(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Year book - Carnegie Institution of Washington"

^q t J^-S, >W*&£-fc 



7);^**.^ r** s/ 



■j - ^7^ 



Digitized by the Internet Archive 

in 2012 with funding from 

LYRASIS Members and Sloan Foundation 



http://www.archive.org/details/yearbookcarne67196768carn 



Carnegie 
Institution 

OF WASHINGTON 



Year Book 67 

1967-1968 



Library of Congress Catalog Card Number 3-16716 
Port City Press, Baltimore, Maryland 



Contents 



page 

Officers and Staff v 

Report of the President 1 

Reports of Departments and Special Studies 1 

Mount Wilson and Palomar Observatories 3 

Geophysical Laboratory 71 

Department of Terrestrial Magnetism 281 

Committee on Image Tubes for Telescopes 387 

Department of Embryology 393 

Department of Plant Biology 471 

Genetics Research Unit 553 

Bibliography 569 

Report of the Executive Committee 571 

Report of Auditors 573 

Abstract of Minutes of the Seventieth Meeting of the Board of Trustees 589 

Articles of Incorporation 591 

By-Laws of the Institution 595 

Index 601 



President and Trustees 



PRESIDENT 

Caryl P. Haskins 

BOARD OP TRUSTEES 

James N. White 
Chairman 

Henry S. Morgan 
Vice-Chairman 

Garrison Norton 
Secretary 



Sir Eric Ashby 
Amory H. Bradford 
Omar N. Bradley 1 
Vannevar Bush 
Michael Ference, Jr. 
Carl J. Gilbert 
Crawford H. Greenewalt 
Caryl P. Haskins 
Alfred L. Loomis 
Robert A. Lovett 
William McC. Martin, Jr. 
Keith S. McHugh 
Henry S. Morgan 
Seeley G. Mudd 2 
William I. Myers 
Garrison Norton 
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 



1 Retired May 3, 1968. 

2 Died March 10, 1968. 



Trustees (continued) 



AUDITING COMMITTEE 



Keith S. McHugh, Chairman 
Alfred L. Loomis 
Juan T. Trippe 



EXECUTIVE COMMITTEE 



Henry S. Morgan, Chairman 
Carl J. Gilbert 
Crawford H. Greenewalt 
Caryl P. Haskins 
Keith S. McHugh 
William I. Myers 
Garrison Norton 
Richard S. Perkins 
Frank Stanton 
James N. White 



RETIREMENT COMMITTEE 



Frank Stanton, Chairman 
Amory H. Bradford 
Garrison Norton 
Richard S. Perkins 



COMMITTEE ON ASTRONOMY 



FINANCE COMMITTEE 



Richard S. Perkins, Chairman 
Crawford H. Greenewalt 
Alfred L. Loomis 
Keith S. McHugh 
Henry S. Morgan 



Crawford H. Greenewalt, Chairman 
Amory H. Bradford 
William McC. Martin, Jr. 



COMMITTEE ON BIOLOGICAL SCIENCES 

Alfred L. Loomis, Chairman 
William I. Myers 
Charles P. Taft 



NOMINATING COMMITTEE 



Carl J. Gilbert, Chairman 
Crawford H. Greenewalt 
Keith S. McHugh 
James N. White 



COMMITTEE ON TERRESTRIAL SCIENCES 

Juan T. Trippe, Chairman 
Richard S. Perkins 






Staff 



MOUNT WILSON AND 
PALOMAR OBSERVATORIES 

813 Santa Barbara Street 
Pasadena, California 91106 

Horace W. Babcock, Director 

Halton C. Arp 

Ira S. Bowen, Distinguished 

Service Member 
Edwin W. Dennison 
Armin J. Deutsch 
Jesse L. Greenstein 
Robert F. Howard 
Robert P. Kraft » 
Robert B. Leighton 
Guido Munch 
J. Beverley Oke 
Bruce H. Rule 
Allan R. Sandage 
Wallace L. W. Sargent 
Maarten Schmidt 
Arthur H. Vaughan, Jr. 
Olin C. Wilson 
Harold Zirin 
Fritz Zwicky 2 



GEOPHYSICAL LABORATORY 

2801 Upton Street, N.W. 
Washington, D.C. 20008 

Philip H. Abelson, Director 
Peter M. Bell 
Francis R. Boyd, Jr. 
Felix Chayes 
Gordon L. Davis 
Gabrielle Donnay 
Joseph L. England 
P. Edgar Hare 
Thomas C. Hoering 
Thomas E. Krogh 
Gunnar Kullerud 
Donald H. Lindsley 
J. Frank Schairer 
Hatten S. Yoder, Jr. 



DEPARTMENT OP 
TERRESTRIAL MAGNETISM 

62p Broad Branch Road, N.W. 
Washington, D.C. 20016 

Ellis T. Bolton, Director 
L. Thomas Aldrich, Associate 
Director 
Merle A. Tuve, Distinguished 

Service Member 
Roy J. Britten 
Louis Brown 
Dean B. Cowie 
Scott E. Forbush 
W. Kent Ford, Jr. 3 
Stanley R. Hart 
David E. Kohne 
Richard B. Roberts 
Vera C. Rubin 
I. Selwyn Sacks 
T. Jefferson Smith 4 
John S. Steinhart 
Kenneth C. Turner 



1 Resigned July 31, 1968. 

2 Retired June 30, 1968. 

a On leave of absence to August 31, 1967. 

4 On leave of absence October 1, 1967, to March 31, 1968. Resigned June 30, 1968. 



Staff {continued) 



DEPARTMENT OF PLANT BIOLOGY 

Stanford, California 94305 

C. Stacy French, Director 

Olle Bjorkman 

Jeanette S. Brown 

Jens C. Clausen, Emeritus 

David C. Fork 

William M. Hiesey 

Malcolm A. Nobs 

James H. C. Smith, Emeritus 



DEPARTMENT OF EMBRYOLOGY 

115 West University Parkway 
Baltimore, Maryland 21210 

James D. Ebert, Director 
David W. Bishop x 
Bent G. Boving 
Donald D. Brown 
Igor B. Dawid 
Robert L. DeHaan 
Elizabeth M. Ramsey 



GENETICS RESEARCH UNIT 

Cold Spring Harbor 
New York 11724 

Alfred D. Hershey, Director 
Barbara McClintock, Distinguished 
Service Member 
Elizabeth Burgi 



Cytogenetics Laboratory 
Ann Arbor, Michigan 



Helen Gay 



1 Retired June 30, 1967. 



Staff (continued) 



OFFICE OF ADMINISTRATION 

1530 P Street, N.W., Washington, D.C. 20005 

Caryl P. Haskins President 

Edward A. Ackerman Executive Officer 

James W. Boise Bursar; Secretary -Treasurer, Retirement Trust; 
Executive Secretary to the Finance Committee 

Marjorie H. Walburn Assistant to the President 

Donald J. Patton Director of Publications 

Sheila A. McGough Editor 

Kenneth P. Henard Assistant Bursar; Assistant Treasurer, 
Retirement Trust 

Pamela W. Thomas Assistant Editor 

Joseph M. S. Haraburda Assistant to the Bursar 

A. Gerald Thompson Assistant to the Director of Publications 

Marshall Hornblower Counsel 



STAFF MEMBERS IN SPECIAL SUBJECT AREAS 

Tatiana Proskouriakoff 
Anna 0. Shepard x 

1 Retired June 30, 1968. 



Staff (continued) 



RESEARCH ASSOCIATES OF THE CARNEGIE INSTITUTION 

Mateo Casaverde 

Lima, Peru 

Richard A. Chase 

Johns Hopkins University 

Louis B. Flexner 

University of Pennsylvania 

Irwin Konigsberg 

University of Virginia 

J. D. McGee 

Imperial College of Science and Technology, University of London 

Jan H. Oort 

University of Leiden 

Harry E. D. Pollock 

Carnegie Institution 

Reynaldo Salgueiro 

La Paz, Bolivia 

Shigeji Suyehiro 

Japan Meteorological Agency 



Former Presidents and Trustees 

PRESIDENTS 

Daniel Coit Gilman, 1902-1904 John Campbell Merriam, 1921-1938; 

Robert Simpson Woodward, 1904-1920 President Emeritus 1939-1945 

Vannevar Bush, 1939-1955 





TRUSTEES 




Alexander Agassiz 


1904-05 


Seth Low 


1902-16 


George J. Baldwin 


1925-27 


Wayne MacVeagh 


1902-07 


Thomas Barbour 


1934-46 


Andrew W. Mellon 


1924-37 


James F. Bell 


1935-61 


Margaret Carnegie Miller 


1955-67 


John S. Billings 


1902-13 


Roswell Miller 


1933-55 


Robert Woods Bliss 


1936-62 


Darius O. Mills 


1902-09 


Lindsay Bradford 


1940-58 


S. Weir Mitchell 


1902-14 


Robert S. Brookings 


1910-29 


Andrew J. Montague 


1907-35 


John L. Cadwalader 


1903-14 


William W. Morrow 


1902-29 


William W. Campbell 


1929-38 


William Church Osborn 


1927-34 


John J. Carty 


1916-32 


James Parmelee 


1917-31 


Whitefoord R. Cole 


1925-34 


Wm. Barclay Parsons 


1907-32 


Frederic A. Delano 


1927-49 


Stewart Paton 


1916-42 


Cleveland H. Dodge 


1903-23 


George W. Pepper 


1914-19 


William E. Dodge 


1902-03 


John J. Pershing 


1930-43 


Charles P. Fenner 


1914-24 


Henning W. Prentis, Jr. 


1942-59 


Homer L. Ferguson 


1927-52 


Henry S. Pritchett 


1906-36 


Simon Flexner 


1910-14 


Gordon S. Rentschler 


1946-48 


W. Cameron Forbes 


1920-55 


David Rockefeller 


1952-56 


James Forrestal 


1948-49 


Elihu Root 


1902-37 


William N. Frew 


1902-15 


Elihu Root, Jr. 


1937-67 


Lyman J. Gage 


1902-12 


Julius Rosenwald 


1929-31 


Walter S. Gifford 


1931-66 


Martin A. Ryerson 


1908-28 


Cass Gilbert 


1924-34 


Henry R. Shepley 


1937-62 


Frederick H. Gillett 


1924-35 


Theobald Smith 


1914-34 


Daniel C. Gilman 


1902-08 


John C. Spooner 


1902-07 


John Hay 


1902-05 


William Benson Storey 


1924-39 


Barklie McKee Henry 


1949-66 


Richard P. Strong 


1934-48 


Myron T. Her rick 


1915-29 


William H. Taft 


1906-15 


Abram S. Hewitt 


1902-03 


William S. Thayer 


1929-32 


Henry L. Higginson 


1902-19 


James W. Wadsworth 


1932-52 


Ethan A. Hitchcock 


1902-09 


Charles D. Walcott 


1902-27 


Henry Hitchcock 


1902 


Frederic C. Walcott 


1931-48 


Herbert Hoover 


1920-49 


Henry P. Walcott 


1910-24 


William Wirt Howe 


1903-09 


Lewis H. Weed 


1935-52 


Charles L. Hutchinson 


1902-04 


William H. Welch 


1906-34 


Walter A. Jessup 


1938-44 


Andrew D. White 


1902-03 


Frank B. Jewett 


1933-49 


Edward D. White 


1902-03 


Samuel P. Langley 


1904-06 


Henry White 


1913-27 


Ernest 0. Lawrence 


1944-58 


George W. Wickersham 


1909-36 


Charles A. Lindbergh 


1934-39 


Robert E. Wilson 


1953-64 


William Lindsay 


1902-09 


Robert S. Woodward 


1905-24 


Henry Cabot Lodge 


1914-24 


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. 



XI 



Report of 
the President 



Mankind is now in one of its rare moods of shifting its outlook. The mere com- 
pulsion of tradition has lost its force. It is the business of philosophers, students, 
and practical men to re-create and re-enact a vision of the world, conservative and 
radical, including those elements of reverence and order without which society 
lapses into riot, a vision penetrated through and through with unflinching ration- 
ality. Such a vision is the knowledge which Plato identified with virtue. 

Alfred North Whitehead — Introduction to Business Adrift, 

by W. B. Donham, 1931 



There are inherent in the scientific method ethical values which, if they were to 
command common assent, would have an immense stabilizing value on public 
affairs. I cannot do more than catalogue them, here. There is the scientist's undog- 
matic attitude to truth and error; his acknowledgement that truth is always changing, 
approaching an asymptote of perfection, but never reaching it. And error? Error 
is not evil: it is only discarded approximations to truth. . . . 

And lastly there is the internationalism of science: it is the one remaining force of 
cohesion between peoples who may be at war on every other issue. Whether the man 
who upsets my scientific theories is my countryman or my enemy is immaterial: 
the theories are upset all the same. Perhaps, then, it is through reflection on the 
ethical and moral implications of science and technology, that is to say the influences 
of these studies on humanism, that a country may make its greatest contribution to 
civilization. 

Sir Eric Ashby — "Machines, Understanding, and Learning, Reflections on 
Technology in Education," The Graduate Journal, University of Texas, 1967 



We live in a world of unreality and dreams. To give up our imaginary position 
as the center, to renounce it, not only intellectually but in the imaginative part of 
our soul, that means to waken to what is real and eternal, to see the true light and 
hear the true silence. . . . It is a transformation analogous to that which takes place 
in the dusk of evening on a road, where we suddenly discern as a tree what we had 
at first seen as a stooping man; or where we suddenly recognize as a rustling of 
leaves what we thought at first were whispering voices. We see the same colors; we 
hear the same sounds, but not in the same way. 

Simone Weil — Waiting for God, 1959 



This is a year of watershed. In our swift-moving and protean world 
every new year must carry a particular significance, must bring fresh 
opportunities, and at the same time must carry forward and renew a pano- 
rama of already existing concerns. This is as true in the realm of science as 
in any other segment of national, or indeed of world, affairs. It is a normal 
situation, and one which we have long since learned to expect. 

Yet by any standard, and from any aspect, the year just past and the 
year ahead may be viewed in retrospect as marking a dividing range in 
many of the greatest concerns of our society and our time — in the very 
evolution, indeed, of our nation. Surely, in terms of major political and 
social tides in both national and international affairs, the critical nature of 
the past year, and its aspect of watershed, are so compellingly evident as 
to make further emphasis superfluous. But as a rising ocean tide affects 
the character and course of every creek and river and winding estuary 
that enters it, and is in turn affected by their flow, so the events of the past 

3 



Jf CARNEGIE INSTITUTION 

year — and the shape of deeper social trends that these events reflect — have 
deeply scored their marks in other areas as well, areas such as science, that 
preoccupy us both as individuals and as a nation, and that are importantly 
and indissolubly of the nexus of those events. It is somewhat striking to 
reflect that in this year, the two hundredth anniversary of the birth of that 
great French mathematical physicist Jean Baptiste Joseph Fourier at 
Auxerre in Yonne, we find the affairs of science as sensitively responsive 
to elements of their larger social background, as intimately reflecting them 
and bounded by them, as Fourier found in a very different context of place 
and time and circumstance — and, so finding, managed personally to span 
the gulf between them during the years of his scientific greatness and of the 
French Revolution. For American science, as for other national concerns, 
the shape of the year's events has indeed had some aspects of a watershed 
whose outlines and influence could significantly affect the contours of our 
scientific landscape for some time to come, or even, indeed, permanently. 
How we respond over the next few years to very real crises in the scientific 
affairs of the nation could leave an important and perhaps enduring imprint 
on our fortunes. 



A century and a half ago, Thomas Jefferson wrote, in a letter to John 
Adams, "I like the dreams of the future better than the history of the past." 
As in so many other aspects of his life and work, it seems very clear today 
that Jefferson, in that single sentence, epitomized with extraordinary 
clarity a creed of the young nation that quite literally had the quality of a 
faith. It was a faith that is hard to define precisely. But perhaps it can best 
be expressed as a consuming passion for the wonder of the world — both the 
natural world and the world of man — and a total dedication to growth and 
movement and aspiration, which taken together sum to the quality of hope 
which today is surely our most precious, and perhaps our most threatened, 
common possession, the very root of our vitality as a people, the quality 
to which men over all the world will respond, particularly in our time, as 
they will to nothing else. And there was a third facet to that faith. It 
comprehended a real reverence, with all the power and the inspiration that 
the fact of reverence conveys. 

Through the centuries since Jefferson's death the society that he knew, 
shaped and motivated in good measure by that faith, has grown to propor- 
tions of which he could hardly have dreamed. It has also, quite obviously, 
changed in structure and character in ways that would have quite defied 
his imagination. New problems have replaced the old at the forefront of 
our concerns. New views of nature and of man have swept like swift, re- 



REPORT OF THE PRESIDENT 5 

current tides over the nation, as they have over a world society transformed 
by experience and development beyond anything that a Jeffersonian society 
could have comprehended in its most soaring nights of imagination. Super- 
ficially it sometimes appears that we are so different a people today, buffeted 
by such new orders of challenge and demand, beset by such new kinds of 
imperatives for our very survival let alone our further growth, that bonds 
with that past have dissolved and vanished. 

Yet a strong case can be made for the contrary view. The basic precepts 
around which a society crystallizes, upon which the foundations of a nation 
are laid, the precepts which have supported and guided such a people to 
maturity and unexampled power, and also to genuine greatness, can be 
discarded only at the most serious peril to the cohesiveness and effectiveness 
and indeed to the basic health of that society. In social as in biological 
evolution, the fundamental elements of structure and of function must be 
retained at all costs, however rapid the pace of evolution in more peripheral 
elements. To deviate from this principle is to jeopardize survival itself — a 
fact illustrated everywhere in the living world, and recognized at a deep 
and pervading level by most societies and nations, including our own. 
The evidence is very strong indeed — and deeply salutary for our future — 
that our devotion to growth and vitality and aspiration, and our reverence 
for these sustaining elements of hope, and the incentives to individual 
excellence that they imply, continue importantly to mold our ethos as a 
people, despite threats and impediments and even denials that abound on 
every hand. And it is especially noteworthy that there has been no national 
concern, from the time of Jefferson to this, where these elements of faith and 
hope and devotion to movement and growth have been more vividly 
exemplified than in the philosophy underlying and motivating the scientific 
way. 

But it would be idle, and worse, to ignore the threats to that ethos of 
growth and vitality and aspiration and reverence posed by the nature of our 
time, or to minimize their dangers. Some seem intrinsic to that passion, 
which has possessed us from the beginning, to dominate our physical 
environment and to use it to the utmost, indeed to consume it with an 
unexampled deadly effectiveness and often enough, indeed, with an unex- 
ampled greed. A prescient observer has pointed out that in any closed 
system, as our environment essentially is, there cannot be overweening 
mastery and organization in one part without correlative disordering and 
destruction elsewhere, unless special thought and effort are tirelessly given 
to the problem. We are only now beginning to implement that basic prin- 
ciple, so crucial to our continuing welfare, in any telling way. As we have all 
become suddenly aware, the ethos of movement and growth upon which 
we founded our society, and which has brought us to our present stature, 



6 CAENEGIE INSTITUTION 

also has its obverse potential. The most sobering task for our time, perhaps, 
is to learn how so to live and plan that the following generation, too, can 
have its chance. For, as Coleridge so truly said: "in today already walks 
tomorrow." 



Nowhere in our national life are the hazards of the times more explicitly 
illustrated than in the field of science itself. In this year of change, it is 
worthwhile to give special attention to some examples of such scientific 
hazards. 

Since the time of Newton, an unusually rapid rate of growth has been one 
of the hallmarks of science. As Derek Price presciently pointed out some 
years ago, the volume of scientific work in the world, whether measured by 
the number of scientific papers issued, or by the number of scientific 
journals, or by the number of scientists themselves, has been increasing at 
an exponential rate such as to approximately double every ten to fifteen 
years, multiplying itself roughly tenfold two or three times each century. 
This is not a new relation. It has clearly held true over the centuries since 
Newton. Neither is it in any way unique in kind. For it seems probable 
that the rate at which all knowledge has expanded- — and perhaps the ac- 
celeration of many other human activities as well, over the last century at 
least — has also been exponential. But, as Price has suggestively demon- 
strated, there seems to be a special structuring to the research front of 
science that has resulted in a growth rate far in excess of those of other 
human activities. An illustration of this that comes close to home involves 
the growth of the population of those men and women holding scientific 
and technical degrees in the United States over the century from 1800 to 
1900. It has been estimated that in 1800 there were approximately one 
thousand individuals in the country who held such degrees. In 1850 the 
number was about ten thousand. By 1900 it had swelled to the neighbor- 
hood of one hundred thousand. Currently it is believed that the number 
of those trained in science and technology — the scientific pool — has in- 
creased by twenty percent since 1964, and its present size is estimated as 
high as eight hundred thousand or more men and women. By such criteria, 
rapid growth seems of the essence of the scientific way. 

John R. Piatt has provided particularly vivid illustrations of this growth 
rate in quite a different context — that of the products of scientific thought 
embodied in some fields of contemporary technology. Toward the end of 
the third decade of this century, for instance, atomic particles could be 
accelerated to speeds equivalent to approximately five hundred thousand 
volts. During the century's fourth decade, that ceiling climbed to about 



REPORT OF THE PRESIDENT 7 

twenty million electron volts. In the fifth decade, it reached approximately 
a half billion. By the middle 1960s it stood at about thirty billion, and 
current aspirations, of course, reach much further, to two hundred, and 
then possibly to one thousand billion electron volts or more. Over a con- 
siderable period, the energies of such machines have increased by a factor 
of ten approximately every seven years. 

Another and more immediately relevant indicator of the dynamic 
growth that is so characteristic a quality of both science and technology 
is the proportion of our national income that has been expended in its 
support over the last decades. In 1940, for example, the total national 
budget for research and development was reckoned at $74 million. A decade 
later, after World War II, it had burgeoned to approximately $1.2 billion. 
In 1953, at the close of the Korean war, it had reached $3 billion, and in 
1956 it amounted to approximately $3.5 billion. It is interesting to notice 
the steady and considerable, but not explosive, expansion over the years of 
the Korean conflict. Then came Sputnik, and by 1959 the part of the 
national budget labeled for research and development had reached $5.8 
billion. Four years later, in 1963, it amounted to $11.9 billion, a sum greater 
than the total funds spent in this category over all the years from the 
American Revolution to and through World War II, and amounting to 
nearly fifteen percent of the entire expendable Federal budget. By the 
following year, the figure had grown to $14.6 billion. Over the eventful 
decade preceding 1964, the estimated Federal budget for research and devel- 
opment had expanded by more than four hundred percent, and a projection 
of continued growth at essentially this extraordinary rate, so kindling to 
our tradition of coupling faith and hope with dynamic movement and 
dynamic growth, had become widely assumed in the communities concerned 
with both research and development. For those communities, as for the 
nation at large, it was a heady period. 

Then, in 1965, there came a sharp turn. The Federal budget allocation 
for science for that year showed an increase over the preceding one of only 
$0.2 billion. The causes of this sudden leveling seem initially to have been 
largely internal. In part, there was concern in the Congress that a rise 
so steep as that of the preceding decade could not be projected much 
further without producing a serious imbalance in the total pattern of 
national commitments. Partly it was caused, clearly, by the need strongly 
felt in the Executive branch to impose some structure on the extraordinary 
tangle which this precipitate expansion had created — to try to find some 
ordering of priorities. That task, of course, has proved of overwhelming 
difficulty. Indeed, in the field of basic research, with its fundamental and 
proper unpredictability, this particular problem, by its very nature, may 
never be solved. 



8 CARNEGIE INSTITUTION 

By 1966, external factors of immense magnitude had been added to these 
internal pressures for reduction of the growth rate. The Federal budget 
for research and development rose to $16.1 billion in that year, but in- 
creased only to an estimated $16.7 billion for 1967, to about $16.9 billion 
in 1968, and to approximately $17.2 billion for 1969. At no time, to be 
sure, had there been a formal reduction in budgeted funds. But after 1964 
the rate of growth, once of the order of twenty percent per year, shrank to 
an average of about two and a half percent. It is to be remembered, too, 
that the purchasing power of the research dollar itself had been falling 
constantly for more than a decade and a half, because of the combined 
effects of inflationary forces on salaries in science and on the purchase 
price of equipment, and, most important, because as any science advances, 
the sophistication and power and complexity and hence the cost of the 
instrumentation required tends constantly to augment. Thus Dael Wolfle 
has calculated that one dollar spent for research in 1965 had roughly the 
purchasing power of eighty-two cents in 1960, of fifty-seven cents in 1955, 
and of forty-one cents in 1950. So it is a genuine question whether there has 
been any real increase in Federal funding of research and development, 
considered together, since 1965. 

This year, to the powerful budgetary constraint of the Vietnamese war 
have been added two further constraining forces. Each threatens to impose 
very severe limitations on the Federal support of both science and tech- 
nology. Yet each is undeniable in its extraordinary importance for the 
nation. The first is the $6 billion cut in national expenditures, which may 
finally amount to something in the neighborhood of $7 billion, and is likely 
to fall especially heavily on Federal support for research and development 
in the next year or two. The second constraint results from the vast demands 
which must be made on our revenues, for as many years ahead as we can 
see, by the plight of the cities and the vital need to restore our environ- 
ment. It is difficult to see where funds for substantial expansion of the 
Federal budget for research and development, taken together, can come 
from for several years ahead. Indeed, it is hard to see how an actual decline 
can be avoided. 

Now there can be no doubt that, during the period of explosive expansion 
in Federal support for research and development, a great deal of money 
was spent, and a great deal of energy and time were consumed, in projects 
that were less than well conceived. It is possible to argue with some weight 
that a period of consolidation, a period in which we must address ourselves 
to the difficult questions of priorities that have never really been answered, 
a period in which we must make the hard choices that we dislike to make — 
and have always disliked to make in the context of rapid and chaotic 
growth that has shaped our society from the beginning — can redound to 
our real benefit. 



REPORT OF THE PRESIDENT 9 

Furthermore, as Piatt has sensitively pointed out, growth curves of the 
kind that we have witnessed in the power of accelerators, or, for another 
example, in speeds of transportation, have many of the characteristics of a 
chain reaction, well illustrated by a flame. And chain reactions, by their 
very nature, do not go on forever — or if they do, they consume all before 
them. Accelerators cannot continue to increase in power indefinitely, else 
the revenues required to build them would exceed the capacities of any one 
nation, or even a combination of nations. Speeds of travel cannot increase 
indefinitely, for at velocities of less than twenty thousand miles per hour 
we have left the earth and are in orbit — where, indeed, we are already. 
Perhaps we are now face-to-face, at least temporarily, with some problems 
of this kind in the Federal funding of research and development. 



In this context, a challenging further dimension for thought and analysis 
has recently been opened by some new and suggestive findings of Derek 
Price. In a series of penetrating studies, he has been able to show that the 
contribution of the various nations of the globe to the world's store of 
scientific information per se, as measured by the share of the world's 
scientific papers in various fields annually issuing from them, is remarkably 
coordinate, not with their total populations, not with their own estimates 
of the funds which their governments expend in research and development — 
which may vary from less than 1 percent of their annual budgets to the high 
of approximately 3.5 percent of the gross national product reckoned for our 
own country — but, remarkably enough, with their overall national wealth. 
That proportion turns out to be extraordinarily uniform among all nations 
which are making significant contributions to the global accumulation of 
scientific knowledge. Whether their wealth be large or small, their expendi- 
tures for research, defined as the effort to create and publish new knowledge, 
appear to approximate 0.7 per cent of the gross national product for each of 
them. In our own case, the disparity between that figure and our own present 
estimate of 3.5 percent of our gross national product spent for research and 
development is accounted for largely by expenditures for development, 
expenditures directed primarily to the utterly essential, but different, 
objectives of technological growth and particularly national defense. 

But, although this figure of 0.7 percent, as the minimum expenditure 
required of any nation that is significantly contributing to the world 
store of scientific knowledge, seems generally consistent among the con- 
tributing countries at the moment, it has not been — and the indications 
are that it will not in future be — constant with time. Price finds, for instance, 
that in 1956 it amounted to approximately 0.3 percent, roughly half the 
present figure. In the absence of important damping influences, the present 
0.7 percent is projected to double approximately every decade. 



10 CARNEGIE INSTITUTION 

But of course such a projection suggests an ultimate picture that is quite 
impossible. To express our current stated commitment of Federal funds to 
research and development, this figure of 0.7 percent of the G.N. P. must 
be multiplied by about five. If the same ratio were to hold in the future, 
then this projection would require us, by the year 2000, to spend roughly 
one half to two thirds of our gross national product on research and develop- 
ment, a situation which, literally interpreted in terms of our present con- 
cepts of such categories, would clearly be absurd. So it is evident that we 
may well experience some real and serious deceleration of proportional, 
though not perhaps of absolute, growth rate in the Federal support of 
research and development between now and the end of the century what- 
ever external circumstances prevail; and that, whether we do or not, we 
may have to evolve some really new conceptions of the whole issue. 



The first warnings of this challenge may be with us today. For it may be 
true that some real immediate and practical benefits can be derived from 
a relatively short period of forced consolidation in Federal support of 
the national scientific effort, with its compulsion to face hard assessments, 
which we do not like to make and almost certainly would not make without 
such pressures. But it is evident that prolonged stagnation would pose 
hazards of major proportions, and that these dangers, moreover, would 
mount swiftly with time. It is well worth our pondering the nature of some 
of those hazards, and considering carefully ways in which they might be 
overcome, or circumvented, or at the very least alleviated. 

Greatest of all the dangers of course, and underlying all of them, is the 
risk to spirit and vitality and hope that deceleration must always bring 
to a people who from the beginning have identified all three so much with 
growth and expansion and change. But there are others of a more practical 
kind. 

In the unremitting competition of our planet, it is crystal clear that any 
nation which permits its scientific resources to wither, or even to diminish, 
over any considerable period of time is ipso facto gravely compromising 
its position in the world. And the greatest of these resources, of course, is the 
human one, that "pool" of the scientifically trained within the population, 
and, most important, the new generation of the gifted young just now enter- 
ing upon their lives' work, who over the next decade will be manning our 
research frontiers. One of the gravest dangers of the deceleration that 
confronts us is that, unless we manage most carefully, it is likely to be 
particularly damaging to just this group. For it is, after all, a ubiquitous 
human tendency to cling especially tenaciously to the known and the tested 
when the going is difficult and uncertain. That temptation is no less strong 



REPORT OF THE PRESIDENT 11 

in the fields of science than elsewhere. Some time ago, for instance, the 
House of Representatives approved a reduction in the funding of the 
National Science Foundation by one fifth — from the $500 million requested 
for the 1969 budget to $400 million, while the Senate approved a budget 
only $10 million higher. Under pressures of this kind, as it becomes impos- 
sible for the granting agencies supporting scientific research to fund more 
than a fraction of the excellent proposals placed before them, the temptation 
to choose in favor of mature established investigators, with high reputations 
and fine records of successful research already publicly supported, over un- 
known but potentially innovating newcomers is understandably great. There 
is a persuasive immediate logic to this course. Yet these are exactly the 
choices that can be particularly dangerous to our scientific future. For- 
tunately, those presently conducting the affairs of the Federal granting 
agencies seem aware of this danger, and clearly they are making consider- 
able efforts to moderate it. 

But the threat to that most precious scientific resource that we have, our 
population of nearly mature scientific trainees who have shown brilliant 
promise, the predoctoral candidates and the candidates for postdoctoral 
fellowships, stems from other quarters too. The training grants of the 
National Aeronautics and Space Administration, for example, which were 
originally established to replenish qualified young talent from the universities 
as the space effort drained them away, and which since 1961 have supported 
work for the Ph.D. by more than a thousand young men and women, have 
dropped to a twentieth since 1966. A survey of one hundred nineteen 
physics departments in universities, recently conducted by the American 
Institute of Physics, revealed that sixteen percent of their senior personnel 
had lost all government research support this year, and the figure was 
expected to rise to twenty-one percent next year. Until 1964, appropriations 
for the National Institutes of Health rose between fifteen and thirty percent 
a year, a figure that was sometimes almost embarrassingly large. But since 
that year support has almost leveled off, and the gulf between resources 
and opportunities is certainly as great this year as it is in other sectors of 
Federally supported research. Such indicators, it is evident, are but the 
iceberg's tips, visible hints of possibly greater problems ahead. 

If, then, the flame must be damped for some time to come, how can we 
manage the containment to the least disadvantage? How can we live with 
diminished growth without stifling developing buds beneath the hardening 
bark, and so gravely compromising our future? That is indeed a major 
question for our time. There are, of course, no clear answers at present. 
But it is surely imperative to think hard about some specific features of the 
situation, and more broadly about the nature of research itself, now and in 
the future. 



12 CARNEGIE INSTITUTION 

First of all, it is worth reflecting that one of the most traumatic features 
of the deceleration that we are witnessing may not he so much in its actual 
magnitude as in the suddenness and unexpectedness of the event. Since science 
is, above all, a cumulative discipline, where new knowledge typically grows 
within the structure of knowledge already won, the basic requirement for 
its health is not to be measured so much in amount of support as in its 
stability and continuity. Unlike many other activities, the level of produc- 
tivity of the scientific way is neither strictly proportional to the amounts 
spent on it, nor can science, like an activity of entrepreneurial character, 
be rapidly expanded or contracted in response to sudden and swift changes 
in the base of support. The adjustments that must be made to effect 
either expansion or contraction on any major scale are much too basic 
and massive and time-consuming for this. Moreover, there is a "critical 
volume" for every successful scientific endeavor, however modest it may 
be — a "critical volume" of facilities and equipment, but above all of the 
number of fine minds engaged — below which it simply becomes ineffective. 
So unexpected fluctuations in the base, major swings from under- to over- 
commitment, uncertainty and unpredictability, are inherently ruinous — 
far more harmful than a somewhat parisimonious, yet assured and con- 
sistent, undergirding. This salient element of a policy for support of science 
has been well known for many years in many places. Awareness of it, both 
in this country and abroad, has motivated many suggestions and some 
policy, such as the trend to extend the terms of Federal research grants 
from one year to three years or five years; or such as the weighing of insti- 
tutional versus project grants that has occurred in many Federal agencies 
over several years, and has recently been a particular concern of, among 
others, the Wescoe Committee reporting to the National Institutes of 
Health. But it is one thing to manage matters so that these hazards of 
instability can be minimized on a relatively small, or local, scale. It is 
quite something else to do it effectively over the whole front of research 
support. 

Perhaps, during the next years, we should concentrate less on trying to 
determine realistic ceilings for our research expenditures, and give more 
thought to how we might provide assured floors to our support of research. 
Just as universities and private research institutes have in recent years 
attempted to order their affairs so that during times of austerity their 
tenured faculties or staffs continue to be supported with their own research 
assured at levels at least sufficient to provide continuity, so, perhaps, in the 
domain of Federal funding we ought to consider very carefully what mini- 
mal orders of support could, with a fair degree of probability, be assured 
over several years to come. Psychological adjustments to that kind of floor — 
bringing some sense of security and of the possibility of longer-range 



REPORT OF THE PRESIDENT 13 

planning as much as anything else — might catalyze hope and verve to a 
degree that should not be underestimated. 

The analyses of Price and his colleagues may provide both a central idea 
and a specific basis around which such an approach could be built. For if, 
as he has demonstrated, the figure of 0.7 percent of G.N.P. for support 
of research is at present an operational one among the nations currently 
making significant contributions to the world's store of scientific knowledge, 
perhaps this suggests such a natural "floor" for the Federal support of 
science itself. It is notable, of course, that this proposed "floor" in fact 
represents only one fifth of the estimated 3.5 percent of our gross national 
product that we have actually committed at present to research and 
development considered together. But the total 3.5 percent merges the 
support of what, as Price has cogently emphasized, are in many respects 
two quite different activities — research, "basic" and "applied" on the one 
hand, and technology on the other. Their distinctiveness in this context, 
and equally their close relationships in other frames of reference, are worth 
scrutiny and emphasis. 

Often, of course, there is a close-knit functional connection, and a vital 
one, between research and technological development. One need only cite 
the drama of the transistor to illustrate this point, or the evolution of 
penicillin, from its initial discovery, to the development of extraordinarily 
high-yielding strains of the parent mold selected from among X-ray- 
induced mutants of Penicillium notatum, to the wholly artificial syntheses 
of the drug recently announced. And the vital quality of this relationship 
is just as clearly illustrated by the limitations imposed on one partner by 
the absence of the other. It may well be, for instance, that computer 
technology is approaching a serious intellectual barrier at present that can 
only be surmounted by further research in three or four related branches of 
fundamental science. 

The development of the maser and the laser offers one of the best ex- 
amples of how intimate and powerful this sort of linkage can be. As 
Charles Townes has recently and cogently pointed out, the relationship 
here is particularly striking not only because of the extraordinary ad- 
vances in technology which have resulted — advances of a different order 
of magnitude than could possibly have been achieved, or even imagined, 
through "straight-line" technological approaches to the same goals. A 
reverse effect is equally arresting. For here not only did major scientific 
conquests bring revolution to whole areas of technology. Reciprocally, these 
conquests were themselves stimulated and supported by contemporary 
technological advance. Under the pressures of World War II, the technique 
of microwave oscillators was developed to the point where a new branch 
of physics, microwave spectroscopy, the study of the interaction between 



14 CARNEGIE INSTITUTION 

gaseous molecules and microwaves, was generated and productively de- 
veloped, at first primarily in industrial laboratories. Yet the commitment 
to that course of research soon waned there. The basic invention of the 
maser in 1951 issued instead from Columbia University, to be followed very 
shortly by similar developments at the University of Maryland and the 
Lebedev Institute in Russia. Clearly, it was the fusion of the essentially 
"pure" research field of microwave spectroscopy — itself an outgrowth of 
intensive applied work originally having predominantly military ends in 
view — with applied electronics that provided the setting from which the 
maser and the laser could emerge. 

At the far end of the scale of development, the practical consequences, 
present and potential, of the maser principle are of course so utterly extra- 
ordinary in their scope and power and variety as already to be legendary: 
from its applications to transoceanic communications; through satellites 
of hitherto unattained orders of range and precision; to the taking of 
scientific measurements of hitherto unrealized sensitivity; to instrumenting 
the world's most precise clock, so accurate that if it were to run three 
hundred thousand years its estimated error would be about one second ; to 
the production, by the laser, of light beams of intensities millions of times 
greater than ever before achieved, for the first time making it possible to 
reflect a beam of light from the moon's surface and receive it again on earth; 
to the precise drilling of holes in materials as refractory as diamond; to the 
provision of a linear measure of enormous precision applicable to the most 
conventional operations in civil engineering on the one hand and to the 
laying out of the course of the first great linear accelerator on the other; 
to ultrahigh-speed photography and holography ; to new modes of high-speed 
processing, storage, and retrieval of information; to the performance of 
delicate eye surgery, notably operations for detached retinae — the range and 
scope and novelty of the uses of the principle seem virtually limitless. And 
the end, clearly, lies far in the future. 

Rather similar circumstances, of course, surrounded the development of 
the principle of the transistor, which has had such enormous application. 
Again a technological environment, coupled with the atmosphere and the 
freedom characteristic of basic research, made possible advances in funda- 
mental understanding that transformed the technical practices of the day, 
as undoubtedly they will those of the future. 

Thus the evidence is crystal clear that research and development can be, 
and in some strikingly successful enterprises have been, wedded in extra- 
ordinarily intimate working partnerships, with spectacular consequences 
for both. But it is equally clear that functionally the two activities are by 
no means identical. There are real and deep-lying differences between them, 
differences in both character and mode and in the environments best suited 



REPORT OF THE PRESIDENT 15 

for their prosecution. In the broad area of national support of research and 
development, it is rare in practice that these two kinds of activity are not 
distinct, and indeed they are rarely responsive to the same kinds of 
relationship to Federal support. 

Volumes have been written about the differences between "basic" and 
"applied" research. In Great Britain, at the national level, the two categories 
are actually administered by separate government departments. Yet such 
a distinction, if it is at all real, is not strongly reflected in the training or the 
capability or, basically, in the philosophy of the investigator, nor in the 
kinds of environment demanded for his best work. At most, it can only 
touch his stated objectives. But it does seem evident that there is a sharp 
contrast between science and technology. Between an activity having 
the primary goal of the winning, publication, and storage of new knowl- 
edge, and an activity oriented to developing, perfecting, and distribut- 
ing specific technical items, there do seem to be basic differences that are 
sharply reflected in both the orientation and the preferred methodology of 
the professional in the two fields. It is well to keep these differences in mind 
when viewing the 0.7 percent of G.N.P. that appears as the "standard" 
support among nations making significant contributions to the world's 
stock of published information, and, contrastingly, the higher percentage of 
the U.S. G.N.P. allocated for the support of science, which combines ex- 
penditures for research with the much more massive sums for technological 
development, whose fruits are primarily of practical application in national 
concerns, public and private. Thus the United States spends considerably 
more in the Federal support of research and development than does any 
European country. It is this combined total that we reckon as our 3.5 per- 
cent of G.N.P. which compares with about 2.5 percent for Great Britain, 
and 2 percent for The Netherlands, the current leaders on the European 
scene. But of two dollars so spent in our own country, one goes for research 
in space and defense. The total reflects very imperfectly our actual Federal 
commitment of dollars devoted to the winning of new knowledge. 



All this suggests that there may be real merit in a formal budgetary 
separation of the estimated Federal support of research from that of devel- 
opment. Such a move might make possible the formal establishment of a 
realistic and more stable "floor" for research support. It might well be set 
in the general region of this 0.7 percent proportion of the G.N.P. The 
concept of such a separation, of course, is not new. It was, in fact, put into 
effect several years ago. In retrospect, there seems little doubt that this was 
an excellent move. It sharpened concepts. Equally important, it made pos- 
sible a continuing real, if modest, increase in the support of research per se, 



16 CARNEGIE INSTITUTION 

well into the lean years. This policy may well have been related, though in- 
directly, to an evolution illustrated by the following figures. While it has 
been calculated that in 1953 somewhat less than seven percent of the total 
Federal commitment to research and development went to research, by 1965 
that figure had increased to an estimated eleven percent. In 1966-1967 it 
was possible, even in the face of stringently tightened circumstances, to 
increase the separated research budget by more than ten percent — main- 
taining roughly the same rate of growth for research that had characterized 
the combined research and development budget in some earlier years. It 
may never be possible to estimate the real effectiveness of such budgetary 
separation in cushioning the worst effects of the shocks attendant upon 
deceleration. But it is clear that its benefits had special impact upon young 
research talent just emerging on the scene. 

In the new Federal budget for 1969, this separation of research and de- 
velopment has been abandoned. Is not this move, far from being progressive, 
actually a retrogressive one? Should it not be reversed? 



