABORATORY
635
l/ELDCITY
i m / sec
Fig. H7. The iron-57 resonant absorption of
a synthetic glass, B7 (c/. caption of Fig. 116).
Chemical analysis of the glass is given in Table
47.
proximate amounts of olivine, glass, and
ilmenite are 28, 66, and 6 wt %, respec-
tively. The high proportion of glass in
the soil is consistent with the close simi-
larity of the chemical analysis of the
bulk soil and the electron microprobe
analysis of an individual glass bead
(Table 44).
Separate B shows that olivine repre-
sents approximately 41% of the total
resonant absorption area. This result is
in contrast to the minor amounts of oli-
vine grains observed in thin section.
However, the studies of Reid et al.
(1973) and Prinz, Dowty, and Keil
(1973), among others, show that the
opaque grains are chemically similar to
the orange glass but have partially re-
crystallized to ilmenite and olivine. Some
glass fragments also contain euhedral
olivine phenocrysts (e.g., Roedder, 1973).
The presence of olivine in the glass
spheres and as a crystallization product
would account for the predominance of
olivine in this fraction (Fig. 115).
The 57Fe absorption spectrum of
orange glass (separate C) at 298°K is
shown in Fig. 116. The spectrum was
fitted assuming a single Fe2+ doublet.
The residual of Fig. 116 shows significant
deviation from the data points in the
regions — 0.85, +0.85, and 1.55 mm/sec.
Furthermore, the results (Table 47) show
that the line widths are considerably
broader (*— O.6-0.7 mm/sec) than in a
crystallized silicate (e.g., typical widths
of a pure phase pyroxene are 0.27—0.30
mm/sec) , and the spectrum is further
characterized by a strong asymmetry of
the peak intensity — the intensity of the
negative velocity peak is approximately
20% greater than the positive component.
.-. •■■■>.•-
l/ELDCITY
Fig. 118. The iron-57 absorption of a syn-
thetic glass, AS-15. The solid line is a least-
squares fit assuming three Lorentzians (nine
line variables, one background variable). The
fit is almost statistically acceptable. The most
intense absorption is due to Fe3+, with a less
intense peak at 1.81 mm/sec, due to Fe2+.
Chemical analysis of the glass is given in Table
47.
636
CARNEGIE INSTITUTION
TABLE 47. Isomer Shifts, Quadrupole Splittings, and Line Widths of Orange Glass
74220,61 . and Synthetic Glasses at 298° K
Sample
Crystal Site
FWHH,*
mm /sec
Isomer Shift, f
mm /sec
Quadrupole
Splitting,
mm/sec
Separate C (74220, 61)
B7, synthetic glass §
AS- 15, synthetic glass] |
Fe2+
Fe2+
Fe2+
Fe3+
0.64J-0.78
0.57J-0.72
0.67
0.66J-0.58
1.05
1.04
0.32
1.99
1.97
1.15
* Full width at half-peak height.
t Referred to metallic Fe at 298°K.
X First-stated value refers to the negative velocity component.
§ Synthetic glass prepared at 1330°C, /o2 = 10"10-5 atm (Si02, 42.85; A1203, 8.08; Ti02, 6.33; CaO,
14.44; MgO, 12.84; FeO, 15.37 wt %). J. Akella, personal communication.
|| Synthetic glass prepared at 1500°C, atmospheric pressure (Si02, 20.20; Ti02, 18.16; Fe203, 30.02;
CaO, 31.62 wt %). H. G. Huckenholz, personal communication.
The integrated intensities differ by less
than 2%, however, and are within experi-
mental error. At 77 °K this characteristic
of the Fe2+ doublet is preserved, suggest-
ing that, the uneven heights cannot be
adequately explained by relaxation ef-
fects. The effect of the peak heights and
broad line widths seems to be character-
istic of other Fe2+ glasses and, for ex-
ample, tektites (e.g., Marzolf, Dehn, and
Salmon, 1967; Boon, 1971; Boon and
Fyfe, 1972; Evans and Leung, 1973). A
plausible explanation is that the glass
spectrum consists of a superimposition of
several Fe2+ doublets and that the com-
puter fit reported here is simply an aver-
age fit to the data points. Consequently
there will be deviations of the least-
squares fitted line from the data, as sug-
gested by the plot of the residual. In the
absence of details of the glass structure,
the data do not justify a more compli-
cated fit than that presented here. The
general features of the glass spectrum
could be explained by changes in the
quadrupole splitting (separation of com-
ponent peaks of an Fe2+ doublet) that
were significantly larger than changes in
the isomer shift (center of gravity of the
absorption relative to zero velocity).
Then the lower velocity line would tend
to be narrower than the positive velocity
peak, and the absorption in this lower
velocity range would be concomitantly
larger. In fact, the areas of the compo-
nent peaks are the same within the ex-
perimental errors.
The spectrum for the Apollo 17 glass is
closely similar to that of a synthetic glass
(B7) prepared at 1335°C and /02 -4011
(Fig. 117, Table 47) . In synthetic glasses
studied by Boon (1971) and Boon and
Fyfe (1972) , it was possible to fit an Fe2+
doublet in addition to the broad doublet
observed in 74220 and B7 glasses. This
doublet was characterized by a smaller
peak separation but similar isomer shift.
It was suggested that this absorption cor-
responded to Fe2+ in a second distinct
octahedral site in the glass. Furthermore,
Boon and Fyfe interpreted the crystal-
field spectra that showed a peak at 5000
cm"1 in terms of Fe2+ in 4-fold sites. In
this study, 57Fe absorption peaks due to
lower Fe2+ coordination and more than
one distinct octahedral site were not re-
solved in the data. A preferred interpre-
tation is that the absorption peak at
approximately 5000 cm-1 in the crystal-
field spectra can be attributed as well to
octahedrally coordinated ferrous iron.
The average quadrupole splittings and
isomer shifts for the lunar glass 74220,61
and synthetic glass B7 are A = 2.00 and
1.97 mm/sec, I.S. = 1.05 and 1.04 mm/
sec, respectively. These values are in
qualitative agreement with the range of
values at 298 °K for the octahedral sites
GEOPHYSICAL LABORATORY
637
in chain silicates (Table 46) . In particu-
lar, the distorted M2 site in pyroxenes
shows the closest comparison. On this
basis, it is reasonable to assume that
Fe2+ sites in these glasses are predomi-
nantly pseudo-octahedral.
The position of Fe3+ in these glasses
can be inferred from the spectrum of
synthetic glass (AS-15) prepared at
1500°C in air (Fig. 118, Table 47). The
spectrum shows an intense doublet with
A = 1.15 mm/sec and I.S. = 0.32 mm/
sec and a less intense peak at — 1.8 mm/
sec. It can be expected that the Fe3+
coordination and bonding (and for that
matter Fe2+) will be different for a glass
and a crystal of the same composition
(e.g., Kurkjian and Bushanan, 1964).
Nevertheless, crystal data are used in
the interpretation of the iron coordina-
tion in glasses discussed in this study.
In this connection, synthetic ferri-diop-
sides (Hafner and Huckenholz, 1971)
provide a well-established example of
ferric iron in both tetrahedral and octa-
hedral sites (Table 46). In comparison,
the synthetic glass AS-15 is characterized
by broad line widths (0.6—0.5 mm/sec
for the lower velocity peaks and — 0.7
mm/sec for the peak at 1.8 mm/sec) but
can be reasonably coded in terms of a
single Fe3+ doublet and a small contri-
bution due to Fe2+. Data from Table 46
suggest that the absorption doublet due
to ferric iron in AS-15 has its closest
analogy to iron in an octahedral en-
vironment.
Significantly, comparison of Figs. 116
and 118 shows that distinct Fe3+ peaks
are not observed in the Apollo 17 glass.
The Fe3+/Fe2+ ratio is clearly <0.1.
Optical Absorption Spectra
Crystal-field spectra of several of the
individual glass spheres were measured
and compared with a synthetic glass of
similar composition. Figures 119 and
120 show optical absorption spectra for
the lunar orange glass and its synthetic
analogue, the Fe- and Ti-rich glass. The
Wove number, 10 cm
26
20
10 4
1
II II
11 1 1 1
\ Apollo
1 1 1
74220,61
\e
200
A\ y Orange
/ \f glass
/ \ FeO 22. 18
/ \ Tio2 8.70
1 S "N \
/ \ \
/ \\
1 \
\
ib 1 1
/ ^>\
1 \
\ /
' Green \\ .~
100
\y /
glass \ \ —
1
K i
i \ i
/
FeO 17.68 \ \
la
i \ I i
/
/
Ti020.25 vx\
1
1 ix 1
/
\
1
\
/
n
ll
1 1 i
\ i ii
II II
400
500
600
Wavelength, nm
1000
1500 2000
Fig. 119. Unpolarized absorption spectra of
orange and green glasses from sample 74220,61.
(To show detail, absorptions greater than 250
cm-1 have been replotted on the indicated scale ;
i.e., d is a continuation of e, c is a continuation
of d, b is a continuation of c, a is a continuation
of b. Curves of a, b, c, and d are off the indi-
cated scale.)
Wove number, 10 cm
20 15 10
\ Synthetic Orange glass
\
\
\
/^~^r 1350 X
/ \
500
Wavelength, nm
1000
Fig. 120. Unpolarized absorption spectra of
synthetic orange glass quenched at log Po2 =
—8.1, —9.1, 1400°C. (As in Fig. 119, absorption
greater than 200 cm-1 has been replotted on
the indicated scale; i.e., / is a continuation of g,
d is a continuation of e, c is a continuation of
d, b is a continuation of c, a is a continuation
of b. Curves a-d and / are off the indicated
scale.)
638
CARNEGIE INSTITUTION
57Fe Mossbauer resonance of the orange
glasses has demonstrated that the co-
ordination sites of iron are the same in
both Ti-free and orange glass. Optical
spectra are also the same in both types of
glass. Quenched in air, the synthetic
glass showed strong Fe3+ absorption. The
strong "blue" shoulder decreases in in-
tensity as Po2 is lowered but less rapidly
than it would if it contained Fe3+ alone,
because of an intravalence charge-trans-
fer interaction between Fe2+ and Ti4+,
Ti3+. The shoulder is of minimum in-
tensity at Poo = 10"91 and then increases
as the Fe2+ bands decrease when P02 is
set below 10"11. The increase in the ab-
sorption edge simultaneously with the
decrease in the band caused by Fe2+ at
1 /xm suggests that some Fe2+ is being
reduced to Fe°. Apparently Fe° and Ti3+
are interacting and producing the ab-
sorption.
Interpretation of the Oxidation State of
Lunar Orange Glass
The absorption spectra in Fig. 120 are
for conditions of Po2 between 10"81 and
10"91. This range provides a crude
bracket for the Po2 of lunar orange glass
by comparison with the spectra shown in
Fig. 119. If this estimate is correct, the
lunar orange glass is relatively oxidized
by several orders of magnitude compared
with most samples of lunar rock (Sato
and Hickling, 1973). That is not to say
that the lunar orange glass is grossly
oxidized. On the contrary, ferric iron is
present in an amount less than 1 wt %,
the detection limit of the 57Fe Mossbauer
technique. The intriguing factor is that a
Po2 of 10"8 or 10"9, although chemically
reduced, is close to the stability field of
magnetite in the temperature range
1300°-1400°C (Muan and Osborn, 1965).
This factor may prove important in ex-
plaining a source for the moon's mag-
netization at the surface.
The orange glass sample has suffered
surface exposure to the reducing effects
of the solar wind for less than 30 m.y.
(Kirsten et al., 1973). Significantly, the
glass, which is evidently a sample of an
extensive unit on the moon, may have
preserved its oxidation (and volatile con-
stituents) since it was originally formed
3.7 billion years ago (G. J. Wasserburg,
reported orally at the Fourth Lunar Sci-
ence Conference, Houston, Texas, 1973).
Rust Alteration of the Apollo 16
Rocks*
L. A. Taylor ,t H. K. Mao, and P. M. Bell
The existence of water on the moon,
either as a free phase or as a component
bound chemically in a mineral, has been
the subject of considerable controversy
and speculation. Both hydrogen and
oxygen occur on the moon, but positive
identification of water or hydroxyl has
hitherto been lacking.
The problem of identification of water
in returned Apollo 16 rocks is less con-
troversial. On the basis of the results of
the present study, it has been established
that most returned Apollo 16 rocks con-
tain both free and combined water. Me-
tallic phases are observed with heavy
coatings of rust, and rust has worked its
way into cracks and crevices of non-
metallic mineral phases. What is again
the subject of speculation and perhaps of
future controversy is the origin of the
water. However, evidence that meteoritic
phases are unstable in the earth's at-
mosphere and are subject to rust is
abundant. The present results may pro-
vide a measure of the degree of meteorite
contamination in Apollo 16 rocks.
Rocks from almost every station vis-
ited during the Apollo 16 mission contain
rust-colored alteration. Samples 66095,11,
* This research was supported by National
Aeronautics and Space Administration grants
NGR-15-005-164 (to Taylor) and NGL-09-140-
012 (to Bell). ARPA grant DAHC-0213 pro-
vided partial support of the electron micro-
probe at Purdue University.
t Department of Geosciences, Purdue Univer-
sity.
GEOPHYSICAL LABORATORY
639
80, and 81, and 67455,48 have been ex-
amined in detail and the discussion is
directed toward selected results obtained
from this study that possibly apply to
all rusted Apollo 16 rocks.
Sample 66095 is a glassy breccia con-
sisting of about 50% plagioclase, 35%
olivine, 3% opaques, and 10-15% relict
clasts that are troctolitic. The plagioclase
contains 95 ± 2 wt % anorthite, and the
olivine composition is 80.5 ± 1 wt %
forsterite, both in the matrix and in the
relict clasts.
This rock contains a spotty, nonuni-
form distribution of native FeNi metal
grains, mainly as clusters up to 1-2 mm
across with the intervening areas contain-
ing little or no native metal. The regions
in and around the metal grains are highly
fractured. The majority of the metal
(and some troilite) has been oxidized to
various degrees and commonly contains
rims of goethite, possibly coexisting with
hematite and magnetite. The goethite
extends for considerable distances along
the numerous cracks and forms a prevail-
ing reddish brown stain. In one area,
where a late-stage eutectic Fe-FeS mix-
ture has filled all the cracks in the frac-
tured region around the metal, the FeNi
l
i
'■ill 1
_Q
*• 66095, 1 1
O
O
NT '
■4- 1.0
— <
c
t:
CD
c
(J
--~/
i Meteoritic
i Metals
Q_
•''• \
+= 0.5
/' * • •
'
\ • •»» • • • • • • , -
i ■, • *. ./
— i 1 1 1 i ■
4 8 12
Weight Per Cent Nickel
Fig. 121. Compositions of native FeNi metal
grains in 66095,11. The dashed lines represent
the boundaries separating meteoritic versus
lunar metal compositions as determined by
Goldstein and Yakowitz (1971).
metal is completely fresh, showing no
visible signs of alteration. Apparently
the area was protected from the solutions
or vapors.
Many of the rusted FeNi grains in
various Apollo 16 rocks are intergrown
with schreibersite (12.4-13.2 wt % Ni,
0.03-0.13 wt % Co) , (Fe,Ni)3P, and very
minor amounts of taenite. The textures
are virtually identical with the meteoritic
assemblages described by Goldstein and
Yakowitz (1971). In addition, a mineral
is observed in 66095 that has the proper-
ties of cohenite, Fe3C, and contains 1.1-
1.4 wt % Ni. Carbon was not verified
directly but was calculated by difference.
Identification was made on the basis of
its greater hardness and slightly yellow
color when compared with kamacite, and
by its weak anisotropism. The presence
of this phase in 66095 has been noted by
El Goresy, Ramdohr, and Medenbach
(1973).
The compositions of 83 metal grains
are shown in Fig. 121. The dashed lines
represent the boundaries for "meteoritic"
metal compositions of approximately
>4.0 wt % Ni and <1.0 wt % Co, as
given by Goldstein and Yakowitz (1971).
The criteria for meteoritic vis-a-vis lunar
origin for iron-nickel cobalt alloys are
yet to be firmly documented, but it is
notable that, without exception, native
FeNi metal compositions plot well within
the limits. This is also true for metal
grains in several other rusty rocks (Pa-
dovani and Carter, 1973). However,
another rock, 67455, which was recently
examined, contains a similar assemblage
of goethite and many oxidized metal
grains but has unusual metal composi-
tions (Fig. 122). The compositions plot
as a relatively tight cluster compared
with most metal compositions from lunar
igneous rocks, where a large spread in
metal ratios is usually found. The 67455
metals may well constitute another mete-
oritic source, for they have compositions
similar to certain howardites and eucrites.
Certain samples of 66095 (e.g., 80)
640
CARNEGIE INSTITUTION
1
1
1 1
2.0
-
-t—
o
A
67455,48
O
O
CD
O
%_
CD
1.5
1.0
<
LL
r
o>
0.5
1
i
1 1
Weight Per Cent Nickel
Fig. 122. Compositions of native FeNi metal
grains in 67455,48. See Fig. 121 for explanation.
also contain several sphalerite grains,
(Zn,Fe)S (El Goresy, Ramdohr, and
Medenbach, 1973). The numerous sphal-
erite grains that were analyzed in the
present study contain 28 ± 2 mole %
FeS regardless of the degree of advanced
oxidation throughout the specimen (80
versus 81 or 11) and therefore are prob-
ably not related to the oxidation. Sphal-
erite is not entirely uncommon in mete-
orites (Ramdohr, 1973), and these com-
positions, although high in FeS by ter-
restrial standards, have been reported in
certain meteoritic sphalerites (A. El
Goresy, personal communication, 1970) .
On the basis of the mineral assemblage
of kamacite, taenite, schreibersite, co-
henite, and sphalerite, and of composi-
tions of the FeNi metal grains (Figs.
121 and 122) , it would appear that these
goethite-bearing rocks may contain an
appreciable meteoritic contribution.
An attempt was made to correlate vis-
ually the "degree" of oxidation of the
metal grains with their composition. The
scale used ranges from 0 for completely
fresh to 7 for extremely oxidized metal —
i.e., oxidation produced extensively on all
sides of the metal grain. Figure 123
shows that there does not appear to be
such a correlation for the sample as a
whole. However, within most of the in-
dividual clusters of FeNi grains there is
usually such a correlation, as also shown
in Fig. 123. The metals with the lowest
Ni contents have undergone the most oxi-
dation and vice versa, indicating that the
oxidation was locally isolated rather than
pervasive.
Figure 123 is a plot of all the metal
compositions and therefore tends to ob-
scure a relationship between degree of
oxidation and the Ni content of an indi-
vidual in a cluster. The amount of oxida-
tion is usually less on the side of a metal
grain where troilite is present (Fig. 124) ,
a finding compatible with those of Wil-
liams and Gibson (1972) (also R. J. Wil-
liams, personal communication, 1972).
The presence of schreibersite does not ap-
pear to affect oxidation.
An important observation in the rusty
rocks is the occurrence of 1.5—4.6 wt %
chlorine in goethite. In addition, the
cracks and regions associated with the
oxidation are chlorine-rich. Figure 125
shows C1K« scans for FeNi grains, which
became highly altered after polishing.
Notice that chlorine is intimately associ-
ated with the metal and is present in the
numerous cracks as well.
These observations strongly suggest the
"Degree" of Oxidation
Fig. 123. The nickel content of native FeNi
metal grains versus the "degree" of oxidation of
the metal. 0, completely fresh; 7, totally oxi-
dized. Solid diamonds represent metals within
the same cluster. See text for further explana-
tion.
GEOPHYSICAL LABORATORY
641
Fig. 124. Selected photomicrographs of "rust" in 66095. The same portion of the sample in
reflected (a) and transmitted (b) light, respectively; likewise for c and d. Notice the gray goe-
thite, which appears ruby-red to reddish brown in transmitted light. Native FeNi is white; tro.
tan; silicates, dark gray.
former existence of the meteoritic mineral
lawrencite, FeCl2, a phase known to
rapidly deliquesce and oxidize upon en-
tering the earth's atmosphere. The proc-
ess of oxidation of lawrencite in humid
air includes the formation of ferric
chloride, ferrous hydroxide, and a highly
oxidizing acidic solution, as the original
lawrencite is finally consumed. The proc-
ess is extremely rapid ; a polished section
begins to alter within seconds in the
earth's atmosphere.
642
CARNEGIE INSTITUTION
67455
b
Fig. 125. Electron microprobe x-ray scans of two highly oxidized native FeNi metal assem-
blages in rock 67455. The two separate parts of the same thin section are labeled a and b.
GEOPHYSICAL LABORATORY
643
Cracks associated with the oxidation
are reported to contain an Fe,Zn chloride
and an Fe,Zn sulfate (El Goresy, Ram-
dohr, and Medenbach, 1973) . It was ini-
tially thought that these salts may have
been the products of fumarolic activity
and that a reaction between the Fe,Zn
chloride salt and the troilite (FeS) pro-
duced the sphalerite.
The sphalerite grains are uniform in
composition throughout the sections ex-
amined. If the hypothesis of fumarolic
origin (El Goresy, Ramdohr, and Meden-
bach, 1973) is correct, it would be an
unusual coincidence if the Fe,Zn chloride
had reacted with the troilite to the same
degree for each grain. This possibility is
unlikely in that the metal clusters have
reacted at isolated areas, excluding an
all-pervasive oxidation mechanism. Fur-
thermore, much of the troilite does not
contain sphalerite, even in areas of ex-
treme oxidation, and does not show
textural evidence for such a reaction.
Commonly sphalerite grains appear en-
tirely enclosed in troilite. It is more
likely that HC1 solutions produced by
the hydration and oxidation of lawrencite
have reacted with sphalerite, adding Zn,
Fe, and sulfur (possibly in the form of
sulfate) to the solution. Precipitation of
iron and zinc chlorides and sulfate salts
from these solutions is probably the cause
of their existence in Apollo 16 rocks.
Grieve and Plant (1973) stated that
the formation of goethite was accom-
plished on the moon since it is present
within a glass-coated rock sample
(64455). Moisture still may have been
accessible, and it is readily observed that
samples 66095 and 67455 are hydrating
and oxidizing in the laboratory today.
Dr. I. Friedman (personal communica-
tion, April 1973) performed deuterium
analysis on approximately 15 g of 66095
and found that the water extracted is
similar in all respects to terrestrial water.
All evidence available at this time sug-
gests that oxidized and hydrated phases
in the Apollo 16 rusty rocks may be the
direct result of meteoritic and postlunar
contamination, although a lunar origin
of the oxidation is not entirely disproved.
An Analytical Study of Iron in
Plagioclase from Apollo 16 Soils
64501, 64502, and 64802; Apollo 16
Rock 66095; and Apollo 15 Rock 15475
P. M. Bell and H. K. Mao *
Plagioclase crystals were selected from
three Apollo 16 soils (64501, 64502,
64802), Apollo 16 rock 66095, and Apollo
15 rock 15475 to search for ferric iron and
to study the chemical zoning of ferrous
iron. The Apollo 16 samples are known
to contain oxidation alteration (Taylor,
Mao, and Bell, this Report) . If iron in
Apollo 16 plagioclase is not oxidized, this
would tend to confirm the view that the
alteration is surficial and terrestrial in
origin. Furthermore, unsystematic zon-
ing of ferrous iron would suggest that the
Apollo 16 rocks were chemically reduced
before they were collected on the lunar
surface (Mao and Bell, this Report) . To
test these suggestions, polarized crystal-
field spectra were measured on the indi-
vidual crystals, which were then analyzed
by electron microprobe.
The crystal-field spectra of all plagio-
clases in this study are similar in form to
the Luna 20 plagioclases (shown in Fig.
141, Mao and Bell, this Report). The
Fe3+ band of the polarized plagioclase
spectrum (shown in Fig. 94, Bell and
Mao, this Report) is not present, suggest-
ing that the crystals are relatively poor
(<0.01 wt % of the total) in ferric iron.
However, there is a linear relationship
between the average iron content and the
absorption coefficient of a strong polar-
ized crystal-field band at 1250 nra in the
Luna 20 plagioclase that was sampled
from a soil not reportedly oxidized or
hydrated (Mao and Bell, this Report).
The iron values of plagioclase crystals
from the Apollo 15 and the altered Apollo
* Work supported in part by National Aero-
nautics and Space Administration grant NGL
09-140-012.
644
CARNEGIE INSTITUTION
16 samples fall off this line, a factor that
suggests the presence of ferric iron be-
cause the relationship assumes all iron is
present as Fe2+ (Fig. 126).
Similarities between Apollo 16 soils
and the Russian Luna 20 samples make it
expedient to use a common base for cor-
relation between crystal-field effects and
electron microprobe analyses of iron in
constituent plagioclase (Bell and Mao,
this Report). All plagioclases studied in
these Apollo 15 and 16 samples are anor-
thite and contain variations in iron con-
centration that are not systemically lo-
calized to a "core" or "rim." Iron values
range from a few hundredths of a weight
per cent to several tenths of a weight
per cent in various parts of a single
plagioclase crystal, in much the same
way as in the Luna 20 plagioclase. The
metallic iron alloy crystals oriented crys-
tallographically in several of the Luna
20 plagioclase crystals are not found
in the Apollo 15 and 16 plagioclases, but
otherwise there is no apparent difference
in the iron content (oriented inclusions of
an unidentified silica-rich phase are noted
in several plagioclase crystals of soil
64501, but no metallic phases other than
globular masses are identified) . Table 48
gives the major-element analyses ob-
tained by electron microprobe (calcu-
lated as oxides), indicating the actual
ranges in iron concentration.
The implications of this study are that
the iron in the plagioclases is reduced like
that in the Luna 20 plagioclases. The
nonsystematic zoning of iron described
above has been attributed to reduction
during impact. There is a suggestion that
Fe3+ in these crystals may be related to
alteration, but the amount present is
small. These data support the hypothesis
that most or all of the alteration of the
Apollo 16 samples did not occur on the
moon.
20
15 -
E
o
e"
c
O
LO
<\J
O
a
o -
10 -
P- 5 —
<
-
1 1 < 1 1 1 1 1
1 1 y 1
-
-
-
-
j^ , \^rc\\ Oji
_
f* I IDOUI . j4
•h 15475.22
—
14163.33 /
_
—
L22003.l.2c/
—
1 W 1
-
_
—
/ 66095.51.2
L20AI0/ •
i *
—
s^ '
y^66095.5l.l
L20A5/ *6450l.l
■*X-2fi '64502.1
_
/ #64802.I3 0Lake Co.
- J
y*^ Plagioclase
'l_22003.l.2a
1 1 1 1 1 1 1 1
1 1 .
-
0.2
0.3
0.4
0.5
0.6
Total Fe , weight percent
Fig. 126. Absorption coefficient of the a-polarized absorption band in plagioclase at 1250 nm
(plotted in cm"1). Included are plagioclases from Luna 20, Apollo 14, 15, and 16, and a terrestrial
plagioclase from Lake County, Oregon. Solid dots are average compositions; brackets show the
range of composition within a single crystal. Points falling off the Luna 20 line probably contain
some iron as Fe3+, which does not contribute to the absorption coefficient of the 1250-nm band.
GEOPHYSICAL LABORATORY
645
TABLE 48. Average Electron Microprobe Analyses of Selected Plagioclase Crystals, wt %
Sample No.:
15475,22
64501,1
64502,1
64802,13
66095,51,1
66095,51,2
Cr203
0.00
0.00
0.01
0.00
0.00
0.00
MnO
0.00
0.00
0.01
0.00
0.00
0.00
FeO
0.66
0.18
0.20
0.16
0.20
0.33
(range)
0.49-0.89
0.16-O.21
0.13-0.24
0.10-0.20
0.15-0.28
0.27-0.39
NiO
0.00
0.00
0.01
0.00
0.00
0.00
Na20
1.21
0.36
0.38
0.32
0.33
0.27
MgO
0.29
0.05
0.11
0.07
0.07
0.11
A1203
31.76
35.12
34.80
34.98
35.57
35.17
Si02
47.91
45.73
45.60
44.61
45.81
45.33
K20
0.02
0.00
0.00
0.00
0.00
0.00
CaO
16.85
18.41
18.82
19.20
18.82
19.18
Ti02
0.05
0.01
0.05
0.02
0.02
0.02
Totals
98.74
99.86
100.04
99.36
100.84
100.42
Petrology of Apollo 16 Lunar
Highland Rocks
F. N. Hodges and I. Kushiro
A detailed study including petrogra-
phy, electron microprobe analysis, and
high-pressure melting experiments has
been made on four Apollo 16 samples —
60025, 60315, 68416, and 62295. The re-
sults of the study and discussions were
presented by Hodges and Kushiro
(1973).
In brief, cataclastic anorthosite 60025,
consisting predominantly of plagioclase
(An98-An94) with small amounts of
chromite-rich spinel (Chr6iSp + Her27
Ulvoio mole %), augite (Ca45Mg37Fei8) ,
and two distinct hypersthenes (Ca3Mg6o
Fe37 and Ca3Mg52Fe45) , probably formed
by accumulation of calcic plagioclase;
small amounts of trapped interstitial
liquid crystallized to plagioclase, augite,
hypersthene, and pigeonite. Subsolidus
reequilibration prior to brecciation re-
sulted in the breakdown of pigeonite to
augite and relatively iron-rich hyper-
sthene. Recrystallized breccia 60315,
characterized by large poikilitic grains
(up to 2 mm) of orthopyroxene (e.g.,
Ca4Mg8iFei5), probably formed near the
base of a large impact ejecta blanket.
The initial matrix material of the brec-
cia, possibly derived in part from con-
siderable depths, must have contained a
small amount of interstitial melt. Crys-
tallization of orthopyroxene poikiloblast
concentrated the melt, into which K, Ti,
Fe (relative to Mg) , and minor elements
were fractionated. Zones of greater melt
concentration crystallized with a well-
developed diabasic texture. The compo-
sitions of coexisting pyroxenes suggest a
crystallization temperature near 1100°C.
Fine-grained, intersertal, feldspathic
basalt 68416, consisting of plagioclase
(An98-An85) , highly zoned pyroxenes —
orthopyroxene (Ca5Mg82Fe13-Ca6Mg70
Fe24), pigeonite (e.g., CaioMg60Fe30),
subcalcic augite of varying Ca content,
and augite (Ca4oMg46Fei4-Ca32Mg38Fe3o)
— and olivine (Fo69-Fo6i) , was most
likely formed by rapid crystallization,
probably of an impact-melted plagioclase
cumulate.
Spinel troctolite, consisting predomi-
nantly of olivine (Fo94-Fo89), plagioclase
(An95-An9i), and spinel (Chr 2-4 mole
%), is one of the most magnesian lunar
samples returned and was probably
formed by rapid crystallization of a
basic magma that may have been derived
by partial melting of a plagioclase- and
spinel-bearing peridotite under anhy-
drous conditions. The presence of
xenocrystic calcic plagioclase (An98-
An94) and chromian spinel (Chr 9-16
mole %) suggests contamination by pre-
existing anorthositic material.
646
CARNEGIE INSTITUTION
Magmas similar in composition to
62295, or one nearer the low-pressure
three-phase boundary olivine-spinel-
plagioclase, may be parental to a wide
variety of highland rocks, including
anorthosite, feldspathic basalt, and gab-
broic anorthosite. The presence of a wide
compositional gap between highland-
type rocks and mare basalts, the signifi-
cant difference in Fe/Mg ratios of
liquidus olivine and pyroxene in the two
rock types over a wide pressure range,
and the general paucity of the pyroxene
component in highland rocks indicate
that highland-type rocks and mare ba-
salts were derived from distinctly dif-
ferent source materials.
Liquidus Phase Relations of Apollo 15
Mare Basalt 15016 *
F. N. Hodges and I. Kushiro
Melting relations of rock 15016, a
coarse-grained, vesicular basalt, have
been determined at pressures from 4 to
* Supported by National Aeronautics and
Space Administration grant NGR 09-140-017.
16 kbar for comparison with those of
highland basalts. The rock consists of
zoned olivine (Fo72-Fo7), pigeonite-ferro-
pigeonite (Ca7Mg6oFe33-CaiiMg37Fe52) ,
subcalcic augite-subcalcic ferroaugite
(e.g., Caor)Mg49Fe26-Cai9Mgi4Fe67), hed-
enbergitic clinopyroxene (e.g., Ca39Mg3
Fe58) , plagioclase (An93.9Abe.oOro.1-
An8o.7Abi5.7Or3.6) , ilmenite, chromite
solid solution, chromian ulvospinel, cris-
tobalite, and other minor constituents.
Mafic minerals constitute more than 70
vol % of the sample.
Experiments were conducted in a solid-
media, high-pressure apparatus with
graphite capsules, using techniques simi-
lar to those applied to rock 62295
(Hodges and Kushiro, 1973). The results
are shown in Fig. 127. Olivine (Fo70-
Fo72 at 6 kbar) is the liquidus or near-
liquidus phase up to approximately 11
kbar. The second phase to crystallize is
spinel solid solution (e.g., Sp24Chr67 Ulvo9
mole %), followed by pigeonitic clino-
pyroxene (Ca7Mg67Fe26 at 1240°C, 6
kbar) . Clinopyroxene may be on the
liquidus above 15 kbar.
1300
0
o
(D
i_
&1200
E
CD
1100
Cpx+L
Ol+L
^=#^l
m Sp+ Cpx + L
Px + Sp + Ol
0l+Cpx+P|+5p
+ L ^
Ol+Cpx+P[ + Sp /
0 2 4 6 8 10 12 14 16
Pressure, kb
Fig. 127. Results of melting experiments on Apollo 15 mare basalt 15016.
GEOPHYSICAL LABORATORY
647
The phase relations are similar to those
of rock 62295; however, olivine coexist-
ing with 15016 melt is much more Fe-rich
(Fo72 compared with Fo93) and spinel is
much richer in both Fe and Cr. If magma
of 15016 composition was formed by
partial or complete melting in the lunar
interior and not as a result of olivine
accumulation as proposed by Humphries,
Biggar, and O'Hara (1972), the source
material must have contained olivine
considerably more iron-rich than the
source material of highland crystalline
rocks.
Crystallization of Pyroxenes in
Apollo 15 Mare Basalts*
/. Kushiro
Apollo 15 basalts 15016, 15476, and
15545 — three different mare-type basalts
— have been studied petrographically and
analyzed with the electron microprobe.
Rock 15016 is a coarse-grained vesicular
basalt consisting of zoned olivine (Fo73-
Fo7), pyroxenes (described below),
plagioclase (An94.oAb6.oOro.o-An8o.7Abi5.7
Or36), ilmenite, chromite solid solution,
chromian ulvospinel, cristobalite, and
minor constituent minerals. Rock 15476
is a pigeonite porphyry consisting of
large prismatic phenocrysts of pigeonite
('—/2 X 10 mm) with rims of subcalcic
augite set in a variolitic groundmass.
Plagioclase An93.oAb7.o-An87.9Abi2Oro.i
occurs in the groundmass and as inclu-
sions in the rims of pyroxene phenocryst.
Rock 15545 is a coarse-grained olivine
basalt containing olivine (Fo56-Fo7),
pyroxene (described below) , and plagio-
clase (An9i.oAb8.5Oro.5-An79.oAbi8.oOr3.o)
The results of the analyses indicate that
large phenocrysts of pigeonite with rims
of subcalcic augite in rock 15476 started
crystallization under lunar subsurface
conditions ; subsequently the magma was
extruded on the lunar surface, followed
by rapid crystallization of iron-rich
* Supported by National Aeronautics and
Space Administration grant NGR 09-140-017.
pyroxenes and plagioclase under condi-
tions of very low oxygen fugacity. It was
found that pyroxenes with Ti/Al ratios
greater than 0.5 also have Fe/(Mg -f
Fe) ratios greater than 0.5 in these three
basalts.
The electron microprobe analyses of
pyroxenes in these rocks are shown in
Fig. 128. The large pigeonite phenocrysts
in rock 15476 (open circles) show a
smooth compositional variation from
low-Ca core (Ca5Mg69Fe26) to relatively
high-Ca rim (Cai2Mg58Fe30) with a slight
increase 01 Fe/Mg (A -» B in Fig. 128) .
The pigeonite is overgrown discontinu-
ously by subcalcic augite (Fig. 129).
The compositional gap between pigeonite
(B) and subcalcic augite (C) is about
20 atomic % Ca. In the subcalcic augite
Ca increases slightly (C -* D) and then
decreases with increase of Fe (D -> E) .
The subcalcic augite is rimmed by fer-
ropigeonite and subcalcic ferroaugite with
a discontinuity (E -» F). These composi-
tional variations are very similar to those
observed by Boyd and Smith (1971)
along the direction normal to (110) of a
pyroxene phenocryst in Apollo 12 basalt
12021. Subcalcic augite, as well as core
pigeonite in rock 15476, is mostly free of
plagioclase inclusions or intergrowths,
but ferropigeonite and subcalcic fer-
roaugite are intergrown with thin an-
hedral plagioclase crystals (Fig. 129) ,
suggesting that ferropigeonite or sub-
calcic ferroaugite crystallized rapidly
with plagioclase. The groundmass py-
roxene is subcalcic ferroaugite, which
changes its composition toward the
CaFeSioO(rFeSi03 join.
Pyroxenes in rocks 15016 and 15545
appear to have compositional ranges sim-
ilar to those in rock 15476; because of
the lack of large euhedral crystals in
these specimens, however, the composi-
tional trends are not well demonstrated.
Pigeonites in these rocks are more Ca-
and Fe-rich than the pigeonite core in
pyroxene phenocrysts in rock 15476. The
late-stage pyroxenes in rock 15016 show
648
CARNEGIE INSTITUTION
o 15476
• 15016
+ 15545
1425
50
Atomic per cent
Fe
Fig. 128. Ca-Mg-Fe plot of the electron microprobe analyses of pyroxenes from Apollo 15 mare
basalts 15476, 15016, and 15545. Solid lines show continuous zonings in a single crystal, and dashed
lines show discontinuous zonings. Trend from A to E is observed along one direction of a single
composite grain, and that from A to iron-rich pigeonite is observed along a different direction of
the same grain. Open circles plotted off the above trend lines are analyses of different pyroxene
crystals in rock 15476. Solid square shows composition of pigeonite formed at 1200 °C under dry,
low-Po2 (10~13 atm) run at 1 atm (Kushiro et al., 1971). Compositional trend of pigeonite in
Apollo 12 basalt 12065 is from Kushiro et al. (1971), and compositional range of iron-free pigeon-
ite is from Kushiro (1972c).
Fig. 129. Photomicrograph of a zoned pigeonite crystal with rims of subcalcic augite and ferro-
augite in pigeonite porphyry 15476. A, augite; P, pigeonite; PI, plagioclase.
GEOPHYSICAL LABORATORY
649
diverse trends, some trending toward fer-
rosilite with crystallization of pyroxfer-
roite and others trending toward heden-
bergite. The relatively iron-rich pyrox-
enes in these samples, including ground-
mass pyroxene in rock 15476, are proba-
bly the result of nonequilibrium, rapid
crystallization.
The Al-Ti relations of the analyzed
clinopyroxenes are shown in Fig. 130. The
large pyroxene phenocryst in rock 15476
shows a linear increase of Al and Ti with
a Ti/Al ratio of 1/6, from a low-Ca
pigeonite core to subcalcic augite. At the
core boundary there is a sharp break in
Al, corresponding to the sharp break in
Ca, shown by point D in Fig. 128. This
break is probably due to the crystalliza-
tion of plagioclase, as mentioned by
Bence and Papike (1972). The rim fer-
ropigeonite and subcalcic ferroaugite as
well as groundmass pyroxenes are more
depleted in Al.
The groundmass pyroxenes in rock
15476 and iron-rich pyroxenes in the
other two rocks plot above the line Ti/Al
= %. In the Ca-Mg-Fe diagram these
"excess-Ti" (Ti/Al > y2) pyroxenes plot
on the iron-rich side of the line Fe/(Mg
+ Fe) - 0.5 (Fig. 128) . Based on the
assumption that Ti4+ does not substitute
for Si, it is suggested that pyroxenes hav-
ing a Ti/Al ratio greater than % would
contain Ti3+ because if Ti is all Ti4+ the
charge balance cannot be maintained be-
tween tetrahedral and octahedral sites
(Boyd and Smith, 1971 ; Kushiro, Ikeda,
and Nakamura, 1972; Bence and Papike,
1972). Therefore, if the analyses are
accurate and the above assumption is
correct, groundmass pyroxenes in rock
15476 and iron-rich pyroxenes in rocks
15016 and 15545 must have crystallized
under very reducing conditions. As men-
tioned above, rock 15476 contains large
phenocrysts of pigeonite with rims of
subcalcic augite set in a variolitic ground-
mass, a texture indicative of rapid crys-
tallization. It is likely that pigeonite
started its crystallization beneath the
lunar surface and grew relatively slowly
from the melt. A portion of the subcalcic
augite would have formed under the same
conditions. Before completion of the re-
action, the magma was extruded on to the
lunar surface, where it crystallized under
very reducing conditions. Plagioclase
crystallized rapidly, and iron-rich py-
roxenes crystallized metastably. The cool-
ing history of this rock, including sub-
0.05-
h-
0.05
0.10
0.15
Al (0 = 6)
Fig. 130. Ti-Al relations of the analyzed pyroxenes from rocks 15476, 15016, and 15545. Symbols
as in Fig. 128.
650
CARNEGIE INSTITUTION
solidus cooling, is discussed by Virgo
(this Report) on the basis of Mg-Fe
ordering and structure of pigeonite.
The pigeonite cores in rock 15476 have
a wide compositional range (7 atomic %
Ca). Phenocrystic pigeonite crystals in
Apollo 12 basalt (pigeonite porphyry)
12065 have a similar compositional range
(6 atomic % Ca) (Kushiro et al, 1971),
as shown in Fig. 128. It is noted that the
compositional range of iron-free pi-
geonite on the join CaMgSi206-MgSi03
(Fig. 128) is 6 atomic % Ca along the
solidus (1425°-1385°C) (Kushiro,
1972c) . It may be possible that the field
of pigeonite at the solidus extends with
a width of 6-7 atomic % Ca from the
iron-free join CaMgSi206-MgSi03 to
the relatively iron-rich portion of the
system CaMgSi206-MgSi03-FeSi03-
CaFeSi206 to include the composi-
tions of natural lunar and terrestrial
pigeonites. In most terrestrial basic rocks,
however, the Ca content of pigeonite
is much more limited than that shown
in Fig. 128. This limited Ca content
is probably due to the lower crystalliza-
tion temperatures of pigeonite in most
terrestrial basic magmas than in the
lunar mare basaltic magmas. To examine
this possibility it is necessary to deter-
mine the compositional field of pigeonite
in the system MgSi03-CaMgSi206-
CaFeSi206-FeSi03 over a range of
temperatures.
Crystallization and Sttbsolidus
Cooling History of Apollo 15
Basalts 15076 and 15476
D. Virgo
Exsolution phenomenon, domain size,
and Fe2+-Mg exchange between the Ml
and M2 crystal sites in P2x/c clinopyrox-
ene provide criteria for determining the
subsolidus cooling history of their host
rock. Coarse-grained pigeonite crystals
from lunar rocks provide ideal samples
for experimental studies pertaining to
these solid state processes. In this report
the results of x-ray diffraction and the
57Fe Mossbauer measurements on pi-
geonite crystals from two Apollo 15
basalts, 15076 and 15476, are discussed.
Description of Basalts 15476 and 15076
Rock sample 15476 consists of large
euhedral pigeonite phenocrysts (up to 15
mm in length) set in a distinctly fine-
grained variolitic groundmass, consisting
of plagioclase, ilmenite, and clinopyrox-
ene. The pigeonite core is Wo5En69Fs26
to Wo9En61Fs30, but the rims are sharply
zoned to subcalcic ferroaugite or zoned
to ferropigeonite and subcalcic ferro-
augite. Kushiro (this Report) argues for
a rapid quenching of a melt containing
pigeonite phenocrysts. Sample 15076, on
the other hand, contains zoned pigeonite
crystals (core is Wo6En6GFs28) set in a
medium-grained subophitic groundmass.
This texture suggests a relatively slower
cooling rate during the quenching inter-
val. Experimental studies (Humphries,
Biggar, and O'Hara, 1972) strongly sug-
gest that the different samples do not
belong to a single lava flow.
Chemical and Structural Characterization
of 15076 and 15476 Pigeonites
Electron microprobe analyses of pol-
ished thin section 15476,32 are discussed
by Kushiro (19726 and this Report).
Selected crystals were also analyzed from
polished thin section 15076,69. Analyses
for individual grains are plotted in Fig.
131. The corresponding Ti-Al relation-
ships are plotted in Fig. 132. Crystals G4,
G6, and G5 have small pigeonite cores
(^^Wo5En68Fs27) but are heterogene-
ously zoned to subcalcic ferroaugite com-
position (— ^Wo24Eni0Fs66, G5) or zoned
to subcalcic augite (^Wo32En45Fs23)
and rapidly zoned over narrow rims to
subcalcic ferroaugite composition
(— Wo22En7Fs71, G4). One crystal, G6,
shows oscillatory zoning in the subcalcic
augite field similar to that reported for
12021 pigeonite (Boyd and Smith, 1971).
GEOPHYSICAL LABORATORY
651
50 Hd
Fig. 131. Ca-Mg-Fe plot in mole per cent of selected analyzed pyroxene grains from polished
thin section 15076,69. Symbols: grain 6 (solid circles), nos. 1-12 correspond to a core-rim tra-
verse; grain 4 (crosses), nos. 1-8 correspond to a core-rim traverse; grain 5 (solid squares), nos.
3-8 correspond to a core-rim traverse; grain 1 (open triangles), nos. 1-9 correspond to a rim-rim
traverse; grain 2 (solid triangles), nos. 1-7 correspond to a rim-rim traverse. Lines connect spot
analyses in each grain.
Crystals Gl and G2 have subcalcic augite
cores that show rapid zoning to subcalcic
ferroaugite compositions. The range of
chemical composition from the core to
the rim of the pigeonite phenocrysts is
best interpreted as a rapid crystal-growth
phenomenon such that the equilibrium
phase relationships are obscured (Virgo,
this Report). The Ti-Al relationships
plotted in Fig. 132 show that the cores
of both the pigeonite crystals and sub-
calcic augite have similar Ti-Al ratios
of -~/l:6. Beyond this stage of crystalli-
zation the Ti-Al ratio rapidly changes
and approaches the 1:2 line. This abrupt
change is attributed to the beginning of
CD
0.05
0.04 -
0.03
0.02-
0.01
, e a a
I 8 ■ I >
A Vll A2 ^ 8 J» ''
12 ^ 7
Ti:AI = l:2
Ti: Al = |;6
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.1 1
Al(0=6)
Fig. 132. Ti-Al relations of selected analyzed pyroxene grains from polished thin section 15076,69.
Symbols as in Fig. 131. Values are based on six oxygens.
652
CARNEGIE INSTITUTION
plagioclase crystallization (Bence and
Papike, 1972). The outer rims of both
distinct pyroxene types are characterized
by Ti-Al ratios > 1:2, suggesting the
occurrence of Ti3+ in the pyroxene crystal
structure. This stage corresponds to crys-
tallization of pyroxene more iron-rich
than compositions along the join
En42Fs58-Wo35En29Fso(5. In summary
then, the crystallization sequence in this
rock can be depicted as beginning crys-
tallization of pigeonite and subcalcic
augite, followed by plagioclase crystalli-
zation. The last stage of crystallization,
corresponding to the outer rims of the
pyroxene crystals, was presumably very
rapid and occurred under significantly
lower reducing conditions. These results
are somewhat similar to those described
for pyroxene crystallization in 15476,32.
Electron microprobe analyses of the
separated pigeonite fractions are given in
Table 49 in terms of end-member com-
positions in mole per cent. These values
are in agreement with the core composi-
tions reported here from the data on the
polished thin section and those of
Kushiro (this Report) .
Selected crystals of both pigeonite frac-
tions were examined by x-ray diffraction,
using the precession technique (Ohashi
and Finger, 1973). Sample 15476 shows
no visible exsolved augite spots along the
expected planes (001) and (100) after
exposures of up to 90 hours. Further-
more, the b reflections {h + k is odd) are
significantly more diffuse than the a
reflections (h -4- k is even). In contrast,
pigeonite crystals in sample 15076 show
diffuse streaking that originates from
pigeonite reflections along both a* and c*
and points toward the expected positions
of the augite reflections. Jagodzinski
and Korekawa (1973) interpreted such
streaking in the same sample as due to
thin exsolution lamellae with average
thickness of not more than one or two
unit cells. This fine-scale exsolution is
reminiscent of the electron micrographs
of Champness and Lorimer (1971), who
attributed such structures to spinoidal
decomposition. Visual examination indi-
cates that the diffuseness of the class-6
reflections in comparison to the a reflec-
tions is similar to that for crystals of
sample 15476.
Site Occupancy of Fe2+-Mg in the Ml
and M2 Pyroxene Sites Determined by
the 57Fe Mossbauer Technique
The Fe2+-Mg distribution between the
Ml and M2 crystal sites was determined
by 57Fe Mossbauer analysis at 77°K
using techniques previously reported
( Year Book 71, p. 607) . The temperature
dependence of the Fe2+-Mg exchange was
determined by heating experiments on
the natural crystals. These experiments
were carried out in evacuated quartz
tubes under controlled temperatures in
a calibrated platinum furnace. The
heated samples were rapidly quenched in
liquid nitrogen.
The least-squares fits of the pigeonite
spectra (assuming four Lorentzian lines
with no constraints) were considered as
statistically acceptable. Typical x2 values
of the natural and heated samples were
'w1.5 units per channel. Figures 133 and
134 show the spectra of the natural sam-
ples 15076 and 15476, respectively. The
spectral data from the least-squares fit-
TABLE 49. Chemical Composition
of Pigeonite Separates
Average Composition
Wo En Fs
Standard Deviations
Number
of Spot
Analyses*
Pigeonite
Wo En Fs
15076
15476
5.8 65.8 27.8
6.5 65.5 28.0
0.6 1.5 0.9
0.3 1.7 0.3
13
10
Na, Mg, Si, Ca, Ti, Cr, Mn, Fe determined.
GEOPHYSICAL LABORATORY
653
15D76 CPXNRTURRL
-4.0
-2.0 .0
1/ELDCITY
2.0
i/se:
4.0
Fig. 133. The 57Fe spectrum of pigeonite sep-
arate from basalt 15076 at 77 °K. Outer peaks,
Fe2+ doublet at Ml; inner peaks, Fe2+ doublet
at M2. The solid line is a least-squares fit of
Lorentzian lines to the uncorrected analyzer
counts. The deviations of the solid line from
the data (divided by the square root of the
background) are plotted as "residual below the
spectra."
ting are given in Table 50. The nuclear
quadrupole splitting and isomer shifts
at the Ml and M2 sites are given in
Table 51. These hyperfine parameters are
in agreement with previously reported
values (Hafner, Virgo, and Warburton,
19716; Schiirmann and Hafner, 1972).
No changes in splittings and shifts were
observed after the heat treatments. There
is no evidence of Fe3+.
Data on site occupancies and calcu-
lated distribution coefficients are given in
Table 52. It should be noted that the
height ratios were taken as a measure
of the area ratios of the doublets, since
statistical errors of the heights in the
least-squares fits were significantly
smaller than those of the apparent widths.
The site occupancies were determined as-
suming Ca2+ was exclusively located at
the M2 sites and th? average "ferrosilite"
content of the crystals was distributed
between the Ml and M2 sites according
to the respective Fe2+ distribution num-
bers. The presence of octahedral Al, Ti,
and Cr in the M sites was ignored in this
analysis since the site preference of these
elements is not definitely known.
The k value determined for 15476 pi-
geonite is 0.082, which is in agreement
with the value of 0.072 determined from
the x-ray refinement by Ohashi and
Finger (1973), considering the slightly
different bulk chemical compositions
used by them.
The exchange energies determined from
the heating experiments in sample 15076
at 750° and 920°C are 3.95 and 4.01
kcal/mole, respectively. These values are
15^76 CPX.NRTURRL
■2.0 .0
l/ELDCITY
Fig. 134. The 57Fe spectrum in pigeonite
separate from basalt 15476 at 77 °K (see caption
of Fig. 133).
654
CARNEGIE INSTITUTION
TABLE 50. Clinopyroxenes from Apollo 15: Peak Heights and Widths at 77 °K
Pigeonite
Ai'
B^
I FWHH,f
mm/sec
FWHH,f
mm/sec
FWHH,t
mm/sec
I FWHH,t
mm/sec
15076, natural
15076, 750°C, 2 days
15076, 920°C, 1 day
15476, natural
0.0132 0.301
0.0191 0.303
0.0309 0.299
0.0080 0.295
0.0736 0.313
0.0765 0.309
0.0984 0.307
0.0481 0.342
0.0644 0.346
0.0654 0.342
0.0858 0.344
0.0383 0.387
0.0147 0.264
0.0198 0.293
0.0325 0.288
0.0082 0.248
* Ai and A2 refer to the Ml doublet; Biand B2, to the M2 doublet.
t Full width at half-peak height.
in agreement with previously reported
measurements on pigeonite crystals, con-
sidering the lower calcium content of
sample 15076. Values of 4.7 and 4.1 kcal/
mole were obtained for samples 14053
(En58Fs3oWoi2) and 12021 (En59Fs32
Wo9), respectively (Schurmann and
Hafner, 1972). In this connection, Haf-
ner, Virgo, and Warburton (19716) had
suggested a linear relationship between
the exchange energy AG°E and the cal-
cium content of the pyroxene. Further-
more, assuming a constant AG°E value
over the temperature range 500°-1000°C
(Virgo and Hafner, 1969), the values of
k equal to 0.08 and 0.09 for pigeonites
15476 and 15076 correspond to equilib-
rium temperatures of 520° and 560 °C,
respectively. These values are signifi-
cantly less than the critical temperature
for ordering (Tc = 600°-810°C for
P21/c pigeonites) and suggest extremely
slow cooling over the temperature range
Tc-Tq {Tq is the annealing temperature
below which further ordering is not pos-
sible and is ^480 °C in orthopyroxenes ;
Virgo and Hafner, 1969) and over a long
geological time.