In this context of the future growth of the scientific effort and of the 
significance of that growth for the nation, another interesting characteristic 
of the scientific way may prove important. Price has pointed out that over 
a span of years almost great enough to set a tradition, only about one fifth 
of the young postgraduates in science, newly equipped with doctoral degrees 
in their fields, have returned to academic environments. This "surplus" in 
the production of scientists suggests that, if we are careful to protect our 
resources of teaching in the future, we can legitimately anticipate that a 
goodly number of scientifically well-trained men and women will become 
available each year who will be free — and often, no doubt, eager — to enter 
fields of direct practical and social concern related to their specialties — 
sometimes, indeed, much broader than their specialties. It is from the 
ranks of such young people particularly that we may expect to draw a 
substantial proportion of those who must plan for the courses of science in 
the nation in the future — a planning process which we can no longer escape, 
a prospect which, much as we inherently dislike it, we can no longer in good 
conscience deny. But at least equally important, this "surplus" of trained 
scientists constitutes a highly significant reservoir of talent and effort 
available for some of the immense and urgent practical tasks that surely lie 
ahead for this generation and the next: the problems of the cities, the 
problems of the preservation and the building of our total environment that 
so preoccupy our thinking but cannot, to the degree that is so supremely 
necessary, commandeer our action today. So we may expect that a con- 
siderable proportion of the talent which will be developed today and to- 



REPORT OF THE PRESIDENT 17 

morrow with the aid of Federal funds formally designated for "scientific" 
purposes will, in the years of their maturity, repay the nation manyfold in 
areas that today we may think of as quite apart from those of science in 
the stricter sense. 

Indeed, it is probably difficult to overestimate the weight of the effective 
talent that people so trained could bring to national problems of this order, 
provided that the limitations as well as the capacities inherent in such 
training are kept soberly in mind. It must always be remembered, of course, 
that this definition of "surplus" can only be valid if the deployment of 
trained men and women is to the same general fields as those in which 
they were prepared. What must be guarded against is the all-too-common 
and all-too-human tendency to regard a training in science — or indeed in 
scientific technology — as providing the tools for attacking these ques- 
tions in a general sense. To imagine, for instance, that the vast and inter- 
locked congeries of problems involved in attempts to rescue deteriorating 
environments from further degradation can be solved by scientific skills 
alone, or even in principal part by them, is to gravely underestimate the 
nature of those problems, and at the same time to misapprehend both the 
strengths and the limitations of the scientific way. It is a temptation that 
planners and implementers alike must firmly and consistently resist. 

Such considerations, of course, lay very special emphasis on the signifi- 
cance of the teacher of science for the whole future of our nation, and 
further emphasize how very important it is, from the standpoint of Federal 
expenditures, that the deceleration in the rate of support to the universities 
for science teaching — teaching being, of course, the inseparable partner of 
academic research — be minimized as far as it possibly can be over the 
coming years. These same considerations further suggest that the signifi- 
cance of scientific skills in national affairs in the future may come to be even 
broader than we conceive it to be at present. This may offer the strongest of 
all imperatives to guard against reducing the Federal support of science, 
now and in the future, so far as we possibly can. 



There is a further characteristic of scientific research that should make it 
easier to accommodate continuing growth in its support over the lean years 
ahead, or at least substantially to cushion the shock of deceleration, pro- 
vided the budgets for research and development can once again be separately 
considered. It is now generally recognized that the costs of research, by and 
large, are far smaller, both relatively and absolutely, than those of corres- 
ponding programs of development. For in ways as fundamental to our 
spiritual welfare and growth as to our material substance, science and the 
humanities make common cause. In an era when the work of vast teams 



18 CARNEGIE INSTITUTION 

has replaced that of the individual in so many areas, for many of the sciences 
it is still as true as it is for the humanities that original and powerful indi- 
viduals remain the most significant finders of fresh paths. Abetted by the 
training and experience that is theirs, and in the case of science by the 
modern instrumentation which plays so key a role in scientific advance; 
immensely helped, again particularly in the case of science, by students or 
assistants or colleagues who form with them the close-knit cohesive working 
groups that continue to constitute the most stimulating and effective 
environments for creative research, nevertheless it is still basically these 
exceptional individual investigators who, in their thousands, pioneer, and 
will continue to pioneer, the new ways. In each new age, it is such youthful 
individuals recently entered upon research who typically create for their 
generation fresh visions of the world, visions rooted in and owing much to 
those of their teachers, but nonetheless distinctive and responsive to the 
new circumstances of the time. The evidence for this, both historical and 
contemporary, is compelling. In coming years our society will need to rely 
even more heavily on these recruits to guide us across the difficult watershed 
that is the terrain of our time. We must do everything within our power to 
assure to the coming generations the tools, the aids, the working environ- 
ments, and above all the individual scope for individual thought that they 
will require. 

Surely one of the most important things that we can do is to preserve 
and enhance and increase, to the greatest extent that we are able, those 
centers of excellence in both formal and practical research training that 
are so critical to preparation for scientific greatness. Hans Krebs has 
recently emphasized how vital such centers are, and what unequalled 
opportunity they can offer to the gifted young individual at a critical period 
in his career, molding him to the stature of those exceptional older men with 
whom he associates. A specific and remarkable example of this is provided 
by the life of Krebs himself. Destined to become a Nobel laureate in 
chemistry, he had the opportunity, during that critical period of develop- 
ment spanning his twenty-fifth to his twenty-ninth years, of working closely 
with Otto Warburg, and indeed of collaborating in a subsidiary capacity in 
the very investigation that brought the Nobel award to Warburg in 1931. 
And it is striking that the career of Warburg himself followed almost 
exactly the same course. As a young man, he was associated closely with 
Emil Fischer, who in 1902 had been awarded the Nobel prize for his work 
on the chemistry of sugars. Fischer, in turn, had worked with Adolf von 
Baeyer, who, for his discoveries in the chemistry of dyestuffs, particularly 
for his synthesis of indigo, also received the Nobel award, in 1905. And so 
the genealogy continues, back to the period when the formal acclaim of 
Nobel awards, with all their shortcomings as well as their virtues, was yet 



REPORT OF THE PRESIDENT 19 

unknown. For von Baeyer was a pupil of Kekule, most famous today, per- 
haps, for his vision of the structure of the benzene molecule. And Kekul6, in 
turn, studied with Liebig, who may fairly be said to have laid the founda- 
tions of organic chemistry. Liebig, again, had worked in Paris in the labora- 
tory of Gay-Lussac, who discovered some of the fundamental laws of the 
behavior of gases. Later Liebig referred to this experience as a critically 
determining one for the whole course of his life. Gay-Lussac, in his turn, 
learned especially from Berthollet, the pioneer in theories of combustion, 
who, among other pioneering achievements, abandoned the notions of 
phlogiston that in his time were still so entrenched. And Berthollet, finally, 
was a pupil of the great Lavoisier. 

Arne Tiselius has spoken movingly of this same relationship, again from 
the depths of personal experience. "To most research workers," he has 
written, "the decisive factor in preparing their minds is obviously their 
impressions and experiences during their university years, particularly if 
they have the good fortune of having a great scientist as their teacher. 
This was so in my case and it should be obvious to all those familiar with the 
work in physical and biochemistry in Sweden how much I owe to The 
Svedberg — a great personality and a great friend. ... In my experience, I 
have come to the conclusion (also expressed by many others in similar 
activities) : in the support of fundamental research, the individual research 
worker is more essential than the research project, when judging priorities." 

In every instance, as both Krebs and Tiselius emphasize, this association 
of pupil with teacher occurred at a period of research training roughly 
corresponding to the years that today would be comprehended in the 
doctoral and postdoctoral periods. And this was no casual or part-time 
connection. It was intense and prolonged. Upon the intimacy of this rela- 
tionship, as upon the qualities of both master and pupil, its unique effec- 
tiveness depended. It also depended, very clearly, upon the environment 
in which the experience took place. A glance at the history of the centers of 
training of great scientists through two centuries is enough to show that 
it is in the free and flexible and rather small research group, centered about 
a few figures of real stature both in research and teaching, which, in this 
context, are of course essentially identical, that the great figures of the 
following generation have developed, and continue to develop. From Liebig's 
laboratory no less than sixty leaders of science were "propagated" in 
succeeding generations from his day to ours, and more than thirty of these 
have been Nobel laureates. 

This situation is as contemporary as it is historical. Krebs has pointed 
to some interesting statistics concerning the sources of highly distinguished 
science — and scientists — in Great Britain over approximately the past 
twenty years. While in the earlier part of this period the universities clearly 



SO CARNEGIE INSTITUTION 

were a main source of scientific excellence, more recently the lead has passed 
in considerable measure to the research institutes, where teaching and 
administrative duties are traditionally light and time is available to concen- 
trate upon substance. Thus, of thirty-two new Fellows elected to the Royal 
Society in March, 1967, only thirteen had conducted the work decisive 
to their selection within universities, and some of these had held research 
professorships there. The British Medical Research Council, in whose 
laboratories Kendrew, Perutz, Wilkins, Crick, and Watson worked, among 
others, and where the crucial discoveries about the structure of DNA were 
consummated, operates with something like five percent of the research 
funds available to the universities and with perhaps one tenth the number 
of scientists. Yet, of a total of eighteen Nobel awards in all scientific areas 
made in Britain since 1950, no less than seven came to members of the 
staff of the Medical Research Council — an astonishingly large share. It 
would be hard to find more striking evidence of the essential place of the 
research institution and the flexible center of excellence, whose members 
are adequately protected from the grueling demands for the diversion of 
time and energy that are so omnipresent with us today. As Krebs has 
stressed, in our age, and particularly in these years of watershed, the 
cultivation of excellence in science is no academic exercise, but a funda- 
mental source of national strength, both economic and political. 



There can be no more vivid nor contemporary tribute to the continuing 
relevance and great importance of the individual, of the close-knit research 
group, and of the research center of distinction in which the individual 
works, and no greater satisfaction of that "need to know" that lies so close 
to the center of our beings and constitutes such a vital element of our 
spiritual heritage, than that afforded by any set of vignettes, however 
briefly and lightly sketched, of a randomly selected few of the advances in 
the understanding of nature that have marked the year just past. 

In astronomy, for instance, the strange quasars, the quasi-stellar sources 
that were first detected through the recording of their radio emissions at 
the British radio observatory at Jodrell Bank, and then, after being identi- 
fied as optically observable objects, were discovered at the Mount Wilson 
and Palomar Observatories to have extraordinary redshifts, continue to 
pose profound questions at the deepest level of cosmology. For the astonish- 
ing values of their redshifts, if interpreted in the usual fashion, indicate 
that they lie at such vast distances that, so far, no explanation satisfactory 
to all astronomers has been offered as to how they can be visible at all. 
And this year there has come the discovery of yet a new quasar, apparently 
more distant from us even than the farthest hitherto found, with a redshift 



REPORT OF THE PRESIDENT 21 

of 2.358, once again the greatest ever observed. It has been several times 
suggested, as a possible resolution of this dilemma of the quasars, that per- 
haps they are not in fact so distant, that a part or all of their enormous 
redshifts may be correlated with conditions other than distance, and that 
perhaps, after all, they are rather local objects. This would indeed explain 
their visibility. But studies of the numbers and the distribution of the radio- 
quiet quasars in the sky — made during the year just past, again at the 
Mount Wilson and Palomar Observatories — seem to eliminate that 
hypothesis. So the mystery continues. 

At the same time, once again from the Mount Wilson and Palomar 
Observatories, there has come this year a new determination of the rate of 
expansion of the universe, and new explorations, too, have been undertaken 
which now make a convincing argument, not only that the expanding 
universe was initiated by a gigantic "big-bang" explosion, but that the 
history of the universe may go back much further than this. In fact, the 
universe may have been alternating between expansion and contraction 
to a dense mass, then expansion again, with a total time cycle of approxi- 
mately 8 X 10 10 years, an interval that defies the imagination. 

Now to the mystery of the quasars is added another : that of the pulsating 
radio stars, or pulsars. The first of these was discovered at Cambridge in 
December of 1967. It emitted bursts of energy at radio frequencies with a 
period of approximately one and one third seconds, a period which is 
remarkably constant, to the order of one part in ten million. In rather rapid 
succession, three more similar objects were recorded: two with virtually the 
same highly constant period as the first, the third with a shorter period, 
also highly constant, of about a quarter of a second. A fourth, the first to 
be recorded from an American source, was detected at the Harvard Obser- 
vatory. It, too, pulsed with a frequency maintained with extraordinary 
precision, at intervals of about seven tenths of a second, each pulse lasting 
for about one fiftieth of a second. Ten pulsars are now known — two dis- 
covered by the Australians, one at Arecibo, one at Harvard, and the re- 
mainder in England. 

What are these strange objects, and what can account for both the 
regularity and the comparatively high frequency of the energy bursts which 
they emit? Because of the delaying effect of electrons along a radio path, an 
effect to which lower frequency radio waves are especially sensitive, the 
portions of the energy pulse that lie in lower frequencies may reach the 
earth seconds later than those at higher frequencies, and this delay can be 
used as a measure of the distance of the objects. Calculations of this kind 
have led to general agreement that, in marked contrast to the quasars, 
pulsars are, in astronomical terms, relatively nearby objects, ranging per- 
haps from fifty to three hundred light-years away. Thus they lie within 



22 CARNEGIE INSTITUTION 

our own galaxy— and indeed may be neighbors of our solar system within 
the Milky Way. 

Four nights of intensive examination with the two-hundred-inch tele- 
scope on Palomar Mountain failed to confirm the existence of any visible 
object corresponding to Pulsar CP 1919, one of the Cambridge discoveries 
which had earlier been suspected of emitting light in the visible range. 
What then can be the nature of these dark objects, and what the secret of 
their extraordinarily regular periodicity, or, even more remarkable, of the 
extremely short intervals between pulses? Highly rhythmic phenomena, 
such as the movements of a pair of binary stars around one another or such 
as, for that matter, the spin of the earth, are of course common and charac- 
teristic phenomena in celestial nature. But if the rhythm of the radio bursts 
from a pulsar is to be explained as being due to its rotation — as though, 
indeed, it were a rotating searchlight whose beam swept the earth once in 
every quarter second, or even once in every one and one third seconds— 
a far higher rate of revolution would be required than is commonly met 
with in the cosmos. Such an extraordinary rate of revolution in turn de- 
mands that the quasars be extremely small and extremely dense bodies. 
These qualifications might be fulfilled by a "white dwarf" — a star which, 
in a cataclysmic burst of energy as a nova, has burned its nuclear fuel and 
collapsed to a small, dense, and ultimately cold body. It has been urged that 
the spinning neutron stars, of extremely small diameter and intense magnetic 
fields, might be even more likely candidates. It has been proposed that the 
plasma stream from the surface of such a star may spiral out along magnetic 
field lines, its tangential velocity increasing as its radius increases until, 
when this velocity equals that of light, a burst of synchrotron radiation is 
emitted in a narrow cone, sweeping the earth on each rotation of the star. 
But at this frontier of knowledge there is still great doubt, and no theory 
has yet won general acceptance. 



Just one hundred years ago this year Louis Pasteur, having moved to 
the Ecole Normale as director of scientific studies, at last was able to 
proclaim definitively that the fermentation of wine results from the action 
of minute, invisible living organisms within it, and that when a fermentation 
fails to proceed normally, it is because the necessary organism either is 
absent or has not grown properly. It was in 1868 that Pasteur announced 
that he had isolated the causative organism of the disease of silkworms that 
was ruining the French silk industry, and that he had found a method of 
detecting infected stock, and of preventing contagion. 

It is interesting to reflect a century later how Pasteur, were he living 
at this hour, would have regarded the discovery of viruses, or the whole 



REPORT OF THE PRESIDENT 23 

complex of problems revolving about the coding and translation of infor- 
mation in heredity, to which the bacilli that he detected have themselves 
contributed so much information. How would he have reacted to the 
achievement of the laboratory copying of the genome of an infectious RNA 
virus, using naturally occurring biological templates and enzymes, by Sol 
Spiegelman at the University of Illinois? How, again, would he have reacted 
to the extension of that dramatic accomplishment this year: the copying 
of the genome of a virus with a DNA core, likewise using information and 
enzymes derived from living cells — a virus which then replicated itself with- 
in a living cell — by Goulian, Kornberg, and Sinsheimer? And most im- 
portant, one wonders what Pasteur, whose gifts of prophecy were so dra- 
matically exemplified in that work of a century ago, would have to say 
about the implications and the possible consequences of these contemporary 
achievements for the greatest goal, which lies yet in the future — the goal 
of the genuinely artificial creation of genetically competent DNA. 

What would Pasteur as a chemist have had to say about the remarkable 
accomplishment reported this year from the Max Planck Institute for Cell 
Chemistry in Munich, in which the fatty acid synthetase, obtained from 
yeast, with a molecular weight of 2.3 million, was successfully crystallized? 
This huge molecule, which acts almost like a chemical factory in catalyzing 
the immensely complex synthesis of fatty acid building blocks within the 
yeast cell, is by far the largest structure of its type ever to be crystallized 
intact. The achievement for the first time provides an opportunity to 
make an X-ray structural analysis of an extensive and highly organized 
"community" of enzymes, and suggests the possibility, at some future 
time, of really knowing the constitution of this functioning molecular 
"organism," and of understanding the way that its parts are operationally 
related. 

Again, what would Pasteur have thought of the complete synthesis of 
the substance glucagon at the Max Planck Institute for Albumen Research, 
also in Munich, announced during the year just past? Glucagon is an essen- 
tial hormone of the pancreas, appearing to play a role in the gastrin- 
secretion-histamine mechanism, and perhaps also in liver function. It is 
sufficiently rare, and sufficiently difficult to extract from its natural source, 
that it has always been in short supply. Quite apart from the technical 
brilliance of the synthesis itself, supplies of the synthetic substance which 
it may make available could have a marked impact upon the treatment of 
hyperglycemic shock, among other things. 

What would Pasteur the chemist have thought of the accomplishment 
involved in a complete synthesis of penicillin from so common and appar- 
ently unlikely a substance as isobutyraldehyde, yielding, finally, an isomer 
of penicillin V, announced this year from the laboratories of the Stevens 



£4 CAENEGIE INSTITUTION 

Institute of Technology? What would he have thought of the synthesis of 
artificial polypeptides by the solid phase method reported last year from 
the Rockefeller University? Until about 1932, it was difficult to synthesize 
polypeptides at all. Xow, with the new method, it becomes feasible to add 
new a min o acids to a growing chain, one by one. Thus it seems possible 
that a chain of almost any length and composition can be synthesized. 
When one considers the impressive biological and medical significance of 
many known polypeptides, such as, for example, vasopressin, which raises 
the blood pressure by contracting capillaries and arteries; or eledoisin, 
isolated from a Mediterranean species of octopus, a substance ten thousand 
times more potent in affecting rate of heart beat than quinidine, at present 
the drug most used for this purpose and itself recently synthesized; or 
physalaemin, derived from the frog Xenopus, a powerful inducer of abnor- 
mally high blood pressure, the potentialities of the new method become 
even more impressive. 



It will be just eleven decades ago in July of 1969 that a co mmu nication 

entitled "On the Tendency of Species to Form Varieties; and on the Per- 
petuation of Varieties and Species by Means of Selection" was jointly 
presented to the Linnean Society by Charles Darwin and Alfred Russel 
Wallace. A little more than a year later there issued a publication called 
On the Origin of Species by Means of Xatural Selection, or the Preservation 
of Favoured Races in the Struggle for Life — and the theory of evolution 
had been placed before the world. How excited, and perhaps amazed, 
Darwin would be, were he among us today, to review the most recent 
evidence that his basic principle can be adumbrated as vividly at the levels 
of cells, and of genes, and indeed of biological molecules, as at those of the 
entire organism and the populations of organisms with which he dealt. 
Perhaps he would be even more moved to see that, more than a century later, 
new discoveries continue to be made, new facets of his insights uncovered, 
at the very levels of life that most concerned him — the levels of organisms 
and their integrated populations. 

How deeply he would have been interested in the exciting demonstra- 
tions, first made at the Carnegie Institution and now widely extended, that 
replication of certain genetic loci, sometimes hundreds, or thousands, and, 
in at least one instance, a million times, seems to occur frequently in the 
chromosomes of higher many-celled organisms throughout the animal and 
the plant kingdoms. What would have been his conjectures about the 
adaptive quality of this extraordinary, and until recently unsuspected, 
phenomenon, so widespread in the a nim al and vegetable kingdoms? Had 
he been familiar with the knowledge, discovered only in very recent years, 



REPORT OF THE PRESIDENT 



that in higher vertebrates there may exist several chemical forms of the 
blood protein-pigment hemoglobin, each specified by a single gene locus; 
with the fact, for instance, that a change, brought about through mutation 
of the genetic message instructing it, in a single one of the 514 amino acids 
of which hemoglobin is composed (a change from glutamic acid to valine 
in the 6-position of /3-hemoglobin) could produce the pathological hereditary 
condition of sickle cell anemia, how deeply interested he surely would 
have been! Had he known, further, as we now know, that individuals 
carrying the gene determining this same sickle cell anemia, when they are 
heterozygous for it, are actually more resistant to certain forms of malaria, 
and so better adapted for survival than normal persons in localities where 
such malarias are rampant, though less well adapted in regions free of them, 
would he have seen this as one of the most poignant examples of survival 
equilibrium in polymorphic populations? How the universality of his theory 
at all levels of life would have been driven home to him ! 

In that primitive fishlike animal the lamprey, there is only one form of 
hemoglobin. Would Darwin have speculated, as a discerning scientist did 
this year, that, in the course of the evolution of early vertebrates, replica- 
tion of the single locus for hemoglobin might have had the effect of providing 
some new organism with "spare" genetic copies: evolutionary insurance, 
as it were, allowing normal functions to be preserved through the continuing 
control of the original, unchanged genetic locus while new hemoglobin- 
modifying mutations, derived from one or more genetic copies, could be 
"experimented with," leading ultimately, perhaps, to the establishment 
of a multiple series of hemoglobin pigments in the blood, such as we find 
today in higher vertebrates, with the physiological versatility that this 
provides? Would he have looked on this case, now so well documented be- 
cause of our new-found knowledge of the details of hemoglobin structure and 
its relation to genetic constitution, as a possible paradigm for similar kinds 
of evolution elsewhere? How would Darwin, who was keenly aware of the 
great evolutionary questions involved in the evident differences in rates of 
evolution of different organisms — problems that continue to baffle us so 
today — have related the phenomenon of genetic replication to this 
mystery? 

How Darwin would have delighted in Spiegelman's brilliant demon- 
stration and use of the principle of natural selection in an entirely new 
domain — at the level of the information stored in the cores of infective 
RNA viruses. By providing natural strands of RNA in the test tube with 
supplies of nucleotides to serve as raw materials for the synthesis of new 
strands, together with stores of the replicating enzyme replicase, new RNA 
strands were readily produced. By maintaining an abundant supply of 
both raw materials and enzyme, and then transferring newly synthesized 



26 CARNEGIE INSTITUTION 

strands of the RNA to fresh tubes as soon as they appeared, it was possible 
to carry out what in fact was an experiment in intensified Darwinian selec- 
tion, of the type often undertaken at the level of multicellular organisms, 
such as fruit flies. This process, repeated through seventy-five "generations" 
of new RNA, resulted in a population of strands which could replicate far 
faster than the originals. The striking point then demonstrated was that 
these RNA strands, selected through seventy-five passages of intense 
competition for rapidity in replication, had changed markedly in composi- 
tion. More than eighty percent of the information originally carried by 
the virus strand had been lost. Supplied with abundant available sub- 
strate for RNA synthesis and in the presence of adequate supplies of the 
replicating enzyme, freed from the necessity of coding for the structures 
necessary to the natural virus, such as the protein coat the particle must 
have to maintain an independent existence during its passage from one 
host cell to another, and the protein "apparatus" requisite for attachment 
of the virus to a new host cell, the selected RNA strands appeared to have 
"shed" a large part of the information code providing for these features. 
The "new" viruses in fact proved less infective than the originals. What 
had been conserved, however, and increased, was the capacity for rapid 
replication. So these strands of RNA had indeed demonstrated as truly an 
adaptive response to intensive selection as Darwin ever described among 
man's domestic animals. In further experiments of a basically similar kind, 
the triphosphate in the medium was lowered to a level barely usable by the 
normal virus. Triphosphate is vitally essential. Under pressure of this 
deprivation, a variant form of RNA, capable of utilizing triphosphate at 
lower concentrations than the normal form, appeared and rapidly dominated 
the culture, again in a pattern of Darwinian selection. A possible practical 
application of work of this kind, using some of the artificially modified 
types of RNA as tools to combat natural RNA viruses in infection by 
competing with them for substrates, is for the future to reveal. How that 
possibility, too, would have excited the men of Darwin's generation, had 
they been able to apprehend it! 

It was just a century in 1968 since Gregor Johann Mendel was elected 
Abbot of the monastery of Briinn, and his most active scientific work there- 
by brought to a close, three years after he had presented that momentous 
two-part paper before the Natural Science Society, the paper that laid the 
very foundations of genetics. In the library of the monastery there have 
been unearthed editions of Darwin's writings with their margins liberally 
annotated by Mendel, presumably during the period of his most intense 
work in the monastery gardens. So it is clear that the man who first grasped 
the very concept of those genetic elements whose mutational changes 
provide the basic raw material for organic evolution must also have thought 



REPORT OF THE PRESIDENT 27 

deeply about the nature and the pattern of that organic evolution. One 
cannot help wondering, in this connection, how Mendel may have speculated 
about the processes of evolution in those genetic mechanisms themselves. 
Over the years since Mendel's death many answers to such questions have 
been set forth, and much evidence uncovered to demonstrate that there 
has indeed been a highly complex evolution at the level of chromosomes, 
and even of genes — an evolution as complex and reticulate and beautiful 
as that at the level of the gross forms of organisms. But how, one may 
finally ask, did the mechanisms for information storage and retrieval them- 
selves develop and become elaborated, in the very earliest evolution of life? 

Here the rejoinders must still all be speculative. For, while the essence of 
genetic replication, in which identical copies of preexisting genetic material 
are made, inheres in reactions involving base-pairing — the fundamental 
nature of which we can reasonably hope eventually to comprehend even 
though we are still vastly ignorant of all its details — real understanding of 
the processes of translation, by which the genetic information is ultimately 
given tangible expression in the structure of proteins, is more difficult. How 
were the particular "rules" developed that now govern the relationships 
between the sixty-four possible trinucleotides and the arrangement of the 
amino acids in proteins— rules apparently universally followed throughout 
the panoply of living things? We are far indeed from answers to these ques- 
tions today. And until such answers are found, we will be able to say little 
that is definite about the way in which the processes of translation evolved. 
What a new dimension Darwin would have found in this field of evolution- 
ary thought! 

Again, how Darwin's keen eye for form and his sense of its significance 
throughout the living world would have been challenged by those investiga- 
tions of development and its mechanisms that constitute one of the most 
exciting frontiers of modern biology. One can imagine, for example, how 
the phenomenon of "programmed cell death" in development would have 
arrested his attention: the fact that the orderly, predetermined death of 
certain cells, occurring at the very time that neighboring cells are rapidly 
proliferating, and playing a vital role in molding the organism's ultimate 
form and structure and function, is truly the final "acting out" of instruc- 
tions coded originally within the DNA of the whole organism, and repre- 
sents a genuine example of specialization. How much Darwin might have 
been tempted to speculate about such examples of extreme specialization 
in terms of natural selection. Perhaps he would have pondered especially 
the following circumstance. Whereas an individual of unusually adaptive 
endowment living as an independent element in a polymorphic, panmictic 
breeding population commonly increases in numbers more rapidly than its 
fellows, so that its type may come to form a larger proportion of the whole 



28 CARNEGIE INSTITUTION 

as the generations pass, specialized high adaptive efficiency in the cell or 
organism that is tightly bound within a dynamically interacting, close-knit 
biological or social complex, in which the constituent living units are deeply 
and immediately interdependent, may have precisely the opposite conse- 
quence in evolution, drastically limiting the numbers of such a cell or 
organism and rigidly fixing the numerical proportions of the whole. 

And surely Darwin would have taken special interest in findings of this 
year in areas of inquiry that were already very familiar to him in the context 
of natural selection and that one might have imagined would long ago have 
been exhaustively explored. In the Origin of Species Darwin had much to 
say of the description, then but newly given, of mimetic butterflies associ- 
ated with their models along the Amazon, and of reports of similar cases of 
mimicry brought by Wallace from Malaya. Both Bates and Wallace had 
interpreted their findings as we would interpret them today, and Darwin, 
in a chapter of the Origin called "Analogical Resemblances," discusses at 
some length and with evident satisfaction the ways in which these remark- 
able instances of mimicry both confirm and extend the theory of natural 
selection. But how delighted Darwin might be, were he here today, to 
reflect that, after more than a century's exploration of this subject by a 
multitude of investigators, the year just past has revealed yet more examples 
of a new, and intricately wonderful, order of this phenomenon. In a genus of 
solitary wasps long known for the striking mimicry of a more powerful 
solitary hunting form exhibited by some of its members, it has recently 
been discovered that several species go much further. They show a dual, 
sex-limited kind of mimicry that simultaneously involves two very different 
models. The males mimic species of social wasps common within their 
range. The females mimic a solitary hunting wasp likewise abundant in 
their territory. Thus these versatile organisms wring " double dividends," 
as it were, from the protective resemblances so well known to Darwin, and 
offer fascinating examples of yet another variation of the pattern of natural 
selection at the level of living things where the theory of natural selection 
itself was first adduced. 



One of the arresting features common to this sampling of contemporary 
advances in our understanding of nature, and in our power to exploit 
newly discovered features of the natural world, is that all of them, despite 
their wide diversity in subject and approach, were achieved either by 
individual scientific leaders or, more often, by rather limited numbers of 
outstandingly trained and gifted investigators, working in those small, 
intense teams that so characteristically form the optimum environments for 
the highest scientific productivity. Moreover, although some of the investi- 
gations demanded very costly instruments, such as powerful telescopes, the 



REPORT OP THE PEESIDENT 29 

material needs of the majority were relatively modest. The order of mone- 
tary support required for each of these achievements was, by usual standards 
of development, truly minuscule. It is clear that the best research is often 
extraordinarily economical, in both money and the actual number of 
men and women involved — though not of the training and quality of 
these investigators. 

This characteristic economy becomes important to reckon with in the 
context of the present. Even under relatively severe economic constrictions, 
it should be possible, given sufficient care and imagination and insight, so 
to manage matters over the next few years that we do not unduly disrupt 
that flow in the creation of new knowledge that is so vital for our future, 
both spiritually and materially. The wisdom that we shall require here 
must evidently comprehend, among other things, a clear vision of the 
separate roles of research and development in our nation. It must also 
include as precise an estimate as we can achieve of their relative costs. 
It bears reiteration, in this connection, that the attempt made in past 
years to distinguish the costs of research and of development in Federal 
accounting takes on a special significance today. If this attempt were 
continued in future, it might well be found that, even if a fairly considerable 
reduction in Federal funds available for research and development, taken 
together, proved necessary for the coming year or two, support for research 
per se over the next critical years might not have to be reduced so severely, 
if indeed at all. 



But there are other dangers, of a more deep-lying and subtle kind, that 
may threaten us in the climate of deceleration of the next few years. They 
will demand vigilant and understanding attention. One of them stems 
from apprehensions about science that undoubtedly lie dormant in our 
society. 

Misunderstanding of the real nature of the scientific way is an ever-lurking 
danger with us, as it is all over the world. But the intensity of the problem 
varies greatly with circumstances. In a time like the present the threat is 
apt to be particularly severe. In such periods our concepts of the relation- 
ship of science to society are all too likely to polarize about either of two 
extremes. At one pole, science may come to be commonly regarded as an 
activity quite apart from the mainstream of society — even, in part at least, 
as a culturally enriching luxury, desirable, to be sure, but essentially 
remote from the principal currents of national concern. If this characteriza- 
tion gains wide acceptance in a society, it can be menacing. For, as Graham 
Hough has recently remarked in a different context: "In good times it is 
possible to pursue a multitude of remote and unpractical activities without 
feeling the painful sense of their disconnection with the actual and the 



30 CARNEGIE INSTITUTION 

present. . . . One of the functions of civilization is to provide room for 
innumerable pursuits that have no direct connection with the daily business 
of living. For in good times there are all manner of unseen threads that 
link the old with the new, and make even scholarship and criticism a part 
of the national life. The past can live in the present without a violent distor- 
tion of its real nature. . . . Today this seems hardly possible." 

In the absence of a keen appreciation of the relevance of even the most 
abstract science to the broader concerns of the society that supports it; 
in the absence of a clear understanding that it is in just those situations 
where viewpoints from the past and of the present are dangerously sundered 
that the scientific way can provide one of the most powerful of cohesive 
forces, one of the most durable of bridges across the gap, Hough's analysis, 
originally applied to other areas of culture, can only too easily become, for 
science, self-fulfilling prophecy. With this danger there is another that may be 
yet more serious. This view of science is all too apt to set off its own reaction: 
the demand for secularization ; the demand that science be plunged too fully 
into the mainstream; the demand that science, if it is to be supported by 
society, should immediately and demonstrably be usable by society, and 
should be deliberately cultivated for its pragmatically anticipated results. 
One of the dangerous aspects of this line of thought, of course, is that once 
the consequences of research are generally regarded as being predictable in 
this way research then logically becomes subject to the same sort of regula- 
tion and control, the same kinds of efficiency ratings, that are properly 
applied to activities with predictable results everywhere in the society. 
Precisely because science is intrinsically unpredictable, its whole essence, 
both as a spiritual and a material resource, can easily be destroyed if this 
sort of misapprehension generally obtains. 

This conceptual battle was fought and, on the whole, won, many years 
ago. Yet it is never really ended. It is in times like the present that we must 
pay special attention to the dangers of losing ground once gained. 



But critical times like these may, at the opposite extreme, exacerbate a 
precisely opposite misconception of the social impact of science, which can 
carry an equal danger. This is the presumption that science has become 
all too socially relevant; that, with technology, it has in our day quite 
literally captured the world, and is proceeding on a course of its own, 
Frankenstein-like, largely independent of social direction and largely 
escaped from social control. Titles of such recent books as The Temporary 
Society, The Technological Society, and The Accidental Century reflect this 
apprehension well. 



REPORT OF THE PRESIDENT 31 

There are undoubtedly powerful and readily understandable reasons for 
this feeling. Conditioned as we often are to associate the technologies of 
World War II and of the Sputnik era primarily with science, we are at a 
point where the possibilities that now seem accessible to us, whether or not 
they can or will ever be realized, seem far more radical than ever before. 
As we peer into the future, we can, without any undue stretch of credibility, 
see ourselves taking the first steps toward navigation of our solar system 
before many years have passed. When Johannes Kepler, in The Dream, 
written in 1609 and first published in Latin in 1634, discussed trips from 
the planet earth to its moon in future, this must have seemed wildly vision- 
ary to his contemporaries, as indeed the title of the work suggests that he 
himself regarded it. Yet how soberly prophetic, in many ways, this work 
seems today, particularly when the author warns that solar radiation will 
constitute a major hazard for space travelers. Only a bit further in the 
future seem to he the many possible consequences of biological and be- 
havioral control, with their vast potential social impacts, which, though 
we may sometimes be inclined to overdraw them because we simply cannot, 
at this stage, assess them accurately, have nevertheless received some 
emphasis in suggestive experimental work during this very year. Here, of 
course, we are dealing with what is basically technology, or technology 
closely interwoven with science. 

When we speak of the possibilities of future science, we are on yet more 
uncertain ground, where imagination and hope and fear have less to tram- 
mel them, because of the unpredictability that is so deeply characteristic 
of the scientific way. It is indeed a basic unpredictability. For, as Sir Peter 
Medawar has cogently pointed out, to predict a really new idea is to have 
that new idea, and if the idea has already occurred, of course, it can no 
longer be the subject of prediction. A quotation from the great English 
physiologist J. S. Haldane, the father of J. B. S. Haldane, is suggestive in 
this connection. In 1931 he wrote concerning the modes by which people 
were just beginning to suspect that the chromosomes operated, and by 
which heredity mediated development. Two investigators of the day, 
Ewald Hering and Richard Semon, "assume," says Haldane, ". . . that 
germ-cells are furnished with a system of engrams, functioning as guide- 
posts to all the normal stages of development. . . . But," Haldane continues, 
"this theory has quite evidently all the defects of other attempts at mecha- 
nistic explanations of development. How such an amazingly complicated 
system of signposts could function by any physico-chemical process or 
reproduce itself indefinitely often is inconceivable." That was published 
only thirteen years before the pioneering experiments of Oswald Avery and 
his colleagues, in which mutant strains of bacteria were genetically trans- 
formed to the normal type by the action of purified DNA from the normal 



32 CARNEGIE INSTITUTION 

form, were to lay the observational foundations from which have grown 
our contemporary concepts of the nature of these very engrams. 

There is a further psychological factor which can be important in this 
general context. It may well be that the imperatives to adjust to tech- 
nological change in our time are actually no greater than those that faced 
our predecessors. Nonetheless, they differ importantly. In our time not 
only is there reasonable promise of developments far more radical and 
more unfamiliar than any we face now and with those developments a threat 
of probing far more deeply and meaningfully into our most closely guarded 
personal domains. In addition, many of those anticipated innovations, and 
the futures that they imply, have been in our minds now for quite a long 
time. It seems quite possible that we are living with a cumulative effect 
of a long-continued awareness and expectation which, under the right 
circumstances, can be subtly and profoundly disconcerting to us. 

Our reactions range, often with abrupt swings, from anxious doubts to 
what may be worse — over-credulity: an exaggerated faith in the powers 
of science, and more especially of technology, to work miracles in our 
society far beyond the limits of their capacity or relevance. Both these 
extremes of feeling arise from misconceptions which we must particularly 
guard against in such times as the present. Overdrawn credulity, especially, 
can lead to errors that may be costly not only in money but in terms of 
the careers of men and women. Furthermore, the cost of such mistakes is 
not to be measured merely in immediate and concrete losses, heavy though 
these may be. With our immense powers of recuperation, our superb 
capacities to rectify our errors, such consequences can in general be miti- 
gated with time. What is more truly serious is that memories of such mis- 
takes, once made, are apt to linger, and to exacerbate the oscillations from 
over- to under-valuation of the scientific mode. In extreme cases, this 
effect can be most harmful. 



There is a final source of unease, primarily concerning technology, that 
affects us all in greater or less measure, and which unless, once again, we 
are vigilant, could have subtle and dangerous social consequences. Our 
own society is perhaps more vulnerable to this particular danger than any 
other on the globe — in part because we have so eagerly embraced tech- 
nology and live by its products to so great a degree. We are continually 
assailed, of course, by the problems of comprehending and coping with the 
sometimes too-localized over-plenty that massive and highly advanced 
technologies, applied in the sphere of daily living, can bring, with the social 
stresses and dislocations that frequently accompany them. But beyond this, 
it is probably no exaggeration to say that the technologies which so pervade 



REPORT OF THE PRESIDENT S3 

our everyday lives have, cumulatively over the years, come to pose real 
threats to our concepts of reality. Urban or suburban life absorbs a larger 
and larger fraction of our nation. More and more people are shut off from 
nearly all daily contact with the natural environment, living in homes and 
working in buildings well isolated from the outside world, moving between 
them, often enough, in sealed and air-conditioned vehicles, enjoying relaxa- 
tion through representations and using products that alike are predomi- 
nantly humanly conceived and executed. We may well come to conclude with 
Ian Jarvie that our all-embracing modern technology is indeed assaulting 
the most fundamental classifications of our experience. We may well ask 
with him "What remains of the basic distinction between the natural and 
the artificial when we are confronted with modern technology?" 

Does this blurring of distinctions matter? There is much to suggest that 
it may matter very seriously indeed. For when such a confusion is com- 
bined with the pressure of all the imagined things that science and tech- 
nology may bring us in the near future, a situation could be created in which 
the fundamental dividing line between imagination and reality becomes 
very badly blurred. There is evidence that this is happening on a significant 
scale in our society today, and its effects could be grave. For there is nothing 
more painfully frustrating than to have embraced the imagined on the 
assumption that it is real, to have wholeheartedly pinned one's vision of the 
future on something that is at best half plausible and never finally materi- 
alizes. One then feels cheated; one feels that what is in fact impossible has 
been willfully withheld. This is a peculiarly acute and dangerous kind of 
suffering. The threat of experiencing it can only be turned aside by a realign- 
ment of the very processes of apprehension and judgment, by the inward 
reestablishment of a firm boundary between what actually exists, or has a 
reasonable probability of existing, and what is, in essence, fictitious. This is 
a kind of stringent therapy which we, in common with a large share of the 
rest of the world, surely need. Should the situation grow worse over the 
coming years, it could threaten seriously the integrity, and indeed the 
safety, of our society. 

Both technology and science have significant parts to play in resolving 
this crisis. The role of technology, which is essentially to act as a bridge, is 
perhaps especially vital. The part of science in this combination is clearly 
that of stabilizer. For the natural sciences early defined "reality" by an 
extension in precision and percipience and control of the traditional mode by 
which laymen also defined reality — as that which could be experimentally 
verified. Verification, of course, originally meant confirmation by the senses. 
It also meant verification, not by the senses of one man only, but by those of 
many. Likewise it implied consistency through time, in the sense that the 
same processes undertaken at another time brought consequences which 



84 CABNEGIE INSTITUTION 

could be apprehended in the same way. This is still an acceptable basic 
sketch of the process by which verification is secured in the natural sciences, 
though the tests of interpretation are now, of course, far more sophisticated. 
It is implicit in our thinking that such tests are practical rather than 
philosophical. Indeed, as Martin Johnson long ago remarked, at a philo- 
sophical level we probably should regard verification, and so "reality," pri- 
marily as a measure of communication among scientists. But the practical 
tests, over a great part of the vast front of science, remain in this general 
mold — a mold, of course, which continues to be the one in which the criteria 
of "reality" for most of our everyday affairs are set; where, too, these 
criteria provide the most powerful and reliable medium of individual 
communication available to us. 

When we enter truly social realms, however, we encounter a very different 
situation. Here, as Donald MacKay has observed so truly, the comparatively 
invariant criteria of the senses no longer suffice. The questions asked of 
social science, and for which the social sciences are seeking the answers, 
cannot be regarded as either "true" "or "valid" in the conventional sense. 
For the content of these questions is sensitively responsive to the inter- 
pretation of the receiver of the information as well as to that of the sender. 
Here the attitude of the audience, the impressions, beliefs, extrapolations 
that may enter, all have a highly operative function in the defining of 
"reality," and their influence may greatly bias our conceptions of it. 

Now this approach to "reality" is necessarily quite different from that of 
the natural sciences. Indeed, it can sometimes lie — and properly— quite 
outside their tradition, and if the integrity and effectiveness of the natural 
sciences are to be preserved, it probably must remain so. And yet we live 
in a world where, with increasing frequency, these two definitions of "reality" 
are placed in apposition, or are actually conjoined around the same sub- 
stantive concerns. And those concerns commonly involve technology. 
This, perhaps, is one of the significant causes of the erosion of the once- 
clear line between imagination and reality, or, perhaps more accurately, 
the growing confusion of different kinds of "reality" that so haunts us to- 
day. There is no clear and present way out of the dilemma. But there can 
be no doubt that trying to fathom its unknowns and contradictions will 
pose a challenge of very high priority over the years ahead, and our answers 
to this challenge will have extraordinary practical as well as theoretical 
importance for our society. 



One final concern must be deeply significant for our time. Just before 
Christmas in 1954, the first real success in transplanting a vital organ from 
one individual to another — a human dream ever since stories of sixteenth- 



REPORT OF THE PRESIDENT SB 

century Europe had erroneously credited Gasparo Tagliacozzi, the great 
Italian surgeon, with this accomplishment — was achieved when a kidney 
was removed from a Boston twin and transplanted to his brother. Since 
that time, more than twelve hundred kidney transplants have been at- 
tempted, with a rising percentage of success. In December of last year the 
first transplantation of a human heart was made. Although the patient 
quickly succumbed, the operation was shortly followed by a second which 
was more successful. Some nine months later, in August of the year just past, 
the twenty-seventh such operation had been completed, and the condition 
of the patient reported as satisfactory. A lung has recently been trans- 
planted in an Edinburgh hospital. Transplantations have been made of 
liver and pancreas. Thanks to intensive research on such antiimmunogenic 
substances as antilymphocyte globulin, designed to suppress the normal 
rejection mechanisms of the body to foreign objects; thanks to improved 
methods and enhanced efficiencies in tissue typing and in the matching of 
appropriate donors and recipients, it seems quite possible that the trans- 
plantation of unpaired as well as paired vital organs between unrelated 
individuals may ultimately achieve a recognized position in surgical therapy 
comparable to the use of artificial hearts and kidneys to which we are 
accustomed today. 

With all the brilliant work that has been done, and the brilliant, if 
limited, successes so far achieved, that day is not yet. But the potentialities 
have been so clearly indicated that, with our penchant for running beyond 
the event in imagination, we are already in the process of psychologically 
adjusting to a very new situation. And that adjustment, of course, brings 
with it all kinds of difficult questions that are new for us. When do we have 
the moral right to remove the vital organs from a body : at what point can we 
morally consider a person to be dead? What is to be considered a "success" 
in terms of the patient-recipient? How long must he survive the operation? 
What condition of body and mind, and what order of social value, must he 
show promise of achieving to make us feel that the operation, however 
brilliantly successful in a technical sense, has been humanly justified? These, 
and a host of associated problems equally difficult and challenging, will 
surely be with us as far into the future as we can see, and they will be 
answered differently by various people at various times and places. 

But there is a deeper and more general issue underlying these matters 
that we must surely keep most critically in mind. In times of crisis, in a 
period when the experience of many kinds of violence to human life and 
the memory of that experience are all too vivid with us; in a period when 
the recollection of the catastrophic destruction of some of the most cherished 
monuments of human will and effort and of human reverence remains all too 
close, there is some real risk that the tremendous recent advances in the 



86 CAENEGIE INSTITUTION 

sciences of life, with the new possibilities that they hold for the betterment 
of the human condition — of which the history of organ transplant forms 
an especially dramatic chapter — might yet be accompanied, bit by bit and 
almost imperceptibly, by a trend to hold human life itself less sacred than 
it has been in the best periods of our past. There is the associated possibility 
that the capacity for reverence itself, perhaps our most precious moral and 
social quality, might slowly be eroded. 

Clearly there exist powerful bulwarks in our society against these dangers. 
They inhere primarily in those vital qualities of hope and devotion to move- 
ment and expansion that have always been so much at the forefront of our 
thought and feeling. In that context, science itself, with its own dedication 
to those qualities, with its total dependence upon them for its own welfare 
and integrity, must form one of the most stable and steadfast of such bul- 
warks. This, perhaps, is the greatest reason why we should give the most 
critical thought, in this year of watershed, to protecting and cultivating, in 
every way that we can conceive, the vitality, and above all the verve, of 
the scientific way. 

One may long, as I do, for a gentler flame, a respite, a pause for musing. But 
perhaps there is no other peace for the artist than what he finds in the heat of combat. 
"Every wall is a door," Emerson correctly said. Let us not look for the door, and 
the way out, anywhere but in the wall against which we are living. Instead, let us 
seek the respite where it is — in the very thick of the battle. . . . Some will say that 
this hope lies in a nation; others, in a man. I believe rather that it is awakened, 
revived, nourished by millions of solitary individuals whose deeds and works every 
day negate frontiers and the crudest implications of history. As a result, there 
shines forth fleetingly the ever threatened truth that each and every man, on the 
foundation of his own sufferings and joys, builds for all. 

Albert Camus — Resistance, Rebellion, and Death, 1957 



The Year in Review 



And by that destiny to perform an act 
Whereof what's past is prologue, what to come 
Is yours and my discharge. 



— Shakespeare, The Tempest, 
Act II, Scene 1, lines 260-262 



Dr. James D. Ebert, Director of the Department of Embryology, states 
in the Introduction to his report this year: "... it has been uncommonly 
difficult to summarize work already completed while conveying an idea of 
new directions. The problems encountered in reaching a proper balance of 
'old' and mew' are illustrated in . . . our activities." Dr. Ebert's words 
apply equally well to the programs of all the Departments of the Institution, 
and perhaps to those of any vital research organization. 

Like a "normal" biological population, a research organization's interests 
at a given time are composed of old and new, the budding and the closing, 
the ascendant and the waning. In a sense the great questions, like the 
analysis of the structure and content of the cosmos, or the analysis of the 
particulate composition and structure of energy and matter, are timeless 
and at once both old and new. When we examine the progress of any field 
of inquiry and its content at any one time, Dr. Ebert's observations do 
indeed apply. 

The distinction is operationally important, for in these days of rapid com- 
munication the time span between the new and the old may be relatively 
short, as a "lode" is soon mined by hundreds of scientists. Or rapid senes- 
cence may supervene, or premature abandonment of once-flourishing fields 
of inquiry occur. And fields may even be "rejuvenated," as the tide of 
interest turns again and energizes them. 

Dr. Ebert's remarks call special attention to the nature of all our Year 
Books. They are, above all, progress reports and not statements of con- 
clusions. Year Book 67 is no exception. In summarizing the content of its 
individual reports this year it seems appropriate to stress explicitly what 

87 



88 CARNEGIE INSTITUTION 

has always been implicit in these Reviews : that progress in the work of the 
Institution, as elsewhere, represents a combination of the old and the new, 
of the emergent problem and the matured field. 

In a delightfully worded and penetrating Introduction to his report this 
year Alfred D. Hershey, Director of the Genetics Research Unit, aptly 
characterizes the essence of most progress in science. He remarks that 
"discovery and description of natural phenomena are ends in themselves" 
to practicing scientists. "One notices a phenomenon that looks interesting. 
After suitable thought and labor, one performs an experiment that ought to 
be instructive. What it actually does is to turn up a second phenomenon not 
related in a simple way to the first. This sequence of events makes for lively 
research. It does not provide explanations ... we have to consider the 
possibility that (this) sequence . . . may prove endless, giving an indefinitely 
complicated description of nature. . . . This seems to be the state of affairs 
in research on the structure of matter at both ends of the scale of magni- 
tude. For somewhat similar reasons, what goes on in atomic nuclei, outer 
space, and living cells might prove equally refractory to explanation." That 
is, the new in science does not often produce broad generalizing explanations. 
It only gives a slightly different view of the infinite than was there before. 
A number of such views are provided by the Institution's work in astronomy 
and astrophysics, nuclear physics, geophysics and geochemistry, and bio- 
physics, biochemistry, and biology during a very rewarding year. 

Astronomy and Astrophysics 

Astronomy surely ranks among the most ancient of all scientific subjects. 
At the same time it is a scientific subject for which there is as yet not even 
a theoretical end to the imagined. 

The year's work at the Mount Wilson and Palomar Observatories can 
once again be summarized in categories that have become familiar to 
readers of these Reviews. 

Work in one of these categories, that of solar astronomy, reaches back 
close to the origin of the Mount Wilson Observatory. Indeed the founder of 
that Observatory, the late George Ellery Hale, selected Mount Wilson as 
its site at least in part because of the peculiar advantages it offered for solar 
observation. Solar study has continued without interruption at the Mount 
Wilson Observatory ever since 1904. The sun's magnetic properties have 
been a subject of investigation since 1908, when Dr. Hale discovered the 
magnetic field of sunspots. Solar physics has been a major research interest 
of Horace W. Babcock, present Director of the Observatories, who was the 
co-inventor of the now widely used solar magnetograph. Robert Howard, 



REPORT OF THE PRESIDENT 39 

Harold Zirin, and their colleagues thus continue a program which now has 
more than half a century behind it. 

Study of the magnetic properties of the sun continues to be an important 
part of the solar program today. Solar magnetograms, which have been a 
feature of the Observatories' operations since 1953, now are reduced by 
computer analysis. They provide the basis for calculations of the integrated 
magnetic-field strength and fluxes for each day's observation. The magneto- 
graph may also be used in the "velocity mode" for producing "Doppler- 
grams," or maps of differential velocity patterns on the sun. Studies of the 
differential solar rotation with the aid of isotachial 1 plots have been initiated. 
The first results show no large-scale meridional currents. 

In collaboration with V. Bumba of the Ondrejov Observatory of Czecho- 
slovakia, Howard found that low-latitude patterns of magnetic fields on 
the sun persist for periods as long as one or two years with rotation rates 
of about 27 days. The 27-day period of these long-lived magnetic "streams" 
on the sun undoubtedly is connected with recurrent magnetic disturbances 
in the vicinity of the earth. 

Howard and J. M. Wilcox, of the University of California, Berkeley, 
have studied the relation of the interplanetary magnetic field with that of 
the sun's photosphere. The solar field is found to have a pattern different 
from that which may be expected from the usual solar differential rotation. 
For example, the rotationally most advanced part of the field is at 15° north 
latitude, not at the equator, as would be projected in differential rotation 
in the photosphere. 

Study of the solar events of another year produced one of the more 
interesting insights on the solar process. H. Zirin and Mrs. D. R. Lackner, 
a NASA postdoctoral fellow, studied in detail the records of the great 
flare of August 28, 1966, one of the largest proton flares observed in the 
present solar cycle. This flare appeared qualitatively different from a 
normal solar flare in the very large size of the area over which energy was 
released, as though a chain reaction of energy release had occurred. From 
the observed area of the flare and assumptions as to its height, it was 
possible to make models of the emitting material. The X-ray emission and 
microwave bursts traceable to the flare can only be explained by assuming 
that they occurred at temperatures greater than a billion degrees. 

Laboratory study also provided the first convincing identifications from 
an astrophysical source of the "forbidden" lines of neutral sulfur in the 
Fraunhofer spectrum, whose presence was first suggested in 1948 by I. S. 
Bowen. D. L. Lambert, in collaboration with J. P. Swings and N. Grevesse 
of the University of Liege, Belgium, has provided an accurate wavelength 

1 An isotach is a contour line denoting equal velocity. 



Ifi CARNEGIE INSTITUTION 

for two neutral sulfur forbidden lines. The observed intensities in the solar 
spectrum are in good agreement with predictions based upon theory. 

Pulsars. The most remarkable astronomical discovery announced during 
the year was that of the rapidly pulsating radio source in the constellation 
Vulpecula, announced by A. Hewish and his collaborators at Cambridge 
University in February 1968. 2 This discovery brought to astronomers' 
attention a whole new class of objects with very unusual properties. The 
most remarkable of these is the radiation of brief pulses of radio energy 
with a period of extremely high constancy. Thus the name "pulsar." The 
pulse recurrence interval for this first pulsar was measured at 1.3372795 
seconds by F. D. Drake and collaborators at the Arecibo Observatory in 
Puerto Rico. 3 

Soon after announcement of the first pulsars, H. Arp, A. Sandage, J. 
Kristian, and others of the Mount Wilson and Palomar Observatories at- 
tempted to identify them optically. A major part of the search has centered 
on the field of CP 1919, the first pulsar with an accurately determined radio 
position. The fields of sources CP 0950 and CP 1133 also were examined, 
but with negative results. An object of magnitude 17.5, apparently a star, 
in the field of the pulsar CP 1919 was searched for periodic fluctuations in 
brightness with the aid of a signal-averaging computer. No periodicity in the 
light was detected. However, measurements and detailed analysis of pulsar 
data are continuing at the Observatories because of the obvious interest 
in the physics of pulsars. Further, it is thought that more precise measure- 
ment of astronomical distances than ever before achieved may be possible, 
as well as the establishment of the plasma distribution in the galaxy with 
greater accuracy and in more detail. 4 

Quasi-stellar Sources. An equally if not more remarkable class of objects 
is that of the quasi-stellar source, first identified optically at the Mount 
Wilson and Palomar Observatories in 1962. They continue to present some 
of the more difficult problems encountered in the recent history of astronomy 
and have aroused at least one lively controversy as to their characteristics 
and position. This controversy centers on the distance of the quasars. Are 
some or all of them "local," that is, within or relatively near to our 
Galaxy? Or are they at cosmological distances, as the large redshifts of many 
quasars indicate? Members of the staff of the Observatories believe they 
have adduced convincing evidence during the year that the quasars are 
not "local." 

2 A. Hewish, S. J. Bell, J. D. H. Pilkington, P. F. Scott, and R. A. Collins, "Observation of a 
Rapidly Pulsating Radio Source," Nature, 217, 1968, p. 709. 

3 F. D. Drake, E. J. Gunbermann, D. L. Jauncey, J. M. Cornelia, G. A. Zeissig, H. D. Craft, Jr., 
"The Rapidly Pulsating Radio Source in Vulpecula," Science, 3 May 1968, p. 503. 

4 Drake, et al., op. cit., p. 507. 



REPORT OF THE PRESIDENT Jf.1 

Quasi-stellar sources may be divided among radio-emitting sources and 
radio-quiet objects. All are optically blue. A. Sandage completed during 
the year photoelectric photometry of 300 blue objects and concluded that a 
major change in the nature of blue objects occurs at the visual magnitude 
designated 15. Radio-quiet quasars begin to occur at about this magnitude 
and increase in frequency with increasingly fainter apparent magnitudes. 

Study of the quasar component of the blue objects investigated became 
sufficiently precise during the year to afford a first tentative estimate of the 
number of radio-quiet quasars in the sky. On the basis of Sandage's photo- 
electric photometry and spectrometry by A. Braccesi and R. Lynds, an 
estimate of 0.4 QSOs per square degree was made to blue magnitude 17.1, 5 
per square degree to magnitude 18.5, and 100 per square degree to 21.5. 
A quantitative growth relation can be calculated from the observed distri- 
bution by magnitude. Although the determination of the growth function 
for the radio-quiet quasars is tentative, Sandage and W. J. Luyten, of the 
University of Minnesota, are convinced that the number of such objects 
is very large. The suggested function for the growth relationship results 
in an estimate of 10 million quasars over the entire sky to a faintness of 
magnitude 22. Sandage believes that such a large number of objects having 
either a constant space density or a positive density gradient outward, 
virtually eliminates the "local" hypothesis for quasar distances. 

An important distinguishing feature of quasi-stellar sources is their 
optical variation with time. Rapid fluctuations of nonthermal radiation 
(continuous emission not attributable to stars) in quasars over periods of 
days is cited by some astronomers in support of a contention that quasars 
are "local." It is also suggested that there may be another explanation for 
the observed redshift. 

To examine the validity of this position a search was started for nonther- 
mal radiation in objects whose cosmological distances are not disputed. 
A kind of galaxy known as the N-type displays nonthermal radiation, and 
has been shown by Sandage to lie at cosmological distances. 

Sandage and J. B. Oke found that the N-type galaxy 3C 371 varied 
optically by nearly a factor of 2 in intensity between June 1966 and Sep- 
tember 1967. Oke found variations of 10 to 15 percent within only a few 
days. Later observations by Sandage on other N-type galaxies also revealed 
significant variations in brightness. It was concluded that large variations 
of radiation within a matter of days cannot in themselves contradict the 
cosmological distance scale of quasars. 

M. Schmidt, observing objects identified by E. T. Olsen of the University 
of Michigan, continued spectroscopic observation of quasi-stellar radio 
sources. A most interesting discovery was the quasi-stellar source 4C 25.5, 
identified by Olsen as a blue stellar object of visual magnitude about 17.5. 



42 CABNEGIE INSTITUTION 

It has an emission-line redshift of 2.358, the largest yet recorded. If the 
redshift is an indicator of relative distance from the earth, once again we 
may record the observation of an object still farther from the earth than 
the farthest known the year before. 

Compact Galaxies and Groups of Galaxies. Further evidence of the enor- 
mous variety and great richness of the sky in classes of astronomical objects 
is given in investigations of compact galaxies and chain galaxies. F. Zwicky, 
Arp, Oke, W. L. W. Sargent, and G. Neugebauer have made spectroscopic 
investigations of several hundred compact galaxies. They have found that 
among the compact galaxies are objects that match the spectra, colors, and 
absolute magnitudes of other extragalactic objects originally considered 
unique, like the quasars. Observations by Zwicky, Arp, Sargent, and T. 
Fairall demonstrate that there exist compact galaxies that form a con- 
tinuous link between ordinary galaxies and quasars. One of these, the "iron 
galaxy" (I Zw 0051 = 12), with an absolute magnitude of —23.1, is con- 
sidered to be as luminous as some quasi-stellar sources. 

Compact galaxies in some cases have been found in clusters which contain 
dozens to hundreds of objects. Three of the more remarkable of these 
clusters, known by their numbers in the Zwicky list, (Zw CI) 1710, 1849, and 
0658, contain respectively 345, 324, and 189 galaxies each. A feature of 
these clusters appears to be a large velocity dispersion. Zw CI 1710, for 
example, has two galaxies with individual velocities of recession of 23,820 
kilometers per second and 21,440 kilometers per second. 

Other clusters of galaxies contain fewer members. Sargent began a study 
of motions of galaxies in small groups. One of the clusters, VV 172, 5 contains 
five galaxies in an almost linear chain, isolated from others of similar bright- 
ness (Plate 1). The mass of the whole system is equivalent to at least 400 
billion solar masses. Image-tube spectra obtained of each galaxy in the 
chain show redshifts of approximately 16,000 kilometers per second for four 
of the galaxies, but one of almost 37,000 kilometers per second for the fifth. 
The four galaxies have a systematic trend of redshift indicating their 
rotation as a system. The anomalous fifth must be either a background 
galaxy, or one that was expelled from the chain with tremendous energy 
about 100 million years ago. VV 172, incidentally, is not unique. Two other 
galactic clusters, Stephan's Quintet and the triple system IC 3481, also 
have an anomalous galaxy, but in these the redshift is much smaller than 
in the remainder of the group. VV 172 is the first to be found with an 
anomalous galaxy apparently receding at such a rapid rate from its system. 

An Unusual Near Neighbor in Space. Arp and J. Greenstein have made 
some observations of the star Wolf 359, one of the nearest neighbors to the 

5 From the discoverer, B. A. Vorontsov-Velyaminov, whose study of the National Geographie- 
Palomar Observatory Sky Survey identified a number of unusual astronomical objects. 



Plate 1 



Report oj the President 




Plate 1. The chain of five galaxies known as VV 172. The second galaxy from the top of the 
chain has a redshift more than twice that of any one of the other four galaxies. (From H. Arp, 
Atlas of Peculiar Galaxies, Pasadena, 1966: No. 329.) 



REPORT OF THE PRESIDENT 43 

earth, with the image-tube spectrograph. On the very first observation they 
found this very faint star to have a flare. Wolf 359 is a member of what is 
called the "old disk population" of stars in the Galaxy. Such strong 
chromospheric activity is unexpected among these stars, and particularly 
in one of such low luminosity. An additional unusual feature was discovered 
in observations by Neugebauer and E. Becklin which showed unusual 
brightness in the infrared. The star has an apparent visual magnitude of 
13.3, but a magnitude of 5.8 at the 3.4-micron wavelength. The infrared 
brightness may be explained by low temperature, but the temperature 
would have to lie far below that expected for any of the "main-sequence" 
stars, where Wolf 359 was thought to fit. These observations suggest that a 
part of the luminosity-temperature function generally accepted by astron- 
omers may be subject to reinterpretation. 

Radio Observations in the Southern Hemisphere. A significant tentative 
conclusion about the Magellanic Clouds, our Galaxy's nearest neighbors, 
was derived from southern hemisphere radio observations. Aided by the 
installation of a new low-noise parametric amplifier and a complete digital 
data acquisition system, the recently established Carnegie-Argentina 
Radio Astronomy Station at La Plata continued investigation of the dynamic 
history of the system comprising our Galaxy and the two Magellanic Clouds. 
Using equations of motion, observed radio velocities, and assumed velocities 
perpendicular to the line of sight, computer studies of the possible history 
of the Galaxy and the Magellanic Clouds were undertaken. Two conclusions 
may be drawn. One is that the radial velocities of the material between the 
two clouds are consistent with an explosive origin for the two clouds. The 
second is that the Magellanic Clouds are probably not an obligate twin 
system, but are most likely independent celestial bodies passing close to 
one another. K. C. Turner, of the Department of Terrestrial Magnetism, 
concludes "there seem to be no considerations other than aesthetic ones 
which require the clouds to be permanent pair." 

Refining the Tools for Analyzing the Cosmos. One of the most fascinating 
questions in all astronomy, if not in all science, is that involving the history 
and processes of the cosmos. As W. H. McCrea has observed, modern 
cosmology is in a supremely interesting state. 6 The original static model of 
the universe proposed by Einstein in 1917 has long since been superseded, 
and the more recent steady-state model of Bondi, Gold, and Hoyle has been 
called into serious question. There remains the "big-bang" hypothesis, 
whose main predictions "appear to have been almost suddenly found to be 
fulfilled." 7 However, in spite of the satisfactory fit to the big-bang model that 
large-scale computing seems to be giving recent observations, McCrea notes 

6 W. H. McCrea, "Cosmology after Half a Century," Science, 21 June 1968, p. 1295. 

7 Ibid. 



44 CARNEGIK INSTITUTION 

that there still remain both empirical and conceptual difficulties about the 
acceptance of such a model. 8 

One other possible interpretation is that of an oscillating universe, in a 
sense a variant of the big-bang model. The question of choice among alter- 
native models still remains. The concept of a "fireball" (or "big-bang") 
origin for the observed universe seems to be generally accepted, although 
the date of this epochal event is still uncertain. 

The key to a decision among known models and to elaboration of the 
features of whatever cosmological model is chosen lies in refinement of the 
parameters H , qo, and a used in its construction. These are, respectively, 
Hubble's constant (ratio of speed of recession to distance), the deceleration 
parameter, and the density parameter. 

Sandage developed during the year a new method of determining the 
Hubble constant. By calculating the apparent magnitude-redshift relation 
for the brightest elliptical galaxies in clusters at distances lying beyond the 
effects of "local" inhomogeneities in our part of the universe, he has arrived 
at a determination of the constant that narrows considerably the range 
between possible upper and lower limits. These he states to be 75 and 50 
kilometers per second megaparsec. 9 The reciprocal of the constant may be 
stated in years. Taken together with the deceleration parameter (q ) of +1, 
the new determination gives an estimate of the time since the "fireball" 
of the big-bang as about 11 billion years. Sandage notes that such an esti- 
mate lies within the range of the estimated age of globular clusters of stars 
(about 10 billion years) and the presently determined age of the heavy 
chemical elements (greater than 6.6 billion years) that make up the earth's 
crust. 

The value of the Hubble constant is of such obvious importance to 
cosmology that Sandage, in collaboration with J. Kristian, has already 
begun a check on the new method by making a second and different type 
of determination, in which the angular sizes of ionized hydrogen patches 
in nearby spiral (Sc) galaxies are measured on high-quality photographs 
made with the 200-inch telescope. 

In collaboration with J. B. Oke, Sandage has also attacked the problem 
of defining more accurately the deceleration parameter (q ). The problem 
is to determine the correction necessary for measured intensities of red- 
shifted galaxies to compensate for the redshifting of the continuum energy 
distribution through the measuring bands. As part of this determination 
Oke and Sandage examined the energy distribution in eight giant E (ellipti- 
cal) galaxies. They found it to be identical for all eight of the galaxies, thus 
supporting the view that elliptical systems are similar in their stellar content. 

8 Ibid., pp. 1297-1299. 

9 One megaparsec equals 3.26 million light years. 



REPORT OF THE PRESIDENT J^5 

These data were compared with new photoelectric measurements for 
galaxies in 32 clusters. The function predicted from the study of the giant 
galaxies agreed with the photoelectric measurements over the entire range 
of redshifts with a probable error of less than ± 0.02 magnitude. Such agree- 
ment, they say, shows that no significant change in energy distribution has 
occurred through evolution of the stellar content of the E galaxies during a 
period of 2.5 billion years. The results are consistent with a new estimate by 
Sandage that the change in magnitude over time in such systems is less 
than 0.05 magnitude per billion years, as determined from study of the 
evolutionary pattern of relevant stars. 

The ultimate purpose of the energy distribution corrections is a decision 
between a "big-bang" (ever-expanding) model of the universe, and an 
oscillating model. The new data had been given a preliminary analysis by 
Sandage and J. D. Peach. They report that the data favor the oscillating 
model, and suggest a total cycle of time between fireballs of about 80 
billion years. 

Improved Instrumentation in Astronomy. In keeping with a long tradition 
the scientists of the Institution have continued to be alert to possible im- 
provements in the instrumentation of astronomical observation. In these 
days, when the pressure for observing time generated by the increasing 
number of astronomers has intensified greatly, the potential contributions 
of improved instrumentation are enormously greater than ever before. At 
least two of the activities of the Mount Wilson and Palomar Observatories, 
extension of photoelectric data systems using digital recording, and the 
Carnegie Southern Observatory, bear special mention. So also does the 
continued work on image tubes for telescopes, a particular interest of the 
Department of Terrestrial Magnestism to which the Observatories also 
have contributed. 

One of the longest continued single projects in the Institution's history 
has been that of the Committee on Image Tubes for Telescopes. The Com- 
mittee was started at the Department of Terrestrial Magnetism in February 
1954 to investigate the possible use of electronic techniques to supplant or 
supplement the prevailing photographic techniques of recording observa- 
tions from optical telescopes. First supported by a grant from the Carnegie 
Corporation of New York and then by grants from the National Science 
Foundation, the Committee, with the cooperation of the Radio Corporation 
of America and other industrial laboratories, has had outstanding success 
in developing a satisfactory image tube intensifier for optical telescopes. 
The first successful two-stage cascaded tubes, developed in the RCA 
laboratories, became available at the end of 1963. Since that time 35 tubes 
have been distributed by the Committee and the National Science Founda- 
tion to astronomers on a priority basis, 14 of them within the last year. 



46 CARNEGIE INSTITUTION 

Meanwhile, further technical development of the tubes and associated 
optics or other instrumentation has continued. The RCA laboratory now 
has under development an improved version of the image tube that promises 
to increase the image resolution obtained on tubes of the present design. 

One notable event during the year was the first use of the image tube on 
the 200-inch Palomar telescope as part of a new Cassegrain image tube 
spectrograph designed by I. S. Bowen. For equivalent resolution, the new 
spectrograph gave a speed gain of about 15, measured in the blue part of 
the spectrum, as compared to the prior "unaided" instrument. The use of 
this spectrograph thus can substantially improve the effectiveness of the 
200-inch by permitting the scheduling of more observations. 

A most important advance in the mode of operating the three major 
telescopes of the Observatories has been designed and gradually installed 
over the last few years. Photoelectric data systems using digital recording 
now take care of a majority of the observing nights on the 60-inch telescope 
(57 percent) and on the 100-inch (68 percent), and now are used for nearly 
a third of the nights on the 200-inch. The tedious and time-consuming task 
of installing digital data cables and many coaxial cables on the 200-inch 
telescope was essentially completed during the year. About 120 digital con- 
ductors now link that telescope with the data room. An additional 56 cables 
were installed to operate a multichannel spectrum scanner at the Cassegrain 
focus. 

The data systems already have contributed notably to efficiency and 
flexibility in the use of the telescopes. For the future, use of integrated 
circuit modules, high-capacity, compact, low-cost computers, and direct 
lines to large computers at distant laboratories offer proven possibilities 
for further steps towards more effective use of these valuable instruments. 

A third important step towards effective instrumentation in optical 
astronomy was taken by the Institution's Board of Trustees during the 
year in approving the establishment of an optical observatory to be located 
in central Chile and administered by the Mount Wilson and Palomar 
Observatories. Thus the first fruits of the careful site survey investigations 
that have been conducted for several years past are now in sight. Engineer- 
ing activities are now proceeding. 

Nuclear Physics 

Nuclear physics has held the attention of some Institution scientists since 
the early 1920s. Ellis T. Bolton, Director of the Department of Terrestrial 
Magnetism, relates in the Introduction of his report the story of one exciting 
event in the history of that interest, when Richard Roberts, Lawrence 
Hafstad, and Robert Meyer gave one of the early demonstrations of nuclear 



REPORT OF THE PRESIDENT Jfl 

fission by bombarding uranium 238 with neutrons from lithium and deu- 
terium in the Department's Van de Graaff accelerator. The successful 
demonstration was conducted in the presence of Niels Bohr, Enrico Fermi, 
Gregory Breit, L. Rosenfeld, and Edward Teller. Although it was unknown 
at the time, prior demonstrations of fission had been conducted 12 days 
before by Frisch in Copenhagen, and three days previously by Dunning 
and his colleagues at Columbia University. Bohr witnessed the Roberts- 
Hafstad-Meyer experiment without knowing that Frisch had already 
successfully completed his experiment. 

Within the last few years interest in nuclear physics has revived at the 
Department of Terrestrial Magnetism. Descriptions will be found in the 
Year Book of experiments conducted during the past year by L. Brown, 
C. Petitjean, and W. Trachslin on the lithium-helium transmutations 
[ 7 Li(p,o;) 4 He and 6 Li(p, 3 He) 4 He] with the use of a polarized proton beam. 
The reactions studied proved to be strongly sensitive to polarization. 
Brown, Petitjean, and Trachslin note that the energy range shown in the 
lithium-helium transmutation experiments "covers a region where the 
structure of the intermediate nuclei 8 Be and 7 Be is still uncertain after 30 
years of research with unpolarized protons." Thus it appears possible that 
atomic nuclei as complicated as that of the beryllium atom could be de- 
scribed in terms not much more complicated than for the helium atom. 

Another finding of interest to these investigators concerned an earlier 
helium-to-helium experiment {Year Book 66). R. G. Satchler, of Oak 
Ridge National Laboratory, has found that the DTM measurements of 
4 He(p,p) 4 He fit a concept of nuclear structure known as the optical model, 
which hitherto has applied only to nuclear scattering at high energies on 
medium and heavy nuclei. The application of the optical model to a light 
element is significant, explain Brown, Petitjean, and Trachslin, because 
"The fit of theory to data . . . had to be made with a reduced number of 
parameters, since the usual complex potentials 10 must be replaced with real 
ones because no inelastic processes 11 are possible." The possibility definitely 
is opened that concepts of nuclear structure can be somewhat simplified, 
permitting a choice among the known theoretical alternatives, and possibly 
a single system of statistical probabilities. 

Geophysics and Geochemistry 

Research in geophysics has as long and as distinguished a record in the 
Institution as research in astronomy. Plans for a geophysical laboratory and 
an "International Magnetic Bureau" were presented in the second Year 

10 Of overlapping levels in the nucleus. 

11 An inelastic process is one in which the bombarded nucleus is left in an "excited" state. 



48 CAENBGIE INSTITUTION 

Book (1903), and the "Department of International Research in Terrestrial 
Magnetism" commenced operation on April 1, 1904. The geophysical pro- 
gram of the Institution has continued uninterrupted ever since, through its 
two departments, the Geophysical Laboratory, and the Department of 
Terrestrial Magnetism. The principal objective of our geophysical research 
is knowledge of the history and composition of the earth — the enormously 
complex universe of rock and molten material lying under us. 

The investigative programs themselves are indicative of the complexity 
of the subject with which they are concerned. The search for understanding 
of the earth's properties and processes in the 1967-1968 programs included 
identification of the mineral components of the earth; identification, descrip- 
tion, and laboratory reproduction of the processes of rock formation, energy 
flow, and structural formation; and the relations of the planet to its im- 
mediate environment in space. 

Identification of New Minerals and Other Components of the Earth's Crust. 
Nothing illustrates better the earth's complexity than the number of known 
minerals found in its crust. More than 2500 minerals already are known. 
To the layman it may come as a mild surprise that today, after thousands 
of years of mineral identification, new minerals are being discovered. Indeed, 
three of them were discovered or identified at the Geophysical Laboratory 
during the year. 

A. El Goresy discovered a new allotropic form of carbon, which he and 
Gabrielle Donnay studied by means of X rays. There are three other known 
forms, diamond, graphite, and lonsdaleite. The newly discovered mineral 
has a hexagonal form, a metallic gray-to-white reflection color, and is slightly 
harder than graphite. It was found in polished sections of "shocked" graphite 
gneisses from the Ries meteorite crater in Germany. It is to be called chaoite 
in honor of Dr. Edward Chao of the United States Geological Survey. 

The second new mineral was discovered by Gabrielle Donnay. It has been 
named ewaldite, and is a complex sodium-calcium-barium carbonate found 
in three-dimensional intergrowth with a previously known mineral, 
mckelveyite, from the Green River formation of the United States. 

The third new mineral, now called sonoraite, was discovered by Richard 
R. V. Gaines, of the Universidad Nacional Aut6nomo de Mexico, in the 
Moctezuma tellurium-gold mine in the state of Sonora, Mexico. It is an iron 
tellurite occurring as highly lustrous, yellowish-green rosettes. The crystal 
structure was determined by Gabrielle Donnay in cooperation with J. 
Zemann of the University of Vienna, and J. M. Stewart of the University of 
Maryland, and a chemical analysis was made by M. H. Hey of the British 
Museum. 

One of the more intriguing problems in the identification of rock com- 
ponents concerns evidences of life in the ancient rocks. For several years 



REPORT OF THE PRESIDENT Ifi 

P. H. Abelson and his colleagues at the Geophysical Laboratory have been 
seeking these evidences. They have been particularly interested in the 
fatty acid and amino acid traces to be found in sedimentary rocks and their 
relations to modern fatty acids and amino acids. Problems lie not only in the 
identification and measurement of life-derived materials, but also in detecting 
contamination of ancient rocks by more recent additions. 

The contamination problem was illustrated in an analysis made this year 
of the amino acid content of the 1.9 billion-year-old Gunflint iron formation 
of Ontario, where microfossils embedded in chert were discovered about 
15 years ago. Other investigators have reported finding amino acids in the 
chert that they believed to have been residual in it for 1.9 billion years. 
However, because of the conditions under which the chert specimens were 
found, and the presence of specific amino acids not considered stable over 
time, Abelson and Hare replicated the experiments. They concluded from 
a number of analytical results, such as the observed ratio of alloisoleucine 
to isoleucine, that the amino acids found in the chert are of recent origin. 

The organic geochemistry group of the Geophysical Laboratory made 
significant progress during the year in identifying the products with life 
origins that are stable under geologic conditions. The group also developed 
further the laboratory-simulated production of these compounds. The com- 
bination of gas-liquid chromatography with a mass spectrometer, used by 
T. C. Hoering for analysis of molecular structure, aided much in gaining 
new understanding. As Dr. Abelson explains in his Introduction to the 
Geophysical Laboratory report, this combination gives unique information 
on molecular structure even when only tiny samples are available. 

Using these methods, Hoering analyzed samples of marine sediment from 
the continental shelf off Southern California for fatty acid content and 
found a surprisingly large content of the branched-chain acids, particularly 
of the isoprenoid type. Such acids, although known in lipides, are usually 
considered to be relatively rare. 

In separate experiments Hoering, P. E. Hare, and R. M. Mitterer suc- 
ceeded in reproducing in the laboratory the pattern of amino acids and 
other life compounds to be found in ancient sediments. Hoering's experi- 
ments were a continuation of a set started last year on the application of 
elevated temperatures to a recent marine sediment. Analyzing the porphy- 
rins that resulted from this treatment, he found the distribution of molec- 
ular weights very similar to that in materials separated from ancient 
sediments. His data suggest that alkylation and transalkylation may be 
important organic reactions occurring in sedimentary rocks in addition to 
hydrogenation, dehydrogenation, cyclization, and hydrolysis. 

Hare and Mitterer experimented with the incubation of fragments of 
modern moluscan shells at temperatures from 185° to 90°C for periods from 



50 



CARNEGIE INSTITUTION 



one day to three months in varying environments. Their experimental 
procedure produced changes in the amino acid content almost identical with 
those seen in fossils (Fig. 1). The detailed data on changes under different 
conditions over time suggested potential uses for amino acid ratios in fossil 
study. If the temperature is known, an approximate age can be calculated. 
On the other hand, incubation temperature may be inferred for material 
whose date of origin is known. Such knowledge could be used in stratigraphic 
correlation, as in connection with deep-sea sediment cores. 

Identification and Reproduction of Rock-Forming Processes. Much of the 
research interest of the Geophysical Laboratory now centers on studies of 
phase relations and other mineralogical attributes that may give clues 
about the composition and structure of the crust, mantle, and core of the 
earth. They are directed towards not only information concerning the 
composition of the earth in all its complexities, but also an understanding 
of how it became what it is. Laboratory experiment has now become a 
prime source of such information, but the older tradition of field study for 
understanding the earth by no means has been abandoned. 




Temperature, °C 



Fig. 1. Comparison of reaction times for laboratory-observed conversion of isoleucine to allioiso- 
leucine (lower right) with radio carbon-dated fossil shell materials (upper end of solid line), 
materials from an Antarctic deep-sea core (A), and Upper Pleistocene materials from Wailes Bluff, 
Maryland, locality. The plot shown is for an Arrhenius equation that relates the rate constant of 
a reaction to the absolute temperature. The data obtained by Hare and Mitterer in their laboratory 
experiments show that extrapolation of laboratory-observed values to fossil specimens may be 
possible. 



REPORT OF THE PRESIDENT 

1700 1 1 



51 



1600 



1000 



30 kb 




100 95 90 85 80 *75 l 70 65 



Fe 



Weight per cent Fe 



FeS 



Fig. 2. Temperature-composition section for 
the Fe-FeS system at 30 kb pressure. Increasing 
sulfur content of a mixture of iron and sulfur 
shown with distance to the right on the diagram. 
The eutectic temperature (minimum melting 
temperature) for the liquid is near 1000°C. The 
same melting temperature has been observed at 
1 atmosphere of pressure, but 3.5 percent more 
sulfur is required to reach the eutectic tempera- 
ture. This observation suggests that even a 
small amount of sulfur at the earth's core may 
have a significant influence on its melting 
temperature. 

From a series of experiments conducted by P. R. Brett of the U. S. Geo- 
logical Survey and P. M. Bell of the Geophysical Laboratory, an interesting 
inference can be made about the temperature of the earth's core. It has 
been known for some time that a mixture of iron and sulfur had a lower 
melting temperature than iron alone under atmospheric conditions. Brett 
and Bell experimented with the Fe-FeS system under 30 kb of pressure. 
The minimum melting (eutectic) temperature was found to be almost 
exactly the same at 30 kb as at one atmosphere (about 990°C). However, 
the amount of sulfur needed to reach the eutectic temperature in the mix- 
ture was 3.5 percent less at 30 kb than under one atmosphere (Fig. 2). 
These results suggest that even a small amount of sulfur, if present in the 
earth's core, would lower the melting temperature of the iron there signifi- 
cantly. The difference between the melting temperature of 100 percent 
iron and the eutectic iron-sulfur combination is more than 600°C. 

A unique and imaginative method of studying deep-lying rocks is de- 
scribed in the report of H. O. A. Meyer and F. R. Boyd on mineral inclusions 
in diamonds. The high temperatures and pressure required for the formation 
of diamonds under natural conditions are such that it seems most likely 



52 CARNEGIE INSTITUTION 

that diamonds have come from depths below 100 kilometers, that is, from 
far below the base of the earth's crust. The chemical compositions of other 
minerals sometimes included within natural diamonds were determined 
quantitatively for the first time by Meyer and Boyd with an electron probe. 
The inclusions are far too small to have been studied by the classic methods 
of analytical chemistry. The olivine, garnet, chromite, and diopside in- 
clusions within the diamonds studied were found to be grossly similar to 
those from ultramafic nodules in the kimberlite where the diamonds occur. 
All of the inclusions were discovered to be monomineralic (i.e., one mineral 
per inclusion), suggesting that the diamonds were formed by a solid-state 
process rather than by crystallization from a magma. Another interesting 
observation is that the garnet and chromite inclusions were unusually rich 
in chromium, whereas a single diopside inclusion was found to contain un- 
usually little chromium. Furthermore, other characteristics of the diopsides 
imply a relatively low temperature for the formation of the diamonds, 
probably 950°C. Boyd observes that this temperature is low in comparison 
to the range in which diamonds have been synthesized in the laboratory. 

The distribution of water in the earth's crust and mantle and its function 
in rock-forming processes have long been subjects of geophysical and geo- 
chemical study. Theories of the genesis of magmas must make allowances 
for the manner in which water is distributed within the earth's mantle. 
The occurrence of water within the deep mantle is a problem of special 
interest. The minerals known as amphiboles, hydrous compounds stable 
at temperatures up to 1000°C at low pressure, have been proposed as mineral 
phases that could hold water deep within the mantle. M. C. Gilbert, a 
Fellow at the Geophysical Laboratory, experimented with the pressure 
effect on amphibole stability, and showed that calcic amphiboles are not 
likely to be stable at depths greater than 70 to 100 km. Gilbert concludes 
that amphiboles may be significant in the petrology of the lower crust and 
upper mantle, but cannot be the water bearers at greater depths. One more 
constraint upon the composition of the mantle thus has been discovered. 

The effects of water comprised an important part of another Geophysical 
Laboratory study on the genesis of basalt magmas — very important sources 
of rocks in the earth's crust. H. S. Yoder, Jr., and I. Kushiro studied the 
melting relations of enstatite (MgSi0 3 ) and forsterite (Mg 2 Si0 4 ) in the 
presence of water. Their results suggest that silica-saturated magmas can 
be formed in the presence of water at considerable depths in the upper 
mantle by partial melting of enstatite or fractional crystallization. These 
results are considered significant indicators of the origins of silica-rich 
magmas, such as the andesitic magmas in mountain-building zones, thought 
to be formed by fractional crystallization of basalt magmas at high water 
pressures. 