Limits on the Cooling History of
Pigeonite Basalts 15076 and 15^76
The crystallization of 15476 is initially
characterized by the fast growth of large
pigeonite crystals and at lower tempera-
tudes by rapid, heterogeneous crystalli-
zation as rims on the phenocrysts and a
finer grained groundmass. The lack of
augite exsolution in the phenocrysts is
also consistent with rapid crystallization
and cooling through the temperature
range — 1200°-950°C (Ross, Huebner,
and Dowty, 1973) and with the relatively
low total calcium. In the temperature
range 1000°-950°C, pigeonite of this
composition will undergo the C2/c ->
P2i/c transition (Prewitt, Brown, and
Papike, 1971). The b reflections in this
sample are quite diffuse compared with
the a reflections, suggesting small P21/c
domains. It should be noted that this
result is significantly different from those
for samples quenched rapidly from
TABLE 51. Clinopyroxenes from Apollo 15: Quadrupole Splittings
and Isomer Shifts at 77 °K
Quadrupole
mm/
Splitting,
sec
Isomer Shifts,*
mm/sec
Pigeonite
Ml
M2
Ml M2
15076, natural
15076, 750°C, 2
15076, 920°C, 1
15476, natural
days
day
3.05
3.01
3.03
3.01
2.14
2.13
2.13
2.13
1.28 1.26
1.27 1.25
1.28 1.26
1.28 1.25
* Referred to metallic iron at 298°K.
GEOPHYSICAL LABORATORY
655
d
'8
<v
o
O
d
o
d
Q
d
CQ
O
O
ft
02
CP
d
CD
X
o
Si
>>
o
.d
U
d -
"S.s
•+S CD
2 o
c3
o
d
a
d
o
o
O
as
o3
<N
d
»o
o
+-5
H
d
hJ
hH
pq
«!
Eh
b£
CD
bfi
c3
"53 ^h
CD
PL-
CS fH
(M
+
C^l
+
CO OS OS <N
OS CO 00 00
O i— i i— i o
dodo
I- b- 1^ o
i— I <— I H CO
r— I t— I i— I ,-H
© © o o
TfH ^ »?JH Tj"l
dodo
o o o o
£r o t^ c<«
OOOiCH
OS 00 00 OS
dodo
00 iO CO 00
CO h iC iO
dodo
00 O 00 CO
"* O CO (N
H(N(Nh
dodo
TJ no
d o o d
+? © O -g
d i> os d
CD CO CD CO
t^ b- b- t>-
o o o «*
iO iO iO iO
o
CD
OS
00
s
cq
^h
CO
OS
Tfl
tH
rt<
CO
0)
d
d
d
d
-d
.£
bC
CO
o
CO
OS
+
OS
<N
-f
00
CD
o
i— i
l-H
o
fe
Ol
0)
-d
+3
43
cu
d
o
a>
>
M
'53
a
CD
Ph
a>
-d
s
O
73
CD
S
cc3
O
656
CARNEGIE INSTITUTION
slightly above the C2/c -> P2i/c transi-
tion (Brown et al, 1972; Brown, Papike,
and Prewitt, 1972 ; Takeda and Williams,
1972). They found sharp class-6 reflec-
tions corresponding to extremely large
domains in compositionally different pi-
geonites quenched rapidly from the C2/c
field. Certainly a slower cooling process
to below 950 °C, resulting in many nu-
cleation sites for the domains in pigeonite,
would be consistent with the preliminary
results from this study. In fact, the cool-
ing rate through the C -» P transition
may have been comparable to that for
the Mull pigeonite (diffuse class-6 re-
flections and domain sizes of ^200 X
100 A in the a-b plane) .
Beyond this point, a complicated cool-
ing history is inferred. Certainly the do-
main size, the absence of unmixing, and
the low temperature of the Fe/Mg
equilibration appear to be incompatible
with a single cooling cycle. One means
of explaining these observed features in
the natural crystal is by a subsequent
heating event, such as contact with a
superimposed lava flow, in which the tem-
perature was raised, but not high enough
to initiate unmixing or growth of the
domains, and then was very slowly
lowered.
Pigeonite sample 15076 is composi-
tionally and structurally similar to
15476. A cooling history similar to that
of 15476 is necessarily inferred, although
the coarser grained matrix and the de-
tection of at least the beginning of ex-
solution may suggest a somewhat slower
cooling rate through the quenching
interval.
A multiple cooling history was also
inferred from the cation diffusion char-
acteristics of the Mull pigeonite (Brown
et al, 1972). Possibly in support of this
interpretation, it is noteworthy to point
out that although the Mull andesite has
a distinct phenocrystic texture, the
groundmass is at least microcrystalline.
Reheating processes for lunar rocks
might also be attributed to shock effects.
Excessive shock effects are required, how-
ever, to raise the temperature to -~/1000oC
(Dundon and Hafner, 1971), and these
lunar basalts do not exhibit any sign of
such events. It is suggested, therefore,
that the multiple cooling history inferred
from these diffusion studies is inherent
from the initial magma extrusion on the
lunar surface and that the subsequent
excavation process from craters has not
disturbed the original cooling history of
these basalts.
Optical Absorption Studies of the
Russian Luna 20 Soil*
P. M. Bell and H. K. Mao
In February 1972 the Russian un-
manned spacecraft Luna 20 landed on the
moon near Mare Crisium and close to the
northern perimeter of Mare Fecunditatis
(3°32'N, 56°33'E; Vinogradov, 1973).
Samples from this lunar highland site,
which is representative of the oldest type
of exposed lunar surface, were obtained
automatically by drill and subsequently
returned to earth by spacecraft. As a re-
sult of an international agreement on the
exchange of data and samples between
the United States and Soviet govern-
ments, a small group of American lunar
investigators were allocated a total of
2.036 g of the Luna 20 sample for study.
Of this sample, 0.05 mg was made avail-
able for study at the Geophysical Labo-
ratory. The samples were soil-cored from
shallow depths (less than 1 m). Judging
from the nature of Apollo soil samples,
these samples are representative mixtures
of meteorite impact ejecta from many
parts of the moon.
Study of the Luna 20 sample provided
an opportunity to obtain data on the
composition of a part of the lunar high-
lands without the assistance of manned
* The data reported here are part of a study
by Adams et al. (1973). Work supported in
part by National Science Foundation grant GA
22707 and National Aeronautics and Space Ad-
ministration grant NGL 09-140-012.
GEOPHYSICAL LABORATORY
657
03
CO
>>
S-H
o
-a
a
O
o
CM
a
o
o
GO
o
• f— I
CO
o
CO
•rH
CO
rb
a
CD
rQ
O
Sh
Oh
O
CD
CO
CD
S-i
CD
CO
PQ
H
T3
CD
o
go
CD
fafi
<!
CD
G
CD
X
o
Sh
>>
a
o
43
o
o
"3b
03
3
00
o
CN
O
o
o
CO T3
© «tf
CM
CO
©
CM
CM
<
©
CM
OS
o
CM
►J
CO
<N rH
o3
CO
CM
CM
o
CM
o
CM
si
CD
h-}^
CM O f-h
r-3 °
CD
CDIOCO^MOOOOIOOCO
CD^OJOOCOMNOIN^
lOOOOOOOl^i-H
© © ©
hcjO-1*
CO (M 00
i— i O i— i
O 00 rJH
O CM O
>OON
omcooo
CM
COi*COOONcDOOcO(N
MMffiOOCCCOHO^CS
"* O O rH
© © OS
CM
O rH O
(MmffiiOHMCJiCOHON
©NHoqioNooqq^
coddcoddNO^ddd
rH i^J
O i— i CO r- 1 CM OS ^f
©COCOOO^CO
O l> ^H
COOOiO
O O iO
CM
Tt1 o
rH O
^cOiOCOOININMO^'f
OH^qoNOfflooo
ioddNdd>oddddd
rH H^ CO
iO
©CO00HHrHrHrH<M©t^r- I
qqcoqcoNNoqoqq
000000^ HH'daid
CO H^ rH
OOOcdCOCOHNiOOlN
qooq^HMNONO
'oddddddio'oocjio
CO H/l rH
lOOHONiflNCOMiOH
PpCMOiOpcMOOrHCMp
CM HH rH
OONHCOCOCJ5NOHO
OprHOCMrHlCCMpCMiP
Moddddd>o'oddd
CO "* rH
lOOOliONCOOOHNOO
HH^oo<OHecq>oN
iModioddooio^d't'd
CM "^ rH
HOOCOiOnOffltOINOlT)*
COCMtOOrHO'OrHOH^lO
1— I rH rH rH rH
OOHMOHCOhOON'-i
ppcMp-^rHiqcoppp
CO H^ rH
OWOMNNCDONht)<
oohoioohcooooo
(NodddddiO'odood
CO H/t rH
O
Ph
■°l9oo^q5oo6
O M i_2 CD "■« ™ 1—4 — -1 — . f? W .fh
00
CM
00
0
o
©
00
©
o
IO
d
©
OS
os
OS
CO
OS
00
OS
00
d
©
o
00
OS
d
©
OS
t^
06
OS
06
OS
OS
OS
O
o
o
o
CD
O
fl
CD
^5
_CD
'ft
a
CO
0
o
. — I
6 *>■
^r CM
£ o
pq ■*
co ►**•
.22 £
co O
'5c
C
03
H tf
658
CARNEGIE INSTITUTION
support in space. The present study in-
cludes (1) a mineralogical determination
of the sample, based on a cooperative
effort that included spectral data ob-
tained by earth-based telescope; (2)
measurements of the diffuse spectral re-
flectance of the bulk sample; and (3)
measurements of polarized crystal-field
spectra of individual grains of glass and
crystal selected from the bulk sample.
Supporting chemical analyses were made
by electron microprobe, and the com-
bined effort yielded valuable data on the
chemistry and petrology of the Luna 20
site. The spectral method appears prom-
ising for analysis of other parts of the
surface of the near side of the moon and
for the study of other solar system bodies
that will not be explored in the near
future.
Analysis of the small volume of Luna
20 sample was a test of the sensitivity of
the techniques employed. It was also a
test of how representative the aliquots
were of the whole and, more significantly,
of the lunar soil at the Luna 20 landing
site. Miniaturization of techniques for
lunar sample analysis (Bell and Mao,
Year Book 72) was entirely adequate for
the task. Whether the results are mean-
ingful in terms of representing extensive
areas of the lunar highlands is a more
nebulous question. The present data
show that the part of the sample studied
represents the bulk sample returned
(total sample 50 g; bulk chemical analy-
sis ir Vii) )gradov, 1973). Comparison of
the dai with those of a well-studied
lunar site, such as the Apollo 16 site,
strongly suggests that the sample is a
well-mixed sample of the lunar highlands.
Representative microprobe analyses of
single grains are given in Table 53. The
data show the soil to be rich in anorthite,
evidently from anorthositic rocks that
are much lower in iron and higher in
calcium and aluminum than the crystals
and glasses from lunar basalts. Studies
of small rock fragments (K. L. Cameron
et al., 1973) indicate that there are troc-
tolitic and noritic variations of anortho-
site and that lunar basaltic rocks are
represented as well.
Optical absorption spectra show min-
eralogical variations as shifts in intensity
and energy of crystal-field bands of iron.
Figure 135A shows the transmission
spectra of glass fragments. The spectra
are closely related to the amount of iron
150
£ 100
9- 50
Colorless glass L-20-AII
FeO 5.49%, Ti02 0.28%
B
500
1000 1500
Wavelength, nm
2000
Fig. 135. (A) Unpolarized transmission spec-
tra of Luna 20 glass fragments. Most Luna 20
glasses are colorless. Curve A is for a brown
fragment unusually rich in iron and titanium
that may be of mare origin. Curve B glass is
colorless but still high in iron, illustrating that
titanium is essential for strong absorption below
0.6 fim.
50-
40
£ 30
"■e 20
Luna 22002, 2, 4d
Orthopyroxene
/ \
500
1000 1500
Wavelength, nm
2000
Fig. 135. (B) Polarized transmission spectra
of an orthopyroxene crystal from the Luna 20
soil. (See analysis 2,4d in Table 53.) Principal
absorptions near 1 and 2 jxm arise from Fe2+.
GEOPHYSICAL LABORATORY
659
0.6
0.4
0.2
Luna 20 A8
Pigeonite-augite
500
1000
1500
Wavelength, nm
2000
Fig. 135. (C) Polarized transmission spectra
of a zoned pigeonite crystal from the Luna 20
soil. (See analysis L20A8 in Table 53.)
40
o 20
1 1 '
i i i
i i i i i | i i i i i i
Luna 20 A9 olivine -
- \\
/
■ v\
\ \"v
/'
, i ,
i . . . . i i
500
1000 1500
Wavelength, nm
2000
Fig. 135. (D) Polarized absorption spectrum
of a low-iron olivine (Table 53) from the Luna
20 soil. The principal absorption bands near
1 fim, which are due to Fe2+, are notably weak.
and titanium (Bell and Mao, 1972). The
correlation of the iron content and the
intensity of the 1.05-/*m band is shown in
Fig. 136. Figure 135 B-E shows the
polarized spectra of pyroxenes, an oli-
vine, and representative plagioclases.
The absorptions of transmitted light are
closely related to the crystalline structure
and to the amount of iron and titanium.
Summarizing the analogous reflection
spectra of bulk lunar samples, Fig. 137A
shows a plot of the wavelength of the
crystal-field band near 1 ^m versus the
band near 2 /ma of several samples from
the Apollo missions and from the Luna 20
mission. The wavelength dependency of
the bands on iron causes a marked sepa-
ration between samples from the high-
lands and samples from the mare basins.
The Luna 20 sample shows the effect of
mixture.
Telescope spectra were obtained for
much larger samples (10-20 km diameter
areas of the lunar surface) but were in-
terpreted from the spectra on single
grains and bulk samples. Figure 137B
shows a diffuse reflectance spectrum of
the Luna 20 bulk sample normalized on
the same base (an area designated MS-2,
located in Mare Serenitatis) as telescope
spectra from several areas of the moon
shown on the same plot. The diffuse
bands indicate not only the mineralogical
components but also the amount of glass.
Compared with the other telescope spec-
tra shown in Fig. 137B, it is possible to
note the Luna 20 mixture relative to
unmixed mare and highland craters. The
relative exposure of Luna 20 is consistent
with a mixed source, part of which may
a Spectra
of anorthite
0.43%
FeO
A 14163,33
800 1000 1200
Wavelength, nm
1400
1600
Fig. 135. (E) Polarized (a) spectrum of an
iron-rich plagioclase from the Luna 20 soil
(1,2c in Table 53) compared with a similar
spectrum of an Apollo 14 plagioclase. Most
Luna 20 plagioclases have a very low iron con-
tent. The absorption arises from Fe2+.
660
CARNEGIE INSTITUTION
E
o
a
a>
a
o
o
c
o
o
CO
_Q
<
150-
00-
50-
0
1 !
• 1 | | 1
i ' !
1 1 « 1 | 1 ' 1 ' |0 ' 1 ■
-
• /
/ ° -
+ /
J •
-
/ o
/ • • °
-
•
/ •
• /
o
-
/•
-
-
/ +
-
—
o Apollo 14163,33 -
/
• Apollo 15601,94
+ Luna 20
/ . 1
.ill.
i , 1
■ i . i l i . i i i , i
10 15
FeO, weight per cent
20
Fig. 136. Plot of the absorption coefficients of the 1.05 ^m band of lunar glasses versus their iron
content. The wavelength of the Fe2+ band remains the same but the depth of the band increases
with the iron content.
GEOPHYSICAL LABORATORY
661
.02
,^
1.00
E
=L
" — "
.98
X
h-
e?
.98
<£
\±i
_l
LU
.94
>
<
.92
.90
ROCKS, BRECCIAS
cn2> SOILS
® LUNA 20 SOIL
■ LUNA 20 PYROXENE
CRYSTAL
JF-
— of
\W>P^
.8
I Q ? 0 ? I
WAVELENGTH (^m)
2.2
2.3
Fig. 137. (A) Plot of the wavelength of the center of the absorption band near 1 /im versus that
of the band near 2 /xm for lunar samples. Numbers refer to Apollo missions. The Luna 20 soil is
similar in average pyroxene composition to mature soils at other highland sites. Individual py-
roxenes from the Luna 20 soil range from orthopyroxene (2,4d) to pigeonite (L20A8) to augite
(2,c). Some of the calcic pyroxenes may be contaminants from the lunar maria.
662
CARNEGIE INSTITUTION
.00
iC 1.00
LP)
1*1-
<
Q
LU
_l
<
u
CO
>
I-
U
LU
.00
.00
I III
u- 1.00
LU r
a: L
0.94
0.90
0.86
0.82
0.78
0.74
0.70
0.66
0.62
i i i i i i i i i i i i i i
♦ MARE CRISIUM l/MS-2
«- APOLLO I6/MS-2
• LUNA-20/MS-2
4 PROCLUS l/MS-2
■ ARISTARCHUS/MS-2
■5 ll\
la1*
I.I..I
I1 1
tf
«'ll» J
• • • •
t
lift
I1!
*0ll
I
-i
_
L
in
i i i
i i i
0.3 0.5 0.7 0.9
WAVELENGTH (/t)
Fig. 137. (B) The relative spectral reflec-
tivity curves for the same areas used in Fig.
137A.
have been ejected from the relatively
undisturbed crater Apollonius C (Vino-
gradov, 1973).
The results show a workable correla-
tion between laboratory measurements of
lunar samples and telescope data. It is
possible to estimate in a first approxima-
tion from these data that the sample is
from a highland site of moderate ex-
posure. The mineralogy is mostly
anorthositic with a mixture of fragments
derived from lunar basalt. The determi-
nations are mineralogical ; however,
chemical compositions can be deduced
that are close to those actually deter-
mined on the Luna 20 sample.
Luna 20 Plagioclase: Crystal-Field
Effects and Chemical Analysis
of Iron
H.K.Mao and P. M.Bell*
Anorthite crystals from loose, fine-
grained material of the Luna 20 sample
(Bell and Mao, this Report) were se-
lected for detailed examination by elec-
tron microprobe and by the crystal-field
technique. The purpose was to analyze
the properties of iron, a minor element in
lunar plagioclase, because of its high
sensitivity to the oxidation state of the
chemical environment in which anorthite
crystallizes (Bell and Mao, this Report) .
It is known that the oxidation-reduction
conditions during crystallization and
cooling on the moon exert strong control
on the chemical path through which
lunar rocks are formed. Furthermore, it
is known from studies of Apollo rocks
that the chemical atmosphere was ex-
ceedingly reducing (Po2 — 10"12 to 10"15
atm for lunar rocks, compared with P02
= 10"5 to 10"8 atm for terrestrial equiva-
lents). What was not known was the ex-
tent to which chemical reduction could
occur in silicate minerals after crystalli-
zation, a factor that could shed light on
the environment during the 3-4 billion
years since the lunar rocks formed.
In the course of studying the Luna 20
plagioclases, it was discovered that the
crystals were zoned in an extraordinary
* Work supported in part by National Science
Foundation grant GA 22707 and National Aero-
nautics and Space Administration grant NGL
09-140-012.
GEOPHYSICAL LABORATORY
663
Ml
ft-w: ftr* #»!:
B
Fig. 138. Photomicrographs of oriented crystals of metallic iron alloy in Luna 20 plagioclase.
(A) Reflected light (width of inclusions = 0.14 ,um). (B) Transmitted light (same inclusions
shown in A).
664
CARNEGIE INSTITUTION
pattern, and that fine, rod-like inclusions
of iron alloy occurred aligned crystal-
lographically. Both zoning and free-iron
inclusions have apparently formed dur-
ing reduction of the plagioclases whether
the iron originally crystallized in plagio-
clase or was introduced later. The obser-
vations have provided support for
theories of chemical behavior on the
moon (Bell and Mao, 1973) .
Photomicrographs of the metal inclu-
sions in one of the plagioclase crystals are
shown in transmitted and reflected light
in Fig. 138 (after Bell and Mao, 1973).
The inclusions appear to be individual
crystals that are observed to be parallel,
whether they are exposed on the surface
of the section or plunge out of focus into
the section. By moving the electron
microprobe beam in steps of 1 ^m across
an exposed inclusion, it was possible to
obtain approximate concentrations of
iron and nickel, the only elements de-
tected. Figure 139 shows a plot of iron
content versus nickel content at various
overlapping steps across the inclusion.
On the basis of a linear fit, nickel was
found to be present at a level of approxi-
mately 1-2 wt %. Major-element analy-
ses of the host plagioclases are given by
Bell and Mao (this Report) . The plagio-
clase in Fig. 138 contains an average of
0.12 wt % iron, but the iron is strongly
zoned.
All the plagioclase crystals studied
were zoned in iron, apparently unrelated
to the iron inclusions and to any syste-
matic trend. In some cases, the FeO con-
tent varied from 0.05 to 0.43 ± 0.01 wt
% (Fig. 140). Isolated zones enriched in
iron appear to have formed at random.
Conceivably the iron could have mi-
grated, perhaps during sufficiently intense
conditions of meteorite impact.
Zoning effects tend to be averages by
the crystal-field technique, which is sen-
sitive to the low concentrations of iron
of these crystals. Figure 141 shows a plot
of the absorption coefficient of the a-
polarized crystal-field band in the Luna
20 plagioclases and in plagioclases from
Apollo 14 and 15 samples. This plot not
only provides a good correlation of iron
versus absorption coefficient (Fig. 140)
but also shows that at these levels of
concentration, iron occupies a unique
position in the plagioclase structure.
Calculations of octahedral and tetra-
0.5
CD
O
^_
Q.
0.4 -
0.3 -
CD
"cd 0.2
.*:
o
0.
1 1 1 1 1 1 1 1 1
Luna 22003, 3e
1 i i i i i i
i
•
plagioclase
•
-
—
' ^<^
—
•
-
^s^ •
-
p^f i i. i i i i i i
•
i .... i .
Iron , weight per cent
Fig. 139. Plot of nickel content versus iron content of a metallic crystal inclusion in host Luna
20 plagioclase (analyses by microprobe ; varying amounts of iron and nickel caused by overlap of
electron beam onto host plagioclase).
GEOPHYSICAL LABORATORY
665
0.2 0.3 0.4
Total Fe , weight percent
0.6
Fig. 140. Plot of absorption coefficient (cm-1) of crystal-field band at 1250 nm in lunar plagio-
clase versus weight percentage of FeO determined by electron microprobe.
a spectra
of anorthite
Wavelength, nm
Fig. 141. Alpha crystal-field spectra of Luna
20 plagioclase and plagioclase from Apollo 14
and 15 soils.
hedral crystal-field splitting for iron
(J0rgensen, 1962a, 6) do not yield satis-
factory results in this case, because the
band energy is intermediate. Possibly
there is averaging between the six-fold
calcium site and the four-fold aluminum
site (Bell and Mao, 1973), both of which
are distorted in plagioclase. The absorp-
tion is strongly polarized, clearly sug-
gesting a distorted site.
From the present study it is established
that iron has migrated within the lunar
plagioclase, judging from the unsyste-
matic nature of the chemical zoning. The
iron alloy crystals may be segregations
by exsolution, although simultaneous
epitaxial growth during crystallization
cannot be ruled out. The observations
suggest that major chemical reduction of
iron occurred after crystallization, during
a chemical history unlike that of any
known plagioclase.
666
CARNEGIE INSTITUTION
Volatilization of Iron-Bearing
Silicates in the Presence of Carbon
M. G. Seitz
Chemical fractionation caused by dif-
ferences in volatility of elements and
their compounds appears to have played
a major role in establishing abundance
patterns in material of the solar system.
Volatile fractionation is considered im-
portant both in the initial cooling and
condensation of the elements from the
solar nebula (Lord, 1965) and in subse-
quent reheating such as may have oc-
curred during lunar volcanism (Storey,
1973).
Mass transfer of the alkalies sodium
and potassium to the vapor phase is
known to occur from basalts heated in
vacuum. At temperatures of 1500 °C,
vapor species such as iron and silicon
oxides have been detected from heated
lunar basalts (DeMaria et al., 1971).
Volatilization experiments were per-
formed to determine the conditions re-
quired to fractionate these latter ele-
ments. Because carbon is a common con-
stituent of the solar system, studies were
performed on silicates both in inert en-
vironments and in the presence of graph-
ite. Langmuir-type vaporization experi-
ments (Margrave, 1959) were performed
in which a 20-mg sample was heated in
crucibles under vacuum. The experi-
ments replicate volatile loss, for example,
from natural material ejected into
vacuum during volcanism or heated from
close passage near the sun. Because
equilibrium between the condensed and
vapor phases is not maintained, the re-
sults are not directly applicable to pre-
dicting the composition of a condensate
derived from a gas.
Samples used in these experiments
were a synthetic high-alumina basalt
having a composition near that of lunar
rock 14310, a basalt from the Juan de
Fuca ridge, and a fragment of the St.
Severin meteorite. The first two basalts
have relatively high iron content. The
third sample was selected to represent
materials of chondritic composition. The
basalts were fused to form glasses as
starting material. After being fused at
1450 °C, the chondrite consisted mostly
of glass with some remaining olivine
crystals and was run in this partially
crystalline state.
Metal crucibles were made from foils
rolled to form cylinders 7 mm in diame-
ter, 4 cm long, and open at both ends,
which were placed in a horizontal carbon
resistance furnace. The furnace was
evacuated in a bell jar to below 1 X 10"5
torr, using an oil-diffusion pump and a
mechanical pump. Pressure was moni-
tored by means of an ionization gauge
below the base of the bell jar. The fur-
nace was then heated, and the tempera-
ture, which was monitored with a thermo-
couple placed m the center of the crucible,
was estimated to be within 20 °C of the
sample temperature. Because of the like-
lihood that the thermocouple was con-
taminated by the vapor species, a new
thermocouple was used for each run. Run
times were systematically varied from 10
to 60 minutes.
Tungsten and platinum formed com-
pounds such as CaWC>4 and FeW04 or
absorbed iron and were unsuitable as
crucibles. No reactions between rhenium
and silicate were found from analyses of
the silicate and crucible after heating,
and, therefore, rhenium crucibles were
used in the subsequent work. Runs with
graphite were made by mixing graphite
with the powdered sample and placing
the mixture in the rhenium crucible.
After evaporation, the composition of
the residue and the vapor that was con-
densed on an aluminum substrate were
analyzed by electron microprobe. No
large-scale inhomogeneities were ob-
served, and the composition of the
sample was taken from the average of
three or more analyses. A 10% change in
the concentration of elements listed in
Table 54 compared with the starting ma-
terial was considered significant.
GEOPHYSICAL LABORATORY
667
TABLE 54. Evaporation of Chondritic Material in the Presence of Carbon
Starting Material
1350°C, 10 min
1400°C, 20 min
Si02 42.92
47.02 (37.7)*
50.62 (45)*
A1203 2.33
4.13 (0)
5.12 (0)
FeO 22.17
2.18 (94)
0.08 (99.8)
MgO 26.41
43.95 (6)
42.65 (26.4)
CaO 2.04
3.62 (0)
3.91 (11.1)
Na20 1.00
0.0 (100)
0.0
K20 0.10
0.0 (100)
0.0
Percentage lost.
Detectable losses of sodium and potas-
sium were observed from the basalt
samples held in the partially molten
state at 1150°C for 20 minutes. The com-
plete loss of these alkalies occurred
within 10 minutes from all samples at
1350 °C. In subsequent heatings to
1450 °C for periods of 20 minutes, how-
ever, no loss of an additional element was
noted. These observations are in accord
with the known volatility of metal oxides
that constitute silicates.
In the presence of carbon, however,
significant volatilization of the major
elements occurred from all samples at
temperatures below 1400°C. Residue
analyses of the meteorite sample are
given in Table 54. The percentage of
oxide lost to the vapor was calculated
from the residue composition and is given
in parentheses. The high loss of iron and
silicon at 1350°C is typical of the basalt
samples as well. No free metal or car-
bides were found in the residue. Analyses
of the condensate produced at 1350 °C
confirmed that iron and silicon formed
the principal components of the vapor
after loss of the alkalies. No voids or
bubbles were observed in the quenched
glass, and evaporation was believed to
occur only from the melt surface. From
the experimental data, mass-loss rates for
silicon and iron at 1350 °C were deter-
mined to be 1.7 X 10"5 and 3.1 X 10"5 g
cm-2 sec"1, respectively. Evaporation at
1400 °C for 20 minutes resulted in al-
most complete loss of iron accompanied
by significant losses of silicon, magne-
sium, and calcium. No loss of aluminum
was observed from any of the samples.
At the higher temperatures used in
these experiments no crystals were de-
tected in the quenched residual basalt.
Thus volatilization occurred rather sim-
ply from the melt to the vapor phase.
The temperature range investigated,
however, was within the melting interval
of the chondrite so that during volatili-
zation the sample consisted of olivine,
other fine-grained crystals, and liquid.
The solid phases are expected to con-
tinuously change their character in re-
sponse to changes in chemistry of the
melt due to evaporation. In spite of this
complication, however, no basic differ-
ences between vaporization from the ba-
salts and the chondrite were observed.
In the presence of carbon, the increased
volatility of the major elements appears
to be directly related to the reducing con-
ditions caused by the carbon. The
evaporation rate of silicon as Si02 is
lower than measured in these experi-
ments. Alternatively, silicon as SiO has
an evaporation rate considerably higher
than measured. It is likely, therefore,
that a partial reduction of silicon is
responsible for the observed volatility.
Similarly, the evaporation rates of cal-
cium and magnesium are expected to in-
crease upon partial reduction. Assuming
that free iron is present in a homogene-
ous melt with unit activity, its mass
transfer rate at 1350 °C may be calcu-
lated to be 5.4 X 10"5 g cm-2 sec"1. This
value is close to the measured rate.
668
CARNEGIE INSTITUTION
The observed evaporation rate of iron
and silicon may be high enough to have
influenced the evolution of magmas on
atmosphere-free bodies. For example, in
the presence of carbon a well-mixed
magma layer 20 m deep at 1350 °C would
be depleted of iron in one year. Possible
evaporation residues in nature are lunar
highland rocks that have low Fe/Mg
ratios and high calcium and aluminum
contents. However, these materials also
contain alkalies that would have been
lost before iron and thus cannot be de-
rived from iron-rich material solely by
vapor fractionation. A similar conclusion
is reached for iron-poor chondrules such
as those found in the Allende carbonace-
ous chondrite, which contain 0.1% Na20.
Volatile fractionation does produce ele-
ment trends that are observed in mete-
orites. For example, Ahrens (1970) has
pointed out that although the concentra-
tions of calcium and aluminum vary in
individual chondrites, the calcium-to-
aluminum ratio by weight remains con-
stant at 1.08. Moreover, this ratio is not
characteristic of the major calcium- and
aluminum-bearing phases of these mete-
orites. Analogously, it is seen from Table
54 that the calcium-to-aluminum ratio is
preserved in the chondritic residue with
over 40% by weight of the meteorite frac-
tionated to the vapor. Other element
trends exist within a single group of
meteorites. For example, Keil (1968) de-
scribed the low iron concentrations and
low silicon-to-magnesium ratios in the
Type II enstatite chondrites relative to
the Type I enstatite chondrite. The Type
II specimens are also higher in calcium
than the Type I specimens. From the
experimental results given here, these
same trends would be obtained in mate-
rial produced from Type I material by
slight volatile fractionation.
These experiments indicate that vola-
tile fractionation of silicates is enhanced
under reducing conditions where elements
are lost in the sequence Na, K, Fe, Si,
Mg, Ca with increasing temperature.
Calcium and aluminum appear to be the
most refractory elements. The volatiliza-
tion rates of iron-bearing silicates in the
presence of graphite are high enough to
have affected the evolution of magmas
on atmosphere-free bodies.
SYSTEMATIC PETROGRAPHY
Development and Modification of the
Information System RKNFSYS
F. Chayes
The rock information system
RKNFSYS, being developed as a proto-
type of one form of quasipublic service
that could provide sophisticated reduc-
tions of petrographic data at reasonable
cost with currently available technology,
has been described in previous Year
Books: a review of its design and a
discussion of the practical advantages of
a system of this type were given in Year
Book 71. A new edition of the data base
of the system, containing 11,160 analyses
drawn from 634 references, was gen-
erated late in July 1972, and Version III
of the User's Manual, including the table
of contents of this edition (RKSP72) and
announcements of a few new reduction
services noted in last year's Report, was
distributed in September. The important
normative variable jemg, which deter-
mines the jo content of ol and the js
content of hy and di, has previously been
available only in the printout of program
NORM; and subroutine CIPW, which
computes all norms used by the system,
has previously yielded only standard
weight percent norms. In October jemg
was made a full system variable so that
it can now be used in any of the data
reductions provided by the system. At
the same time subroutine CIPW was
modified to permit optional generation of
weight percent, oxide molecular, or
cationic norms; and calling programs
were modified to read, record, and trans-
mit the user's option in this respect.
GEOPHYSICAL LABORATORY
669
A major improvement in the system's
ability to form linear combinations of
the types so common in petrographic
work was implemented in January. This
and the recently reduced priority of
routine extension of the system data base
are by far the most important develop-
ments of the year; they are discussed
separately in the following sections.
Present Status and Possible Expansion
of the Information Content of the
Data Base
The data base is now large enough to
provide ample input for the development
and testing of system reduction programs
and is in continual use for this purpose.
It is also large enough to be of consid-
erable substantive value — it is in fact a
major research resource. A satisfactory
data base for igneous petrology as a
whole, however, should not be confined
to Cenozoic volcanic rocks and probably
should contain more information than
the source name and bulk chemical
analysis of each rock included.
Although relaxation or elimination of
coverage restrictions might materially
increase operating costs by increasing the
size of the data base, it would not neces-
sarily involve fundamental redesign of
the system. But expansion of informa-
tion content — to include, for instance,
mineralogy, texture, structure, and age —
would indeed require fundamental re-
design. Such redesign will ultimately
have to be undertaken, and one of the
more distressing consequences of its suc-
cess will be the subsequent need for
complete reexamination of the references
from which the existing data base has
been drawn. From this point of view,
routine extension of the present data ac-
quisition activity, which is all that can
be sustained at current levels of support
and staffing, seems uneconomic. In the
generation of an electronic data base
from published materials the major in-
vestment of time and effort is precisely
in the conventional library search, and if
a comprehensive search is to be made in
the reasonably near future, there seems
little point in making an incomplete one
now. The question to be decided is
whether the time has come to plan and
implement a thorough search. Routine
expansion of the system data base, a
major activity since National Science
Foundation support of the program be-
gan in 1969, will receive rather low
priority pending decision of this question,
concerning which discussion is solicited.
In the meantime, intermittent atten-
tion is being given to ways in which the
information content of the data base
might be expanded. Most published rock
analyses are accompanied by some infor-
mation about the mineralogy of the
analyzed materials. Quantitative modes
are still so rare that there would be little
point in attempting to incorporate them
in the data base. Qualitative modes —
listings of the essential minerals — of
analyzed rocks are quite common, how-
ever, and a data bank that included
them would clearly be superior to one
that did not. Indeed, this is the sense in
which the word "mode" was first used
in petrography, and the extent of con-
cordance— i.e., of qualitative agreement
— between normative and modal assem-
blages has always been a matter of inter-
est and concern to petrologists. To cite
only one example, it has long been known
that olivine-bearing basalts are not in-
frequently (^-normative. Recent work
(Chayes, 1972) indicates quite clearly
that a considerable majority of norma-
tive non-feldspathoidal Cenozoic basalts
are (^-normative rather than oZ-norma-
tive. Whether as a basis for petrological
speculation or for purposes of designing
or justifying modifications of normative
combining conventions, it would be use-
ful to know how often olivine appears on
the lists of essential minerals in these
^-normative basalts. The only way this
information can be obtained at present is
by restudy of the source references, but it
could easily be extracted from an appro-
670
CARNEGIE INSTITUTION
priately expanded data base, and there
is no doubt that the data base could be so
expanded. In similar vein, although
radiometric ages for analyzed specimens
are still exceedingly scarce, the descrip-
tions accompanying most published anal-
yses usually specify stratigraphic age,
and this information would often be of
interest to the user of the system if it
were available.
It is to be remembered, however, that
for the present and at least the near
future the major expense of operating
any system of this type will be in the
form of charges for active mass storage.
Could the data base be expanded to in-
clude qualitative mineralogy and strati-
graphic age without physical enlargement
sufficient to lead to critical increases in
storage charges? It seems quite likely
that this could be done; indeed, it ought
to be possible to pack qualitative miner-
alogical and stratigraphic age data into
a single machine word of 36 bits, the
length currently available. Preliminary
consideration suggests that no more than
25 or 30 mineral names would be required.
Each of these could be assigned a par-
ticular bit, this bit to be set if the mineral
is recorded in the source description and
left unset otherwise. The remaining un-
assigned bits would probably be sufficient
for the storage of similarly coded strati-
graphic age information. If it proved
desirable to record phenocryst and
groundmass mineralogy separately, a
single 36-bit word would be inadequate
for the recording of mineralogical assem-
blage and stratigraphic age, but two
would be ample.
Each analysis is now assigned 16 words
of storage, so that extension of the infor-
mation content in this fashion would in-
volve an expansion of 6% to 12% in the
size of the base. Even this, however,
would be unnecessary if the rather exten-
sive reorganization facilitating it also
provided for elimination of currently un-
used space. Specifically, although the
existing system routinely allots space for
storage of Zr02, Cr203, and C02 in each
analysis, these components are rarely de-
termined. Just how rarely was not ap-
preciated at the time the present system
format was developed, but in 11,160
analyses contained in the current version
of the base, there are only 248 Zr02 de-
terminations, 374 of Cr203, and 1866 of
C02. The system should be redesigned so
that storage is allotted for minor elements
only when actually needed. In sum, en-
largement of the information content to
include mineralogical assemblage, strati-
graphic age, and perhaps even some
aspects of texture and structure, need not
unrealistically increase either the physi-
cal size of the data base or the cost of
maintaining it in active mass storage.
In any such expansion of information
content the overriding requirements are
that transfer of material from source ref-
erence to data base should not be too
difficult and that exploitation of the data
base should be as simple and straight-
forward as possible. The first of these is
by far the more important, for if the data
were actually incorporated in the base
their extraction from it without undue
inconvenience to users could almost cer-
tainly be arranged. Exploratory work
suggests that although development of a
reasonably simple scheme for transfer-
ring age and modal data from printed
page to data bank would be troublesome
and demanding, it would not be for-
bidding.
An Improved Function Compiler }
Subroutine TRNEVL, Incorporated
in the System Work-File Generator
The most important retrieval program
of the system scans the data base and
generates from it work files containing
data to be reduced by other system pro-
grams. This scanning program, the key-
stone of the system, is designed to per-
form the electronic equivalent of the
conventional library search that neces-
sarily precedes nonelectronic reduction of
GEOPHYSICAL LABORATORY
671
data gathered from the reference and
journal literature. A sufficiently powerful
work-file generator makes it possible to
do the actual data reduction with pro-
grams that are simple, small, and, accord-
ingly, relatively inexpensive to operate.
Linear combinations of variables
copied or computed from the raw analy-
ses are often the object of petrochemical
study, and the system work-file generator
has always had limited capability for
forming and evaluating very simple func-
tions of this kind. The 1972 version of
the generator, for instance, could form
combinations of the type:
(diXi ± djXj ± . . .)/(dmxt
dnXn
.),
the a's being numerical coefficients, the
x's the identification numbers of either
oxide weight percentages, oxide molar
proportions, or normative parameters.
This feat was initiated by a punched card
input so complex as to lead to frequent
operator error. Dissatisfaction with both
the limited repertoire of combinations that
could be generated in this fashion and the
complexity of card input required at op-
eration time culminated in the coding of
a new function generating subroutine that
is essentially a rudimentary compiler.
The new function generator, written in
FORTRAN, accepts 80-character FOR-
TRAN expressions that contain any com-
bination of the operators +> — , *, / and
**, but contain neither function refer-
ences, subroutine calls, nor parentheses of
levels greater than one. Acceptable ex-
pressions may contain any combination
of numerical operands with literal op-
erands known to the system, i.e., the sys-
tem names of oxide weight percentages,
oxide molar proportions, normative mole-
cules, and the normative index femg.
As many as ten functions, each read in
from a single card, may be requested in
the building of any work file. Functions
are coded on initial entry to the generator
and are evaluated for each current analy-
sis on subsequent entries. If the last
operand of an expression is numerical,
the operator that precedes it may be an
inequality. Depending on whether or not
this is so, each current value of the ex-
pression is added to the "sort battery"
or "save vector" of the analysis from
which it has been computed.
The new function generator, which was
begun almost on impulse but soon turned
into a major programming assignment,
greatly increases the power and simplifies
the operation of the system. Copies of it
are available, but interested readers will
realize that in addition to recognizing
only literal operands used in RKNFSYS,
the generator takes advantage of the spe-
cific hardware and software available.
Readers using computers other than one
of the Univac 1100 series will probably
find it useful only as a general guide.
A Relation between Silica Saturation
and Oxidation State in Class II,
(ol -f- hy) Normative Basalts
F. Chayes
Other things being equal, the kinds and
amounts of major normative components
yielded by the conventional calculation
schedule vary materially with fluctua-
tion in the molar ratio Fe203/(Fe203 +
FeO) . Petrologists encountering unduly
high values of the latter occasioned by
surficial or fumarolic alteration ordi-
narily refrain from normative calcula-
tion. There is no general agreement
about the level of oxidation sufficient to
warrant this abstinence, and it is often
tolerated in the first instance only be-
cause in many unmetamorphosed regions
reasonably fresh material can be found if
the search seems worthwhile.
In certain metamorphic terranes, on
the other hand, a relatively high level of
oxidation not immediately connected
with weathering or decomposition may
characterize rocks of many kinds over
large areas. In the study of such mate-
rials, as in the treatment of surficially
altered materials so expensive that re-
sampling is out of the question (e.g., sub-
672
CARNEGIE INSTITUTION
marine basalts), the practice of routinely
adjusting analyses by numerically reduc-
ing some part of the reported Fe203 to
FeO prior to normative calculation is be-
coming increasingly common. The pur-
pose of the adjustment is usually to re-
store the sample composition to its condi-
tion prior to alteration or metamorphism.
To the extent that this earlier condition
is defined by theory or speculation it may
be argued that the normative consequence
of the adjustment is of no importance
or that the norm obtained from the ad-
justed analysis is actually preferable to
that obtained from the raw data. There
are certainly specific occasions when one
or other of these arguments is quite
legitimate.
To the extent that it is thought to be
justified by petrographic evidence, how-
ever, the adjustment is often unrealistic.
It has long been realized, for instance,
that salic alkaline rocks, whatever their
level of silica saturation, tend to be
rather highly oxidized even when there
is no reason to suspect that they have
been weathered or otherwise altered. It
is perhaps not so widely recognized that
normatively feldspathoidal basalts are
also more highly oxidized than those
whose norms are free of feldspathoid.
And it does not seem to have been noted
before that silica-saturated basalts in
which ol > hy tend to be more highly
oxidized than those in which ol < hy.
This latter effect is neither strong enough
nor simple enough to generate strong,
positive correlation between ol/hy and
Fe203/(Fe203 + FeO), but in the ba-
salts of the oceanic islands and basins,
the only ones for which calculations are
available at this writing, it seems to be
easily significant. Evidence bearing on
the matter is summarized in Table 55;
part A contains for all relevant analyses
in the current version of the data base,
the joint frequencies of ol/hy greater and
less than 1, and of Fe203/(Fe203 +
FeO) greater and less than 0.15, the mean
value of this (molar) ratio.
TABLE 55. Relation between Low and High
Values of X = ol/hy and the Molar Ratio
Y = Fe203/(Fe203 + FeO) in (ol,hy)
Normative Basalts of the Oceanic
Islands and Basins
A. All Data
Y
X
>1
<1
>y
<y
76
81
77
136
X2 = 5.596
B. Analyses with
ol <25
Y
X
>1
<1
>y
<y
56
52
75
136
X2 = 7.849
C. Analyses with
Y - y <2sy
Y
X
>1
<1
>y
<y
68
81
65
136
X2 = 6.424
D. Analyses with
ol <25 and Y - y <2sy
Y
X
>1
<1
>y
<y
49
52
65
136
X2 = 7.486
Rocks whose norms are very rich in ol
are likely to be accumulates, and those
characterized by really extreme oxida-
tion are likely to have undergone exten-
sive surficial alteration. Specimens of
either category ordinarily provide little
information about the oxidation state of
fresh, relatively undifferentiated basalt.
Accordingly, the former are excluded
from the tally in part B of the table, the
latter from the tally in part C, and both
from that in part D. For all four data
GEOPHYSICAL LABORATORY
673
sets x2 is in excess of the 0.02 point, and
in two of the four, those from which
analyses excessively rich in ol have been
excluded, it is comfortably past the 0.01
point, the test in each case being against
the hypothesis that oxidation, as char-
acterized by the molar Fe oxide ratio,
and silica saturation, as characterized by
the ol/hy ratio, are independent. There
is thus a very strong suggestion of some
nonrandom association between the two,
and the partitioned frequencies in all
four parts of the table indicate that speci-
mens with above-average oxidation ratios
are relatively more abundant among ba-
salts in which ol > hy. Yet adjusting the
analysis of a saturated basalt by reducing
some of its Fe203 to FeO prior to norma-
tive calculation increases its apparent
ol/hy ratio, making it more similar to
that encountered in rocks in which the
Fe203/FeO ratio ordinarily is high.
It may be objected that highly oxidized
basalts in which ol » hy are "really"
ne-normative materials that happen to
project in the hy field only because of
oxidation, so that an oxide adjustment
that returns them to the ne field is not
inappropriate. In fact, however, reason-
able adjustment of the oxidation ratio of
a basalt analysis will only rarely shift its
projection from the hy to the ne field, and
such shifts as are generated in this way
will often, perhaps usually, introduce
anomalies quite as severe as the one they
are thought to remove. In one such re-
cently proposed adjustment, for instance,
the weight percentage of Fe203 in excess
of (Ti02 + 1.5) is converted to FeO
prior to normative calculation. Be-
cause of their high Ti02 content, alkali
basalts will very rarely be shifted from
the hy field to the ne field by this trans-
formation ; such shifts are far more likely
to occur in Ti02-poor subalkaline basalts,
and ne-normative basalts low in Ti02
are exceedingly uncommon.
Now the adjustment of specific analy-
ses for specific clearly stated reasons may
certainly be appropriate. Carried out by
rote or as a matter of convention, how-
ever, adjustment of published analyses
tends to blur the increasingly unclear
distinction between observation and in-
ference. There is no more justification
for routine ex post facto adjustment of
the Fe203/FeO ratios of analyses culled
from the literature than there would be
for similar treatment of, for instance,
their K20/Na203 or Al203/Si02 ratios.
A Further Note on the Distribution
of Silica-Saturation Components in
Cenozoic Basalt
F. Chayes
In view of the lack of information
about — or even definition of — the
"shapes" of distributions of variables in
multicomponent space, argument about
whether a particular projection distorts
an unprojected sample or parent distri-
bution is in principle endless. In prac-
tice, however, the choice of meaningful
projections is usually small, and if a
particular distribution characteristic is
detected in one projection, it is some-
times possible to plot the data in another
that is clearly less favorable to it. Al-
though the failure of a particular prop-
erty to survive such a transformation
does not ipso facto discredit either the
property or the projection in which it
emerged, persistence of the property in
successively less favorable projections is
of obvious interest. When, for instance,
a sample frequency distribution exhibits
more than one maximum in a particular
projection, the projected distance be-
tween maxima, scaled in class widths, is
clearly of critical importance, for as it
decreases, the resolving power of the pro-
jection must also decrease. A strong
tendency toward bimodality may persist
under contractions of this distance suffi-
cient to suppress or eliminate a weak or
adventitious one.
The evidently bimodal distribution of
normative silica-saturation components
in Cenozoic basalt (Chayes, 1972) is a
674
CARNEGIE INSTITUTION
case in point. This was first noted in
plots of sample frequencies in the Tilley-
Muir projection (Tilley and Muir, 1962) ,
which consists of three equilateral tri-
angles, the central one sharing an edge
with each marginal one. Strong maxima
appeared in each marginal triangle near
its shared edge. In the computer program
used to reduce the data, each triangle
was partitioned into 100 subtriangles.
Thus the same projection space and
grouping interval were used for ne-
normative, {ol + %) -normative and
(/-normative basalts; this situation is
very favorable for the detection of max-
ima in the ne and q fields. Whether this
projection characteristic is considered
desirable because it reduces the proba-
bility of failure to detect a real difference
or dangerous because it increases the
probabilitiy of "detecting" differences
that do not exist depends to some extent
on prior conviction about the nature and
origin of basalt. Both are true; as in
many practical situations, power against
error of the first kind varies inversely
with protection against error of the sec-
ond kind.
It was pointed out in the original dis-
cussion (Chayes, 1972, p. 292) that the
bimodal distribution of silica saturation
components in Cenozoic basalt is quite as
evident in the Coombs (1963) projection,
in which both the area allotted to the
ol-di-hy ternary and the distance be-
tween the maxima are much reduced. In
the Coombs projection as actually pro-
posed by Coombs, i.e., with molecular
norms, the distance between the ol-di and
the hy-di joins is halved. It has now
been found that the bimodal character of
the sample distribution survives a fur-
ther halving of this distance. Such a re-
duction results from the substitution of
cationic for molecular norms in the
Coombs projection. The di-hy join then
intersects the ol-q join at ol15 rather than
ol50, and the areas available for project-
ing q-, {ol -j- %)-» and ne-normative
basalts are in the ratio 3:1:4, as opposed
Fig. 142. Distribution of silica saturation
parameters in 3663 analyses of Cenozoic basalts
(including all ankaramite, atlantite, basalt,
basanite, ciminite, crinanite, diabase, dolerite,
hawaiite, lamproite, limburgite, mugearite,
murambite, sakalavite, tephrite, teschenite,
tholeiite, trachybasalt, trachydolerite, vesuvi-
ite) extracted from data base RKSP72. Class
widths are 10% for each of any pair of ternary
components. Each class within contour n con-
tains at least n2 analyses.
to 1:1:1 in the Tilley-Muir and 1:1:2 in
the original Coombs projection.
As Figs. 142 and 143 show, the distri-
bution of silica-saturation components in
Cenozoic basalt retains its markedly bi-
modal character in this clearly unfavor-
able projection. Figure 142 is the con-
toured sample frequency distribution of
saturation parameters in the 3663 Ceno-
zoic basaltic volcanics whose analyses
are contained in the current version of
the data base of RKNFSYS. Norms with
Fig. 143. Data as in Fig. 142 except that 639
analyses with ol > 25 and/or Thornton-Tuttle
index > 45 have been omitted. Grouping and
contouring conventions as in Fig. 142.
GEOPHYSICAL LABORATORY
675
more than 25% ol or a Thornton-Tuttle
index greater than 45, of which there are
639, were excluded from the tally leading
to Fig. 143. As in the original discussion,
the contouring in both diagrams is on a
square-root basis, each unit area lying
within the area bounded by the nth con-
tour containing at least n2 analyses.
Even though in the critical region the
entire distance between the di-ol and
di-hy joins is less than three class widths,
this is enough to show a marked decrease
in frequency with increase in the ratio
ol/hy.
It is to be stressed that the bimodality
found in the distribution of silica-satura-
tion parameters need not extend to other
compositional variables. In particular, it
remains to be demonstrated whether any
essential raw oxide behaves in similar
fashion, and it is clear that most of them
do not. Indeed, if, as suggested, the
silica-saturation state of basalt magma
reflects the influence of pressure on the
initial melt composition, it is conceivable
that the immediate parents of particular
under- and oversaturated basalt magmas
might even be of identical composition.
In any event, the distribution of silica-
saturation components of basalt is gov-
erned far more by the covariances of the
oxides than by their means and vari-
ances.
Multivariate Frequency Information
on Rock Chemical Data
T. N. Irvine
The usual type of frequency informa-
tion obtained from large samples of rock
chemical data consists of either univari-
ate distributions or joint bivariate distri-
butions of variables that relate to one or
more of the constituents for which data
are recorded. For many petrological
problems, however, one would like to
know the relative frequencies of whole-
rock compositions as defined by the inde-
pendent variations of all the constituents
reported in a conventional chemical anal-
ysis. A method for determining this kind
of joint multivariate distribution and
projecting it onto simple graphs has been
developed and is described here with
examples of its application.
Method
A working solution to the basic prob-
lem of determining a joint multivariate
frequency distribution has been achieved
by means of two computer programs,
FREQ and MODES. The problem of
graphing the distribution has been par-
tially resolved by a third program,
PRJECT.
In program FREQ, the range of each
constituent to be treated as a variable is
divided into classes of equal width that
can be identified by rank number ; in this
way the multidimensional chemical space
that contains the analyses is effectively
divided into a series of cells distinguished
by combinations of rank numbers.* The
current version of the program accommo-
dates 14 different constituents. As the
rank numbers are computed, they are
packed into successive parts of a pair of
computer words, which is then entered as
a single unit into a sorting routine that
places it in order with other pairs in a
descending unsigned binary sequence.
When complete, the sequenced list is in-
spected, and by comparison of successive
entries, an inventory of the occupied cells
and the number of analyses that each
contains is prepared and filed in tempo-
rary storage for use in the other pro-
grams.
Program MODES determines the clus-
tering or "modality" of cells having a
specified minimum content of analyses.
A mode is visualized as a continuous
grouping of these cells such that the mid-
point composition of every member is
within one class width of the midpoint of
* This method was suggested by Dr. Chaves,
who has generously given much assistance
and counsel toward its implementation and has
supplied data from his RKNFSYS library for
experimentation.
676
CARNEGIE INSTITUTION
at least one other member in terms of
every constituent. By corollary, the es-
sential distinction between two modes is
that all the cells in one must differ from
all the cells in the other by at least one
class-width unit of at least one
constituent. The program identifies the
different modes and lists the cells they
contain. This information is of interest
in itself and is useful in contouring the
frequency diagrams.
Program PRJECT controls the basic
plotting operations necessary to con-
struct bivariate (ternary) projections of
the multivariate distribution. Ideally,
this can be accomplished by plotting a
series of graphs showing the midpoint
compositions of cells containing different
minimum numbers of analyses. The plot
points on each graph should tend to clus-
ter according to the modality of the dis-
tribution at the cell density level repre-
sented, and lines drawn to encircle the
clusters will represent projected silhou-
ettes of the multidimensional modes and
may be treated as frequency contours. In
practice, more satisfactory contours are
obtained if the actual analyses contained
in the cells are plotted, rather than just
the cell midpoint compositions. More
points are shown so the clusters are easier
to define, and the contours can be drawn
to delimit the modes as they are actually
occupied by data rather than merely en-
circling hypothetical midpoints. After
the projected contours are delineated,
they are transferred to a single graph to
form the frequency diagram.