REPORT OF THE PRESIDENT 58 

An interesting tool of theoretical significance has been developed by W. B. 
Bryan of the Geophysical Laboratory. The study of an intrusive igneous 
rock suite evolved from a magma, or possibly the mixing of two or more 
magmas, involves complex relations in crystal-melt equilibria, assimilation, 
etc. Bryan has written a computer program that will generate an estimate 
of the composition of the magma at any arbitrarily chosen stage of evolution. 
A large number of alternative hypotheses may thus be tested rapidly, as 
well as the capacity of an inferred mechanism to account for observed 
trends in the evolution. 

Studies of Meteorites and Meteorite Effects. Several interesting studies of 
meteorites and their impact were undertaken during the year at the Geo- 
physical Laboratory. P. M. Bell and E. Chao have experimented for the 
first time with the formation and annealing of synthetic dense glasses 
composed of feldspars. Such glasses are observed at the surface of the earth 
in the impact zones surrounding meteor craters. They are formed by the 
high-pressure shock-wave that accompanies the meteoric impact, and they 
are extraordinarily dense. A comparison of such impact products with the 
products of other explosive events, such as volcanic eruptions, is expected 
to be useful in the analysis of samples from the moon or other planets when 
they become available. 

G. Kullerud and A. El Goresy have explained one of the anomalies of 
meteorite mineralogy as compared to terrestrial assemblages: the lack of 
equilibrium between minerals in the same assemblage. The lack of equilib- 
rium need not indicate a lack of differentiation in meteoritic bodies, as 
previously believed, but can be more readily explained as the effect of shock 
in at least some instances. 

Geophysical Field Studies. Although laboratory experiment has become 
the dominant interest of geophysical research workers in the Institution, the 
long tradition of field observation that began with the Department of 
International Research in Terrestrial Magnetism in 1904 continues as a 
highly useful approach to new knowledge about the interior of the earth. 
The program of seismological observations of the Department of Terrestrial 
Magnetism depends upon field operation. An unusual geophysical field 
investigation of heat flow from the earth's interior was started in 1966 and 
completed this year. 

J. S. Steinhart, S. R. Hart, and T. J. Smith of the Department of Terres- 
trial Magnetism report this year on their regional survey of heat flow from 
the earth's crust as measured from the bottom of Lake Superior. Working 
from the U. S. Coast Guard Cutter Woodrush, measurements were recorded 
in 1967 of sediment temperature gradients and thermal conductivities of 
these sediments at 95 heat-flow sites. The data made available from these 
measurements made possible the compilation of a regional heat-flow map of 



5 Jf CARNEGIE INSTITUTION 

Lake Superior (Fig. 3). Steinhart and his colleagues describe it as a "signif- 
icant advance in our understanding of regional heat-flow patterns." 

The range of heat-flow values found by Steinhart, Hart, and Smith is 
surprisingly large, varying by a factor of almost three, in some places 
within very short distances. In general the variations of heat flow in the 
Lake correlate rather well with the known geologic, gravity, and magnetic 
features of the Lake bottom and surrounding shores. Volcanic rock areas, 
for example, are highly magnetic, relatively dense, and have a low heat flow 
(being low in radioactivity). Using gravity, magnetic, and heat-flow data, 
Steinhart, Hart, and Smith have constructed a "near-surface" model of 
the geologic structure of the Lake Superior basin. The model shows rela- 
tively thick (10-kilometer) sections of nonmagnetic, moderately dense, and 
quite radioactive sediments with high heat flow in the central and eastern 
sections of the lake. The western lake bottom is described as one of non- 
radioactive, very light, nonmagnetic sediments of lesser thickness (5- 
kilometer). Such rocks are known on the surface of the Apostle Islands in 
the western part of the Lake. 

The geologic structure model is considered to explain the main features 
of regional gravity, magnetic and heat-flow variations "in terms consistent 
with the known or inferred geology." The investigators have found that 
there is a close relationship between gravity and heat flow in areas dominated 
by crystalline rocks. Although they recognize that heat-flow variations also 
are influenced by geological complexities at deeper levels, such as variations 
in crustal thickness, they consider it reasonable to ascribe most of the heat- 
flow variations observed on the Lake Superior bottom to variations in the 
radioactivity of the upper crustal rocks (within 10 kilometers). This study, 
confirming the hypothesis that relatively dramatic variations in energy flow 
from the interior of the earth to its surface can occur within a few kilo- 
meters, shows that horizontal variations in the crustal structure of the 
earth, no less than depth variations, are important arbiters of the earth's 
geophysical features. 

Biochemical, Biophysical, and Biological Investigations 

In the essay to which I alluded at the beginning of this Review, Alfred 
Hershey speaks of "the DNA revolution." He remarks that it is "one of a 
small class of successful scientific revolutions, and should renew significantly 
our confidence that the structure of the universe is eventually decipherable." 
Hershey further observes that revelation of the structure of DNA has "set 
off a numerous and continuing train of successes in the analysis of the 
living machine." 




2 o 

M 

< O 



!H 


rt 

a 


d. 


lO 


'3 





oS T3 

rt m 

« OJ 

<+-< »-• 

O Ph 

a ® 

P -d 



K 



a oi 

O rH 

O a> 

. -a 
a += 
.2 a 

<D —I 



a 

<B jg: 



CP O 

<v > 



3-S 



5- o 

S2* 



<u 






56 CARNEGIE INSTITUTION 

The DNA revolution itself, in the terms of this report, is no longer "new." 
However, each succeeding year brings forth fresh evidence that the frontal 
wave of analyses which it started is continuing extremely actively. The 
impact of this wave, first fully evident on a broad scale in the Institution 
during our last report year, continued strongly in three of the depart- 
ments undertaking research on biological problems: Dr. Hershey's own 
Genetics Research Unit, the Biophysics Section of the Department of 
Terrestrial Magnetism, and the Department of Embryology. 

By far the longest-continued program in the Institution directly focused 
on the story of DNA is that of the Genetics Research Unit. In the conclud- 
ing section of his report, entitled "The T4-X Universe," Hershey gives 
some elegant insights that bridge a decade of thought and experiment in an- 
alyzing the DNA message. T4 and X, of course, refer to two viruses that are 
favorite materials for DNA analyses. In his account Hershey begins with 
the classical Crick-Watson DNA model and from it derives a model struc- 
ture for T4 DNA, known to be characterized by terminal repetition and 
circular permutation, attributes not contemplated in the classical model. 
The complexity of even this simplest of all life forms is disclosed in Hershey's 
observation that "T4 DNA molecules exist as isomeric systems representing, 
perhaps, 200,000 permuted structures." On the other hand, X phage is 
known in only one molecular form. DNA molecules from X phage particles 
are all alike. 

Centering attention on terminal repetition and permutation, Hershey 
concludes, "DNA language is not human language and information theory 
of the message variety is not serviceable for both . . . DNA language is a 
chemical language obeying special topological rules, rules that are enforced 
by the special kind of twofold redundancy usually found in its messages." 
Hershey takes some pains to elaborate this critical difference because the 
term "redundancy" derives from the theory of open messages, and he con- 
siders it inappropriate to DNA language theory. Instead of the term 
"terminal redundancy" Hershey prefers "joining sequences" (Fig. 4). 

After describing the experimental evidence from both T4 and X phages, 
Hershey says "The nature of the T4-X DNA message suggests that DNA 
replication depends on at least four categories of genetic elements: one or 
more starting sequences; one or more enzymes . . . directly responsible for 
DNA synthesis, including something that recognizes starting sequences; 
joining sequences ; and a mechanism for the regeneration of joining sequences 
from joined sequences." The latter element can be supplied either by a 
cutting enzyme or by a hypothetical device for measuring lengths of DNA. 

Hershey continues with a description of experimental evidence for the 
starting sequences and joining sequences in X phage, noting that "in X 
clustering of genes of related function is not a principle but a fact." How- 



REPORT OF THE PRESIDENT 57 

AB MN Z 



A'B' M'N' Z' 



AB MN 



M'N' Z'A" 



ZA MN 



N' Z'A' M'N' 

Fig. 4. Alfred Hershey's illustration of DNA "language." Three bihelical DNA molecules are 
represented as information diagrams. Each letter signifies a nucleotide sequence of arbitrary- 
length. The prime of each letter indicates the corresponding complementary sequence on the 
second strand of a helix. Structure I conforms to the classic Crick-Watson model, but is not known 
to exist. It is an open sequence with a twofold redundancy of language in a code of paired comple- 
mentarity. Structure II contains slightly more information per nucleotide than Structure I, and has 
the sequence Z' A' which is missing from Structure I. Joining the complementary ends of Struc- 
ture II results in a closed sequence with exactly twofold redundancy. The extra information in 
Structure II thus is the potentiality of closure. Structure II does exist, as the DNA of X phage. 
Structure III has the same information content as Structure II, but differs in mass to the extent 
of the additional N and N'. It can also form a closed sequence of exactly twofold redundancy. 
It differs further from the preceding two structures in sentence order. This illustrates the possi- 
bility of permutations, 26 of which would be permitted by the alphabet used in the illustration. 
Structure III and its cyclic permutations represent T4 DNA, which is characterized by terminal 
repetition and circular permutation. 

Using these illustrations Hershey comments upon the difference between human language and 
DNA language. "Consider a message A through Z in which sentence Z is a paraphrase of A. Then, 
for purposes of human communication, we have A-Z = A-YA = AY, and Z serves only to mini- 
mize errors of transmission. In the language of DNA chemistry, A-Y and A-YA are not equivalent, 
and open sequences of the type A-Y may well be meaningless. Furthermore, we can write A- 
YA = N-YA-N for DNA language, a rule implying a degree of independence in the meaning of 
individual sentences not common in human messages. ... In short, DNA language is a chemical 
language obeying special topological rules, rules that are enforced by the special kind of twofold 
redundancy usually found in its messages." 

ever, also "in X, the joining sequences at the molecular ends form the most 
obligatory sort of functional unit but are scarcely clustered in the genetic 
map." He adds his opinion that the clustering of genes is a radical principle 
in relation to genetic combination, a useful device, but one perhaps seldom 
exploited in chromosomes at large. Hershey concludes that the radical 
clustering principle first suggested by R. A. F. Fisher nearly 40 years ago 
has a conservative counterpart, an "unclustering principle." Referring to 
evolution, he says that "history is inscrutable to us precisely because, as 
Pareto noticed some years back, it mixes incompatible principles." Nor 
should one underestimate the significance of the adventitious in evolution, 
as when "E. coli, in persuading X to center its joining sequence 12 in the 

12 The adjacent genes on either side of the joining sequence are symmetrical; development may 
proceed from either side. 



58 CARNEGIE INSTITUTION 

prophage map, just made the best of a second-rate opportunity." Hershey's 
"unclustering principle" may indeed be at least partly dependent on the 
random event. 

Isolation of a Gene Cluster. Another striking illustration of the continuing 
nature of the DNA revolution is presented this year in the work of David 
Kohne of the Biophysics Section of the Department of Terrestrial Mag- 
netism. Kohne has actually succeeded in chemically isolating certain genes. 
He has isolated and purified the nucleotide sequences in DNA that specify 
the ribosomal RNAs in the bacteria E. coli and Proteus mirabilis. By 
coincidence, Kohne treats one example of the genetic phenomenon of 
clustering to which Hershey referred. In his experiments this year he isolated 
ribosomal RNA cistrons (clusters composed of several functionally related 
genes) from the two bacterial species. The RNA cistrons may be regarded 
as the fraction of the total bacterial cell DNA that codes, that is, carries 
the message for, the formation of ribosomal RNA. This is believed to be the 
first instance of the essentially complete purification of deoxyribonucleotide 
sequences characteristic for any specific gene function. 

This isolation of the ribosomal RNA cistron depended on knowledge 
gained earlier at the Biophysics Section of the reassociation kinetics of 
bacterial DNA 13 and on the capability achieved there to separate reasso- 
ciated DNA from unreassociated DNA. It provides a particularly good 
example of the extraordinary effectiveness of a combination of intimate 
knowledge of a difficult technology, painstaking experimental work, and 
comprehensive imagination. 

Kohne's basic procedure may be summarized in two stages. First came 
the reaction of nonradioactive ribosomal RNA with radioactive DNA under 
controlled conditions such that very little DNA-DNA reassociation oc- 
curred, although all but a small fraction of the DNA complementary to 
the ribosomal RNA formed rRNA-DNA hybrids. Second came the separa- 
tion of the reassociated and unreassociated DNA by passing the incubation 
mixture through an hydroxy apatite column. Under predetermined con- 
ditions the rRNA-DNA adsorbs onto the hydroxyapatite while the unreas- 
sociated DNA passes on through the column. 

By these methods Kohne found that E. coli contains about five separate 
ribosomal RNA cistrons per cell, and Proteus mirabilis has about six such 
cistrons. Thermal stability studies of the reassociated cistrons suggested 
that the isolated cistrons have a higher guanine-cytosine content than does 
"bulk" DNA of the whole genome of E. coli (about 55 percent as compared 
to 52 percent). (Fig. 5.) The same appears to be true for Proteus mirabilis. 

13 The recombination in vitro of separate DNA Btrands into their normal double-stranded 
configuration. 



REPORT OF THE PRESIDENT 



59 




1 10 

Cgt (mole sec/liter) 



100 



1000 



Fig. 5. The reassociation kinetics of a mixture of radioactive E. coli rRNA cistrons and E. coli 
unlabeled bulk DNA. C t represents the product of the concentration of DNA and the time of 
incubation (moles of nucleotide X seconds per liter). Curve C (solid line) is a reaction curve for 
bulk DNA. Curves A and B are theoretical reaction curves calculated for situations where: 
A. All the rRNA cistrons are identical; and B. Four of the rRNA cistrons are identical and one is 
different. 

The open circles represent the observed reassociation of rRNA cistrons in David Kohne's 
experiments to determine the heterogeneity of rRNA cistrons. 



The existence of multiple rRNA cistrons in each bacterial cell has raised 
a question as to the possible heterogeneity of the ribosomal RNA. Isolation 
of the rRNA cistrons has made it feasible to examine this question of hetero- 
geneity. Further reassociation experiments by Kohne indicate that probably 
three quarters of the cistrons in E. coli and Proteus mirabilis are very 
similar to one another. A high degree of homogeneity therefore appears to 
exist in the ribosomal RNA gene clusters in these species. 

Comparison of the ribosomal RNA cistrons in different species was the 
next logical problem. Again using similar hybridization techniques, small 
quantities of radioactive labelled E. coli ribosomal RNA cistrons were 
reassociated with unlabelled DNA from other bacterial species. The species 



60 CARNEGIE INSTITUTION 

examined exhibited more than 90 percent relatedness between their 
ribosomal RNA cistrons as compared to a 4 to 40 percent relatedness be- 
tween the whole DNAs of the same species. Kohne observes in his Year 
Book report that the ribosomal RNA-cistron relatedness indicates that 
these cistrons apparently have been "strongly conserved during bacterial 
evolution and . . . the great majority of nucleotide sequences have diverged 
at a much faster rate than the rR-cistrons." This confirms independently 
previous findings of other investigators. 

Kohne's experiments clearly demonstrate that the small amounts of 
repeated nucleotide sequences which are involved in a known physiological 
function, that of protein synthesis, occur in the bacterial genome. However, 
it is not yet determined whether all of the member sequences are used in 
the synthesis. The experiments further show that theoretically any gene 
may be chemically identified, and that directed "repair" or "engineering" 
of specific gene structure is not impossibly visionary. Further, the technique 
developed in these experiments provides a powerful tool for fractionating 
DNA so as to detect the expression of that DNA during the life cycle of a 
cell or organism. It may have great usefulness in the study of differential 
gene expression occurring in developing organisms. I turn next to the Institu- 
tion's recent work in precisely this field. 

Development of the Individual Organism. Inquiry into the modes of develop- 
ment and differentiation of the individual many-celled organism has been 
a long-term interest within the Institution, having been a continuous part of 
the program of the Department of Embryology for more than 60 years. 
Today, in spite of the long history of attention to the subject, "the orderly 
morphogenesis of the fertilized egg into complex and highly differentiated 
multicellular organisms" still presents some of the most formidable prob- 
lems on the frontier of molecular biology. 14 Now that ethologists, or be- 
haviorists, see the need for embarking on "a full program of experimental 
embryology of behavior," 15 the prime importance of embryology and 
developmental biology among the life sciences can no longer be questioned. 

The complexity of the highly coordinated processes that ultimately 
produce a mature organism has been eloquently described in Arthur 
Koestler's The Ghost in the Machine. Koestler reminds us that "it takes 
fifty-six generations of cells to produce a human being out of a single 
fertilized egg cell." He enumerates the steps involved: the multiplication 
of cells, the subsequent growth of daughter-cells, cell specialization, and 
the shaping of the organism. Developmental mechanisms, Koestler argued, 
require multilevel analyses. 

14 Gunther S. Stent, "That Was the Molecular Biology That Was," Science, 26 April 1968, p. 
394. 

15 N. Tinbergen, "On War and Peace in Animals and Man," Science, 28 June 1968, p. 1416. 



Plate 2 



Report of the President 




Plate 2. Illustrations of Robert DeHaan's microelectrode recording activity of seven-day chick 
embryo heart cells that have been 48 hours in culture. Scale equals 100 microns. A. Electrode 
penetrating a myocardial cell in a group of seven cells in contact. B. Unsuccessful impalements. 
C. Successfully impaled myocardial cell. 



REPORT OF THE PRESIDENT 61 

Perhaps the easier and certainly the more popular level for research in 
our day is the molecular and the cellular. The problems confronting the 
research worker whose attention focuses on the supracellular levels remain 
formidable, even though "some reasonable molecular mechanisms for 
cellular differentiation can at least be imagined." 16 

One Staff Member of the Department of Embryology, Robert DeHaan, 
has consistently kept this focus in his research. In a lecture presented at the 
annual symposium of the Society for Developmental Biology he emphasized 
that it is necessary to explore the processes of development not only at the 
molecular level and the levels of supramolecular aggregates, but also at the 
level of cells and their parts and of tissues if we are ultimately to achieve a 
real understanding of organogenesis and the emergence of a new individual. 
In this context, DeHaan has made steady progress on several fronts in his 
continuing study of the developing heart. 

DeHaan and his co-workers have contributed particularly to our under- 
standing of the origins of the heart-forming mesoderm — that is, the location 
and boundaries of prospective cardiac tissue and to the organization of that 
tissue within those boundaries. His study has further included investiga- 
tion of the movements of tissue through which a flat sheet of mesoderm is 
converted into a primitive tubular heart. It has involved the microanatomy 
and behavioral properties of embryonic heart cells as seen both in the intact 
heart and in tissue culture, and the adhesive relations of those cells. Most 
recently he has been interested in the functional properties and population 
dynamics of the pacemaking cells, those heart cells which beat spon- 
taneously. 

During 1967-1968, DeHaan continued this multilevel approach. Prob- 
ably the most important event during the year was a technical advance, 
by which more accurate observation of attributes of the pacemaking cells 
can be made. DeHaan and Sheldon Gottlieb, an undergraduate student at 
Johns Hopkins University, succeeded in impaling and electrophysiologically 
analyzing isolated embryonic heart cells (Plate 2). Their achievement almost 
certainly will make possible more rapid advance in further study of the 
origin and characteristics of pacemaker cells. 

Often in scientific progress technical advances provide powerful keys to 
ultimate understanding. Several further accomplishments of this kind at 
the Department of Embryology during the year illustrate this situation. 

Peter Tuft of the University of Edinburgh, who spent the year as a 
Carnegie Fellow at the Department, achieved a technical innovation that 
should facilitate significant further advances of knowledge at the levels of 
cell and tissue. Tuft and B. G. Boving studied the uptake of water by the 

16 Stent, op. cit. 



62 CABNEGIE INSTITUTION 

rabbit blastocyst, a thin-walled sphere of cells resulting from cleavage that 
outwardly appear to be undifferentiated. Study of the liquid content of the 
blastocyst is considered important because the accumulation of fluid in the 
blastocyst cavity is the stimulus for the equidistant spacing of the blasto- 
cytes (undifferentiated cells) and, more generally, because it exemplifies 
an important feature of embryogenesis. The blastocyst forms the first of a 
series of relatively large, transient, liquid-filled intercellular cavities arising 
in embryonic development. The formation of this first cavity is astonishingly 
rapid. Four days after mating the rabbit blastocyst begins to expand, 
and within another four days its volume has increased ten-thousandfold. 
Nearly all of the increase in volume is due to the uptake of water. 

Tuft's and Boving's technical advance comprises a modification of existing 
techniques for determining freezing points, adapting them to the microscale 
required for studying individual blastocytes. Their technique enabled them 
to make reliable determinations of the properties of water in the blastocyst 
and its maternal environment. Their findings demonstrate convincingly 
that the accumulation of liquid in the blastocyst does not result from a 
passive flow of water, but requires energy-coupled reactions for its formation. 

A majority of the studies at the Department of Embryology were at the 
molecular level, where technical advances also played an important role 
during the year. One such advance was the development by Donald Brown, 
Igor Dawid, and others, of a method of isolating large amounts of ribosomal 
DNA from bulk DNA. 

Brown, Dawid, and their colleagues have shown that the oocytes (early- 
stage eggs) of five amphibian species as well as those of an echiuroid worm, 
Urechis caupo, contain many extra copies of the genes for the ribosomal 
RNAs known as 28S and 18S. For example, an oocyte nucleus of the frog 
Xenopus laevis contains about 1000 times as many genes for 28S and 18S 
ribosomal RNA as does the nucleus of a somatic (body) cell. The oocytes 
of these animals all synthesize large quantities of ribosomes for storage and 
the extra gene copies apparently act as templates in this synthesis. The 
genes are active only in oocytes and no longer function after the first division 
of the cell (meiosis) and ovulation. 

This event is of considerable biological interest because it is so highly 
specific. Only about 0.1 to 0.2 percent of the germ cell's DNA has been 
selected for repeated copying. Knowledge of the mechanism of replication 
therefore is being sought. Alternative hypotheses as to the mechanism 
postulate changes in the nucleotide sequence of the genes for the ribosomal 
RNA, subtle cues in the structure of the nucleolar organizer site of the 
chromosomes, and the involvement of polymerase enzymes. 

An important clue may lie in the finding reported this year by Brown, 
Dawid, and Ronald H. Reeder, a Fellow of the Helen Hay Whitney Founda- 



REPORT OF THE PRESIDENT 63 

tion, that the extrachromosomal copies of ribosomal DNA (rDNA) are 
subtly different from the rDNA associated with the chromosomes in somatic 
cells. The extra copies of rDNA in the germinal vesicle (oocyte nucleus) 
have a higher buoyant density than somatic-cell DNA. Explaining this 
difference in buoyant densities is considered of major importance by the 
Director of the Department, and is the present focus of research interest of 
the Brown-Dawid group. 

Brown, Dawid, and Reeder recently coupled the now-standard cesium 
chloride technique for obtaining density gradients with an initial precipita- 
tion of DNA by means of poly lysine. Poly lysine precipitates adenine- 
thymine-rich DNA before DNA rich in guanine-cytosine. More than 90 per- 
cent of the bulk DNA from Xenopus laevis can be precipitated by polylysine 
under conditions where at least 50 percent of the rDNA remains in solution. 
This method has permitted the purification of large amounts of pure rDNA 
from somatic-cell DNA. Evidence can now be sought (and is being sought) 
as to why the buoyant densities of the two rDNAs differ. 

Two earlier technical advances, that of cloned-cell culture and that of 
somatic-cell hybridization, were extended and applied effectively in several 
experiments at the molecular level during the year at the Department of 
Embryology. Mary Weiss, a Fellow of the U. S. Public Health Service, 
cooperating with Boris Ephrussi (University of Paris) and L. J. Scaletta 
(Case Western Reserve University) has studied the nature of the relation 
between a viral genome and the host cell transformed by that virus. One 
inquiry has concerned her and her colleagues. Is the tumor virus actually 
integrated into the genome or does it persist in a cell as an episome-like free 
particle? An extraordinary — and highly useful — property of hybrid cells 
is that all the chromosomes of one "parent" may ultimately be eliminated 
from the descendant hybrid cells. Earlier experiments by Dr. Weiss and 
H. Green showed that man-mouse hybrid cells in subsequent divisions 
undergo rapid and extensive segregation of human chromosomes with no 
accompanying loss of the mouse chromosome of the genome. Taking advan- 
tage of this phenomenon, Dr. Weiss transformed human cells by a tumor 
virus, SV 40 , and then hybridized these cells with normal mouse cells. After 
a suitable number of generations, tests for the viral antigen (T) show it to be 
absent in hybrid cells that have lost most or all of their human chromo- 
somes. This is taken as strong evidence that the viral genome was actually 
integrated with the human chromosomes of transformed cells. 

Two additional technical advances opened new vistas in the techniques 
of cell culture. In collaboration with Carnegie Fellow H. G. Coon, Dr. Weiss 
obtained substantial increases in the spontaneous mating rate of hybridizing 
cell lines by assisting their fusion with inactivated virus. Employing an 
inactivated Sendai virus and two mouse "L" cell lines known as A9 and 



64 CARNEGIE INSTITUTION 

CI 1 D, Drs. Coon and Weiss found that artificial induction of fusion did 
indeed assist the formation of these propagating hybrid strains. The same 
increase in virus-assisted mating rates was observed in combinations of the 
cell line CI 1 D and normal rat liver cells, a hybrid that should be of con- 
siderable future laboratory interest. 

The rat liver cells used for the virus-assisted hybridization themselves 
represented an important technical advance. They were obtained during the 
year by Dr. Coon, who cloned liver cells from infant rats to produce in 
culture typical epithelial pavement cells. Cells which have been serially 
subcultured show the same characteristics of phase-dense cytoplasm and 
tightly packed nuclei. Individual cells form colonies upon transfer, and they 
produce some of the normal serum antigens, but it is not yet known whether 
or not these subcultured cells are capable of other complex functions of 
normal liver cells. 

Plant Biology 

Most of the remaining biological investigations of the Institution concern 
plants. The Department of Plant Biology is primarily concerned with two 
aspects of the life of plants. Its interest in the first, experimental taxonomy — 
the relation of plant speciation and physiology to environment — can be 
traced to the earliest years of the Institution, having begun in the Desert 
Laboratory of the Institution at Tucson, Arizona, in 1903. The second is 
more strongly focused on a single biological process, the fragile, complex, 
immensely important system of photosynthesis. The biochemical investiga- 
tions of photosynthesis were commenced at the Institution's Department 
of Botanical Research at Tucson in 1910. After 58 years of work, the major 
questions still remain unsolved, and the objective of these investigations 
remains substantially the same, although, of course, significant progress has 
been made in the intervening half century. 

C. Stacy French, Director of the Department of Plant Biology, discusses 
some of the reasons for the measured progress that biochemists have experi- 
enced in this field and why full understanding of the processes of photo- 
synthesis is yet to come. Dr. French says, "As the basic questions about 
(photosynthesis) became more clearly formulated the more easily definable 
problems were taken up by chemists and physicists, usually with the expec- 
tation of a quick breakthrough from the application of their more sharply 
defined methods of experimentation and interpretation. . . . We seem to 
have passed through the period of naivete in the application of the more 
rigorous disciplines to biological problems. Rather than being considered a 
limitation on precise experimentation the diversity of species and of phy- 
siological states of living plants is now being exploited. . . . Excellence of 



REPORT OF THE PRESIDENT 65 

chemical and physical procedures combined with weak biology, or the con- 
verse situation, is no longer a characteristic of our subject." Dr. French 
may take a certain amount of satisfaction in the trend that he has reported 
for his field, because few research groups concerned with photosynthesis 
elsewhere in the world have so consistently supported biophysical-bio- 
chemical research and physiological-environmental studies side by side as 
has his own. 

Research during the year at the Department of Plant Biology also 
illustrated vividly, once again, the influence of technical advances upon 
research results. The model of the photosynthetic mechanism that is 
gradually evolving still has many uncertainties because the components of 
the systems concerned have not been isolated in vitro in their natural, 
functional form. One of the more pressing problems involving the isolating 
of components is that of separating the parts of chloroplasts that drive the 
distinct yet highly interrelated photochemical reactions. Detergents like 
Triton X-100 and digitonin have been widely used to obtain chloroplast 
fractions for experiment. Although the detergents greatly facilitate the 
separation of the fractions, they have the disadvantage of becoming incor- 
porated in the fraction isolated and may damage the activity of the isolated 
pigment and bias the experimental measurement of the activity. 

This year Jean-Marie Michel and Marie-Rose Michel- Wolwertz, Fellows 
at the Department, devised a procedure for separating chloroplast fractions 
that avoids the use of detergents. Their method uses a needle- valve homog- 
enizer developed at the Department in 1950 with a buffered salt solution, 
followed by centrifugation in a sucrose concentration gradient. The Michels' 
method was tested on both spinach chloroplasts and those of Chlorella 
pyrenoidosa. Tests made of the absorption spectra and activities of the three 
fractions obtained in this way showed that the first (low-density) fraction 
displayed photosynthetic activity and spectral characteristics of the com- 
ponent known as system 1, as previously identified from action spectra of 
five plants. The two "heavier" fractions showed absorption spectra and 
photochemical characteristics of the component known as system 2. Thus, 
the method has been shown to separate the two principal postulated systems 
of the photosynthetic process. 

By coincidence, a natural plastome mutant fractionation of the two 
photochemical systems was discovered by Ulrich Heber, a Carnegie Cor- 
poration Fellow, in mutant leaves of the evening primrose, Oenothera. Some 
of the leaves Professor Heber found to have only system 1 functioning, 
whereas others had only system 2 activity. Chloroplast particles from 
mutant system 1 leaves were found to have similar absorption spectra to 
those of the system 1 fractions isolated from spinach or Chlorella. The 
chloroplasts of the mutant system 2 leaves resemble the system 2 fraction 



66 CARNEGIE INSTITUTION 

from normal plants in photochemical and spectral properties. The detection 
and analysis of these mutants emphasizes Dr. French's comment about the 
combination of biochemistry and biology that now characterizes research 
on photosynthesis. 

Action and absorbance spectra are important means of measuring and 
identifying activity in the photosynthetic system. Their correct interpreta- 
tion is an enduring problem in photosynthesis research. The curves which 
record the spectra are always composites, and the correlation of elements in 
the shapes of the curves with postulated components of the system has been 
an important source of the understanding thus far achieved about the 
mechanism of photosynthesis. 

This year the Department experimented with a new means of analyzing 
spectral curves. For the first time a digital computer was employed in this 
task. This was made possible by the installation of a cable connection be- 
tween the Department and the Computation center of Stanford University. 
French, J. S. Brown, L. Prager, and M. Lawrence describe in the Depart- 
ment's report their first efforts at analyzing by computer the spectra of 
natural chlorophyll complexes. Curve analysis may show that evidence 
previously assumed to be indicative of an individual chlorophyll component 
may in fact represent only variations in bandwidth of chlorophyll com- 
ponents having the same peak wavelengths. This appears to be true, for 
instance, for spectral differences between the low-density and high-density 
Chlorella fractions obtained by the Michels' new method. 

On the other hand, curve analysis may also suggest the existence of 
hitherto unrecognized components. Such may be the case for the absorbance 
spectrum of the mutant system 1 Oenothera leaves described above. The 
original absorbance curves and the computer analysis of the curve are 
shown in Figure 6. The wide "tail" of the original absorbance curve may be 
due to a component having a 708-nm (7080 angstroms) peak. French and 
his colleagues state, however, that the accuracy of the records is not yet 
sufficient to do more than suggest the presence of such a long wavelength 
component. Nonetheless, it is obvious that the new method of spectral 
curve analysis is promising enough to contribute to the objective described 
by Dr. French in the Introduction to his report : "To find out how many . . . 
forms of chlorophyll exist, (and) which of them are part of the two photo- 
chemical systems of photosynthesis." 

The intricacies of identifying portions of even one set of the components 
of the photosynthetic system are well illustrated by questions involving the 
yellow pigments known as carotenoids. All photosynthesizing plants contain 
carotenoids, and a variety of answers have been proposed as to their func- 
tions. Some, but not all, of the carotenoids may be light absorbers, passing 
on their absorbed energy to chlorophyll. They are thought also to act as 



REPORT OF THE PRESIDENT 



67 



i 1 1 r 




J i i i 



600 



650 700 

Wavelength, nm 



a 

c 
o 



< 



t i i r 



1 | i r 



C36-C37 




W . 5 =I3.5 



600 



650 
Wavelength.nm 



700 



Fig. 6. The upper diagram gives the absorption spectra measured at 77° K for chloroplast 
particles isolated for Oenothera (primrose) mutants having only photosystem 1 or photosystem 
2 activity. The lower diagram indicates the possible additional components in photosystem 1 
missing from photosystem 2. 



internal light niters, protecting chlorophyll from bleaching, and to partici- 
pate in the oxygen-evolving step of photosynthesis. Extending a study begun 
earlier, David Fork continued the study of carotenoid responses in the 
yellow-green alga, Botrydiopsis, which lacks chlorophyll b, a pigment that 
interferes with the measurement of carotenoid changes. Dr. Fork found in 
experiments that activation of either system 1 or system 2 of this alga gave 
identical changes in the carotenoids. He found that the changes were in the 



68 CARNEGIE INSTITUTION 

same direction regardless of which of the two photosystems was activated. 
He concludes that the same carotenoid pigment is modified by both of the 
photosystems, and that it is not a link between the two main photochemical 
reactions. 

Experimental Taxonomy. For many years the experimental taxonomy 
group in the Department of Plant Biology has studied the effects of climatic 
and atmospheric differences dependent on altitude and latitude, concen- 
trating particularly on an examination of covariance with altitude of species 
and race characteristics in the monkey flower, Mimulus. The investigation 
has involved the comparative examination of a number of physical and 
physiological characteristics in hybrid progenies from a number of con- 
trasting forms of Mimulus, growing in very different altitudinal environ- 
ments. 

Extensive study of 16 genetic markers that distinguish the parent races 
used in the Mimulus hybrids has been carried out on thousands of plants 
over a 20-year period. They have shown that ability to survive in specific 
environments is correlated to a significant degree with certain combinations 
of morphological characteristics. These characteristics can be recognized 
as early as the third generation of progeny among the hybrids. The third 
generation hybrids, indeed, begin to approach the status of recognizable 
ecological races. 

An illustration of the versatility of Mimulus as a botanical research 
material is provided in the report by W. M. Hiesey, M. A. Nobs and 0. 
Bjorkman on photosynthesis in an amphiploid Mimulus. They note that 
amphiploidy (the inheritance of all chromosomes from two parent species) 
has played an important evolutionary role in the synthesis of new species 
from ancestral forms of restricted ecological or geographical distribution. 
Such species frequently are successful in invading new environments not 
occupied by the parent species. Measurements of the photosynthetic rates 
of M. lewisii from 10,500 feet elevation in the Sierra Nevada of California 
and of M. nelsonii from Mexico with that of their amphiploid descendants 
showed that the amphiploid had an equal or higher rate of photosynthesis 
than its parents at lower temperatures (22°C or below) . On the other hand, 
the rate at higher temperatures (38°C) was a third or less that of either of 
its parents. Hiesey, Nobs, and Bjorkman consider the measurements 
significant in that a combination genome of two species occurring through 
amphiploidy appears to have the capacity of notably altering some basic 
physiological functions. 

Bjorkman, E. Gauhl, Hiesey, and Nobs also report on the beginning 
of a study designed to compare photosynthetic responses of plants from 
widely contrasting latitudes. As their material they chose the liverwort 
Marchantia polymorpha, one of the simplest of green plants. It ranges from 



REPORT OF THE PRESIDENT 69 

the arctic polar regions to the equator, and its relatively simple structure 
favors studies of its photosynthesis in vivo. Dr. Bjorkman developed during 
the year a rapid method of measuring photosynthetic rates in liverwort 
tissue by means of an oxygen electrode. Bjorkman's method makes possible 
study in vivo of variations in photosynthetic activity, including those re- 
sulting from rapid changes of temperature, light intensity, and other causes. 

During the summer of 1968 Bjorkman and Gauhl collected Marchantia 
specimens from the Point Barrow area of Alaska as well as from California. 
Laboratory study of the effects of temperature and oxygen concentration 
levels on photosynthesis in California specimens of Marchantia was com- 
menced during the year. 

Dr. Bjorkman plans to compare the properties of some selected enzymes 
known to be of key importance in the metabolism of green plants among 
the three general types of Marchantia. One of these enzymes is carboxydis- 
mutase (ribulose-1, 5-diphosphate carboxylase), whose activity Bjorkman 
studied during the year in goldenrod (Solidago), plantain (Plantago), and 
other species. His studies were directed toward a comparison of species 
normally growing in shade with those normally growing in full sunlight. 
They demonstrated quite strikingly the much higher carboxydismutase 
activity in the plants adapted to full sunlight. Bjorkman is now developing 
methods for preparing extracts of this and other enzymes, like phosphori- 
bulokinase, and citrate synthase, the condensing enzyme, from Marchantia 
tissues. 

These studies would seem to have the potentiality of carrying investiga- 
tions of ecology and of speciation to a new level of biochemical refinement, 
demonstrating further Dr. French's comment that biochemistry and biology 
are now achieving an increasingly effective working relationship. 

Observations on Nervous System and Brain Function 

It is appropriate to end this discussion of the year's activities with the 
mention of a field comparatively new for the Institution. Gunther Stent 
recently observed, "There now seems to remain only one major frontier 
of biological inquiry for which reasonable molecular mechanisms still can- 
not be even imagined: the higher nervous system. Its fantastic attributes 
continue to pose a problem as hopelessly difficult and intractably complex 
as the hereditary mechanism did a generation ago." 17 

For several years, with the aid of the Carnegie Corporation, the Institu- 
tion has sponsored a modest number of fellowships for study of this most 
fascinating, most difficult, and perhaps most important of all biological 
fields. Such fellowships have been held by both senior and postdoctoral 

17 Stent, op. cit., p. 395. 



70 CARNEGIE INSTITUTION 

research workers at McGill University, Johns Hopkins University, the 
National Physical Laboratory (Teddington, England), and the University 
of Sussex. For a longer period Dr. Richard B. Roberts, working in the Bio- 
physics Section of the Department of Terrestrial Magnetism, has been 
collaborating with Drs. L. B. and J. B. Flexner at the University of Penn- 
sylvania in an experimental program in this general field. 

Already the fellowships have produced some significant accretions to our 
knowledge of the functioning of nervous systems. Such was Dr. Thomas B. 
Mulholland's discovery of the relation of the alpha rhythm in the brain to 
the oculomotor function during his tenure as a Carnegie Fellow at the 
National Physical Laboratory in 1964-1965. This discovery called into 
question the use of the alpha rhythm as a measure of alertness. The signif- 
icance of the discovery was attested at the International Conference on 
Attention in Neurophysiology conducted at Teddington on October 3-5, 
1967, with the partial support of the Institution. It was described by one of 
the principal participants, Dr. H. H. Jasper of Montreal, as a landmark in 
the field. 

Dr. Roberts and his colleague, A. V. Rake, report in this Year Book on 
their long-term studies of the mechanisms of memory in the brain. Roberts 
and Rake report that there appear to be three forms of memory: a long- 
term memory persisting over the lifetime of an individual, having a basis 
similar to inherited behavior patterns and requiring protein synthesis for 
formation and maintenance; a very short-term memory of some seconds 
duration, probably based on electrical reverberation, and erasable with 
electroconvulsive shock; and an intermediate-term memory of several 
hours duration, postulated to have a nonprotein chemical basis, but capable 
of surviving electroconvulsive shock. Drs. Roberts and Rake suggest that 
the latter two forms of memory, the very short-term and the intermediate- 
term, together make up a system having different properties from those 
exhibited by either type operating alone. The combination of the two types 
they designate as short-term memory. 

They postulate a model of short-term memory in which an initial burst 
of reverberation causes chemical changes that facilitate further reverbera- 
tion and greater chemical change. The reverberation is then quenched by 
inhibition of the cells concerned. Chemical tracers would decay during the 
periods of quenching. A succeeding burst of reverberation would restore or 
increase the chemical trace. Duration of short-term memory therefore would 
depend on both the initial strength of the chemical trace and on the fre- 
quency of bursts of reverberation. 

Experiments on laboratory mice during the year by Roberts and Rake 
were directed toward: (1) observation of biochemical changes correlated 
with learning; and (2) attempts to affect learning and memory by drug 



REPORT OF THE PRESIDENT 71 

injection. All experiments directed toward observing biochemical correlates 
of learning in live animals had negative results. Attention therefore was 
shifted to cell culture of neurons, which is now in progress. Experiments on 
the behavioral effects of drugs were more promising. Injection with cyclo- 
heximide (a protein synthesis inhibitor) and pentabarbitol (a depressant or 
reverberation inhibitor) produced severe memory impairment, although 
neither drug alone showed such an effect. Although Roberts and Rake do 
not regard their results as conclusive they hope that this kind of experi- 
mental approach ultimately may give some indication of the duration of 
short-term memory and the time required for the establishment of a protein- 
based long-term memory. 

This year the Institution was favored at its Annual Trustees' Lecture on 
May 3, 1968, by a report from one of the most imaginative of the students 
of nervous system function, Professor Donald MacKay of the University of 
Keel in England. MacKay chose to discuss the eventual impact upon the 
personal aspect of human nature of "mechanical" answers from investiga- 
tions like those reported by Roberts and Rake. As Dr. MacKay put it, 
"What status would be left to our conscious experience in a mechanistic 
scheme of brain function?" 

Stent, in the Science article published the week prior to Dr. MacKay's 
lecture, raised essentially the same question: " . . . the higher nervous 
system does . . . present the most ancient and best-known paradoxes in the 
history of human thought : the relation of mind to matter, or of free will to 
determinism." 18 Stent believes that the old question has renewed scientific 
importance because "increasing numbers of veteran molecular biologists of 
the informational . . . school are now turning toward the nervous system .... 
But it is also possible that study of the nervous system is bringing us to the 
limits of human understanding, in that the brain may not be capable ... of 
providing an explanation for itself . . . there exist processes which, though 
they clearly obey the laws of physics, can never be explained." 19 

Professor MacKay remarked: "... even though our brains were as 
mechanical as clockwork we, as we know one another as individuals, are 
in a certain essential sense, indeterminant and free. . . . trying to link 
experience onto brain mechanism as if it were a kind of additional invisible 
extra cause, is rather like trying to find room for the equation being solved 
by a computer in the electronic theory of its workings. . . . The mystery of 
our being doesn't consist in the mechanical inexplicability of our brains. . . ." 

Here indeed is a new frontier for scientific investigation, with significant 
problems, as Lord Adrian said at the Conference on Attention in Neuro- 
physiology, sufficient for centuries. 

18 Ibid. 

19 Ibid. 



72 CARNEGIE INSTITUTION 

Losses . . . 

It is with a sense of deep personal loss, and an even greater sense of the 
loss to the Institution, that I record here the death of Seeley Greenleaf 
Mudd, a Trustee and great friend of the Institution. In his passing, in 
Pasadena on March 10, 1968, at the age of 73, the Institution lost a wise 
leader and counselor. 

There are men who lift the age they inhabit so that all men walk on higher 
ground. Dr. Mudd was surely one of these. Generous, resourceful, humane, 
a man of limitless energy and dedication, he inspired all who knew him. 

Dr. Mudd became a Trustee of the Institution in 1940 and remained on 
its Board until his death. As a member of the Astronomy Committee, he 
was very active in the affairs of the Mount Wilson and Palomar Observa- 
tories. As a Trustee also of the California Institute of Technology, he forged 
notable links between the two institutions that jointly operate the Observa- 
tories. 

Dr. Mudd also was a Trustee of Pomona College, of Stanford University, 
the University of Southern California, and the National Fund for Medical 
Education. He served as physician, scholar, educator, advisor to govern- 
ments and institutions. As a student, he first enrolled at Stanford Uni- 
versity, later transferring to Columbia, where he received a B.S. degree in 
1917. He received the M.D. from Harvard in 1924 and after completing 
house officership at Massachusetts General Hospital, entered practice in 
cardiology in Los Angeles. He was an associate in cancer research at the 
California Institute of Technology from 1931 to 1935, and from 1935 to 
1945 he was professor of radiation therapy at the same institution. From 
1941 to 1943 he served as dean of the medical school at the University of 
Southern California. 

Dr. Mudd was a Diplomate of the American Board of Internists, a mem- 
ber of the Aesculapian Club, the Los Angeles Academy of Medicine, the 
National Research Council Advisory Committee on Aviation Medicine 
(1942-1945), the Pacific Interurban Clinical Club, the Radiological Society 
of North America, and the American Heart Association. He was also 
Director of the Finney-Howell Research Foundation in Baltimore from 
1936 to 1946, and later Director of the Good Hope Hospital Association of 
Los Angeles. To these many organizations, as to the Carnegie Institution, 
he gave his services unstintingly. All of us who knew him and worked with 
him will miss him deeply. 

The year has been further saddened by the passing of several scientists 
who had made indispensable contributions to the professional life of the 
Institution. 



REPORT OF THE PRESIDENT 73 

On April 8, 1968, the Institution lost Dr. Harold D. Babcock. He had 
served as a Staff Member of the Mount Wilson and Palomar Observatories 
for almost 40 years (1909-1948). A solar physicist of the highest distinction, 
he was a pioneer, with George Ellery Hale, in solar research. Among his many 
published contributions, his authorship, with others, of Revision of Rowland's 
Preliminary Table of Solar Spectrum Wave-Lengths, with an Extension to the 
Present Limit of the Infra-Red, and, with Charlotte E. Moore of The Solar 
Spectrum X 6600 to X 18495 may be especially noted. His life and work 
provided an outstanding chapter in the research history of the Institution. 
His son, Dr. Horace W. Babcock, is currently Director of the Observatories. 

Dr. Robert B. Sosman, who worked as a physicist and served as Assistant 
Director at the Geophysical Laboratory from 1908 to 1928, died on October 
30, 1967. He was a pioneer in American ceramics and in 1953 won the Albert 
Victor Bleininger Award, the highest honor a ceramicist can receive. In 
1937 Dr. Sosman served as President of the American Ceramic Society. 
Co-originator of the Day and Sosman high-temperature scale, he was the 
author of the definitive study The Properties of Silica. 

Dr. Carl G. Hartman, a Staff Member at the Department of Embryology 
from 1925 to 1941, died on March 5, 1968, at the age of 88. A biologist whose 
discoveries in embryology helped develop the principles of birth control, 
Dr. Hartman will long be remembered for his work and his personal contri- 
bution to the Institution. 

Another long-time associate in the Department of Embryology, Dr. George 
W. Bartelmez, author of many of the articles in the well-known series 
Contributions to Embryology, also died this year. Dr. Bartelmez' remarkable 
career encompassed two major areas in embryology: the development of the 
reproductive tract, and the development of the embryonic nervous system. 
In the latter area he laid the groundwork for what is clearly emerging as a 
major field of study. 

Miss Henrietta Swope, a Research Fellow at the Mount Wilson and 
Palomar Observatories since 1952, formally completed a most distinguished 
career with the Institution this year. At its April meeting the American 
Astronomical Society announced the presentation of the Annie Jump 
Cannon Award to Miss Swope. The award commemorates the Harvard 
astronomer Annie Jump Cannon, herself a renowned investigator of 
variable stars. 

A devoted friend of the Observatories, Miss Swope studied numerous 
cepheid variable stars in several dwarf galaxies and in M 31, the great 



74 CARNEGIE INSTITUTION 

Andromeda Galaxy. Her precise and detailed investigation of the periods, 
amplitudes and colors of these variable stars led to calibration of the distance 
moduli of the galaxies in which they occur. One of her most significant 
contributions was the calibration — the most accurate to date — of the dis- 
tance of the Andromeda Galaxy. This measurement has become the "celes- 
tial yardstick" by which other distances in the universe are calculated. 

Dr. Anna O. Shepard, a most distinguished Staff Member in the former 
Department of Archaeology, retires this year. Since the formal closing of 
the Department in 1958, she has served as a special research associate. She 
holds a unique position in archaeology as a ceramics technologist, and is 
the author of many authoritative works on this subject, which she has 
pioneered. Her book Ceramics for the Archaeologist is a classic. 

Dr. David W. Bishop, a Staff Member of the Department of Embryology 
since 1952, also retires this year. A specialist in human reproductive systems 
and spermatology, Dr. Bishop did much pioneering work at the Department. 



. . . and Gains 

In the year just past, the Institution has been particularly fortunate in 
gaining three new Trustees of great experience and reputation, all elected 
to the Board on May 3, 1968. Michael Ference, Jr., Robert Morgan Penn- 
oyer, and William Matson Roth are our newly elected Board Members. 

Michael Ference, Jr., was born in Whiting, Indiana, in 1911. He received 
his B.S. degree in 1933 from the University of Chicago. After obtaining his 
Ph.D. from the same institution in 1936, Dr. Ference became a teacher 
of physics at the University. During the war he served as civilian consultant 
to the Air Force. After spending some years with the Signal Corps England 
Laboratories, first as Chief Scientist, then as Technical Director, in 1953 
Dr. Ference entered his present association with the Ford Motor Company, 
as Chief Scientist. He served successively as Associate Director, Director, 
and Executive Director, and since 1958 has been Vice-President of Research 
for the Ford Motor Company. 

Dr. Ference is a member of the President's Science Advisory Committee, 
and the National Academy of Sciences. He is a Trustee of the RAND 
Corporation, the Argonne Universities Association and Case Western 
Reserve University, a member of the Board of Governors of Wayne State 
University, and a member of the Governor's Scientific Advisory Board of 



REPORT OF THE PRESIDENT 75 

Michigan. He is also an advisor to the Group Weather Modification Com- 
mittee of the National Science Foundation, and a member of the Research 
and Development Advisory Council. 

Robert Morgan Pennoyer was born in New York City in 1925. After 
receiving his undergraduate education at Harvard College and serving two 
years in the U. S. Naval Reserve, he studied law at Columbia University. 
In 1951 he was admitted to the New York State Bar. He was associated with 
the law firm of Davis Polk, Wardell, Sunderland & Kiendl in New York 
City until 1953, when he became Assistant U. S. Attorney for the Southern 
District of New York. Two years later he joined the staff of the Office of the 
Secretary of Defense. Since 1958 Mr. Pennoyer has been Associate, then 
Partner, in the firm of Patterson, Belknap & Webb in New York City. 

His trusteeships include the Union Theological Seminary, the Metro- 
politan Museum of Art, and the American Place Theatre. He serves on the 
Board of Visitors of Columbia Law School, and is a member of the Council 
on Foreign Relations, Inc., and a director of the Association of Harvard 
Alumni. 

William Matson Roth was born in San Francisco in 1916. He was 
graduated from Yale University in 1939. In 1947 he became associated with 
the Barber Oil Company, and the following year joined the Honolulu Oil 
Corporation, of which he is now a member of the Board of Directors. Mr. 
Roth has been Vice-President of Finance, and Director of the Matson 
Navigation Company and Subsidiaries in San Francisco, and Chairman of 
the Board of the Pacific National Life Assurance Company. In 1963 he 
became Deputy Special Representative for Trade Negotiations for the U. S. 
Government. At present he is serving the Federal Government as Ambassa- 
dor Without Portfolio, in the capacity of Special Representative for Trade 
Negotiations. 

Mr. Roth has been a Trustee of the Committee for Economic Devel- 
opment, and a Regent of the University of California. 

It gives me the greatest pleasure to record the following honors which 
have come to Staff Members of the Institution during the past year. 

Dr. Barbara McClintock, Distinguished Service Member of the Genetics 
Research Unit, was awarded the honorary degree of Doctor of Science by 
the University of Missouri for "distinguished service and leadership in 
research and education." 

Dr. Philip H. Abelson, Director of the Geophysical Laboratory, received 
the honorary degree of Doctor of Humane Letters from the University of 
Puget Sound. 



76 CARNEGIE INSTITUTION 

Dr. Horace W. Babcock, Director of the Mount Wilson and Palomar 
Observatories, was elected a corresponding member of the Socie^te* Royale 
des Sciences de Liege. 

Dr. James D. Ebert of the Department of Embryology was elected to 
membership on the Board of Directors of Oak Ridge Associated Universities. 
He was invited to deliver a National Institutes of Health Lecture, and, 
in addition, offered a series of E. B. Wilson Lectures at the University of 
Texas and gave the Yamagiwa Memorial Lecture at the Cancer Institute 
of Tokyo. 

Dr. Allan R. Sandage of the Mount Wilson and Palomar Observatories 
was elected to the Organizing Committee of the Commission Scientifique de 
Physique des Instituts Internationaux, Solvay. He gave the Halley Lecture 
at Oxford University and spoke on radio galaxies and quasars to the Inter- 
national Astronomical Union in August 1967. At a meeting of the Ritten- 
house Society honoring his studies on cosmology and his work on such 
recently discovered cosmic objects as quasars, Dr. Sandage was given the 
Rittenhouse Medal. 

Dr. William M. Hiesey of the Department of Plant Biology was awarded 
the Certificate of Merit of the Botanical Society of America "in recognition 
of distinguished achievement in and contributions to the advancement of 
botanical science. Ecological physiologist, ... a pioneer in elucidating the 
genecological nature of species; he has done much to encourage and help 
students in all areas of plant science." 

Dr. Felix Chayes of the Geophysical Laboratory served as President of 
the Mineralogical Society of America during the year 1967-1968. 

Dr. Robert Howard of the Mount Wilson and Palomar Observatories 
with Dr. V. Bumba of the Ondrejov Observatory received the Astronomy 
Prize from the Czechoslovak Academy of Sciences in November 1967 for 
their work on large-scale magnetic fields. 

Dr. Olle Bjorkman of the Department of Plant Biology was awarded the 
degree of Doctor of Philosophy "with great distinction" at the University 
of Uppsala in December 1967. This degree is awarded only to outstanding 
scholars, and confers the title of Member of the Faculty at the University 
of Uppsala. 



Reports of Departments 
and Special Studies 



Mount Wilson and Palomar Observatories 

Geophysical Laboratory 

Department of Terrestrial Magnetism 

Committee on Image Tubes for Telescopes 

Department of Embryology 

Department of Plant Biology 

Genetics Research Unit 



Mount Wilson and Palomar 
Observatories 



Operated by Carnegie Institution of Washington 
and California Institute of Technology 

Pasadena, California 



Horace W. Babcock 
Director 



OBSERVATORY COMMITTEE 

Horace W. Babcock 
Chairman 

Carl D. Anderson 
Jesse L. Greenstein 
Robert B. Leighton 

Allan R. Sandage 
Olin C. Wilson 



Carnegie Institution Year Book 67, 1967-1968 



Contents 



Introduction 

Observing Conditions 

Physics of the Sun 

Large-scale solar magnetic fields . 

Magnetograms 

Active regions 

Ha filaments and magnetic fields . 

Studies of velocity fields on the solar 
surface 

The solar _ cycle 

Solar activity 

K-line cinematography .... 

Extreme ultraviolet 

Composition of the solar photosphere. 

Coronal physics 

Forbidden lines in the Fraunhofer 

spectrum 

Planets and the Moon 

Mars 

Jupiter 

Satellites of Jupiter 

Planetary photography .... 
Stellar Spectroscopy and Photometry . 

White dwarfs 

Dwarf Me stars 

Model atmospheres of M-dwarf 
stars 

Spectroscopy and photometry of cool 
stars 

Temperature, gravity, and mass of 
t Herculis, Vega, and Sirius . 

Early-type main-sequence stars . 

Faint blue stars ...... 

Spectroscopic binaries 

Peculiar A stars 

Variable stars 

Measurements of H and K flux . 

Infrared scanning 

Classification 

Blue stragglers 

Stellar composition 

He 3 in magnetic stars 

Absolute Spectrophotometry . 

Star Clusters 

Gaps in the giant branch of M 15 . 

M 92 

M 69 

NGC 6553 

NGC 6171 

NGC 6981 

NGC 7099 

Blue stars 

R associations 

Praesepe cluster 



7 Classification system for galactic 

9 globular clusters 30 

9 Interstellar Gas and Gaseous Nebulae . 30 

9 Interstellar absorption lines ... 30 

10 Orion nebula 31 

10 Planetary nebulae 32 

11 Crab nebula 32 

Infrared Sky Survey 33 

11 Observational Cosmology .... 33 

12 New determination of the Hubble 

12 constant 33 

13 New values for the redshift K-cov- 

13 rection 34 

\\ Galaxies 34 

Distance to the Virgo cluster . . 34 

1 _ Galaxies in groups 35 

Association of galaxies and radio 

sources 36 

1^ Markarian blue galaxies .... 36 

|p Seyfert galaxies 36 

:H Infrared observations 36 

Dwarf systems 36 

1 ' Polarization 36 

r!L Radio galaxies 37 

Radio observations of peculiar 

.„ galaxies 37 

Catalog of galaxies and of clusters 

..q of galaxies 38 

Compact galaxies 38 

,q Clusters of compact galaxies ... 39 

,q Sizes of rich clusters of galaxies . . 40 

9f) Area of the sky covered by clusters 

™ of galaxies 40 

2i The Galaxy 40 

21 Supernovae 41 

22 Absolute magnitudes 41 

23 Frequency 41 

23 Photometry and light curves . . 41 

24 Spectra 42 

24 Pulsating Radio Sources 42 

26 X-Ray Sources 43 

26 Scorpius X-l 43 

27 Cygnus X-2 44 

27 Quasi-Stellar Sources 45 

27 Spectroscopic observations ... 45 

27 Radio-quiet quasars and white 

28 dwarfs 45 

28 Optical variations of quasars and 

28 N-type galaxies 46 

29 Spectral energy distribution ... 47 

29 Polarization 47 

29 Parkes source 0237-23 .... 48 

29 Source 3C 287 48 



Theoretical Studies 

Stability of rotating systems . 

Dissociation equilibrium of H~ . 

Nonthermal radiation 

Variations of extragalactic sources . 

Guest Investigators 

Astroelectronics Laboratory 

200-inch data system 

Other data systems 

General laboratory projects . 
Instrumentation 

Optical design 

Image-tube spectrograph .... 

Multichannel spectrometer . . . 



49 Polarimeter 59 

49 60-inch photometric telescope . . 59 

50 Modification of Mount Wilson 

50 60-inch telescope 60 

50 Solar instrumentation 60 

51 Robinson solar telescope .... 60 
56 Photoheliograph for Apollo ... 61 
56 Photographic Laboratory .... 61 
co Site Investigation 62 

Z Chile 62 

?5 Big Bear solar observatory ... 62 

5g Bibliography 63 

58 Staff and Organization 68 



INTRODUCTION 



This year marks the centenary of the 
birth of George Ellery Hale, founder of 
the Kenwood, Yerkes, Mount Wilson 
and Palomar Observatories. It is also 
the 60th anniversary of his discovery of 
the magnetic field of sunspots. A unique 
figure in science, Hale, with his extraordi- 
nary imagination, insight, and organiza- 
tional ability is responsible, more than 
any other individual, for the great ad- 
vances in astronomy that we are privi- 
leged to see in the 20th century. Inno- 
vator and builder of great telescopes — 
the Yerkes 40-inch refractor (1897), the 
Mount Wilson 60-inch (1908) and 100- 
inch reflectors (1918), and the 200-inch 
on Palomar Mountain (1948), each the 
largest of its time — Hale also set goals 
that inspired and guided his co-workers 
and his followers. It is fortunate that the 
story of Hale's life is now readily avail- 
able through the recent biography, Ex- 
plorer of the Universe, by Helen Wright 
(1966). 

Hale wrote in 1905, "The prime object 
of the [Mount Wilson] Solar Observa- 
tory is to apply new instruments and 
methods of research in a study of the 
physical elements of the problem of 
stellar evolution." 

As Harold Zirin has said, "Only a 
profound mind could produce such a goal 
for a solar observatory; only such goals 
and programs, coupled with Hale's 
talents, could make it successful." 

How gratified Hale would have been 
as a result of the action of the Trustees 
of the Carnegie Institution this year 
in approving the establishment of a 
major new observatory to be located 
in central Chile, and how enthusiastic 
about the expanded goals of the Mount 
Wilson and Palomar Observatories, whose 
instruments, with new methods of re- 
search, are now applied in a study of 
the physical elements and evolution of 
the universe itself. 



The quasi-stellar sources (quasars) re- 
tain a position in the front rank of in- 
terest and attention notwithstanding the 
challenges raised by X-ray stars and the 
newly discovered pulsating radio sources. 
Among the recently identified quasars 
observed spectroscopically by Maarten 
Schmidt in the year just past is 4C 25.5. 
This is a blue object, starlike in appear- 
ance, of visual magnitude about 17.5. 
Two broad emission lines in its spectrum 
at A4082 and A5200 are identified by 
Schmidt as Lyman a of hydrogen (A = 
1215) and A, 1549 of triply ionized car- 
bon. The redshift (A-A )/A o is 2.358, 
the largest so far observed. An unusual 
feature is the appearance of absorption 
lines on the longward side of the cor- 
responding emission lines ; the absorption 
lines have a redshift of 2.3683, indicat- 
ing that the absorbing material is mov- 
ing toward the emitting parts of the 
source at about 900 km/sec. 

The quasar with the richest known 
spectrum and the sharpest lines is 
Parkes 0237-23. Greenstein and Sargent, 
analyzing an excellent spectrogram 
made with the 200-inch telescope, have 
compiled a very selected list of 49 ab- 
sorption lines for most of which they 
have measured equivalent widths, half- 
widths, and central depths. Using a sys- 
tematic identification procedure devised 
by Professor John Bahcall, the three 
investigators found that it was necessary 
to segregate the absorption lines into 
five separate sets characterized by dif- 
ferent redshifts: 2.2015, 1.6706, 1.6560, 
1.5132, and 1.2642. The authors conclude 
that it is at present impossible to decide 
whether the lines arise in intergalactic 
space, in galaxies in the line of sight, or 
in clouds that have been expelled from 
the object and are moving at speeds up 
to 0.3 c relative to the source of the 
emission lines. 

Additional confirmation of the optical 



8 



CARNEGIE INSTITUTION 



variation of quasars on a time scale of 
days has been made. Such rapid fluctua- 
tions of the nonthermal radiation (con- 
tinuous emission not attributable to 
stars) constitute one of the phenomena 
that have been cited by some who hold 
that quasars are at the "local" and not 
the "cosmological" distances indicated 
by their redshifts. This argument against 
the cosmological distance scale for 
quasars has fallen, for it has been found 
independently by both Sandage and Oke 
that certain N-type galaxies (specifically 
3C 371) changed by a factor of nearly 2 
in optical intensity in a little over a year. 
It is generally conceded that N-type 
galaxies, like others, are at distances 
represented by their redshifts in the 
conventional Hubble diagram. 

Another strong argument for cosmo- 
logical distances of quasars has emerged 
from the recent estimates of the great 
number of the radio-quiet quasars and 
the rate at which the number increases 
to fainter limits. Sandage has been able 
to arrive at a first tentative estimate of 
the number of radio-quiet quasars per 
square degree to three increasingly faint 
levels of apparent magnitude. Based on 
photometry with the 200-inch telescope 
of a field centered on Selected Area 57 
and on earlier work, it is estimated that 
there are 0.4 radio-quiet quasars per 
square degree to a limit B = 18.1 (ap- 
parent magnitude in blue light) , 5 per 
square degree to Z? = 18.5, and 100 per 
square degree to 5 = 21.5. These figures 
suggest that the number of quasars over 
the whole sky to -5 = 22 is very large, 
probably approaching 10 7 . Sandage be- 
lieves that such a large number of objects 
having either a constant space density 
or a positive density gradient outward 
eliminates the local hypothesis for quasar 
distances. 

It was early in 1968 that a group of 
Cambridge radio astronomers headed by 
Dr. A. R. Hewish announced the dis- 
covery of four rapidly pulsating radio 
sources of remarkably stable frequency 
of the order of 1 second. The radio posi- 



tion of one of these "pulsars," CP 1919, 
is close to a relatively conspicuous opti- 
cal object that was suggested by Ryle 
and Bailey as a possible candidate for 
identification. 

Astronomers at Mount Wilson and 
Palomar, as at many other observatories, 
immediately started efforts at optical 
identification, but no peculiar objects are 
found near the radio positions of three 
of the pulsars, and for CP 1919 careful 
observations with the 200-inch telescope 
by Kristian, Sandage, Schmidt, Snellen, 
and Westphal, using the digital data re- 
cording and harmonic analysis, yielded 
no evidence of optical pulsation of the 
suggested blue source. The probability 
is that the known pulsars are so faint as 
to be beyond the reach of even the 200- 
inch telescope for effective observation, 
if not for detection. On this problem, 
as for the quasars and others, one can 
anticipate increasing demands for in- 
struments of the greatest practicable 
light-gathering power. 

The Observatories have been fortunate 
in having as a guest investigator Dr. 
Willem J. Luyten of the University of 
Minnesota, an astronomer who has 
specialized in the determination of the 
proper motions of stars. Some years ago, 
Luyten pointed out that the 48-inch 
schmidt telescope is a nearly ideal in- 
strument for proper-motion work. The 
National Geographic Society-Palomar 
Observatory Sky Survey, completed un- 
der the supervision of R. Minkowski with 
the 48-inch in 1957, constitutes an ex- 
cellent set of first-epoch plates for such 
measurements. In recent years, Luyten 
has obtained many second-epoch plates 
of the sky with the same instrument, 
and on this program a total of 752 re- 
peat plates have now been taken; only 
184 remain. 

During the year, Luyten and his asso- 
ciates at Minnesota examined 20 pairs of 
schmidt plates, finding more than 11,000 
proper-motion stars, among which are 
several degenerate objects of unusually 
high velocity. Soon an automated pro- 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



9 



gram of blinking all the plates for de- 
tection and measurement of proper- 
motion stars will be commenced by 
Luyten, using a machine especially con- 
structed for the purpose by the Control 
Data Corporation under a grant from 
the National Aeronautics and Space Ad- 



ministration. It is to be expected that a 
very large amount of basic data on stel- 
lar motions will result. This is one of the 
extended large-scale programs of a "sus- 
taining" nature that are important for 
the orderly advance of astronomy. 



OBSERVING CONDITIONS 



Mount Wilson received 26.98 inches 
of precipitation during the year. The 64- 
year average is 35.76 inches. Total snow- 
fall was 27 inches. The highest tem- 
perature reached 90 °F and the lowest, 
on December 14, was 9°F. 

At Palomar Mountain the total rain- 
fall was 24.63 inches, with a snowfall 
of 40.8 inches. A minimum temperature 
of 10°F was recorded. A major earth- 
quake centered at Ocotillo Wells, some 
33 miles from the Palomar Observatory, 
occurred on April 8, 1968. The telescopes 
and buildings shook intensely but no 
equipment damage was noted. A few 



rock slides were in evidence on the moun- 
tain slopes. 

As limited by weather conditions, the 
hours worked with the major telescopes, 
of a possible 3800 nighttime hours, were 
as shown in Table 1. 





TABLE 1. Observations 




Telescope 


Complete 
Nights 


Partial 

Nights 


Total 

Hours 

Worked 


60-inch 
100-inch 
200-inch 


213 

237 
238 


68 
51 
62 


2486 
2586 
2554 



PHYSICS OF THE SUN 



Routine solar observations were made 
by Thomas Cragg, Robert Howard, and 
Merwyn Utter on 323 days. The records 
of various kinds made between June 1, 
1967, and May 31, 1968, were as follows: 

Direct photographs 333 

Ha spectroheliograms, 30-foot focus 592 

K2 spectroheliograms, 30-foot focus 601 

Full-disk magnetograms 272 

Fine-scan magnetograms 1207 

Sunspot drawings 270 

Magnetic classifications of sunspot 
groups were made visually on 225 days 
during the year. 

Synoptic charts of solar magnetic fields 
observed at Mount Wilson between 
August 1959 and July 1966 were pub- 
lished in the form of an atlas by the 
Carnegie Institution of Washington 
(Publication 626). The authors are 
Robert Howard, Dr. V. Bumba of the 



Astronomical Institute of the Czecho- 
slovak Academy of Sciences, Ondrejov, 
and Sara F. Smith of the Lockheed Solar 
Observatory. Funds for the reduction of 
the data and the publication of the atlas 
were provided in part by the Office of 
Naval Research. 

Large-Scale Solar Magnetic Fields 

Howard, on a six-month leave of ab- 
sence at the Ondrejov Observatory, 
Czechoslovakia, collaborated with Dr. V. 
Bumba in a study of the large-scale 
magnetic fields of the sun as observed in 
recent years at Mount Wilson. They 
found that low-latitude patterns of mag- 
netic fields persist for intervals as long 
as one or two years, with rotation rates 
of very nearly 27.0 days (synodic). 
Many regions contribute magnetic flux 
to make up these patterns, and the long 



10 



CARNEGIE INSTITUTION 



lifetimes of the patterns indicate that 
regions tend to occur at a preferred phase 
in the 27-day period for long intervals 
of time. The 27.0-day period is un- 
doubtedly connected with recurrent geo- 
magnetic disturbances. The patterns that 
seem correlated with these recurrent 
disturbances are preferentially of posi- 
tive polarity (over the seven-year period 
studied) and precede the large Unipolar 
Magnetic Regions, which during this 
time were predominantly negative. Fur- 
ther study of the development of activity 
in the long-lived magnetic "streams'' is 
planned. 

Dr. John M. Wilcox of the Space Sci- 
ences Laboratory, University of Cali- 
fornia, Berkeley, and Howard have com- 
pared an interpolated pattern of the 
interplanetary magnetic field with the 
photospheric magnetic field. This com- 
parison reveals that there is a pattern 
in the solar field which extends over a 
wide range of latitude on both sides of 
the solar equator. This pattern shows 
characteristics that differ from the usual 
solar differential rotation in three re- 
spects: (1) If this particular inter- 
planetary sector pattern persisted for 
two years, as seems likely, then the sec- 
tor boundaries would be stretched in 
longitude by differential rotation at a 
much larger amount than is observed. 
(2) Differential rotation would lead to 
the rotationally most advanced part of 
the feature being located at the equator, 
while the most advanced part of the 
feature is actually at about 15° north 
latitude. (3) The evolution with time 
of the photospheric sector pattern (dur- 
ing 1964) is distinctly different from 
that to be expected simply from differ- 
ential rotation. 

Magnetograms 

The computer programs that reduce 
the magnetic and other data from the 
daily full-disk magnetograms were 
largely rewritten during the year in order 
to make them faster and more eco- 



nomical. The daily magnetograms, which 
require about 45 minutes to obtain at 
the telescope, and which are made with 
a square aperture 17 arc-sec on a side, 
continue to be reduced by velocity as 
well as magnetic data. For each day's 
observation, the differential rotation of 
the sun is calculated in the form of the 
coefficients of a sin 2 -f sin 4 expansion in 
the solar latitude. This project continues 
as a long-term study of possible changes 
in the differential rotation with the solar 
activity cycle. In addition, integrated 
magnetic-field strengths and fluxes are 
calculated for polar and other latitude 
regions for each day's observation. This 
large amount of material has not yet 
been analyzed in detail. This program is 
supported in part by the Office of Naval 
Research. 

From the residuals in the least-squares 
equation for the differential rotation, it 
is possible to construct a "Dopplergram" 
or isotach contour map of the solar 
surface. The observations for these pro- 
grams were completed during the year 
and the first few test Dopplergrams were 
obtained. Preliminary results give no 
evidence for the existence of meridional 
or Rossby-type large-scale currents down 
to the present limiting line-of-sight 
velocity of 100 meters/sec. 

Active Regions 

Mrs. Sara F. Smith of the Lockheed 
Solar Observatory, Burbank, and 
Howard continued a study of the mag- 
netic classifications of solar active re- 
gions. The Mount Wilson magnetograms 
were used to investigate the configura- 
tions of the associated magnetic fields. 
Such data are preferable to the sunspot 
magnetic-field configurations because the 
sunspots are relatively short-lived and 
do not always reflect an inherent mag- 
netic complexity in a region. Truly com- 
plex regions show a great deal of flare 
activity. A further study of the orienta- 
tions of magnetic axes in active regions 
has been undertaken in collaboration 



MOUNT WILSON AND PALOMAK OBSERVATORIES 



11 



with Mrs. Dora Lackner. In particular, 
regions of reversed polarity will be in- 
vestigated because in a preliminary 
study it appeared that such configura- 
tions were associated with the produc- 
tion of many flares. 

Rust continued his magnetographic ob- 
servations of active regions in hopes of 
detecting magnetic-field fluctuations in 
the photosphere at the time of solar 
flares. The magnetograph at the 150- 
foot tower, operated on the A5250 line of 
Fe I, was programmed to scan entire ac- 
tive centers in 5 to 10 minutes with 10 
arc-sec resolution. During the summer 
observation season, over 2000 magneto- 
grams were thus obtained and reduced on 
the IBM 7094 computer at Caltech. 
Funds for this extensive reduction pro- 
cess were provided by the Office of Naval 
Research. First results indicate that the 
magnetic energy in an active region in- 
creases in the half hour preceding a flare. 
During flares, the magnetic energy de- 
creases and the energy losses are com- 
parable to estimated flare energies of 
about 10 30 ergs. Contour maps of the 
longitudinal components of the photo- 
spheric fields indicate that flares start 
near small (<10") magnetic features 
that are of polarity opposite to that of 
the surrounding magnetic fields. Rust 
calls these features "satellite sunspots" 
because they usually appear clustered 
around the outer edge of the penumbra 
of a large visible sunspot. Satellite sun- 
spots themselves are not always visible 
in integrated light, but they are quite 
apparent on maps of the magnetic fields 
and they are the seats of Ellerman bombs 
and of many surges observed in Ha 
photographs. The fields of satellite sun- 
spots weaken simultaneously with flares 
and explosive surges. Rust is continuing 
to analyze these data at the Sacramento 
Peak Observatory. 

Ha Filaments and Magnetic Fields 

Rust took advantage of the excellent 
solar seeing conditions present at the 



150-foot tower in the early summer 
mornings to make several maps at 3" 
and 5" resolution of the magnetic fields 
near large, quiescent filaments. He has 
compared the magnetic flux measured 
in the photosphere near filaments with 
that measured in filaments when they 
appear as bright prominences at the 
solar limb. The limb measurements were 
obtained by the observers at the High 
Altitude Observatory's Climax, Colorado, 
station. The observations show that the 
flux available for filament support on the 
Kippenhahn and Schliiter prominence 
model is about equal to that actually 
measured in the prominence. The mag- 
netic-field distribution in the photosphere 
under quiescent filaments seems to de- 
termine filament shape, position, and 
stability on the solar disk. 

Studies of Velocity Fields on the Solar 
Surface 

Howard, in collaboration with Dr. 
J. M. Wilcox and Mr. Andrew Tanen- 
baum of the Space Sciences Laboratory, 
University of California, Berkeley, has 
investigated velocity fields on the solar 
surface by means of a newly developed 
technique. Using the magnetograph and 
guiding system at the 150-foot tower 
telescope at Mount Wilson, they scanned 
one straight line on the sun's disk back 
and forth for several hours. The digitized 
magnetic, velocity, brightness, and posi- 
tion data are plotted in such a way that 
one can follow the velocity or other 
quantities for each point on the scan 
line. Thus an accurate one-dimensional 
picture of the field may be constructed. 
A large amount of data was accumulated 
during the summer and only a prelimi- 
nary reduction has been carried out. The 
reduction of amplitudes of the oscillatory 
motions in the photosphere in a plage 
region is quite evident in these observa- 
tions, as is the absence of oscillations in 
sunspots. In quiet regions there appears 
to be a structuring of the oscillatory 
field. There are some places where the 



12 



CARNEGIE INSTITUTION 



amplitudes are consistently high, and 
others where the amplitudes are con- 
sistently low. This study is still in 
progress, and it is hoped that it may be 
possible to determine which quiet-sun 
features are associated with high ampli- 
tudes and which with low amplitudes. 

The Solar Cycle 

Leighton continued his investigation, 
begun in 1965, of a magnetokinematic 
model of the solar cycle. It seems capable 
of reproducing many of the properties 
of the solar cycle with a precision equal- 
ling that with which the properties are 
now known. This comprehensive model, 
soon to be published, represents a sig- 
nificant advance in our understanding 
of the sun. 

Solar Activity 

A number of interesting and novel 
events were observed with the photo- 
heliograph on Robinson Laboratory. In 
particular, on January 29, 1968, an erup- 
tive arch prominence was observed in Ha 
light directly on the disk through its 
entire trajectory. The eruptive promi- 
nence began with a spiraling motion and 
rose slowly (10-50 km/sec) to a height 
of approximately 200,000 km above the 
disk. It then descended toward a nearby 
sunspot group with velocities up to 200 
km/sec. The falling material produced 
chromospheric brightening. Most inter- 
esting is that the eruptive arch appeared 
bright against the disk near the top 
of its trajectory. Because this requires 
high density, as well as high temperature, 
it is suggested that shocks are developed 
through the top of the orbit. These prob- 
lems are currently under study by 
Zirin. 

In another case, a quiescence promi- 
nence was observed to brighten without 
any nearby flare but with obvious density 
and temperature increase ; it erupted up- 
ward in a surge, and then reappeared 
in its former state an hour later. 



A number of interesting solar flares 
have been observed, each giving new 
insight into the flare process. Some flares 
have been observed with rise times as 
short as 10 seconds. There appears to be 
considerable interaction between neigh- 
boring spot groups in setting off flares, 
and a number of records suggesting ex- 
change of material between neighboring 
spot groups have been obtained. The 
Robinson photoheliograph has been 
operated every clear day by Messrs. 
Bohlin, Lorenz, Lambert, and Zirin, as- 
sisted by several Caltech undergraduates. 

Mrs. Dora Russo Lackner, NASA 
postdoctoral fellow, and Zirin made a 
detailed investigation of the large flare 
of August 28, 1966, which was observed 
in Ha at Mount Wilson. This was one of 
the largest proton flares in the present 
solar cycle. This large flare appeared 
qualitatively different from the normal 
small solar flare in that energy was re- 
leased over a very large area of the 
sunspot. It was as though a small pre- 
liminary eruption set off the chain energy 
release through the entire group. The 
rise of the flare was characterized by a 
rapid elongation of two bright strands, 
which then separated rapidly with a 
velocity of 150 km/sec. The time of in- 
crease in X-ray emission agreed well 
with the period of elongation and separa- 
tion of the Ha strands. From the ob- 
served area of the flare and assumptions 
on the height, it is possible to make 
models of the emitting material. It is 
found that the X rays and the micro- 
wave bursts must be explained by tem- 
peratures, or at least particle energies, 
greater than 10 9 degrees (or greater than 
80 keV) . If these energetic electrons are 
spread over the entire volume, they 
make up about 10 -3 of the material (i.e., 
their density is 10 8 /cm 3 ) ; if the hard 
X rays and microwave bursts are pro- 
duced in a small kernel, then the density 
in the kernel must be much higher, and 
a true temperature of 10 9 degrees might 
be achieved there. The radio emission 
may be explained by synchrotron action, 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



13 



with a cutoff due to absorption by the 
corona, or possibly by the Razin effect 
as has been advocated by Ramaty and 
Lingenfelter (1967). 

K-Line Cinematography 

The Lockheed Solar Observatory 
kindly lent a birefringent filter for the 
K-line which, with the new guider, has 
been used for observations of chromo- 
spheric structure and active regions at 
the Hale Solar Laboratory. These in- 
vestigations have been pursued by 
Julian, Lackner, Zirin, and Eric Persson, 
a Caltech graduate student. Although 
cinematography of active regions has 
not been particularly interesting, beauti- 
ful details may be observed in the 
chromosphere. Films of chromospheric 
oscillations have been obtained and it 
is confirmed that the K-line shows the 
horizontal oscillations similar to those 
seen in Ha. The oscillations, however, 
are much more complex, with both bright 
and dark regions showing motion. The 
bright faculae at the edges of the K-net- 
work remain fixed during this oscillation. 

Observations of the Ha oscillation in 
the chromosphere have been obtained 
also. Zirin and Lambert, using the 150- 
foot solar tower at Mount Wilson, ob- 
tained a series of movies of the Ha line 
at the limb of the sun. These films con- 
firm that there is substantial measurable 
velocity in the horizontal oscillation of 
the chromosphere. The periods of these 
oscillations are around 180 seconds. It 
appears the amplitudes may be as high 
as 10 km/sec. It is hoped to combine the 
Ha and K-line observations soon for a 
more comprehensive understanding of the 
chromospheric oscillation. 

Extreme Ultraviolet 

Much information about the upper 
atmosphere of the sun is contained in 
spectrograms in the extreme ultraviolet, 
and in recent years these have increased 
in quality and reliability. Zirin has been 
studying their interpretation. He finds 



that the determination of abundances 
of elements from the ultraviolet spectra 
is difficult and contradictory, particu- 
larly for lines in the transition zone. 
Some of Zirin's results differ from the 
abundance determinations of Pottasch 
(Astrophys. J., 187, 945, 1963) and Athay 
(Astrophys. J., 145, 784, 1966). It ap- 
pears that considerably more study is 
required before the transition zone can 
be understood. On the other hand, abun- 
dance determinations for coronal ions 
from the extreme ultraviolet appear 
reliable. 

Recent spectroheliograms in the ex- 
treme ultraviolet line obtained at the 
Naval Research Laboratory are a rich 
source of information on the structure 
of the chromosphere. In particular, most 
of the transition-zone lines show a flat 
brightness distribution across the disk. 
Only a very narrow and sharp limb- 
brightening near the limb is seen. There- 
fore, the normal concept of a secant 
distribution is invalid. Zirin has de- 
veloped a model in which roughness and 
inhomogeneity of the chromosphere ex- 
plain this flat distribution of emission. 

Another interesting aspect of the ex- 
treme ultraviolet spectroheliograms is 
that He II radiation at A304 is observed 
in the corona. Calculation of the ioniza- 
tion equilibrium of helium shows that 
there is always a moderate amount of 
singly ionized helium present in the 
corona. Therefore, one may predict in- 
teresting effects in the A304 resonance 
line, such as reflection by coronal clouds 
or ultraviolet emission from plages. Simi- 
lar effects are observed in Lyman a, and 
Zirin has proposed that the self-reversal 
of Lyman a is, in fact, due to absorption 
in the corona. 

Kozlovsky and Zirin have calculated 
the ionization equilibrium of oxygen 
in the corona. This was aimed at solv- 
ing the contradiction that the VI 
lines are extremely strong, yet they be- 
long to an intermediate ionization level 
not observed to be strong in the transi- 
tion zone. They found that because of 



14 



CARNEGIE INSTITUTION 



the lithiumlike configuration there is a 
considerable amount of VI at tempera- 
tures up to 2xl0 6 °K. Therefore, a large 
part of the VI in fact comes from the 
corona rather than the transition zone. 
This has been confirmed by the observa- 
tion of the secant distribution in this 
line by the Orbiting Solar Observatory 
experiment of Harvard College Observa- 
tory. 

In conjunction with the effort to un- 
derstand the chromosphere from studies 
of limb-brightening in the ultraviolet, 
Simon and Zirin have attempted to gain 
information from the radio emission of 
the sun. A number of scans of the sun at 
3 cm and 13 cm were obtained by Dr. 
Alan Moffet with the Goldstone antenna 
of the Jet Propulsion Laboratory. In 
addition, data on brightness distribution 
have been collected from other sources. 
When the effect of solar active regions is 
eliminated, the brightness distribution is 
flat at almost all frequencies. This means 
that the sun is essentially rough at all 
these frequencies. Investigation of these 
matters is continuing. 

A new step in our understanding of 
the outer solar corona and zodiacal light 
was obtained in photographs made by 
Surveyors V and VI. Zirin participates as 
a member of the working committee of 
astronomy, and Bohlin is actively work- 
ing with the results. After the sun sets 
behind the lunar horizon, it has been 
possible to take long exposures of the 
solar corona with various Polaroids. 
Photographs have been obtained show- 
ing the K-corona up to 60 radii from 
the center of the sun. 

Composition of the Solar Photosphere 

Lambert has continued work on a 
comprehensive reinvestigation of the 
composition of the solar photosphere. 
New results for the alkaline earths — 
magnesium, calcium, strontium, and 
barium — have been obtained in coopera- 
tion with Dr. B. Warner of the University 
of Texas. The abundances are 



log JV(Mg) = 7.48 
log 2V(Sr)= 2.82 



log N(Cn) = 6.33 
log N(B&) = 1.90 



on the standard scale where log N(H) — 
12.00. Nine trace elements were studied 
in collaboration with Drs. B. Warner 
and E. A. Mallia of the Department of 
Astrophysics in Oxford, England. The 
results are 



log 2V(Zn) = 4.42 
log #(Ca) = 2.07 
logiV(Hg)<3.0 



log JV(Ga)= 2.84 
log JV(In)= 1.71 

logiV(Ti) <02 



log iV(Ge) = 3.32 
log JV(Sn) = 1.71 
logiV(Pb) =1.90 

The isotopic abundance ratio C 12 /C 13 
was derived by Lambert from a measure- 
ment of the isotopic line Ri, 2 a (13) of 
the CH A 2 A-X 2 tt band. This line was 
discussed recently by Richter and Tonner 
(Z. Astrophys., 67, 155, 1967). A new 
measurement of the equivalent width 
was obtained from scans made with the 
Oxford spectrometer. The derived photo- 
spheric abundance ratio is C 12 /C 13 = 
150 ±30, which exceeds the terrestrial 
ratio (C 12 /C 13 = 89) by almost a factor 
of 2. This result is in agreement with a 
prediction made by Fowler, Greenstein, 
and Hoyle (Geophys. J. Roy. Astron. 
Soc, 6, 148, 1961). This investigation 
and work in progress on isotopic lines in 
the C 2 Swan band was carried out in 
collaboration with Dr. Mallia. 