Discussion
The frequency information obtained
by the above method is numerically very
different from that derived in the conven-
tional way; it may or may not appear
graphically different, depending on the
nature of the distribution. The numerical
differences arise because with multiple
variables the number of class units or
cells potentially available for occupancy
tends to be extremely large, even when
the class widths for the individual vari-
ables are relatively coarse; hence the
number of analyses per occupied cell
tends to be small even in the regions of
maximum concentration. No definite
value can be attached to the number of
available cells, but an impression of its
magnitude may be obtained by consider-
ing the ranges of the variables. If k{ is
the number of class widths necessary to
cover the range of variable i, then for n
independent variables the number of cells
is nominally n k{. If the variables are
normalized (as to 100%), the number is
reduced but is still at least (l/kimax)
n kh where /cimax is the largest value of k{.
Thus for 11 variables, even with as few as
four class widths per variable there could
be as many as 410, or more than a million,
cells into which the data might fall. The
number of possibilities is further reduced
by correlations between variables but
will still be very large unless the correla-
tions are exceptionally strong.
A rather surprising feature of joint
multivariate distributions is that despite
the large number of cells potentially
available for occupancy, the number of
modes in a distribution tends to be small.
The basic reason is that the number of
immediate neighbors around each cell in-
creases exponentially with the number of
variables, the relation being (3n — 1).
Thus in the above example there are
more than a million cells, but each of
these cells has more than 177,000 imme-
diate neighbors, so those that are occu-
pied are likely to be conterminous. In
general, if the variables have class widths
as large as two standard deviations, the
possibility of more than one mode is vir-
tually ruled out. Class widths approach-
ing this size are commonly necessary to
obtain significant distributions when
dealing with 10 or more variables (see
below).
Perhaps the most novel feature of the
frequency diagrams proposed here is that
in their preparation the usual sequence
of constructional operations is essentially
GEOPHYSICAL LABORATORY
677
reversed. In making conventional fre-
quency diagrams, the data are projected
onto the graph first (in principle at least,
if not in reality) , and then their distribu-
tion is counted. In the present method,
the distribution is counted first and then
is itself projected. This difference leads
to some interesting points of contrast. In
conventional frequency diagrams the dis-
tribution shown for a given sample
changes with the projection, not because
it is being viewed from different direc-
tions but because it pertains only to the
variables represented. In the projections
described here, the distribution is always
the same, regardless of the variables
plotted, be they prime variables, liiH,rr
combinations, or ratios, and regardless of
whether the graph is univariate or bi-
variate. The distribution changes in ap-
pearance with the projection but only
because of the change in view. Basic
features such as numbers of modes and
mode density are constant.
A first consideration in applying the
method relates to the choice of variables.
In practice this is likely to be governed
in large degree by the data available, but
a general guideline is to include all varia-
bles that are believed to be essential or
relevant to the problem under investiga-
tion.
A more difficult problem concerns the
choice of class widths. This matter re-
lates to, but is not strictly controlled by,
procedures for testing a sample distribu-
tion for statistical significance. Two such
procedures have been examined. In one,
the incidence of cells containing more
than one analysis is tested by x2 against
Poisson expectation. This test is highly
sensitive, however, and imposes only an
extreme lower limit on the volume of a
cell.
The second procedure stems from a
method originated by Kamb (1959) for
contouring petrofabric diagrams. It re-
quires that the volume of a cell be suffi-
ciently large that <r/E < 1/3, where E
is the average number of analyses per
cell and o- is the standard deviation of the
number that would fall in the cell under
random sampling of an appropriate
model population. There being no defi-
nite number of cells available for occu-
pancy by the total number of analyses,
N, the number of occupied cells, M, is
used instead and only the distribution of
analyses (N — M) is considered. In
these circumstances, E = (N — M)/M
and o- = \/(N — M)/M. The essential
requirement therefore is M < N/10. But
this condition is generally much too re-
strictive for the present problem. For
example, in dealing with 10 variables in
rock groups such as basalt or andesite,
even using samples that are large by cur-
ren oetrographic standards, it would re-
quire cells so large that the subdivision
of the sample would probably be of little
interest. For practical purposes it is use-
ful only to set an upper limit on cell
volume, but it serves to make the impor-
tant point that to obtain highly signifi-
cant, detailed joint frequency distribu-
tions for numerous variables one must
have extremely large samples.
In an empirical approach to the prob-
lem, exploratory work with 10 variables
suggests that relatively stable distribu-
tion patterns are obtained when M <
N/2, and it appears that this condition
can generally be met with class widths
that are substantially less than twice the
univariate standard deviations of the
variables. Such widths are coarse by
usual standards but are of interest.*
With fewer variables, of course, the
widths could be reduced.
It will be noted that the above discus-
sion relates only to the volume of the cell,
not to the dimensions of its sides; one
must still choose the specific combina-
tion of widths to be used. The most per-
plexing aspect of this problem is whether
* Because such large class widths are required,
a modification of program FREQ is planned
whereby the analyses will be counted in over-
lapping cells. This procedure will be analogous
to that by which petrofabric data are counted
and will smooth the frequency data as in a
moving average.
678
CARNEGIE INSTITUTION
minor constituents and constituents of
small variance should be treated in the
same way as major variables. In theory
there is nothing to require this, but in
practice there is probably less chance of
prejudicing the results if the class widths
are related in some systematic way to the
ranges or standard deviations of their re-
spective variables. On the other hand,
the option of making some class widths
much coarser than others may in many
cases be the only way to give a variable
any recognition at all without precluding
the possibility of obtaining a stable dis-
tribution.
Some First Applications
The applicability of the present
method depends on the question posed by
the problem under investigation. If the
question concerns the most frequent
values or relationships of specific varia-
bles in the sample, then more standard
procedures are appropriate. But if it is
of the type, "What are the most frequent
compositions in the sample?" or "What is
the relationship of these variables in the
most frequent compositions?", then
multivariate information of the kind de-
scribed here is probably desirable.
The problem that prompted this study
is one discussed previously by Chayes
(1972) concerning the frequency distri-
bution of the normative silica-saturation
components in basalt. The conventional
approach shows a bimodal distribution
in some projections (see the preceding
section by Chayes) but not in others (for
example, in the Ol'-Ne'-Q' diagram used
by Irvine and Baragar, 1971). It was
TABLE 56. Cell-Content Distributions for Three Groups of Cenozoic Basalts Based
on the Joint Frequencies of 10 Oxide Constituents
Frequency, in
Number of Cells
Number of Analyses
Hawaiian
Submarine
in Cell
Basalts
Basalts
All Other Basalts
A*
A*
A*
B*
1
277
82
2582
1565
2
28
10
152
279
3
11
2
23
86
4
3
5
3
41
5
5
1
2
17
6
1
1
6
7
2
2
5
8
1
5
9
1
1
1
5
10
1
4
11
1
2
12
2
13
1
14
15
1
17
2
19
1
20
1
21
1
23
1
26
1
No. of occupied cells
334
103
2764
2023
No. of analyses
512
156
2992
2992
* Class widths: (A) Si02, A1203, FeO, MgO, CaO, 2 wt %; Fe203, 1.5%; Na20, K20, Ti02, 1%;
P205, 0.5%. (B) Si02, A1203, FeO, MgO, CaO, 3 wt %; Fe203, 2.25%; Na20, K20, Ti02, 1.5%;
P205, 0.75%.
GEOPHYSICAL LABORATORY
or
SUBMARINE BASALTS
679
ALL OTHER BASALTS
Opx
Cationic norms
Fig. 144. Normative projections of the joint frequency distributions of 10 oxides in three groups
of Cenozoic basalt. Contours are in numbers of analyses per cell, based on data and class widths
summarized in Table 56.
considered that the joint frequency dis-
tribution of all the constituents that are
ordinarily used in calculating the norm
might yield a more definitive result.
The investigation was based on the
same sample of 3660 analyses of basaltic
rocks described in the previous section of
this Report, but it was found expedient
to divide the sample into three parts —
Hawaiian, submarine, and all other ba-
salts. The results, which are based on
10 oxides, are summarized in Table 56
and in Ol'-Ne'-Q' projections in Fig. 144.
Averages of the analyses falling in the
most populous cells for the Hawaiian and
submarine basalts are listed in Table 57.
The class widths denoted A in Table 56
are roughly one standard deviation for
the oxides in the Hawaiian and sub-
marine basalt groups. They were chosen
as being about the largest values that
would give useful subdivision of each
variable and were accepted because they
revealed interesting frequency informa-
tion. It was obvious, however, that they
were not broad enough for the group "all
other basalts," so for this group each
width was increased by half, as in B.
The frequency distributions defined for
the Hawaiian and submarine basalts are
outstanding in showing pronounced uni-
modal concentrations of analyses. Al-
though the 512 analyses from Hawaii are
distributed among 334 cells, 109, or about
21%, are contained in only 10 cells con-
terminous in one mode. For the sub-
marine basalts 28 of 156 analyses, or
about 18%, fall in 4 mutually neighbor-
ing cells, and 48 of the analyses fall in 9
conterminous cells containing at least 4
analyses each. Considering that the
probability of a cell containing more than
one analysis is very small, these modes
are tremendous concentrations of data,
a point that is underlined by the fact that
no comparable mode was found among
all the 2992 analyses of all other basalts.
(By the same class widths, A, the domi-
nant mode shown by the other basalts is
formed by only 2 cells containing only 9
and 4 analyses!) This was the prime
reason for isolating the Hawaiian and
submarine basalts: If they had been
combined with the others, they would
dominate any distribution pattern, no
matter what the class widths, and so
might tend to obscure additional features
of interest. A second reason is that the
Hawaiian and submarine basalt mode
compositions must obviously be impor-
tant in themselves.
As regards the original problem, that
of the general modality of basalt, the
most pertinent frequency distribution is
680
CARNEGIE
INSTITUTION
TABLE 57.
Commonest Compositions
among
Hawaiian and Submarine Basalts Based
on the Joint Frequencies of 10 Oxides *
No. of
Analyses
Si02
A1203
Fe203
FeO
MgO
CaO
Na20
K20
Ti02
P205
Hawaiian Basalts
8
52.20
14.27
1.82
9.25
6.79
10.49
2.34
0.44
2.14
0.25
9
52.13
14.24
1.36
9.78
6.92
10.50
2.33
0.39
2.13
0.23
4
50.41
14.90
4.91
7.15
7.02
10.50
2.22
0.30
2.29
0.29
5
50.63
14.89
3.24
8.32
7.19
10.57
2.15
0.36
2.40
0.26
7
50.95
15.11
3.68
7.53
7.02
10.65
2.11
0.34
2.33
0.28
4
50.91
15.23
3.81
7.49
7.39
9.90
2.21
0.35
2.44
0.28
23
50.97
14.44
2.13
9.11
7.06
10.84
2.34
0.43
2.45
0.25
5
51.08
14.31
1.30
9.54
7.35
11.16
2.32
0.42
2.32
0.21
13
51.40
13.73
2.00
10.65
5.39
9.31
2.78
0.76
3.57
0.40
10
50.47
13.30
2.29
9.23
8.58
10.70
2.20
0.45
2.53
0.26
5
51.64
13.66
2.27
9.00
8.46
10.24
2.25
0.35
1.92
0.20
11
50.66
13.65
2.35
9.53
6.64
10.57
2.47
0.59
3.25
0.30
15
51.15
13.79
2.14
9.18
7.49
10.86
2.31
0.44
2.40
0.24
7
50.91
13.80
1.22
10.28
7.28
10.83
2.33
0.48
2.64
0.24
5
50.98
13.60
1.35
9.73
8.45
10.69
2.22
0.42
2.32
0.23
6
50.79
13.80
1.31
9.84
7.62
11.17
2.26
0.46
2.49
0.26
4
49.07
15.29
3.79
8.27
7.14
10.93
2.20
0.34
2.70
0.27
5
49.82
13.38
1.36
9.87
8.89
11.10
2.20
0.49
2.63
0.24
Submarine Basalts
4
50.76
17.31
3.87
4.87
7.10
11.27
3.21
0.23
1.26
0.13
4
50.29
16.17
1.68
6.85
8.97
11.58
2.70
0.30
1.31
0.15
5
50.67
16.79
1.86
7.21
7.39
11.58
2.88
0.20
1.33
0.10
7
50.95
17.07
1.05
7.59
7.47
11.41
2.87
0.14
1.33
0.11
9
50.23
14.97
1.93
8.62
8.43
11.05
2.81
0.17
1.64
0.14
4
50.69
15.20
1.95
9.09
7.33
10.92
2.84
0.16
1.67
0.14
7
50.37
15.66
1.25
8.59
8.67
10.76
2.80
0.16
1.60
0.14
4
51.07
14.87
1.46
9.39
7.51
10.94
2.82
0.12
1.67
0.15
4
49.22
16.25
2.10
6.71
9.33
11.43
2.80
0.50
1.48
0.18
* The data are averages for cells containing four or more analyses and are sequenced according
to their rank-number combinations; class widths as in Table 56.
that for the group "all other basalts,"
illustrated in Fig. 144. This distribution
has been shown by means of program
MODES to be essentially unimodal.
Somewhat larger class widths possibly
should have been used, but the number
of modes cannot be increased by in-
creasing cell volume; the only change
that might be effected is to shift the peak
(or, more properly, the core) of the mode
from the quartz-normative region where
it is presently located to the nepheline-
normative region. Rearranging the class
widths might introduce a second mode,
but from general inspection of the data
this possibility seems unlikely.
One of the principal uses anticipated
for the method is to summarize the chem-
ical characteristics of various rock types
or suites and possibly of certain mineral
groups, such as the amphiboles. The
summary of data on Hawaiian and sub-
marine basalts in Table 57 is an example.
This relatively brief listing is an abstract
of 668 analyses showing the most fre-
quent combinations of 10 different oxides.
Such tables are convenient for compara-
tive purposes and should be useful in
other ways, for example in selecting rock
compositions for experimental study. The
technique of identifying an analysis by a
rank-number combination is also the
basis of a method for searching a large
file of data for compositions of a par-
ticular type.
Another potential application is in the
GEOPHYSICAL LABORATORY
681
correlation of chemical variations in vol-
canic rocks with experimentally deter-
mined liquidus boundaries. For the
Hawaiian basalt data at least, it can be
shown that the most common composi-
tions represent magma that was rela-
tively free of suspended crystals. The
method may also prove useful for select-
ing the most nearly primary compositions
from files of analyses of volcanic rocks
that have been slightly altered through
metamorphism.
Determining Null Values of Ratio
Correlation by Simulation
F. Chayes
Ratios and relations between ratios are
of paramount importance to the penolo-
gist (for a review, see Chayes, 1949) , and
procedures for calculating first-order ap-
proximations of simple ratio correlations
are well known (see, for instance, Chayes,
1971, pp. 11-24). Any such expression
can be transformed into one in which the
only terms are the coefficients of varia-
tion of the numerator (s) and denomina-
tor (s), as in the original formulation of
Pearson.
If, for instance, Yt = X\/X2 and Yj =
Xlf where the parent parameters of the
Xls are fii, /x2, o-i2, a22, and pi2 = 0, the
null correlation between the Y's given by
the "8 method" is
Pi) = /A2CTl/ (jU.22(Tl2 4" fXl2(T22)1/2j (1)
and division of numerator and denomina-
tor of the right side by mfio transforms
this to
wsd/tdHW)172,
(2)
where Cn = an/ixn, n = 1, 2, is the parent
value of the coefficient of variation of Xn.
The approximation is adequate only if
the Cn are sufficiently small that second
and higher order powers of them are neg-
ligible, and in much geochemical work
this is not so. Simulation experiments
readily define what is meant here by
"sufficiently small" in a practical sense
and may also be used to generate appro-
priate null values of p when one or more
of the coefficients of variation of the raw
data are large enough to make equation 1
or its analogues for other forms of ratio
correlation unreliable. A computer pro-
gram suitable for experimental work of
this sort was developed for a National Sci-
ence Foundation Institute on Geostatis-
tics held at Chicago Circle in June of
1972 and will be described in the proceed-
ings of the Institute.*
The program generates a set of four
random numbers from a parent popula-
tion uniformly distributed in the range
(0,1), transforms these to pseudonormal
deviates with zero mean and unit vari-
ance, and adjusts each with the mean and
variance assigned it at operation time to
produce the current set of "observed
values" of Xa, Xb, Xc, Xd. From these,
all possible simple ratios are formed; the
elements of the vector of terms, ratios,
squares, and cross products are stored in
cumulators; and the process is repeated
until the requested number of items, i.e.,
of sets of "observed values," has been sup-
plied and processed. The covariance ma-
trix is then computed from the cumulated
sums, sums of squares, and sums of cross
products, and the requested correlations
are printed out. Since the means and vari-
ances of the "observed variables" are
assigned separately at operation time,
the effect of any combination of coeffi-
cients of variation on simple ratio corre-
lation may be evaluated experimentally.
Lacking specific instruction to the con-
trary, the program generates the "ob-
served variables" from parents that are
uncorrected. The user may, however,
assign correlation (s) of arbitrary size
and sign to any pair, or to any two mu-
tually exclusive pairs, of these variables.
* Listings can be obtained either from the
writer or from Professor Richard McCammon.
Department of Geology, University of Illinois
at Chicago Circle, Box 4348, Chicago, Illinois
60680.
682
CARNEGIE INSTITUTION
Thus by specifying
/*a — t*c — M ¥= Pb
<7a — 0"c = °"d v2^ °"6
Pcd¥=®
and by requesting sample statistics for
correlations between ratios of common
denominator, he can evaluate the effect
of correlation between the numerators on
the Pearsonian "spurious" correlation
when numerators and denominators have
coefficients of variation C and Cb, re-
spectively.
The present program is suitable for use
in connection with major constituents in
which C « 1. It contains no protection
against negative "observed values" and
so does not permit realistic modeling of
the J-shaped distributions characteristic
of trace constituents and certain minor
elements, in which, even though no nega-
tive values occur, o- > ft, so that C > 1.
Preliminary experimentation with this
program confirms an earlier suggestion
(Chayes, 1971) that for ratios formed
from uncorrelated terms homogeneous in
C, linear approximations of p are good if
C < 0.1 and are still fair if C < 0.15.
For 0.15 < C < 0.35, however, the simu-
lated ratio correlations differ widely both
from the relevant first-order approxima-
tions and from zero; in this range either
higher order approximation or experi-
mental evaluation of null values against
which to test observed simple ratio corre-
lations will be essential. Values of C in
this troublesome range are often en-
countered in essential element distribu-
tions.
BIOGEOCHEMISTRY
A Comparison of Melanoidin
and Humic Acid
T. C. Hoering
Humic acid is defined as the mixture of
organic matter that can be extracted
from a sediment or a soil by dilute alka-
line solutions and then precipitated by
acidifying the extract. This complex ma-
terial represents a significant fraction of
the abundant polymeric organic matter
of high molecular weight contained in
sedimentary rocks. Humic acid was
chosen for the study of this major reser-
voir of organic matter because it can be
isolated readily from the inorganic min-
erals with which it is associated. Al-
though it has been studied for many
years, its molecular constitution is not
known. It bears no resemblance to the
components of living organisms from
which it is formed.
The lignin from woody plants has been
considered by many workers to be a
major precursor of humic acid. Waks-
man (1938) proposed the ligno-protein
hypothesis to account for organically
bound nitrogen, and this concept has
formed the basis for most studies of its
structure. This hypothesis presents a
number of difficulties, however, in ac-
counting for the occurrence and prop-
erties of natural humic acid (Stevenson
and Butler, 1969).
Maillard (1913) was one of the first to
note the similarities between the prop-
erties of humic acid and the products of
the condensation of sugars and amino
acids, which he called melanoidin. En-
ders (1943) used more sophisticated
chemical techniques to point out other
similarities. Abelson and Hare (Year
Book 68, pp. 297-303; Year Book 69, pp.
327-334) expanded on this approach to
the humic acid problem. Modern meth-
ods of chemical instrumentation were
applied to the analysis of products. Syn-
thetic melanoidins were prepared, and a
remarkable parallelism was noted be-
tween the rates of reaction of pure amino
acids with synthetic and natural humic
acid. A striking similarity between the
electron-spin-resonance spectra of the
two materials was observed, and further-
more the changes in these spectra upon
treatment with reducing agents and with
GEOPHYSICAL LABORATORY
683
molecular oxygen were identical. These
experiments established the extremely
reactive character of humic acid and
strengthened the belief that melanoidin
formed an essential part of the natural
material. The experiments described be-
low extend this comparison to a much
wider range of properties and support
the melanoidin concept for the origin of
humic acid.
Experimental
The humic acid studied in this work
was isolated from the humate, cemented
sands of northwest Florida. The geolog-
ical occurrence and chemical aspects of
these deposits have been described by
Swanson and Palacas (1965). They in-
terpreted the material as soil organic
matter that has migrated downward with
water into pure quartz sands. A change
of acidity caused the precipitation of
humic acid and tightly cemented the
blocks of sand. The organic matter be-
haves in all respects like humic acid.
Little clay or silt is present, and large
quantities of material with a low ash
content can be isolated readily. The pro-
cedure of R. G. Malcolm (personal com-
munication, 1972) was used and consisted
of the following steps: extraction with
sodium pyrophosphate, acidification and
centrifugation, solution in dilute sodium
hydroxide and high-speed centrifugation,
filtration through Millipore filters with
0.45 jam pores, dialysis through a collo-
dion membrane against water for 2 days,
passage through columns of ion-exchange
resins saturated with hydroxide or hy-
drogen ions, and freeze-drying. Soft,
fluffy powders are obtained, which have
less than 0.5% ash and are readily solu-
ble in dilute alkalies.
Melanoidins were synthesized by re-
acting d-glucose and amino acids at a
12:1 weight ratio in solutions buffered to
pH 8. In some cases, an excess of solid
calcium carbonate was used to hold the
acidity constant. Individual prepara-
tions were made with a single amino acid
TABLE 58. Elemental Composition of
Melanoidin and Humic Acid, wt %
Melanoidin*
Humic
Acidt
Florida
HumateJ
C
H
N
O
54.73
5.17
5.04
35.06
46.2-64.8
4.1-5.9
0.7-5.1
31.8-47.3
55.0
4.4
1.4
38.5
* Manskaya and Drozdova, 1968.
t Schnitzer and Khan, 1972.
| Swanson and Palacas, 1965.
such as glutamic acid, lysine, or tyrosine
or with an equimolar mixture of glycine,
alanine, glutamic acid, lysine, and leu-
cine. The temperature of the reactants
was held just below the boiling point in
order to minimize foaming. The mixture
turned bright yellow within an hour after
reaching temperature and after a few
hours was dark colored. It was held at
temperature for 4 days, after which it
was cooled and acidified with hydro-
chloric acid. The gelatinous precipitate
was centrifuged and washed until free of
chloride ion. It was dialyzed for several
days and then freeze-dried. A brown
powder resembling the humic acid prep-
arations was obtained.
Freshly precipitated humic acid and
melanoidin form gels that have similar
appearance, color, texture, and solubility
in dilute alkali. These gels bind great
quantities of water and undergo a
twenty-fold reduction in volume when
dried at 110°C. Table 58 shows that
melanoidin and a wide range of humic
acid have comparable elemental abun-
dances, although melanoidin tends to
have a higher nitrogen content. Table 59
TABLE 59. Organic Functional Groups in
Humic Acid and Melanoidin
Humic
Melanoidin Acid
Carbonyl group, wt %*
Phenolic hydroxyl, wt %*
Total hydroxyl, wt %*
Total acidity, meq/gf
2.6
6.53
9.6
7.2
2.4
6.90
8.6
6.7
* Enders and Theis, 1938.
f This work.
684
CARNEGIE INSTITUTION
shows that the two have similar contents
of organic functional groups that can be
determined by simple volumetric meth-
ods. The infrared absorption spectra of
the pair, shown in Fig. 145, are domi-
nated by a strong band at 2.9 /xm due to
hydrogen-bonded hydroxyl groups. The
other broad unresolved bands are indica-
tive of a wide range of functional groups,
including carboxyl and carbonyl. The
spectra have a general similarity, sup-
porting the data of Table 59.
Humic acid binds metallic ions
strongly, presumably through ionic bonds
with carboxyl groups and through coor-
dinate bonding with hydroxyl or carbonyi
groups. The following experiments show
that melanoidin behaves likewise. Humic
acid and melanoidin preparations were
stirred at room temperature for 12 hours
with an excess of 0.1 molar Fe2(S04)3.
The mixtures were centrifuged and
washed with water until the supernatant
solution was free of iron. The solids were
dialyzed for several days and then freeze-
dried. The humic acid, which had ini-
tially less than 0.35 wt % iron, contained
4.2%. The melanoidin had 1.8 wt %
iron. The Mossbauer absorption spectra
using 57Fe radiation were measured by
100
£=
90
O
"co
80
CO
70
F
c/">
HO
1_
50
-\ —
.
40
r
CD
M)
<^>
i_
20
CI)
LL
10
0,
^iW-*-^
\ A
JIAT Id «n .
1/
Humic acid_
3 4 5 6 7 8 9
Wavelength in micrometers
10
Melanoidin
"I 1 1 1 r
4 5 6 7 8
Wavelength in micrometers
10
Fig. 145. The infrared absorption spectra of melanoidin and humic acid. The spectra were
measured on a 0.5% organic substance in a potassium bromide pellet. The melanoidin was syn-
thesized from glucose and a mixture of five aliphatic amino acids.
GEOPHYSICAL LABORATORY
685
IRON IN MELANOIDIN
IRON IN HUMIC ACID
.000
\.
-4.2
.0 2.1
VELOCITY mm/sec
4.2
.000
• ft & c-
/ ; ■
.010
> •
•; $ \:
020
1; '
•
— 1 1 1 — 1
-4.2
21
.0 2.1
VELOCITY mm/sec
4.2
Fig. 146. The Mossbauer absorption spectra of the iron bonded into humic acid and melanoidin.
The melanoidin was prepared from glucose and glutamic acid.
Dr. David Virgo of this Laboratory, and
the results are shown in Fig. 146. The
doublet at —0.156 and +2.182 cm/sec
shows clearly that divalent iron is bonded
into both the humic acid and the melanoi-
din, although trivalent iron was used as
a starting material. The materials are
reducing agents that reduce ferric anions
and then form complexes. The shoulder
at +0.591 cm/sec is due to approxi-
mately 5% bonded ferric iron. This ab-
sorption peak has its corresponding mem-
ber of a doublet hidden under the large
adsorption at — 0J56. The positions of
the peaks due to ferrous iron correspond
to complex ions having a large degree of
ionic bonding. The peaks are broader
than those given by ferrous iron in crys-
talline solids and are probably the result
of a variety of binding sites in the or-
ganic complex.
During the course of the reaction of
amino acids with glucose to produce
melanoidin, a number of volatile com-
pounds of low molecular weight are pro-
duced. In particular, when phenylalanine
was reacted with glucose in an inert
atmosphere, a 10% yield of phenylace-
taldehyde was obtained. When the reac-
tion was carried out in air, a 20% yield
of phenylacetic acid resulted. The path-
way of this reaction will be discussed
later. When natural humic acid was
refluxed with phenylalanine, the same
reaction occurred, with a similar yield of
the aldehyde or acid. The products were
identified by mass spectrometry, infrared
spectrometry, and gas chromatographic
retention times as compared with authen-
tic compounds.
The classical method for determining
structural components of natural and
synthetic organic polymers is through
chemical degradation into smaller, recog-
nizable molecules. In some cases, major
structural features of the polymer can be
seen by fitting the pieces back together.
This approach has not been particularly
successful for humic acid. Either the de-
gradative reactions do not proceed far
enough to give molecules that can be
identified, or they proceed too far and
give very simple substances with little
structural information. Most attempts
described in the literature have given low
yields of complex mixtures. Indeed, the
slow progress toward understanding the
structure of humic acid is due in part to
686
CARNEGIE INSTITUTION
this difficulty. Therefore, several new
approaches to the chemical degradation
of humic acid and melanoidin were at-
tempted.
When the materials are treated with
elemental chlorine in water solution, a
vigorous reaction occurs. Some hydrogen
chloride is released, and the mixture soon
bleaches and becomes bright yellow.
When the clear solution is made alkaline
and more chlorine is added to form the
mild oxidant sodium hypochlorite, a sub-
stantial yield of chlorinated organic acids
is made from the synthetic and natural
materials after a short reflux period. Ap-
proximately 30% of the starting poly-
mers are converted to acids soluble in
ethyl ether. Methyl esters were pre-
pared as derivatives of the acids, and the
mixture was separated by silica-gel
chromatography using hexane-ethyl ether
eluants of increasing polarity. Figure
147 shows gas chromatograms of the
esters eluted by 20% ether from each.
The mass spectra of the individual com-
pounds giving rise to the peaks in the
chromatograms were measured. In the
case of the eighteen corresponding peaks
numbered in Fig. 147, the mass spectra
were identical or were due to homologues
of each other. An inspection of the mass
spectra shows the compounds to be
methyl esters of chlorinated aliphatic or
heterocyclic acids. Further identification
of the esters is proceeding by means of
the mass spectra and infrared spectra to
determine their significance in terms of
the structure of humic acid and melanoi-
din.
Melanoidin and humic acid are reac-
tive materials. Many of the degradative
procedures described in the literature
have been drastic, and it is no surprise
that low yields of simple molecules are
obtained. An attempt was made to pro-
tect and stabilize the polymers prior to
degradation. They were reacted with
dimethyl sulfate in alkaline solution to
convert hydroxyl groups to methyl ethers,
which are considerably more stable upon
oxidative attack. There is reason to sus-
pect that the polymers contain carbon-
to-carbon double bonds. This point in a
molecule is particularly susceptible to de-
gradation. The methylated material was
hydrogenated by 3% sodium amalgam in
alkaline solution under an inert atmos-
phere. Both substances went through a
similar series of color changes, starting
with black, which turned to red, then deep
yellow, and finally bright lemon yellow. On
acidification, the black polymers changed
to light gray. The reduced materials
were then oxidized by chromic acid in
hot 3 molar sulfuric acid. About 20% of
the starting material was converted to
ether-soluble acids. These were con-
verted to their methyl esters and given a
preliminary separation by silica-gel-
column chromatography. Figure 148
shows gas chromatograms of the esters
eluted by 20% ethyl ether in hexane and
gives the identification of several of the
peaks. The presence of branched chain
dicarboxylic acid in the products from
melanoidin is especially interesting be-
cause this sample had been prepared from
glucose and glutamic acid, both of which
contain only straight carbon chains. It is
unlikely that the chemical treatment
would produce carbon branching as an
artifact. It is more likely that some of
the cross-linking of the polymer is due to
carbon-to-carbon bonding. This is the
first evidence that such bonding occurs.
Humic acid and melanoidin from glu-
cose and glutamic acid were oxidized by
copper oxide in strong alkali at 170 °C for
3 hours, and the resulting acids were ex-
tracted. Their trimethylsilyl derivatives
were prepared and separated by gas
chromatography, and their mass spectra
were measured. The differences in the
suites of products obtained were signifi-
cant. The humic acid yielded small
amounts of aromatic (benzenoid) acids
and phenolic acids. These compounds
were missing in the product from mela-
noidin.
GEOPHYSICAL LABORATORY
687
Chlorination Products of
Melanoidin from Glucose and Glycine
50°
Column Temperature
CD
CO
d
o
a.
CO
or
o
a>
a3
O
Chlorination Products
of Humic Acid
50 '
Column Temperature
160'
Fig. 147. Gas chromatogram of methyl esters of chlorinated acids produced by chlorination of
melanoidin and humic acid. The fraction of the esters eluted from silica gel by 20 volume %
ethyl ether in hexane is shown. A 100-foot by 0.010-inch capillary column coated with DEGS sub-
strate was temperature-programmed from 75° to 150°C and then held isothermally. The mela-
noidin was prepared from glucose and glycine.
688
CARNEGIE INSTITUTION
Product of Chromic Acid
Oxidation of Humic Acid
Column Temperature
CD
(f)
C
o
a.
en
CD
or
o
CD
Product of Chromic Acid
Oxidation of Melanoidin
75c
Column Temperature
220*
Fig. 148. The gas chromatogram of methyl esters produced by the chromic acid oxidation of
methylated and reduced humic acid or melanoidin. The fraction of the esters eluted from silica
gel by 20% ethyl ether in hexane is shown. A 100-foot by 0.010-inch capillary column coated with
SP-1000 substrate was temperature-programmed from 75° to 220 °C at a rate of 4° per minute and
then held isothermally. The melanoidin was prepared from glucose and glutamic acid.
GEOPHYSICAL LABORATORY
689
This difference is important because
such aromatic and phenolic material has
been interpreted by some as being due to
lignin residues contained in humic acid.
The problem was studied in the following
manner. Samples of melanoidin were
sealed in glass tubes under an inert at-
mosphere and heated to 170° and 220 °C
for periods up to 3 days. The heated ma-
terial was then oxidized with copper
oxide as before. Benzenoid acids could
now be detected, but the yield of phenolic
acids was small. It is conceivable that
some aromatic structures in humic acid
are produced in the early stages of the
diagenesis of organic matter, and do not
necessarily represent features of starting
compounds.
Discussion
The number of comparable properties
between humic acid and synthetic me-
lanoidin is large, and it seems likely that
at least a part of the natural material has
originated from the reaction of a carbo-
hydrate and an amino acid. If this is
true, then it is possible to gain new in-
sight into the structure and reactions of
humic acid.
The chemistry of melanoidin forma-
tion has been studied intensively by food
technologists. The browning of dehy-
drated foodstuffs is caused by such reac-
tions. Some of the color, flavor, and
foaming of beer is due to melanoidin.
Hodge (1953) has integrated a large
number of facts about the material into
a consistent mechanism. A simplified
outline of his mechanism for the Mail-
lard reaction, which leads to melanoidin
formation, is shown in Fig. 149.
Several key points bear on the relation-
ship of melanoidin to geochemically sig-
nificant materials. The first is the irre-
versible Amadori rearrangement, which
converts two of the most abundant con-
stituents of living organisms, amino acids
and carbohydrates, into rearranged prod-
ucts. The products are probably not
metabolized as readily as their precurs-
Initial Stage of Maillard Reaction
0
C-H
H-C-OH
HO-C-H
H-C-OH
H-C-OH
H-C-OH
i
H
0
H 9_0H
HN-C-H
+ R
N-SUBSTITUTED
Glycosylamine
Amadori
Rearrangement
Intermediate Stage of Maillard Reaction
l-Amino- l-Deoxy- 2-Ketose
Dehydration
HC — CH
II M H
HC C-C=0
V
Furfural
Fission
Acetol
Pyruvaldehyde
Diacetyl
etc.
Dehydration
0
C-OH
d-0
I
CH
n
CH
i
c=o
C-OH
ii
0
Reductone
Final Stage of Maillard Reaction
Interreaction and Polymerization
of Decomposition Products and
Starting Reactants
T
Melanoidins
Dark Nitrogenous Polymers and
Copolymers
Fig. 149. An outline of the pathway for the
Maillard reaction to produce melanoidin.
Adapted from Hodge (1953).
ors. Organic matter is effectively re-
moved from the biological carbon cycle
and preserved.
The rearrangement product rapidly
degrades into a very wide range of reac-
tive substances, which continue to react
with each other and with starting reac-
tants. Many of the products of this
second step of the Maillard reaction are
known to polymerize to dark solids. Pre-
690
CARNEGIE INSTITUTION
sumably the polymer contains structural
elements based on furfural, reductones,
and the fission products. This may ac-
count for the lack of success of many of
the chemical degradation procedures that
have been applied in the past on humic
acid. Such structures are reactive and
would certainly be degraded into prod-
ucts with only a few carbon atoms. The
model indicates clearly that milder, more
specific degradation reactions will have
to be applied before much insight is
gained into the molecular structure of
humic acid.
The remarkable reaction of humic acid
with amino acids to yield aldehydes can
now be clarified. This reaction is similar
to that caused by the reagent ninhydrin,
which is specific to amino acids. It occurs
only with molecules containing conju-
gated or activated carbonyl groups. The
reductones shown in Fig. 149 are exam-
ples. They are known constituents of
melanoidins and hence are most likely
present in humic acid.
The melanoidin model predicts that
humic acid should have a structural com-
ponent based on furfural or on the furan
ring. This five-membered, heterocyclic
ring has some stability under proper con-
ditions and should persist. No definite
evidence for its presence in humic acid
has been obtained. Clearly, future re-
search should focus on this problem.
The approach in this work assumes
that melanoidin and humic acid are typ-
ical precursors of the polymeric organic
matter in sedimentary rocks. This as-
sumption is obviously an oversimplifica-
tion, since such polymers contain struc-
tural elements that have been derived
from the lipid fraction of living orga-
nisms. Ogner and Schnitzer (1970) have
shown that humic acid is an efficient
scavenger for hydrophobic molecules
such as lipids, presumably because of its
great surface area and its loose network
structure, which is responsible for gel
formation.
Separation of Amino Acid Optical
Isomers by Gas Chromatography
P. E. Hare and T. C. Hoering
All the common amino acids (except
glycine) found in proteins contain at
least one asymmetric carbon atom and
can therefore exist in one of two config-
urations that are mirror images or optical
isomers. Designated as the d- and l-
amino acid isomers, they are not separa-
ble, in general, by chromatographic tech-
niques.
There are four common amino acids
that contain a second asymmetric carbon
atom. In addition to the d- and L-amino
acid isomers there are two isomers,
termed diastereoisomers, that are not
mirror images. Diastereoisomers (termed
allo-amino acids) have different physical
properties and can generally be resolved
by chromatographic techniques. Success
has been achieved in resolving the dia-
stereoisomers of isoleucine, hydroxypro-
line, threonine, and hydroxy lysine by
ion-exchange chromatography.
Proteins in living organisms consist al-
most exclusively of L-amino acids. Race-
mization is the process by which an op-
tically active form (e.g., L-amino acid
isomer) is transformed into an optically
inactive form (e.g., an equal mixture of
l- and D-amino acid isomers). The term
epimerization is used for the transforma-
tion of diastereoisomers.
An earlier study (Hare and Abelson,
Year Book 66; Hare, 1969) used specific
enzymes to determine the d- and L-amino
acids present in some fossil materials.
The method was based on the difference
of amino acid concentrations before and
after enzyme treatment and did not
measure the proportion directly.
The present study evaluates a tech-
nique developed by Pollock and Oyama
(1966) that involves the formation of
amino acid derivatives with two asym-
metric carbon atoms. The technique
yields a direct measurement of the pro-
portions of the d- and L-amino acid iso-
GEOPHYSICAL LABORATORY
691
mers present. In principle it can be
applied to all amino acids, but in prac-
tice difficulties are encountered with the
hydroxyl amino acids (threonine, serine,
and tyrosine) and the basic amino acids
(lysine, histidine, and arginine) .
To resolve optical isomers of amino
acids with only a single center of asym-
metry from each other, it is necessary to
introduce a second center by making a
suitable derivative. Pollock and Oyama
(1966) succeeded in using an optically
active alcohol (+)-2-butanol that will
react with the carboxyl group of amino
acids to form a stable ester. Trifluoroace-
tic anhydride is used to react with the
amino group to form the volatile tri-
fluoroacetate derivative. Because the
amino acid derivative has an asymmetric
carbon atom in the ( + )-2-butanol (des-
ignated d), the d- and L-optical isomers
of the amino acid itself will form dia-
stereoisomers similar in principle to iso-
leucine and alloisoleucine. The L-amino
acid isomer will form an LD-diastereoiso-
mer with the D-butanol whereas a d-
amino acid isomer will form a DD-dia-
stereoisomer that can be separated from
the LD-isomer, since they are not mirror
images. A mixture of d- and L-amino
acid isomers will form both dd- and ld-
isomers in proportion to their original
composition.
The procedure for preparing the de-
rivatives is modified from that of Kven-
volden, Peterson, and Pollock (1972).
The sample is dried under a stream of
dry nitrogen at 50 °C. Methylene chlo-
ride is added to form an azeotropic mix-
ture with any traces of water as it is
taken to dryness. Approximately 250 /xl
of anhydrous ( + )-2-butanol, 4 I in
HC1, is added to the amino acid hydro-
chloride, and the vial, sealed with a
Teflon-lined cap, is heated for IV2 hours.
The excess butanol-HCl is then evapo-
rated under a stream of nitrogen at 50 °C,
250 /xl of trifluoroacetic anhydride (25%
by volume in methylene chloride) is
added, and the capped vial is heated for
5-10 minutes at 100 °C. The mixture is
cooled in an ice bath before opening and
taken just to dryness with a stream of
nitrogen at 0°C. The residue is dissolved
in methylene chloride and stored at
— 20 °C for later use.
Table 60 shows the results when the
derivatives of standard amino acid mix-
tures are analyzed for the proportion of
d- and L-isomers. The values in column 1
indicate that in a standard of pure l-
amino acids, 1.3 to 2.0% D-isomer is pres-
ent, owing to the small fraction of ( — ) -2-
butanol present as an impurity. Column
2 shows the effect of heating the L-amino
acids to 110° C for 22 hours in 6 N HC1, a
TABLE 60. Percentage of D-Amino Acid Isomers in Standard Mixtures of
L-Amino Acids with Various Treatments
1
2
3
4
5
6
L-Stand-
7
L-Stand-
L-Stand-
ard,
L-Stand-
L-Stand-
ard,
ard,
3 days,
L-Stand-
ard,
ard, 7 hr,
1 day,
3 days,
161°C,
d- and Li-
Amino Acid
ard
hydrolyzed
161°C
161°C
161°C
hydrolyzed
Standard
Aspartic acid
1.4
3.1
46.2
49.7
49.5
Glutamic acid
1.4
2.3
5.9
13.7
28.4
29.3
49.7
Proline
1.5
3.2
46.0
50.2
50.1
49.5
Phenylalanine
1.7
1.8
32.2
49.9
49.5
50.1
50.3
Alanine
2.0
2.5
24.8
42.1
49.7
49.0
49.5
Valine
2.0
2.2
17.0
37.6
49.8
50.2
50.2
Leucine
1.3
2.1
21.3
42.9
50.1
49.9
49.7
Isoleucine
2.0
2.3
18.2*
46.8*
57.7*
57.5*
* D-alloisoleucine.
692
CARNEGIE INSTITUTION
procedure used routinely in the hydroly-
sis of peptides and proteins. There is an
increase in the D-amino acid content of
0.1% for phenylalanine to 1.7% for
aspartic acid and proline. In columns 3,
4, and 5 are the results of heating an
L-amino acid standard at pH 8.5 at
161 °C for 7 hours, 1 day, and 3 days,
respectively. L-aspartic acid and L-pro-
line are racemized relatively rapidly,
whereas L-glutamic acid, which is similar
chemically to aspartic acid, racemizes at
the slowest rate. In 3 days at 161 °C
glutamic acid reached only 28% d-
isomer, whereas each of the other amino
acids was completely racemic (50% d-
and 50% L-isomers). Glutamic acid
under slightly alkaline conditions forms
a lactam, pyroglutamic acid, stabilizing
it against both decomposition and race-
mization. The hydrolyzed aliquot of the
3-day sample in column 6 shows insig-
nificant changes in the D-amino acid
proportions compared with the unhydro-
lyzed sample. Column 7 shows the re-
sults for a standard mixture of six race-
mic amino acids that are known to have
equal d and l contents. The maximum
deviation from the theoretical 50% is
±0.5% (Fig. 150).
Table 61 lists the results for five sam-
ples of organic geochemical interest.
When a sample of shell material from a
recent (—15 years) Mercenaria is pre-
pared (including hydrolysis), significant
amounts of D-amino acids are present,
ranging from 1.5% D-alloisoleucine to
8.7% D-aspartic acid. Most of the d-
amino acids probably come from the
hydrolysis step of peptide-bound mate-
rial, although some may be from the
natural racemization-hydrolysis proc-
esses operative for 15 years. A 1000-
year-old (±200 years, C-14 date) speci-
men of Mercenaria shows a substantial
increase in D-amino acids. D-valine is
present in the smallest amount (4.2%),
and D-aspartic acid, in the highest
amount (21.4%). A Mercenaria sample
from a Pleistocene locality, Wailes Bluff
('—'80 ,000 years old), shows a further in-
crease in D-amino acid contents. Again,
D-valine is lowest (16.4%) and D-aspartic
acid is highest (33%). A much older
sample of Mercenaria from the Upper
Miocene (5-10 X 106 yr) Drum Point
locality shows nearly 50% D-amino acids.
Only valine and proline are significantly
less than racemic. Also included are some
values for unhydrolyzed samples. These
values are in parentheses and for the
1000- and 80,000-year-old samples show
significantly higher levels of D-amino
acids. A sample of Foraminifera from
JOIDES site 148 at 230 meters depth
shows nearly racemic amounts of d-
amino acids in the free amino acid frac-
tion. Only valine (21.6%) is far from
the equilibrium value of 50%.
In an earlier Report (Hare, Year Book
70) it was shown that hydrolysis of pro-
teins under strongly alkaline conditions
results in nearly racemic mixtures of d-
and L-amino acids. The same treatment
on free L-amino acids showed only small
amounts of racemization. It is proposed
that during the diagenesis of proteins in
mollusk shells or in tests of Foramini-
fera, the natural slow hydrolysis of pep-
tide bonds under slightly alkaline condi-
tions (• — pH 8) gives rise to free amino
acids with substantial proportions of d-
amino acid isomers. If racemization oc-
curred primarily after the free amino
acids were released from the proteins,
then it might be expected that the pattern
of racemization would follow the pattern
for the heated L-amino acid standard
shown in Table 60. L-glutamic acid race-
mizes much more slowly than any other
amino acid at pH 8.5. In the natural sam-
ples, valine is the slowest amino acid to
racemize, whereas glutamic acid shows
complete racemization in the Miocene
sample, in which there is only 37% d-
valine. Hydrolysis showed that glutamic
acid had formed a lactam structure in the
shell and that it was substantially race-
mized. The data in Table 60 on free l-
amino acids suggest that if free L-gluta-
mic acid had resulted from the hydrolysis
of the shell protein, then a lactam of
GEOPHYSICAL LABORATORY
693
E
3?
o ©
© '-H
t_ 08
<u — i
£ *
c ^
'% o
a; .2
Pi 09
o B
CD S-i
02
gg
aS
O
O
Q3 >
QJ
. +s
02 aS
T3 O
•g O
g S
.2 s
S o
w aS
CO "
> a
T3
QQ
►*» £
^j a;
C3 O
2 "3
o to
- s
O aS
IT'S
■? §
cn -7
A o
+ s
w O
O ^
aS O
u O
»o
T3
0)
s
s
aS
!~
bC
O
E S
aS £J C
O -C
s- "if
as -5
a §•
a
J3
o
(jopejap uoiiDzmoj auuD|j) asuodsa^j
'-1 OS
—
694
CARNEGIE INSTITUTION
TABLE 61. Percentage of D-Amino Acid Isomers in Shells of Recent and Fossil
Mercenaria and Foraminifera from a JOIDES Core
Mercenaria
Foraminifera
Upper
Amino Acid
Recent
1000 Years Old
~80,000 Years Old
Miocene
JOIDES
Aspartic acid
8.7
21.4 (27.2)
33.0 (43.4)
47.1
(46.5)
Glutamic acid
4
11.6 (15.4)
20.0 (27.8)
49.0
(45.6)
Proline
4.6
13.6 (23.4)
28.3 (41.9)
44.6
(49.9)
Phenylalanine
3.8
12.0 (23.1)
27.0 (43.2)
49.5
(47.4)
Alanine
2.8
13.9 (21.8)
31.6 (41.7)
49.6
(51.2)
Valine
2.3
4.2 (12.8)
16.4 (30.4)
37.4
(21.6)
Leucine
3.1
9.1 (17.4)
19.2 (37.0)
(48.8)
Isoleucine
1.5*
9.5*(22.3*)
21.6*(38.5*)
54*
(55.3*)
* D-alloisoleucine.
Values in parentheses are for unhydrolyzed samples.
L-glutamic acid would have formed that
should have been highly resistant to race-
mization. The data of Table 61 suggest
that when free glutamic acid was formed
from the slow, natural hydrolysis of the
shell protein, substantial amounts of
D-glutamic acid were already present.
The determination of the relative pro-
portions of amino acid optical isomers
has significance in organic geochemical
studies. Amino acids synthesized abio-
logically have equal amounts of d- and
L-isomers. Biologically synthesized pro-
teins are polymers of L-amino acids, so
it would seem possible to distinguish the
mode of origin of an amino acid assem-
blage by this criterion. From previous
work (Hare and Abelson, Year Book 66)
and from the present study, however, it
is clear that during fossilization of shells
virtually all the L-amino acids racemize
extensively.
A preliminary study of a sample of
clay material from the same sample as
the Foraminifera in Table 61 shows a
completely different picture. Care was
taken to eliminate all carbonate, includ-
ing Coccoliths. Hydrolysis with 6 N HC1
and desalting with 2 N NaOH isolated a
mixture of amino acids almost entirely of
the L-conflguration. If these are not re-
cent contaminants it would indicate that
it is possible by clay-protein or clay-
amino acid interactions, or both, to pre-
serve the L-configuration of amino acids
for geologically significant periods of
time. It is necessary, therefore, to evalu-
ate the source material before making
any conclusions as to the rates of dia-
genesis of amino acids.
NEW TECHNIQUES, EQUIPMENT, AND
CALIBRATIONS
A Computer-Automated, Single-
Crystal X-Ray Diffractometer
L. W. Finger, C. G. Hadidiacos, and Y. Ohashi
The automation of single-crystal dif-
fractometers has reached the stage where
obsolete equipment is being replaced.
This report describes a flexible, inexpen-
sive, and easily programmed system de-
signed to implement such a change and
attempts to explain some of the reasons
for the options chosen. Although the
constraints of a limited budget do not
constitute the ideal criteria for selection
of equipment, cost reduction may be ac-
complished with no sacrifice of perform-
ance.
The diffractometer that has been re-
placed is a Supper-Pace equi-inclination
system operated in the mode specified by
GEOPHYSICAL LABORATORY
695
Burnham (Year Book 65, pp. 283-285).
This device was reasonably reliable and
had been upgraded by replacement of
subsidiary electronics; however, it was
not well suited for the study of crystal
structures at high temperatures, a
planned major research effort in this
Laboratory.
The input instructions for any diffrac-
tometer depend upon the lattice geom-
etry. If there is no on-line computer in
the system, then the necessary sequence
of operations must be prepared on a
separate computer. This procedure poses
no serious problem for collection of data
at room temperature as the unit-cell
parameters may be measured before
mounting the crystal in the instrument;
however, the change of parameters with
heating requires a recalculation of the in-
put instructions. For equi-inclination
geometry, a second difficulty occurs be-
cause the crystal rotation axis must be
parallel to one of the base vectors of the
crystallographic system. As a crystal of
low symmetry is heated, a rotation of the
axes may occur, necessitating an orienta-
tion correction. This geometry also is
very susceptible to the effects of multiple
diffraction, which is a major source of
systematic error in the measurement of
single-crystal intensities. Each of these
difficulties may be minimized by the use
of a four-circle diffractometer with an
on-line computer programmed to deter-
mine the lattice constants and the rela-
tionship between the crystallographic
and diffractometer axes after the crystal
is mounted on the instrument. In addi-
tion, the effects of multiple diffraction
are reduced by aligning the crystal in an
arbitrary orientation.
Although there was little freedom in
the choice of the diffraction geometry as
outlined above, the major details of the
axis-control electronics were not so easily
chosen. Many similar systems use en-
coding devices that digitize the absolute
position of the angle; however, experience
with the Supper-Pace diffractometer and
the automated electron microprobe
(Finger and Hadidiacos, Year Book 70,
pp. 269-275, and Year Book 71, pp. 598-
600), as well as other single-crystal dif-
fru^orneters (Alperin and Prince, 1970;
Basing et al, 1968; and C. T. Prewitt,
oersonal communication) , has shown
that lulse counting or relative encoding
with precision stepping motors provides a
reliable and inexpensive means of posi-
tioning. If acceleration and deceleration
techniques are employed, very rapid
pulse rates (on the order of 1000/sec)
may be used. Because the step-counting
approach involves significantly less elec-
tronics and expense, it has been selected
for this system.
In the selection of a computer and as-
sociated peripherals, the system designer
has virtually unlimited freedom. It is
highly desirable to be able to access the
programs for different functions such as
crystal orientation, lattice constant de-
termination, and data collection without
loading paper tapes or other such opera-
tions. This program accessing is usually
accomplished by the addition of a disk
or drum storage unit to the computer
(Alperin and Prince, 1970; Corfield,
1969; Lenhert, 1972); however, the
equivalent function may be performed in
a dedicated computer system by having
all necessary programs simultaneously
available in the memory. The relative
feasibilities of these two approaches are
dependent upon the total storage re-
quired. A memory-resident system is
ample for present requirements, and such
a configuration has been constructed us-
ing a Digital Equipment Corporation
PDP-11/20 computer with 12,000 words
of storage. The major reason for the se-
lection of this computer was the previous
experience in programming and inter-
facing the PDP-11 used on the auto-
mated electron microprobe. Figure 151 is
a block diagram of the total system.