Coronal Physics 

Recent studies of the coronal-line spec- 
trum have emphasized that differences 
in composition appear to exist between 
the photosphere and the corona. In an 
attempt to resolve this outstanding prob- 
lem, Lambert has begun a thorough 
evaluation of the methods employed in 
the determination of the coronal com- 
position and the possible processes that 
might be responsible for the composition 
difference. A. G. W. Cameron (Astro- 
phys. Letters, 1, 35, 1967) suggested that 
the overabundance of heavy elements in 
the solar corona was due to an enhanced 



MOUNT WILSON AND PALOMAE OBSERVATORIES 



15 



rate of ejection of ions from the under- 
lying chromosphere and that radiation 
pressure through line absorption was 
the responsible factor. A quantitative 
study of this proposal has shown that 
the effective force produced by radiation 
pressure is insufficient to explain the ob- 
served enhancement. 

As a continuation and final summary 
of work begun at the High Altitude Ob- 
servatory, Bohlin is investigating the 
evolution, structure, and associated phe- 
nomena of white-light solar coronal 
streamers. The data for this study are 
direct photographs of the corona from 
both the 1965 eclipse and three un- 
manned, stratospheric balloon flights of 
an externally occulted coronagraph dur- 
ing 1964-1965 (the Coronascope II pro- 
ject under the direction of Dr. Gordon 
Newkirk, High Altitude Observatory, 
and assisted by Bohlin). Complementing 
these direct photographs are nearly daily 
photoelectric scans of the inner K-corona 
(r<l% Re) made with the High Alti- 
tude Observatory K-coronameter in 
Hawaii. Cross-identification of features 
seen in the photographs and by the 
K-coronameter led to a sample of eight 
coronal streamers whose positions on the 
solar disk could be found with consider- 
able certainty. 

This study shows that streamers co- 
rotate with the sun and exhibit consider- 
able stability in basic configurations 
over periods of several rotations. While 
rather slow evolution was the general 
rule for coronal structure during this 
period of solar minimum, occasionally 
growth and decay of major streamers 
within periods of 14 days were detected. 
The lifetime of one high-latitude "hel- 
met" streamer was 4 to 5 solar rotations ; 
another streamer overlying a complex 
region of activity lasted for over 7 ro- 
tations. One equatorial streamer that 
overlay an intense, flaring active region 
lived for only one rotation and was 
clearly related to an intense geomag- 
netic storm as well as to magnetic and 



solar-wind velocity structure in inter- 
planetary space. 

The average coronal streamer could 
be adequately described by a three- 
dimensional density model having a 
radially dependent core density which 
decreases as a Gaussian perpendicular 
from the core line. The 1/e-density cross- 
section contour at r~l% Ro was 
~25° in latitude by —'40 o in longitude. 
Calculations showed that such a feature 
will remain visible for about four days on 
either side of limb passage. The most 
promising method of predicting coronal 
streamer structure lay in calculating the 
magnetic potential lines of force above 
the solar surface, using as a source the 
surface fields as observed by the Mount 
Wilson magnetograph. 

This program is being extended by 
Bohlin to study the quasi-stationary pat- 
terns of coronal structure in 1965, both 
as related to solar activity and also to 
the "sector structure" of the interplane- 
tary medium observed by satellites. 
These quasi-stationary coronal patterns 
are being averaged to provide a basis 
for the derivation of a nonsymmetric 
density model of the entire inner corona 
during mid-1965. 

Bohlin has also begun analysis of the 
Jet Propulsion Laboratory photographs 
of the extended solar corona observed at 
sunset from the lunar surface by Sur- 
veyor moon-probe satellites. These pho- 
tographs should eventually provide in- 
tensity and polarization observations of 
the extended (10<,r<50 Ro) corona 
and zodiacal light, which has hitherto 
been poorly observed. 

Forbidden Lines in the Fraunhojer 
Spectrum 

The possible presence of [S I] and 
[Ca II] lines in the Fraunhofer spectrum 
was first suggested in 1948 by Bowen 
(Rev. Mod. Phys., 20, 109). Recent 
laboratory measurements of forbidden 
lines and the ultraviolet spectrum have 
provided accurate wavelengths for the 



16 



CARNEGIE INSTITUTION 



S I forbidden lines which arise from 
transitions within the 3s 2 3p 4 ground 
configuration. Lambert, in collaboration 
with Drs. J. P. Swings and N. Grevesse 
of the Institut d'Astrophysique in Liege, 
Belgium, has carried out a search for 
[S I] lines in the solar spectrum. Two 
lines were successfully identified and 
are the first convincing identifications 
of [SI] lines in an astrophysical source. 
Their observed intensities are in good 
agreement with predictions based upon 
theoretical transition probabilities and 
the sulphur abundance derived by 
Lambert and Warner (Monthly Notices 
Roy. Astron. Soc, 138, 181, 1968). 



The [Ca II] transition 4s 2 S 1/2 — 
3d 2 D 3/2 was searched for by Lambert in 
collaboration with Drs. Mallia and 
Warner. The line was identified in the 
Fraunhofer spectrum for the first time. 
The observed intensity is in approximate 
agreement with a prediction based upon 
a theoretical estimate of the transition 
probability and the revised Ca abun- 
dance derived by Lambert and Warner. 
Accurate observations were not pos- 
sible because the line is blended with a 
stronger telluric line. This study con- 
firms the independent identification of 
the [Ca II] line by Grevesse and Swings 
(C.R. Acad. Sci. Paris, 266, 110, 1968). 



PLANETS AND THE MOON 



Mars 

The reduction and analysis of the 
measurements of the absorption lines in 
the weak 5v 3 band of C0 2 at A8690 in 
the spectrum of Mars, carried out inter- 
ferometrically by Mtinch at the 1967 
opposition (Year Book 66, p. 260), have 
been completed. Individual or average 
tracings of the R8, RIO, or R12 rota- 
tional lines were fitted by least squares 
to the observed instrumental profile, 
which approximates the shape of an 
Airy function with half-intensity width 
of 0.038 A. The average equivalent width 
for these lines, reduced to the center of 
the disk, is 3.67 + 0.032 mA. It is believed 
that this result, involving some 60,000 
pulse counts per spectral element, is the 
most accurate measurement of a weak 
absorption line ever carried out in any 
astronomical source except the sun, and 
demonstrates the great power and pos- 
sibilities of Fabry-Perot spectroscopy. 

Jupiter 

The intensity of absorption lines in 
the NH 3 band at A6450 in the spectrum 
of Jupiter is being measured by Munch 
at the 100-inch coude by means of Fabry- 
Perot interferometers, as a function of 



position over the planetary disk. It is 
hoped to derive from these measure- 
ments information about the mechanism 
of line formation and scattering proper- 
ties of the atmosphere. 

Difference spectra in spectral regions 
containing strong CH 4 bands, between 
the various zones and bands in the 
Jupiter disk, are also being obtained by 
Munch with the 100-inch coude scanner 
under medium resolution (AA=3.5 A). 
From these measures it is possible to 
derive in principle the concentration of 
the scattering solid particles relative to 
the gas at various heights over the differ- 
ent areas. Further observational work 
on this problem is needed to find out 
whether it will lead to profitable results. 

Satellites of Jupiter 

Low-resolution spectra of the Gallilean 
satellites of Jupiter in the regions AA.1.4- 
1.7 and 2.0-2.4 p have been obtained by 
Munch and Neugebauer at the east-arm 
focus of the 200-inch telescope for the 
purpose of verifying previous reports re- 
garding the existence of considerable 
variations in the reflectivity of the vari- 
ous satellites. Preliminary analysis of 
the data, now in the process of reduction, 
confirms such variations. 



MOUNT WILSON AND PALOMAR OBSBEVATOEIES 



17 



Planetary Photography 

Murray continued experiments aimed 
at improving photography of the planets. 
Photographs of Mercury were made 
shortly after sunrise at the coude focus 



of the 100-inch telescope using both full 
and reduced (45-inch) aperture. Multi- 
ple printing of one set of negatives by 
Dr. Bradford Smith of New Mexico 
State University yielded recognizable 
detail on the disk of the planet. 



STELLAR SPECTROSCOPY AND PHOTOMETRY 



White Dwarfs 

In observations of the Lowell proper- 
motion stars, with special concentration 
on those of moderately small proper mo- 
tion, Greenstein discovered approxi- 
mately 40 more white dwarfs. In enor- 
mous preponderance are those of spectral 
type DA, although several new helium- 
rich stars have been found. An interest- 
ing discovery is a group of three rela- 
tively bright sdO subdwarfs, GD 298, 
299, and 300. 

At the opposite end of the white-dwarf 
sequence lie the cool degenerate stars, 
for which there have been extensive 
searches by Eggen and Greenstein. It is 
now clear that the difficulty of discovery 
of these objects is based on a genuine 
physical effect — the rapid cooling as the 
interior solidifies into a lattice, so that 
no more free energy is available from the 
nucleons. This rapid cooling would be 
reflected in a drop in the frequency as a 
function of absolute visual magnitude, 
since the stars would rapidly reach large 
bolometric corrections. Spectra obtained 
by Greenstein of suspected candidates 
for red degenerate stars have shown in al- 
most every case that the stars are weak- 
lined K or even M stars with no out- 
standing peculiarities. The ultraviolet ex- 
cess found by Eggen may be explained 
in large part by the line blanketing being 
much less in sdK stars than in the main- 
sequence stars. Of 23 stars suspected by 
Eggen to have very large space motions 
if they are main-sequence stars of 
luminosity given by their B — V color, 2 
were yellowish degenerate stars of spec- 
tral type DC, and 1 was DA. All others 
were classified as sdG, sdK, or dMp. 



However, the change in derived absolute 
magnitude using the ultraviolet excess 
was insufficient to reduce the mean space 
motions of these stars to an acceptable 
value. Of 19 stars for which tangential 
velocities could be derived, 5 had space 
motions greater than 500 km/sec. The 
most reasonable interpretation of these 
extraordinary large motions, which are 
not borne out by the observed low- 
dispersion radial velocities, is that the 
stars are, in fact, subluminous and lie 
between the main sequence and the red 
degenerate stars. The motions become 
reasonable if the stars lie at least 1.5 mag 
below the main sequence; they may, in 
fact, lie as much as 3 mag below the main 
sequence. 

Tsuji computed the effective line 
blanketing in a star of 4000° effective 
temperature and very low metal abun- 
dance. This star approaches the black- 
body line if its metal abundance is less 
than about 10~ 2 that of the sun, and is 
essentially on the blackbody line at 10 -4 
solar metal abundance. The change in 
B — V is small, however, so that Eggen's 
selection of the high-velocity proper-mo- 
tion stars has, in fact, disclosed a group 
that is intermediate in luminosity be- 
tween main-sequence and fully degen- 
erate stars. Almost all these stars are 
faint, and none of this type has yet been 
found among those analyzed spectro- 
scopically. The lines are quite sharp, 
although there is a chance that strong 
lines may, in fact, be broadened by high 
pressure, which would be consistent with 
very low metal abundance. In such stars, 
Rayleigh scattering becomes a dominant 
opacity source because of the weakness 



18 



CARNEGIE INSTITUTION 



of the negative hydrogen-ion contribu- 
tion. 

Dwarf Me Stars 

An observation made with the image- 
tube spectrograph concerns the star Wolf 
359, one of our nearest neighbors in 
space, with a parallax of 0.4 arc-second. 
This very faint star, with M F =+16.6, 
was found by Arp and Greenstein to show 
a spectroscopic flare on the first observa- 
tion. Three subsequent spectra showed 
a normal dMe spectrum with strong hy- 
drogen-line and K-line emission. During 
the flare, the hydrogen lines broadened 
very greatly and the neutral helium lines 
appeared. The velocities may have be- 
come slightly more negative. The star 
is a member of the old disk population 
with space motions U,V,W, of + 27, — 45, 
and —20 km/sec. Wolf 359 was noted to 
have had a visual flare by Prof. H. U. 
Sandig, but no other observations are 
available. It is unexpected to find a 
strong chromospheric activity persisting 
in a very old disk star, and especially 
one of such low luminosity. However, 
observations by Neugebauer and Becklin 
have further compounded the mystery of 
this unusual object. They observed the 
infrared brightness at 1.6, 2.2, and 3.4 
microns on several dates. They did not 
find any significant variation. On the 
other hand, they found that this star 
of apparent visual magnitude 13.3 had 
magnitude 5.8 at 3.4 /x. This extremely 
large infrared color index can possibly 
be explained by low temperature, ap- 
proximately 2200 °K, far below that ex- 
pected for any main-sequence star even 
of spectral type dM8. The TiO bands 
are not outstandingly strong in the blue- 
green portion of the spectrum. Combining 
the information from the infrared with 
UBV photometry by Eggen, one finds 
that the object radiates 10 4 times more 
strongly at 3.4 /x than in the ultraviolet 
at 0.4 fi. Thus the shape of the lower 
end of the main sequence is very much 
at issue. 



Krzeminski contained photoelectric 
UBV observations of a sample of dMe 
stars to establish the character of their 
light variability. New observations of 
BD + 34° 106 gave a period of 2.170 
days for sinusoidal light variation. 
Change in amplitude and phase shift 
between 1966 and 1967 observations was 
found in HDE 234677, although the 
period of light variations of 3.836 days 
remains constant. 

Model Atmospheres of M-Dwarf Stars 

In order to construct model atmo- 
spheres of M-dwarf stars, Tsuji has 
evaluated molecular-line opacities such 
as are due to vibration-rotation bands 
of H 2 and CO, pure rotation bands of 
H 2 0, and electronic bands of TiO, MgH, 
CaH, and SiH. The effect of line struc- 
ture on molecular opacity is approxi- 
mately taken into account by the prob- 
ability distribution function correspond- 
ing to the Elsasser-band model. This is 
characterized by two parameters, the 
mean absorption coefficient and the mean 
line separation at each wavelength inter- 
val. The absorption coefficient of the col- 
lision-induced vibration-rotation bands 
of H 2 is calculated, and it turns out to 
be of comparable importance with the 
absorption of Hg (ff) around 2 /x in M- 
dwarf stars. The Rosseland mean opacity 
at relatively high pressures is also cal- 
culated. Atomic-line absorptions, how- 
ever, are not considered. 

Based on these opacity sources, a 
model atmosphere of M-dwarf stars with 
r 6 = 3000°K, log g = 4.8, v t =2 km/sec, 
and solar metal content has been cal- 
culated by Tsuji. The resulting model in 
radiative equilibrium shows very low 
surface temperature near 1100°K. The 
emergent flux calculated from this model 
shows appreciable flux excess around 1 p 
and 1.6 /x, while it shows general flux 
deficiency in the infrared. This is because 
1 /x and 1.6 /x are the windows of molecu- 
lar absorptions, and because their spec- 
tral regions correspond to the maximum 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



19 



of Planck radiation of 3000°K. As a 
whole, the opacity of cool atmospheres is 
high in the infrared because of the strong 
absorptions by H 2 and CO, while it is 
relatively low below 1 \l even if TiO and 
other molecular absorptions are taken 
into account. For this reason the red- 
dwarf stars are blue as compared with 
the blackbody radiation. If this fact is 
taken into account in the analysis of 
Johnson's infrared multicolor photom- 
etry, the temperature scale of M-dwarf 
stars appears to be lower by 200 °K or 
more than that estimated without con- 
sidering the model effect. Then the effec- 
tive temperature at type M4V is about 
3000°K rather than 3200°K. This study 
is being extended to several other effec- 
tive temperatures, and the effect of con- 
vective energy transport is considered. 
Some opacity data are also improved. 

Spectroscopy and Photometry of Cool 
Stars 

Coude spectrograms of the M2V star 
HD 95735 have been obtained by Tsuji 
in the near infrared as well as in the 
blue and yellow. These plates are to be 
analyzed by the use of model atmo- 
spheres now under calculation. The Mira 
variable, o Ceti, was observed at several 
phases between August 1967 and Febru- 
ary 1968 in the near infrared. 

Tsuji observed several K-giant stars, 
including one Ba II star, £ Capricorni, 
in the near infrared mainly for a study 
of the red system of CN and possibly 
for investigation of the abundance of C 13 . 

To supplement the analysis of high- 
dispersion spectrograms of two R-type 
stars, HD 156074 and HD 182040, taken 
by Greenstein with the 200-inch tele- 
scope, photoelectric scanning of the spec- 
tra of these stars is being carried out 
with the 60-inch Cassegrain scanner. 
Molecular opacities such as are due to C 2 , 
CN, and CH, which are important in 
these stars, are investigated by the same 
method adopted to calculate the molecu- 
lar opacity for M stars. Photoelectric 



scans of several cool carbon stars are 
also being attempted. 

Temperature, Gravity, and Mass of r 
Herculis, Vega, and Sirius 

Heintze derived from high-dispersion 
spectrograms profiles of the B aimer 
lines of the B5 IV star t Her (HD 
147392; a (1900) = 16 h 16™7, 8(1900) = 
+ 46°33') and compared them with 
theoretical predicted line profiles. Mono- 
chromatic continuum magnitudes of the 
star were also observed and compared 
with monochromatic continuum fluxes 
derived from models. The effective tem- 
perature of t Her is found to be 14500 ± 
500°K and log g = 3.5 ±0.2. It is pointed 
out that t Her is not reddened, that the 
rotational velocity is very low, and that 
departures from LTE very likely occur in 
the hydrogen spectrum. 

For a Lyrae, Heintze found the effec- 
tive temperature to be 10,000 °K and 
log g = 3.6; furthermore, he concluded 
that the mass is 1.3 ±0.4 solar masses. 
Finally, he provides some evidence that 
Sirius A itself is a binary system. An 
additional conclusion is that the true 
Balmer jump (BJ) of a Lyr is 1™43, and 
that for B-type stars the following re- 
lation is valid: BJ = 1.20 (U-B) +1M5. 
The Balmer jump derived from Balmer- 
line-blanketed models of Mihalas (1966) 
seems to agree with the observations. 
Those derived in 1965 require a correc- 
tion of -0T06. The U-B colors com- 
piled by Mihalas (1966) require a cor- 
rection of +0T07 5 , while those derived by 
him in 1965 need a correction of +0?02 5 . 

Early-Type Main-Sequence Stars 

With the scanner on the 60-inch tele- 
scope at Mount Wilson, Kodaira is 
studying the continuous radiative flux 
(AA3300-6000) of B3 V stars brighter 
than about the fifth magnitude and north 
of —15° declination. He is also using the 
X-spectrograph to obtain plates for 
studying the variation of radial velocity 
of these stars. 



20 



CARNEGIE INSTITUTION 



Computer programs for the quantita- 
tive interpretation of early-type spectra 
by means of model-atmosphere tech- 
niques are being constructed by Scholz, 
and some are already in use. 

High-dispersion spectra of several 
stars (0,B,F) were taken. An investiga- 
tion of t Scorpii (BO V) and A Leporis 
(B0.5 IV) will be completed soon by 
Scholz in collaboration with J. Hardorp 
of Hamburg. The evaluation of plates of 
HD 58343 (B3 Ve), which is supposed 
to be a rapid pole-on rotator, has been 
started with Kodaira. Models of the a 
Lyr atmosphere proposed recently are 
being checked by comparison with pub- 
lished and unpublished observations. 

Faint Blue Stars 

Sargent and Dr. Leonard Searle of the 
Mount Stromlo Observatory completed 
work on the spectra of 30 of the brighter 
stars in Feige's list of blue stars at high 
galactic latitudes. The spectra were ob- 
tained earlier at the Lick Observatory, 
and had a dispersion of 48 A/mm. The 
spectrograms were classified on a system 
that used measured equivalent widths, 
line profiles, and colors in addition to 
visual inspection (Astrophys. J., 152, 443, 
1968). 

The main result of this work was the 
discovery that Feige 86, a B star with 
weak helium lines of a type commonly 
found in the galactic halo, has additional 
peculiarities in its spectrum, including 
an abnormally strong line of P II. The 
star has a large proper motion and, on 
kinematic grounds, must be a horizontal- 
branch star of Population II. Sargent and 
Searle concluded that the existence of 
Feige 86 means that the weak helium 
lines in the spectra of halo blue stars 
can no longer be taken as an indication 
of a low helium abundance in the ma- 
terial from which the star formed. Fol- 
lowing this work, Greenstein and Sargent 
began a survey of the spectra of halo B 
stars down to MV~11 in order to see if 
other peculiar stars could be found. They 



used the Palomar coude spectrograph at 
18 A/mm. It was found that BD + 
5°2468, a ninth magnitude object having 
the UBV colors appropriate to a B5 star, 
has an abnormal spectrum similar to 
that of a A4500-Si-type Ap star. This star 
has a fairly large proper motion that 
leads to a transverse velocity of over 300 
km/sec if it is on the main sequence. 
BD + 5°2468 provides additional evi- 
dence that peculiar A and B stars of the 
kind found in Population I also exist in 
Population II. 

In their work on the brighter Feige 
stars, Sargent and Searle studied the 
properties of several B stars with ab- 
normally broad H lines. They found that 
these sdB stars always show an abnor- 
mally small ratio of the strengths of the 
singlets to the triplets among the He I 
lines. The sdB stars sometimes, but not 
always, have weak He I lines for their 
colors; the singlet-triplet anomaly exists 
irrespective of the absolute strengths of 
the He I lines. In order to obtain more 
understanding of this effect, Sargent is 
cooperating with Drs. Searle and Bas- 
chek of the Mount Stromlo Observatory 
in a detailed spectrophotometric study of 
AC + 9°6-12. This is a ninth magnitude 
sdB star for which Sargent has obtained 
18-A/mm Palomar coude spectrograms. 

Sargent and Greenstein have continued 
their search for interesting objects of the 
halo population of spectral type B. These 
are selected from lists by Feige and 
Luyten, who found many proper-motion 
stars and other sources. A few objects of 
special interest may be noted. These are 
stars with weak helium lines, notwith- 
standing a photoelectric color that sets 
them close to the main sequence at spec- 
tral types B5 or earlier. One is — 12° 6299 ; 
another, +5°2468. The latter may be a 
spectrum variable of the silicon type 
belonging to the halo population. 

Spectroscopic Binaries 

Deutsch has now found approximate 
periods for two other A stars that are 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



21 



interesting spectroscopic binaries. One is 
51 Tauri, an A8 star on the Hyades 
main sequence. This is a double-line 
spectroscopic binary with a period of 
about 12.3 years. The secondary has a 
spectral type near GO. Visual observers 
should take the opportunity to search for 
it during the five years beginning in 
1971, when its separation will be of the 
order of 0'.'15. The other spectroscopic 
binary is 53 Camelopardalis, an A2p star 
in which H. W. Babcock has found a 
strong magnetic field that reverses with a 
period of 8.03 days. The star also exhibits 
large-amplitude spectrum variation and 
small-amplitude radial-velocity varia- 
tion, both with the same period. Veloci- 
ties from Babcock's plates and Deutsch's 
show also that 53 Cam is a single-line 
spectroscopic binary with an orbit of 
large eccentricity and a period of about 
6.7 years. Babcock has drawn attention to 
certain relatively small irregularities in 
the magnetic cycle, in which the phase of 
the magnetic curve appears at some 
epochs to lead the phase in a strictly peri- 
odic ephemeris and at some epochs to lag. 
These irregularities prove to be too large 
for attribution to the effects of light- 
transit time across the binary orbit; they 
appear to be caused by actual deforma- 
tions of the field in the stellar reversing 
layer. 

Peculiar A Stars 

Sargent, working in cooperation with 
Drs. S. Strom and K. Strom of the 
Harvard College Observatory, completed 
an abundance analysis of HD 204411, an 
unusual Ap star. Its main peculiarity is a 
deficiency of carbon by a factor of about 
20, together with slight widening of the 
iron peak. 

Mrs. A. I. Sargent, Greenstein, and 
Sargent completed a study of the spec- 
tra of about 15 Ap stars in the red 
region of the spectrum. This was aimed 
particularly at a study of the neon 
abundance. Lines of this element are not 
found in normal stars cooler than about 



B7. Only a few Ap stars are hotter than 
this. However, for six stars that were 
sufficiently hot for Ne I lines to be ex- 
pected if the element is in normal abun- 
dance, the following results were found: 
Ne was absent in 3 stars of the A4200-Si 
group; Ne lines slightly stronger than 
normal were found in 3 B stars with weak 
He I lines. 

Miss Judith Cohen has studied micro- 
photometer tracings and wavelength 
measures of numerous coude plates in 
the visual spectrum of a 2 Canum Venati- 
corum. Greenstein and Deutsch obtained 
these spectrograms at various phases 
during several cycles of this well-known 
spectrum variable and magnetic reverser, 
which has a period of 5.5 days. Con- 
current observations of other kinds were 
obtained by Conti, Danziger, Oke, and 
Sargent. Miss Cohen's discussion is the 
first to yield satisfactory identifications 
of most lines in the visual region of this 
rich spectrum. Like earlier investigators, 
she assigned each line to one of three 
different systems that vary distinctively 
in intensity and radial velocity. Her work 
establishes that lines of Gd III and CI II 
occur, and that their intensity and veloc- 
ity variations are indistinguishable from 
those of Gd II, Eu II, and other singly 
ionized rare earths. Dr. P. Swings of 
the Institut d'Astrophysique, Belgium, 
has reported a similar result for Gd III 
and several other doubly ionized species 
that produce lines in the region AA.3070- 
3300. Miss Cohen questions earlier attri- 
butions of lines to Hg II and Pb II, and 
she finds no evidence for P II. 

Variable Stars 

Z Camelopardalis, the prototype of a 
subclass of dwarf novae, has been studied 
spectroscopically and photometrically by 
Kraft, Krzeminski, and Dr. G. S. Mum- 
ford of the Tufts University. The star is 
found to be a double-line spectroscopic 
binary [sdBe + G(?)] with an orbital 
period of 0.289845 days and mass ratio 
near unity ; masses of the components are 



22 



CARNEGIE INSTITUTION 



around 1 99? O. Although eclipses have 
not been detected, the light curves show 
a number of features that recur with 
the same period as the spectroscopic one. 
Of these, the most pronounced is a 
"shoulder" reminiscent of that in U 
Geminorum and VV Puppis. The moment 
of spectroscopic conjunction, when the 
blue star is behind, comes on the descend- 
ing branch of the shoulder and is ac- 
companied by a (U — B) excess amount- 
ing to 0.1 mag. The mean colors at 
maximum light are B — 7=0.00, U — B = 
— 0.80, and at minimum are B — V = 
+ 0.50, U-B= -0.75. The model ad- 
vanced previously by Kraft (Adv. Astron. 
Astrophys., 2, 43, 1965) for old and 
dwarf novae appears entirely applicable 
to Z Cam. 

Garrison has completed his two-year 
classification dispersion survey of long- 
period variable stars brighter than 9th 
magnitude and north of —30°. Approxi- 
mately 900 spectrograms of 150 Mira 
variables were obtained. Spectra were 
taken of some stars at all phases, but for 
most stars preference was given to maxi- 
mum and premaximum phases. This pro- 
gram was carried out in collaboration 
with Deutsch and P. C. Keenan, who are 
continuing to obtain spectrograms at 
20 A/mm in the blue and violet for the 
study of systematic anomalies in various 
absorption lines and bands. One of the 
results of this two-year combined pro- 
gram is the discovery of AlO emission 
in R Piscium, R Corvi, R Cassiopeiae, 
and RT Librae. In the past this rare phe- 
nomenon had been seen on only two 
occasions — Mira in 1924 and R Serpentis 
in 1960. A discussion of the anomalous 
behavior of the AlO bands in Mira 
variables is in preparation. 

The distance of the 1?9 classical 
Cepheid SU Cassiopeiae was determined 
by Racine using nearby field stars that, 
like the variable, are seen to illuminate 
reflection nebulae and are presumably 
connected with the same dust cloud. Its 
absolute magnitude was found to be 
(My) = —2.45 ±0.2 mag, in good agree- 



ment with predictions based on modern 
calibrations of the period-luminosity re- 
lation. 

In June 1967, Deutsch found that the 
bright combination variable CH Cygni 
had sustained another outburst like that 
first seen in 1963. The hot continuum 
was invisible on plates of March 1961 
and September 1966, as were the per- 
mitted emission lines of hydrogen and 
helium. The M star appears to change 
its radial velocity slowly, as though in 
an orbit with period of the order of 10 
years. 

Measurements of H and K Flux 

Wilson has continued his photoelectric 
measurements of the flux at the center 
of stellar H and K lines. More suitable 
blue-sensitive photomultipliers were ac- 
quired and have been in operation since 
August 1967. About this time also, 
changes were made in the observational 
program. Numerous Stromgren-Perry 
stars were dropped; those retained were 
the stars with known H-K emission, some 
others with measured flux only slightly 
less than these, and a selection of stars 
with minimum flux to serve as standards, 
i.e., as a check on the observations. Added 
to the program were a number of later- 
type main-sequence stars, mostly known 
to have H-K emission components and 
extending from G5 to M2. 

For the earlier-type stars, F5 to G2, 
there have been so far no unambiguous 
variations in H-K flux. Hence it may be 
concluded that any cyclical changes for 
these objects must be of long period 
compared to the interval of slightly over 
two years during which they have been 
observed, or else that the technique used 
is insufficiently sensitive. 

Some of the later-type main-sequence 
stars show variations in H-K flux which 
are probably real. The time covered by 
the observations of these objects has, 
however, not been long enough to reveal 
any certain evidence of variations of a 
cyclical nature. Enough has now been 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



23 



done to indicate that this project is 
necessarily long range. Unfortunately, 
there is no real evidence to indicate ulti- 
mate success in uncovering stellar ana- 
logues of the solar-activity cycle. 

Infrared Scanning 

Hyland is using the infrared scanner 
in collaboration with E. Groth and H. 
Butcher, Caltech physics students, at the 
Cassegrain foci of the 24-inch, 60-inch, 
and 100-inch telescopes to scan K and 
M giants in the 1.5-1.8 /x. and 2.1-2.5 /j. 
ranges to obtain integrated band in- 
tensities for the vibration rotation bands 
of CO. Initial results for a Tauri, /? 
Geminorum, a Ursae Majoris, and S Vir- 
ginis, with a resolution of 32 A for the 
1.5-1.8 [x region and 67 A for the 2.1-2.4 
fi region are promising. They show that 
accurate band strengths can be obtained. 
The theoretical interpretation of these 
band strengths still presents a major 
problem, the solution of which, however, 
should yield, in combination with high- 
dispersion spectra, dependable values of 
the C/O ratio for late-type stars. 

Classification 

The spectrum of the very peculiar star 
HD 147010, which is a member of the 
upper Scorpius complex of stars, gas, and 
dust, has been studied by Garrison using 
a high-dispersion plate taken by Munch. 
The Fe II, Cr II, Si II, and Ti II lines 
are similar in strength to those of an 
early A-type supergiant, while the H 
and He I lines are similar to those of a 
late B-type giant. 

The spectrum of NGC 2024 No. 1, an 
extremely heavily reddened star in the 
Orion association, has been classified by 
Garrison. He has found it to be identical 
to that of 6 2 Orionis B (B0.5 Vp with 
very broad hydrogen lines). A compari- 
son of the photometry for the two stars 
indicates that the ratio of total to selec- 
tive absorption is less than 4, contradict- 
ing previously suggested higher values 



for Orion. The resulting absolute magni- 
tudes for the two stars are consistent 
with the appearance of their spectra. 

The cluster Tr 37 is very close to the 
position of the very-high-luminosity red 
supergiant /x Cephei. In order to in- 
vestigate the relationship between //. Cep, 
the cluster, and the association I Cephei, 
Garrison has taken classification disper- 
sion spectrograms of stars in and around 
Tr 37. The stars in this cluster also have 
an important bearing on the question 
of the ratio of total to selective absorp- 
tion since H. L. Johnson has suggested 
that it is higher in this region than 
normal. 

Garrison and Schild have finished clas- 
sification of the spectra of all southern 
OB stars earlier than B8 in the bright 
star catalog. Some B8-A2 supergiants 
were also included. The plates were 
taken by Dr. W. A. Hiltner using a 
1-prism classification spectrograph and 
the 16-inch telescope at Cerro Tololo. 

Stellar spectra are usually classified 
on the basis of suitable line ratios, since 
estimates of these can be very sensitive 
to temperature changes but relatively in- 
sensitive to abundance anomalies, lumi- 
nosity effects, or the densitometric prop- 
erties of the spectrograms. In classify- 
ing M giants, however, it is usually 
necessary to assign spectral types on the 
basis of the absolute intensities of TiO 
bands rather than on the basis of band 
ratios. At 10 A/mm, however, tempera- 
ture-sensitive line ratios can still be 
isolated in the spectra of K and M giants. 
Deutsch, Wilson, and Keenan have found 
that these confirm the spectral types 
based on TiO. Although some M giants 
have moderately high velocities — of the 
order of 120 km/sec — and reveal some 
evidence of mild metal-deficiency line 
weakening, their spectral types from TiO 
bands nevertheless are reliable tempera- 
ture indicators. One by-product of this 
collaboration is the development of a 
simple null method, based on lines judged 
to be of equal strength, for determining 



24 



CARNEGIE INSTITUTION 



g and M stars of known excitation tem- 
perature and chemical composition. An- 
other by-product is the discovery that in 
some, perhaps all, of the high-velocity 
M giants the Balmer lines have variable 
intensity. On many plates they are con- 
spicuously stronger than in standard 
giants of the same spectral type. 

Blue Stragglers 

Continuing his study of blue stragglers 
in old open clusters, Deutsch has found 
that Fagerholm 190 in M 67 is a single- 
line spectroscopic binary with a period 
of 4.198 days and a velocity amplitude 
of about 18 km/sec. He also has incon- 
clusive evidence for duplicity in other 
blue stragglers in the same cluster. These 
results lend support to the McCrea hy- 
pothesis for the formation of such stars 
by evolution of close binary stars with 
mass exchange. The results also militate 
against the view that the blue stragglers 
are single stars that have returned to 
the main sequence after passing through 
the red-giant phase of stellar evolution. 
If the McCrea hypothesis is indeed cor- 
rect, the rotational velocities of 50-100 
km/sec, reported by Deutsch in previous 
Year Books, can no longer be adduced 
as evidence in support of a Dicke-type 
model for the rotation of solar-type stars. 
However, Sargent's mass determinations 
for three other blue stragglers still tend 
to favor the idea that these stars have 
no close companions and have already 
been red giants. 

Stellar Composition 

The study of peculiar elemental abun- 
dances in unevolved stars has led to an 
unexpected result concerning the K 
dwarfs discovered by Spinrad some years 
ago to have exceptionally strong metallic 
lines. In HD 75732 and HD 145675, 
which proved to be essentially normal 
Kl main-sequence stars, Greenstein and 
Oinas have found that not only are the 
metallic lines enhanced, but a large num- 



ber of molecular features are detectable 
both in the spectrum and in the spectro- 
photometric scans. The CN molecule is 
strong, as is the Swan system of C 2 . The 
latter bands are well resolved, but would 
not, unfortunately, be strong enough to 
permit detection of C 13 atoms if they were 
present. The great strength of carbon 
features in unevolved stars cannot be 
caused by internal nucleosynthesis. The 
possibilities are (1) that the oxygen 
abundance in these stars is abnormally 
low, so that the C/O ratio exceeds unity, 
or (2) both carbon and oxygen are en- 
hanced in abundance, with carbon more 
enhanced than oxygen. In any case, it is 
apparent that the strong metallic lines 
and the change of the C/O ratio from its 
solar value in old disk-population stars 
must be connected with a spotty chemi- 
cal element concentration in the inter- 
stellar gas out of which these stars were 
formed. The possibility exists that the 
high abundance of oxygen already noted 
in Arcturus and the high abundance of 
carbon here detected are connected with 
explosions in super-massive stars formed 
early in the history of the disk. Oinas 
will continue to measure and observe 
further stars of this type for carbon, 
oxygen, and metallic-line peculiarities. 
A complete and accurate analysis has 
been made of three horizontal-branch 
A stars by Kodaira, Greenstein, and 
Oke. Kodaira had already published a 
model-atmosphere analysis of HD 
161817. The two additional stars, HD 
86986 and HD 109995, had been observed 
spectroscopically by Greenstein at Palo- 
mar and spectrophotometrically by Oke. 
Since the fine analysis of HD 161817 
was available, and since the temperatures 
of the three stars were nearly the same, 
it was possible to make an accurate dif- 
ferential analysis of HD 96986 and HD 
109995 with respect to HD 161817. The 
effective temperatures are near 7800°K, 
and the surface gravity is low (log 
gr = 3.0) . Rotation is small. The mean 
metal deficiences with respect to the sun 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



25 



were, in the logarithm, —1.54 for HD 
86986, -1.76 for HD 109995, and -1.16 
for HD 161817. Thus the horizontal- 
branch A stars of very high velocity show 
a range of metal deficiency from a factor 
of 15 below the sun to 60. The systematic 
trends among the abundance deficiencies 
are similar to those already noted for 
the very metal-poor giant HD 122563. 
The a-particle nuclei (Mg, Al, Si) are 
least deficient; those of the s-process 
(Sr, Y, Cr, Ba) are the most deficient. 
The differences in abundance anomalies 
are greatest in the most metal-poor star. 
A tentative analysis of the carbon and 
oxygen content, using lines very near the 
limit of detectability, seems to indicate 
that both carbon and oxygen are in 
high abundance relative to metals, i.e., 
not very much lower than in the sun. 
This is in agreement with earlier obser- 
vations of the slightly metal-deficient red 
giants in which oxygen showed a sur- 
prisingly high abundance. The masses 
deduced for the stars are below 1 solar 
mass and the space motions over 300 
km/sec. 

From data obtained by Conti, Dan- 
ziger, and herself, Ann Merchant Boes- 
gaard has found that the beryllium 
content of F-G dwarf stars increases with 
increasing B — V. For the same stars, the 
lithium content decreases with B — V. 
The decreases in lithium can be under- 
stood in terms of convective depletion. 
The apparent increase in beryllium, how- 
ever, requires either (1) the postulation 
of a new depletion or production mecha- 
nism, or (2) a variation in the spallation 
production parameters, e.g., duration 
and energy spectrum of bombarding 
particles. 

Mrs. Boesgaard has completed an 
analysis of the isotopic abundances of 
magnesium in 10 late-type stars from 
the 0, band of MgH. The stars studied 
give evidence for Mg 25 and Mg 26 in 
amounts similar to the terrestrial abun- 
dance ratios relative to Mg 24 . 

A spectrogram exposed in the ultra- 



violet to A310G of £ Capricorni, a hot 
Ba II star, has been obtained by Mrs. 
Boesgaard for the purpose of searching 
for the strong Tc II lines in that spectral 
region. There can be no positive identifi- 
cation of Tc II, however, since this re- 
gion of the spectrum is very crowded 
with atomic lines. An upper limit to the 
line strengths will be estimated. 

Palomar high-dispersion spectrograms 
of BD + 39°4926 (c/. Oke, Greenstein, 
and Gunn, in Stellar Evolution, p. 399, 
R. F. Stein and A. G. W. Cameron, eds., 
Plenum Press, New York, 1966) are 
being analyzed by Kodaira. They show 
numerous O I, C I, and weak ionized 
metallic lines. Tentative results suggest 
that this star has the physical parameters 
log g = l and e = O.7, normal oxygen 
and carbon abundances, and a moderate 
deficiency in metallic elements. 

A curve-of-growth analysis has been 
undertaken by Hyland for several hot 
Si A4200 stars from material obtained at 
Mount Stromlo Observatory. From 
spectra of the blue-wavelength region 
only it has been found that in each case 

(1) the excitation temperature is only 
slightly lower than the effective tempera- 
ture obtained from photoelectric scans, 

(2) the derived value of the micro- 
turbulence is very low, and (3) the 
chromium-iron ratio is greatly enhanced 
compared with normal stars. These re- 
sults are in direct opposition to those 
obtained by Searle et al. (Astrophys. J., 
145, 141, 1966) from a consideration of 
the ultraviolet spectra of similar peculiar 
stars, and may be due to an extreme 
depth dependence of the relevant quanti- 
ties through the atmospheres. 

The silicon star HD 221006 has been 
found by Hyland to be a spectrum varia- 
ble showing extreme variations of its 
Si II lines. No definitive period has yet 
been obtained for the variations. All 
measured lines of Si II vary in phase 
with one another, and Si III A4552 also 
varies in the same phase. The lines of 
Mg II and Fe II appear to vary slightly 



26 



CARNEGIE INSTITUTION 



and in phase with Si II, whereas C II 
A4267 varies slightly in antiphase with 
the other elements. He A4471 does not 
show any systematic variation and is 
always very weak for the high 
(~16,000°K) effective temperature of 
the star. It has been estimated that 
N He /N H <0.01. The value is in good 
agreement with the helium abundances 
Hyland obtained for silicon stars in 
galactic clusters, where it was shown 
that the anomalously low helium abun- 
dance was purely atmospheric. 

From a study of the absolute energy 
distributions and Hy profiles of 40 B 
stars, Hyland has determined an effec- 
tive-temperature scale as a function of 
Q=(U-B) -0.72 (B-V). This scale 
agrees well with the fundamental tem- 
peratures of Hanbury-Brown et al. 
{Monthly Notices, Roy. Astron. Soc, 
137, 393, 1967) and with the scale pro- 
posed by Morton and Adams (Astro- 
phys., J., 151, 61, 1968) , but predicts con- 
siderably higher temperatures than those 
found by Heintze (Bull. Astron. Inst. 
Neth., 20, 1-25, 1968). This scale has 
been used to derive helium abundances 
for 9 cluster B stars and leads to a mean 



result of N H e/N H = 0.11, in good agree- 
ment with most recent direct and in- 
direct determinations. 

He 3 in Magnetic Stars 

Zirin has studied the A10830 line in a 
number of magnetic stars in order to con- 
firm the result of Sargent and Jugaku on 
the presence of He 3 in 3 Centauri A. The 
infrared image tube of the 72-inch camera 
of the 200-inch was used with a new 
transfer lens composed of two Canon 
f/0.9 camera lenses face to face that has 
made possible considerable gain in speed. 
The presence of the He 3 line in 3 Cen A 
is confirmed. He 3 is also found in e Ursae 
Majoris. Both He 3 and He 4 are found in 
yi Virginis. The helium line could not be 
detected in a 2 Canis Venaticorum, a 
Andromedae, /3 Coronae Borealis, and 
several other magnetic variable stars. 
The A10830 line is favorable for such 
observations because it is the strongest 
of helium lines. Also, the He 3 -He 4 shift 
is about 1 A. 

The A10830 line has been studied in 
various other stars also, one of the most 
interesting results being the observations 
of A10830 emission in R Scuti. 



ABSOLUTE SPECTROPHOTOMETRY 



A program of absolute calibration of 
the flux of a Lyrae is being continued 
by Oke and Schild. All equipment for 
the calibration, including the 4-inch 
Newtonian telescope, the 200-inch prime- 
focus scanner, the data system, and the 
platinum blackbody furnace, is now in 
operation at Palomar. A number of ob- 
servations of the furnace and a Lyrae 
are now in hand, but reductions are not 
yet complete. One problem being en- 
countered is that the extinction along 
the horizontal path between the source 
and the 4-inch telescope seems to be 
greater than was expected. This extinc- 
tion is being measured by means of a 
quartz-iodine lamp that is operated al- 
ternately on the roof of the powerhouse 



and atop a tower located midway be- 
tween the powerhouse and the 4-inch 
telescope. From the difference between 
these two measurements, with allowance 
made for the reciprocal square decrease 
of brightness and distance, the horizontal 
extinction can be deduced. Careful mea- 
surements of this extinction are being 
made each night in which absolute cali- 
bration work is done. 

Schild is measuring absolute fluxes of 
bright B and A stars to visual magnitude 
3.5 with special emphasis on the stars 
near the galactic poles, which are pre- 
sumably not appreciably reddened. These 
data will be used for a careful compari- 
son with detailed models when the ab- 
solute calibration is available. Several 



MOUNT WILSON AND PALOMAE OBSERVATORIES 



27 



bright stars visible from the northern 
hemisphere for which direct radius de- 
terminations are available from the in- 
terferometer work by Dr. R. Hanbury- 
Brown at the Narrabri Observatory are 
also being observed. 

The 4-inch telescope and scanner are 
also being used by Oke and Schild to 
measure absolute fluxes of bright ellipti- 
cal galaxies, especially NGC 4472. From 
these data, the i£- corrections needed in 
the luminosity-redshift relation will be 
derived. It is important to observe the 
elliptical galaxies in their entirety since 
population differences between the nuclei 
and outer regions are thought to exist. 



The 4-inch telescope is an ideal instru- 
ment with which to make such measure- 
ments, since it has a large field and 
a chopper with which the contribution of 
the nightsky radiation can be continu- 
ously monitored. 

Visvanathan and Oke completed the 
observation of 25 stars in the Pleiades 
cluster. Absolute fluxes in selected 50-A 
bands in the region of AA3400-8000 
and Hy line strengths have been obtained 
with an S20 cell at the 60-inch telescope. 
For the bright early-B stars, the line 
strength of the He I A4472 was observed 
with the coude scanner of the 100-inch 
telescope. 



STAR CLUSTERS 



Gaps in the Giant Branch of M 15 

The newly modified Sartorius iris 
photometer has been used to measure a 
color-magnitude diagram for the globu- 
lar cluster M 15 that is sufficiently pre- 
cise so that fine structure has begun to 
appear. Katem measured 664 uncrowded 
stars brighter than V=17 on ten plates 
taken by Sandage in each of two colors. 

The major new result of the investiga- 
tion is the appearance of a distinct gap, 
about 0.5 mag wide in V, centered at 
A7 = 0.9 mag above the horizontal 
branch. Other deeper but narrower gaps 
are probable at V = 15.5 and V— 16.1. 
With respect to a smooth distribution, 
the deficiency of stars in the major gap 
is 4.2 times the standard deviation. In 
addition, a similar structure at the same 
distance above the horizontal branch also 
seems to occur in previous color-magni- 
tude diagrams of other globular clusters, 
including M 5, M 2, M 3, NGC 7006, 
NGC 362, NGC 5897, and perhaps M 53. 

Kristian found it possible to estimate 
that the intrinsic width of the giant 
branch in M 15 is less than ±0.01 mag 
in B — V at constant luminosity by com- 
paring the known measuring errors with 
observed dispersion in the color-magni- 



tude diagram. This corresponds to a 
surface temperature difference of AT/ 
T<±0.01 and to a difference in radius 
of AR/R< ±0.02. 

The presence of gaps along the giant 
sequence suggests that stars do not 
evolve at a uniform rate along this 
sequence. If real, the gaps provide con- 
straints on models of nuclear burning in 
interior shells as evolution proceeds. In 
order further to confirm the results found 
for M 15, a similar study of the globular 
cluster M 3 has been started. 

M92 

Using photographic plates and a pho- 
toelectric sequence obtained by Sandage 
with the 200-inch reflector, Hartwick has 
made star counts to a limiting magnitude 
V ~ 21.5 in the globular cluster M 92. 
These counts are being used to obtain 
semi-empirical evolutionary tracks with 
the object of obtaining the helium con- 
tent of M 92. 

M69 

Hartwick and Sandage have obtained 
a color-magnitude (C-M) diagram for the 
strong-line globular cluster M 69. A pho- 
tometric sequence was set up in the 



28 



CARNEGIE INSTITUTION 



cluster by means of a photographic trans- 
fer from NGC 6356, in which a photo- 
electric sequence exists. The C-M dia- 
gram shows a stubby red horizontal 
branch and a relatively gently sloped 
giant branch resembling 47 Tucanae in 
both of these respects. With an adopted 
reddening of 0.2 mag and an assumed 
magnitude of M v =+0.5 for the hori- 
zontal branch, the true distance modulus 
for M 69 is (m-M) = 15.2 ±6.5. The 
membership of four long-period variables 
near the top of the giant branch seems 
certain. 

NGC 6553 

Using photographic plates obtained by 
Sandage, Hartwick has obtained a color- 
magnitude diagram for the globular 
cluster NGC 6553. This cluster is of in- 
terest because it has been classified by 
W. W. Morgan as being among the 
strongest-lined of all the galactic globu- 
lar clusters. The cluster is projected on 
an extremely rich star field, making pho- 
tometry difficult. A sequence was set up 
by photographic transfer from NGC 
6356, and the scale and zero point of the 
sequence were established by Racine by 
photoelectric observations of three of 
the stars with the 100-inch telescope. 
While the topology of the C-M diagram 
of NGC 6553 is similar to that of M 69, 
the giant branch is less steeply sloped. 
By adopting a reddening of 0.71 mag and 
an absolute magnitude for the horizontal 
branch of .MY =+0.5, a true distance 
modulus of (m — M) = 14.42 ±0.5 was 
found. The membership of two long- 
period variables near the top of the giant 
branch, as well as one RR Lyrae star, 
seems certain. 

Using photographic plates calibrated 
with photoelectric standards, Dickens has 
been studying the following globular 
clusters in three colors (U, B, and V). 
The observations were made with the 
100-inch and 60-inch telescopes. 



NGC 6171 

Work on the RR Lyrae variables in 
this cluster is nearing completion. With 
the help of Philip Rust, a Caltech stu- 
dent, more than 10,000 star measure- 
ments were made on 80 plates using the 
digitized Sartorius iris-diaphragm pho- 
tometer. The program included many 
horizontal-branch stars as well as RR 
Lyrae variables. Computer programs 
were written and used to produce light 
curves for the variables, and also color- 
magnitude and two-color diagrams for 
the variables at mean light and for the 
horizontal-branch stars. Analysis of the 
residuals of comparison stars also mea- 
sured indicates that magnitudes and 
colors averaged over more than 20 plates 
have a mean error ~0.01 mag. 

There is quite a sharp separation in 
color between variable and nonvariable 
stars along the horizontal branch. There 
is also a complete separation in color 
between a- and c-type variables, with 
a gap of about 0.05 mag between the 
reddest c-type and the bluest a-type. 
This is in marked contrast with metal- 
poor clusters, such as w Centauri, that 
show an overlap in color between the two 
types of variable star. The mean period- 
amplitude relationship for NGC 6171 
occurs at considerably shorter periods 
than that for variables in metal-poor 
clusters. Since NGC 6171 appears to be 
metal rich, its variables can presumably 
be identified with those in the field with 
small values of the line strength param- 
eter As. These delineate a mean period- 
amplitude relation at shorter periods 
than those of large As (metal poor) . Since 
NGC 6171 contains large amplitude 
variables with periods ~0.4 day, these 
variables probably represent the cluster 
counterparts of the short-period, large- 
amplitude a-type variable stars previ- 
ously known to occur only in the field. 

NGC 6981 

The color-magnitude and two-color 
diagrams have been obtained by Dickens 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



29 



for stars in and above the horizontal 
branch, which occurs at V = 16.9. Al- 
though the integrated spectrum of the 
cluster is G3, the color-magnitude dia- 
gram does not appear to be typical of 
very metal-rich clusters in that it has 
nearly equal numbers of stars on each 
side of the variable gap. Provisional 
values of the reddening and ultraviolet 
excess are E(B-V) =0.07 and 8(U-B) 
= 0.20. 

NGC 7099 

Dickens has obtained the color-magni- 
tude and two-color diagrams for stars 
brighter than 7 = 17. The horizontal 
branch occurs at 7 = 15.5, which, using 
the provisional value of E(B — V) =0.10 
from UBV measurements of field stars, 
indicates a distance of 9 kpc, assuming 
M v =+0A for the RR Lyrae variables. 
The cluster appears to be extremely metal 
poor (integrated spectrum A7) with a 
heavy concentration of stars on the blue 
side of the variable gap and an ultra- 
violet excess of S(U- B) =0.28. The hori- 
zontal-branch stars lie to the left of the 
standard relationship in the U—B/B — V 
diagram. 

Blue Stars 

Sargent continued earlier work on the 
spectra of blue stars in globular clusters 
(Year Book 66, p. 275). Greenstein 
(Astrophys. J., 152, 431, 1968) suggested 
that the sharp K lines of Ca II found by 
Sargent and others in the spectra of 
globular-cluster blue stars and attributed 
to the interstellar medium might be due 
to the unresolved background of late- 
type stars in the clusters. Sargent found 
preliminary evidence against this hy- 
pothesis in that a blue star in M 15, 
twice as far from the center of the cluster 
as those previously observed, still has a 
strong, sharp K line. 

R Associations 

Racine has initiated a photometric and 
spectroscopic study of R associations in 



an attempt to investigate the evolution- 
ary characteristics of their illuminating 
stars as well as the law of interstellar 
reddening and absorption. The observa- 
tions are essentially completed for two 
groups of reflection nebulae in Monoceros 
and Canis Majoris. Monoceros R2 (a = 
6 h 05 m ; S=-6°20', 1900) contains some 
30 members with a range of nearly 1 mag 
in E(B-V). Canis Majoris (a = 7 b 00 m ; 
S=-ll°20', 1900) is possibly associated 
with Canis Majoris OBI, and is set 
within a long arc of H II emission ter- 
minated by a vast reflection nebula il- 
luminated by HD 53623. Among the 
probable members of CMa Rl one finds 
the N star W Canis Majoris and the R 
Coronae-type variable Z Canis Majoris 
(MWC 165) . This star, at the tip of the 
very small nebula NGC 2327, shows 
strong P Cygni emission over a peculiar 
F-type spectrum. Becklin and Neuge- 
bauer observed the star in the infrared 
and found an excess reminiscent of R 
Monoceros. Other stars possibly related 
to this group are the R star EM Monoc- 
eros and the P Cygni star HD 51480. 
The final analysis of the data for 
these two R associations is in progress 
and observations of similar groups are 
being made in the summer Milky Way. 

Praesepe Cluster 

The effect of axial rotation on the 
colors and magnitudes of stars in Prae- 
sepe was further examined by Dickens, 
Kraft, and Krzeminski. As mentioned in 
Year Book 66, (p. 276), deviations 
8(U-B) for a fixed (B-V) are found to 
be less well correlated with Y= (V sin i) / 
(V sin i) than is the case in the Hyades, 
and correlation found by Strittmatter 
between excessive brightness and (V sin 
i) 2 is poorly confirmed. A study of 
Stromgren-type photometric parameters 
made available by Crawford suggests 
that the unexpectedly poor correlation 
found for the Praesepe stars may be a 
result of a very slight differential red- 
dening. 



30 



CARNEGIE INSTITUTION 



Classification System for Galactic 
Globular Clusters 

In view of recent suggestions by S. 
van den Bergh (Astron. J.., 72, 70, 1967; 
Publ. Astron. Soc. Pacific, 79, 460, 1967) 
and Sandage and Wildey on the possibil- 
ity that the helium abundance affects the 
distribution of stars along the hori- 
zontal branch in globular clusters, Hart- 
wick has proposed a two-dimensional 
classification in which the two indices are 
the quantity (B — V) 0>g defined by Sand- 
age and Smith (Astrophys. J., 144, 886, 
1966) and a parameter S defined as the 
slope of the line joining the intersection 
of the horizontal and giant branches and 
a point on the giant branch 2.5 mag 



brighter. According to theoretical compu- 
tations of Demarque, Giesler, and Iben, 
both of the above indices depend on the 
helium as well as on the metal abundance 
of the cluster. The dependences were 
separated by plotting each quantity 
against the Morgan class, which is essen- 
tially an indicator of metal abundance 
only. When the two indices for each of 
the 27 globular clusters with available 
C-M diagrams are plotted, a two-dimen- 
sional progression of horizontal-branch 
shape is found as one proceeds along a 
path of constant Y, varying Z, and a 
path of varying Y, constant Z, thus con- 
firming the suggestion of van den Bergh, 
Sandage, and Wildey. 



INTERSTELLAR GAS AND GASEOUS NEBULAE 



Interstellar Absorption Lines 

Munch and Vaughan have continued 
their analysis of high-resolution inter- 
ferometric observations of interstellar 
absorption lines in the spectra of nearby 
early-type stars. A curve-of-growth 
analysis of the sodium D lines had indi- 
cated a typical internal rms velocity 
dispersion, for apparently single com- 
ponents, of as low as 0.5-1.0 km/sec. 
Thus hyperfine splitting must be taken 
into account in interpreting both the 
curve of growth and the observed pro- 
files. One result of considerable interest 
is the possible existence of several un- 
usually massive interstellar sodium 
clouds, which may tend to show lower 
turbulence than the more common, less 
massive (or more highly ionized) clouds. 
In connection with this work, a new 
larger-aperture Fabry-Perot etalon de- 
signed to give improved resolution in 
the violet has been acquired recently. 
Further observations of interstellar Ca II 
(H and K) , CN, and CH + in this region 
are planned. 

Vaughan and Danziger used the 100- 
inch coude interferometric scanner in an 
attempt to obtain an improved upper 



limit on the abundance of interstellar 
lithium in the directions of £ Ophiuchi 
and a Cygni. The observations, which 
are currently being reduced, were con- 
centrated in five selected bands of 2 km/ 
sec width on and near the expected radial 
velocities of interstellar absorption. On 
the basis of a recent study of the H I 
clouds near £ Oph by Herbig (Z. Astro- 
phys., 68, 243, 1968) , it is considered that, 
if the hoped-for detection limit of W\< 
0.5 mA at A6707 can be reached, a nega- 
tive result would imply an interstellar 
lithium-silicon abundance ratio not 
greater than the chondritic value. 

The spectrum of £ Oph has been 
scanned photoelectrically by Munch 
through a Fabry-Perot interferometer 
around the expected wavelength A8757.69 
A of the R0 line in the (2, 0) band of 
the Phillips system of the C 2 molecule. 
So far no line stronger than 5 mA has 
been detected at an 80% confidence level, 
but it is planned to continue the observa- 
tions in order to set a sharper upper limit. 

The interstellar Ca II lines have been 
noted for some time to be present in stars 
of high-galactic latitude. Since a vertical 
path through the normal layer of gas at 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



31 



the galactic pole is supposed to traverse 
only 100 parsecs, the presence of the 
K line on low-dispersion spectra of halo 
blue stars was rather remarkable. The 
line has now been measured by Green- 
stein in 33 halo stars of various spectral 
types and distances from the galactic 
plane. In stars of moderate galactic 
latitude, intensities reach 0.55 A; if the 
absorbing layer is assumed to be plane- 
parallel, the equivalent width of the 
interstellar K line reduced to the galactic 
poles is found to range from 0.35 A to 
0.10 A. There is some indication that the 
line still increases in strength 500 pc 
above the galactic plane. There are nine 
stars whose spectra have been obtained 
at sufficient dispersion so that velocities 
are meaningful. After correction for solar 
motion and the normal galactic rotation 
terms, the residual velocities of the 
interstellar clouds at high latitudes are 
found to run from +7 to —34 km/sec. 
Of the nine, only one has a positive 
residual cloud velocity. The preponder- 
ance of negative velocities is similar to 
that found in the 21-cm line observa- 
tions of high-latitude neutral hydrogen. 
However, the velocities are very much 
lower, presumably reflecting the higher 
density of the clouds. The line is also 
observed to be surprisingly strong in very 
blue globular-cluster stars ; spectra at 90 
A/mm are available for a few. There is 
an apparent paradox in that the K line 
sometimes gives a high velocity similar 
to that of the globular cluster itself. It is 
possible that there is some contamination 
from the background light of red sub- 
giants on the slit of the spectrograph. 
In M 13, the star Barnard 29 observed 
at 18 A/mm showed a star velocity of 
— 244 and a K-line velocity of — 33 km/ 
sec, suggesting that the K line was inter- 
stellar in origin. In other cluster stars it 
is clear that the K line is partly stellar 
and partly interstellar in M 13. Where 
the cluster velocity is about —245 km/ 
sec, the K-line velocity observed runs 
from -141 to -33. 



Orion Nebula 

The Orion Nebula has been mapped ex- 
tensively by Greenstein with emphasis on 
the spectral region near the Balmer series 
limit. Coude spectrograms at 18 A/mm 
were used. Because of velocity broaden- 
ing, the lines begin to merge near princi- 
pal quantum number n = 35. However, 
the shape of the blended intensity curve 
in this region, and also near n = 75 where 
the lines are completely blended, and 
just beyond the series limit have all 
been measured. Plates have also been 
taken to show the doublet ratio of the 
[O II] lines of AA3726 and 3729. In co- 
operation with Prof. Goldberg at 
Harvard, and a student, Don Haler, an 
attempt has been made to fit the run of 
intensity of the higher members of the 
Balmer series to obtain the values of 
b n (the occupation numbers) for high 
values of n. This is quite sensitive to 
electron density, at densities below n e = 
10 4 , and is also somewhat sensitive to 
temperature if the electron temperature 
is very low. General agreement with the 
expected curves computed by Haler has 
been found ; new collisional cross sections 
have been derived for this purpose. A 
quite low electron temperature has been 
derived for NGC 7027, suggestive of a 
very efficient cooling mechanism. Gold- 
berg, on the basis of elaborate computa- 
tions of collisional excitation cross sec- 
tions, has suggested that the cooling 
mechanism is the forbidden emission of 
the singly and doubly ionized metals of 
the iron group. There is a complicating 
link between the derived electron tem- 
perature and electron density, but in the 
Orion Nebula the electron density can 
be checked by the variation of the A3726/ 
A3729 ratio on the assumption that the 
hydrogen and the forbidden oxygen 
lines are produced in the same region. In 
the very inhomogeneous planetary nebula 
this method may not be so successful. 

Neugebauer, Becklin, and Hilgeman 
have continued infrared studies of the 
Orion Nebula and other regions of pos- 



32 



CARNEGIE INSTITUTION 



sible star formation. In Orion, 2.2- and 
1.65-/* scans were made on the 60-inch 
telescope of the region surrounding the 
infrared point source, the nonthermal 
OH source (observed by Raimond and 
Elliasen, 1967), and the 20-ju, nebula 
(Kleinman and Low, 1967). The scans 
showed a large concentration of 2.2-//, 
radiation in the region and an additional 
point source of infrared radiation. The 
source has the same color but is fainter 
than the first Orion-nebula infrared 
source found. No image of the source 
is seen on photographic plates. Low-reso- 
lution infrared spectral scans of the 
Orion nebula were also made with the 
24-inch telescope on Mount Wilson. De- 
tection of the Brackett-y line of hydro- 
gen was made. The strength of this line 
will be useful in providing information 
about the radiation from the nebula, as 
well as the physical characteristics of 
the nebula itself. 

Observations at 2.2 and 1.65 /x were 
also made on the 200-inch and 60-inch 
telescopes at the positions of three non- 
thermal OH sources, W3, W51, and W75. 
Infrared radiation was detected near all 
three sources. The sources at W75 and 
W51 appear extended, whereas the 
source near W3 is pointlike. None of the 
infrared sources associated with OH 
emission can be identified with a visible 
object on the Palomar Sky Survey 
schmidt prints. 

In their survey of the internal kine- 
matics of the Orion Nebula, Wilson and 
Munch discovered that in certain areas 
the emission lines become very broad or 
even double (Year Book 56, p. 55). In 
order to establish the origin of the large 
mass motions revealed in this fashion, 
Munch has started a detailed study of 
line profiles in representative areas of the 
nebula, using Fabry-Perot interferom- 
eters of various resolving powers. The 
much greater efficiency of the photo- 
electric detector and larger throughput 
of the interferometer allow the study 
of lines fully two orders of magnitude 



fainter than was possible photograph- 
ically. The observed complex profiles of 
the line [0 III] A4363 and their com- 
parison with those of the Nl and N2 
lines have shown that the electron tem- 
perature is essentially the same in the 
two gas masses that give rise to the 
double lines. If the line-doubling is a 
manifestation of a strong shock wave, 
then this shock is nearly isothermal. Fur- 
ther observations in the [0 II] and [0 I] 
lines are planned in order to obtain in- 
formation about densities and ionization 
conditions in the various line components. 

Planetary Nebulae 

The internal motions in the planetary 
nebula NGC 6543, which on visual in- 
spection has a markedly helical form, 
have been studied by Munch on multi- 
slit plates taken at the 200-inch coude 
with 4.5 A/mm dispersion. The radial 
velocities of various points in the nebula 
that cover a range between —100 and 
— 25 km/sec suggest that the nebular 
material is contained on two helical sur- 
faces nearly symmetrical with respect 
to the central star and moving away 
from it. The observed arms are not iso- 
lated features but are condensations on 
these helical sheets, which contain suffi- 
cient matter to shield the outer regions 
from the ultraviolet radiation of the 
central star, as required by a pronounced 
ionization stratification. 

Crab Nebula 

Further study of the motions of the 
line-emitting filaments in the Crab 
Nebula were carried out by Virginia 
Trimble [Year Book 66, p. 277). The 
measurement of proper motions for fila- 
ments with known radial velocity has al- 
lowed the construction of three-dimen- 
sional models for the nebula. The mo- 
tions of the filaments together with other 
nondynamical data indicate a most prob- 
able distance to the object of 2.02 kpc. 
The proper motion and radial velocity 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



33 



derived for the nebula as a whole lead 
to the conclusion that the nebula is mov- 
ing 112 km/sec faster than the galactic 
rotation at its position. The proper mo- 
tion of the so-called central star trans- 
lates into an equally unlikely space 
motion, but this result, as well as the 
motion of the whole nebula, is quite un- 
certain. The likelihood of a physical con- 
nection between the star and the nebula 
has thus been diminished, but not com- 
pletely ruled out, as a consequence of 
these measurements. 



The line-emitting filaments have been 
found to be distributed throughout the 
nebula rather than on a thin outer en- 
velope, as previously supposed. The mo- 
tions of the filaments are nevertheless 
largely radial, each having a velocity 
nearly proportional to its distance from 
the expansion center. The mean devia- 
tion from this proportionality is 0.01 
arc-sec/yr, or 70 km/sec, an amount 
compatible with the scatter of the fila- 
ments around the expansion center in the 
year 1140 A.D. 



INFRARED SKY SURVEY 



The infrared survey of the sky funded 
by the National Aeronautics and Space 
Administration was completed by Neuge- 
bauer and Becklin with the assistance 
of Edward Groth. A catalog of about 
5500 objects, essentially complete to 
K = 3, is being prepared for publication 



this summer. Descriptions of the instru- 
mentation and preliminary results have 
been given in previous Year Books. It is 
currently planned to transport the tele- 
scope from Mount Wilson to Cerro 
Tololo in Chile, and thus to survey the 
entire sky. 



OBSERVATIONAL COSMOLOGY 



New Determination of the Hubble 
Constant 

A new method to determine the value 
of the expansion rate of the universe 
has been used by Sandage. Difficulties 
with previous methods have centered 
around a possible local non-Hubble flow 
in the expansion field for redshifts in the 
range 0<cAA/A o < 4000 km/sec due to 
perturbations of the local nonhomoge- 
neous distribution of matter. The new 
method circumvents this difficulty by 
(1) reading the apparent-magnitude- 
redshift relation (the Hubble diagram) 
for first-ranked E galaxies in clusters at 
redshifts beyond the local anisotropy 
(cAA/A ~ 10,000 km/sec), (2) using 
the fact that such cluster galaxies are 
remarkably constant in absolute lumi- 
nosity [o- {My) ~ ±0.26 mag], and (3) 
calibrating My of the first-ranked E 
cluster galaxy with NGC 4472, which is 



the brightest E galaxy in the Virgo 
cluster. 

The distance to the Virgo cluster was 
found from Racine's photometry of 200- 
inch plates of globular clusters in NGC 
4486 (M 87) taken by Sandage. Racine 
showed that the luminosity function 
for these clusters is steep and begins near 
£ = 21.3 in M 87. If the brightest of 
these star clusters has the same My as 
the brightest globular cluster in M 31 
at M v = —9.83, then the apparent modu- 
lus of the Virgo cluster in blue light 
(X~ 4400 A) is m-M= 31.1. The abso- 
lute magnitude of NGC 4472 is deter- 
mined to be M B =— 21.68, which, when 
combined with a new Hubble diagram 
obtained by Sandage for 42 first-ranked 
E galaxies in clusters, gives for the 
Hubble constants = 75.3-1! km/sec Mpc, 
or H- 1 = 13-U xlO 9 years. 

This determination gives an upper 
limit because the brightest globular 



34 



CARNEGIE INSTITUTION 



cluster of the 2000 clusters in M 87 is 
likely to be brighter than the most lumi- 
nous cluster in M 31 with only 250 
globular-cluster members. Taking this 
statistical effect into account, it seems 
possible that 77 could be as low as 50 
km/sec Mpc, or 77 _1 could be as large as 
19xl0 9 years. The importance of this 
result is that it removes the difficulty 
with the time scale given by older esti- 
mates of 77 that were as high as 120 km/ 
sec Mpc (77- 1 = 8.1 X 10 9 years) . If H~ x = 
19xl0 9 years, and if the deceleration 
parameter is q ~ +1, then the time 
since the fireball is 0.571 77; 1 =llxl0 9 
years, which is consistent with the age 
of the globular clusters (T ~ 10xl0 9 
years) and the age of the heavy chemical 
elements (T^O.OxlO 9 years) that make 
up the earth's crust. 

Work by Kristian and Sandage has 
continued on a second method of finding 
77 by use of the angular sizes of ionized 
hydrogen patches in nearby Sc galaxies. 
The data are not yet sufficient for a solu- 
tion. The value of 77 is of such obvious 
importance to all theories of cosmology 
that all efforts to check the new method 
are desirable. 

New Values for the Redshift 
K-Correction 

The corrections to measured intensities 
of redshifted galaxies to compensate for 
the shifting of the continuum energy 
distribution through the measuring bands 
must be accurately known before the 
deceleration parameter, q , can be mea- 
sured via the Hubble diagram. 

Oke and Sandage completed a new 
determination of K B , K v , and K B for 
the broadband B, V, and R magnitude 
systems as a function of redshift in the 
range 0<AA/A a <0.30. New measure- 
ments of the energy distribution of 8 



giant E galaxies were completed in the 
wavelength range 3200 A<A <11,000 A 
with a spectral resolution of 50 A. It 
was found that /(A) was identical for 
all eight E galaxies, supporting the view 
that elliptical systems with i!7 F; > — 18 
mag are similar in their stellar content. 

The individual results were combined 
to form a mean 7(A) function that was 
numerically redshifted through the pho- 
tometric sensitivity functions, S(\), of 
the three broadband filter functions. 

New photoelectric measurements of 
the B — V color of the first several 
brightest galaxies in 32 clusters, whose 
redshifts range from 0<AA/A <0.20, 
were compared with the calculations. 
The predicted K B — K v = f{A\/X ) agree 
with the observations over the entire 
redshift range to better than ± 0.02 mag, 
showing the absence of the Stebbins- 
Whitford effect, at least to this level of 
accuracy. Such agreement shows that 
no change in the shape of 7(A) has 
occurred due to evolution of the stellar 
content of these E galaxies during the 
look-back time of -O.ISV 1 ~ 2.5 xlO 9 
years. These data provide constraints on 
the change of integrated luminosity of 
E galaxies due to evolution of the stellar 
content. The results are consistent with 
a new estimate by Sandage that dV / 
dt <^.0.05 mag/10 9 yrs from considera- 
tions of the evolutionary pattern of the 
relevant stars. 

The new 7£- corrections have been used 
to refine the Hubble diagram reported 
in previous Year Books. The ultimate 
purpose is to determine the deceleration 
parameter for a decision between an 
ever-expanding model or an oscillating 
universe. The new data, which have been 
analyzed in a preliminary fashion by 
Peach and by Sandage, favor the oscillat- 
ing case with a total cycle time of 
about 80 X 10 9 years. 



GALAXIES 

Distance to the Virgo Cluster and magnitudes for the globular-star 

Racine has used plates of M 87 taken 



with the Hale telescope to obtain colors 



clusters associated with this giant ellipti- 
cal galaxy, which is a member of the 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



35 



Virgo cluster of galaxies. These globular 
clusters in M 87 are found to have a 
mean intrinsic color ((B — V))= +0.74, 
which is not significantly different from 
the mean colors of globular clusters in 
the Galaxy or in M 31. The apparent- 
luminosity function of the M 87 clusters 
rises steeply at 5 = 21.2 mag, and ap- 
pears to reach a maximum at .6 = 23.2 
mag. By comparing with the luminosities 
of the globular clusters in the Galaxy 
and in M 31, the distance of the Virgo 
cluster of galaxies is found to be 12 Mpc, 
and a value of 79 ±12 km/sec/Mpc is 
obtained for the Hubble constant. 

Galaxies in Groups 

Sargent has begun a study of the mo- 
tions of galaxies in small groups by ob- 
taining spectra of galaxies in four sys- 
tems discovered by Vorontsov-Velya- 
minov— VV 144, 150, 165, and 172. 
(These objects are all illustrated in Arp's 
Atlas of Peculiar Galaxies.) VV 144, 
which is also I Zw 1122 + 54, consists of a 
nucleus that E. M. and G. R. Burbidge 
found earlier to have a broad Ha emis- 
sion line in its spectrum. A jet projecting 
from the nucleus has a prominent knot 
halfway along it. Image-tube spectra of 
the nucleus of W 144 confirmed that 
it resembles a Seyfert galaxy. There are 
broad emission lines of hydrogen and 
much sharper emission lines of [0 II] 
and [0 III] in the blue region of the 
spectrum. Spectrograms were obtained 
with the slit along the jet of W 144. 
Although the jet is easily visible to the 
eye at the telescope, no trace of its spec- 
trum was found, so that it clearly does 
not radiate a predominantly emission- 
line spectrum. On the other hand, the 
knot has an absorption-line spectrum 
similar to that of an A5 star. The red- 
shift of the knot is 6467 km/sec, as com- 
pared to 6169 km/sec for the nucleus. 
This leads to a minimum mass for the 
system of about 2 x 10 11 Wl O . 

VV 150 and W 165 both consist of 
irregular chains, each containing four 
galaxies. The spectra of these objects 



have not yet been measured, but it is 
evident from visual inspection that in 
each system the random velocities of the 
component galaxies do not exceed a few 
hundred kilometers per second. However, 
this is not the case for VV 172. This 
system consists of an almost linear chain 
of five galaxies about 1 arc-min in pro- 
jected length and fairly isolated from 
other galaxies of similar brightness. At 
least two image-tube spectra have been 
obtained of each galaxy in the chain. 
If we denote the galaxies by A, B, C, D, 
and E, proceeding north to south along 
the chain, the redshifts are found to be 
7^ = 16,070, V B = 36,880, V = 15,820, 
V D = 15,690, and 7^ = 15,480 km/sec. 
Thus galaxy B has a redshift that is 
20,000 km/sec greater than those of the 
other four galaxies. This galaxy is the 
only one that shows an emission line, 
A3727 of [O II], in addition to H and K 
absorption lines. Neglecting galaxy B, 
the other four galaxies in W 172 have a 
systematic trend of redshift along the 
chain, indicative of rotation. This leads 
to a lower limit for the mass of the 
whole system of ~4xl0 lx WlO. It is 
clear that galaxy B is not gravitationally 
bound to the system if its redshift is a 
true Doppler shift. Preliminary calcu- 
lations indicate that the sharpness of the 
emission and absorption lines in galaxy 
B rule out the possibility that its excess 
redshift is gravitational. We are left 
with the possibilities either that galaxy 
B is a background galaxy or that it 
has been expelled from the chain with a 
kinetic energy of about 10 60 ergs about 
10 7 years ago. Sargent intends to in- 
vestigate other small groups of interact- 
ing galaxies in the future. However, it is 
already clear that VV 172 is not unique. 
It merely displays in an extreme form a 
kind of phenomenon already discovered 
in Stephan's Quintet and in the inter- 
acting triple system IC 3481, Anon, 
IC 3483. In both these last two cases, 
the anomalous galaxy has a much smaller 
redshift than the rest of the group. 



36 



CARNEGIE INSTITUTION 



Association of Galaxies and Radio 
Sources 

Arp continued his investigation of the 
apparent association of objects with at- 
tention to the surroundings of the 13 
brightest E galaxies. He concluded that 
10 of these galaxies appear to be mem- 
bers of chains of galaxies, and that where 
radio-source extensions from the central 
galaxy are well marked, the lines of these 
extensions coincide with the line of di- 
rection of the chain. 

Markarian Blue Galaxies 

Following a visit by Dr. E. Khachikian 
of the Burakan Astrophysical Observa- 
tory, Armenia, U.S.S.R., Sargent began 
a program to obtain spectra of blue 
galaxies discovered by Markarian which 
had not previously been observed. Satis- 
factory spectra were obtained of 
Markarian Nos. 18, 20, 21, 29, 38, 43, 
47, and 48, of which Nos. 20, 38, 39, and 
43 revealed sharp emission lines. 

Seyfert Galaxies 

Oke and Sargent have done further 
work on NGC 4151. The forbidden 
lines and cores of the hydrogen lines are 
formed in clouds with ^ = 5000 cm -3 
and T e = 20,000 °K; these clouds fill only 
a small fraction of the total volume of 
the nucleus. The intensities of the ob- 
served coronal lines can be explained 
by postulating a hot gas with T e = 10 a °K, 
which fills the entire nucleus of the 
galaxy and for which ^ = 100 cm -3 . The 
cool clouds and hot medium are nearly 
in pressure equilibrium. The origin of 
the very broad wings of the hydrogen 
lines and the ionized helium line A4686 
is still uncertain. They could be pro- 
duced by a very dense gas in violent mo- 
tion or possibly by electron scattering. 

Infrared Observations 

Neugebauer and Becklin also have be- 
gun a program of measuring Seyfert 



galaxies, compact galaxies, radio gal- 
axies, and related objects at 1.65 and 
2.2 fi. About 12 of these galaxies have 
been measured so far. Several show an 
infrared excess suggestive of a non- 
thermal source of radiation. 

Dwarf Systems 

In combining the period-luminosity 
curves of variables in the Leo II system 
with those of other dwarf systems, Miss 
Swope noticed that the variables brighter 
than the usual RR Lyrae cluster varia- 
bles found in these systems may form 
a period-luminosity curve of their own 
that falls somewhere between that of 
the classical Cepheids of Population I 
and the W Virginis-type Cepheids found 
in globular clusters. However, the ma- 
terial is still too limited to permit deriv- 
ing the slope of the curve. But both the 
Ursa Minor system of van Agt and the 
Leo II system seem to indicate that this 
curve extends to variables with period in 
the usual RR Lyrae range but brighter 
than the other RR Lyrae variables in 
the system. Otherwise the work on the 
Leo II system still awaits a photoelectric 
sequence in order to determine the photo- 
visual scale and whether there is any 
reddening in the system. 

Polarization 

Sandage and Visvanathan have ob- 
served photoelectrically at the prime 
focus of the 200-inch telescope the large- 
scale filaments of M 82 for polarization 
and colors in the U, B, V system. The 
apertures used were 17" and 28" in 
diameter. A preliminary analysis shows 
the following: (1) The percentage of 
polarization varies from 13% to 32%. 
(2) The position angle of the electric 
vector is found to be perpendicular to 
the filamentary structure. (3) Colors 
show large ultraviolet excess and they fit 
with the energy spectrum of synchrotron 
radiation. 

A program by Visvanathan is in prog- 
ress to observe polarization in Seyfert 



MOUNT WILSON AND PALOMAB OBSERVATORIES 



37 



galaxies, N-type galaxies, and compact 
galaxies. A number of optically variable 
quasi-stellar sources, N-type galaxies, 
and compact galaxies are being observed 
for polarization and colors. 

Visvanathan and Oke have separated 
the continuum of the Seyfert galaxy into 
thermal and nonthermal components 
from the polarization observations and 
scanner data. The polarization data were 
obtained with different apertures and the 
use of U, B, V, and R niters at the 100- 
inch and 60-inch telescopes. Polarization 
is high in the ultraviolet and gradually 
decreases toward longer wavelengths. 
Scanner observations from ultraviolet to 
infrared have been obtained at the prime 
focus of the 200-inch telescope. By as- 
suming that the nonthermal component 
has no wavelength dependence of polar- 
ization and that the observed change in 
percentage polarization with A is due to 
the addition of unpolarized galaxy radia- 
tion, the continuum of NGC 1068 is 
separated into thermal and nonthermal 
components. The derived nonthermal 
component is smooth and similar to the 
energy distribution of the quasi-stellar 
source 3C 273 from 0.33 to 1.0 /*. 

Radio Galaxies 

Oke discovered that the nucleus of the 
N-type radio galaxy 3C 371 changed 
by more than 1 mag during the past two 
years and could change by 0.10-0.15 mag 
over intervals of a few days. The abso- 
lute luminosity is approximately 10% 
of that of 3C 48 if 3C 48 is assumed to 
be at a cosmological distance. It can be 
concluded that rapid variability cannot 
now be used in an argument against the 
cosmological nature of quasars. Using the 
spectral-energy distribution and the ob- 
served strengths of the H and K lines 
and Mg II line, it is possible to separate 
the radiation of the background galaxy 
from that of the nonthermal variable 
nucleus. 

Oke, Sargent, Neugebauer, and Beck- 
lin have made scanner observations and 



infrared measures of two Zwicky com- 
pact galaxies, I Zw 1727 + 50 and 3C 120. 
The first of these is variable in light and 
color, and there must be a substantial 
nonthermal component. The fact that 
the energy distribution flattens in the 
infrared suggests, however, that the stel- 
lar contribution is by no means small. 
The spectral-energy distribution in 3C 
120 continues to increase rapidly in the 
infrared, indicating a strong nonthermal 
contribution to the observed radiation. 
Oke is continuing to obtain spectral- 
energy distributions of radio galaxies 
such as NGC 1068, NGC 1275, and 
Cygnus A. The last is an extremely red 
object, even when corrections are made 
for reddening, and its large optical size 
suggests that there must be a large stellar 
contribution to the radiation. The ac- 
curacy of the observations is not yet 
sufficient to indicate whether the H and 
K lines are present in absorption. 

Radio Observations of Peculiar Galaxies 

Most of the Zwicky compact galaxies 
do not appear in the catalogs of radio 
sources, despite the fact that many of 
these galaxies have spectra that re- 
semble those quasi-stellar objects (e.g., 
I Zw 0051 + 12) or radio galaxies. It is 
possible that many of these galaxies are 
radio sources which, however, are op- 
tically thick at frequencies below 1000 
Mc/s where the conventional surveys, 
such as the 4C, have been carried out. 
Accordingly, Sargent collaborated with 
Dr. A. T. Moffet of the California Insti- 
tute of Technology in a search for radio 
emission from peculiar galaxies. Obser- 
vations were made with the Owens Val- 
ley interferometer at a base line of 500 
feet EW at wavelengths of 21 cm and 
3 cm. The objects observed included 50 
galaxies in Zwicky's lists and 25 Mark- 
arian blue galaxies. At 3 cm none of the 
peculiar galaxies were detected, indicat- 
ing an upper limit to the radio flux of 
about 0.2 flux units. At 21 cm a few ob- 
jects were detected, including Markarian 



38 



CARNEGIE INSTITUTION 



3 which is in the 4C catalogue. However, 
none of the galaxies with broad emission 
lines in their optical spectra and with 
colors similar to quasi-stellar objects 
were detected at either wavelength. 

Catalogue of Galaxies and of Clusters of 
Galaxies 

The work on all six volumes covering 
the unobscured areas from Dec —3° to 
+ 90° has now been completed. The 
manuscripts of the last two volumes, 
Nos. IV and VI, will be sent to the 
printer within the next few months. An 
important contribution to these volumes 
was made by Herzog with his invention 
of a simple graph from which values of 
the precession in all parts of the sky can 
immediately be measured off. It is ex- 
pected that the finished volumes will 
become available at the end of the year. 
The compilation has been produced un- 
der the direction of Zwicky by Herzog, 
Kowal, Gates, Barbon, Fairall, and 
others, with funding by the National Sci- 
ence Foundation. 

Compact Galaxies 

While surveying the 48-inch schmidt 
plates for clusters of galaxies to be in- 
cluded in the Catalogue, Zwicky has 
continued to select the more obvious 
compact galaxies. In addition to the five 
lists previously issued by Zwicky, two 
more lists (VI and VII) have been as- 
sembled which contain 222 and 938 ob- 
jects, respectively. Altogether the seven 
lists contain 2309 objects. Of these, sev- 
eral hundred have been investigated di- 
rectly and spectroscopically by Zwicky, 
Arp, Oke, Sargent, and Neugebauer at 
Palomar, and, working with Zwicky, by 
R. Barbon and T. Fairall using the 
large reflectors at the Kitt Peak and 
McDonald Observatories. Some of the 
main results found are that among the 
compact galaxies are objects which 
match the spectra, colors, and absolute 
magnitudes of other extragalactic objects 



that were originally thought to be unique, 
such as the Seyfert galaxies, N galaxies, 
and quasars. 

Most important for cosmological 
theory is the fact that some of the 
extremely compact galaxies that appear 
starlike when photographed with the 
200-inch have larger, and occasionally 
very much larger, redshifts than the 
supposedly brightest cluster galaxies of 
the same magnitude. Examples are: 

VII Zw 1354.3 + 6457. An almost star- 
like emission-line object even with the 
200-inch telescope. Estimated m pg = l§A, 
U-B= -0.8, 5-7= +0.6, and sym- 
bolic velocity of recession 7 S = 2100 km/ 
sec, that is, indicative absolute magnitude 
M pg =-20.2. 