The portions of the hardware related
directly to the diffractometer are a step-
ping motor and control for each of four
axes, a scintillation detector, a single-
channel analyzer, a rate-meter and strip-
696
CARNEGIE INSTITUTION
29 MOTi
WITH 5
REDUCER
J>
w MOTOR
WITH 5
REDUCER
X MOTOR
DIRECT
DRIVE
D-
0 MOTOR
DIRECT
DRIVE
D-
DETECTOR
AND POWER
SUPPLY
STRIP-CHART
RECORDER
MOTOR
CONTROL
X-RAY
SHUTTER
ATTENUATOR
AND FILTER
SELECTORS
(2)
20 MOTOR
CONTROL
oj MOTOR
CONTROL
\ MOTOR
CONTROL
0 MOTOR
CONTROL
SINGLE
CHANNEL
ANALYZER
RATE-METER
FURNACE
CONTROL
CRYSTAL
OSCILLATOR
REAL-TIME
CLOCK
f 100 HZ CLOCK
TIMER AND
PRESET
SCALER
□D
O
O
TELETYPE
HIGH SPEED
PAPER TAPE
READER
WRITE ONLY
MAGNETIC
TAPE
PDP-11/20
CPU
12 K
MEMORY
Fig. 151. The block diagram for the computer-automated, high-temperature, single-crystal x-ray
diffractometer. The diffractometer and x-ray generator are not shown. See Fig. 152 for details of
the furnace control.
chart recorder, a furnace controller, a
timer and scaler, and four programmable
relay closures. The relay closures are
used to control the x-ray shutter, the
strip-chart motor drive, and two attenua-
tor assemblies. The interface also con-
tains a crystal oscillator, which provides
the incrementation signals for the timer,
the time reference for the motor-drive
operations, and the real-time clock data.
The circuits, with the exception of the
furnace controller, are either standard or
extremely specific and will not be dis-
cussed in detail here.
Temperature Control
The circuit diagram for the program-
mable temperature controller is shown in
Fig. 152. The emf output of a Pt-PtlO%-
Rh thermocouple is amplified by a low-
noise, wide-band, chopper-stabilized am-
plifier (Al) with a gain of 1000. This de-
vice has the necessary voltage, tempera-
ture, and noise stability to accurately
convert the thermocouple emf's to the
usable range of 0 to 10 volts. The output
of the amplifier is input to a digital volt-
meter, which may be observed by the
operator; in -addition, this value is input
to the computer.
Control of the temperature has been
provided by wiring a field-effect, transis-
tor-input operational amplifier (A2) as
a comparator. In the manual mode, the
voltage on the reference input is con-
trolled by potentiometer R8. In the auto-
matic mode, this input is supplied by the
digital-to-analogue converter, resulting
in computer control. The other input of
the comparator is connected to the am-
plified thermocouple output; therefore,
the output of A2 is proportional to the
temperature error of the furnace. The
error signal is displayed on the galvan-
GEOPHYSICAL LABORATORY
697
o
a
e
o
2
5
Si
S
CD
H
tc
&>
Si
CD
cd
a
o
o
03
CD
CD
CD O
o3 "cq
8 2
I fl
2 .2
bD co
CD ^
£ e
o3 cu
bb S
.2 CD
T3 J3
CD gj>
H 2
^ Ph
bb .
o
698
CARNEGIE INSTITUTION
ometer Ml and amplified by A3. The
resulting potential is applied to the power
control circuitry, which utilizes a pre-
cision voltage regulator, jxsl 723, to con-
trol the amount of current passed by
transistors Qi and Q2. The regulator is
suitable not only for external control of
its output voltage but also for compensa-
tion of variations in the supply voltage,
Vcc. Thus this circuit is an externally
controlled, regulated power supply. Re-
sistors R14 and R15 are used as collector
current equalizers so that the transistors
can be operated in parallel without being
perfectly matched. Resistor R16 limits
the maximum current to the furnace
to about 6 amperes. Meters M2 and M3
display the current and voltage applied
to the heater.
The controller system described above
has not been thoroughly tested, and exten-
sive discussions of thermal stability, for
example, would be premature. Prelimi-
nary testing of the unit, however, has
shown a temperature variation of about
±2° at 700°C with a usable range of 50°
to 950°C. The sensitivity of the digital-
to-analogue converter used is approxi-
mately 0.3 °C per input unit, which
should be satisfactory.
Programs
A major, often overlooked expense of
computer-automated systems lies in the
preparation of suitable programs. This
cost is minimized by the use of a high-
level language in the programming.
Disk-oriented systems usually employ a
compiled language such as FORTRAN,
but this is not generally available for
memory-resident systems. On the other
hand, the interpretive language BASIC is
available for the PDP-11 and has been
used for this system. In addition to the
BASIC code, a small routine to link the
interpreter to the diffractometer inter-
face electronics has been prepared in
PDP-11 machine language. The total
programming time required to prepare a
system to orient the crystal, determine
lattice constants, and measure intensities
was approximately 150 hours — a very
reasonable time expenditure.
The activities performed on the dif-
fractometer may be divided into two
broad categories, intensity collection and
preliminary analysis. The new system
includes some novel approaches to each
aspect.
Intensity Measurements
Intensity measurement usually con-
sists of sequentially generating the Miller
indices for all the diffracting planes of
interest, driving the motors to the appro-
priate positions, measuring backgrounds
for a fixed time, and then scanning the
peak to measure the integrated intensity.
One difficulty with this procedure is that
the amount of time involved in meas-
uring all reflections is essentially con-
stant. Using counting statistics as a
measure of the uncertainty of the meas-
urement, the fractional error for the in-
tense peaks will be much smaller than for
the weak ones. The net result is that
many of the latter data are below the
level of significance ; increasing the meas-
urement time for all reflections would be
prohibitive, however, and might not im-
prove the situation because the optimum
condition would require a different ap-
portionment of time spent on the peak
and the background. Various aspects of
this problem have been studied by Ham-
ilton (1967) and Shoemaker (1968).
Their results are not in the exact form
required and have been rederived.* It
can be shown that the proper choices of
times for the intensity measurement are
U
B> + yB'P'
m2(P' — B')2
and
tP =
P' + yBT'
m2{P' — B')2
(1)
(2)
* The complete derivation is available upon
request.
GEOPHYSICAL LABORATORY
699
where tB is the total background time, tP
is the total scan time, Bf and P' are the
count rates at the background and peak
positions, respectively, and m is the de-
sired fractional error in the intensity. In
the event that P' < B', equations 1 and 2
are no longer valid, but the correct peak
and background times must be equal to
half the total time to be spent.
The proper application of equations 1
and 2 to a data-collection program re-
quires complete computer control over
the interface (Shoemaker, 1968). In the
system described here, the timer may be
preset to any desired time; similarly the
scan rate may be set to virtually any
value lower than 8° per minute.
Although this procedure appears to
provide very satisfactory results, consid-
erable work remains to evaluate it defi-
nitively. Several assumptions in the de-
rivation must be substantiated. In addi-
tion, the results of crystal-structure re-
finements from both methods of data col-
lection on the same crystal must be com-
pared.
Preliminary Analysis
In addition to the usual routines for
determining unit cell and crystal orien-
tation, this system includes some special
programs. One that automatically
searches reciprocal space to locate reflec-
tions is very useful when the crystal is
first mounted on the unit. The operator
may use the resulting table of diffractom-
eter angles and intensities to refine the
orientation of the crystal.
The second type of special program
available is useful in the study of diffuse
scattering such as exhibited in the
(h + &)-odd reflections of pigeonite.
This diffuseness has been used to esti-
mate the sizes of the antiphase domains
and to estimate thermal histories (Mori-
moto and Tokonami, 1969; Clark, Ross,
and Appleman, 1971; Brown, Papike,
and Prewitt, 1972; Hamil, Ghose, and
Sparks, 1973; and Brown and Wechsler,
1973) . One program is used to scan
through a reflection in any arbitrary di-
rection. The resulting strip-chart record
may be used to measure the half-width of
the peak. A second program is similar
to the procedure described by Hamil,
Ghose, and Sparks (1973). It performs
point counts of the intensity in the vicin-
ity of the reflection. From these data, a
detailed description of the intensity dis-
tribution may be computed (Ohashi and
Finger, in preparation) .
Summary
Although the diffractometer system de-
scribed above has been in service for only
a limited time, it has been shown to be
very easy to use for both the crystallog-
rapher and the programmer. It has also
been extremely reliable. A recent data-
collection sequence which lasted nearly
a month was completed without error. In
the future, this system should evolve
into an extremely powerful tool for the
study of crystal structures, thermal
states, and the x-ray behavior of mate-
rials at elevated temperatures.
Modal Analysis on the Automated
Electron Microprobe
F. Chayes
With the computerized monitoring and
control capability provided by Finger
and Hadidiacos {Year Book 70, pp. 269-
275; Year Book 71, pp. 598-600), the
routine operations involved in modal
analysis could now be performed on the
electron microprobe without operator
intervention. Construction of a BASIC
program that attempts to exploit this
possibility was begun during the last part
of the report year. Full operating data
will not be available for some time, but
the general design of the program will be
of interest to readers who have access to
automated electron microprobes. Given
computer monitoring of the motion of the
specimen holder, the notion of systema-
tically occupying a series of grid points
700
CARNEGIE INSTITUTION
almost suggests itself, the critical ques-
tion being the nature of the observation
to be made at each point.
For a sound modal analysis one wants
identifications at many grid points. Un-
less these identifications can be made
rapidly, the procedure will be grossly un-
economic. Even a qualitative analysis
requires a couple of minutes, and this is
far too much. In the current version of
the program, identification is accom-
plished by a form of ranking conveni-
ently described as a "ternary dominance
index." Each spectrometer is peaked for
a particular element, in the fashion pro-
posed by Chodos and Albee (1971), with
the microscope focused on the mineral in
which the element in question is most
abundant. (In a granite, for instance,
quartz would be used to peak on Si, micro-
cline to peak on K, and plagioclase, per-
haps, to peak on Ca.) The ranks of the
dominance index are defined for each ele-
ment by two scale factors applied to the
count recorded by each spectrometer at its
peak setting on the relevant internal
standard. Denoting the larger of the
factors by L, the smaller by S, and the
count recorded at the conclusion of the
peaking operation by C, a grid point at
which a count greater than LC is scored
is assigned a dominance index of 1; a
count intermediate between LC and SC
leads to a dominance index of 2 ; a count
less than SC to one of 3. For each grid
site such an index is recorded for each
of the three spectrometers, in fixed order.
There are thus 27 possible ternary
dominance indices. Some of these — in
the present program as many as 7 — may
be used as specific identifiers. During the
actual analysis the ternary dominance
index found at each grid point triggers
incrementation of the contents of the ap-
propriate element of a storage vector.
When counts have been made at all
points of the assigned grid, the relative
frequencies of specifically designated ter-
nary dominance indices are separately
listed, and the contents of the others are
pooled. In a granite analysis, for in-
stance, if spectrometers were assigned, in
order, to Ca, Si, and K, the dominance
index for quartz would be (313) and that
for potassium feldspar (321). If only
these two are requested, the tallies for all
other indices are pooled and reported as
"others."
Preliminary experience indicates, as is
hardly surprising, that identification by
ternary dominance index may be satis-
factory for some minerals and unsatis-
factory for others. It may fail because
the elements on which it is based do not
differentiate certain minerals from each
other: If, as in the example above, for
instance, the spectrometers are peaked on
Ca, Si, and K, the minerals magnetite,
ilmenite, and rutile will all have the
ternary dominance index (333) . This
might sometimes be remedied by using
ambiguous indices to trigger repeaking of
one or more of the spectrometers on ele-
ments that would distinguish the miner-
als in question from each other, but such
a modification would increase running
time materially if the ambiguities in-
volved abundant minerals. The opposite
case, in which what the petrographer or-
dinarily regards as a single mineral
yields more than one dominance index,
is probably more important if only be-
cause it will so often affect major min-
erals, e.g., the variably sericitized and
epidotized plagioclase of most two-feld-
spar granites. Simple program modifica-
tions would permit the operator to in-
struct the program to pool certain of the
dominance indices into specific combina-
tions before the final step, in which un-
specified indices are pooled into "others."
In addition to bona fide ambiguities
such as these, which would persist even if
the observations could be made without
error, the dominance index may lead to
misidentification because the area irradi-
ated at a particular grid point is not
monomineralic, because it contains cleav-
age cracks or other imperfections, or be-
cause focus is not properly maintained
GEOPHYSICAL LABORATORY 701
during traverse from one grid point to Fluorescent Method for the Analysis
another. The importance of the first two of Amino Acids and Peptides
factors can scarcely be evaluated unless p E Hare
the third is in good control, and in this,
as in other applications of the electron The new reagent fluorescamine (Uden-
microprobe, the consequence of loss of friend et al., 1972) promises to bring
focus may be catastrophic. In the present about significant advances in the detec-
program the operator may interrupt at tion and quantitation of submicrogram
any time to adjust focus, but the constant amounts of amino acids, peptides, and
surveillance this requires largely sacri- proteins. The compound forms highly
flees what should be one of the major fluorescent derivatives of NH2 groups
advantages of automation, viz., unat- that can be detected by sensitive fluores-
tended operation. Further, although ex- cence spectroscopy with activation at 390
perience to date suggests that intermit- nm and fluorescence at 490 nm. In many
tent adjustment of focus may easily be- cases it is possible to distinguish between
come haphazard, there is no assurance free amino acids and peptides by varying
that it is properly random; on the other the pH of the reaction because peptides
hand, adjustment of focus at every grid retain a high degree of sensitivity over a
site, though perhaps admissible in special wide pH range (5-11) whereas free
circumstances, is too time-consuming for amino acids are sensitive only over a rela-
routine use. Ideally, focusing capability tively narrow pH range (8.5-9.5). A
should be incorporated in the program ; comparison of two successive runs at pH
this is now being attempted. 6 and 9, for example, shows peptides in
With or without automatic focusing, it both runs but reveals free amino acids
is clear that the procedure will not be only in the run at pH 9.
particularly rapid. Allowing a second per The system developed at the Geophys-
point for counting and almost as much ical Laboratory (Hare, 1966, 1969, 1972,
for translation of 0.5 mm between points, and Year Book 70) for the analysis of
the maximum speed cannot much exceed amino acids and peptides has been based
30 points per minute. In the current on the reaction of ninhydrin with amino
version of the program, computing over- acids. This system is at least three orders
head adds the better part of another sec- of magnitude more sensitive and one
ond per point, so the actual rate is a little order of magnitude faster than the orig-
less than 25 points per minute. A skilled inal commercial amino acid analyzers
petrographer working on optimum mate- (Spackman, Stein, and Moore, 1958).
rial with a conventional microscope Most of the improvement in sensitivity
should be able to do much better. But is the result of reducing the volume of
fatigue is an important limiting factor the ion-exchange columns,
in nonelectronic modal analysis, the time It has been possible to modify the im-
required increases very rapidly with de- proved instrument in such a way that
crease in grain size or increase in identifl- either ninhydrin or fluorescamine can be
cation difficulties, and for combinations used to monitor the effluent stream from
of these reasons many materials of the ion-exchange column. It has also
considerable petrographic interest are been possible to operate two columns
quite unsuited for conventional modal simultaneously, monitoring one with
analysis no matter how much time the ninhydrin and the other with fluores-
analyst is prepared to spend on them. A camine. A block diagram of this system
reliable, smoothly operating program for is shown in Fig. 153.
modal analysis by electron microprobe For the fluorescent detection of the
would be a useful auxiliary weapon in the effluent, an Aminco fluorocolorimeter was
petrographer's arsenal. used with a 4-watt lamp source. With
702
CARNEGIE INSTITUTION
Nitrogen —
Pressure
(~500 p.s.i.
Automatic
Valves
PH
3.25
PH
4.25
PH
5.2
PH
10. 1
Polystyrene
Buffer Reservoirs
Nitrogen —
Pressure
H 00 p.s.i)
Stainless
Steel
Ninhydrin
Reagent
Reservoir
(-20°C)
S
Flowmeter
■CZh
and Back
Pressure
Column
20-Port
Sample
Injector
Ion-Exchange
Column (55°C)
25 cm X 2 mm
8fi Resin
2-Po
Sam
Injector
ort I
pie £n
Reaction
Coil(IOO°C)
Stream
Splitter
Valve
Colorimeter
Recorder
N2 or He H50 p.s.i)
Borate
Buffer
pH9.5
Fluorescamine
in Acetone
U
Fluorimeter Recorder
Fig. 153. Block diagram of amino acid analyzer with both ninhydrin and fluorescamine detec-
tion systems. Stream splitter valve allows the collection of unreacted material from the column
while a fractional part of the column effluent is passing through fluorimetric detection system.
this detector the fluorescamine method is
approximately as sensitive as the nin-
hydrin method (Fig. 154).
The claim made for greatly increased
sensitivity with fluorescamine (Uden-
friend et al., 1972) is somewhat mislead-
ing. The ninhydrin and fluorescamine
systems were not monitoring similar col-
umns as they were in the present study.
An increase in the intensity of the source
will produce a similar increase in the
intensity of the fluorescence so it may
well be possible to increase the sensitivity
of the fluorescamine system by a signifi-
cant factor.
At the present levels of sensitivity
there are some significant advantages in
the use of fluorescamine. The reaction
with amino acids is essentially instan-
taneous at room temperature and thus
requires no long coil-heating time as does
ninhydrin (10 minutes at 100 °C). As
can be seen in Fig. 154, this technique
improves resolution and simplifies the
system considerably.
It has been possible to develop a
simple stream-splitting device that en-
ables one to collect most of the effluent
while continuously monitoring a part of
the effluent with fluorescamine. The efflu-
ent from the column is passed through a
three-port manifold to the mixing mani-
fold. The side-port is connected to a
short length of tubing (the collecting
tube) and a valve consisting of a Teflon
tube drawn through a Kel-F body fitted
with two nylon set-screws to constrict
the tubing. The amount of the column
effluent from 0 to 90% can be adjusted to
flow through this tube, while the remain-
der of the column flow is forced into the
mixer with fluorescamine and detected in
the fluorocolorimeter. It is easy to adjust
the lengths of tubing so that a com-
pound's fluorescent reaction product is
making a peak on the recorder at the
same time it is emerging unreacted from
the collector tube (Fig. 153). It can be
put, into a fraction collector for further
characterization.
There are some disadvantages in the
use of fluorescamine. It is not sensitive
to the amino acids proline and hydroxy-
proline since these are not primary
amines. It is possible to program another
step in which an oxidizing agent converts
proline and hydroxyproline to primary
amines, which will then in turn react with
fluorescamine. Only limited success was
achieved in detecting proline by the use
of fluorescamine. Another disadvantage is
its insolubility in aqueous buffers, which
GEOPHYSICAL LABORATORY
703
TIME (MINUTES)
A NINHYDRIN - WITH 6mm PATH FLOWCELL
w^JoC!t
B FLUORESCAMINE- 1 NANOMOLE
£8
JU U
z
=3 Z>
o
Crt
!
I
1
j
! i
i !
! !
1
1
i
1
*
^u
C. FLUORESCAMINE - 20 PICOMOLES
~i i r
30 40 50
TIME (MINUTES)
a g
a. 2
TO ^
d *
'§ 2
03 g
.d CD
O
03 r2
CU v3
*s'5
^-v CD
CU O
is
O o3
CD CP
2 £P
'3 2
o 2
u. -d
>» «
** 5-
CD ^
d
a-
cp d
T'g
2 3
r! CO
O (h
CD O
O ^3
§1
-3 %
CD
"fls
S-. -d
-o 2
rv o3
^ CU
o
•2 3
2^
o3 rt
£> -d
co
£ 03
O
d
cd
T3
d
o3
-3 g
.2 2
ri 03
H QJ
o
CU
43 cu
fc 2
2~
go
CO ^^
CU
S-i
o <u
d.2
50 -3 -J
^ S .22
-M ft 5-1
■.H CU
^ "° "5
*! § 2
d
o3 O
-4-> b*
ft bu
o3
CO .£i
m d
03 d
2
03
cu
bfi
2d
CU p] --1
2 O CU
°3 2 2
m I ©
O) d CO
-d * co
-^ in CU
d d
CD
o -Q
o3 03
O cu
>d o
d
o
'bfi
CU
—
io cu 03
2£:S
bfi cu
w -i M
CU d • —
«^ 2
O 03 o3
704
CARNEGIE INSTITUTION
necessitates an organic solvent and a
careful selection of relative flow rates to
prevent precipitation in the lines.
The amino acid analyzer at this Labo-
ratory has proved to be a versatile instru-
ment. It has been adapted for silica
analysis (see Frantz and Hare, this Re-
port) and can probably be adapted to
almost any method requiring the addi-
tion of reagents and the subsequent
monitoring of a reaction product that
absorbs light energy or produces fluores-
cence.
The Analysis of Silica at Submicro-
gram Levels Using a Flow-Cell
colorimetric technique
J. D. Frantz and P. E. Hare
Progress in the experimental study of
hydrothermal mineral-solution equilibria
is greatly dependent on the development
of techniques by which small amounts of
aqueous fluids can be analyzed. These
microanalytical techniques are necessi-
tated by the relatively small amounts of
solution used in most hydrothermal ex-
periments (<100 /xl). Refractory ele-
ments such as silicon and aluminum are
especially difficult to analyze since they
are commonly present at extremely low
concentrations in the fluid (Morey and
Hesselgesser, 1951). A microanalytical
technique for dissolved silica has now
been developed that utilizes a continuous
flow system in which the absorbance of
silica molybdate complexes is measured
in a flow-cell colorimeter similar to that
developed for amino acids at this Labo-
ratory (Hare, 1972). The sensitivity and
precision of this technique are compar-
able to those obtained from the standard
static molybdate-yellow method (Strick-
land, 1952), except that the new system
requires only 1 yxl of sample rather than
the normal 1 to 10 ml.
In the past a great deal of effort has
been devoted to the colorimetric analysis
of silica in aqueous solutions (Mullin and
Riley, 1955) . The most widely accepted
method uses ammonium molybdate to
form silica molybdate complexes that
have absorbance in the visible light. Re-
duced complexes (using a reducing agent
such as Na2S03) absorb blue light; un-
reduced species, yellow light. Absorb-
ances, and therefore relative silica con-
centrations, are determined using a
standard colorimeter with constant vol-
ume cells. One difficulty with this
method has been caused by color insta-
bility due to the presence of at least two
complexes, one of which is metastable.
The present method of analysis incorpo-
rates these principles, but instead of the
normal static instrument, it uses a flow-
cell colorimeter, which eliminates the
effects of color instability.
The system basically consists of a con-
stant flow of an ammonium molybdate
solution from a polystyrene reservoir
(A, Fig. 155) through Teflon tubing (32
gauge) into two flow colorimeters (G,
H) . The reagent flow is driven by pres-
surized argon and monitored using a flow
meter (C) . Sample injection is ac-
complished by inserting a 1 to 5 /*1 Dade
disposable micropipette into the ammo-
nium molybdate stream (E, F). Inser-
tion of the pipette is facilitated by the use
of two three-way valves (D) , whereby
the injector is bypassed when the pipettes
are exchanged. The accuracy of this in-
jection method is ±1% in volume. In
order to ensure the formation of the silica
molybdate complexes, the sample and
surrounding solution flow through 40 feet
of tubing immersed in boiling water be-
fore entering the colorimeters (G) . Col-
orimeters with 6- and 60-mm path
lengths were used in order to cover a
large range of silica concentration.
The details of the flow colorimeters, as
described by Hare (Year Book 70, pp.
268-269), are shown in Fig. 156. The
0.020-inch inner bore of a black Teflon
cylinder is sealed at either end with clear
glass plates. Diagonal holes into the
chamber provide for entry and exit of the
solution through 32-gauge Teflon tubing.
GEOPHYSICAL LABORATORY
705
Fig. 155. Schematic diagram of the flow system. (A) Reagent reservoir. (B) 32-gauge Teflon
tubing. (C) Flow meter. (D) Three-way valve. (E) Sample injector. (F) Volupette micro-
pipette. (G) 40-foot coil of Teflon tubing. (H) 6-mm colorimetric cell. (I) 60-mm colorimetric
cell.
A miniature lamp with a focusing lens
(Chicago Miniature Lamp Works CM
20-3) and an interference filter (Oriel
G-572-4000) provides a monochromatic
light source of 400-nm wavelength at one
end of the cylinder. A photoconductor
cell (Clariex CL902L) is placed at the
other end. A constant 2.8 volts is sup-
plied by a Sorensen power supply (QSA
10-1.4) , which provides a 0.75-mV output
using a 60-mm cell. The response has
proven extremely stable and yields pre-
cise determinations of changes in absorb-
ance.
Figure 157 illustrates the recorder re-
sponse (in transmittance) to the addi-
-+
To Recorder
and Integrator
Fig. 156. Flow-cell colorimeter. (A) Delrin
body. (B) Interference filter. (C) Focusing
miniature lamp. (D) Black Teflon cylinder;
inner surface lined with Teflon tubing. (E)
Photoconductor cell. (F) Clear glass plate. (G)
Delrin screw. (H) 32-gauge Teflon tubing.
Fig. 157. Recorder chart showing response to
the addition of silica standards ranging from 1
to 1000 parts per million. Bottom set of peaks
corresponds to use of 6-mm flow cell ; top set,
60-mm cell.
706
CARNEGIE INSTITUTION
tion of 1-jA aliquots of silica standards
ranging from 1 to 1000 parts per million
(prepared from dried sodium silicate
pentahydrate). The bottom and top sets
of peaks correspond to the flow cells with
6- and 60-mm path lengths, respectively.
The asymmetrical shape of the peaks
may be the result of friction-induced
turbulence in the Teflon tubing. The
areas under the peaks generated from the
60-mm cell are represented in Fig. 158.
They were calculated using a logarithmic
scale for heights and an arbitrary linear
unit to measure half-widths with the area
of the blank subtracted from each. The
surprisingly close approximation to a
Beer's law response over the concentration
range is apparent. The resolution of this
method can probably be increased by
several orders of magnitude by the addi-
tion of an acid to control pH at 2 to 3 or
by the use of a blue visible light source
in conjunction with the addition of a re-
ducing agent (Mullin and Riley, 1955).
This technique has proved extremely
valuable in the study of hydrothermal
reactions involving aqueous silica. The
general method should be applicable to
100 200 300 400 500 600 700 800 900 1000
PPM SILICA
Fig. 158. Plot of peak-height area versus
silica concentration in parts per million. Inset
provides expanded view of low-concentration
field. See text for explanation of units in de-
scribing areas.
the analysis of many other elements for
which colorimetric techniques are avail-
able. One exciting application might be
the analysis of various anions that may
be found in fluid inclusions in igneous
and metamorphic minerals.
Albite- Jadeite-Quartz Equilibrium :
A Hydrostatic Determination
James Fred Hays and P. M . Bell
The solid-media, high-pressure appa-
ratus developed by Boyd and England
(Year Book 57, and 1960, 1963) is now
routinely employed in geochemical stud-
ies at some thirty or more laboratories.
Unfortunately, there is a significant un-
certainty associated with pressures at-
tained in this apparatus, owing to the
so-called friction correction that must be
applied. Many attempts have been made
to estimate the magnitude of this friction
correction, but there has been no uni-
versal agreement on appropriate calibra-
tion methods and at present a wide range
of pressure-cell configurations, run pro-
cedures, and calibration methods are in
use. Perhaps the most complete study of
this problem is that of Johannes et al.
(1971), in which investigators from six
different laboratories, including the Geo-
physical Laboratory, undertook a study
of the albite- jadeite-quartz equilibrium
at 600 °C, using identical starting mate-
rials and criteria for equilibrium but dif-
ferent pressure cells, run procedures, and
calibration methods. Apparent pressures
for the equilibrium (without friction cor-
rections) ranged from about 15 kbar to
more than 18 kbar. After each investiga-
tor had applied a friction correction to
his own results, the corrected values were
found to lie within a band of about 1
kbar in width centered at 16.2 kbar. It
was not clear whether this apparent
agreement validated the calibration
methods used or whether a shared sys-
tematic bias was involved.
In order to resolve this question and to
provide an accurate calibration point for
GEOPHYSICAL LABORATORY
707
other users of solid-media apparatus, this
equilibrium was studied again at 600 °C,
this time using the Birch-Bridgman
nitrogen-gas apparatus (Robertson,
Birch, and MacDonald, 1957), in which
the pressure is hydrostatic and can be
monitored directly by means of a Man-
ganin resistance coil. The starting mate-
rial was the same as in the earlier study :
a 1:1 mixture (by weight) of crystalline
reactants and products sealed with water
in a platinum tube. Comparison of x-ray
diffraction peak intensities and optical
microscopy was used to determine the ex-
tent and direction of reaction. Run times
of 12 hours proved adequate in each case
to produce positive evidence of reaction.
Jadeite and quartz were observed optic-
ally in the products of all runs on the
high-pressure side of the equilibrium, al-
though it was evident that some quartz
had dissolved in the vapor phase (Boett-
cher and Wyllie, 1969b).
The albite starting material used, as in
the earlier study, was high albite, synthe-
sized hydrothermally from a mixture of
oxides by Johannes at 750 °C and 2 kbar
in a one-week run. Cell parameters for
this material are given in Table 62. No
change in cell parameters was noted after
12 hours at 600 °C and 16 kbar. The re-
action studied is therefore high albite =
jadeite + quartz. It is not clear that
high albite is the most stable structural
form at 600 °C. The evidence summar-
ized by R. C. Newton and Smith (1967),
however, and the data shown in Fig. 160
(below) suggest that the structural state
of the starting feldspar has little effect
on the pressure of the equilibrium at
TABLE 62. Cell Parameters of High-Albite
Starting Material
620
610
£ 600
590
580
-
-
High
Albite
f
r"
I
I
-
i
i
/
/
/
i
■j
Jadeite + Quartz
1
1
I
15
16
17
18
Pr
essure, kb
a 8.153 A (±0.003)*
b 12.862 A (±0.002)
c 7.114 A (±0.002)
a 93.60° (±0.02)
j8 116.50° (±0.02)
7 89.96° (±0.02)
* Stated uncertainties are least-squares
residuals.
Cell refinement by Guy Hovis using powdered
sample, Ni-filtered Cu« radiation, and silicon
internal standard.
Fig. 159. Bracketing runs on equilibrium at
600° C. Size of symbol gives uncertainty limits;
dashed symbols include allowance for pressure
error resulting from uncertainty in run temper-
ature.
600 °C, although a pronounced effect on
the slope of the pressure-temperature
curve is to be expected.
Results of two runs bracketing the
equilibrium are shown in Fig. 159. The
size of the boxes is a measure of the ex-
perimental 'precision and reflects mainly
fluctuations in temperature and pressure
during the runs (±5°C, ±0.12 kbar).
Errors in measurement of thermocouple
emf and Manganin coil resistance are
negligible.
Pt-Ptl0%Rh thermocouples were not
individually calibrated and an uncer-
tainty of ±3° is expected from this
source based on the manufacturer's cali-
bration. No correction for the effect of
pressure on thermocouple emf was made.
Since the P-T slope of the equilibrium
curve (with high albite) is about 30
bar/deg, the temperature uncertainty
implies a corresponding uncertainty in
the equilibrium pressure at 600 °C of
±0.15 kbar, indicated by the dashed ex-
tensions of the boxes in Fig. 159.
A significant uncertainty arising from
nonlinearity of the Manganin resistance
pressure scale remains to be evaluated.
The Manganin coil was calibrated at low
pressures by comparison with a well-
calibrated Heise bourdon-tube gauge that
had been calibrated against a dead-
weight gauge. The coil was found to
708
CARNEGIE INSTITUTION
620
This work
/
610
— NTS
x /
o
\/HK K
BL
0,-600
*
x«» •
• •
« 590
/ NS
/ BW
NK
6
•" 580
Johannes et al /
/,
l
1
15 16
17
Pressure, kb
18
Fig. 160. Comparison with previous results.
HK, Hlabse and Kleppa (1968); NS, R. C.
Newton and Smith (1967); BW, Boettcher and
Wyllie (19696); K, Kelley et al. (1953); BL,
Birch and LeComte (1960) ; NK, M. S. Newton
and Kennedy (1968).
have a pressure coefficient of 2.24 X 10~3
kbar-1. If this coefficient can be assumed
constant to 16-17 kbar, then the bracket
is centered at 16.4 kbar, and each
bracketing run has a total pressure un-
certainty of ±0.3 kbar. If, as seems
likely, the Manganin pressure scale is
nonlinear (Babb, 1963), then all pres-
sures given here should be increased by
about 1 % . Attempts are now being made
to calibrate the coil at the bismuth I— II
transition to evaluate this effect, but
absolute calibration in the range 16-18
kbar is not available at this time.
The tentative value of 16.4 ± 0.5 kbar
for the equilibrium pressure at 600 °C is
compared with previous work in Fig. 160.
The rectangle includes all the corrected
values reported by Johannes et al. ( 1971 ) .
The values given by Hlabse and Kleppa
TABLE 63. Enthalpy of Reaction :
High Albite = Jadeite + Quartz
A#°873 = +4423 db 900 cal
Atf °873 = +2879 ± 2000 cal
A#°873 = +3955 ± 800 cal
(1)
(2)
(3)
1 . This work based on an equilibrium pressure
of 16,500 ± 500 bars. A£873 and A7298 from
Robie and Waldbaum (1968).
2. From tabulated enthalpies in Robie and
Waldbaum (1968).
3. Hlabse and Kleppa (1968). Extrapolated
to 873°K.
(1968) and by Kelley et al. (1953) are
calorimetric predictions. The values
given by R. C. Newton and Smith (1967)
and by Boettcher and Wyllie (19696) are
experimental determinations using solid-
media apparatus with low albite and al-
bite glass, respectively, as starting mate-
rials. The determination by Birch and
LeComte (1960) is an extrapolation to
lower temperature from experiments car-
ried out at high temperatures in a gas
apparatus identical with the present ap-
paratus. R. C. Newton and Smith (1967)
have shown that it is possible to fit Birch
and LeComte's high-temperature re-
versals with a curve that also passes
through the 600 °C bracket if account is
taken of the effect of Al-Si disorder in
albite on the slope of the equilibrium
curve.
The equilibrium pressure at 600 °C can
be combined with entropy and volume
data for high albite, jadeite, and quartz
tabulated in Robie and Waldbaum
(1968), and a value for the enthalpy of
reaction at 873°K can be calculated.
This value is shown in comparison with
others obtained using calorimetric meth-
ods in Table 63.
Incongruent Melting of Pure
Diopside
/. Kushiro
Melting relations of pure diopside at 1
atm have been studied in detail, and its
incongruent melting behavior has been
confirmed. Electron microprobe analyses
of diopside solid solution and coexisting
glass have established the solidus and the
liquidus near the diopside composition at
temperatures above 1383 °C. For the
calibration of a thermocouple with diop-
side within an error of less than 0.5°, it
is necessary to obtain a run in which the
amount of crystals is less than 20%.
Melting relations near the diopside
composition were reported last year (Ku-
shiro, 1972c, Fig. 4; Year Book 71,
p. 605) , and the incongruent melting of
GEOPHYSICAL LABORATORY
709
pure diopside was shown on the basis of
one run at 1390 °C. Because pure diop-
side has been used as a secondary stand-
ard in the temperature scale, it is impor-
tant to know its melting behavior in
detail. In the present study, the melting
relations of pure diopside synthesized in-
dependently by J. F. Schairer and by
D. H. Eggler were determined, using both
the quenching method and electron
microprobe analysis. The temperatures
of the runs were calibrated by the liqui-
dus temperature of pure diopside
(1391.5°C).
At temperatures above at least 1375 °C,
diopside solid solution and liquid co-
existed. The amount of liquid increases
rapidly at temperatures above 1385°C,
and at about 1390 °C the ratio of glass
and diopside solid solution is nearly 1.
Coexisting diopside solid solution and
glass in the quenched run products were
analyzed with the electron microprobe by
the method described by Finger and
Hadidiacos (Year Book 70, pp. 269-
275). The results are shown in Fig. 161.
Diopside solid solution is closer in com-
position to pure diopside at lower tem-
perature; however, it is less diopside-rich
than Di98En2 (wt %) at temperatures
above 1375°C. Biggar and O'Hara
(1969) observed, in the melting of pure
diopside, a small amount of glass even at
temperatures as low as 1350 °C, suggest-
ing that even at this temperature pure
diopside is not stable, and the liquid is
significantly off the join diopside-wol-
lastonite.
The electron microprobe analyses show
that the diopside solid solution contains
a small amount (<5 wt %) of forsterite
component and that coexisting glass con-
tains excess silica over the compositions
of the join MgSi03-CaSi03. In Fig. 161
the compositions of diopside solid solu-
tion and glass are projected onto the join
MgSi03-CaSi03 along the lines of equal
Ca content.
The analyses of the glasses show a
smooth liquidus curve, which is very
close to the liquidus determined by
Schairer and Bowen (1942) by the
quenching method (dashed curve in Fig.
161). The diopside solid solution occurs
as sharp euhedral crystals, and no quench
crystals are observed (Fig. 162A). On
1395 -
1390
u
o
| 1385
(D
Cl
E
H- 1380
1375
10
MgSi03
Liq
1391.5°
Schairer & Bowen
CaMgSi206 10
Weight per cent
CaSiO-
Fig. 161. Liquidus and solidus of diopside solid solution determined by electron microprobe
analyses.
710
CARNEGIE INSTITUTION
Fig. 162. Photomicrographs of quenched run products of diopside (A) 1390.5°C; (B) 1382°C
(poorly quenched; Q, quench crystals).
the other hand, one run, which was made
at 1382 °C and poorly quenched (the cap-
sule did not drop immediately into mer-
cury), contains a small amount of fine-
grained, irregular-shaped quench crystals
between large crystals of diopside solid
solution (Fig. 162B). The glass in this
run product has the composition CaO
40.7, MgO 3.80, Si02 54.9 wt %, which
lies in the pseudowollastonite liquidus
field and is far from the expected com-
position. Quenching, therefore, must be
particularly rapid when electron micro-
probe analysis is applied to the determi-
nation of liquid composition.
On the basis of the equilibrium dia-
gram (Fig. 161), temperature can be
determined from the ratio of diopside
crystals and glass even below the liquidus
for pure diopside composition; for ex-
ample, the ratios estimated are about
1:1, 2:1, 3:1, and 4:1 at 1390.5°, 1389°,
1388°, and 1387°C, respectively. For the
calibration of a thermocouple using the
diopside liquidus within an error of less
than 0.5°, however, it is necessary to ob-
tain a run in which the amount of diop-
side crystals is less than 20%. The tem-
perature of this run should be between
1391° and 1391.5°C.
STAFF ACTIVITIES
Conference on Geo chemical Transport
and Kinetics
Forty-two scientists, working in di-
verse fields ranging from ceramics and
metallurgy to geochemistry and field
geology, met to discuss "Geochemical
Transport and Kinetics." The CIW-
supported conference was organized by
Drs. Albrecht W. Hofmann (Department
of Terrestrial Magnetism) , Hatten S.
Yoder, Jr. (Geophysical Laboratory),
Bruno J. Giletti (Brown University),
and Richard A. Yund (Brown Univer-
sity) , and was held at Airlie House, in
Warrenton, Virginia, from June 4 to 6,
1973.
Rates of diffusion, vapor transport, ex-
solution, and irreversible chemical reac-
tions are of fundamental importance to
the understanding of geochemical proc-
esses in the earth's crust and mantle.
Little is known at present about these
rates and the factors controlling them.
The need for systematic research on rate
processes is widely recognized, but until
GEOPHYSICAL LABORATORY 711
recently, progress in this field was dis- mineral assemblages, and metasomatic
appointingly slow. Modern instrumenta- zoning in skarns).
tion has opened new opportunities for the The conference discussion made it ap-
solution of many of the problems that parent that research on chemical trans-
were difficult to study in the past. For port and kinetics has made a significant
example, it is now possible to analyze beginning in the earth sciences. In order
diffusion gradients on the scale of a few to encourage others to join in this work,
micrometers with the electron micro- the Carnegie Institution of Washington
probe, and the electron microscope re- will publish a volume of the proceedings
solves incipient exsolution lamellae only of the conference. The book is being as-
150 angstroms wide. sembled and edited by the organizers and
The attending scientists presented a will contain both research contributions
wide range of views on the subject: One and review papers presented at the con-
participant discussed "vacancy wind" in ference.
solids; another held up an object and The following scientists participated in
explained that "this is a rock." Sixteen the conference: Thomas J. Ahrens (Cali-
contributions were concerned with solid- fornia Institute of Technology), David
state diffusion, with particular emphasis E. Anderson (University of Illinois),
on common rock-forming minerals such Peter M. Bell (Geophysical Laboratory),
as feldspar, olivine, quartz, mica, and C. Ernest Birchenall (University of Del-
calcite. Introductory reviews on diffu- aware), John Brady (Harvard Univer-
sion mechanisms in simple crystals and sity) , Donald M. Burt (Rijksuniversiteit
on diffusion in sulfides illustrated the te Utrecht, Netherlands), Bruce H. T.
difficulties one encounters when measur- Chai (Yale University), Robert N. Clay-
ing diffusion in crystals even though their ton (University of Chicago), Alfred R.
structure is comparatively simple. Cooper (Technische Universitat Claus-
Other topics of discussion were diffu- thai, Germany), George W. Fisher (The
sion theory in multi-component systems, Johns Hopkins University) , Raymond C.
diffusion related to geochronology kinet- Fletcher (Department of Terrestrial
ics of calcite recrystallization, kinetics Magnetism) , Kenneth A. Foland (Uni-
of exsolution and disordering in feld- versity of Pennsylvania) , John D. Frantz
spars, equilibrium and mass transfer (Geophysical Laboratory), Fred A. Frey
among minerals and aqueous solutions at (Massachusetts Institute of Technology) ,
high temperatures and pressures, and dif- Bruno J. Giletti (Brown University),
fusion in granitic melts. Julian R, Goldsmith (University of Chi-
An evening session dealt with the use cago) , Harry W. Green II (University
of hydrogen, oxygen, strontium, and ar- of California, Davis), Stanley R. Hart
gon isotopes as tracers in geological proc- (Department of Terrestrial Magnetism) ,
esses. The final day of the conference Harold C. Helgeson (University of Cali-
was devoted to transport in sedimentary fornia, Berkeley) , Arthur H. Heuer (Case
and metamorphic rocks. The contribu- Western Reserve University), Albrecht
tions covered a number of theoretical W. Hofmann, (Department of Terrestrial
aspects (steady-state diffusion in meta- Magnetism), J. T. Iiyama (Centre de
morphism, recrystallization under non- Synthese et Chimie des Mineraux,
hydrostatic stress, models of infiltration France), Thomas E. Krogh (Geophysical
and diffusion metasomatism, and volatile Laboratory) , Ikuo Kushiro (Geophysical
flux during metamorphism) , as well as Laboratory), Frank T. Manheim (Woods
geological observations (diffusional per- Hole Oceanographic Institution) , John
meability of sediments, the relationship R. Manning (U.S. National Bureau of
of metamorphic differentiation to vein Standards) , Robert H. McCallister (Pur-
712
CARNEGIE INSTITUTION
due University), J. D. C. McConnell
(Cambridge University, England), D.
James Misener (University of British
Columbia), Philip M. Orville (Yale Uni-
versity) , Malcolm Ross (U.S. Geological
Survey), Danny Rye (Yale University),
Herbert R. Shaw (U.S. Geological Sur-
vey) , Yuch-Ning Shieh (Purdue Univer-
sity), Philip J. Sipling (Brown Univer-
sity), Hugh P. Taylor, Jr. (California
Institute of Technology), Rosemary J.
Vidale (State University of New York,
Binghamton), Alain Weisbrod (Geo-
physical Laboratory) , David R. Wones
(U.S. Geological Survey), Hatten S.
Yoder, Jr. (Geophysical Laboratory),
and Richard A. Yund (Brown Univers-
ity).
Washington Crystal Colloquium
The Washington Crystal Colloquium,
an informal monthly assembly of crys-
tallographers from the Washington area,
met eight times during the report year.
The following lectures were presented:
"Crystallography and structural chem-
istry of calcium phosphates," by Brian
Dickens (National Bureau of Stand-
ards), October 20, 1972.
"Degree of ionicity in crystals? Semi-
quantitative information from diffraction
data," by P. Coppens (State University
of New York at Buffalo), November 10,
1972.
A discussion of photoreactions with
"Nucleic acid bases," by Judith L. Flip-
pen and "Tyramines and a steroid," by
Daniel S. Jones (Naval Research Labo-
ratory) , December 15, 1972.
"Flash x-ray diffraction studies and
phase transitions," by Quinton Johnson
(Lawrence-Livermore Laboratories) ,
January 12, 1973.
"Comparison of radial distribution
functions (RDF's) from glass and crys-
tal structures," by John Konnert (Naval
Research Laboratory), and "Determina-
tion of heavy-metal bond distances using
RDF's from powder data," by Carl
Quicksall (Georgetown University), Feb-
ruary 23, 1973.
"Mathematical characterization of
twinning," by A. Santoro (National Bu-
reau of Standards), March 23, 1973.
"High-low phase transitions in clino-
pyroxene," by Yoshikazu Ohashi (Geo-
physical Laboratory), April 27, 1973.
"X-ray diffraction and electron micros-
copy of biological membranes," by
Harvey Pollard (National Institutes of
Health) , June 8, 1973.
Seminar Series
The regular seminar series met 14
times during the report year. The speak-
ers included invited guests, Staff Mem-
bers, and Fellows. The following lec-
tures were presented:
"Amphibole stability — theory and ex-
periment," by D. H. Eggler (Geophysical
Laboratory) , October 19, 1972.
"Chemical petrology of kimberlites,"
by F. R. Boyd (Geophysical Labora-
tory), November 2, 1972.
"Degree of ionicity in crystals? — semi-
quantitative information from diffrac-
tion data," by P. Coppens (State Univer-
sity of New York, Buffalo) , November
10, 1972.
"Pedogenesis, strontium isotope geol-
ogy and tectonics of Mount Ararat, a
rhyodacite volcano," by R. St. J. Lam-
bert (University of Alberta), November
21, 1972.
"Kinetics and mechanisms of exsolu-
tion in the nepheline-kalsilite and diop-
side-enstatite systems," by R. H. McCal-
lister (Purdue University), December 7,
1972.
"Petrology of the Snake River ba-
salts," by G. Stone (Geophysical Labora-
tory) , December 14, 1972.
"Garnet zoning — some diffusion mod-
els", by G. R. Buckley (University of
Illinois) , December 18, 1972.
"Petrologic implications of some co-
rona structures in high-grade metamor-
phism," by W. L. Griffin (University of
Oslo) , January 12, 1973.
GEOPHYSICAL LABORATORY
713
"Duke Island ultramafic complex," by
T. N. Irvine (Geophysical Laboratory),
February 15, 1973.
"Melting of iron at pressures to 160
kbars," by J. Liu (University of Ro-
chester), February 22, 1973/
"Ca-Mg distribution between garnet
and fassaitic clinopyroxene," by S. Banno
(University of Manchester), March 13,
1973.
"An example of hydrothermal evolu-
tion in low-pressure metamorphism : the
Mayres pegmatite (Massif Central,
France)/' by A. Weisbrod (Geophysical
Laboratory), March 22, 1973.
"Petrology and geochemistry of the
ophiolites of the Antalya Napes, Taurus
Range, Turkey. Constitution of the an-
cient Tethysian oceanic crust," by T.
Juteau (Institut National Polytechnique,
France), May 31, 1973.
"Telescope petrology: reflection spec-
troscopy of the surfaces of solar system
objects," by T. McCord (Massachusetts
Institute of Technology), June 21, 1973.
Field Studies
P. M. Bell, F. R. Boyd, T. N. Irvine,
and H. S. Yoder, Jr., participated in a
field trip to the Canadian Northwest
Territories, featuring the Muskox intru-
sion and Coppermine River lavas. The
excursion, which was part of the program
of the 24th International Geological Con-
gress, was led by Irvine and W. R. A.
Baragar of the Geological Survey of
Canada and included 19 participants.
The region is a major area of flood basalt
magmatism, and the Muskox intrusion is
a well-developed stratiform igneous com-
plex of ultramafic, gabbroic, and grano-
phyric rocks.
D. H. Eggler led a field trip to the
Virginia Dale ring-dike complex of Colo-
rado and Wyoming in conjunction with
the Rocky Mountain Section Meeting of
the Geological Society of America. After-
ward, he and Dr. P. Ragland of the Uni-
versity of North Carolina collected sam-
ple material from this Precambrian com-
plex for a geochemical study of magma
mixing and hybridization.
F. N. Hodges participated in a field
trip in connection with the Annual Meet-
ing of the Geological Society of America,
featuring Precambrian igneous and meta-
morphic rocks along the Gunflint Trail
in northern Minnesota. In April he ex-
amined Tertiary alkalic intrusions in the
trans-Pecos region of Texas and New
Mexico in the company of geologists from
the University of Texas, the University
of Toronto, and the Geological Survey of
Canada.
T. C. Hoering participated in a field
trip to Walton County, Florida, where he
collected the humate-cemented sands de-
scribed elsewhere in this Report.
T. N. Irvine spent part of August in
the Juneau area of Alaska collecting
Mesozoic volcanic rocks described in this
Report. In September he attended the
Penrose Field Conference concerned with
the petrology and structural relations of
ophiolites. Five ultramafic complexes in
Oregon and California were visited on
this 10-day excursion sponsored by the
Geological Society of America.
T. E. Krogh did field work in two parts
of Ontario. In the French River area he
collected samples of layered paragneiss
for an investigation of chemical inter-
actions and trace-element redistribution
during amphibolite-facies metamor-
phism; and in the Sudbury area he col-
lected volcanic and intrusive rocks from
the lower Huronian section for age de-
termination.
J. M. Mattinson spent part of July
1972 in Newfoundland, collecting ophio-
litic, granitic, and volcanic rocks for geo-
chronologic studies. The ophiolites of
Newfoundland are among the oldest
known in the world. Mattinson was ac-
companied by W. R. Church of the Uni-
versity of Western Ontario, R. K.
Stevens of Memorial University, New-
foundland, and others.
K. Muehlenbachs journeyed to Iceland
in August 1972, to collect volcanic rocks
714
CARNEGIE INSTITUTION
for oxygen isotopic studies. He visited
areas of acidic volcanism at Snaefellsnes,
Kroksfjordur, Myvatn, and Landmanna-
laugar. In March 1973 he returned to
Iceland to collect hawaiite and mugearite
from the new eruption on Heimaey
Island.
G. T. Stone was co-leader of a field
conference in the Snake River Plain
sponsored by the U.S. Geological Survey
in August. After the conference he did
preliminary collecting of Pliocene to
Holocene basaltic lavas in the eastern
part of the plain and in southeastern
Oregon as part of an expanding study of
the Snake River basaltic province.
A. Weisbrod made a field trip to Ver-
mont in the company of J. B. Thompson,
Jr., and H. D. Holland of Harvard Uni-
versity, to inspect greenschist and upper
amphibolite facies metamorphic series.
H. S. Yoder, Jr. collected at the type
locality of melilite "basalt" at Hochbohl,
near Owen, Germany, and sampled the
eclogites of the spectacular Otztal near
Langenfeld, Austria, in the company of
Professor H. G. Huckenholz (Munich).
He also took part in a field trip led by
Professor W. Wimmenauer (Freiburg)
and sponsored by the Deutsche Mineral-
ogische Gesellschaft to the famous
Kaiserstuhl along the Rhine River. Pro-
fessor Werner Schreyer (Bochum)
showed him the unusual pillow basalts of
Dillenburg, Germany.
Exhibits
Exhibits illustrating some of the cur-
rent research at the Geophysical Labora-
tory were opened to the public in May
1973 at the Carnegie Institution of
Washington's Annual Meeting of the
Board of Trustees. Featured was a study
of the "orange soil" returned by the
Apollo 17 mission. Other subjects in-
cluded lunar pyroxenes, organic sub-
stances in deep-sea sediments, mixing of
two contrasting magmas, melting of gar-
net and spinel peridotites, uranium dis-
tribution between coexisting crystals and
silicate melt, distribution of rubidium
and strontium in Precambrian meta-
morphic rocks, and mineral paragenesis
and distribution of skarn deposits.
Lectures
During the report year Staff Members
and Fellows were invited to present lec-
tures and participated in symposia and
other extracurricular activities as fol-
lows :
P. M. Bell lectured at the Department
of Geological Sciences, Virginia Poly-
technic Institute and State University;
the West Indies Laboratory of Fairleigh
Dickinson University at St. Croix; the
Department of Earth and Space Sciences,
State University of New York at Stony
Brook; the Division of Geological and
Planetary Sciences, California Institute
of Technology; and the Department of
Geological Sciences, Harvard University.
He served as a member of the Lunar Sci-
ence Review Board of the Lunar Science
Institute and as a member of the Lunar
Sample Analysis Planning Team of the
National Aeronautics and Space Admin-
istration Manned Spacecraft Center.
F. R. Boyd, Jr., delivered his presi-
dential address on "Progress and prob-
lems in understanding kimberlites" to
the Geochemical Society at its annual
meeting in November 1972 at Minneap-
olis. He also lectured on "The pyroxene
geotherm" to the Geological Society of
Washington.
D. H. Eggler addressed seminars at the
Geology Departments of the University
of North Carolina, Brooklyn College,
and Colorado State University.
L. W. Finger lectured at the weekly
Chemistry Department seminar series at
the University of Maryland and gave a
week-long series of informal talks at the
Geology Department, University of Cali-
fornia at Davis.
J. D. Frantz gave an invited talk on
"Mineral solution equilibria" at the
weekly Chemistry Department seminar
series at the University of Maryland.
GEOPHYSICAL LABORATORY 715
P. E. Hare lectured on his work on sity, and at the Geological Society of
amino acid diagenesis in deep-sea sedi- Washington.
ments at the Chemistry Department of K. Muehlenbachs participated in the
the University of Maryland. He also Gordon Research Conference on the
reported on the development of tech- Chemistry and Physics of Isotopes at the
niques for determining amino acid iso- Holderness School, Plymouth, New
mers to the Conference for the Analysis Hampshire.
of Carbon in Carbonaceous Chondrites Y. Ohashi gave an invited lecture and
and Returned Lunar Samples held at the seminar at the Department of Geological
Lunar Science Institute in Houston, and Geophysical Sciences, Princeton
Texas. University, on May 21-22, 1973.
B. J. Hensen addressed the Geology M. G. Raymond participated in the
Department at Rutgers University and Gordon Research Conference on the
the Geological Society of Washington. Chemistry and Physics of Solids held at
T. C. Hoering served on the Panel on New Hampton, New Hampshire.
Orientations in Geochemistry of the U.S. M. G. Seitz gave a lecture on "Ura-
National Committee on Geochemistry, nium and thorium partitioning in a sim-
Earth Sciences Division, National Acad- pie basalt system" on April 12, 1973, to
emy of Sciences. He participated in the the Division of Geological and Planetary
Gordon Research Conference on Geo- Sciences, California Institute of Tech-
chemistry held at the Holderness School, nology, and delivered an address to the
Plymouth, New Hampshire, and gave an Geological Society of Washington on
invited paper on "Carbohydrate-amino particle-track registration from natural
acid condensation products as a model samples. He participated in a sympo-
for natural humic acids" at the national sium on the scientific applications of par-
meeting of the American Association of tide tracks in solids at the annual meet-
Petroleum Geologists, held on May 15, ing of the American Nuclear Society in
1973, at Anaheim, California. Las Vegas, Nevada, and joined in a dis-
T. N. Irvine lectured on "The Duke Is- cussion seminar on February 8, 1973, at
land ultramafic complex, Alaska" to the the National Bureau of Standards on
Geological Society of Washington and current analytical studies using particle-
the Department of Geology at the Vir- track etching techniques,
ginia Polytechnic Institute and State G. T. Stone lectured on the "Geology
University. and petrology of the Snake River ba-
T. E. Krogh participated in the Euro- salts" at a seminar at the Department of
pean Colloquium of Geochronology, held Geological Sciences, Virginia Polytech-
in Heidelberg, Germany, in September nic Institute and State University, on
1972, where he presented a paper on "The March 7, 1973.
effects of regional metamorphism on ura- A. Weisbrod gave invited addresses on
nium-lead systems in zircons and on "Equilibrium between feldspars and so-
rubidium-strontium systems in whole lutions in the Mayres pegmatite" to the
rocks." He also lectured to the geochem- Department of Geology, Princeton Uni-
istry group at the Swiss Federal Institute versity, and on "Fluid inclusions and pe-
of Technology at Zurich, Switzerland, trology of a low-pressure metamorphic
and to members of the Geological Sur- pegmatite" to the Department of Geol-
vey of Finland at Helsinki. ogy, Harvard University.
I. Kushiro served as an Associate Edi- H. S. Yoder, Jr., presented a major
tor of the Journal of Geophysical Re- address on "Melilite stability and para-
search. genesis" before the 50th Anniversary
H. K. Mao lectured at the Department Meeting of the Deutsche Mineralogische
of Geological Sciences, Harvard Univer- Gesellschaft in Karlsruhe, Germany. He
716
CARNEGIE INSTITUTION
also spoke at Baldwin-Wallace College,
Berea, Ohio, and at the Harvard Club of
Washington, D.C. His presidential ad-
dress before the Mineralogical Society of
America was entitled "Contemporane-
ous basaltic and rhyolitic magmas." He
served on an Evaluation Panel for the
Inorganic Materials Division of the In-
stitute for Materials Research, National
Bureau of Standards, and the Geosciences
Advisory Panel for the Los Alamos Sci-
entific Laboratory.
D. M. Burt, F. Chayes, B. J. Hensen,
T. N. Irvine, I. Kushiro, R. N. Thomp-
son, and H. S. Yoder, Jr., participated in
the meetings of the International Geo-
logical Congress held in August 1972 at
Montreal, Canada.
J. Akella, D. H. Eggler, B. J. Hensen,
F. N. Hodges, I. Kushiro, M. G. Seitz,
and G. T. Stone participated in the Inter-
national Conference on Distribution and
Partition of Trace Elements and Origin
of Volcanic Rocks held at Newport,
Rhode Island, October 2-4, 1972, under
the sponsorship of the American Geo-
physical Union, the University of Rhode
Island, and the International Association
of Volcanology and Chemistry of the
Earth's Interior.
J. Akella, P. M. Bell, F. R. Boyd, Jr.,
F. N. Hodges, I. Kushiro, H. K. Mao, D.
Virgo, and H. S. Yoder, Jr., participated
in the National Aeronautics and Space
Administration Fourth Lunar Science
Conference held at the Lunar Science In-
stitute near Houston, Texas, from March
5-8, 1973.
Petrolo gists' Club
Three meetings were held during the
62nd year of the Petrologists' Club. The
lectures presented were:
"Lunar spinels: subsolidus reactions
and primary compositional variations,"
by S. E. Haggerty (University of Massa-
chusetts), December 19, 1972.
"Nature of magmas from pyrolite in
island arc and oceanic rift environ-
ments," by D. H. Green (Australian Na-
tional University), March 12, 1973.
"Apollo 17: Taurus Littro revisited,"
by R. Pepin (University of Minnesota),
May 15, 1973.
BIBLIOGRAPHY
Bell, P. M., and H. K. Mao, Crystal-field effects
of iron and titanium in selected grains of
Apollo 12, 14, and 15 rocks, glasses, and fine
fractions, Proc. Third Lunar Sci. Con]., Geo-
chim. Cosmochim. Acta, Suppl. 3, Vol. 1,
The MIT Press, Cambridge, pp. 545-553,
1972 (G.L. Paper 1607).
Bell, P. M., and H. K. Mao, Optical and chem-
ical analysis of iron in Luna 20 plagioclase,
Geochim. Cosmochim. Acta, 37, 755-759, 1973
(G.L. Paper 1621).
Boyd, F. R., see Meyer, H. 0. A.
Bryan, W. B., see Ewart, A.
Burt, D. M., The facies of some Ca-Fe-Si skarns
in Japan, Int. Geol. Congr., 24th, Montreal,
Sect. 2, 284-288, 1972 (G.L. Paper 1614).
Carpenter, B. S., see Seitz, M. G.
Chayes, F., Looking through rocks, in Statistics:
A Guide to the Unknown, J. M. Tanur and
others, eds., Holden-Day, Inc., San Francisco,
pp. 362-371, 1972 (G.L. Paper 1597).
Chayes, F., Effect of the proportion transforma-
tion on central tendency, J. Int. Ass. Math.
Geol. 4, 269-270, 1972 (G.L. Paper 1610).
Ewart, A., and W. B. Bryan, Petrography and
geochemistry of the igneous rocks from Eua,
Tongan Islands, Geol. Soc. Amer. Bull., 83,
3281-3298, 1972 (G.L. Paper 1616).
Finger, L. W., see Prince, E.
Ikeda, Y., see Kushiro, I.
Krogh, T. E., A low-contamination method for
hydrothermal decomposition of zircon and
extraction of U and Pb for isotopic age de-
terminations, Geochim. Cosmochim. Acta,
37, 485-494, 1973 (G.L. Paper 1618).
Kushiro, I., Effect of water on the composition
of magmas formed at high pressures, J. Pe-
trology, 13, 311-334, 1972 (G.L. Paper 1605).
Kushiro, I., Determination of liquidus relations
in synthetic silicate systems with electron
probe analysis: the system forsterite-diop-
side-silica at 1 atmosphere, Amer. Mineral.,
57, 1260-1271, 1972 (G.L. Paper 1608).
Kushiro, I., Origin of some magmas in oceanic
and circum-oceanic regions, Tectonophysics,
17, 211-222, 1973 (G.L. Paper 1620).
Kushiro, I., Y. Ikeda, and Y. Nakamura, Pe-
trology of Apollo 14 high-alumina basalt,
GEOPHYSICAL LABORATORY
717
Proc. Third Lunar Sci. Con]., Geochim. Cos-
mochim. Acta, Suppl. 3, Vol. 1, The MIT
Press, Cambridge, pp. 115-129, 1972 (G.L.
Paper 1606).
Mao, H. K., see Bell, P. M.
Mattinson, J. M., Preparation of hydrofluoric,
hydrochloric, and nitric acids at ultralow lead
levels, Anal Chem., 44, 1715-1716, 1972 (G.L.
Paper 1604).
Mattinson, J. M., Ages of zircons from the
Northern Cascade Mountains, Washington,
Geol. Soc. Amer. Bull, 83, 3769-3783, 1972
(G.L. Paper 1613).
Meyer, H. 0. A., and F. R. Boyd, Composition
and origin of crystalline inclusions in natural
diamonds, Geochim. Cosmochim. Acta, 36,
1255-1273, 1972 (G.L. Paper 1609).
Nakamura, Y., see Kushiro, I.
Prince, E., and L. W. Finger, Use of constraints
on thermal motion in structure refinement of
molecules with libra ting side groups, Acta
Crystallogr., Sect. B, 29, 179-183, 1973 (G.L.
Paper 1617).
Seitz, M. G., R. M. Walker, and B. S. Carpen-
ter, Improved methods for measurement of
thermal neutron dose by the fission-track
technique, J. Awl Phys., 44, 510-512, 1973
(G.L. Paper 1612).
Smith, D., Stability of iron-rich pyroxene in the
system CaSi03-FeSi03-MgSi03, Amer. Min-
eral, 57, 1413-1428, 1972 (G.L. Paper 1611).
Walker, R. M., see Seitz, M. G.
Yoder, H. S., Jr., Contemporaneous basaltic
and rhyolitic magmas, Amer. Mineral, 58,
153-171, 1973 (G.L. Paper 1619).
REFERENCES CITED
Abbott, M. J., Aenigmatite from the ground-
mass of a peralkaline trachyte, Amer. Min-
eral, 52, 1895-1901, 1967.
Adams, J. A. S., P. M. C. Barretto, and R. B.
Clark, Radon-222 loss from zircons and
sphenes: stochastic considerations of discord-
ant lead isotopic ages (abstract), Abstr. with
Programs (Geol. Soc. Amer.), 4, 430, 1972.
Adams, J. B., P. M. Bell, J. E. Conel, H. K.
Mao, T. B. McCord, and D. B. Nash, Visible
and near-infrared transmission and reflec-
tance measurements of the Luna 20 soil,
Geochim. Cosmochim. Acta, 37, 731-743, 1973.
Adler, D., Insulating and metallic states in
transition metal oxides, Solid State Physics,
21, 1-115, 1968.
Ahrens, L. H., The composition of stony mete-
orites, IX, Abundance trends of the refrac-
tory elements in chondrites, basaltic chon-
drites and Apollo 11 fines, Earth Planet. Sci.
Lett., 10, 1-6, 1970.
Ahrens, T. J., T. Takahashi, and G. F. Davis, A
proposed equation of state of stishovite, /.
Geophys. Res., 75, 310-316, 1970.
Akella, J., and F. R. Boyd, Partitioning of Ti
and Al between coexisting silicates, oxides
and liquids, Proc. Fourth Lunar Sci. Con].,
Geochim. Cosmochim. Acta, Suppl. 4, The
MIT Press, Cambridge, in press, 1973.
Allen, J. C, P. J. Modreski, C. Haygood, and
A. L. Boettcher, The role of water in the
mantle of the earth : the stability of amphib-
oles and micas, Int. Geol. Congr., 24th, Mont-
real, Sect. 2, 231-240, 1972.
Alperin, H. A., and E. Prince, A time-shared
computer system for diffractometer control,
J. Res. Nat. Bur. Stand., Sect. C, 74, 89-95,
1970.
Anderson, A. T., Oxidation of the Lablache
Lake titaniferous magnetite deposit, Quebec,
J. Geol, 76, 528-547, 1968.
Anderson, A. T., R. N. Clayton, and T. K.
Mayeda, Oxygen isotope geothermometry of
mafic igneous rocks, J. Geol, 79, 715-729,
1971.
Anderson, D. L., and 0. L. Anderson. The bulk
modulus-volume relationship for oxides, J.
Geophys. Res., 75, 3494-3500, 1970.
Anderson, G. M., and C. W. Burnham, The
solubility of quartz in super-critical water,
Amer. J. Sci., 263, 494-511, 1965.
Atkins, F. B., The pyroxenes of the Bushveld
Igneous Complex, central Transvaal, Ph.D.
thesis, Oxford University, 1965.
Babb, S. E., Jr., Some notes concerning Bridg-
man's Manganin pressure scale, in High
Pressure Measurement, A. A. Giardini and
E. C. Lloyd, eds., Butterworths, Washington,
D.C., pp. 115-124, 1963.
Bailey, D. K., and J. F. Schairer, The system
NazO-ALOs-Fe-Os-S^ at 1 atmosphere, and
pedogenesis of alkaline rocks, J. Petrology, 7 ,
114-170, 1966.
Bailey, E. D., et al, Tertiary and post-Tertiary
geology of Mull, Loch Aline, and Oban,
Mem. Geol. Surv. Scotland, 1924.
Baker, E. H., The calcium oxide-carbon dioxide
system in the pressure range of 1-300 atmos-
pheres, J. Chem. Soc. London, 464-470. 1962.
Bancroft, G. M., R. G. Burns, and A. G. Mad-
dock, Applications of the Mossbauer effect to
silicate mineralogy, Part 1, Iron silicates of
known crystal structure, Geochim. Cosmo-
chim. Acta, 31, 2219-2246, 1967.
Bancroft, G. M., P. G. L. Williams, and R. G.
Burns, Mossbauer spectra of minerals along
718
CARNEGIE INSTITUTION
the diopside-hedenbergite tie line, Amer.
Mineral, 56, 1617-1625, 1971.
Banks, P. O., and L. T. Silver, Evaluation of
the decay constant of uranium-235 from lead
isotope ratios, J. Geophys. Res.,71 , 4037-4046,
1966.
Banks, R. J., Geomagnetic variations and the
electrical conductivity of the upper mantle,
Geophys. J. Roy. Astron. Soc, 17, 457-487,
1969.
Banno, S., Garnet-pyroxene equilibrium in
granulite facies rocks and inclusions in kim-
berlite and alkali basalt, Jap. J. Geol. Geogr.,
36, 23-35, 1965.
Banno, S., Classification of eclogites in terms
of physical conditions of their origin, Phys.
Earth Planet. Interiors, 3, 405-421, 1970.
Banno, S., and Y. Matsui, Eclogite types and
partition of Mg, Fe, and Mn between clino-
pyroxene and garnet, Proc. Jap. Acad., 41,
716-721, 1965.
Bassett, W. A., and L. Ming, Disproportiona-
tion of Fe2Si04 to 2FeO + Si04 at pressures
up to 250 kbar and temperatures up to
3000° C, Phys. Earth Planet. Interiors, 6, 154-
160, 1972.
Bassett, W. A., and T. Takahashi, Dispropor-
tionate of Fe2Si04 to 2FeO + Si02 at high
pressure and temperature (abstract), Trans.
Amer. Geophys. Union, 51, 828, 1970.
Bell, P. M., and H. K. Mao, Crystal-field effects
of iron and titanium in selected grains of
Apollo 12, 14, and 15 rocks, glasses and fine
fractions, Proc. Third Lunar Sci. Con]., Geo-
chim. Cosmochim. Acta, Suppl. 3, Vol. 1, The
MIT Press, Cambridge, pp. 545-553, 1972.
Bell, P. M., and H. K. Mao, Optical and chem-
ical analysis of iron in Luna 20 plagioclase,
Geochim. Cosmochim. Acta, 37, 755-759, 1973.
Belozerskii, G. N., V. N. Gitsovitch, A. N.
Murin, Yu. P. Khimich, and Yu. M. Yakov-
lev, Determination of the parameters of
quadrupole interaction in yttrium iron gar-
nets, Pis'ma Zh. Eksp. Teor. Fiz., 11, 173-177,
1970 [JETP Lett., 11, 106-108, 1970].
Bence, A. E., and J. J. Papike, Pyroxenes as
recorders of lunar basalt petrogenesis : chem-
ical trends due to crystal-liquid interaction,
Proc. Third Lunar Sci. Con]., Geochim. Cos-
mochim. Acta, Suppl. 3, Vol. 1, The MIT
Press, Cambridge, pp., 431-469, 1972.
Berg, H. C, D. L. Jones, and D. H. Richter,
Gravina-Nutzotin belt — Tectonic significance
of an Upper Mesozoic sedimentary and vol-
canic sequence in southern and southeastern
Alaska, U.S. Geol. Surv. Prof. Pap., 800-D,
D1-D24, 1972.
Biggar, G. M., and M. J. O'Hara, Solid solutions
at atmospheric pressure in the system CaO-
MgO-Si02 with special reference to the insta-
bilities of diopside, akermanite, and monti-
cellite, Progress in Experimental Petrology,
Natural Environment Research Council,
Manchester: Edinburgh, First Report, 86-96,
1969.
Birch, F., and P. LeComte, Temperature-
pressure plane for albite composition, Amer.
J. Sci., 258, 209-217, 1960.
Birle, J. D, G. V. Gibbs, P. B. Moore, and
J. V. Smith, Crystal structure of natural
olivines, Amer. Mineral, 53, 807-824, 1968.
Boettcher, A. L., The system CaO-Al203-Si02-
H20 at high pressures and temperatures, J .
Petrology, 11, 337-379, 1970.
Boettcher, A. L., B. O. Mysen, and J. C. Allen,
Techniques for the control of water fugacity
and oxygen fugacity for experimentation in
solid-media high-pressure apparatus, J . Geo-
phys. Res., in press, 1973.
Boettcher, A. L., and P. J. Wyllie, The calcite-
aragonite transition measured in the system
CaO-C02-H20, /. Geol, 76, 314-330, 1968.
Boettcher, A. L., and P. J. Wyllie, The system
CaO-Si02-C02-H20 ; III, Second critical end-
point on the melting curve, Geochim. Cosmo-
chim. Acta, 33, 611-632, 1969a.
Boettcher, A. L., and P. J. Wyllie, Phase rela-
tionships in the system NaAlSi04-Si02-H20
to 35 kilobars pressure, Amer. J . Sci., 267,
875-909, 1969b.
Boon, J. A., Mossbauer investigations in the
system Na20-FeO-Si02, Chem. Geol, 7, 153-
169, 1971.
Boon, J. A., and W. S. Fyfe, The coordination
number of ferrous ions in silicate glasses,
Chem. Geol, 10, 287-298, 1972.
Bowen, N. L., The crystallization of haplo-
basaltic, haplodioritic and related magmas,
Amer. J. Sci., Jfi, 161-185, 1915.
Bowen, N. L., Progressive metamorphism of
siliceous limestone and dolomite, J . Geol, 48,
225-274, 1940.
Bowen, N. L., Phase equilibria bearing on the
origin and differentiation of alkaline rocks,
Amer. J. Sci., 243, 75-89, 1945.
Bowen, N. L., and J. F. Schairer, The system
MgO-FeO-Si02, Amer. J. Sci., 29, 151-217,
1935.
Bown, M. G., A new amphibole polymorph in
intergrowth with tremolite: clino-anthophyl-
lite? (abstract), Amer. Mineral, 51, 259-260,
1965.
Boyd, F. R., Garnet peridotites and the system
CaSiOa-MgSiOa-AlaOs, Mineral. Soc. Amer.
Spec. Pap., 3, 63-75, 1970.
Boyd, F. R., The pyroxene geotherm, Geo-
chim. Cosmochim. Acta, in press, 1973.
GEOPHYSICAL LABORATORY
719
Boyd, F. R., and J. L. England, Apparatus for
phase-equilibrium measurements at pressures
up to 50 kb and temperatures up to 1750 °C,
J. Geophys. Res., 65, 741-748, 1960.
Boyd, F. R., and J. L. England, Effect of pres-
sure on the melting points of diopside,
CaMgSbOe, and albite, NaAlSisOs, in the
range up to 50 kilobars, J. Geophys. Res., 68,
311-323, 1963.
Boyd, F. R., and P. H. Nixon, Origin of the
ilmenite-silicate nodules in kimberlites from
Lesotho and South Africa, in Lesotho Kim-
berlites, P. H. Nixon, ed., Cape Town, in
press, 1973.
Boyd, F. R., and J. F. Schairer, The system
MgSiOa-CaMgSisOe, J. Petrology, 5, 275-309,
1964.
Boyd, F. R., and D. Smith, Compositional zon-
ing in pyroxenes from lunar rock 12021,
Oceanus Procellarum, J. Petrology, 12, 439-
464, 1971.
Bradley, R. S., A. K. Jamil, and D. C. Munro,
The electrical conductivity of olivine at high
temperatures and pressures, Geochim. Cos-
mochim. Acta, 28, 1669-1678, 1964.
Briden, J. C, Recurrent continental drift of
Gondwanaland, Nature, 215, 1334-1339, 1967.
Brown, G. E., J. J. Papike, and C. T. Prewitt,
The relationship of domain structure in
pigeonites to thermal history and the P2i/c
<=^ C2/c phase transition (abstract), Abstr.
with Programs (Geol Soc. Amer.), 4, 458-459,
1972.
Brown, G. E., and C. T. Prewitt, High-tempera-
ture crystal chemistry of hortonolite, Amer.
Mineral, 58, 577-587, 1973.
Brown, G. E., C. T. Prewitt, J. J. Papike, and
S. Sueno, A comparison of the structures of
low and high pigeonite, J. Geophys. Res., 77,
577^-5789, 1972.
Brown, G. E., and B. A. Wechsler, Crystallog-
raphy of pigeonites from basaltic vitrophyre
15597, Proc. Fourth Lunar Sci. Con]., Geo-
chim. Cosmochim. Acta, Suppl. 4, The MIT
Press, Cambridge, in press, 1973.
Brown, G. M., Pyroxenes from the early and
middle stages of fractionation of the Skaer-
gaard intrusion, east Greenland, Mineral.
Mag., 31, 511-543, 1957.
Burnett, D., M. Monnin, M. Seitz, R. Walker,
and D. Yuhas, U-Th distributions and fission-
track dating of lunar samples, Proc.
Second Lunar Sci. Con]., Geochim. Cosmo-
chim. Acta, Suppl. 2, Vol. 2, The MIT Press,
Cambridge, pp. 1503-1519, 1971.
Burnham, C. W., and N, F. Davis, The role of
HaO in silicate melts; I, P-V-T relations in
the system NaAlSiaOs-Hi-O to 10 kilobars and
1000° C, Amer. J. Sci., 270, 54-79, 1971.
Burnham, C. W., J. R. Holloway, and N. F.
Davis, Thermodynamic properties of water
to 100 °C and 10,000 bars, Geol. Soc. Amer.
Spec. Pap., 132, 1970.
Burnham, C. W., Y. Ohashi, S. S. Hafner, and
D. Virgo, Cation distribution and atomic
thermal vibrations in an iron-rich ortho-
pyroxene, Amer. Mineral., 56, 850-876, 1971.
Burns, R. G., Mineralogical Applications o]
Crystal-Field Theory, Cambridge University
Press, Cambridge, 1970.
Burns, R. G., M. G. Clark, and A. J. Stone,
Vibronic polarization in the electronic spectra
of gillespite, a mineral containing iron (II)
in square planar coordination, Inorg. Chem.,
5, 1268-1272, 1966.
Bush, W. R., S. S. Hafner, and D. Virgo, Some
ordering of iron and magnesium at the octa-
hedrally coordinated sites in a magnesium-
rich olivine, Nature, 227, 1339-1341, 1970.
Busing, W. R., R. E. Ellison, H. A. Levy, S. P.
King, and R. D. Roseberry, The Oak Ridge
computer-controlled X-ray diffractometer,
ORNL-4143, Oak Ridge National Laboratory,
Oak Ridge, Tennessee, 1968.
Butler, P., Mineral compositions and equilibria
in the metamorphosed iron formation of the
Gagnon region, Quebec, Canada, /. Petrology,
10, 56-101, 1969.
Cameron, A. E., D. H. Smith, and R. L. Walker,
Mass spectrometry of nanogram-size samples
of lead, Anal. Chem., 41, 525-526, 1969.
Cameron, K. L., J. J. Papike, A. E. Bence, and
S. Sueno, Petrology of fine-grained rock frag-
ments and petrologic implications of single
crystals from Luna 20 soil, Geochim. Cosmo-
chim. Acta, 37, 775-793, 1973.
Cameron, M., S. Sueno, C. T. Prewitt, and J. J.
Papike, High-temperature crystal chemistry
of acmite, diopside, hedenbergite, jadeite,
spodumene, and ureyite, Amer. Mineral., 58,
594-618, 1973.
Cannillo, E., F. Mazzi, J. H. Fang, Paul D.
Robinson, and Y. Ohya, The crystal structure
of aenigmatite, Amer. Mineral., 56, 427^446,
1971.
Carmichael, D. M., On the mechanism of pro-
grade metamorphic reactions in quartz-bear-
ing pelitic rocks, Contrib. Mineral. Petrol.,
20, 266-267, 1969.
Carmichael, I. S. E., The petrology of Thing-
muli, a Tertiary volcano in eastern Iceland,
J. Petrology, 5, 435-460, 1964.
Champness, P. E., and G. W. Lorimer, An elec-
tron microscopic study of a lunar pyroxene,
Contrib. Mineral. Petrol, 33, 171-183, 1971.
Chayes, F., On ratio correlation in petrographv,
J. Geol, 57, 239-254, 1949.
720
CARNEGIE INSTITUTION
Chayes, F., Ratio Correlation, The University
of Chicago Press, 1971.
Chayes, F., Silica saturation in Cenozoic basalt,
Phil. Trans. Roy. Soc. London, Ser. A, 271,
285-296, 1972.
Cnen, J., Petrology and chemistry of garnet
lherzolite nodules in kimberlite from South
Africa, Amer. Mineral., 56, 2098-2110, 1971.
Chtdos, A. A., and A. L. Albee, Quantitative
microprobe analysis and data reduction using
an on-line mini computer (abstract), Pro-
ceedings of the Sixth National Conference on
Electron Probe Analysis, Electron Probe So-
ciety of America, p. 15A, 1971.
Church, S. E., Lead and strontium isotopes and
their bearing on the genesis of andesitic
magmas in the Cascades, unpublished Ph.D.
dissertation, University of California, Santa
Barbara, 1970.
Clark, J. R., D. E. Appleman, and J. J. Papike,
Crystal-chemical characterization of clinopy-
roxenes based on eight new structure refine-
ments, Mineral. Soc. Amer. Spec. Pap., 2, 31-
50, 1969.
Clark, J. R., M. Ross, and D. E. Appleman,
Crystal chemistry of a lunar pigeonite, Amer.
Mineral., 56, 888-908, 1971.
Clark, S. P., Jr., Absorption spectra of some
silicates in the visible and near infrared,
Amer. Mineral., 42, 733-742, 1957.
Clark, S. P., Jr., and A. E. Ringwood, Density
distribution and constitution of the mantle,
Rev. Geophys., 2, 35-68, 1964.
Clarke, R. S., Jr., E. Jarosewich, B. Mason, J.
Nelen, M. Gomez, and J. R. Hyde, The
Allende, Mexico, meteoritic shower, Smith-
son. Contrib. Earth Sci., 5, 1970.
Clayton, R. N., and T. K. Mayeda, The use of
bromine pentafluoride in the extraction of
oxygen from oxides and silicates for isotopic
analysis, Geochim. Cosmochim. Acta, 27, 43-
52, 1963.
Clementi, E., Ab initio computations in atoms
and molecules, IBM J. Res. Develop., 9, 2-19,
1965.
Cohen, M. L., P. J. Lin, D. M. Roessler, and
W. C. Walker, Ultraviolet optical properties
and electronic band structure of magnesium
oxide, Phys. Rev., 155, 992-996. 1967.
Coleman, R. G., D. E. Lee, L. B. Beatty, and
W. W. Brannock, Eclogites and eclogites:
their differences and similarities, Geol. Soc.
Amer. Bull, 76, 483-508, 1965.
Coombs, D. S., Trends and affinities of basaltic
magmas and pyroxenes as illustrated on the
diopside-olivine-silica diagram, Mineral. Soc.
Amer. Spec. Pap., 1, 227-250, 1963.
Corbato, F. J., and A. C. Switendick, Integrals
for diatomic molecular calculations, in Meth-
ods in . Computational Physics, Vol. 2, B.
Adler and S. Fernbach, eds., Academic Press,
New York, pp. 155-179, 1963.
Corfield, P. W. R., A computer-controlled
diffractometer (abstract), Acta Crystallogr.,
Sect. A, 25, Suppl., S71, 1969.
Craig, H., Standard for reporting concentrations
of deuterium and oxygen-18 in natural waters,
Science, 133, 1833-1834, 1961.
Currie, K. L., The reaction 3 cordierite ^ 2
garnet + 4 sillimanite + 5 quartz as a geo-
logical thermometer, Contrib. Mineral. Pe-
trol, 33, 215-226, 1971.
Dallwitz, W. B., D. H. Green, and J. E. Thomp-
son, Clinoenstatite in a volcanic rock from
the Cape Vogel area, Papua, J. Petrology, 7,
375-403, 1966.
Davis, B. T. C, and F. R. Boyd, The join
MgaSiaOe-CaMgSizOe at 30 kilobars and its
application to pyroxenes from kimberlites,
J. Geophys. Res., 71, 3567-3576, 1966.
Davis, B. T. C, and J. L. England, The melting
of forsterite up to 50 kilobars, J. Geophys.
Res., 69, 1113-1116, 1964.
Dawson, J. B., The genesis of kimberlite, Dia-
mond Res., 71, 2-8, 1971.
Dawson, J. B., and J. B. Hawthorne, Mag-
matic sedimentation and carbonatitic differ-
entiation in kimberlite sills at Benfontein,
South Africa, /. Geol. Soc. London, 129, 61-
85, 1973.
Dawson, J. B., and D. G. Powell, Mica in the
upper mantle, Contrib. Mineral. Petrol, 22,
233-237, 1969.
Dawson, J. B., and A. M. Reid, A pyroxene-
ilmenite intergrowth from the Monastery
mine, South Africa, Contrib. Mineral. Petrol,
26, 296-301, 1970.
DeMaria, G., G. Balducci, M. Guido, and V.
Piacente, Mass spectrometric investigation of
the vaporization process of Apollo 12 lunar
samples, Proc. Second Lunar Sci. Conf., Geo-
chim. Cosmochim. Acta, Suppl. 2, Vol. 2, The
MIT Press, Cambridge, pp. 1367-1380, 1971.
Dempster, A. N., and R. Richard, Regional
geology and structure, in Lesotho Kimber-
lites, P. H. Nixon, ed., Cape Town, in press,
1973.
Dietz, R. S., and J. C. Holden, The breakup of
Pangea, Sci. Amer., 223, 30-41, 1970.
Dobson, D. C, J. W. Linnett, and M. M. Rah-
man, Mossbauer studies of the charge trans-
fer process in the system ZnxFe3 - *04, J. Phys.
Chem. Solids, 31, 2727-2733, 1970.
Donnelly, T. W., J. J. W. Rogers, P. Pushkan,
and R. L. Armstrong, Chemical evolution of-
the igneous rocks of the eastern West Indies :
An investigation of thorium, uranium, and
GEOPHYSICAL LABORATORY
721
potassium distributions, and lead and stron-
tium isotopic ratios, Geol. Soc. Amer. Mem.,
130, 181-224, 1971.
Dowty, E., and. D. H. Lindsley, Mossbauer
spectra of synthetic hedenbergite-ferrosilite
pyroxenes, Amer. Mineral., 58, 850-868, 1973.
Dowty, E., M. Ross, and F. Cuttitta, Fe2+-Mg
site distribution in Apollo 12021 clinopyrox-
enes: evidence for bias in Mossbauer meas-
urements, and relation of ordering to exsolu-
tion, Proc. Third Lunar Sci. Con]., Geochim.
Cosmochim. Acta, Suppl. 3, Vol. 1, The MIT
Press, Cambridge, pp. 481^92, 1972.
Drickamer, H. G., and C. W. Frank, Electronic
Transitions and the High Pressure Chemistry
and Physics of Solids, Chapman and Hall,
London, 1973.
Duba, A., Electrical conductivity of olivine, J.
Geophys. Res., 77, 2483-2495, 1972.
Dundon, R. W., and S. S. Hafner, Cation dis-
order in shocked orthopyroxenes, Science,
174, 581-583, 1971.
Eberhard, E., Michbarkeit und strukturelle
Eigenschaften von synthetischem Cordierit in
den Systemen (MnO, FeO, MgO, CaO)-
ALOa-SiC^-H^O bei verhaltnismassig niedrigen
Temperaturen und Drucken, Z. Kristallogr.,
117, 171-183, 1962.
von Eckermann, H., The petrogenesis of the
Alno alkaline rocks, Bull. Geol. Inst. Uppsala,
40, 25-36, 1961.
Eckhart, D., K. Larsen, and T. Madden, Long-
period magnetic fluctuations and mantle elec-
trical conductivity estimates, J. Geophys.
Res., 68, 6279-6286, 1963.
Eggler, D. H., Water-saturated and undersatu-
ated melting relations in a Paricutin andesite
and an estimate of water content in the
natural magma, Contrib. Mineral. Petrol., 34,
261-271, 1972a.
Eggler, D. H., Amphibole stability in H20-
undersaturated calc-alkaline melts, Earth
Planet. Sci. Lett., 15, 28-34, 1972b.
El Goresy, A., P. Ramdohr, and 0. Medenbach,
Mineralogy of Apollo 16 sample 66095, Earth
Planet. Sci. Lett., in press, 1973.
Emslie, R. F., The geology of the Michikamau
intrusion, Labrador, Can. Geol. Surv. Pap.,
68-57, 85 pp., 1970.
Enders, C, Wie entsteht der Humus in der
Natur? Die Chemie, 56, 281-292, 1943.
Enders, C, and K. Theis, Die Melanoidine und
ihre Beziehung zu den Huminsauren,
Brennst.-Chem., 19, 402-407, 1938.
Ernst, W. G., Synthesis, stability relations, and
occurrence of riebeckite and riebeckite-
arfvedsonite solid solutions, J. Geol., 70, 689-
736, 1962.
Euler, F., and J. A. Bruce, Oxygen coordinates
of compounds with garnet structure, Acta
Crystallogr., 19, 971-978, 1965.
Evans, B. J., and L. Leung, Fe5T Mossbauer stud-
ies of tektites (abstract), Trans. Amer. Geo-
phys. Union, 54, 350, 1973.
Evans, B. W., Pyrope granet^piezometer or
thermometer? Geol. Soc. Amer. Bull., 76,
1295-1300, 1965.
Ewart, A., and W. B. Bryan, Petrography and
geochemistry of the igneous rocks from
Eua, Tongan Islands, Geol. Soc. Amer. Bull.,
83, 3281-3298, 1972.
Faure, G., P. M. Hurley, and J. L. Powell, The
isotopic composition of strontium in surface
water from the North Atlantic Ocean, Geo-
chim. Cosmochim. Acta, 29, 209-220, 1965.
Findlay, D. C, Origin of the Tulameen ultra-
mafic-gabbro complex, southern British Co-
lumbia, Can. J. Earth Sci., 6, 399-425, 1969.
Fink, L. K., Jr., Evidence for the antiquity of
the Lesser Antilles island arc, Trans. Amer.
Geophys. Union, 51, 326-427, 1970.
Fink, L. K., Jr., Bathymetric and geologic
studies of the Guadeloupe region, Lesser An-
tilles island arc, Marine Geol, 12, 267-288,
1972.
Fisher, G. W., The application of ionic equi-
libria to metamorphic differentiation: an
example, Contrib. Mineral. Petrol., 29, 91-
103, 1970.
Fisher, J. R., and E-an Zen, Thermochemical
calculations from hydrothermal phase equi-
librium data and the free energy of H20,
Amer. J. Sci., 270, 297-314, 1971.
Fleming, E. H., Jr., A. Ghiorso, and B. B.
Cunningham, The specific alpha-activities
and half-lives of U234, U235, and U236, Phys.
Rev., 88, 642-652, 1952.
Fox, P. J., E. Schreiber, and B. C. Heezen, The
geology of the Caribbean crust: Tertiary
sediments, granitic and basic rocks from the
Aves Ridge, Tectonophysics, 12, 89-109, 1971.
Friedman, I., and R. L. Smith, The deuterium
content of water in some volcanic glasses,
Geochim. Cosmochim. Acta, 15, 218-228, 1951.
Garlick, D. G., Oxygen isotope fractionation in
igneous rocks. Earth Planet. Sci. Lett., 1,
361-368, 1966.
Gibb, F. G. F., Clinopyroxene crystallization
from basic magma of alkaline affinity, Nature
(London), Phys. Sci., 230, 153-154, 1971.
Giletti, B. J., and J. L. Kulp, Radon leakage
from radioactive minerals, Amer. Mineral.,
40, 481-496, 1955.
Goldich, S. S., C. E. Hedge, and T. W. Stern,
Age of the Morton and Montevideo gneisses
and related rocks, southwestern Minnesota,
Geol. Soc. Amer. Bull, 81, 3671-3696, 1970.
722
CARNEGIE INSTITUTION
Goldstein, J. I., and H. Yakowitz, Metallic in-
clusions and metal particles in the Apollo 12
lunar soil, Proc. Second Lunar Sci. Con].,
Geochim. Cosmochim. Acta, Suppl. 2, Vol. 1,
The MIT Press, Cambridge, 1971.
Green, D. H., Composition of basaltic magmas
as indicators of conditions of origin; applica-
tion to oceanic volcanism, Phil. Trans. Roy.
Soc. London, Ser. A, 268, 707-725, 1971.
Green, D. H., Magmatic activity as the major
process in the chemical evolution of the
earth's crust and mantle, Tectonophysics, 13,
42-71, 1972.
Green, D. H., Contrasted melting relations in
a pyrolite upper mantle under mid-oceanic
ridge, stable crust and island arc environ-
ments, Tectonophysics, 17, 285-297, 1973a.
Green, D. H., Conditions of melting of basanite
magma from garnet peridotite, Earth Planet.
Sci. Lett., 17, 456-465, 1973b.
Green, D. H., and A. E. Ringwood, The genesis
of basaltic magmas, Contrib. Mineral. Pe-
trol, 15, 103-190, 1967a.
Green, D. H., and A. E. Ringwood, The sta-
bility fields of aluminous pyroxene, peridotite
and garnet peridotite and their relevance in
upper mantle structure, Earth Planet. Sci.
Lett., 8, 151-160, 19676.
Green, D. H., and A. E. Ringwood, An experi-
mental investigation of the gabbro to eclogite
transformation and its petrological applica-
tions, Geochim. Cosmochim. Acta, 31, 767-
833, 1967c.
Green, D. H., and A. E. Ringwood, Mineralogy
of peridotitic compositions under upper man-
tle conditions, Phys. Earth Planet. Interiors,
3, 359-371, 1970.
Green, H. W., A C02 charged asthenosphere,
Nature, 238, 2-5, 1972.
Grieve, R. A. F., and A. G. Plant, 64455, an
ellipsoidal glass-coated highland basalt pro-
jectile (abstract), in Lunar Science IV, J. W.
Chamberlain and C. Watkins, eds., Lunar
Science Institute, Houston, pp. 317-319, 1973.
Grove, T. L., D. Walker, J. Longhi, E. N.
Stolper, and J. F. Hays, Petrology of rock
12002 from Oceanus Procellarum (abstract),
in Lunar Science IV, J. W. Chamberlain and
C. Watkins, eds., Lunar Science Institute,
Houston, pp. 323-325, 1973.
Grover, J. E., D. H. Lindsley, and A. C. Turn-
ock, Ca-Mg-Fe pyroxenes: subsolidus phase
relations in iron-rich portions of the pyroxene
quadrilateral, Abstr. with Programs (Geol.
Soc. Amer.), 5, 521-522, 1973.
Hafner, S. S., and H. G. Huckenholz, Moss-
bauer spectrum of synthetic ferridiopside,
Nature (London) Phys. Sci., 233, 9-11, 1971.
Hafner, S. S., B. Janik, and D. Virgo, State and
location of iron in Apollo 11 samples, in
Mossbauer Effect Methodology, Vol. 6, I. J.
Gruverman, ed., Plenum Press, New York,
pp. 193-207, 1971.
Hafner, S. S., D. Virgo, and D. Warburton,
Oxidation state of iron in plagioclase from
lunar basalts, Earth Planet. Sci. Lett., 12,
159-166, 1971a.
Hafner, S. S., D. Virgo, and D. Warburton,
Cation distribution and cooling history of
clino pyroxenes from Oceanus Procellarum.
Proc. Second Lunar Sci. Conf., Geochim.
Cosmochim. Acta, Suppl. 2, Vol. 1, The MIT
Press, Cambridge, pp. 91-108, 19716.
Hald, N., A. Noe-Nygaard, and A. K. Peder-
sen, The Kroksfjordur central volcano in
north-west Iceland, Acta Naturalia Islandica,
2, no. 10, 29 pp., 1971.
Hallimond, A. F., Optically uniaxial augite
from Mull, Mineral. Mag., 17, 97-99, 1914.
Hamil, M. M., S. Ghose, and R. A. Sparks,
Anti-phase domains in a lunar pigeonite:
determination of the average shape, size and
orientation by X-ray diffraction (abstract),
Trans. Amer. Geophys. Union, 54, 499, 1973.
Hamilton, D. L., C. Wayne Burnham, and
E. F. Osborn, The solubility of water and
effects of oxygen fugacity and water content
on crystallization in mafic magmas, J. Pe-
trology, 5, 21-39, 1964.
Hamilton, R. M., Temperature variation at
constant pressures of the electrical conduc-
tivity of periclase and olivine, J. Geophys.
Res., 70, 5679-5692, 1965.
Hamilton, W. C, On the isotropic temperature
factor equivalent to a given anisotropic tem-
perature, Acta Crystallogr., 12, 609-610, 1959.
Hamilton, W. C, A simple diffractometer looks
at experimental design (abstract), American
Crystallographic Association, Minneapolis
Meeting, Program and Abstracts, 34, 1967.
Hare, P. E., Automatic multiple column amino
acid analysis — the use of pressure elution in
small bore ion-exchange columns, Fed. Proc,
Fed. Amer. Soc. Exp. Biol, 25, 709, 1966.
Hare, P. E., Geochemistry of proteins, peptides,
and amino acids, in Organic Geochemistry,
G. Eglinton and M. T. J. Murphy, eds.,
Springer-Verlag New York Inc., 438-463, 1969.
Hare, P. E., Ion exchange chromatography in
lunar organic analysis, Space Life Sci., 3, 354-
359, 1972.
Hasebe, K., N. Fujii, and S. Uyeda, Thermal
processes under island arcs, Tectonophysics,
10, 335-355, 1970.
Heald, E. F., J. Naughton, and I. L. Barnes,
The chemistry of volcanic gases; use of equi-
librium calculations in the interpretation of
GEOPHYSICAL LABORATORY
723
volcanic gas samples, J. Geophys. Res., 68,
545-557, 1963.
Hedge, C. E., and J. F. Lewis, Isotopic com-
position of strontium in three basalt-andesite
centers along the Lesser Antilles arc, Contrib.
Mineral. Petrol, 32, 39-47, 1971.
Hemley, J. J., Some mineralogical equilibria in
the system KzO-ALOs-SiCVILO, Amer. J.
Sci., 257, 241-270, 1959.
Hensen, B. J., and D. H. Green, Experimental
study of cordierite and garnet in pelitic com-
positions at high pressures and temperatures;
I, Compositions with excess alumina-silicate,
Contrib. Mineral. Petrol, 33, 309-330, 1971.
Hermance, J. F., and L. R. Grillot, Correlation
of magneto-telluric, seismic, and temperature
data from southwest Iceland, J. Geophys.
Res., 75, 6582-6591, 1970.
Herzenberg, C. L., R. B. Moler, and D. L.
Riley, Mossbauer instrumental analysis of
Apollo 12 lunar rock and soil samples, Proc.
Second Lunar Sci. Con]., Geochim. Cosmo-
chim. Acta, Suppl. 2, Vol. 3, The MIT Press,
Cambridge, pp. 2103-2123, 1971.
Hess, H. H., Pyroxenes of common mafic mag-
mas, Parts I and II, Amer. Mineral, 26, 515-
535, 573-594, 1941.
Hill, D. P., and L. C. Pakiser, Seismic refrac-
tion study of crustal structure between the
Nevada test site and Boise, Idaho, Geol. Soc.
Amer. Bull, 78, 685-704, 1967.
Hill, R. E. T., and A. L. Boettcher, Water in
the earth's mantle; melting curves of basalt-
water and basalt-water-carbon dioxide, Sci-
ence, 167, 980-981, 1970.
Hlabse, T., and O. J. Kleppa, The thermo-
chemistry of jadeite, Amer. Mineral, 53,
1281-1292, 1968.
Hodge, J. E., Chemistry of browning reactions
in model systems, Agr. Food Chem., 1, 928—
943, 1953.
Hodges, F. N., and I. Kushiro, Petrology of
Apollo 16 lunar highland rocks, Proc. Fourth
Lunar Sci. Con]., Geochim. Cosmochim. Acta,
Suppl. 4, The MIT Press, Cambridge, in
press, 1973.
Hofmann, Albrecht, Chromatographic theory of
infiltration metasomatism and its application
to feldspars, Amer. J. Sci., 272, 69-90, 1972.
Holloway, J. R., The system pargasite-ELO-
C02: a model for melting of a hydrous min-
eral with a mixed-volatile fluid; I, Experi-
mental results to 8 kbar, Geochim. Cosmo-
chim. Acta, 37, 651-666, 1973.
Holloway, J. R., and C. W. Burnham, Melting
relations of basalt with equilibrium water
pressure less than total pressure, J. Petrology,
13, 1-29, 1972.
Holmes, A., A contribution to the petrology of
kimberlite and its inclusions, Trans. Geol
Soc. S. Africa, 39, 379-428, 1937.
Hoppe, H. J., Chemische und mikroskopische
Untersuchungen an islandischen Gesteinen,
Chem. Erde, 2, 42^-437, 1938.
Housley, R. M., and R. W. Grant, Electrostatic
forces, molecular-orbital theory, and the elec-
tric field gradient in yttrium iron garnet,
Phys. Rev. Lett., 29, 203-206, 1972.
Housley, R. M., R. W. Grant, A. N. Muir, Jr.,
M. Blander, and M. Abdel-Gawad, Moss-
bauer studies of Apollo 12 samples, Proc.
Second Lunar Sci. Con]., Geochim. Cosmo-
chim. Acta, Suppl. 2, Vol. 3, The MIT Press,
Cambridge, pp. 2125-2136, 1971.
Hsu, L. C, Selected phase relationships in the
system Al-Mn-Fe-Si-O-H : a model for gar-
net equilibria, J. Petrology, 9, 60-83, 1968.
Huebner, J. S., M. Ross, and N. Hickling, Par-
tial melting of pyroxenes and origin of Mare
basalts (abstract), in Lunar Science IV, J. W.
Chamberlain and C. Watkins, eds., Lunar
Science Institute, Houston, 397, 1973.
Hughes, H., The electrical conductivity of the
earth's interior, Ph.D. thesis, University of
Cambridge, Cambridge, England, 1953.
Humphries, D. J., G. M. Biggar, and M. J.
O'Hara, Phase equilibria and origin of Apollo
15 basalts etc., in The Apollo 15 Lunar Sam-
ples, J. W. Chamberlain and C. Watkins,
eds., Lunar Science Institute, Houston, 103-
107, 1972.
Iiyama, J. T., Recherches sur le role de l'eau
dans la structure et le polymorphisme de la
cordierite, Bull Soc. Fr. Mineral. Cristallogr.,
83, 155-178, 1960.
Ingalls, R., Electric-field gradient tensor in fer-
rous compounds, Phys. Rev., A, 133, 787-794,
1964.
Irvine, T. N., The ultramafic complex and re-
lated rocks of Duke Island, southeastern
Alaska, Ph.D. thesis, California Institute of
Technology, Pasadena, 1959.
Irvine, T. N., The Duke Island ultramafic com-
plex, southeastern Alaska, in Ultramafic and
Related Rocks, P. J. Wyllie, ed., John Wiley
and Sons, Inc., New York, pp. 84—96, 1967.
Irvine, T. N., Petrology of the Duke Island
ultramafic complex, southeastern Alaska,
Geol. Soc. Amer. Mem., 138, in press, 1973.
Irvine, T. N., and W. R. A. Baragar, A guide to
the chemical classification of the common
volcanic rocks, Can. J. Earth Sci., 8, 523-548,
1971.
Irving, A. J., and P. J. Wyllie, Melting rela-
tionships in CaO-MgO-C02 compared with
the solidus for mantle peridotite (abstract),
Trans. Amer. Geophys. Union, 54, 479, 1973.
724
CARNEGIE INSTITUTION
Ito, T., On the symmetry of the rhombic py-
roxenes, Z. Kristallogr., 90, 151-162, 1935.
Ito, T., X-Ray .Studies on Polymorphism, Ma-
ruzen, Tokyo, 1950.
Jackson, E. D., and T. L. Wright, Xenoliths in
the Honolulu volcanic series, Hawaii, J. Pe-
trology, 11, 405^30, 1970.
.affey, A. H., K. F. Flynn, L. E. Glendenin,
W. C. Bentley, and A. M. Essling, Precision
measurement of the half-lives and specific ac-
tivies of U235 and U238, Phys. Rev., C, 4, 1889-
1906, 1971.
Jagodzinski, H., and M. Korekawa, Diffuse X-
ray scattering by lunar minerals (abstract), in
Lunar Science IV, J. W. Chamberlain and C.
Watkins, eds., Lunar Science Institute, Hous-
ton, pp. 409-411, 1973.
Jakobsson, S. P., The geology and petrography
of Vestmann Islands; a preliminary report,
Surtsey Res. Prog. Rep., 4, 113-129, 1968.
Jakobsson, S. P., Chemistry and distribution
pattern of Recent basaltic rocks in Iceland,
Lithos, 5, 365-386, 1972.
Jakobsson, S. P., A. K. Pedersen, J. G. Ronsbo,
and L. Melchior Larsen, Petrology of mugea-
rite-hawaiite : early extrusives in the 1973
Heimaey eruption, Iceland, Lithos, 6, 203-
214, 1973.
Jambor, J. L., and C. H. Smith, Accurate de-
termination of olivine composition using
standard small diameter X-ray powder
camera, Amer. Mineral., 47, 194, 1962.
Johannes, W., P. M. Bell, H. K. Mao, A. L.
Boettcher, D. W. Chipman, J. F. Hays, R. C.
Newton, and F. Seifert, An interlaboratory
comparison of piston-cylinder pressure cali-
bration using the albite-breakdown reaction,
Contrib. Mineral. Petrol, 32, 24-38, 1971.
J0rgensen, C. K., Absorption Spectra and
Chemical Bonding in Complexes, Addison-
Wesley Publishing Company, Reading, Mass-
achusetts, 1962a.
J0rgensen, C. K., Orbitals in Atoms and Mole-
cules, Academic Press, New York, 19626.
Kamb, W. B., Ice petrofabric observations from
Blue Glacier, Washington, in relation to
theory and experiment, J. Geophys. Res., 64,
1891-1909, 1959.
Karig, D. E., Ridges and basins of the Tonga-
Kermadec arc system, J. Geophys. Res., 75,
239-254, 1970.
Kawai, N., and S. Mochizuki, Metallic states
in the three 3d transition metal oxides, Fe203,
Cr203, and Ti02 under static high pressures,
Phys. Lett., 36A, 54-55, 1971.
Keesmann, I., S. Matthes, W. Schreyer, and F.
Seifert, Stability of almandine in the system
FeO-(Fe203)-Al203-Si02-(H20) at elevated
pressures, Contrib. Mineral. Petrol., 31, 132—
144, 1971.
Keil, Klaus, Mineralogical and chemical rela-
tionships among enstatite chondrites, J. Geo-
phys. Res., 73, 6945-6976, 1968.
Kelley, K. K., S. S. Todd, R. L. Orr, E. G.
King, and K. R. Bonnickson, Thermodynamic
properties of sodium-aluminum and potas-
sium-aluminum silicates, U.S. Bur. Mines
Rep. Invest., 4955, 21 pp., 1953.
Kelsey, C. H., and D. McKie, The unit-cell of
aenigmatite, Mineral. Mag., 33, 986-1001,
1964.
Kerrick, D. M., Experimental determination of
muscovite + quartz stability with Ph2o <
Ptotai, Amer. J. Sci., 272, 946-958, 1972.
Kirkham, V. R. D., Snake River downwarp, J.
Geol, 39, 456-482, 1931.
Kirsten, T., P. Horn, D. Hegmann, W. Hiibner,
and D. Storzer, Apollo 17 crystalline rocks
and soils: rare gases, ion tracks, and ages
(abstract), Trans. Amer. Geophys. Union, 54,
595-596, 1973.
Kjartansson, G., Nokkar nyjar C14 aldursakv-
ardanir, Natturujraedingurinn, 36, 126—141,
1967.
Knopf, A., The Eagle River region, southeast-
ern Alaska, U.S. Geol. Surv. Bull., 446, 1-58,
1912.
Kobayashi, Y., and H. Maruyama, Electrical
conductivity of olivine single crystals at high
temperature, Earth Planet. Sci. Lett., 11,
415-419, 1971.
Korzhinskii, D. S., Mobility and inertness of
components in metasomatism (in Russian),
Izv. Akad. Nauk SSSR, 1, 36-60, 1936.
Korzhinskii, D. S., The theory of systems with
perfectly mobile components and processes of
mineral formation, Amer. J. Sci., 263, 193-205,
1965.
Korzhinskii, D. S., Theory of Metasomatic Zon-
ing, Clarendon Press, Oxford, 1970.
Koster van Groos, A. F., and P. J. Wyllie,
Liquid immiscibility in the join NaAlSi308-
Na2C03-H20 and its bearing on the genesis of
carbonatites, Amer. J. Sci., 266, 932-967, 1968.
Kracek, F. C, The cristobalite liquidus in the
alkali oxide-silica systems and the heat of
fusion of cristobalite, J. Amer. Chem. Soc, 52,
1436-1442, 1930.
Krogh, T. E., A low-contamination method for
hydrothermal decomposition of zircon and
extraction of U and Pb for isotopic age deter-
minations, Geochim. Cosmochim. Acta, 37,
485-^94, 1973.
Kuno, H., and K. Nagashima, Chemical com-
positions of hypersthene and pigeonite in
equilibrium in magma, Amer. Mineral., 37,
1000-1006, 1952.
GEOPHYSICAL LABORATORY
725
Kurkjian, C. R., and D. N. E. Bushanan, Moss-
bauer absorption of 57Fe in inorganic glasses,
Phys. Chem. Glasses, 5, 63-70, 1964.
Kushiro, I., Si-Al relation in clinopyroxenes
from igneous rocks, Amer. J. Sci., 258, 548-
554, 1960.
Kushiro, I., Compositions of magmas formed
by partial zone melting of the earth's upper
mantle, J. Geophys. Res., 73, 619-634, 1968.
Kushiro, I., The system forsterite-diopside-silica
with and without water at high pressures,
Amer. /. Sci., Schairer Vol., 261 A, 269-294,
1969.
Kushiro, I., Effect of water on the composition
of magmas formed at high pressures, J. Pe-
trology, 13, 311-334, 1972a.
Kushiro, I., Petrology of some Apollo 15 mare
basalts, in The Apollo 15 Lunar Samples,
J. W. Chamberlain and C. Watkins, eds.,
Lunar Science Institute, Houston, pp. 128-
130, 1972b.
Kushiro, I., Determination of liquidus relations
in synthetic silicate systems with electron
probe analysis: the system forsterite-diop-
side-silica at 1 atmosphere, Amer. Mineral.,
57, 1260-1271, 1972c.
Kushiro, I., Partial melting of garnet lherzolites
in kimberlite at high pressures, in Kimberlites
of the Northern Lesotho Province, Lesotho
National Development Corp., in press, 1973.
Kushiro, I., and K. Aoki, Origin of some
eclogite inclusions in kimberlite, Amer. Min-
eral, 53, 1347-1367, 1968.
Kushiro, I., Y. Ikeda, and Y. Nakamura, Pe-
trology of Apollo 14 high-alumina basalt,
Proc. Third Lunar Sci. Conf., Geochim. Cos-
mochim. Acta, Suppl. 3, Vol. 1, The MIT
Press, Cambridge, pp. 115-129, 1972.
Kushiro, I., Y. Nakamura, K. Kitayama, and S.
Akimoto, Petrology of some Apollo 12 crys-
talline rocks, Proc. Second Lunar Sci. Conf.,
Geochim. Cosmochim. Acta, Suppl. 2, Vol. 1,
The MIT Press, Cambridge, pp. 481-495,
1971.
Kushiro, I., N. Shimizu, Y. Nakamura, and S.
Akimoto, Compositions of coexisting liquid
and solid phases formed upon melting of
natural garnet and spinel lherzolites at high
pressures: a preliminary report, Earth Planet.
Sci. Lett., 14, 19-25, 1972.
Kushiro, I., Y. Syono, and S. Akimoto, Effect of
pressure on garnet-pyroxene equilibrium in
the system MgSi03-CaSi03-Al203, Earth
Planet. Sci. Lett., 2, 460-464, 1967.
Kushiro, I., Y. Syono, and S. Akimoto, Melting
of a peridotite nodule at high pressures and
high water pressures, J. Geophys. Res., 73,
6023-6029, 1968.
Kushiro, I., H. S. Yoder, Jr., and M. Nishikawa,
Effect of water on the melting of enstatite,
Geol. Soc. Amer. Bull, 79, 1685-1692, 1968.
Kvenvolden, K. A., E. Peterson, and G. E.
Pollock, Geochemistry of amino acid enantio-
mers: gas chromatography of their diastereo-
meric derivatives, in Advances in Organic
Geochemistry, 1971, H. von Gaertner and H.
Wehner, eds., Pergamon Press, New York,
pp. 387-401, 1972.
Lambert, I. B., and P. J. Wyllie, Stability of
hornblende and a model for the low velocity
zone, Nature, 219, 1240-1241, 1968.
Lanphere, M. A., and G. D. Eberlain, Potas-
sium-argon ages of magnetite-bearing ultra-
mafic complexes in southeastern Alaska (ab-
stract), Geol. Soc. Amer. Spec. Pap., 87, 94,
1966.
Larson, R. L., and W. C. Pitman, World-wide
correlation of Mesozoic magnetic anomalies,
and its implications, Geol. Soc. Amer. Bull,
83, 3645-3661, 1972.
Lathram, E. H., J. S. Pomeroy, H. C. Berg, and
R. A. Loney, Reconnaissance geology of
Admiralty Island, Alaska, UJ3. Geol Surv.
Bull, 1181-R, 1-48, 1965.
Lawson, A. W., On the high temperature heat
conductivity of insulators, J. Phys. Chem.
Solids, 3, 155, 1957.
Lawson, A. W., and J. C. Jamieson, Energy
transfer in the earth's mantle, J. Geol. 66,
540-551, 1958.
Leake, B. E., Compilation of chemical analyses
and physical constants of natural cordierites,
Amer. Mineral, 45, 282-298, 1960.
Leeman, W. P., and W. I. Manton, Strontium
isotopic composition of basaltic lavas from
the Snake River Plain, southern Idaho, Earth
Planet. Sci. Lett., 11, 420-434, 1971.
Lenhert, P. G., A disc oriented diffractometer
control program (abstract), American Crys-
tallographic Association, Albuquerque Meet-
ing, Program and Abstracts, 58, 1972.
Lewis, H. C, Papers and Notes on the Genesis
and Matrix of the Diamond, T. G. Bonney,
ed., Longmans, Green and Company, London,
72 pp., 1897.
Lord, H. S., Ill, Molecular equilibria and con-
densation in a solar nebula and cool stellar
atmospheres, Icarus, 4, 279-288, 1965.
Lorimer, G. W., and P. E. Champness, Com-
bined electron microscopy and analysis of an
orthopyroxene, Amer. Mineral, 58, 243—248,
1973.
Low, W., and M. Weger, Paramagnetic reso-
nance and optical spectra of divalent iron in
cubic field, Part II, Experimental methods,
Phys. Rev., 118, 1130-1136, 1960.
726
CARNEGIE INSTITUTION
Lubimova, E. A., Theory of thermal state of
the earth's mantle, in The Earth's Mantle,
T. F. Gaskell, ed., Academic Press, New
York, 231-323, 1967.
Lumbers, S. B., Geology of the Burwash area
(west half), district of Sudbury and Parry
Sound, Ont. Dept. Mines, Preliminary Map
P680, 1970.
MacGregor, I. D., The system MgO-Si02-Ti02
and its bearing on the distribution of Ti02
in basalts, Amer. J. Sci., Schairer Vol., 267 A,
342-363, 1969.
MacGregor, I. D., The effect of CaO, Cr203,
Fe203, and A1203 on the stability of spinel
and garnet peridotites, Phys. Earth Planet.
Interiors, 3, 372-377, 1970.
MacGregor, I. D., The system MgO-Al203-Si02 :
solubility of A1203 in enstatite for spinel and
garnet peridotite compositions, Amer. Min-
eral., in press, 1973.
Maillard, L. C, Formation de matieres hum-
iques par action de polypeptides sur les sucres,
C. R.Acad. Sci. (Paris), 156, 148-149, 1913.
Malde, H. E., Fault zone along northern bound-
ary of western Snake River Plain, Idaho,
Science, 130, 272, 1959.
Malfait, B. T., and M. G. Dinkelman, Circum-
Caribbean tectonic and igneous activity and
the evolution of the Caribbean plate, Geol.
Soc. Amer. Bull, 83, 251-272, 1972.
Manskaya, S. M., and T. V. Drozdova, Geo-
chemistry of Organic Substances, translated
and edited by L. Shapiro and I. Breger,
Pergamon Press, New York, 1968.
Mao, H. K., The pressure dependence of the
lattice parameters and volume of ferromag-
nesian spinels and its implications to the
earth's mantle, Ph. D. thesis, University of
Rochester, Rochester, New York, 1967.
Mao, H. K., and P. M. Bell, Electrical conduc-
tivity and the red shift of absorption in
olivine and spinel at high pressure, Science,
176, 403-406, 1972.
Mao, H. K., and P. M. Bell, Polarized crystal-
field spectra of micro particles of the moon,
Amer. Soc. Test. Mater. Spec. Tech. Publ.,
539, 100-119, 1973.
Mao, H. K., T. Takahashi, W. A. Bassett, J. S.
Weaver, and S. Akimoto, Effect of pressure
and temperature on the molar volumes of
wiistite and of three (Fe,Mg)2Si04 spinel
solid solutions, J. Geophys. Res., 74, 1061—
1069, 1969.
Margrave, J. L., Vapor pressure, in Physico-
chemical Measurements at High Tempera-
ture, J. O'M. Bockris, J. L. White, and J. D.
Mackenzie, eds., Academic Press, New York,
pp. 225-246, 1959.
Marzolf, J. G., J. T. Dehn, and J. F. Salmon,
Mossbauer studies of tektites, pyroxenes and
olivines, Advan. Chem. Ser., 68, 61-85, 1967.
Matsui, Y., Y. Syono, and Y. Maeda, Moss-
bauer spectra of natural and synthetic Ca-
rich clinopyroxenes, Tech. Rep. Alfl8, Insti-
tute of Solid State Physics, University of
Tokyo, Japan, 1971.
Mattinson, J. M., Age and evolution of the
Tatoosh volcano-plutonic complex (abstract),
Trans. Amer. Geophys. Union, 54, 494, 1973.
McCallum, M. E., and D. H. Eggler, Mineral-
ogy of the Sloan diatreme, a kimberlite pipe
in northern Larimer County, Colorado, Amer.
Mineral, 56, 1735-1749, 1971.
McDonald, K. L., Penetration of the geomag-
netic secular field through a mantle with
variable conductivity, J. Geophys. Res., 62,
117-141, 1957.
McGetchin, T. R., and J. R. Besancon, Carbon-
ate inclusions in mantle-derived pyropes,
Earth Planet. Sci. Lett., 18, 408-410, 1973.
McGetchin, T. R., and L. T. Silver, Composi-
tional relations in minerals from kimberlite
and related rocks in the Moses Rock Dike,
San Juan County, Utah, Amer. Mineral, 55,
1738-1753, 1970.
McKenzie, D. P., The viscosity of the mantle,
Geophys. J. Roy. Astron. Soc, 14, 297-305,
1967.
McKenzie, D. P., Speculations on the con-
sequences and causes of plate motions, Geo-
phys. J. Roy. Astron. Soc, 18, 1-32, 1969.
McQuarrie, M., Thermal conductivity: VII,
Analysis of variation of conductivity with
temperature for A1203, BeO and MgO, J.
Amer. Ceram. Soc, 37, 91-95, 1954.
Meyer, H. O. A., and D. G. Brookins, Eclogite
xenoliths from Stockdale kimberlite, Kansas,
Contrib. Mineral Petrol, 34, 60-72, 1971.
Miller, A., Distribution of cations in spinels,
J. Appl. Phys., 30, Suppl., No. 4, 24S-25S,
1959.
Mitchell, R. H., Magnesian ilmenite and its role
in kimberlite petrogenesis, J. Geol, 81, 301-
311, 1973.
Miyashiro, A., Evolution of metamorphic belts,
J. Petrology, 2, 277-311, 1961.
Mizutani, H., and H. Kanamori, Electrical con-
ductivities of rock-forming minerals at high
temperatures, /. Phys. Earth, 15, 25-31, 1967.
Modreski, P. J., The phase relations of phlogo-
pite in the system K20-MgO-CaO-Al203-Si02-
H20 to 35 kilobars pressure: a model for the
stability of mica in the upper mantle of the
earth, unpublished Ph.D. thesis, The Penn-
sylvania State University, 1972.
GEOPHYSICAL LABORATORY
727
Modreski, P. J., and A. L. Boettcher, The sta-
bility of phlogopite .+ enstatite at high pres-
sures : a model for micas in the interior of the
earth, Amer. J. Sci., 272, 852-869, 1972.
Montigny, R., M. Javoy, and C. J. Allegre,
Sr87Sr86, K/Rb, and 018/016 ratios in the
Pinde ophiolitic complex, Greece (abstract),
Abstr. with Programs (Geol. Soc. Amer.), 7,
627-628, 1970.
Moorbath, S., H. Sigurdsson, and R. Goodwin,
K-Ar ages of the oldest exposed rocks in Ice-
land, Earth Planet. Sci. Lett., 4, 197-205,
1968.
Moorbath, S., and G. P. L. Walker, Strontium
isotope investigation of igneous rocks from
Iceland, Nature, 207, 837-840, 1965.
Moore, R. K., and W. B. White, Electronic
spectra of transition metal ions in silicate
garnets, Can.. Mineral, 11, 791-811, 1972.
Morey, G. W., and M. Fleischer, Equilibrium
between vapor and liquid phases in the sys-
tem C02-H20-K20-Si02, Geol. Soc. Amer.
Bull, 51, 1035-1058, 1940.
Morey, G. W., and J. M. Hesselgesser, The solu-
bility of some minerals in superheated steam
at high pressures, Econ. Geol, 46, 821-835,
1951.
Morgan, W. J., Convection plumes in the lower
mantle, Nature, 230, 42-43, 1971.
Morimoto, N., The existence of monoclinic
pyroxenes with the space group C52fc-P2i/c,
Proc. Jap. Acad., 32, 750-752, 1956.
Morimoto, N., and N. Giiven, Refinement of
the crystal structure of pigeonite, Amer. Min-
eral, 55, 1195-1209, 1970.
Morimoto, N., and K. Koto, The crystal struc-
ture of orthoenstatite, Z. Kristallogr ., , 129,
65-83, 1969.
Morimoto, N., and M. Tokonami, Domain
structure of pigeonite and clinoenstatite,
Amer. Mineral, 54, 725-740, 1969.
Muan, A., and E. F. Osborn, Phase Equilibrium
among Oxides in Steelmaking, Addison-Wes-
ley Publishing Company, Reading, Massa-
chusetts, 1965.
Muehlenbachs, K., A. T. Anderson, and G. E.
Sigvaldason, The origins of O18-poor volcanic
rocks from Iceland (abstract), Trans. Amer.
Geophys. Union, 53, 556, 1972.
Muehlenbachs, K., and R. N. Clayton, Oxygen
isotope studies of fresh and weathered sub-
marine basalts, Can. J. Earth Sci., 9, 172—184,
1972a.
Muehlenbachs, K., and R. N. Clayton, Oxygen
isotope geochemistry of submarine green-
stones, Can. J. Earth Sci., 9, 471-478, 19726.
Mullin, J. B., and J. P. Riley, The colorimetric
determination of silicate with special refer-
ence to sea and natural waters, Anal. Chim.
Acta, 12, 162-176, 1955.
Murray, C. G., Zoned ultramafic complexes of
the Alaskan type: Feeder pipes of andesitic
volcanoes, in "Studies in Earth and Space
Sciences," R. Shagam and others, eds., Geol.
Soc. Amer. Mem., 132, 313-335, 1972.
Mysen, B. O., and A. L. Boettcher, Melting in
a hydrous mantle: phase relationships in
peridotite-H20-C02 systems (abstract), Abstr.
with Programs (Geol. Soc. Amer.), 4, 608,
1972.
Mysen, B. O., and I. Kushiro, Generation of
andesitic magmas in the upper mantle under
hydrous conditions; a discussion of a paper
by I. A. Nicholls and A. E. Ringwood, "Pro-
duction of silica-saturated tholeiitic magmas
in island arcs." Comments to the reply of
I. A. Nicholls and A. E. Ringwood. Earth
Planet. Sci. Lett., in press, 1973.
Nafziger, R. H., G. C. Ulmer, and E. Woer-
mann, Gaseous buffering for the control of
oxygen fugacity at one atmosphere, in Re-
search Techniques for High Pressure and
High Temperature, G. C. Ulmer, ed.,
Springer-Verlag New York Inc., pp. 9-42,
1971.
Nakamura, Y., Equilibrium relations in Mg-
rich part of the pyroxene quadrilateral. Min-
eral J., 6, 264-276, 1971.
Nakamura, Y., and I. Kushiro, Compositional
relations of coexisting orthopyroxene, pigeon-
ite and augite in a tholeiitic andesite from
Harbore volcano, Contrib. Mineral. Petrol,
26, 265-275, 1970a.
Nakamura, Y., and I. Kushiro, Equilibrium
relations of hypersthene, pigeonite and augite
in crystallizing magmas: microprobe study of
a pigeonite andesite from Weiselberg, Ger-
many, Amer. Mineral, 55, 1999-2015, 1970b.
Newton, M. S., and G. C. Kennedy, Jadeite,
analcite, nepheline, and albite at high tem-
peratures and pressures, Amer. J. Sci., 266,
728-735, 1968.
Newton, R. C, An experimental determination
of the high-pressure stability limits of mag-
nesian cordierite under wet and dry condi-
tions, J. Geol, 80, 398-420, 1972.
Newton, R. C, and J. V. Smith, Investigations
concerning the breakdown of albite at depth
in the earth, /. Geol, 75, 26S-286, 1967.
Nicholson, W. J., and G. Burns, Quadrupole
coupling constant, eq Q/h, of Fe3+ in several
rare-earth iron garnets, Phys. Rev., A, 133,
156&-1570, 1964.
Nixon, P. H., O. von Knorring, and J. M.
Rooke, Kimberlites and associated inclusions:
a mineralogical and geochemical study, Amer.
Mineral, 48, 1090^1132, 1963.
728
CARNEGIE INSTITUTION
Noritomi, K., The electrical conductivity of
rocks and the determination of electrical con-
ductivity of the earth's interior, J. Mining
Coll. Akita Univ., Ser. A, 1, 27-59, 1961.
Ogner, G., and M. Schnitzer, Humic sub-
stances: fulvic acid-dialkyl phthalate com-
plexes and their role in pollution, Science,
170, 317-318, 1970.
O'Hara, M. J., Primary magmas and the origin
of basalts, Scot. J. Geol, 1, 19-40, 1965.
O'Hara, M. J., and E. L. P. Mercy, Petrology
and petrogenesis of some garnetiferous peri-
dotites, Trans. Roy. Soc. Edinburgh, 65, 251-
314, 1963.
O'Hara, M. J., and H. S. Yoder, Jr., Formation
and fractionation of basic magmas at high
pressures, Scot. J. Geol, 3, 67-117, 1967.
Ohashi, Y., High temperature structural crys-
tallography of synthetic clinopyroxenes,
(CaFe)Si03, Ph.D. thesis, Harvard Univer-
sity, Cambridge, 1973.
Ohashi, Y., and C. W. Burnham, Electrostatic
and repulsive energies of the Ml and M2
cation sites in pyroxenes, J. Geophys. Res.,
77, 5761-5766, 1972.
Ohashi, Y., and C. W. Burnham, Clinopyroxene
lattice deformations: the roles of chemical
substitution and temperature, Amer. Min-
eral, 58, 843-849, 1973.
Ohashi, Y., and L. W. Finger, A lunar pigeon-
ite: crystal structure of primitive-cell do-
mains (abstract), American Crystallographic
Association, Storrs (Conn.) Meeting, Pro-
gram and Abstracts, Ser. 2, 1, 169, 1973.
Orville, P. M., Unit-cell parameters of the
microcline-low albite and the sanidine-high
albite solid solution series, Amer. Mineral,
52, 55-86, 1967.
Osborn, E. F., The system CaSi03-diopside-
anorthite, Amer. J. Sci., 2Jfi, 751-788, 1942.
Oxburgh, E. F., Penological evidence for the
presence of amphibole in the upper mantle
and its petrogenetic and geophysical implica-
tions, Geol Mag., 101, 1-19, 1964.
Pabst, A., The crystal structure of gillespite,
Amer. Mineral, 28, 372-390, 1937.
Padovani, E. R., and J. L. Carter, Unusual par-
ticles in Apollo 16 soils (abstract), Trans.
Amer. Geonphys. Union, 54, P-70, 1973.
Pakiser, L. C, and I. Zietz, Transcontinental
crustal and upper-mantle structure, Rev.
Geophys., 3, 505-520, 1965.
Palmason, G., Crustal structure of Iceland from
explosion seismology, Soc. Sci. Islandica, 40,
187 pp., 1971.
Papike, J. J., A. E. Bence, G. E. Brown, C. T.
Prewitt, and C. H. Wu, Apollo 12 clinopyrox-
enes: exsolution and epitaxy, Earth Planet.
Sci. Lett., 10, 307-315, 1971.
Peterman, Z. E., R. G. Coleman, and R. A.
Hildreth, Sr87/Sr86 in mafic rocks of the
Troodos massif, Cyprus, U.S. Geol Surv.
Prof. Pap., 750-D, 157-161, 1972.
Philpotts, A. R., Clinobronzite in the diabase
of eastern Connecticut (abstract), Trans.
Amer. Geophys. Union, 54, 492, 1973.
Philpotts, A. R., E. F. Pattison, and J. S. Fox,
Kalsilite, diopside, and melilite in a sedi-
mentary xenolith from Brome Mountain,
Quebec, Nature, 214, 1322-1323, 1967.
Pitman, W. C, and M. Talwani, Sea-floor
spreading in the North Atlantic, Geol Soc.
Amer. Bull, 83, 619-646, 1972.
Podosek, F. A., and R. S. Lewis, 129I and 244Pu
abundances in white inclusions of the Allende
meteorite, Earth Planet. Sci. Lett., 15, 101—
109, 1972.
Pollock, G. E., and V. I. Oyama, Resolution
and separation of racemic amino acids by gas
chromatography and the application to pro-
tein analysis, J. Gas Chromatogr., 4, 126-131,
1966.
Powers, H. A., A distinctive chemical charac-
teristic of Snake River basalts of Idaho, U.S.
Geol. Surv. Prof. Pap., 400-B, B298, 1960a.
Powers, H. A., Alkalic lava flow, with fluidity
of basalt, in the Snake River Plain, Idaho,
U.S. Geol. Surv. Prof. Pap., 400-B, B297,
1960b.
Presnall, D. C, The geometrical analysis of
partial fusion, Amer. J. Sci., 267, 1178-1194,
1969.
Press, F., Earth models obtained by Monte
Carlo inversion, J. Geophys. Res., 73, 5223-
5234, 1968.
Prewitt, C. T., G. E. Brown, and J. J. Papike,
Apollo 12 clinopyroxenes, high temperature
X-ray diffraction studies, Proc. Second Lunar
Sci. Conf., Geochim. Cosmochim. Acta, Suppl.
2, Vol. 1, The MIT Press, Cambridge, pp. 59-
68, 1971.
Prewitt, C. T., J. J. Papike, and M. Ross,
Cummingtonite : a reversible nonquenchable
transition from P2i/m to C2/m symmetry,
Amer. Mineral, 55, 305-306, 1970.
Price, P. J., Ambipolar thermodiffusion of elec-
trons and holes in semiconductors, Phil. Mag.,
46, 1252-1260, 1955.
Prinz, M., E. Dowty, and K. Keil, Mineralogy
and chemistry of Apollo 17 soils (abstract),
Trans. Amer. Geophys. Union, 54, 605, 1973.
Quijano-Rico, M., and H. Wanke, Determina-
tion of boron, lithium, and chlorine in mete-
orites, in Meteoritic Research, P. M. Mill-
man, ed., D. Reidel Co., Dordrecht, Nether-
lands, pp. 132-145, 1969.
GEOPHYSICAL LABORATORY
729
Ramdohr, P., The Opaque Minerals in Stony
Meteorites, American Elsevier Publishing
Company, New York, 1973.
Raymond, M., Electric-field-gradient calcula-
tions in the aluminum silicates (Al2Si05),
Phys. Rev., B, 3, 3692-3702, 1971.
Reid, A. M., R. W. Brown, G. E. Lofgren,
G. N. Heiken, and G. Moreland, Apollo 17
orange glass, Apollo 15 green glass and
Hawaiian lava fountain glass (abstract),
Trans. Amer. Geophys. Union, 54, 606, 1973.
Reiling, G. H., and E. B. Hensley, Fundamental
optical absorption in magnesium oxide, Phys.
Rev., 112, 1106-1111, 1958.
Ribbe, P. H, H. D. Megaw, and W. H. Taylor,
The albite structures, Acta Crystallogr., Sect.
B, 25, 1503-1518, 1969.
Ribbe, P. EL, and J. V. Smith, X-ray emission
microanalysis of rock-forming minerals, IV,
Plagioclase feldspars, J. Geol, 74, 217-233,
1966.
Ricci, J. E., The Phase Rule and Heterogeneous
Equilibrium, D. Van Nostrand Company,
Inc., New York, 1951.
Richardson, S. W., Staurolite stability in a part
of the system Fe-Al-Si-O-H, J. Petrology, 9,
467-488, 1968.
Rikitake, T., Electromagnetism and the Earth's
Interior, Elsevier, Amsterdam, 1966.
Ringwood, A. E., and J. F. Lovering, Signifi-
cance of pyroxene-ilmenite intergrowths
among kimberlite xenoliths, Earth Planet.
Sci. Lett., 7, 371-375, 1970.
Robbins, M., G. K. Wertherm, R. C. Sherwood,
and D. N. E. Buchanan, Magnetic properties
and site distributions in the system FeCr204-
Fe304(Fe2+Cr2-*Fe3+*04), J. Phys. Chem.
Solids, 32,717-729, 1971.
Robertson, E. C, F. Birch, and G. J. F. Mac-
Donald, Experimental determination of jade-
ite stability relations to 25,000 bars, Amer.
J. Sci., 255, 115-137, 1957.
Robertson, J. K., and P. J. Wyllie, Rock-water
systems, with special reference to the water-
deficient region, Amer. J. Sci., 271, 252-277,
1971.
Robie, R. A., P. M. Bethke, and K. M. Beards-
ley, Selected X-ray crystallographic data ;
molar volumes, and densities of minerals and
related substances, U.S. Geol. Surv. Bull.,
1248, 1967.
Robie, R. A., and D. R. Waldbaum, Thermo-
dynamic properties of minerals and related
substances at 298.15°K (25.0° C) and one at-
mosphere (1.013 bars) pressure and at high
temperatures, £/jS. Geol. Surv. Bull., 1259, 256
pp., 1968.
Roedder, E., Liquid C02 inclusions in olivine-
bearing nodules and phenocrysts from basalts,
Amer. Mineral, 50, 1746-1782, 1965.
Roedder, E., Origin of orange glass spherules
in Apollo-17 sample 74220 (abstract), Trans.
Amer. Geophys. Union, 54, 612, 1973.
Roots, E. F., Geology and mineral deposits of
Aiken Lake map-area, British Columbia,
Geol. Surv. Can. Mem., 274, 1-246, 1954.
Rose, H. J., F. Cuttitta, E. J. Dwornik, M. K.
Carron, R. P. Christian, J. R. Lindsay, D. T.
Ligon, and R. R. Larson, Semimicro X-ray
fluorescence analysis of lunar samples, Proc.
Apollo 11 Lunar Sci. Con]., Geochim. Cosmo-
chim. Acta, Suppl. 1, Vol. 2, Pergamon Press,
New York, pp. 1493-1497, 1970.
Ross, M., J. S. Huebner, and E. Dowty, De-
lineation of the one-atmosphere augite-
pigeonite solvus, Amer. Mineral., 58, in press,
1973.
Ross, M., J. S. Huebner, and N. Hickling, De-
lineation of the orthopyroxene-pigeonite
transition and its bearing on pyroxene phase
relations in lunar rocks (abstract), in Lunar
Science IV, J. W. Chamberlain and C. Wat-
kins, eds., Lunar Science Institute, Houston,
p. 637, 1973.
Ruckmick, J. C, and J. A. Noble, Origin of the
ultramafic complex at Union Bay, southeast-
ern Alaska, Geol. Soc. Amer. Bull., 70, 981-
1018, 1959.
Runciman, W. A., D. Sengupta, and J. T. Gour-
ley, The polarized spectra of iron in silicates,
II, Olivine, Amer. Mineral, 58, 451-456, 1973.
Rutherford, M. J., Phase relations in the system
Al203-Si02-Fe-0-H at Pnuld = 2 kbar (ab-
stract), Trans. Amer. Geophys. Union, 51,
437, 1970.
Sato, M., and N. Hickling, Oxygen fugacity
values of some lunar rocks (abstract), in
Lunar Science IV, J. W. Chamberlain and C.
Watkins, eds., Lunar Science Institute, Hous-
ton, pp. 650-652, 1973.
Savin, S. M., and S. Epstein, The oxygen and
hydrogen isotope geochemistry of ocean sedi-
ments and shales, Geochim. Cosmochim.
Acta, 34, 25-42, 1970.
Sawatzky, G. A., and J. Hupkes, Redetermina-
tion of nuclear quadrupole moments from
hyperfine interactions in compounds, Phys.
Rev. Lett., 25, 100-101, 1970.
Sawatzky, G. A., F. Van der Woude, and A. H.
Moorish, Mossbauer study of several ferri-
magnetic spinels, Phys. Rev., 187, 747-757,
1969.
Scarf e, C. M., W. C. Luth, and O. F. Tuttle, An
experimental study bearing on the absence of
leucite in plutonic rocks, Amer. Mineral, 51,
726-735, 1966.
730
CARNEGIE INSTITUTION
Schairer, J. F., The system K20-MgO-Al203-
Si02: I, Results of quenching experiments on
four joins in the tetrahedron cordierite-
forsterite-leucite-silica and on the join cor-
dierite-mullite-potash feldspar, J. Amer.
Ceram. Soc, 37, 501-533, 1954.
Schairer, J. F., and N. L. Bowen, The binary
system CaSi03-diopside and the relations be-
tween CaSiOs and akermanite, Amer. J. Sci.,
240, 725-742, 1942.
Schairer, J. F., and H. S. Yoder, Jr., The nature
of residual liquids from crystallization, with
data on the system nepheline-diopside-silica,
Amer. J. Sci., Bradley Vol., 258A, 273-283,
1960.
Schatz, J. F., and G. Simmons, Thermal conduc-
tivity of earth material at high temperatures,
J. Geophys. Res., 77, 6966-6983, 1972.
Schieber, M. M., Selected Topics in Solid State
Physics; Vol. 8, Experimental Magnetochem-
istry: Nonmetallic Materials, E. P. Wohl-
farth, ed., John Wiley and Sons, Inc., New
York, 1967.
Schmidt, D. L., Quaternary geology of the
Bellevue area, Blaine and Camas counties,
Idaho, Ph.D. thesis, University of Washing-
ton, 1961.
Schnitzer, M., and S. U. Khan, Humic Sub-
stances in the Environment, Marcel Dekker,
Inc., New York, 1972.
Schreyer, W., and H. S. Yoder, Jr., The system
Mg-cordierite-H20 and related rocks, Neues
Jahrb. Mineral, Abh., 101, 271-342, 1964.
Schiirmann, K., and S. S. Hafner, Distinct sub-
solidus cooling histories of Apollo 14 basalts,
Proc. Third Lunar Sci., Con]., Geochim. Cos-
mochim. Acta, Suppl. 3, Vol. 1, The MIT
Press, Cambridge, pp. 493-506, 1972.
Seitz, M. G., and S. R. Hart, Uranium and
boron distributions in some oceanic ultra-
mafic rocks, Earth Planet. Sci. Lett., in press,
1973.
Seitz, M. G., R. M. Walker, and B. S. Car-
penter, Improved methods for measurement
of thermal neutron dose by the fission track
technique, /. Appl. Phys., 44, 510-512, 1973.
Shand, S. J., The heavy minerals of kimberlite,
Trans. Geol. Soc. S. Africa, 37, 57-68, 1934.
Shankland, T. J., Transport properties of
olivine, in The Application of Modern Phys-
ics to the Earth and Planetary Interior, S. K.
Runcorn, ed., Interscience, New York, pp.
175-190, 1969.
Sharma, R. R., Nuclear quadrupole moment of
Al27 in A1203, Phys. Rev. Lett., 25, 1622-1623,
1970.
Sharma, R. R., Nuclear quadrupole interactions
in several rare-earth iron garnets, Phys. Rev.,
B., 6, 4310-4323, 1972.
Sharma, R. R., and B. N. Teng, Quadrupole
coupling constants of Fe3+ in yttrium iron
garnet, Phys. Rev. Lett., 27, 679-681, 1971.
Shaw, H. R., Theoretical solubility of H20 in
silicate melts: quasi-crystalline models, J.
Geol, 72, 601-617, 1964.
Sheppard, S. M. F., and S. Epstein, D/H and
018/016 ratios of minerals of possible mantle
or lower crustal origin, Earth Planet. Sci.
Lett., 9, 232-239, 1970.
Shiraki, K., Metamorphic basement rocks of
the Yap Islands, western Pacific: possible
oceanic crust beneath an island arc, Earth
Planet. Sci. Lett., 13, 167-174, 1971.
Shmulovich, K. I., Stability of merwinite in the
system CaO-MgO-Si02-C02, Dokl Akad.
Nauk SSSR, 184, 1177-1179, 1969.
Shoemaker, D. P., Optimization of counting
times in computer-controlled X-ray and neu-
tron single-crystal diffractometry, Acta Crys-
tallogr., Sect. A, 24, 136-142, 1968.
Sigurdsson, H., The petrology and chemistry
of the Setberg volcanic region and the inter-
mediate and acid rocks of Iceland, Ph.D.
thesis, Durham University, 1970.
Silver, L. T., The relation between radioactivity
and discordance in zircons, in "Nuclear Geo-
physics," Nat. Acad. Sci.-Nat. Res. Council
PubL, 1075, 34-42, 1963.
Skinner, B. J., Physical properties of end mem-
bers of the garnet group, Amer. Mineral, 41,
428-436, 1956.
Smith, A. G., and A. Hallam, The fit of the
southern continents, Nature, 225, 139-144,
1970.
Smith, D., Stability of iron-rich pyroxene in
the system CaSi03-FeSi03-MgSi03, Amer.
Mineral, 57, 1413-1428, 1972.
Smith, H. L., and A. J. Cohen, Absorption
spectra of cations in alkali-silicate glasses of
high ultraviolet transmission, Phys. Chem.
Glasses, 4, 173-187, 1963.
Smith, J. V., X-ray-emission microanalysis of
rock-forming minerals, VI, Clinopyroxenes
near the diopside-hedenbergite join, J. Geol,
74, 463-477, 1966.
Smyth, J. R., High-temperature single-crystal
X-ray studies of natural orthopyroxenes,
Amer. Mineral, 55, 312, 1970.
Smyth, J. R., Protoenstatite : a crystal-structure
refinement at 1100°C, Z. Kristallogr., 134,
262-274, 1971.
Smyth, J. R., and C. W. Burnham, The crystal
structures of high and low clinohypersthene,
Earth Planet. Sci. Lett., 14, 183-189, 1972.
Smyth, J. R., and R. M. Hazen, The crystal
structures of forsterite and hortonolite at
several temperatures up to 900° C, Amer.
Mineral, 58, 58S-593, 1973.
GEOPHYSICAL LABORATORY
731
Spackman, D. H., W. Stein, and S. Moore,
Automatic recording apparatus for use in the
chromatography of amino acids, Anal. Chem.,
30, 119O-1206, 1958.
Speidel, D. H., Phase equilibria in the system
MgO-FeO-Fe203: the 1300° C isothermal sec-
tion and extrapolations to other tempera-
tures, J. Amer. Ceram. Soc, 50, 243-248, 1967.
Stearns, H. T., Volcanism in the Mud Lake
area, Idaho, Amer. J. Sci., 11, 353-363, 1926,
Stearns, H. T., L. Crandall, and W. G. Steward,
Geology and groundwater resources of the
Snake River Plain in southeastern Idaho, U.S.
Geol. Surv. Water Supply Pap., 774, 268 pp.,
1938.
Sternheimer, R. M., Quadrupole antishielding
factors of ions, Phys. Rev., 130, 1423-1425,
1963.
Stevenson, F. J., and J. H. A. Butler, Chemistry
of humic acids and related pigments, in
Organic Geochemistry, G. Eglinton and
M. T. J. Murphy, eds., Springer- Verlag New
York, Inc., 1969.
Stewart, D. B., and D. von Limbach, Thermal
expansion of low and high albite, Amer. Min-
eral, 52, 389-413, 1967.
Stieff, L. R., T. W. Stern, S. Oshiro, and F. E.
Senftle, Tables for the calculation of lead
isotope ages, U.S. Geol. Surv. Prof. Pap.,
334-A, 40 pp., 1959.
Stone, G. T„ Petrology of Upper Cenozoic ba-
salts of the western Snake River Plain, Idaho
(abstract), Geol. Soc. Amer. Spec. Pap., 87,
168, 1966.
Stone, G. T., Petrology of Upper Cenozoic ba-
salts of the western Snake River Plain, Idaho,
Ph.D. thesis, University of Colorado, 1967.
Stone, G. T., Structural implications of Qua-
ternary lava-dome distribution in the Snake
River Plain, Idaho (abstract), Abstr. with
Programs (Geol. Soc. Amer.), 1, pt. 7, 217-
218, 1969.
Stone, G. T., Highly evolved basaltic lavas in
the western Snake River Plain, Idaho (ab-
stract), Abstr. with Programs (Geol. Soc.
Amer.), 2, 695-696, 1970.
Storey, W. C, Volatilization studies on a
terrestrial basalt and their applicability to
volatilization from the lunar surface, Nature,
241, 154-157, 1973.
Strens, R. G. J., Pressure-induced spin-pairing
in gillespite BaFe(II)Si4Oio, Chem. Commun.,
no. 21, 777-778, 1966.
Strickland, J. D. H., The preparation and
properties of silicomolybdate acid, J. Amer.
Chem. Soc, 74, 862-876, 1952.
Subbarao, K. V., and C. E. Hedge, K, Rb, Sr,
and Sr87Sr86 in rocks from the mid-Indian
Ocean Ridge, Earth Planet. Sci. Lett., 18,
223-228, 1973.
Sueno, S., J. J. Papike, C. T. Prewitt, and
G. E. Brown, Crystal structure of high cum-
mingtonite, J. Geophys. Res., 77, 5767-5777,
1972.
Suzoki, T., and S. Epstein, Hydrogen isotope
fractionation factors (a's) between muscovite,
biotite, hornblende and water (abstract),
Trans. Amer. Geophys. Union, 51, 451, 1970.
Swanson, V. E., and J. G. Palacas, Humate in
coastal sands of northwest Florida, U.S. Geol.
Surv. Bull., 1214-B, 1965.
Takeda, H., Crystallographic studies of co-
existing aluminan orthopyroxene and augite
of high-pressure origin, J. Geophys. Res., 77,
5798-5811, 1972a.
Takeda, H., Structural studies of rim augite and
core pigeonite from lunar rock 12052, Earth
Planet. Sci. Lett., 15, 65-71, 1972b.
Takeda, H., and R. J. Williams, Pigeonites and
augites precipitated from synthetic lunar
melts (abstract), Trans. Amer. Geophys.
Union, 53, 551, 1972.
Taljaard, M. S., South African melilite basalts
and their relations, Trans. Geol. Soc. S.
Africa, 39, 281-316, 1936.
Tannhauser, D. S., Conducting in iron oxides,
J. Phys. Chem. Solids, 23, 25-34, 1962.
Tatsumoto, M., Genetic relations of oceanic
basalts as indicated by lead isotopes, Science,
153, 1094-1101, 1966.
Taylor, D. R., Overlap contribution to the elec-
tric field gradient in an ionic complex, J.
Chem. Phys., 48, 536-537, 1968.
Taylor, H. P., Jr., The zoned ultramafic com-
plexes of southeastern Alaska, in Ultramafic
and Related Rocks, P. J. Wyllie, ed., John
Wiley and Sons, Inc., New York, pp. 96-118,
1967.'
Taylor, H. P., Jr., The oxygen isotope geo-
chemistry of igneous rocks, Contrib. Mineral.
Petrol., 19, 1-71, 1968.
Taylor, H. P., Jr., and R. W. Forester, Low-O18
igneous rocks from the intrusive complexes of
Skye, Mull and Ardnamurchan, western Scot-
land, J. Petrology, 12, 465-497, 1971.
Taylor, H. P., Jr., and J. A. Noble, Origin of
the ultramafic complexes on southeastern
Alaska, Int. Geol. Congr., 21st, Norden, Pt.
13, 175-187, 1960.
Taylor, H. P., Jr., and J. A. Noble, Origin of
magnetite in the zoned ultramafic complexes
of southeastern Alaska, in "Magmatic Ore
Deposits," H. D. B. Wilson, ed, Econ. Geol.
Monogr., 4, pp. 209-230, 1969.
Thompson, J. B., The thermodynamic basis for
the mineral facies concept, Amer. J. Sci., 253,
65-101, 1955.
732
CARNEGIE INSTITUTION
Thompson, J. B., Geochemical reaction and
open systems, Geochim. Cosmochim. Acta, 34,
529^551, 1970.
Thorarinsson, S., Hekla and Katla, the share
of acid and intermediate lava and tephra in
the volcanic products through the geological
history of Iceland, in "Iceland and Mid-
Ocean Ridges," S. Bjornsson, ed., Soc. Sci.
Islandica Rit, 38, pp. 190-199, 1967.
Thorarinsson, S., and G. E. Sigvaldason, The
Hekla eruption of 1970, Bull. VolcanoL, 36,
269-288, 1972.
Thorarinsson, S., S. Steinthorsson, Th. Einars-
son, H. Kristmannsdottir, and N. Oskarsson,
The eruption of Heimaey, Iceland, Nature,
251, 372-373, 1973.
Tilley, C. E., and I. D. Muir, The Hebridean
Plateau magma type, Trans. Edinburgh Geol.
Soc, 19, 208-215, 1962.
Tilton, G. R., Volume diffusion as a mechanism
for discordant lead ages, J. Geophys. Res., 65,
2933-2945, 1960.
Tilton, G. R., C. Patterson, H. Brown, M.
Inghram, R. Hayden, D. Hess, and E. Larsen,
Jr., Isotopic composition and distribution of
lead, uranium, and thorium in a Precam-
brian granite, Geol. Soc. Amer. Bull., 66,
1131-1148, 1955.
Tolland, H. G., and R. G. J. Strens, Electrical
conduction in physical and chemical mix-
tures; application to planetary mantles, Phys.
Earth Planet. Interiors, 5, 380-386, 1972.
Tozer, D. C., The electrical properties of the
earth's interior, Phys. Chem. Earth, 3, 414-
436, 1959.
Turnock, A. C., D. H. Lindsley, and J. E.
Grover, Synthesis and unit cell parameters of
Ca-Mg-Fe pyroxenes, Amer. Mineral., 58,
50-59, 1973.
Udenfriend, S., S. Stein, P. Bohlen, W. Dair-
man, W. Leimgruber, and M. Weigele, Ap-
plications of fluorescamine, a new reagent for
assay of amino acids, peptides, proteins and
other primary amines in the picomole range,
Science, 178, 871-872, 1972.
Ukhanov, A. V., Olivine melilitite from the
diamond-bearing diatremes on Anabar, Dokl.
Akad. Nauk SSSR, 153, 923-925, 1963.
Ulmer, G. C., and W. J. Smothers, The system
MgO-Cr203-Fe203 at 1300° C in air, J. Amer.
Ceram. Soc, 51, 315-319, 1968.
Upton, B. G. J., and W. J. Wadsworth, Basalts
of Reunion Island, Indian Ocean, Bull. Vol-
canoL, 24, 7-24, 1966.
Verhoogen, J., Distribution of titanium between
silicates and oxides in igneous rocks, Amer.
J. Sci., 260, 211-220, 1962.
Vinogradov, A. P., Preliminary data on lunar
soil collected by the Luna 20 unmanned
spacecraft, Geochim. Cosmochim. Acta, 37,
721-729, 1973.
Virgo, D., and S. S. Hafner, Fe2+,Mg order-
disorder in heated orthopyroxenes, Mineral.
Soc. Amer. Spec Pap., 2, 67-81, 1969.
Virgo, D., and S. S. Hafner, Temperature-
dependent Mg,Fe distribution in a lunar
olivine, Earth Planet. Sci. Lett., 14, 305-312,
1972.
Vogt, P. R., C. N. Anderson, and D. R. Bracey,
Mesozoic magnetic anomalies, sea-floor
spreading, and geomagnetic reversals in the
southwestern North Atlantic, J. Geophys.
Res., 76, 4796-4823, 1971.
Wager, L. R., and G. M. Brown, Layered
Igneous Rocks, Oliver and Boyd, Edinburgh
and London, 1968.
Wagner, P. A., The Diamond Fields of South-
ern Africa, The Transvaal Leader, Johannes-
burg, 1914 (2nd impression, C. Struick (PTY)
Ltd., Cape Town, 1971).
Waksman, S. A., Humus: Origin, Chemical
Composition and Importance in Nature, Wil-
liams and Wilkins Company, Baltimore, 1938.
Waldbaum, D. R., High-temperature thermo-
dynamic properties of alkali feldspars, Con-
trib. Mineral. Petrol, 17, 71-77, 1968.
Walker, G. P. L., The Breiddalur central vol-
cano, eastern Iceland, Quart. J. Geol. Soc.
London, 119, 29-63, 1963.
Walker, G. P. L., Acid volcanic rocks in Iceland,
Bull. VolcanoL, 29, 375-406, 1966.
Walter, L. S., Experimental studies on Bowen's
decarbonation series: I, P-T univariant equi-
libria of the "monticellite" and "akermanite"
reactions, Amer. J. Sci., 261, 488-500, 1963.
Watson, R. E., Analytic Hartree-Fock solutions
for 0=, Phys. Rev., Ill, 1108-1110, 1958.
Weertman, J., Coalescence of magma pockets
into large pools in the upper mantle, Geol.
Soc. Amer. Bull., 83, 3531-3532, 1972.
Weill, D. F., and W. S. Fyfe, The solubility of
quartz in H20 in the range 1000-4000 bars and
400°-550°C, Geochim. Cosmochim. Acta, 28,
1243-1256, 1964.
Weisbrod, A., Les conditions du metamor-
phisme dans le Cevennes Medianes, C. R.
Acad. Sci. (Paris), 266, 755-757, 1968.
Weisbrod, A., Petrologie du socle metamor-
phique des Cevennes Medianes, unpublished
doctoral thesis, 1970.
Weller, W. W., and K. K. Kelley, Low-tempera-
ture heat capacities and entropies at 298.15 °K
of akermanite, cordierite, gehlenite and mer-
winite, U.S. Bur. Mines, Rep. Inv., 6343, 1963.
GEOPHYSICAL LABORATORY
733
White, R. W., Ultramafic inclusions in basaltic
rocks from Hawaii, Contrib. Mineral. Petrol.,
12, 245-314, 1966.
White, W. B., and K. L. Keester, Optical ab-
sorption spectra of iron in the rock-forming
silicates, Amer. Mineral., 51, 774-791, 1966.
Willemse, J., and J. J. Bensch, Inclusions of
original carbonate rocks in gabbro and norite
of the eastern part of the Bushveld complex,
Trans. Geol. Soc. S. Africa, 67, 1-87, 1964.
Williams, P. G. L., G. M. Bancroft, M. G.
Bown, and A. C. Turnock, Anomalous Moss-
bauer spectra of C2/c pyroxenes, Nature
(London), Phys. Sci., 230, 149-151, 1971.
Williams, R. J., and E. K. Gibson, The origin
and stability of lunar goethite, hematite and
magnetite, Earth Planet. Sci. Lett., 17, 84—88,
1972.
Winkler, H. G. F., Pedogenesis of Metamor-
phic Rocks, 2nd ed., Springer-Verlag, Berlin,
1967.
Wood, B. J., and S. Banno, Garnet-orthopyrox-
ene and orthopyroxene-clinopyroxene rela-
tionships in simple and complex systems,
Contrib. Mineral. Petrol., in press, 1973.
Wright, F. E., Obsidian from Hrafntinnuhryg-
gur, Iceland: its lithophysae and surface
markings, Geol. Soc. Amer. Bull., 26, 255-286,
1915.
Wright, T. L., M. J. Grolier, and D. A. Swan-
son, Chemical variation related to the stratig-
raphy of the Columbia River basalt, Geol.
Soc. Amer. Bull, 84, 371-386, 1973.
Wyllie, P. J., Melting relationships in the sys-
tem CaO-MgO-C02-H20 with petrological
applications, J. Petrology, 6, 101-123, 1965.
Wyllie, P. J., Experimental data bearing on the
petrogenetic links between kimberlites and
carbonatites, Int. Mineral. Ass. Pap. Proc.
Gen. Meet., 4th, New Delhi, 67-82, 1966.
Wyllie, P. J., Experimental limits for melting
in the earth's crust and the upper mantle,
Amer. Geophys. Union, Geophys. Monogr.
Ser., 14, 279-301, 1971.
Wyllie, P. J., and J. L. Haas, The system CaO-
Si02-C02-H20 : I, Melting relationships with
excess vapor at 1 kilobar pressure, Geochim.
Cosmochim. Acta, 29, 871-892, 1965.
Wyllie, P. J., and O. F. Tuttle, Effect of carbon
dioxide on the melting of granite and feld-
spars, Amer. J. Sci., 257, 648-655, 1959.
Wyllie, P. J., and 0. F. Tuttle, The system
CaO-C02-H20 and the origin of carbonatites,
/. Petrology, 1, 1-46, 1960.
Yoder, H. S., Jr., Experimental studies bearing
on the origin of anorthosite, N . Y . State Mus.
Sci. Serv. Mem., 18, 13-22, 1968.
Yoder, H. S., Jr., Calcalkaline andesites: ex-
perimental data bearing on the origin of their
assumed characteristics, Oreg. Dep. Geol.
Miner. Ind. Bull, 65, 77-89, 1969.
Yoder, H. S., Jr., Contemporaneous basaltic
and rhyolitic magmas, Amer. Mineral, 58,
153-171, 1973.
Yoder, H. S., Jr., and I. Kushiro, Melting of a
hydrous phase: phlogopite, Amer. J. Sci.,
Schairer Vol., 267 A, 558-582, 1969.
Yoder, H. S., Jr., and C. E. Tilley, Origin of
basalt magmas : an experimental study of na-
tural and synthetic rock systems, J. Petrol-
ogy, 8, 342-532, 1962.
Zen, E-an, Thermochemical parameters of min-
erals from oxygen-buffered hydrothermal
equilibrium data, Contrib. Mineral Petrol,
39, 65-80, 1973.
Zies, E. G., A new analysis of cossyrite from the
island of Pantelleria, Amer. Mineral, 51, 200-
205, 1966.
PERSONNEL
Scientific Staff
Director: H. S. Yoder, Jr., Petrologist
Emeritus Research Associate: E. G. Zies,
Chemist
Systematic Petrologist : F. Chayes
Petrologists: F. R. Boyd, Jr., D. H. Eggler,1
T. N. Irvine,2 I. Kushiro, R. N. Thomp-
son,3 A. Weisbrod 1
Physical Chemist: T. C. Hoering
Geophysicists : P. M. Bell, H. K. Mao *
Organic Geochemist : P. E. Hare
Crystallographer : L. W. Finger
Isotope Geochemists: G. L. Davis, T. E.
Krogh
Solid State Geochemist : D. Virgo
Postdoctoral Fellows: J. Akella, University
of California at Los Angeles; D. M. Burt,
1 Temporary appointment from September 1,
1972.
2 Temporary appointment from August 16,
1972.
3 Appointment terminated September 30.
1972, to accept position as Lecturer in Petrol-
ogy at the Imperial College, London, England.
734
CARNEGIE INSTITUTION
Harvard University; 4 J. S. Dickey, Jr.,
Astrophysical Observatory, Smithsonian
Institution; 5 J. D. Frantz, Johns Hopkins
University; 6 B. J. Hensen, Australian Na-
tional University; F. N. Hodges, Univer-
sity of Texas at Austin; 6 K. King, Jr.,
Lamont-Doherty Geological Observatory ; 7
H. K. Mao, University of Rochester; 8
J. M. Mattinson, University of California
at Santa Barbara; 9 K. Muehlenbachs, Uni-
versity of Chicago; Y. Ohashi, Harvard
University; M. G. Raymond, University
of Chicago; M. G. Seitz, Washington Uni-
versity
Predoctoral Fellows: J. I. Hedges, Marine
Science Institute, University of Texas, Port
Aransas; 10 D. J. Misener, University of
British Columbia, Canada; n B. 0. Mysen,
Pennsylvania State University 6
Guest Investigators: R. M. Abu-Eid, Massa-
chusetts Institute of Technology; J. B.
Adams, West Indies Laboratory, Fairleigh
4 Appointment terminated August 31, 1972, to
accept position as Visiting Professor of Geo-
chemistry at the Vening Meinesz Institute, Uni-
versity of Utrecht, Netherlands.
5 Appointment terminated August 31, 1972, to
accept position as Assistant Professor, Depart-
ment of Earth and Planetary Sciences, Massa-
chusetts Institute of Technology.
6 Appointment from September 1, 1972.
7 Appointment from July 1, 1972.
8 Appointment terminated August 31, 1972, to
accept position as Temporary Staff Member
(Geophysicist) from September 1, 1972.
9 Appointment terminated June 30, 1973, to
accept position in the Department of Geological
Sciences, University of California at Santa
Barbara.
10 Appointment from March 1, 1973.
11 Appointment terminated July 31, 1972.
Dickinson University; I. Ermanovics, Geo-
logical Survey of Canada; S. E. Haggerty,
University of Massachusetts; J. F. Hays,
Harvard University; E. Makovicky, Yale
University; R. McNutt, McMasters Uni-
versity, Canada; V. R. Meenakshi, Univer-
sity of South Carolina; A. Muan, Pennsyl-
vania State University; G. T. Stone,
University of Oklahoma; G. C. Ulmer,
Temple University.
Operating and Maintenance Staff
Executive Officer: A. D. Singer
Accountant: C. B. Petry
Editor and Librarian: Dolores M. Thomas
Stenographers : Marjorie E. Imlay, Mabel B.
Mattingly
Clerk and Technician: H. J. Lutz
Electronic Technician: C. G. Hadidiacos
Research Assistant: Helen J. Wolanin 12
Instrument Maker and Technician: G. E.
Speicher
Instrument Makers: C. A. Batten, L. C.
Garver, W. H. Lyons
Carpenter and Thin Section Technician: E. J.
Shipley 13
Machinists: H. P. Compton,14 W. R. Reed
Apprentice Machinist: P. M. Vacchio 15
Building Engineer: H. L. Moore
Custodian and Painter: M. Ferguson
Custodians and Mechanics: L. B. Patrick, D.
Ratliff, Jr.
12 Resigned October 15, 1972.
13 Retired June 30, 1973.
14 Appointment from November 20, 1972.
15 Appointment from November 1, 1972.
Bibliography
July 1, 1972- June 30, 1973
PUBLICATIONS OF THE INSTITUTION
Carnegie Institution of Washington. Year Book
71. Octavo, viii + 764 pages, 248 figures.
December 1972.
Atlas of Galactic Neutral Hydrogen for the
Region 270° < I < 310° ; —7° < b < 2° ,
by Silvia Garzoli. Publication 629. Octavo,
123 pages, 103 figures. November 1972.
Carnegie Institution of Washington Catalogue
1973-1974. Octavo, 77 pages, 26 figures. May
1973.
Carnegie Institution Newsletter. Series 2, Num-
bers 1-4. Issued in October, February, May,
and July.
The Hubble Atlas of Galaxies, by Allan Sand-
age. Publication 618. Quarto, 141 pages, 55
figures. Reprint. October 1972.
Genetic Variations of Drosophila melanog aster,
by Dan L. Lindsley and E. H. Grell. Publi-
cation 627. Octavo, 472 pages, 169 figures.
October 1972.
PUBLICATIONS BY THE PRESIDENT
OF THE INSTITUTION
Philip H. Abelson
Editorials in Science
Volume 177: p. 121, July 14, 1972, The
fourth revolution;
p. 479, August 11, 1972, The
new physics report;
p. 655, August 25, 1972, En-
vironmental quality;
p. 947, September 15, 1972,
Earth sciences and the qual-
ity of life;
Volume 178: p. 12, October 6, 1972, Latin
American aspirations;
p. 355, October 27, 1972, En-
ergy conservation;
p. 701, November 17, 1972, Sci-
ence, technology, and di-
plomacy ;
p. 937, December 1, 1972, Re-
liability of consumer goods;
Volume 179: p. 17, January 5, 1973, Con-
sumer product safety;
p. 233, January 19, 1973, De-
parture of the President's
science adviser;
p. 333, January 26, 1973, Tech-
nological initiatives and po-
litical realities;
p. 431, February 2, 1973, Dis-
covery and evaluation of
resources ;
p. 641, February 16, 1973, Cas-
ualties of governmental re-
organization ;
p. 857, March 2, 1973, Energy
and national security [re-
printed under title "Dealing
now with the energy crisis"
in Current, April 1973, No.
150, pp. 56-57] ;
Volume 180: p. 13, April 6, 1973, Blow-hot,
blow-cold educational pol-
ices;
p. 259, April 20, 1973, Addi-
tional sources of financial
and political support for sci-
ence;
p. 449, May 4, 1973, Spain—
another Japan?;
p. 819, May 25, 1973, Observ-
ing and predicting earth-
quakes ;
735
736
CARNEGIE INSTITUTION
p. 1127, June 15, 1973, Impor-
tation of petroleum.
The President's Page in E($S, Transactions,
American Geophysical Union
Volume 53: p. 683, July 1972, State of the
Union ;
p. 811, September 1972, Stock-
holm and sequelae;
Volume 54: p. 75, February 1973, An addi-
tional mode of communica-
tion for AGU;
p. 123, March 1973, The gov-
ernment and geophysics;
p. 179, April 1973, Need for
better defense of research ;
p. 531, May 1973, Scientific
communication.
Report of the President. Reprinted from
Carnegie Institution of Washington Year
Book 71, 39 pages, December 1972. ,
Environment : a delicate balance oj costs and
benefits. Excerpts from Report of the
President, Carnegie Institution of Washing-
ton Year Book 71, reprinted by American
Electric Power Co., Inc., New York, 6
pages.
Science in the Seventies (see Year Book 71,
p. 718). Excerpts reprinted in AGB Re-
ports, Vol. 15, No. 5, February 1973, pp.
24-29.
PUBLICATIONS BY THE
EXECUTIVE OFFICER
Edward A. Ackerman
American Resources — Their Management and
Conservation. With J. Russell Whitaker.
(Harcourt, Brace and Company, New York,
1951, x + 497 pp.) Reprint edition, Arno
Press, Inc., New York, 1972.
Where is a research frontier? (see Year Book
63, p. 615; Year Book 71, p. 718). Reprint
136, Warner Modular Publications, Inc.,
Andover, Massachusetts, 1973, pp. 1-12.
Land use institutions in the Washington-
Baltimore region — a mirror for metropoli-
tan America. With Robert G. Dyck and
Atlee E. Shidler. Chapter 4 in Environ-
ment : a New Focus for Land-use Planning,
Donald McAllister, editor. National Sci-
ence Foundation, Washington, D.C. In
press.
Toward a national land use information sys-
tem. With Robert H. Alexander. In press.
Administrative Reports
Report of the Executive Committee
To the Trustees of the Carnegie Institution of Washington
Gentlemen :
In accordance with the provisions of the By-Laws, the Executive Committee submits
this report ot the Annual Meeting of the Board of Trustees.
During the fiscal year ending June 30, 1973, the Executive Committee held four meetings.
Printed accounts of these meetings have been or will be mailed to each Trustee.
The estimate of expenditures of the fiscal year beginning July 1, 1973, has been reviewed
by the Executive Committee.
The terms of office of the Vice-Chairman and the Secretary of the Board of Trustees
expire on May 4, 1973. The terms of all Committee Chairmen and the following members
of Committees also expire on May 4, 1973 :
Executive Committee
Henry S. Morgan
Finance Committee
William T. Golden
Keith S. McHugh
Henry S. Morgan
Nominating Committee
Charles H. Townes
Auditing Committee
Keith S. McHugh
Juan T. Trippe
Retirement Committee
Caryl P. Haskins
(appointed to serve until next
meeting of the Board of Trustees
vice Amory H. Bradford — 1974)
William McChesney Martin, Jr., Chairman
May 4, 1978
739
Abstract of Minutes
of the Seventy-Fourth Meeting of the Board of Trustees
The annual meeting of the Board of Trustees was held in the Board Room of the
Administration Building on Friday, May 4, 1973. The meeting was called to order by
Chairman Garrison Norton.
The following Trustees were present: Michael Ference, Jr., Carl J. Gilbert, William T.
Golden, Caryl P. Haskins, William R. Hewlett, Keith S. McHugh, William McChesney
Martin, Jr., Henry S. Morgan, William I. Myers, Walter H. Page, Richard S. Perkins,
William W. Rubey, Frank Stanton, and Juan T. Trippe. The President, Philip H. Abelson,
was also in attendance.
The minutes of the Seventy-Fourth Meeting were approved.
The resignation of Alfred L. Loomis as Trustee was accepted with regret. Dr. Loomis
was designated Trustee Emeritus.
Lewis M. Branscomb was elected a member of the Board of Trustees.
William McChesney Martin, Jr., was elected Vice Chairman of the Board, and William
T. Golden was elected Secretary of the Board, both for terms ending in 1976.
On the recommendation of the Nominating Committee, the following were elected for
one-year terms: William McChesney Martin, Jr., as Chairman of the Executive Committee;
Richard S. Perkins, as Chairman of the Finance Committee; William M. Roth, as Chair-
man of the Nominating Committee; Keith S. McHugh, as Chairman of the Auditing
Committee; and Frank Stanton, as Chairman of the Retirement Committee.
Vacancies in Standing Committees, with terms as indicated in parentheses, were filled as
follows: Henry S. Morgan (1976) was elected a member of the Executive Committee;
William T. Golden (1976), Keith S. McHugh (1976), and Henry S. Morgan (1976) were
elected members of the Finance Committee; Keith S. McHugh (1976) and Juan T. Trippe
(1976) were elected members of the Auditing Committee; William R. Hewlett (1976) was
elected a member of the Nominating Committee; and Caryl P. Haskins (1974) was elected
a member of the Retirement Committee.
The reports of the Executive Committee, the Finance Committee, the Retirement Com-
mittee, and the Auditing Committee were accepted. On the recommendation of the latter,
it was resolved that Arthur Andersen & Co. be reappointed as public accountants for the
fiscal year beginning July 1, 1973.
The annual report of the President was accepted.
741
742 CARNEGIE INSTITUTION
To provide for the operation of the Institution for the fiscal year beginning July 1, 1973,
and upon recommendation of the Executive Committee, the sum of $6,301,945 was appro-
priated. To provide for construction expenses of the Irenee du Pont Telescope project at
the Las Campanas Observatory for the fiscal year beginning July 1, 1973, and upon recom-
mendation of the Executive Committee, the sum of $1,300,000 was appropriated from the
Unrestricted Capital Fund for use at the discretion of the President. To provide for site
development expenses at the Las Campanas Observatory for the fiscal year beginning
July 1, 1973, and upon recommendation of the Executive Committee, the sum of $424,000
was appropriated from the Unrestricted Capital Fund for use at the discretion of the
President.
The Chairman reported the death of Dr. Edward A. Ackerman, Executive Officer of the
Institution. The President and Dr. Haskins spoke of Dr. Ackerman's many contributions
and service to the Institution and Dr. Haskins proposed the following resolution, which
was unanimously adopted by the Trustees :
Be it Resolved, That the Board of Trustees of the Carnegie Institution of
Washington records its deep sense of loss in the death of Edward A. Ackerman,
the Executive Officer of the Institution. In his idealism, in his constancy, in his
devotion, in his high sense of duty and his unflagging service to the Institution,
Dr. Ackerman brought inspiration to all its enterprises. The entire Institution
family will always remember what he gave it through almost fifteen years of
unremitting service.
And Be it Further Resolved, That this resolution be entered on the Minutes
of the Board of Trustees and that a copy be sent to Mrs. Ackerman.
Arthur Andersen 8c Co.
1666 K Street, N.W.
Washington, D. C. 20006
September 7, 1973
REPORT OF INDEPENDENT PUBLIC ACCOUNTANTS
To the Auditing Committee of
Carnegie Institution of Washington:
We have examined the statement of assets and funds of
CARNEGIE INSTITUTION OF WASHINGTON (a nonprofit corporation
chartered by Act of the United States Congress) as of June 30,
1973 and 1972, the related summary statement of changes in funds
for the year ended June 30, 1973, and the supporting exhibits
and schedules on pages 744- through 756. Our examination was made
in accordance with generally accepted auditing standards, and
accordingly included such tests of the accounting records and
such other auditing procedures as we considered necessary in the
circumstances .
In our opinion, the accompanying financial statements
and supporting exhibits and schedules (pages 744 through 756)
present fairly the assets and funds of Carnegie Institution of
Washington as of June 30, 1973 and 1972, and the changes in
funds for the year ended June 30, 1973, resulting from the con-
sistent application of the cash basis of accounting.
/\^JbU^ A^J^^ f &>>
743
CARNEGIE INSTITUTION OF WASHINGTON
STATEMENT A
ASSETS AND FUNDS
JUNE 30, 1973 AND 1972
ASSETS
Cash
Advances
Investments (cost)*, Schedule 2:
Governmental bonds
Nongovernmental bonds
Mortgage
Common stocks
Cash advanced at discretion of Investment Advisor .
Total investments
Land (cost)
Buildings (cost) ,
Total assets ,
FUNDS
Operating Fund, Exhibit 1
Restricted Grants, Exhibit 2
Endowment and Special Funds, Exhibit 3
Land and buildings ,
Total funds
1973 1972
264,871.84 $ 260,407.30
29,724.07
1,499,375.00
30,881,906.25
2,884.99
51,346,547.20
83,730,713.44
763,227.83
2,948,456.10
5 3,740,158.95
(43,492.73)
80,328,643.13
3,711,683.93
7,556.38
1,000,000.00
39,004,386.30
5,891.84
41,662,027.11
(814,112.77)
80,858,192.48
758,912.33
2,912,275.93
$87,736,993.28 $84,797,344.42
$ 3,349,557.73
(33,871.36)
77,810,469.79
3,671,188.26
$87,736,993.28 $84,797,344.42
* Approximate market value on June 30, 1973: $124,770,315; on June 30, 1972: $128,825,707
744
CM
CD Tt © O tr-
00 m © © t-h
U0 CM
1— 1
©
CO
00
00
cm
n
CO
CO*
i6c- ccd Tt
CM O t- O CO
O CO CO to m
in o
O Ol
^t ©^
cm
1— 1
CO
CO
in
9
CO
CO
Oi
o
in
tr-
io
oo" 00 "* co" tr^
o co ■<* n m
H 1C lO CO
t-H Oi
CM t-H
T-H
CM
T— 1
o
CO
©"
CO
co"
co
•hh"
00
«©■
cd"
cm" h
t-H
T-H
00
CN
00
►J GQ
CO
CM
00
00
t-
CD_
co"
in cm
m t-h
in o
© Oi
Tf CO
T-H CX
tr-
ee
in
©
->*
O
tr-
CO
in
©
©"
CO
ex
CO
00
CO
T-H
T-H
t>
CO'
z
o
a
z
co
<
O
z
o
i—i
Eh
CO
Z
a
Z
05
<
co
Q
Z
co
a
u
o
Eh
z
W
<
Eh
CO
<
CO
CO
Eh
Z
w
Ed
Eh
<
CO
CO
t>
ex
o
co
Z
P
•-3
Q
H
O
Z
05
<
>h
EC
£ ?5
~ CO
e co
§ 0 -c:
-q cy !*J
S ja.
trj CO
CD
CM
•~ g 15
s-1
05
CX
©
CD
c-
t>
as-
CD
CO
T-H
tr-
ee
co"
CO
CO
t>
tr^
m
m
CO
CO"
6©-
CM
t>
©
"3
!-3
CO
t>
CO "tf
oq in
o
CO
tr-
in
in t>
CM ©
o_ co_
o
o
CO
co
in
in
t>
m
oo" co"
O CO
t-h m
co"
tr-
CO
t^r
in
o
o
ir^
co
in
as
o
tr-
o o o
o o o
o o o
o o o
O^ CD^ 0_
cn" h in
m co t-h
on ^
CO
CM t-h
TT
rr
CO
CO
CO
t-H
CO
CO
CO
t-
tr-
Tf
T-H
T-H
CO
oo"
oo"
oo"
T-H
T-H
CM
lO
m
CO
CM
CM
o
O0
e»
o
o
CO
t>
CO
m
o
o
CD
t>
CO
-H
m
Q
o o o
o
00
CM
m
O
© o o
©#
ir-
CM
Oi
t>
o o ©
tr-^
in
T-H
00
CO
© o o
CO
CO
o
in
m
O CO o
T— 1
in
CO
t-H
®
cm t-h m
00
i>
o
©
o
m co t— <
CO
th
ex
^
t>
©_ CO
T-H
t>
CO
t-
m"
cm"
oo"
t>
co"
t>
co
t>
ex
OS
CO"
m
m
^^
^-^
t>
CO
co
t>
T-H
CM
CM
cx
CD
Tt<
S
CO
-^
€»
^^
cd
be
1
'a
cd
o
.. CD
CO N
2 3
12 co
05
cd "o
a
c^
QJ
CD
E
6-
CO
+J
co
c
CO
>
c
or
c
CO
-a
^ Q
co
05
-a
co
N
&
o
co"
be
a
3 «3
o o
cc
>
;-
CO
O
CO
cd
s-s
O Oh
CO
c
O
"S **
K3 CO
a -a
o =3
a^
a
<
a
w
CO
e
o
c
o
-£3
CO
<
O &
55 3
cd Js
^ co
co S
c '-h
cd cd
r_D O
o
CO
a
co
s-| co
cd t>
a> ex
>> T-H
be
C3
©"
CO
CO
<D
H
9
-M
3
T3
CO
V.
-5
CO
be
T>
•S
c
CO
§
co"
13
J3
o
-1
C
a)
"ed
Q
Z
CQ
745
CARNEGIE INSTITUTION OF WASHINGTON
EXHIBIT 1 CHANGES IN OPERATING FUND
FOR THE YEAR ENDED JUNE 30, 1973
Balance , July 1, 1972 $ 3,349,557.73
Appropriations, Statement B:
Budget, July 1, 1972 to June 30, 1973 $5,709,567.00
Las Campanas Observatory
Unrestricted Capital 2,052,000.00
Greenewalt Astronomy 361,600.00
Special Instrumentation 15,000.00 8,138,167.00
Total available for expenditures 11,487,724.73
Expenditures:
Salaries 2,987,637.44
Educational and research supplies 1,872,508.27
Building maintenance 547,639.20
Employee benefits, retirement contributions 421,984.47
Equipment 371,532.63
Administrative 365,754.79
Fellowship grants 205,326.28
Financial advisory services 194,472.78
Employee benefits, other 170,836.00
Taxes 156,799.30
Travel 116,885.64
Publications 108,757.20
Consulting fees and insurance. 92,250.13
Awards 33,168.50
Commissary 29,699.71
Rent 17,497.00
Buildings 17,090.05
Shop 16,178.75
Entertainment 11,803.64
Fellowship travel 5,428.50
Land 4,315.50
Total expenditures 7,747,565.78
Balance, June 30, 1973 $ 3,740,158.95
746
EXHIBIT 2
CARNEGIE INSTITUTION OF WASHINGTON
CHANGES IN RESTRICTED GRANTS
FOR THE YEAR ENDED JUNE 30, 1973
Expenditures
Balance Balance
July 1, 1972 Grants Salaries Other June 30, 1973
Aldrich Gift $ 2,800.00 $ 300.00 $ 2,500.00
Carnegie Corporation of
New York $ 955-68 80,000.00 80,639.90 315.78
Department of the
Air Force (3,207.50) 9,943.00 $ 5,967.81 767.69
Jane Coffin Childs
Memorial Fund 9,125.00 9,124.94 .06
Jet Propulsion Laboratory . . 196-68 117.68 79.00
Medical Research Council . . 119.59 119.59
Mosely Gift 10,800.00 10,800.00
Muscular Dystrophy
Association 500.00 134.18 365.82
National Aeronautics &
Space Administration ... . (37,397.43) 175,367.00 25,241.94 129,114.39 (16,386.76)
National Science
Foundation 5,756.10 198,000.00 24,579.88 210,335.02 (31,158.80)
Office of Naval Research .. . (3,661.46) 20,988.00 14,007.56 8,388.98 (5,070.00)
Public Health Service 7,653.00 7,153.00 500.00
Wistar Institute 3,366.98 29,200.00 6,399.96 31,604.85 (5,437.83)
Total ($33,871.36) $544,376.00 $76,197.15 $477,800.22 ($43,492.73)*
'Does not include grants to be received as follows:
National Science Foundation $228,637.55
National Aeronautics & Space Administration 80,150.00
Wistar Institute 23,353.00
Public Health Service 8,000.00
Office of Naval Research 5,070.00
Jane Coffin Childs Memorial Fund 2,187.50
Muscular Dystrophy Association 500.00
$347,898.05
747
CO
8*
Is?
a
-a
O CD
O ""*
003CO
005CD
00
©
© i>
O CO
o 00
O CO lO
O "tf OS
O rH Tf
co
t>
CO
o t>
©" t> >o
O O CD
OHO
CM" Oi"
CM CM
€6-
o" 00
1— 1 T— 1
CO lo
coos
rH tjh
CM CM
co"^"
CM t— l
CM
CD rH CO rH rH T^
OS t- CO CO 00 l>-
H CD Tj? Oi CM Oi
Oi rH LO 00 00 CO
CM LO OS CO_C^C©
CD ©" CM" ©>" Cd" oo"
rH CM rH CO 00
t>
CO
Oi
rH
©
CO
l>
Tf
CM
CD
<*
00
Oi
CM
LO
co_
©"
00
#3-
O
a.
H
CiCMLOOrHOOOiOCO
^oqcDLOt^c^coooq
CO
LO
t>
LO
050'CDCOCOH|>CO^
t- ■* co in t> o oi co m
oo ai cor as r> rH -^ t£ co
O51OHO5C0C0HH
rH^^rH^-'-^^^
CO
Oi
Oi
CN
LO
oo
co
CO
LO
oc
T—
co
CD
0C
CM
00
00
rH
Oi
Oi
LO O rt* CO CM
CO O CM CO 00
t> © LO LO LO
OO OHIO
LO CD Tt* O
crTrH
CD
CO
000050
oqt>o
"tf O* CD O*
rH O COO
Ho^Qoq
hioooo"
rH CM
o
o#
t>
CD
rH
CO
CO
rH
CO
T3
i—
cd
■2 c
CO
"o
e -5
z
Q
z
CO
03
o
Eh
&H
CO
t>
o
Z
<
(35
rH
CO
<
6
H
Oh
CO
o"
co
H
Z
-a
e
co
6^
O
Q
z
*-3
'6
z
g
<
Q
H
Q
In
Z
z
«d
D
H
03
CD CD
S S
H
o
<
w §
CO
Q
H
s> s
Z
1— 1
z
>H
■- • — 1
H
H
H
*—
a
§
Eh
z
OS
CO
O
03
O
CM
<
z
<c
fe
CD Oi
o
LO
©
00
CM
CO
CO"
LO
°i
co"
€©■
o
©
Oi
CD
l>
LO
LO
cm"
o o-
rH OO
oi lo
Oi o
I> LO
cm"cd"
00Tfl>t>H^
r* © rH Oi <N "<t
CD rH CM t>- LO rH
Oi CM rH LO CO f-
CM t> CO CM Oi Tt
Tf
I>
rt
l>
CO
CM
T^
t>
CM
lo ;
LO
LO
CO
t>
LO
co"
«©■
OiCMLOOrHCOCTiOCO
^COCDiOI>l>COOOO
Cr3 0'cDCDCOrHI>COTj5
tr-^fcocMtr-oaicMLO
OOCTiCDaiC^rHTfTfCO
Oi" LO" rH Oi CO" CO" rH rH
o
o
o
t>
o o
o
rH
o
o
Q
T^
o
o
o
CO
CD
o
o
rH
rH
o
LO
rf
CD rH
CO
CO
"tf
&
O rH
O TtJ
o" cri
o o
o^o^
o"rH
s*
O CM
O rH
a
CM LO
'"D
CM CM
#3-
O Oi CD
O C75 LO
O* CO l>
O ^ CO
O rH rH
o" t> co"
O O Oi
o" co"
Oi
CM
CD
00
CM
as
rH
CD
00
o"
rH
CD_
co"
€©■
LO
LO
CD
rH
LO"
LO
CO
O rt<
CM CD
CD CD*
LO t-
LO rH
rH Oi
CD Oi
co t>
CD oi
CD ">*
CO.O^
Oo"oo"
rH Oi
Oi 00 CM Tt< rH CO
CM 00 LO CO Oi Oi
Oi O* Oi rH LO ^
CD O CM CO t- CO
HTthHOO
CO
O
1>
rjH
CD
CO
Oi
CO
LO
CD
Oi
CO ;
rH
TjH
CD_
CO [
€©■
CO CM Oi
■*HCJ
rH O* 00
CM rH CD
Oi_ 00 CD_
LO Oi rH
CM
OM050
CO CO t> 00
CD CO CD LO
00 CO CO CM
cq Tf oq_ Tt
^"o"oo"oi"
co Oi lo
LO
Oi
oi
CD
t>
co
E-
CQ
h-H
X
X
c
CD
® -H
.a I
CU , i -JTh
e &°
H CO
s-/ T3 -*J
O^ H
-3 cc %
<03 c
o
CD
Z
c3
O co o
,-Hl
cu jy -a
CO h-H
c3 rS CQ
OO03
'Sort,
co ^
E-a
o a
o < ^
i o
o
b£
C
►SCO
CO —h -*-'" co • "-1
O cd ® "S co .2 -O
J2 co cu STJ S cu
C 0) O 3 CO .h
P03 03XOiS
C CO ,D _
O CO 3 D- CO
I/)
~- * _D
a co -3
•2 3 o -
<
,±3 cp
CO i^
£ CP
co cr;
c -
CO 0> 24
(_i CO CO
E d
3 c3
J2
rj CO
o •— '
S 13 ^
J7 CO cu o
o oL cu o
SCO H 5:
o
Eh
748
-J»> -~> ^J
ft te S
ft Co
2 co
a. e
ft 3
©
as
i— i
00
CO
t>
CD
as
as
LO
OO
©
LO
H
LO
CD
CN
O
LO
00
©
CD CD
00 CD
CO ©
00
CD
LO
CO
CD
CM
o
O
o
O
i— i
©
l>
LO
LO
o
©
CO
o
CD
00
CN
CO
o
CM
o
O
LO
©
as
o
00
r^
i— 1
CD
CM
T— 1
©
CO
tT
LO
CM
©
CD
LO
T— 1
00
T*
CN
T-H
as
T*
LO
CD
00
CO
I>
co
CO
LO
as
00
LO
i— i
©'
rr
t>
CO
o
El
CO
cd
"*
o
o
LO
CM
t>
CM
T-H
CO
CD
00
CM
o
00
LO
T-H
CO
«*
CO
^t1
©
CO
T-H
Tt<
LO
CO
o
t-
3
LO
00
as
00
CM
t>
as
CD
CO
©
LO
as
LO
CM
t>
o
LO
CO
as
as
T-H
00
CD
LO
©
LO
CO
r^
""a
CD
tr-
00
CM
00
as
CO
Tt<
o
CO
l>
CO
rr
^
i— l
io
00
"tf
o
as
LO
as
CM
as
t— 1
C-
as
Tf
CD
CM
T-H
LO
00
T— 1
CM
CD
CM
OS
t>
CD
00
OS
ft
©
©
LO
as
l>
CO
LO
T-H
X
63
t— i
T-H
CM
00
t>
LO
CD
LO
C-'
€©■
r^
t-H
CM
CM
t-
t-H
Tt<
LO
TF
otf
CM
00
CO
T-H
as
■*
CD
CD
©
t>
T-H
as
O0
Tf
h-
«tf
Tf
©
T-H
CO
i— 1
CO
00
as
t-H
©
CO
CD
©
©
O
LO
00
LO
CO
©
as
LO
CM
LO
©
CD
l>
CD
LO
CD
©
LO
CM
CM
©
CO
©
CM
t-
CM
as
o
CO
CO
as
LO
CM
CO
CM
CO
rH
CM
^■^
"ft
t-H
s— '
T-H
t-H
^^
" — '
CD
^^
'
s-"'
LO
00
t>
CM
oq
CO
o
©
©
©
©^
o
LO
CO
as
co'
co
35
co'
CM
co
z
Q
o
Z
CO
H
P
o
a
fc
T-H
a
o
CO
<
H
63
&
<1
Z
6h
05
63
P
O
z
O
Q
63
u
O
Q
Z
P
>h
63
05
<
05
<
t/3
2
63
Z
2
>h
a
P
GO
63
K
H
63
Z
o
Q
05
O
Q
CQ
63
J
P
Q
63
O
GO
CO
T-H
as
©
©
©
©
©
©
©
©
©
©
©
©
©
©
•2
LO
00
©
CO
©
©
©
o
©
©
©
©
©
©
©
©
"t
©
CM
c^
T-H
CO
©
©
LO
©
©
©
©
©
©
t>
o
LO
CM
©
CM
T-H
c^
LO
©
o
©
©
©
©
©
©
©
as
CM
t-H
CO
CM
CM
00
©
00
©
©
I>
LO
©
©
T-H
ft
CD
CD
CO
©
Tf
00
t-
T-H
©
LO
^P
LO
LO
©
©
00
O
00
Th
©
©
00
CM
©
©
T-H
t>
t>
00
©
LO
CO
ft
a.
as
as
CO
©
t>
CO
©
t-H
CO
T-H
CM
00
^C
€£-
«9-
CD O t>
CD -S -
e
P
a, --s
LO
00
LO
CO
C50
as
E
&
co
as
©
CM
00
©
as
CO
CX)
CO
00
©
LO
CM
CO
©
CD
00"
©
©
©*
©
©"
CO
t>
00
as
c^
©^
©"
©
t>
©
CM
©
©
©
©
CO
00
LO
00
©
00
CM
LO
CM
©
©
©
LO
©
©
©
©
© Xf
© ©
© T-H
co'
LO
CM
©
©
©
©
00
©
LO
LO
©
t>
LO
©
>5
o
as'
CJ
CO
as
Id
CO
as
cs
©
CO
c
o
-t-J
crt
43
cd
>
o
-t-5
'S
6
c
o
u
CO
H
as
cs
as
a
>
as
!*H
_o
cd
>
as
cd
•
p
as
cd
CO
Ch
CO
CO
be
as
CO
©
O
CO
©
O
CO
o
©
cd
o
as
co'
o
as
CO
'a
_o
-4H)
cd
o
CO
as
co'
-IH1
as
co'
■+J
as
>>
O
CO
c
cd
as
o
—
a
ft
a
o
CO
03
>>
CO
o
F
as
c
c
>
a
cd
c3
a
S
cd
n
13
o
ft
be
o
as
05
a.
©
<
©
3
=*5
be
as
©
^
-a
cd
o
©
as
CO
JO
'o
>5
o
CO
o
c3
o
as
Id
"3
as
as
O
as
as
o
cd
'3
c
OS
be
g
—
cd
CS
o
c
a
P,
as
as
o
as
CO
m
3
CO
o
©
c
CO
5*H
«4— 1
c
C
Q.
_
C
c
cd
O
cd
as
d
as
as
as
o
bi
b
o
c
S
J
O
63
CL
O
05
O
O
U
63
63
6-
U
p
©
©
©
©'
co
©'
eft-
$
O
cr-
749
SCHEDULE 2
CARNEGIE INSTITUTION OF WASHINGTON
INVESTMENTS, JUNE 30, 1973
Description Par
Federal Agency Bonds
Federal Home Loan Banks,
Cons., 7.30s, 1983 $1,000,000
Twelve Federal Land Banks,
Cons., 7.30s, 1982 500,000
Total
Foreign Bonds
Alberta Government Telephone Commission,
Deb., 4% s, 1989 700,000
Alcan Aluminum Corporation,
Prom. Note, 43/4 s, 1984 672,000
Aluminum Co. of Canada, Ltd.,
S. F. Deb., 9 % s, 1995 500,000
Churchill Falls (Labrador) Corp. Ltd.,
1st Mtg. Series A, 7 % s, 2007 800,000
IACLtd.,
Sec. Note, 5 % s, 1982 750,000
Quebec Hydro-Electric Commission,
S. F. Deb., 5s, 1988 681,000
Quebec Hydro-Electric Commission,
Deb. Series BN, 9 % s, 1995 500,000
Toronto (Municipality of Metropolitan),
S. F. Deb., 5s, 1979 500,000
Total
Public Utility Bonds
American Telephone & Telegraph Company,
Deb., 8.70s, 2002 925,000
American Telephone & Telegraph Company,
Deb., 8% s, 2000 375,000
Consumers Power Co.,
1st Mtg., 4 % s, 1987 4,000
Minnesota Power & Light Co.,
1st Mtg., 3V8 s, 1975 200,000
Pacific Gas & Electric Co.,
1st & Ref. Mtg. Series BB, 5s, 1989 .... 250,000
Pacific Power & Light Co.,
1st Mtg., 4% s, 1986 250,000
Potomac Electric Power Co.,
Deb., 45/8s, 1982 210,000
Washington Water Power Co.,
1st Mtg., 47/8 s, 1987 300,000
Total
Cost
Approximate
Market
999,375.00
500,000.00
1,499,375.00
$ 991,250
496,875
1,488,125
700,000.00
476,000
672,000.00
551,040
517,267.09
545,000
800,000.00
746,000
750,000.00
585,000
669,082.50
514,155
504,125.95
540,000
498,637.50
446,250
5,111,113.04
4,403,445
925,000.00
994,375
375,500.00
404,531
4,013.38
3,030
200,510.37
184,750
251,050.16
187,813
251,550.67
183,750
212,225.19
170,363
300,000.00
229,125
2,519,849.77
2,357,737
750
CARNEGIE INSTITUTION OF WASHINGTON
mWESTMENTS-Continued
Description Par
Industrial and Miscellaneous Bonds
Atlantic Richfield Co.,
Demand Note $1,175,000
Bankers Trust Co. New York, N. Y.,
Ctf. of Dep., 7.91s, 1973 1,700,000
Bankers Trust Co. New York, N. Y.,
Ctf. of Dep., 8.18s, 1973 825,000
Boeing Co.,
Notes, 6 % s, 1986 740,000
Columbia Broadcasting System Inc.,
Prom. Note, 5Y2 s, 1991 800,000
Commercial Credit Co.,
Note, 35/8s, 1976 400,000
Crown Zellerbach Corp.,
Prom. Note, 45/8s, 1981 225,000
Crown Zellerbach Corp.,
S. F. Deb., 87/8 s, 2000 1,000,000
Erie Mining Company,
1st Mtg. Series B, 4V2s, 1983 355,000
First National City Bank,
Capital Conv. Notes, 4V2 s, 1990 1,000,000
First National City Bank,
Ctf. of Dep., 7.95s, 1973 3,700,000
Fischbach & Moore Inc.,
Conv. Sub. Deb., 43/4 s, 1997 238,000
Four Corners Pipe Line Co.,
Sec. Note, 5s, 1982 116,000
General Electric Co.,
Demand Note 370,000
General Electric Credit Corp. (N. Y.),
Prom. Note, 5s, 1975 500,000
General Motors Acceptance Corp.,
4 % s, 1982 480,000
General Motors Acceptance Corp.,
Deb., 5s, 1977 200,000
General Motors Acceptance Corp.,
Deb., 5s, 1981 200,000
General Portland Cement Co.,
Conv. Sub. Deb., 5s, 1977 150,000
GTE Sylvania Inc. ,
Demand Note 240,000
Halliburton Co.,
Conv. Note, 4s, 1997 1,000,000
Hystron Fibers Inc.,
Note, 5 % s, 1986 934,000
Intl. Harvester Credit Corp.,
Demand Note 87,000
Irving Trust Co. New York, N. Y.,
Ctf. of Dep., 7 5/8s, 1973 500,000
Kaiser Aluminum & Chemical Corp.,
1st Mtg., 5V2s, 1987 193,000
Cost
Approximate
Market
1,175,000.00
$ 1,175,000
1,700,000.00
1,699,779
825,000.00
824,769
740,000.00
634,550
800,000.00
666,000
401,380.57
364,000
225,000.00
196,875
1,026,928.91
1,065,000
344,662.40
256,488
1,075,715.00
2,340,000
3,700,000.00
3,699,445
246,725.00
202,300
116,000.00
102,080
370,000.00
370,000
500,000.00
472,500
390,912.00
387,600
195,000.00
182,500
199,000.00
169,500
154,500.00
136,313
240,000.00
240,000
1,000,000.00
1,275,000
934,000.00
817,250
87,000.00
87,000
500,000.00
499,915
193,000.00
158,260
751
CARNEGIE INSTITUTION OF WASHINGTON
INVESTMENTS-C ontinued
Description
Industrial and Miscellaneous Bonds
—Continued
Kresge (S. S.) Company,
Prom. Note, 47/8 s, 1983
Manufacturers Hanover Trust Co., N. Y.,
Ctf. of Dep., 7.96s, 1973
Mercantile Stores Co., Inc.,
S. F. Deb., 8.70s, 1995
Montgomery Ward Credit Corp.,
Deb., 47/8s, 1980
NCNBCorp.,
S. F. Deb., 8.40s, 1995
Sears Roebuck Acceptance Corp.,
Sub. Deb., 45/8s, 1977
Shell Funding Corp.,
Collat. Tr. Note, 4 % s, 1985
Trailer Train Co.,
4 % s, 1976
United Air Lines Inc.,
Notes, 5s, 1984
Woolworth (F. W.) Company,
Prom. Note, 5s, 1981
Total
Bonds, funds invested
Mortgage
Alfred D. Hershey and Harriet D. Hershey,
5V2s, 1974
Common Stocks
Alcon Laboratories, Inc
AMP Incorporated
Aquitaine Co. of Canada, Ltd
ARA Services, Inc
A T Cross Co
Avon Products, Inc
Baker Industries, Inc
Baker Oil Tools, Inc
Baxter Laboratories, Inc
Charles River Breeding Labs, Inc
Par
Cost
Approximate
Market
583,333.35
$ 583,333.35
$ 469,583
,700,000
1,700,000.00
1,699,524
500,000
500,000.00
525,000
200,000
199,000.00
163,250
500,000
497,500.00
520,000
525,000
511,505.00
469,875
776,000
776,000.00
653,780
194,127.01
194,127.01
183,935
700,000
700,000.00
577,500
449,654.20
449,654.20
367,592
23,250,943.44
23,652,163
32,381,281.25
31,901,470
2,884.99
Shares
2,884.99
Cost
2,885
Approximate
Market
8,000
261,374.30
260,000
36,000
783,265.58
1,566,000
14,000
398,680.60
294,000
11,300
1,516,315.57
1,350,350
5,000
297,500.00
263,750
24,000
1,111,961.35
2,964,000
10,000
350,094.89
190,000
20,000
551,538.00
515,000
18,000
973,093.04
837,000
6,000
221,784.30
165,000
752
CARNEGIE INSTITUTION OF WASHINGTON
mVESTMENTS-C ontinued
Description Shares
Common Stocks
— Continued
Chemed Corporation 8,000
Chesebrough-Pond's, Inc 28,000
Christiana Securities Company 5,221
Coca-Cola Bottling Co. of N. Y 48,000
Coca-Cola Company (The) 35,000
Davis Water and Waste Industries 8,000
Disney (Walt) Productions 29,132
Eastman Kodak Company 35,000
Economics Laboratory, Inc 20,000
Exxon Corporation 23,561
Federated Department Stores, Inc 24,000
First National City Corp 60,416
General Electric Company 30,000
General Motors Corporation 36,169
Gilbert Associates, Inc 1,500
Gilbert Associates, Inc 5,400
Golden State Foods Corporation 3,000
Hughes Supply, Inc 6,000
International Business Machines Corp. . . . 28,750
International Tel & Tel Corp 31,000
Johnson & Johnson 35,000
Kresge (S. S.) Co 33,651
Loctite Corp 5,000
Lowes Companies, Inc 13,334
McDonalds Corp 20,000
Merck & Co., Inc 46,000
Mgic Investment Corporation 15,000
Minnesota Mining & Manufacturing Co 38,000
National Chemsearch Corp 15,000
National Data Corp 5,000
New England Nuclear Corp 5,600
Oshmans Sporting Goods, Inc 9,000
Peabody Galion Corporation 30,000
Penney (J. C.) Company, Inc 35,000
Perry Drug Stores, Inc 2,850
Philip Morris, Inc 20,000
Pickwick International, Inc 20,000
Polaroid Corporation 11,000
Puritan Bennett Corp 3,000
R P Scherer Corporation 11,000
Cost
Approximate
Market
276,762.50
911,861.03
655,683.00
853,643.39
561,147.87
295,337.50
1,060,573.16
418,234.14
751,375.00
671,594.17
582,805.81
348,278.77
376,089.41
1,206,754.99
55,750.00
227,675.00
112,775.00
219,058.10
812,740.24
1,983,098.08
678,921.26
1,272,994.33
235,562.50
809,684.18
1,068,119.94
806,982.24
1,136,227.50
1,916,654.93
1,193,861.66
145,775.00
175,453.23
251,571.35
1,061,300.04
1,967,073.28
38,475.00
2,580,601.05
753,380.76
1,197,431,69
186,575.00
335,800.00
276,000
2,177,000
835,360
876,000
5,009,375
140,000
2,235,881
4,777,500
760,000
2,320,759
930,000
2,477,056
1,721,250
2,405,239
44,250
159,300
45,000
82,500
9,113,750
941,625
3,985,625
1,139,928
248,750
680,034
1,142,500
4,134,250
1,021,875
3,182,500
1,335,000
131,250
205,100
146,250
603,750
2,708,125
14,963
2,445,000
455,000
1,509,750
187,500
357,500
753
CARNEGIE INSTITUTION OF WASHINGTON
INVESTMENTS-Confmaed
Description
Common Stocks
— Continued
Ralston Purina Co
Reynolds & Reynolds Co
Rouse Company
Rubbermaid, Inc
Sears Roebuck and Co
Schlumberger, Ltd
Snap-on Tools Corp
Squibb Corp
Standard Oil Co. of California
Steak & Ale Restaurants of America, Inc.
Texaco, Inc
Times-Mirror Company
Waste Management, Inc
WD 40 Company
Xerox Corp
Common stocks, funds invested
Aggregate Investments . .
Shares
Cost
Approximate
Market
45,000
3,000
6,200
1,400
20,000
60,000
3,000
20,000
33,000
5,600
48,380
56,000
12,000
8,000
25,000
$ 1,979,080.46
132,000.00
165,225.00
107,403.09
1,821,704.03
1,789,328.00
170,500.00
1,862,399.23
2,622,954.86
250,288.84
249,172.89
1,265,083.38
264,000.00
208,587.50
1,799,530.19
51,346,547.20
$83,730,713.44
$ 1,603,125
124,125
84,475
112,000
1,902,500
5,940,000
166,500
2,000,000
2,541,000
147,000
1,657,015
980,000
234,000
116,000
3,890,625
92,865,960
$124,770,315
754
SCHEDULE 3
CARNEGIE INSTITUTION OF WASHINGTON
SUMMARY OF INVESTMENT TRANSACTIONS
FOR THE YEAR ENDED JUNE 30, 1973
Cash awaiting investment, June 30, 1972
Sales and Redemptions
Capital
Gain
Loss
Book Value
Bonds
Mortgage
Common stocks
Realized capital gain, nek-
Statement B
5 8,656.23 $1,795,911.31 $119,569,016.57
3,006.85
8,496,729.68 133,901.83 7,009,503.34
{,505,385.91 $8,505,385.91
1,505,385.91 1,929,813.14 126,581,526.76
6,575,572.77 6,575,572.77
Total sales and redemptions 133,157,099.53
Gifts of securities recorded at fair market value 312,400.00
Cash returned to Investment Advisor (814,112.77)
Cash transferred to Operating Fund (4,002,886.92)
Total 128,652,499.84
Acquisitions
Bonds $111,945,911.52
Common stocks 16,694,023.43
Total acquisitions 128,639,934.95
Cash awaiting investment, June 30, 1973 $ 12,564.89
755
CARNEGIE INSTITUTION OF WASHINGTON
SUMMARY OF SIGNIFICANT ACCOUNTING POLICIES
Cash Basis of Accounting
The Institution maintains its records and the accompanying financial statements have been pre-
pared on the cashreceipts and disbursements basis of accounting. Accrued income and accrued expenses
are not reflected in the financial statements; however, the balance of unexpended appropriations is
reserved for outstanding obligations.
Land, Buildings and Equipment
Expenditures for land and buildings are charged to current operations as paid. When individual
projects have been completed, the related costs are capitalized in the land or buildings account. The
Institution follows the policy of not depreciating its buildings.
Expenditures for equipment, including telescopes, are charged to current operations as paid, and
the cost of equipment is not capitalized.
Retirement Plan
The Institution has a retirement plan in which all staff members are eligible to participate in
accordance with Article 3 of the Plan. Actuarially determined contributions are funded currently by the
Institution, and there are no unfunded past service costs. Benefits under the Plan upon retirement are
dependent upon the investment performance of the Institution's Retirement Trust.
756
Articles of Incorporation
Jfifig-eigjrtlr Congress of i|c Intteb States of America;
&t tfue j&econd Jfossiott,
Begun and held at the City of Washington on Monday, the seventh day of December, one
thousand nine hundred and three.
an* ACT
To incorporate tbe Carnegie Institution of Washington.
Be it enacted by the Senate and House of Representatives of the United
States of America in Congress assembled, That the persons following, being persons
who are now trustees of the Carnegie Institution, namely, Alexander Agassiz,
John S. Billings, John L. Cadwalader, Cleveland H. Dodge, William N. Frew,
Lyman J. Gage, Daniel C. Gilman, John Hay, Henry L. Higginson, William
Wirt Howe, Charles L. Hutchinson, Samuel P. Langley, William Lindsay, Seth
Low, Wayne MacVeagh, Darius 0. Mills, S. Weir Mitchell, William W. Morrow,
Ethan A. Hitchcock, Elihu Root, John C. Spooner, Andrew D. White, Charles
D. Walcott, Carroll D. Wright, their associates and successors, duly chosen, are
hereby incorporated and declared to be a body corporate by the name of the
Carnegie Institution of Washington and by that name shall be known and have
perpetual succession, with the powers, limitations, and restrictions herein contained.
Sec. 2. That the objects of the corporation shall be to encourage, in the
broadest and most liberal manner, investigation, research, and discovery, and
the application of knowledge to the improvement of mankind; and in particular —
(a) To conduct, endow, and assist investigation in any department of
science, literature, or art, and to this end to cooperate with governments,
universities, colleges, technical schools, learned societies, and individuals.
(b) To appoint committees of experts to direct special lines of research.
(c) To publish and distribute documents.
(d) To conduct lectures, hold meetings, and acquire and maintain a library.
(e) To purchase such property, real or personal, and construct such building
or buildings as may be necessary to carry on the work of the corporation.
757
758 CARNEGIE INSTITUTION
(f) In general, to do and perform all things necessary to promote the
objects of the institution, with full power, however, to the trustees hereinafter
appointed and their successors from time to time to modify the conditions and
regulations under which the work shall he carried on, so as to secure the
application of the funds in the manner best adapted to the conditions of the time,
provided that the objects of the corporation shall at all times be among the
foregoing or kindred thereto.
Sec. 3. That the direction and management of the affairs of the corporation
and the control and disposal of its property and funds shall be vested in a board
of trustees, twenty-two in number, to be composed of the following individuals :
Alexander Agassiz, John S. Billings, John L. Cadwalader, Cleveland H. Dodge,
William N. Frew, Lyman J. Gage, Daniel C. Gilman, John Hay, Henry
L. Higginson, William Wirt Howe, Charles L. Hutchinson, Samuel P.
Langley, William Lindsay, Seth Low, Wayne MacVeagh, Darius 0. Mills,
S. Weir Mitchell, William W. Morrow, Ethan A. Hitchcock, Elihu Hoot,
John C. Spooner, Andrew D. White, Charles D. Walcott, Carroll D. Wright,
who shall constitute the first board of trustees. The board of trustees shall
have power from time to time to increase its membership to not more than
twenty-seven members. Vacancies occasioned by death, resignation, or otherwise
shall be filled by the remaining trustees in such manner as the by-laws shall
prescribe ; and the persons so elected shall thereupon become trustees and also
members of the said corporation. The principal place of business of the said
corporation shall be the city of Washington, in the District of Columbia.
Sec. 4. That such board of trustees shall be entitled to take, hold and
administer the securities, funds, and property so transferred by said Andrew
Carnegie to the trustees of the Carnegie Institution and such other funds or
property as may at any time be given, devised, or bequeathed to them, or to such-
corporation, for the purposes of the trust ; and with full power from time to time to
adopt a common seal, to appoint such officers, members of the board of trustees or
otherwise, and such employees as may be deemed necessary in carrying on the
business of the corporation, at such salaries or with such remuneration as they may
deem proper ; and with full power to adopt by-laws from time to time and such rules
or regulations as may be necessary to secure the safe and convenient transaction
of the business of the corporation; and with full power and discretion to deal
with and expend the income of the corporation in such manner as in their
judgment will best promote the objects herein set forth and in general to have
and use all powers and authority necessary to promote such objects and carry out
the purposes of the donor. The said trustees shall have further power from time
ARTICLES OF INCORPORATION 759
to time to hold as investments the securities hereinabove referred to so transferred
by Andrew Carnegie, and any property which has been or may be transferred
to them or such corporation by Andrew Carnegie or by any other person,
persons, or corporation, and to invest any sums or amounts from time to time
in such securities and in such form and manner as are permitted to trustees
or to charitable or literary corporations for investment, according to the laws
of the States of New York, Pennsylvania, or Massachusetts, or in such securities
as are authorized for investment by the said deed of trust so executed by Andrew
Carnegie, or by any deed of gift or last will and testament to be hereafter made
or executed.
Sec. 5. That the said corporation may take and hold any additional
donations, grants, devises, or bequests which may be made in further support of
the purposes of the said corporation, and may include in the expenses thereof
the personal expenses wThich the trustees may incur in attending meetings or
otherwise in carrying out the business of the trust, but the services of the
trustees as such shall be gratuitous.
Sec. 6. That as soon as may be possible after the passage of this Act a
meeting of the trustees hereinbefore named shall be called by Daniel C. Gilman,
John S. Billings, Charles D. Walcott, S. Weir Mitchell, John Hay, Elihu Boot,
and Carroll D. Wright, or any four of them, at the city of Washington, in
the District of Columbia, by notice served in person or by mail addressed to
each trustee at his place of residence; and the said trustees, or a majority
thereof, being assembled, shall organize and proceed to adopt by-laws, to elect
officers and appoint committees, and generally to organize the said corporation;
and said trustees herein named, on behalf of the corporation hereby incorporated,
shall thereupon receive, take over, and enter into possession, custody, and
management of all property, real or personal, of the corporation heretofore known
as the Carnegie Institution, incorporated, as hereinbefore set forth under " An Act
to establish a Code of Law for the District of Columbia, January fourth, nineteen
hundred and two," and to all its rights, contracts, claims, and property of any
kind or nature ; and the several oflicers of such corporation, or any other person
having charge of any of the securities, funds, real or personal, books or property
thereof, shall, on demand, deliver the same to the said trustees appointed by this
Act or to the persons appointed by them to receive the same ; and the trustees
of the existing corporation and the trustees herein named shall and may take
such other steps as shall be necessary to carry out the purposes of this Act.
Sec. 7. That the rights of the creditors of the said existing corporation
known as the Carnegie Institution shall not in any manner be impaired by the
760
CARNEGIE INSTITUTION
passage of this Act, or the transfer of the property hereinbefore mentioned, nor
shall any liability or obligation for the payment of any sums due or to become
due, or any claim or demand, in any manner or for any cause existing against
the said existing corporation, be released or impaired; but such corporation hereby
incorporated is declared to succeed to the obligations and liabilities and to be held
liable to pay and discharge all of the debts, liabilities, and contracts of the said
corporation so existing to the same effect as if such new corporation had itself
incurred the obligation or liability to pay such debt or damages, and no such action
or proceeding before any court or tribunal shall be deemed to have abated or been
discontinued by reason of the passage of this Act.
Sec. 8. That Congress may from time to time alter, repeal, or modify this
Act of incorporation, but no contract or individual right made or acquired shall
thereby be divested or impaired.
Sec. 9. That this Act shall take effect immediately.
President of the Senate pro tempore.
vCcr-xjd /V~-
-e
By-Laws of the Institution
Adopted December 13, 1904- Amended December 13, 1910, December 13, 1912,
December 10, 1937, December 15, 1939, December 13, 1940, December 18, 1942,
December 12, 1947, December 10, 1954, October 24, 1957, May 8, 1959, May 13, 1960,
May 10, 1963, May 15, 1964, March 6, 1967, May 3, 1968, May 14, 1971, and August 31,
1972.
ARTICLE I
The Trustees
1. The Board of Trustees shall consist of twenty-four members with power to increase
its membership to not more than twenty-seven members. The Trustees shall hold office
continuously and not for a stated term.
2. In case any Trustee shall fail to attend three successive annual meetings of the
Board he shall thereupon cease to be a Trustee.
3. No Trustee shall receive any compensation for his services as such.
4. All vacancies in the Board of Trustees shall be filled by the Trustees by ballot at an
annual meeting, but no person shall be declared elected unless he receives the votes of
two-thirds of the Trustees present.
5. If, at any time during an emergency period, there be no surviving Trustee capable
of acting, the President, the Director of each existing Department, and the Executive
Officer, or such of them as shall then be surviving and capable of acting, shall constitute
a Board of Trustees pro tern, with full powers under the provisions of the Articles of
Incorporation and these By-Laws. Should neither the President, nor any such Director,
nor the Executive Officer be capable of acting, the senior surviving Staff Member of each
existing Department shall be a Trustee pro tern with full powers of a Trustee under the
Articles of Incorporation and these By-Laws. It shall be incumbent on the Trustees
pro tern to reconstitute the Board with permanent members within a reasonable time
after the emergency has passed, at which time the Trustees pro tern shall cease to hold
office. A list of Staff Member seniority, as designated annually by the President, shall
be kept in the Institution's records.
6. A Trustee who resigns after having served at least five years and having reached age
seventy shall be eligible for designation by the Board as a Trustee Emeritus. A Trustee
Emeritus shall be entitled to attend the annual meeting of the Board but shall have no vote
and shall not be counted for purposes of ascertaining the presence of a quorum.
ARTICLE II
Officers oj the Board
1. The officers of the Board shall be a Chairman of the Board, a Vice-Chairman, and
a Secretary, who shall be elected by the Trustees, from the members of the Board, by
ballot to serve for a term of three years. All vacancies shall be filled by the Board for
the unexpired term; provided, however, that the Executive Committee shall have power
to fill a vacancy in the office of Secretary to serve until the next meeting of the Board of
Trustees.
2. The Chairman shall preside at all meetings and shall have the usual powers of a
presiding officer.
3. The Vice-Chairman, in the absence or disability of the Chairman, shall perform the
duties of the Chairman.
761
762 CARNEGIE INSTITUTION
4. The Secretary shall issue notices of meetings of the Board, record its transactions,
and conduct that part of the correspondence relating to the Board and to his duties.
ARTICLE III
Executive Administration
The President
1. There shall be a President who shall be elected by ballot by, and hold office during
the pleasure of, the Board, who shall be the chief executive officer of the Institution.
The President, subject to the control of the Board and the Executive Committee, shall
have general charge of all matters of administration and supervision of all arrangements
for research and other work undertaken by the Institution or with its funds. He shall
prepare and submit to the Board of Trustees and to the Executive Committee plans
and suggestions for the work of the Institution, shall conduct its general correspondence
and the correspondence with applicants for grants and with the special advisers of the
Committee, and shall present his recommendations in each case to the Executive Com-
mittee for decision. All proposals and requests for grants shall be referred to the President
for consideration and report. He shall have power to remove, appoint, and, within the
scope of funds made available by the Trustees, provide for compensation of subordinate
employees and to fix the compensation of such employees within the limits of a maximum
rate of compensation to be established from time to time by the Executive Committee.
He shall be ex officio a member of the Executive Committee.
2. He shall be the legal custodian of the seal and of all property of the Institution
whose custody is not otherwise provided for. He shall sign and execute on behalf of
the corporation all contracts and instruments necessary in authorized administrative and
research matters and affix the corporate seal thereto when necessary, and may delegate
the performance of such acts and other administrative duties in his absence to the
Executive Officer. He may execute all other contracts, deeds, and instruments on behalf
of the corporation and affix the seal thereto when expressly authorized by the Board of
Trustees or Executive Committee. He may, within the limits of his own authorization,
delegate to the Executive Officer authority to act as custodian of and affix the corporate
seal. He shall be responsible for the expenditure and disbursement of all funds of the
Institution in accordance with the directions of the Board and of the Executive Com-
mittee, and shall keep accurate accounts of all receipts and disbursements. Following
approval by the Executive Committee he shall transmit to the Board of Trustees before
its annual meeting a written report of the operations and business of the Institution
for the preceding fiscal year with his recommendations for work and appropriations for
the succeeding fiscal year.
3. He shall attend all meetings of the Board of Trustees.
4. There shall be an officer designated Executive Officer who shall be appointed by
and hold office at the pleasure of the President, subject to the approval of the Executive
Committee. His duties shall be to assist and act for the President as the latter may duly
authorize and direct.
5. The President shall retire from office at the end of the fiscal year in which he becomes
sixty-five years of age.
ARTICLE IV
Meetings and Voting
1. The annual meeting of the Board of Trustees shall be held in the City of Washington,
in the District of Columbia, in May of each year on a date fixed by the Executive
Committee, or at such other time or such other place as may be designated by the
Executive Committee, or if not so designated prior to May 1 of such year, by the
BY-LAWS 763
Chairman of the Board of Trustees, or if he is absent or is unable or refuses to act, by
any Trustee with the written consent of the majority of the Trustees then holding office.
2. Special meetings of the Board of Trustees may be called, and the time and place
of meeting designated, by the Chairman, or by the Executive Committee, or by any
Trustee with the written consent of the majority of the Trustees then holding office.
Upon the written request of seven members of the Board, the Chairman shall call a
special meeting.
3. Notices of meetings shall be given ten days prior to the date thereof. Notice may
be given to any Trustee personally, or by mail or by telegram sent to the usual address
of such Trustee. Notices of adjourned meetings need not be given except when the
adjournment is for ten days or more.
4. The presence of a majority of the Trustees holding office shall constitute a quorum
for the transaction of business at any meeting. An act of the majority of the Trustees
present at a meeting at which a quorum is present shall be the act of the Board except
as otherwise provided in these By-Laws. If, at a duly called meeting, less than a quorum
is present, a majority of those present may adjourn the meeting from time to time
until a quorum is present. Trustees present at a duly called or held meeting at which a
quorum is present may continue to do business until adjournment notwithstanding the
withdrawal of enough Trustees to leave less than a quorum.
5. The transactions of any meeting, however called and noticed, shall be as valid as
though carried out at a meeting duly held after regular call and notice, if a quorum is
present and if, either before or after the meeting, each of the Trustees not present
in person signs a written waiver of notice, or consent to the holding of such meeting, or
approval of the minutes thereof. All such waivers, consents, or approvals shall be filed
with the corporate records or made a part of the minutes of the meeting.
6. Any action which, under law or these By-Laws, is authorized to be taken at a
meeting of the Board of Trustees may be taken without a meeting if authorized in a
document or documents in writing signed by all the Trustees then holding office and
filed with the Secretary.
7. During an emergency period the term "Trustees holding office" shall, for purposes
of this Article, mean the surviving members of the Board who have not been rendered
incapable of acting for any reason including difficulty of transportation to a place of
meeting or of communication with other surviving members of the Board.
article v
Committees
1. There shall be the following Standing Committees, viz. an Executive Committee,
a Finance Committee, an Auditing Committee, a Nominating Committee, and a Retire-
ment Committee.
2. All vacancies in the Standing Committees shall be filled by the Board of Trustees
at the next annual meeting of the Board and may be filled at a special meeting of the
Board. A vacancy in the Executive Committee and, upon request of the remaining
members of any other Standing Committee, a vacancy in such other Committee may be
filled by the Executive Committee by temporary appointment to serve until the next
meeting of the Board.
3. The terms of all officers and of all members of Committees, as provided for herein,
shall continue until their successors are elected or appointed.
Executive Committee
4. The Executive Committee shall consist of the Chairman, Vice-Chairman, and
Secretary of the Board of Trustees, the President of the Institution ex officio, and, in
addition, not less than five or more than eight Trustees to be elected by the Board by
ballot for a term of three years, who shall be eligible for re-election. Any member elected
764 CARNEGIE INSTITUTION
to fill a vacancy shall serve for the remainder of his predecessor's term. The presence of
four members of the Committee shall constitute a quorum for the transaction of business
at any meeting.
5. The Executive Committee shall, when the Board is not in session and has not given
specific directions, have general control of the administration of the affairs of the
corporation and general supervision of all arrangements for administration, research,
and other matters undertaken or promoted by the Institution. It shall also submit to the
Board of Trustees a printed or typewritten report of each of its meetings, and at the
annual meeting shall submit to the Board a report for publication.
6. The Executive Committee shall have power to authorize the purchase, sale, exchange,
or transfer of real estate.
Finance Committee
7. The Finance Committee shall consist of not less than five and not more than six
members to be elected by the Board of Trustees by ballot for a term of three years, who
shall be eligible for re-election. The presence of three members of the Committee shall
constitute a quorum for the transaction of business at any meeting.
8. The Finance Committee shall have custody of the securities of the corporation
and general charge of its investments and invested funds, including its investments and
invested funds as trustee of any retirement plan for the Institution's staff members and
employees, and shall care for and dispose of the same subject to the directions of the
Board of Trustees. It shall have power to authorize the purchase, sale, exchange, or
transfer of securities and to delegate this power. It shall consider and recommend to the
Board from time to time such measures as in its opinion will promote the financial interests
of the Institution and of the trust fund under any retirement plan for the Institution's
staff members and employees, and shall make a report at each meeting of the Board.
Auditing Committee
9. The Auditing Committee shall consist of three members to be elected by the Board
of Trustees by ballot for a term of three years.
10. Before each annual meeting of the Board of Trustees, the Auditing Committee
shall cause the accounts of the Institution for the preceding fiscal year to be audited
by public accountants. The accountants shall report to the Committee, and the Com-
mittee shall present said report at the ensuing annual meeting of the Board with such
recommendations as the Committee may deem appropriate.
Nominating Committee
11. The Nominating Committee shall consist of the Chairman of the Board of
Trustees ex officio and, in addition, three Trustees to be elected by the Board by ballot
for a term of three years, who shall not be eligible for re-election until after the lapse
of one year. Any member elected to fill a vacancy shall serve for the remainder of his
predecessor's term, provided that of the Nominating Committee first elected after
adoption of this By-Law one member shall serve for one year, one member shall serve
for two years, and one member shall serve for three years, the Committee to determine
the respective terms by lot.
12. Sixty days prior to an annual meeting of the Board the Nominating Committee
shall notify the Trustees by mail of the vacancies to be filled in membership of the Board.
Each Trustee may submit nominations for such vacancies. Nominations so submitted
shall be considered by the Nominating Committee, and ten days prior to the annual
meeting the Nominating Committee shall submit to members of the Board by mail a list
of the persons so nominated, with its recommendations for filling existing vacancies on
the Board and its Standing Committees. No other nominations shall be received by the
Board at the annual meeting except with the unanimous consent of the Trustees present.
BY-LAWS 755
Retirement Committee
13. The Retirement Committee shall consist of three members to be elected by the
Board of Trustees by ballot for a term of three years, who shall be eligible for re-election
and the Chairman of the Finance Committee ex officio. Any member elected to fill a
vacancy shall serve for the remainder of his predecessor's term.
14. The Retirement Committee shall, subject to the directions of the Board of Trustees,
be responsible for the maintenance of a retirement plan for staff members and
employees of the Institution and act for the Institution in its capacity as trustee
under any such plan, except that any matter relating to investments under any such
plan shall be the responsibility of the Finance Committee subject to the directions
of the Board of Trustees. The Committee shall submit a report to the Board at the
annual meeting of the Board.
ARTICLE VI
Financial Administration
1. No expenditure shall be authorized or made except in pursuance of a previous
appropriation by the Board of Trustees, or as provided in Article V, paragraph 8, hereof.
2. The fiscal year of the Institution shall commence on the first day of July in each
year.
3. The Executive Committee shall submit to the annual meeting of the Board a full
statement of the finances and work of the Institution for the preceding fiscal year and a
detailed estimate of the expenditures of the succeeding fiscal year.
4. The Board of Trustees, at the annual meeting in each year, shall make general
appropriations for the ensuing fiscal year; but nothing contained herein shall prevent
the Board of Trustees from making special appropriations at any meeting.
5. The Executive Committee shall have general charge and control of all appropria-
tions made by the Board. Following the annual meeting, the Executive Committee may
allocate these appropriations for the succeeding fiscal year. The Committee shall have full
authority to reallocate available funds, as needed, and to transfer balances.
6. The securities of the Institution and evidences of property, and funds invested and
to be invested, shall be deposited in such safe depository or in the custody of such trust
company and under such safeguards as the Finance Committee shall designate, subject
to directions of the Board of Trustees. Income of the Institution available for expenditure
shall be deposited in such banks or depositories as may from time to time be designated
by the Executive Committee.
7. Any trust company entrusted with the custody of securities by the Finance Com-
mittee may, by resolution of the Board of Trustees, be made Fiscal Agent of the
Institution, upon an agreed compensation, for the transaction of the business coming
within the authority of the Finance Committee.
8. The property of the Institution is irrevocably dedicated to charitable purposes,
and in the event of dissolution its property shall be used for and distributed to those
charitable purposes as are specified by the Congress of the United States in the Articles of
Incorporation, Public Law No. 260, approved April 28, 1904, as the same may be
amended from time to time.
ARTICLE VII
Amendment of By-Laws
1. These By-Laws may be amended at any annual or special meeting of the Board of
Trustees by a two-thirds vote of the members present, provided written notice of the
proposed amendment shall have been served personally upon, or mailed to the usual
address of, each member of the Board twenty days prior to the meeting.
Index of Names
Aaronson, Marc, 37, 162
Abelson, Philip H., v, 15, 682, 690, 694, 741
publications, 735, 736
Report of the President, 1-38
Abercrombie, M., 89
Abu-Eid, R. M., 423, 734
studies, 564-567
Ackerman, Edward A., vii, 32, 742
publications, 736
Adams, J. B., 734
Adelman, Saul J., 138
publications, 156, 159
Adkins, John M., 162
Adkinson, Bruce, 156, 164
Agassiz, Alexander, vi, 757, 758
Ahrens, T. J., 558, 561, 711
publications, 717
Aitken, Yvonne, 326
Akella, J., 36, 421, 425, 426, 523, 526, 531, 532,
623, 629, 636, 716, 733
publications, 717
studies, 523-526, 623-629
Akimoto, S., 265, 275, 308, 309
publications, 312
Albright, Joseph F., 36, 93
Alcaino, Gonzalo, 149, 155
Aldrich, L. Thomas, 30, 38, 171, 224, 247, 305,
310, 314
publications, 312
studies, 249-252
Allen, David A., 25, 109
publications, 156, 158, 159, 160
Aller, L. H, 141
Amesz, Jan, 326
Andersen, C. A., 224, 310, 315
studies, 280-283
Anderson, D. E., 711
Anderson, J. M.,326
Andre, M., 326
Angel, J. R. P., 138
Aparicio, P., 312
Arnold, L. G., 315
Arp, Halton C, 37, 123, 125, 126, 161
publications, 156, 157, 158
Ashby, Eric, v
Assousa, George E., 38, 173, 179, 314
publications, 312
studies, 181-182, 181-184, 184, 186-190
Ayala, A., 26, 411
Baad, M., 411
Babcock, H. D., 225
Babcock, Horace W., 8, 37, 120, 155, 161, 178
report of the Director, Hale Observatories,
101-164
publications, 156
Bagnuolo, William G., 25, 37,121, 123, 162
publications, 156, 160
Baldwin, George J., vi
Balick, B, 306, 316
studies, 182-184, 184
Bannister, John, 224, 225, 315
studies, 249-252
Banno, S., 713
Baragar, W. R. A., 678, 713
publications, 723
Barbour, Thomas, vi
Barker, D. S., 423, 478
studies, 578-581
Barkley, D. S., 91, 92
Barton, J. M., 611
Baxter, F., 84, 85
Beach, L., 224, 317
studies, 225-226, 249-252
Becklin, Eric E., 10, 25, 37, 103, 113, 114, 127, 162
publications, 155, 156, 157, 158, 159, 160
Bedke, J. R., 153, 163
Bell, James F., vi
Bell, P. M., 36, 423, 426, 427, 527, 552, 555, 558,
562, 563, 574, 575, 576, 629, 630, 631, 643, 644,
658, 659, 662, 664, 665, 711, 713, 714, 716, 733
publications, 716, 717, 718, 724, 726
studies, 574-576, 629-631, 631-638, 638-643,
643-644, 656-662, 662-665, 706-708
Bell, R. A., 146
Benjamin, J., 91
Berrocal, Jesus A., 38
Berry, Joseph A., 37, 325, 336, 410, 411, 412, 414
studies, 393-403, 403-405, 405-407
Billings, John S., vi, 757, 758, 759
Birchenall, C. E., 711
Biroc, Sandra L., 36, 94
Bjorkman, Olle, 23, 26, 37, 322, 325, 403, 410,
411,412,414
studies, 393^03
Bjorn, L.-O., 326
Blackler, Antonie W., 31, 32, 35, 91, 94
Bliss, Robert Woods, vi
Bohm, K.-H., 138
Boise, James W., vii
Bolton, Ellis T., 16, 88, 194, 306, 314
report of the Director, Department of Terres-
trial Magnetism, 165-318
Bonner, Tom I., 88, 184, 315
publications, 312
studies, 204-207
767
768
CARNEGIE INSTITUTION
Borra, Ermanno F., 25,37,111, 112, 136, 162
Boving, Bent G., 93
Bowen, Ira S., 37, 157, 160, 161
Bowen, N. L, 420, 421, 457, 503, 504, 506, 507,
538, 709
publications, 718, 730
Bowes, George, 38, 325, 326, 411, 412, 414
studies, 403-405, 405-407
Boyd, Francis R., Jr., 12, 13, 36, 224, 266, 270,
309, 316, 418, 419, 421, 425, 426, 431, 432, 433,
436, 442, 446, 447, 449, 458, 468, 472, 473, 475,
496, 523, 524, 525, 526, 528, 530, 531, 532, 533,
623, 647, 649, 650, 706, 712, 713, 714, 716, 733
publications, 716, 717, 718, 719, 720
studies, 272-276, 431-445, 523-526, 623-629
Boyden, John E., 162
Boynton, John E., 326
Bradford, Amory H., vi, 739
Bradford, Lindsay, vi
Bradley, Montgomery S., vii
Bradley, Omar N., vi
Brady, J., 711
Brand, L., 89
Branscomb, Lewis M., v, 33, 741
Briggs, Winslow R., 22, 26, 33, 34, 321, 325, 410
Bril, C, 326
Britten, Roy J., 18, 38, 221, 306, 314
studies, 222, 222-223, 223, 224
Brookings, Robert S., vi
Brooks, C, 224, 316
publications, 312
studies, 252-259
Brown, Donald B.,21, 34, 36, 8, 9, 10, 11, 20, 25,
32, 91, 92, 93
publications, 91
studies, 10-14
Brown, Jeanette S., 24, 37, 323, 325, 411, 412, 414
studies, 351-359, 359-361, 407-410
Brown, Louis, 17, 38, 170, 173, 304, 314
publications, 312
studies, 186-190, 190-194
Brownlee, George, 91, 94
Brucato, Robert J., 10, 11, 37, 103, 116, 117, 120,
127, 154, 155, 162
publications, 157, 158, 160, 161
Brueckel, Frank J., 162
publications, 156, 157
Buckley, G. R, 712
Burdin, David, 27
Burger, M., 89
Burns, R. G., 423
studies, 564-567
Burrhus, K. D., 224, 317
Burt, D. M., 36, 711, 716, 733
publications, 716
Bush, Vannevar, vi, 325, 418
Butler, W. L., 365
Byers, M. J., 315
studies, 217
Cader, Andrew, 95
Cadwalader, John L., vi, 757, 758
Calvin, M., 322
Campbell, William W., vi
Carrasco, Luis, 144
Carroll, Dana, 36, 9, 10, 15, 94
studies, 15
Carty, John J., vi
Casaverde, Mateo, 224, 225, 315
studies, 249-252
Catanzaro, B., 393
Chai, B. H. T., 711
Chamberlin, M.E.,221
Champness, P. E., 535, 540, 652
publications, 719, 720, 725
Chase, John W., 94
Chayes, F„ 36, 428, 429, 430, 669, 673, 674, 675,
678, 681, 682, 716, 733
publications, 716, 719
studies, 668-671, 671-673, 673-675, 681-682,
699-701
Cheung, C, 304
Christensen, Clark G., 139, 162
Church, W. R., 713
Clardy, Ken W., 162
Claude, P., 89
Clausen, Jens C, 322, 325
Clayton, R. N., 711
Clube, S. V. M., 139
Cochran, D. B., 221
Cohen, G. N., 18, 171, 306, 307, 316
studies, 207-214
Cohen, Judith G., 139
Cole, Whitefoord R., vi
Coleman, A. W., 91
Colgate, S., 185
Collett, C. F., 221
Colomb, R., 185, 186, 316
Comaford, D., 224, 316
studies, 280-283
Cone, R. A., 89
Conti, Peter S., 139, 140
publications, 157, 160
Coon, Hayden G., 20, 8, 39, 40, 94
studies, 40-42
Coppens, P., 712
Cowie, Dean B., 18, 38, 171, 194, 307, 314
studies, 207-214
Cragg, Thomas A., 162
publications, 157
Crawford, M. L., 224, 310, 316
studies, 290-293
Cruvellier, P., 116, 155
Cuatrecasas, P., 89
Cudworth, Kyle, 140
Culbertson, Jan C, 95
Currie, D. G., 140
Cuyubamba, Arturo, 38, 224, 315
studies, 252-259
Czyzak, S. J., 141
INDEX OF NAMES
769
Danziger, I. J., 151
Davidson, E. H., 206, 221, 306, 316
studies, 222-223
Davidson, N., 24, 45, 91
Davis, B. T. C, 266, 309
Davis, G. L., 15, 36, 224, 289, 295, 299, 310, 311,
312, 316, 424, 602, 605, 733
studies, 601-610, 610-613
Davis, Michael B., 88, 194, 315
studies, 217-221
Dawid, Igor B., 20, 36, 8, 10, 38, 39, 40, 45, 46,
91,93
publications, 91
studies, 38-39, 40-42, 45-47
DeHaan, Robert L., 19, 86, 10, 73, 77, 79, 80, 81,
83,85,89,91,92,93
publications, 91
de Kouchkovsky, Yaroslav, 326
Delano, Frederic A., vi
del Pozo, S., 224, 316
studies, 249-252
Dennison, Edwin W., 11, 12, 87, 154, 155, 161
publications, 157
DeRemer, Sherry, 38
Detchev, Georgi, 37, 326, 414
Devreotes, Peter, 36, 94
Dickens, B., 712
Dickens, R. J., 141
Dickey, J. S., Jr., 36, 734
Dickinson, W. R., 255, 257, 258, 259, 308
Doak, J. B., 178, 317
Dobzhansky, Theodosius, 26, 411
Dodge, Cleveland H., vi, 757, 758
Dodge, F. W., 259
Dodge, William E., vi
Doyle, F., 85
Drachman, Daniel B., 19, 9, 50, 56, 94
studies, 56-60
Dravins, Dainis, 141
publications, 157
Duncan, Douglas, 112
Dunham, P., 89
Durrant, C. J., 109
Durr, Thomas E., 36, 73, 85, 94
Dworetsky, Michael M., 87, 110, 112, 162
publications, 157, 159
Ebert, James D., 18, 19, 21, 27, 34, 86, 9, 93
publications, 91, 92
studies, 64-69
report of the Director, Department of Em-
bryology, 3-95
Ecklund, E. T., 173, 317
studies, 179-180
Edberg, S. J., 104
publications, 157, 158
Edds, M. V., Jr., 91, 92
Edelman, G. M., 47, 89
Edidin, Michael, 89, 94
publications, 92
Eggler, D. H, 36, 419, 420, 446, 456, 457, 459, 463,
470, 472, 475, 476, 477, 496, 498, 709, 712, 713,
714, 716, 733
publications, 721, 726
studies, 446-449, 457-467, 491-495
Elias, Jay, 11,103, 115
Emerson, Robert, 321, 323
Emslie, R. F., 611, 721
England, J. L., 458, 468, 496, 523, 526, 530, 706
publications, 719, 720
Erickson, W. C, 316
Erkes, J. W., 316
studies, 182-184
Erlank, A. J, 316
Ermanovics, I., 316, 734
Erskain, M., 403
Esposito, P., 317
studies, 197-200, 200-204
Evertson, D. W., 224, 225, 308, 316
Fair, Eugene, 11, 154, 163
Fambrough, Douglas M., 19, 36, 9, 10, 80, 89, 93
publications, 92
studies, 50-55, 55-56, 56-60, 63-64
Felsenfeld, G., 91
Fenner, Charles P., vi
Ference, Michael, Jr., v, 741
Ferguson, Homer L., vi
Fewson, C. A., 326
Filloy, Emilio, 185, 315
Finger, L. W, 14, 36, 422, 423, 429, 446, 474, 503,
530, 540, 544, 547, 550, 552, 570, 578, 626, 652,
653, 695, 699, 709, 714, 733
publications, 716, 717, 728
studies, 544-547, 547-551, 569-573, 694-699
Fink, L. K., Jr., 425, 617, 620, 622
publications, 721
studies, 616-623
Finkelstein, A., 89
Fisher, G. W., 711
Flemmings, Ernestine, 7, 94
Fletcher, R. C, 224, 310, 316, 711
publications, 312
studies, 276-280, 283-285
Flexner, J. B., 195, 197, 306, 316
Flexner, Louis B., 93, 195, 197, 306, 316
Flexner, Simon, vi
Flippen, J. L., 712
Flores, A., 224, 316
Foland, K. A., 711
Forbes, W. Cameron, vi
Forbush, S. E., 224, 316
Ford, Glenn A., 324, 414
studies, 361-364
Ford, W. Kent, Jr., 17, 88, 169, 306, 314
publications, 312
studies, 173-178
Fork, David C, 22, 24, 37, 322, 324, 325, 411, 412,
413,414
studies, 368-371, 376-384, 384-388
770
CARNEGIE INSTITUTION
Forrestal, James, vi
Frantz, J. D., 36, 420, 430, 704, 711, 714, 734
studies, 507-515, 704-706
French, C. Stacy, 22, 23, 31, 32, 33, 34, 37, 324,
325, 329, 335, 336, 411, 412, 413, 414
report of the Director, Department of Plant
Biology, 319-414
studies, 336-351, 351-359, 361-364, 365-368
Frew, William N., vi, 757, 758
Frey, F. A., 711
Frogel, Jay A., 146, 150, 151, 155
publications, 157
Fugita, Yoshio, 141
Fulpius, B., 51
Gage, L. P., 92, 93
Gage, Lyman J., vi, 757, 758
Galau, G. A., 221
Gardner, Ernest, 10, 93
publications, 92
studies, 87-89
Garzoli, S., 185, 316
Gasanov, Ralphreed A., 24, 37, 323, 326, 412, 414
studies, 351-359
Gates, David M., 326
Gatley, Ian, 11, 37, 103, 115, 162
Gauhl, E. W., 326
Gehrels, Tom, 142
Gelinas, R., 91
Gershfeld, N. L., 49
publications, 92
Gettrust, J., 316
Gibbs, M., 326
Giesecke, A. A., 224, 316
Gifford, Walter S., vi
Gil, M, 185, 186, 304, 316
Gilbert, Carl J., v, 741
Gilbert, Cass, vi
Gilbert, Scott, 36, 94
Giletti, B. J., 224, 298, 304, 310, 311, 316, 616,
710, 711
studies, 280-283
Gillett, Frederick H., vi
Gilman, Daniel C, vi, 757, 758, 759
Gilula, N„ 89
Goad, J., 304, 316
Goetz, Alexander F. H., 142
Golden, William T., v, 739, 741
Goldberg, R. B., 221
Goldsmith, D., 149
Goldsmith, J. R„ 711
Goniadzki, D., 185, 186, 316
Goodman, M., 312, 316
Gordon, M., 185, 304, 316
Gorshkov, G. S., 259
Govindjee, R., 326
Graham, D. E., 221
studies, 222, 222-223, 223, 224
Graham, J. A., 316
Grauert, Borwin, 38, 224, 311, 315
publications, 312, 313, 588
studies, 285-288, 288-290, 290-293, 293-297,
297-299, 299-302, 302-303
Green, D. H., 266, 267, 272, 309, 716
Green, H. W., II, 711
Green, Richard, 37, 127, 162
Greene, T. F., 148
Greenewalt, Crawford H., v
Greenstein, Jesse L., 10, 37, 108, 109, 117, 119,
140, 162
publications, 157, 158, 159, 160
Gregory, Thomas S., 162
Griffin, R. F., 142
Griffin, W. L., 263, 272, 274, 308, 309, 712
Groshong, R. EL, Jr., 284, 285, 310, 316
publications, 313
Gross, K., 91
Guest, J. E., 257, 258, 308
Guit, B., 304
Gunn, B, 256
Gunn, James E., 37, 123, 127, 128, 133, 135, 162
publications, 157, 158, 160
Hadidiacos, C. G., 429, 446, 474, 503, 530, 695,
699, 709, 734
studies, 694-699
Haefele, Douglas M., 38, 315
Hafner, S. S., 540, 542, 552, 634, 637, 653, 654,
656
publications, 719, 722, 730, 732
Hagar, William G., 26, 38, 324, 326, 414
studies, 361-364
Haggerty, S. E, 716, 734
Hall, A. E., 326
Hall, Harvey Monroe, 322
Hall, L., 224, 311, 316
studies, 293-297
Hallberg, R., 91
Halldal, P., 326
Hand, Cadet, 403
Handler, R., 25
Hanny, R., 311, 313
Haraburda, Joseph M. S., vii
Hardy, Eduardo, 25, 37, 132, 133, 149, 155, 162
Hare, P. E., 15, 36, 429, 430, 682, 690, 692, 694,
701, 704, 715, 733
studies, 690-694, 701-704, 704-706
Harris, J. W. S., 94
Harrison, Robert, 26
Hart, Michael H., 25, 37, 105, 162
Hart, Stanley R., 16, 17, 35, 38, 72, 224, 225, 270,
305, 310, 314, 585, 589, 600, 618, 711
publications, 313, 730
studies, 259-262, 268-270
Hartwick, F. D. A., 142
Hartzell, H. Criss, 36, 10, 50, 80, 94
publications, 92
studies, 55-56
Harvey, Paul M., 11, 37, 103, 115, 163
publications, 156, 158, 160
INDEX OF NAMES
771
Haskins, Caryl P., v, vi, 739, 741
Hay, John, vi, 757, 758, 759
Hay, Robert J., 94
Hays, J. F., 734
publications, 724
studies, 706-708
Heber, U. W., 326
Hedges, J. I., 36, 734
Helgeson, H. C, 711
Helin, Eleanor, 147
Heller, C, 27, 411
Henard, Kenneth R., vii
Henery, M., 221
studies, 222-223
Hendler, R., 91
Hennig, W., 91
Henry, Barklie McKee, vi
Hensen, B. J., 36, 421, 442, 446, 475, 515, 523,
715, 716, 734
publications, 723
studies, 527-534
Herbig, G. H., Ill
Herrick, Myron T., vi
Hertig, Arthur T., 93
Hershey, Alfred D., vii
Heuer, A. H., 711
Hewitt, Abram S., vi
Hewlett, William R., v, 741
Heydenburg, N. P., 170, 186, 190, 313
Hibberson, W., 266, 267, 309
Hickson, Paul, 37, 163
Hiesey, William M., 37, 322, 325, 414
Hietanen, A., 297, 301, 311
Higgins, N. W., 285, 290, 310
Higginson, Henry L., vi, 757, 758
Hille, Bertil, 79
Hinthorne, J. R., 224, 310, 316
studies, 280-283
Hitchcock, Ethan A., vi, 757, 758
Hitchcock, Henry, vi
Hiyama, Tetsuo, 22, 38, 322, 326, 411, 413, 414
studies, 368-371, 371-373, 384-388
Hodges, F. N., 36, 420, 423, 427, 465, 578, 645,
646, 713, 716, 734
studies, 495-497, 578-581, 645-646, 646-647
Hoering, T. C, 15, 36, 429, 713, 715, 733
studies, 682-690, 690-694
Hofmann, Albrecht W., 38, 224, 290, 304, 310,
311, 314, 710, 711
publications, 723
studies, 276-280, 280-283, 297-299, 299-302
Holland, H. D., 714
Hollinger, T., 91
Horn, John, 26
Honjo, Tasuku, 36, 10, 94
publications, 92
studies, 25-26, 31-35
Hoover, Herbert, vi
Hopson, C. A., 285, 288, 289, 290, 310, 311, 425,
613
studies, 616-623
Horak, Ivan, 20, 28, 36, 8, 39, 94
studies, 40-42
Hornblower, Marshall, vii
Howard, Robert F., 25, 37, 104, 105, 162
publications, 157, 158, 159
Howe, William Wirt, vi, 757, 758
Howell, James A., 37, 163
Hoyer, Bill H., 38, 194, 305, 306, 307, 314
publications, 313
studies, 214-217
Huaco, D., 316
Huchra, John P., 37, 118, 120, 127, 163
publications, 158, 160
Huckenholz, H. G., 634, 636, 637, 714
publications, 722
Humphreys, Roberta M., 150, 156
Hutchinson, Charles L., vi, 757, 758
Hyland, Harry, 25
Igiri, Godwin C, 38, 315
Iiyama, J. T., 711
Irvine, T. N., 36, 419, 420, 428, 442, 478, 483,
488, 489, 678, 713, 715, 716, 733
publications, 723
studies, 478-491, 675-681
Isacks, B. L., 233, 238, 307
Isacks, I., 240, 241
Jacobi, G., 326
Jacobson, M., 89
Jakobsson, S. P., 424, 597
publications, 724
studies, 597-598
James, David E, 38, 224, 225, 247, 258, 304, 305,
308, 314
publications, 313
studies, 252-259
Jessup, Walter A., vi
Jewett, Frank B., vi
Johnson, Q., 712
Johnson, R. T., 89
Johnson, T. V, 143
Jones, Burton, 140
Jones, D. S., 712
Jordan, E., 21, 7, 8, 11
Joseph, N., 50
Joy, Alfred H., 32, 161, 162
Juteau, T., 713
Katem, Basil N, 127, 162
publications, 159
Katz, Margaret, 162
Kaushagen, Carol, 36, 10, 39, 94,
studies, 43-44
Keck, D. C, 322
Keenan, Philip C, 143
King, K., Jr., 36, 734
772
CARNEGIE INSTITUTION
Kirshner, Robert P., 25, 37, 103, 118, 163
publications, 157, 158, 160
Klein, Andrew Scott, 95
Klein, W. H., 221
Klitzman, Jack M., 38, 315, 317
Knapp, S. L., 140
Knox, Grace V., 162
Kohne, D. E, 306, 313
Kolzak, Nancy, 95
Konigsberg, Irwin R., 93
Konnert, J., 712
Koo, David C, 143, 144
Kormendy, John, 37, 126, 163
publications, 156, 158
Kowal, Charles, 103, 118, 162
publications, 158, 159, 160
Kraft, R. J. E., 144
Kristian, Jerome, 9, 10, 11, 26, 37, 102, 103, 116,
117, 127, 128, 129, 133, 137, 154, 162
publications, 157, 158, 160
Krogh, T. E., 15, 36, 224, 299, 316, 424, 425, 435,
602, 605, 607, 609, 613, 711, 713, 715, 733
publications, 313, 716, 724,
studies, 601-610, 610-613
Kron, Richard G., 143, 144
Krueger, Caroline R., 20, 36, 73, 85, 94
Kumar, Cidambi K., 38, 306, 315
publications, 313
studies, 173-178
Kundig, W., 89
Kuno, H., 263, 264, 267, 308
Kiintzel, Hans, 10
Kushiro, I., 36, 224, 226, 267, 272, 273, 309, 316,
420, 427, 430, 444, 449, 451, 452, 453, 454,
465, 471, 476, 477, 490, 493, 495, 496, 498,
499, 500, 501, 502, 523, 527, 533, 534, 537,
539, 540, 589, 591, 627, 645, 646, 648, 649,
650, 708, 711, 713, 715, 716
publications, 716, 717, 725, 727, 733
studies, 270-272, 497-502, 502-507, 645-646,
646-647, 647-650, 708-710
Labeyrie, A., 144
Lambert, R. St. J., 712
Langley, Samuel P., vi, 757, 758
Lanning, Howard H., 144, 162
publications, 158
Larsen, E. S., 259
Lawrence, Ernest O., vi
Lawrence, Harry, 336
Lazareff, Bernard, 106
Leder, P., 10, 91
Leffler, Allan T., II, 38, 315
Leighton, Robert B., 37, 162
publications, 158, 159
Leister, Dennis E., 36, 10, 39, 94
studies, 42-43
Lepler, Elliot C, 37, 163
Lewis, B. T. R., 316
Liewer, K. M., 140
Lindbergh, Charles A., vi
Linde, Alan T., 38, 224, 225, 247, 308, 314
publications, 313
studies, 245-246
Lindsay, William, vi, 757, 758
Ling, V., 89
Lister, C. B., 316
Little, C. A., 173, 317
studies, 179-180
Liu, J., 713
Lizardi, Paul, 36, 9, 10, 20, 94
studies, 18-19
Lodge, Henry Cabot, vi
Lodish, H., 91
Loer, Steven J., 37, 143, 163
Loewy, A. G., 92
Longair, Malcolm, 127
Loomis, Albert L., v, 741
Loomis, Don, 11, 154
Loos, E. E., 326
Lovett, Robert A., vi
Loving, H., 85
Low, Seth, vi, 757, 758
Lowen, A. Louise, 162
Lowrance, J. L., 175
MacDougal, D. T., 321, 325
MacGregor, I. D., 449, 472, 500, 523, 526, 530,
531
publications, 726
Mackay, Craig D., 144
MacVeagh, Wayne, vi, 757, 758
Madsen, A., 326
Mahoney, J. W., 316
Makovicky, E., 734
Manheim, F. T., 711
Mantai, K. E., 326
Manning, J. R., 711
Mao, H. K., 14, 36, 422, 423, 426, 427, 496, 527,
552, 555, 558, 560, 561, 562, 563, 574, 575, 576,
629, 630, 631, 643, 644, 658, 659, 662, 664, 665,
715, 716, 733, 734
publications, 716, 717, 718, 724, 726
studies, 552-554, 554-557, 557-564, 564-567,
574-576, 629-631, 631-638, 63&-643, 643-644,
656-662, 662-665
Marchalonis, J. J., 91
Martin, M. A., 316
Martin, William McChesney, Jr., v, 739, 741
Matson, D. L., 143
Matthews, Thomas, 8
Mattinson, J. M., 36, 224, 316, 425, 613, 615, 713,
734
publications, 717, 726
studies, 613-616, 616-623
McBride, T. J., 18
McCallister, R. H., 711, 712
McCallum, M. E., 419, 446
studies, 446-449
McClintock, Barbara, vii, 34
McClure, Robert D., 142
McConnell, J. D. C, 712
INDEX OF NAMES
773
McCord, T., 713
McCord, T. B., 144
McDonald, Terence F., 36, 73, 79, 91, 93, 94
publications, 92
studies, 86-87
McGinnies, William G., 321, 326, 414
McGough, Sheila A., vii
McHugh, Keith S., v, 739, 741
McMahon, D., 326
McNutt, R., 734
Medina, E., 326
Meenakshi, V. R., 734
Mellerup, Susan B., 162
Mellon, Andrew W., vi
Melnick, Jorge, 37, 163
Menke, W., 326
Merriam, John Campbell, vi
Meyer, R. P., 307, 313
Michalitsanos, Andrew, 37, 162
Michel, J. M., 326
Michel-Wolwertz, M.-R., 326
Miller, I. R., 92
Miller, Margaret Carnegie, vi
Miller, Roswell, vi
Miller, R. S., 92
Miller, W. C, 12, 152, 153, 163
Milner, H. W., 323, 325
Mills, Darius 0., vi, 757, 758
Minkowski, R., 101, 118, 119
publications, 157, 158
Mirabel, I. F., 185, 186, 313, 316
Misener, D. J, 36, 712, 734
Misenhimer, H. Robert, 93
Mitchell, S. Weir, vi, 757, 758, 759
Mitchell, Walter E., Jr., 145
Molnar, P., 240, 241,307
Montague, Andrew J., vi
Mooney, H., 26, 27, 403, 410
studies, 393^03
Moore/Ronald, 37, 135, 162
Morgan, Henry S., v, 739, 741
Morras, R., 185, 186, 316
Morrow, William W., vi, 757, 758
Morton, D. C, 145
Motoya, Y., 224, 316
Moyzis, Linda, 20, 36, 73, 85, 94
Muan, A., 555, 638, 734
publications, 727
Mudd, Seeley G., vi
Muecke, Edward C, 82, 94
Muehlenbachs, K., 36, 424, 593, 594, 596, 597,
598, 599, 600, 713, 715, 734
publications, 727
studies, 593-597, 597-598, 59&-601
Miiller, A., 326
Munch, Guido, 37, 106, 107, 108, 116, 155, 162
publications, 158
Murata, Norio, 24, 38, 324, 326, 329, 337, 414
studies, 376-384
Murata, Teruyo, 326
studies, 336-351
Murthy, V. R., 263, 272, 308, 309
Myers, William I., v, 741
Mysen, B. 0., 36, 419, 444, 468, 469, 470, 471, 475,
476, 477, 500, 734
studies, 467-478
Nason, A., 91, 92
Nelson, P., 89
Neufeld, B. R., 221,316
studies, 222, 224
Neugebauer, Gerry, 10, 11, 25, 37, 103, 106, 113,
114, 127, 154, 155, 162
publications, 156, 157, 158, 160
Ney, Edward P., 150, 156
Nicholson, Frank, 336, 414
studies, 393-403
Ninnemann, Helga I., 38, 325, 326, 413, 414
studies, 365-368
Nixon, P. H., 12, 13, 270, 272, 309, 418, 419, 431,
433, 436, 438, 443, 446, 449, 475, 524, 525, 531,
532, 533
publications, 719
studies, 431-445
Nobs, Malcolm A., 37, 322, 325, 336, 337, 414
studies, 393-403
Nojima, T., 92, 93, 94
Noonan, Thomas W., 145
Norris. John, 145, 150, 156 .
Norton, Garrison, v, 741
Obradovich, J. D., 298, 301, 311
Ocola, Leonidas, 224, 247, 315
publications, 313
Oemler, Augustus, Jr., 37, 163
publications, 157, 158
Ohashi, Y., 36, 422, 423, 429, 540, 543, 544, 545,
546, 547, 550, 569, 652, 653, 699, 712, 715, 734
publications, 719, 728
studies, 544-547, 547-551, 569-573, 694-699
Oinas, Valdar, 37, 145, 163
Okada, Hiromu, 16, 38, 171, 224, 307, 315
publications, 313
studies, 226-233, 233-238, 238-245
Okazaki, R., 91
Oke, J. Beverley, 10, 11, 12, 37, 103, 117, 118, 119,
133, 137, 145, 154, 162, 175, 306
publications, 158, 160
Olafsson, G., 224, 316
O'Rahilly, Ronan, 35, 10, 87, 88, 89, 93
publications, 92
Orr, C. W., 92
Orton, Glenn S., 37, 163
Orville, P. M., 712
Osborn, E. F., 497, 503, 504, 638
publications, 722, 727
Osborn, William Church, vi
Ozato, Keiko, 36, 94
Ozawa, E., 92
774
CARNEGIE INSTITUTION
Pagano, Richard E., 36, 9, 93
publications, 92
studies, 47-50
Page, E., 91
Page, Walter H., v, 741
Pagel, B. E. J., 145
Papp, R., 26, 411
Pardue, Mary Lou, 14, 94
Parmelee, James, vi
Parsons, Sidney B., 146
Parsons, William Barclay, vi
Partridge, R. B., 317
Pasachoff, Jay M., 37, 162
publications, 158
Paton, Stewart, vi
Paul, J., 91
Pearcy, R. W., 326
Pena, H., 304
Pennoyer, Robert M., v
Periston, M. V., 103, 118, 126
publications, 157, 158, 159
Peoples, J., 190
Pepin, R., 716
Pepper, George W., vi
Perkins; Richard S., v, 741
Perry, Robert, 38
Pershing, John J., vi
Persson, S. Eric, 146, 150, 151, 156
publications, 157, 159
Peterman, Z. E., 255, 298, 301, 311
Peterson, D. M., 109
publications, 157, 159, 160
Philpotts, J. A., 263, 264, 265, 272, 274, 308, 309
Pichler, H., 253, 255, 257, 258, 308
Pickett J. M., 326
Pitelka, Louis, 38
Plitt, Calvin E., 36, 73, 94
Poe, G. R., 247, 317
Pogo, Alexander, 162
Polinger, I. S., 92
Pollard, H., 712
Pollock, Harry E. D., vii
Poppel, W., 185, 186, 304, 317
Popper, Daniel M., 146
Powell, G., 304
Prata, Stephen W., 37, 162
publications, 159
Prather, Michael, 403
Prather, Nancy, 403
Pratt, N. M., 126
Prentis, Henning W., Jr., vi
Preston, George W., 37, 109, 110, 112, 162,
publications, 159
Pritchett, Henry S., vi
Proskouriakoff, Tatiana, vii
Punnett, T., 324
Pyper, Diane M., 147
Quicksall, C., 712
Quirk, William J., 37, 162
publications, 157, 159
Quiroga, R., 185, 317
Quiroga, Ricardo, 38, 224, 315
studies, 249-252
Ragland, P., 713
Ramirez, J. E., 224, 225, 249, 317
Rash, J. E., 92, 93
Raymond, L. P., 326
Raymond, M. G., 36, 423, 547, 577, 578, 715, 734
studies, 576-578
Rebbert, M., 45, 46
Reeder, Ronald H., 36, 10, 25, 26, 28, 31, 92
publications, 93
studies, 25-26, 26-28, 28-31
Rees, Martin, 135
Reich, E., 51
Rentschler, Gordon S., vi
Ribbe, P., 573, 574, 575
publications, 729
Rice, Nancy R., 38, 194, 304, 306, 314
studies, 197-200, 200-204, 214-217
Richter, Paul H., 147
Rico, M., 221
Ried, A., 326
Rigg, J-, 221
Riley, Malcolm S., 162
Ringwood, A. E., 266, 267, 272, 309
Ritchie, Aileen K., 36, 10, 50, 94
studies, 60-63
Roberts, M.S., 174,317
Roberts, Richard B., 38, 194, 305, 314
publications, 313
studies, 195-197
Rockefeller, David, vi
Rodriguez, A., 224, 317
Roeder, Robert, 14, 25, 93, 94
publications, 93
Rohrer, Urs C., 38, 173, 315
publications, 313
studies, 190^194
Root, Elihu, vi, 757, 758, 759
Root, Elihu, Jr., vi
Roseman, S., 89
Rosenquist, Glenn C., 94
Rosenwald, Julius, vi
Ross, M., 421,712
studies, 535-540
Roth, S., 89
Roth, William M., v
Rubey, William W., v, 741
Rubin, S., 317
Rubin, Vera C., 17, 34, 38, 169, 304, 305, 306, 314
publications, 313
studies, 173-178
Ruiz, Maria Teresa, 37, 162
Rule, Bruce, H., 11, 37, 154, 162
Ruysschaert, J. M., 92, 93
Rye, D., 712
Ryerson, Martin A., vi
INDEX OF NAMES
775
Saa, G., 313
Sachs, Howard G., 19, 21, 28, 36 9, 73, 79, 83, 85,
91, 92, 93, 94
publications, 93
studies, 64-69, 86-87
Sacks, I. Selwyn, 16, 38, 171, 190, 224, 225, 247,
306, 307, 308, 314
publications, 313
studies, 226-233, 233-238, 245-246, 246-247
Salgueiro, Reynaldo, 224, 225, 315
studies, 249-252
Sanchez, P., 403
Sandage, Allan R., 8, 9, 11, 25, 37, 101, 102, 108,
119, 120, 127, 128, 129, 131, 132, 133, 137,
142, 154, 162
publications, 159, 160
Sanders, W. L., 147
Santoro, A., 712
Sapatino, V., 221
Sargent, Anneila, 108
Sargent, Wallace L. W., 9, 25, 37, 102, 103, 118.
120, 121, 123, 124, 125, 162
publications, 156, 157, 158, 159, 160
Satyamurti, S., 19, 9, 50, 94
studies, 56-60
Schairer, J. F., 421, 451, 490, 491, 500, 503, 504,
505, 531, 538, 709
publications, 718, 719, 730
Schechter, Paul L., 37, 120, 163
Schectman, Stephen A., 37, 127, 163,
publications, 160
Schiff, J. A., 326
Schilling, J. G, 16, 172, 224, 308, 317
studies, 259-262
Schmidt, C., 27, 411
Schmidt, Maarten, 9, 37, 101, 102, 127, 134, 162
publications, 159, 160
Schneider, A., 91
Schneiderman, H., 92, 93
Schnetzler, C. C., 265, 308, 309
Schreiber, Ulrich, 24, 37, 323, 326, 414
studies, 327-330
Schreyer, W., 714
Schroeder, John R., 410
Schulman, M. D., 326
Schuster, Hans, 153
Schwartz, P. R., 138
publications, 160
Schweizer, Francois, 144
Scuderi, J., 178, 318
Searle, Leonard T., 9, 11, 25, 37, 102, 103, 118,
121, 123, 134, 154, 155, 162
publications, 156, 158, 159, 160
Seitz, M. G., H, 36, 224, 317, 423, 581, 583, 584,
585, 586, 587, 588, 589, 602, 715, 716, 734
publications, 313, 716, 717, 719, 730
studies, 302-303, 581-586, 586-588, 588-593,
666-668
Sestak, Z., 326
Seyler, R. G., 317
publications, 314
Schaffer, D., 127
Shapley, Harlow, 32
Shaw, H. R., 712
Shepley, Henry R., vi
Sheppard, J. R., 89
Sheridan, J., 89
Shieh, Y.-N., 712
Shimizu, Nobumichi, 17, 38, 172, 224, 308, 309,
315
publications, 314
studies, 262-268, 268^270, 270-272, 272-276
Shoemaker, Eugene M., 147
publications, 158, 160
Shoening, William, 153
Shorthill, R. W, 148
Sigtryggsson, H., 224, 317
Silsbury, J. H., 326
Sim, Elizabeth, 153
Simoni, Diglio, 38, 224, 225, 247, 314
Singer, S. J., 89
Sinha, A. K, 314
Sipling, P. J., 712
Sisken, J. E., 94
Smith, B., 35, 64, 69
Smith, George, 11
Smith, J. H. C, 322, 325
Smith, M. J., 221
Smith, Theobald, vi
Smith, T. Jefferson, 224, 307, 315
studies, 249-252
Smith, W. H., 314
Snoke, J. Arthur, 38, 224, 315
studies, 233-238, 245-246
Snyder, G. L., 259
Soeder, C. J., 327
Somerville, D., 7, 35, 64, 69, 95
Soptrajanova, G., 302, 311
publications, 314
Spoehr, H. A., 321, 322, 323, 325
Spooner, John C., vi, 757, 758
Springer, Max, 10, 73, 80, 81, 83, 85, 94
publications, 93
studies, 86-87
Stambrook, Peter J., 21, 36, 9, 93
publications, 93
studies, 35-37, 64-69, 69-72
Stanton, Frank, v, 741
Steiger, R. H., 285, 288, 289, 290, 310
Stein, W. A., 135, 160
Steiner, E., 314
Stern, Ralph H., 36, 10, 15, 94
Stevens, R. K., 713
Stone, G. T., 424, 598, 600, 712, 714, 715, 716,
734
publications, 731
studies, 598-601
Storey, William Benson, vi
Strain, Harold, 322
Straus, Neil A., 38, 306, 314, 315
776
CARNEGIE INSTITUTION
Strauss, F., 186, 317
Strauss, Z., 186, 317
Streeter, George, 87
Strong, Richard P., vi
Sugimoto, Kazunori, 36, 9, 10, 14, 93, 94
Suyehiro, Shigeji, 224, 225, 308, 315
studies, 246-247
Suzuki, E., 36, 20, 23
Suzuki, Yoshiaki, 10, 20, 22, 23, 92, 93, 94
Swan, Larry, 27
Swope, Henrietta H., 162
Taft, Charles P., v
Taft, William H., vi
Takamiya, Atusi, 24, 26, 37, 323, 327, 363, 413,
414
studies, 330-336, 336-351
Tamayo, L., 224, 225, 317
studies, 249-252
Tamiya, Hiroshi, 323
Tamman, G. A., 101, 120
Tanaka, Katsuo, 37, 106, 135, 162
Tartof, Kenneth D., 94
Taylor, H. P., Jr., 276, 309, 712
Taylor, L. A., 426, 643
studies, 638-643
Temmer, Georges, 170, 186
Thayer, William S., vi
Thomas, John, 27, 411
Thompson, G., 314
Thompson, Harold, 403
Thompson, James, 403
Thompson, J. B., Jr., 714
Thompson, R. N., 36, 716, 733
Thompson, T. E., 92, 93
Thompson, W. J., 314
Thonnard, Norbert, 38, 173, 314
studies, 179-180, 180-181, 181-182
Tilley, C. E., 451, 453, 489, 490, 500, 507, 598,
674
publications, 732, 733
Tilton, G. R., 285, 286, 291, 292, 295, 310, 311,
312, 604, 615
publications, 732
Tinsley, Beatrice, 123
Tomblin, J., 317
Townes, Charles H., v, 739
Triep, Enrique, 38, 224, 315
studies, 249-252
Trippe, Juan T., v, 739, 741
Troughton, John H., 38, 327, 410, 414
Truffa-Bachi, P., 18, 171, 306, 307, 317
studies, 207-214
Tsuchida, N., 20
Tucker, John A., 35, 89, 92, 93, 94
Turner, Edwin, 103, 123, 125
Turner, Kenneth C, 30, 38, 169, 173, 179, 185,
186, 314
publications, 314
Turnrose, Barry E., 37, 163
Tuve, M. A., 17, 38, 179, 190, 224, 314
publications, 314
studies, 173-178, 181-182
Ulmer, G. C, 423, 567, 568, 629, 734
publications, 727, 732
studies, 567-569
Urbach, W., 327
van den Bergh, Sidney, 148
van der Kruit, Pieter, 9, 25, 37, 102, 121, 122,
130, 162
publications, 158
van de Velde, Neltje W., 307, 317
publications, 314
studies, 214-217
Van Rinsvelt, H. A., 17, 170, 317
publications, 314
studies, 186-190
Vaughan, Arthur H., Jr., 11, 12, 25, 37, 111, 154,
155, 162
publications, 157, 160
Veeder, Glenn J., 37, 143, 163
Velez, Camilla, 36, 9, 69, 94
Velusamy, Thangasamy, 38, 315
Vidale, R. J., 712
Vidaver, W. E., 323, 327, 329
Vieira, E. R., 186
Virgin, H. I., 26, 327, 412, 413
Virgo, D., 36, 421, 422, 423, 426, 427, 535, 540,
541, 542, 552, 629, 630, 631, 632, 634, 650,
651, 653, 654, 685, 716, 733
publications, 719, 721, 732
studies, 535-540, 540-544, 567,-569, 631-638,
650-656
Visvanathan, N., 119, 123, 129, 162
publications, 159, 160
Vogel, Hans, 153
Volponi, V., 224, 317
Vorpahl, Joan, 148, 149,
publications, 159, 160
Waddington, Bruce, 37, 163
Wadsworth, James W., vi
Wagner, Manfred, 156, 164
Wagner, M. E., 224, 310, 317
studies, 290-293
Walburn, Marjorie H., vii
Walcott, Charles D., vi, 757, 758, 759
Walcott, Frederic C, vi
Walcott, Henry P., vi
Wallace, Donald G., 38, 194, 315
Wallerstein, George, 149
publications, 157, 160
Warner, John W., Ill, 38, 156, 315
studies, 173-178, 184
Wasserburg, G. J., 286, 310, 312
Webster, D. A., 327
Weed, Lewis H., vi
Weill, D. F., 265, 308
INDEX OF NAMES
777
Weisbrod, A., 36, 420, 421, 515, 517, 520, 522,
712, 713, 714, 715, 733
publications, 732
studies, 507-515, 515-518, 518-521, 521-523
Weiss, Charles, 26
Welch, William H., vi
Wellauer, Peter K., 20, 28, 36, 8, 24, 38, 39
Werner, Michael W., 11, 37, 103, 115, 162
West, M. T., 221
Westbrook, William E., 37, 163
Westphal, James A., 11, 26, 37, 127, 133, 137, 162
publications, 157, 158, 160
White, Andrew D., vi, 757, 758
White, Edward D., vi
White, Henry, vi
White, James N., v
White, Richard E., 37, 162
Wickersham, George W., vi
Wickes, Richard A., 37, 163
Wiessner, W., 327
Wiley, Paul F., 18
Williamson, Robert, 36, 9, 10, 19, 24, 94
Willner, Steven P., 37, 163
publications, 158, 160 ,
Wilson, Olin C, 37, 112, 162
publications, 160
Wilson, Robert E., vi, 153
Wimmenauer, W., 714
Wiser, John, 7
Wolf, Hermann K., 89, 94
Wones, D. R., 712
Woodward, Robert S., vi
Wraight, C. A., 327
Wright, Carroll D., vi, 757, 758, 759
Wu, Guang-Jer, 36, 46, 93, 94
Wynne, C. G., 136, 137
Wynn-Williams, C. G., 25, 114, 127
publications, 156, 157, 158, 160
Yoder, H. S., Jr., 12, 28, 34, 36, 419, 450, 451, 453,
454, 457, 458, 465, 475, 477, 489, 490, 491, 493,
495, 496, 498, 500, 503, 504, 505, 507, 520, 521,
522, 523, 534, 535, 552, 586, 593, 595, 598, 710,
712, 713, 714, 715, 716, 733
report of the Director, Geophysical Labora-
tory, 418-734
publications, 717, 725, 728, 730, 733
studies, 449-457
Yoshikawa-Fukada, Masako, 19, 36, 10, 92, 93,
94
Yund, R. A, 710, 712
Zahner, H., 18
Zappala, Robert R., 10, 37, 110, 111, 113, 117,
137, 162
Zies, E. G., 36, 579, 733
publications, 733
Zirin, Harold, 37, 106, 135, 138, 149, 162
publications, 157, 160
Zwicky, Fritz, 35, 118, 126, 162
publications, 158, 161
Live Music Archive
Librivox Free Audio
Metropolitan Museum
Cleveland Museum of Art
Internet Arcade
Console Living Room
Books to Borrow
Open Library
TV News
Understanding 9/11