II Zw 0430.5 + 0515. A new intrinsi- 
cally bright Seyfert emission-line galaxy 
of variable brightness m = 13.1 to 14.6. 

I Zw 1727.1 + 5015. A variable radio- 
quiet galaxy with featureless spectrum. 

I Zw 0051.0 + 1225. A luminous com- 
pact galaxy with strong Fe II emission 
lines and 7 = 13.90, U-B= -0.80, and 
5-7 = 0.36. With 7 S = 18150 km/sec, 
an indicative absolute magnitude M v = 
— 22.6 is obtained. 

II Zw 2130.0 + 0956. A new very bright 
Seyfert galaxy. With ra p = 14.80 and 7 S = 
18750 km/sec, the indicative absolute 
magnitude becomes M p = —21.6. 

The above data are based on observa- 
tions by Zwicky, Arp, Sargent, and Fair- 
all. These findings demonstrate that there 
exist compact galaxies which form a 
continuous link between the ordinary 
galaxies and the quasars. 

Sargent has made spectroscopic ob- 
servations of 127 objects in Zwicky 's 
first five lists of "Compact Galaxies and 
Compact Parts of Galaxies." About 100 
of these spectra were obtained with the 
image-tube spectrograph at the Casse- 
grain focus of the Hale telescope. The 
rest were obtained with the prime-focus 
spectrograph. Forty-six of the galaxies 
have spectra containing emission lines. 
Usually the emission lines are sharp, 



MOUNT WILSON AND PALOMAR B S EE V AT ORIE S 



39 



but five galaxies have broad emission 
lines similar to those in the spectra of 
Seyfert galaxies. If galaxies in Zwicky's 
lists are selected for observation on the 
basis that they appear blue on the 
Palomar Sky Atlas prints, the propor- 
tion of emission-line objects is greater 
than 50%. Some details of the more 
interesting objects are given below. 

1. The "iron galaxy" I Zw 0051 + 12. 
The spectrum of this object has been 
described by Sargent in Astrophys. J. 
(Letters), 152, L31, 1968. The galaxy 
has a starlike nucleus of m„=:14.0, which 
gives a spectrum composed of broad 
emission lines of H and Fe II in a blue 
continuum. With a redshift of 18,150 
km/sec, this object has M v =— 23.1 and 
is as luminous as some quasi-stellar 
sources if the latter are at cosmological 
distances. 

2. II Zw 2130+09. This is an N-type 
galaxy having a starlike nucleus of mag- 
nitude F=14.6. This object has a spec- 
trum similar to that of a Seyfert galaxy 
with very broad emission lines of H and 
He II and sharper emission lines of 
[O II] and [O III]. However, it is un- 
usual in that weak permitted emission 
lines of Fe II are also visible. This object 
has a redshift of 18,350 km/sec and 
M v = —22.7. There is a preliminary indi- 
cation that Fe II emission lines are indic- 
ative of high intrinsic luminosity, since 
Wampler and Oke have identified these 
lines in the spectrum of 3C 273. 

3. Several compact galaxies, notably 
II Zw 0553 + 03, II Zw 0508-02, II Zw 
0448 + 03, 1 Zw 1535 + 55, 1 Zw 1634 + 52, 
and II Zw 1448 + 35, have very strong 
emission lines principally of H, [O II], 
and [O III] on a blue continuous back- 
ground. All these objects have small red- 
shifts and appear to be intrinsically 
faint. For example, II Zw 0553 + 03 has 

a redshift of 617 km/sec and My 15. 

No stellar absorption lines have been 
seen in the spectra of these objects; the 
mode of excitation of the hot gas re- 
mains obscure. 



Sargent made UBV observations of 
about 30 compact galaxies in an attempt 
to relate the colors to the spectra of these 
objects. The following preliminary re- 
sults were obtained : 

1. The galaxies having very broad 
emission lines have colors that fall in the 
region occupied by QSOs in the (B — V, 
U—B) diagram. 

2. The objects having sharp emission 
lines occupy a band below the QSOs. 
Although they are blue, they do not have 
such pronounced ultraviolet excesses as 
the QSOs. 

3. The objects having weak, sharp 
emission lines and A-type absorption 
spectra fall in the region of the (B — V, 
U—B) diagram occupied by the Magel- 
lanic Irregulars. 

4. The compact galaxies with G-type 
absorption spectra have colors similar to 
ordinary elliptical galaxies. This is borne 
out by observations of the continuous 
energy distribution of a few such objects. 

Objects in Zwicky's list of compact 
galaxies continue to be observed by Arp 
as well as by Sargent. Arp reports a 
nearly stellar blue compact, VII Zw 
1354 + 64, with a Seyfert spectrum with 
the relatively high redshift of 2 = 0.075. 
A blue compact at 17 h 9 m + 64°17' has an 
early-type absorption spectrum and the 
relatively high redshift of 2 = 0.90. Arp 
has taken direct, good-seeing photo- 
graphs and numerous spectra of 3C 371, 
the galaxy first noted by Oke to be 
variable in optical light. Aside from a 
very small, faint haze directly around 
the nucleus, the galaxy is shown to be 
optically stellar. Low-contrast emission 
lines have been recorded in the spectrum, 
and analysis of the material is proceed- 
ing in collaboration with Visvanathan. 

Clusters of Compact Galaxies 

Among the 2309 objects in Zwicky's 
seven lists there are many close double, 
triple, and multiple compact galaxies. 
Most interesting, however, are larger 
groups and clusters of compact galaxies 



40 



CARNEGIE INSTITUTION 



which contain dozens to hundreds of ob- 
jects. Some of the groups found by 
Zwicky are the following: 

Zw CI 0152.0 + 3337. A cluster com- 
posed of very compact galaxies, the 
brightest of which have m p = l7.3. 

Zw Gr 0338.3 + 1510. A beautiful group 
of about 50 compact galaxies, including 
some close pairs. The brightest members 
have m p = 16.3. 

Zw CI 0658.3 + 6320. A rich medium- 
distant cluster with 189 galaxies listed 
in the Catalogue. The redshift of galaxy 
No. 1, that is of Zw CI 0658.3 + 6320-1, 
is V s = 28,600 km/sec. The brightest 
members have m p s* 16.9. 

Zw CI 1710.4 + 6401. A rich cluster of 
345 medium-compact and compact gal- 
axies. The symbolic velocity of recession 
is V s = 23,820 km/sec for member No. 3, 
and 21,440 km/sec for No. 5. The bright- 
est member has about m p = 15.9. 

Zw CI 1849.1 + 7724. An open cluster 
containing about 325 compact galaxies. 
For the brightest members, ra p = 15.7. 

Sargent obtained image-tube spectra 
of the brightest galaxy in each of five 
Abell clusters— Nos. 655, 963, 2029, 2204, 
and 2224 — in order to determine red- 
shifts as part of a long-term program 
in cooperation with Sandage and 
Kristian. Spectra were obtained also of 
galaxies in two clusters of compact gal- 
axies discovered by Zwicky. Two gal- 
axies in Zw CI 1710 + 64 gave redshifts 
of 21,440 and 23,820 km/sec. This is a 
preliminary indication that this cluster 
has an unusually large velocity disper- 



sion. A redshift of 28,600 km/sec was 
obtained for a galaxy in Zw CI 0659 + 63. 

Sizes of Rich Clusters oj Galaxies 

From a preliminary survey (Publ. 
Astron. Soc. Pacific, SO, 218, 1938) with 
the 18-inch schmidt telescope, Zwicky in 
1938 determined the size of the average 
cell occupied by each of the hundred 
nearest rich clusters of galaxies as D = 
7.8 million parsecs, assuming a Hubble 
constant if = 550 km/sec million pc, or 
.D = 43 million pc indicative (for H — 
100 km/sec million pc). Analyzing the 
distribution of 9800 clusters of galaxies 
listed in the six volumes of the Cata- 
logue, Zwicky has now redetermined D 
and obtained D = 40 million pc, in good 
agreement with the value derived in 
1938 and the long-standing prediction 
that no clusters of clusters exist among 
the objects listed in the Catalogue. 

Area of the Sky Covered by 
Clusters of Galaxies 

The projected central areas of the 
rich clusters of galaxies to distances 
corresponding to symbolic velocities of 
recession V s = cz = 75000 km/sec, as listed 
in the Catalogue, in the least obscured 
areas cover 60% of the sky. To distances 
corresponding to 2 = 2 instead of z = %, 
these central parts of the clusters would 
cover the sky five times, or actually 
many more times if the increase in 
density with distance, in agreement with 
the cosmological interpretation of the 
large redshifts of quasars, were taken 
into account. 



THE GALAXY 



Studies of the center of the Galaxy 
at infrared wavelengths have been con- 
tinued by Neugebauer and Becklin. 
With the 60-inch telescope, spatial scans 
at 1.65 and 2.2 fx were made of a region 
2° x 1° around the galactic center. In ad- 
dition to producing much more precise 



color measurements of sources at the 
center, the scans show that most of the 
extended infrared sources lie within y 2 ° 
of that point. In the spring of 1968, simi- 
lar observations were made, also at 5.0 
and 10 p. The data are still being re- 
duced. Low-resolution (~400 A) spec- 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



41 



tral scans of the central "point" source 
at the galactic center were obtained with 
the 200-inch telescope. These scans, 
which cover the wavelength regions be- 
tween 2 and 2.5 n, will be compared in 



detail with similar scans of the extended 
background obtained at the Newtonian 
focus. No strong emission or absorption 
features were observed. 



SUPERNOVAE 



The search for supernovae with the 
Palomar schmidt telescopes, supported in 
large part by the National Science 
Foundation, has been continued. Ten 
supernovae (1 each by Herzog and 
Kowal, 8 by Zwicky) have been dis- 
covered in the period from June 1, 1967, 
to May 31, 1968. In the same period, 
supernovae were found by the members 
of the Committee for Research on Super- 
novae within Commission 28 of the In- 
ternational Astronomical Union. Most 
remarkable among these is the fourth 
supernova found in NGC 6946 since 
1917. The statistical investigation by 
Zwicky of multiple appearances of super- 
novae in certain galaxies indicates that 
there are stellar systems which in a 
given long period produce up to 10 times 
as many supernovae as other systems 
that structurally and in other known 
respects have the same characteristics. 

Absolute Magnitudes 

Indicative absolute photographic mag- 
nitudes of 23 supernovae were deter- 
mined by Kowal and it was found that 
they are grouped about an absolute value 
iW p = 18.6, with a dispersion of about 
AM P = 0.75. Kowal also plotted a new 
symbolic velocity (V s ) —apparent-mag- 
nitude (m p ) relation and found (using 
7 supernovae of type I in the Virgo and 
Coma Clusters) 5 log V s = m p + 3.59, in 
good agreement with the relation first 
derived by Zwicky in 1961 (Publ. Astron. 
Soc. Pacific, 78, 185). From 15 super- 
novae in cluster galaxies and field gal- 
axies, Kowal derives 5 log V s = m p + 327. 



Frequency 

The frequency of supernovae was 
originally determined by Zwicky (Astro- 
phys. J., 96, 28, 1942) from his survey, 
with J. J. Johnson, of 836 galaxies listed 
in the Shapley-Ames Catalogue, using 
the 18-inch schmidt telescope. The net 
result was a frequency of one super- 
nova per 359 years per galaxy, on the 
average. No preference was found for 
any of the various structural types of 
galaxies, except that no supernovae of 
the type II occurred in any spherical 
or elliptical galaxies. Barbon, in col- 
laboration with Zwicky, has now rede- 
termined the average frequency of ap- 
pearance, v. A total of 2144 galaxies 
brighter than ra p = 15.7 listed in the 
Catalogue and 45 supernovae appearing 
in them September 1957 and June 1967 
were considered. The net result was v r = l 
supernova per 316 years per galaxy, in 
close agreement with Zwicky's original 
result. 

Separating the 45 supernovae into 37 
that appeared in cluster galaxies and 8 
in field galaxies, Barbon obtained 

f c i=l supernova per 335 years per 
cluster galaxy 

Afield = 1 supernova per 229 years per 
field galaxy. 

Photometry and Light Curves 

Zwicky, Rudnicki, Kowal, Herzog, 
and Fairall have constructed the light 
curves of about two dozen supernovae, 
which will be published in the near 
future. Photometric data for all known 
supernovae have been compiled by 
Kowal, who also derived indicative abso- 
lute magnitudes for about 40 objects for 



42 



CARNEGIE INSTITUTION 



which the maxima have been observed. 
For type I the maximum photographic 
magnitude is —18.6, and for type II the 
maximum is —6.5, on the average. It is 
anticipated that supernovae can be used 
soon to derive the Hubble constant and 
the second-order term in the magnitude- 
redshift relation. 

Spectra 

Sargent obtained spectra of five super- 
novae with the image-tube spectrograph 
at the 200-inch telescope. The super- 
nova in Anon 0813 + 20 was observed on 



April 25, 1968, and found to be of type I. 
The redshift of the parent galaxy is 
12,800 km/sec. The supernova in Anon 
1256 + 27 was observed April 25 and is 
also of type I. The redshift of the parent 
galaxy is 12,050 km/sec. The supernova 
in NGC 6946 was observed on April 25 
and found to be of type II. On May 16, 
1968, Sargent observed the supernova in 
NGC 4891; it appears to be of type I. 
The supernova in the semicompact gal- 
axy Anon 1404 + 53 was observed the 
same night. Its spectrum was not readily 
identifiable. 



PULSATING RADIO SOURCES 



Following the announcement of the 
remarkable discovery of pulsating radio 
sources by Hewish and his collaborators 
in February 1968, a program was started 
toward their optical identification. 

Photographs of the field of CP 1919 
by Arp include sky-limited Ha inter- 
ference-filter plates with the 48-inch 
schmidt, and limiting exposures in the 
blue, visual, and red regions with the 
200-inch telescope. The best 200-inch 
photograph, made in good seeing, goes to 
about m pg = 23 and is being used by 
various investigators in the attempted 
optical identification of the source. 

Efforts have so far centered on CP 
1919, in part because it was the first 
source with an accurately determined 
radio position, and in part because there 
is a relatively conspicuous optical object 
at that position that has been suggested 
by Ryle and Bailey as a possible candi- 
date. The object shows H and K in 
absorption at a radial velocity of —40 
km/sec, according to Sargent and 
Schmidt. Photoelectric photometry by 
Sandage and Kristian give 7 = 17.5, B — 
7= +1.4, and U-B= +0.8. 

A search for optical variations in 
this and other objects has been started 
by Kristian, Sandage, Schmidt, Grant 
Snellen of the Caltech Computing Cen- 



ter, and Westphal. Pulses from the 200- 
inch prime-focus photometer are re- 
corded directly on magnetic tape, along 
with a 12 KHz timing signal, so that 
the arrival time of each photon can be 
determined. These data are being 
searched for periodic fluctuations in 
brightness, with preliminary results that 
are negative. One 2-hour run on CP 
1919, for example, was played back 
through a 1024 channel signal-averaging 
computer that was cycled at twelve times 
the radio period P, so that periodicities 
of P, IP, ZP, 4P, 6P, or 12P with suffi- 
cient amplitude should have been de- 
tected. The data from the averager were 
then folded into the submultiples of 12P 
and digitally smoothed with filters of 
varying widths in order to improve the 
precision. No periodicity was detected, 
with results ranging from pulse widths 
of 50 msec at 12P with counting statistics 
of 1% to quasi-sinusoidal oscillations at 
2P with counting statistics of 0.2%. 
Measurements and more detailed analy- 
sis of the data are continuing. 

For the other two sources with fairly 
accurate radio positions, photographs of 
the fields have been obtained by Kristian 
with the 48-inch schmidt telescope and 
IHa-J plates. For CP 0950, there are 
only two or three objects close to the 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



43 



plate limit near the radio position. Sev- 
eral objects near the position of CP 1133 
as determined by Moffet and Ekers have 
been investigated spectroscopically by 
Schmidt. None of the stellar objects 



seemed peculiar. The object closest to 
the radio position turned out to be a 
compact 16-17 mag galaxy of early- 
type spectrum. The Balmer lines are seen 
in absorption at a redshift of 11%. 



X-RAY SOURCES 



Scorpius X-l 

Spectroscopic and photometric data 
obtained in 1966-1967 have been ana- 
lyzed in detail and discussed by Kristian, 
Sandage, and Westphal. Eleven spectro- 
grams were remeasured with the exten- 
sive use of digital techniques. The plates, 
some of which were single-trailed, were 
scanned in successive strips with a micro- 
photometer whose photoelectric output 
was amplified and sent by underground 
cable to the Caltech computing center. 
There the signal was converted to 
digital form and stored on magnetic tape 
by a high-speed analogue-to-digital con- 
verter. All subsequent analysis of the 
plate material was done by computer, 
using the digitized data. The digital 
techniques proved quite powerful, both 
for radial velocity measurements and 
faint line detection. 

Radial velocity changes of the order 
of 100 km/sec over several hours were 
found, with the emission lines He II 
A4686 and the Balmer series moving in 
opposite directions. In addition to these 
velocity changes within a single night, 
there also appear to be systematic night- 
to-night changes of the same magnitude 
over a period of a week. The available 
data are not yet sufficient to determine 
whether the velocity changes are due to 
orbital motions or to other effects, such 
as gas streaming. A minimum statement 
is that some systematic motion is pres- 
ent, and in this respect Sco X-l is 
similar to the binary source Cygnus X-2 
(see below) . 

In addition to the prominent lines 
previously mentioned here and else- 
where, many faint emission features 



were noted, including numerous lines of 
He I, He II at AA3858 and 4541, several 
weak lines tentatively identified as due 
to Fe II, a number of weak, broad fea- 
tures between AA4350 and 4600, and sev- 
eral unidentified lines. No convincing 
identifications of forbidden lines were 
made, but the identification problem 
awaits further study. 

A firm lower limit for the distance to 
Sco X-l, based on Wallerstein's in- 
vestigation of the line strengths of inter- 
stellar Ca II K, is 300 pc. An upper limit, 
assuming that Sco X-l is like the bright- 
est old novae, is 1000 pc. Distance 
estimates based on the number of hydro- 
gen atoms in the line of sight and on the 
X-ray optical depth, while more un- 
certain, are consistent with these limits. 

The optical flux of Sco X-l has been 
monitored on five separate nights for 
intervals of 2 to 4 hours, and the object 
was found to be highly and erratically 
variable. Rapid optical flickering with 
amplitudes of a few hundredths of a 
magnitude on a time scale of a few 
minutes is always present, superposed on 
slower continuous variations with ampli- 
tudes of the order of 0.15 mag and occa- 
sional bursts of 0.2 mag lasting for 
intervals of the order of 10 min and hav- 
ing very rapid rise and decay times. 

Two indirect, order-of-magnitude esti- 
mates suggest that the flickering may 
be in some way related to the X-ray 
output. Assuming that the X rays are 
produced in a homogeneous bremsstrah- 
lung source, the cooling time is of the 
order of minutes, which is the char- 
acteristic time of the flicker. If the flicker 
were produced by a low-Q, low-order 
harmonic pulsation of the star, then, 



44 



CARNEGIE INSTITUTION 



for plausible parameters, the acoustic 
energy delivered to the stellar atmo- 
sphere would be of the same order as 
the observed X-ray output, if a mecha- 
nism were available to convert one to the 
other. A test of this idea would be to 
look for a correlation between the ampli- 
tude of the flicker and the X-ray output. 

To improve our understanding of the 
physics of X-ray sources, it is now de- 
sirable to observe such a source simul- 
taneously for several hours in the X-ray 
and optical regions of the spectrum and 
to test for possible correlations of its 
variations in the two regions. An early 
attempt at such an experiment is 
planned. 

Neugebauer, Oke, Becklin, and Gar- 
mire have completed a study of Sco X-l 
by scanning into the infrared. Photoelec- 
tric spectrum scans from AA3300 to 10,800 
and infrared photometric measures at 
1.65 and 2.2 p have been obtained on 
many nights. The observed spectral 
energy distribution /„ is nearly flat from 
0.33 to 1.0 fi and then turns down in the 
infrared, approaching an f v ccv 2 law. 
The observations rule out free-free 
radiation from a transparent gas; they 
also rule out synchrotron radiation un- 
less some self-absorption mechanism 
operates in the visible. The X rays can 
be explained by free-free radiation from 
a gas with a temperature of 5xl0 7 ° K. 
This same mass of gas can also explain 
the observed fluxes in the visible and 
infrared if a reasonable amount of inter- 
stellar absorption is assumed, since it 
can be shown that the optical depth in 
the visible and infrared is large. 

Cygnus X-2 

The object tentatively identified as 
Cyg X-2 has been observed photo- 
metrically and spectroscopically by 
Kristian, Sandage, and Westphal and 
the identification confirmed. 

The photometric data show variations 
similar to those of the first identified 
source, Sco X-l. Broad-band UBV data 



on seven nights showed variations up to 
0.3 mag between successive nights. A 
5-hour run of continuous monitoring, 
using a special two-channel ratio pho- 
tometer developed by Thomas McCord 
of the Division of Geological Sciences at 
Caltech, showed short-peaked oscilla- 
tions of about 0.04 mag and time scales 
of a few minutes, superposed on a longer 
fluctuation period of about 1 hour and 
amplitude 0.10 mag; underlying these 
was a gradual decline of 0.5 mag over 
the 5-hour interval. 

A 200-inch prime-focus spectrogram 
was analyzed by means of the digital 
techniques described above for the Sco 
X-l spectra. Among the remarkable fea- 
tures of this spectrogram are (1) very 
broad hydrogen lines in absorption, with 
a notable absence of damping wings such 
as are found in white-dwarf spectra; 
(2) narrow lines of He II A4686 in 
emission and circumstellar Ca II in ab- 
sorption; (3) a strong asymmetry just 
shortward of Hy and of HS; (4) an ab- 
sorption feature at A4460 that may be 
He II A4471; and (5) a faint emission 
line due to He I A4713. According to 
Greenstein, the presence of both A4713 
and He II A4686 in the same spectrum 
is very rare. The features (1), (3), and 
(4) may indicate the presence in the 
object of gas streams with velocities in 
excess of 700 km/sec. 

The spectrogram is consistent with a 
suggestion made earlier by C. R. Lynds 
and E. M. Burbidge that Cyg X-2 is a 
spectroscopic binary whose components 
are associated with the He II line and 
the absorption lines, respectively. The 
available data are minimal, but the as- 
sumption of circular orbits gives a lower 
limit of 300 km/sec for the relative 
orbital velocity, approximately equal 
masses, a period less than 0.4 Wl days, 
a separation less than 0.01 WIq a.u., 
and a systematic velocity of —250 km/ 
sec. This velocity would reflect almost 
entirely the sun's motion in the Galaxy, 
which would suggest that Cyg X-2 is 
a Population II obj ect. 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



45 



For three independent reasons, it ap- 
pears that the optical flux from the ob- 
ject cannot primarily be nonthermal 
radiation connected with the production 
of the X rays: (1) there are quasi- 
periodic fluctuations as short as a few 
minutes; (2) the central depth is at least 



A e = 0A for the Ca II K shell line, and 
the hydrogen lines have appreciable 
depths; and (3) there is disagreement 
between the observed brightness and 
color of the object and an extrapolated 
bremsstrahlung spectrum fitted to the 
X-ray data. 



QUASI-STELLAR SOURCES 



Spectroscopic Observations 

Spectroscopic observations of quasi- 
stellar radio sources are being continued 
by Schmidt. Most of the sources observed 
have been identified by Edward T. Olsen 
of the University of Michigan from ac- 
curate positions obtained for 4C radio 
sources. More than half of the identifica- 
tions are confirmed by the spectroscopic 
nature of the objects. Relatively few red- 
shifts have been determined so far. 

The most interesting quasi-stellar 
source is 4C 25.5, which was identified 
by Olsen with a blue stellar object of 
visual magnitude around 17.5. Its spec- 
trum shows broad emission lines at 
wavelengths of 4082 and 5200 A. These 
lines are identified as Lyman-a and C IV 
A1549 emission at a redshift of 2.358, 
the largest redshift recorded to date. 
Absorption lines due to Ly-a and C IV 
are seen on the red side of the emission 
lines. The absorption-line redshift is 
2.3683. Apparently the absorbing ma- 
terial is flowing toward the emitting 
parts of the quasi-stellar source with a 
speed of around 900 km/sec. Several 
other absorption lines have not yet been 
identified. 

Schmidt is undertaking a systematic 
spectroscopic study of blue stellar objects 
identified by Sandage and Luyten in 
fields at 8 h 28 m , +18° and 15 h 10 m , +24°. 
The objective is to obtain a sample of 
quasi-stellar sources complete to a limit- 
ing magnitude near 18. This will allow 
a derivation of the space distribution 
similar to that reported in Year Book 66 
(p. 292) for quasi-stellar radio sources 
in the revised 3C catalog. 



Radio-Quiet Quasars and White Dwarfs 

Sandage has completed photoelectric 
photometry of 300 blue objects listed in 
the catalogs of the general survey dis- 
cussed in Year Book 66 (p. 292). Most 
of the objects lie in the magnitude range 
15 < 7 < 19.2. Analysis of the nature of 
these objects was begun from a study of 
the JJ—B, B — V diagram. Two major 
regions of the distribution are along the 
blackbody line and within the F and G 
subdwarf domain. There are virtually no 
stars along the luminosity class III-VI 
line of the TJ—B, B — V diagram, con- 
trary to the situation for blue stars 
brighter than V ~ 15. This result con- 
firms the earlier conclusion that a major 
change in the nature of the blue objects 
occurs near 7 = 15. Although the change 
is not abrupt, it is clearly present. 

Spectra by Schmidt show that all ob- 
jects above the blackbody line with 
0.6>£ — F>0.0 are radio-quiet quasars. 
A number of objects below the black- 
body line are also quasars, but the 
contamination of the sample by nearby 
white dwarfs is also appreciable. 

Sandage and Luyten conclude that 
brighter than V ~ 15 the blue population 
consists of stars that follow the lumi- 
nosity class III-VI line of the TJ—B, 
B — V diagram. These are hot OB sub- 
dwarfs, runaway main-sequence B stars, 
and hot horizontal-branch stars. Fainter 
than V ~ 15, the number of these objects 
decreases drastically due to the steep 
density gradient perpendicular to the 
galactic plane. However, at about this 
magnitude the volume of space sampled 
for white dwarfs with V ~ + 10 and 



46 



CARNEGIE INSTITUTION 



fainter (D<100 pc) becomes large 
enough so that appreciable numbers of 
white dwarfs begin to appear. Radio- 
quiet quasars also begin to occur. 

Work on the quasar component be- 
came sufficiently precise during the re- 
port year to permit a first tentative esti- 
mate of the number of radio-quiet 
quasars per square degree at various 
apparent-magnitude levels. Spectra by 
Braccesi and Lynds of the interloper 
field of Sandage and Veron, together with 
photoelectric photometry by Sandage in 
a 200-inch field of 0.04 square degree 
centered on SA 57, combined with work 
reported last year, suggest that there 
are 0.4 QSO per square degree to B = 18.1, 
5 per square degree to J3 = 18.5, and 100 
per square degree to 5 = 21.5. These are 
consistent with a growth relation 

log N(B) =0.75 5-14.0 

for the number of radio-quiet quasars 
per square degree brighter than magni- 
tude B. The data are not yet of sufficient 
accuracy to determine whether the differ- 
ence of the coefficient in B from the 
constant-density value of 0.6 is signifi- 
cant, but Sandage and Luyten note that 
their result is consistent with Schmidt's 
earlier result, based on a smaller but 
more homogeneous sample of radio 
quasars, that d log N(B)/dB>0.Q. 

Although this determination of the 
growth function of radio-quiet quasars 
is tentative, Sandage and Luyten are 
convinced that the number of such ob- 
jects per square degree at B ~ 22 is 
very large. The suggested function gives 
10 7 quasars over the entire sky to B = 22. 
Sandage believes that such a large num- 
ber of objects that have either a constant 
space density [d log N(B)/dB = 0.6] 
or a positive density gradient outward 
virtually eliminates the local hypothe- 
sis for quasar distances. 

Optical Variations of Quasars and 
N-Type Galaxies 

One of the principal distinguishing fea- 
tures of quasi-stellar sources is their 



optical variation. Rapid fluctuations of 
the nonthermal radiation (continuous 
emission not attributable to stars) in 
quasars on a time scale of days is one 
of the phenomena cited by some astrono- 
mers in their contention that quasars 
are at local distances, contrary to the 
evidence presented by the large red- 
shifts. 

This particular argument against cos- 
mological distances for quasars would 
fall if similar optical variations of non- 
thermal radiation could be discovered in 
objects whose cosmological distances 
were not in dispute. 

In 1963 Burbidge, Burbidge, and Sand- 
age pointed out the similarity of the 
nuclei of Seyfert galaxies and quasars. 
More recently, the nonthermal radiation 
in highly nucleated N-type galaxies, 
isolated by Morgan at the Yerkes Ob- 
servatory, has been recognized to be of 
the same type. Because Seyfert galaxies 
and N-type systems are at distances 
given by their redshifts, the discovery 
of optical variations in the nuclei of 
either of these galaxy types would be 
crucial. 

From repeated photometric observa- 
tions, Sandage, and independently Oke, 
found that the optical flux (3300 A<A< 
10,000 A) of the N-type galaxy 3C 371 
changed by AV ~ 0.7 mag (nearly a 
factor of 2 in intensity) between June 
1966 and September 1967. Oke found 
that the time scale for variations of 10- 
15% was only a few days. 

Sandage showed that the N-type gal- 
axies are at their Hubble distances by 
noting that all members of the class (11 
objects) studied to October 1967 follow 
the redshift-apparent magnitude rela- 
tion (the Hubble diagram) defined by 
radio galaxies not classed as N; and that 
the visual magnitudes, which form the 
abscissa of the Hubble diagram, are not 
contaminated with the nonthermal com- 
ponent by more than 0.3 mag for N gal- 
axies with redshifts less than AA/Ao = 0.3. 
The contamination is less than the 
spread of radio E galaxies about the 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



47 



mean line in the Hubble diagram. Be- 
cause (1) the radio N galaxies contain 
a radio E system underlying the non- 
thermal radiation, (2) the radio E gal- 
axies are constant in absolute luminosity 
to within o- ~ ±0.5 mag, and therefore 
define a meaningful Hubble diagram, 
and (3) there is no dispute over the 
distances of E galaxies, these data show 
that N galaxies are at their Hubble 
distances to within the statistics of the 
E galaxies with which they are com- 
pared. 

Later observations by Sandage in 
August to September 1967 showed that 
the N galaxies 3C 109 and 3C 390.3 also 
vary in brightness by at least 0.3 mag. 
The compact galaxy I Zw 1727 + 50 was 
found to be an optical variable by Sand- 
age, and independently by Oke, Sargent, 
Neugebauer, and Becklin. These dis- 
coveries appear to give clear evidence 
that large variations of radiation on a 
time scale of days cannot, in them- 
selves, contradict the cosmological dis- 
tance scale of quasars because the N-type 
galaxies where such variations have now 
been observed are at their Hubble 
distances. 

Spectral Energy Distribution 

E. J. Wampler at the Lick Observatory 
and Oke have continued to scan 3C 446 
to study the changes in the spectral 
energy distribution with time and to 
measure absolute line intensities of C IV 
A1550 and C III A1909. No line-intensity 
changes have been observed since the 
original outburst in 1966. This suggests 
that the size of the region giving rise 
to the emission lines is at least one light 
year. During the summer of 1967, 3C 
446 dropped back to near minimum 
light. The flux in the red, however, re- 
mained abnormally high; this may indi- 
cate that two distinct energy distribu- 
tions, one of which is variable, are 
present. 

Oke, Neugebauer, and Becklin have 
been carrying out a program to deter- 
mine overall spectral energy distribu- 



tions of many quasi-stellar sources. Oke 
has obtained detailed scan data from 
AA3300 to 6000, and a few selected wave- 
lengths between AA6000 and 10,000. 
Neugebauer and Becklin, with the 200- 
inch telescope, have made measurements 
at 1.65 and 2.2 ^. Data now are avail- 
able for more than 20 objects. In the 
great majority of cases, the fluxes f v 
change smoothly with wavelength, usu- 
ally increasing toward the infrared. In a 
few, as in 3C 273, the spectra are flat 
and abruptly brighten in the infrared. 
Occasional objects show very abnormal 
energy distributions. 

Neugebauer and Becklin have at- 
tempted measurements of 42 quasars; of 
these, 29 have been observed to have a 
K magnitude larger than 15. In the 
region of the sky between R.A. = 13 b 
_0 h -5 h with Dec -30°, 23 of the 26 
known quasars brighter than T=17.0 
have been measured. For this sample 
the average V — K color index equals 3.1 
with a scatter of 0.7. If it is assumed 
the source of radiation is nonthermal, 
then this color index corresponds to a 
spectral index a of 0.6 ±0.3. (The flux is 
proportional to 2~ a .) In general, the 2.2-/x 
flux density lies on or near the line con- 
necting the radio and visual flux 
densities. 

Of the 39 sources measured, 19 were 
measured twice or more at 2.2 p, with 
the result that only two sources showed 
variations larger than those expected 
from normal statistics. These two 
sources, 3C 279 and 3C 345, are known 
visual variables. The brightness of 3C 
345 decreased by a factor of 3 between 
summer of 1967 and the spring of 1968. 
The results of the infrared measurements 
of quasars are being prepared for pub- 
lication in collaboration with Oke. 

Polarization 

Visvanathan has started a program at 
the 100-inch to measure the polarization 
in sources that show nonthermal con- 
tinuum. Fourteen QSS were observed for 
polarization. Seven of the 14 show mea- 



48 



CARNEGIE INSTITUTION 



surable intrinsic polarization in the 
range from 0.014 to 0.087 mag. The 
source 3C 454.3, which is optically varia- 
ble, shows polarization of 0.087 mag, and 
the polarization elements seem to have 
remained the same on the nights ob- 
served. Other QSS that have been ob- 
served on different nights do not show 
variation in P and 9. 

Parkes Source 0237-23 

An elaborate investigation of the ab- 
sorption-line spectrum of this object 
has been completed spectroscopically by 
Greenstein and Sargent, working at 90 
and 190 A/mm. The best spectrum re- 
quired observations on two nights. A 
large number of lines are suspected to be 
present, and a very selected list of 49 
absorption lines in the wavelength range 
AA3300-4800 has been obtained. Equiva- 
lent width, half widths, and central 
depths are given for 38 of these at A 
greater than 3650 A. The first observed 
fact is that the lines are barely resolved, 
if they are resolved, even at this disper- 
sion. In other words, the lines are very 
much sharper than in most other quasi- 
stellar sources. In addition, a large num- 
ber of lines could not be identified with 
the first redshift proposed by Arp, 
Bolton, and Kinman. Work by Green- 
stein and Schmidt a year ago had al- 
ready suggested that two separate red- 
shifts were present at 2 = 2.202 and 1.956. 
However, an attempt to identify the 
full absorption-line list led to great diffi- 
culties, so that the only satisfactory pro- 
cedure, eventually, was the use of a sys- 
tematic method invented by Bahcall, in- 
volving an elaborate computing machine 
program. Bahcall's program permits the 
isolation of candidate redshifts on the 
basis of a comparison of a list of lines 
expected to be present with the observed 
list of lines. Bahcall decided to use only 
lines of singly to quadruply ionized 
common elements, arising from their 
ground state. (The latter assumption is 
required if one expects the lines to be 



formed at relatively low electron den- 
sity.) Bahcall's procedure turned out a 
large number of redshifts; next it was 
necessary to isolate those that seemed to 
be physically reasonable. Therefore, cri- 
teria such as the presence of Lyman-a 
and other strong lines, the identification 
of the strongest lines for each element, 
the identification of those expected to 
be strongest with normal composition, 
etc., had to be made. The redshifts thus 
revealed were 2 = 2.2015, 1.6706, 1.6560, 
1.5132, and 1.3642. In the paper by Bah- 
call, Greenstein, and Sargent, there is no 
strong evidence for an absorption red- 
shift at 1.955 that had been found by 
earlier workers. However, reexamination 
of this possibility subsequent to sub- 
mission for publication suggests that, in 
fact, a shift of 1.9550 may be possible. 
There is some slight evidence for the 
presence of a line from a low excited 
state in one of the shells, but in all the 
others no excited states are present. It is 
impossible to conclude whether the lines 
arise in intergalactic space, in galaxies 
in the line of sight, or in clouds which 
have been expelled from the object and 
which are moving at speeds up to 0.3 
the velocity of light relative to the 
source of the emission lines. 

Source 3C 287 

A report by T. A. Matthews at the 
University of Maryland of time-vary- 
ing nebulosity associated with the quasar 
3C 287 has been investigated and not 
confirmed. Matthews has compared a 
100-inch plate taken in 1965 with a 48- 
inch plate taken in 1966. He reported 
seeing a "bridge" on the 100-inch plate 
between the quasar and a fainter, redder 
star some 3 arc-sec away. On the 48-inch 
plate he reported that the bridge had 
expanded into a nebulosity about 4 sec 
wide, and, assuming expansion at less 
than the velocity of light, deduced an 
upper limit of % -million light years 
for the distance of the quasar. 

Following early reports of this work, 



MOUNT WILSON AND PALOMAE OBSERVATORIES 



49 



Kristian and Peach examined twelve 48- 
inch and four 200-inch plates of 3C 287, 
taken in 1950, 1964, 1967, and 1968, for 
evidence of the nebulosity and failed to 
detect it. In particular, all the 200-inch 
plates, with better resolution and fainter 
limiting magnitude than those on which 
the nebulosity was reported, show the 
images of the quasar and the nearby star 
to be distinctly stellar, with no as- 
sociated nebulosity between or around 
them. The first of these was taken by 
Sandage in 1964, a year before the re- 
ported changes, and the second by Arp in 
1967. 

The other two 200-inch plates, taken 
in 1968 by Kristian and Peach, were 
compared with two 48-inch plates taken 
three weeks earlier. The 200-inch plates 
show no nebulosity. One of the 48-inch 
plates was taken in exceptionally good 
seeing. Because of the smaller scale size 
and the sphero-chromatic aberration of 
the schmidt telescope, the images are 
blended, with a perceptible "waist" be- 
tween them. The second schmidt plate 
was taken on the following night in 
poorer seeing and shows the images even 
more smeared, the waist having disap- 



peared. The image on this plate looks 
similar to that on prints of Matthews' 
plate. From this comparison and an 
examination of the appearance of 3C 
287 on earlier 48-inch plates, some of 
them taken on the same nights and 
others only a month apart, it was con- 
cluded that atmospheric, optical, and 
photographic effects could be sufficient 
to account for the appearance of the 
1966 plate without having to attribute 
any part of the image to nebulosity. 

Several attempts to obtain a 100-inch 
plate of the object were unsuccessful 
because of poor weather. An examination 
of randomly selected plates in the files, 
however, showed that spurious "bridges" 
(as photographic artifacts) between 
pairs similar to 3C 287 and its com- 
panion are very common on 100-inch 
plates. 

It appears also that the faint star may 
have a proper motion. This is based on a 
comparison of the Sky Survey red plate 
with the later 48-inch plates. The quality 
of the image on the Sky Survey plate, 
however, is not good, and this result can 
only be verified in time by a later epoch 
plate. 



THEORETICAL STUDIES 



Stability of Rotating Systems 

The gravitational stability of thin, 
rotating disks of stars with respect to 
small-scale, axisymmetric disturbances 
has been reexamined by Julian, taking 
into account possible overstabilities. 
These overstabilities have been shown 
to be absent whenever the density of 
stars in phase space is a decreasing 
function of the size of the Lindblad epi- 
cycle. The existence in such cases of a 
maximum wavelength of disturbance 
and a maximum r.m.s. random velocity 
beyond which all simple instabilities are 
avoided has been established. On the 
other hand, overstability has been shown 
to occur for no rotation if the distribu- 



tion of stellar velocities in any given 
direction has a sufficiently high relative 
maximum at a nonzero velocity. More- 
over, overstabilities have also been shown 
to arise even in the presence of rotation 
for that velocity distribution wherein 
all the Lindblad epicycles are equal in 
size. This work suggests that stable 
velocity distributions for the stars in disk 
galaxies must be similar to the Schwarz- 
child one. 

Electron cyclotron masers are cur- 
rently being studied by Peter Goldreich 
of the Geological Sciences Division and 
Julian with possible applications to radio 
sources. The effects on the maser phe- 
nomenon of various electron distribution 
functions are being pursued by numerical 



50 



CARNEGIE INSTITUTION 



analysis of the dispersion equation for 
electromagnetic waves in a plasma. 

The structure of the magnetic field 
surrounding a rapidly rotating neutron 
star with a relativistic, co-rotating mag- 
netosphere is being studied by Goldreich 
and Julian with possible application to 
pulsating radio sources. 

Dissociation Equilibrium oj H~ 

The assumption that the source func- 
tion for continuous thermal radiation 
due to the negative hydrogen (H~) ion 
in solar-type stellar atmospheres is 
equivalent to the Planck function has 
previously been justified on the grounds 
that the dissociation equilibrium of the 
H~ ion is governed by the associative 
detachment reaction H~+H— >H 2 + e. 
Lambert and Dr. B. E. J. Pagel of the 
Royal Greenwich Observatory, Herst- 
monceaux, England, have rediscussed 
this assumption on the basis of recent 
experimental and theoretical estimates 
for the rates of the important processes 
controlling the H~ equilibrium. They 
conclude that the assumption is justified 
for the photospheres of main-sequence 
stars and of normal giants, at least for 
those giants in the range of spectral 
types G5 III to K5 III. This conclusion 
is confirmed by an analysis of the ob- 
served infrared color indices. 

Nonthermal Radiation 

Dr. Leonard Searle of the Mount 
Stromlo Observatory and Sargent found 
that the equivalent width of the H/3 
emission line is about 50 A in all extra- 
galactic objects having broad emission 
lines for which this quantity is known. 
They summarized the existing data for 
15 objects, including 6 Seyfert galaxies, 

1 N-type galaxy, 3 Zwicky compacts, 

2 blue stellar objects, and 2 quasi-stellar 
radio sources. The expected equivalent 
width of H/? due to recombinations in a 
gas at 20,000°K, the temperature char- 
acteristically found for these objects, is 



500 A. It has commonly been supposed 
that objects of the kind listed have non- 
thermal optical continua. In order to ex- 
plain why the optical nonthermal radia- 
tion should always weaken the H/3 emis- 
sion line by about the same amount, 
Searle and Sargent suggest that the gas 
in these objects is ionized by the non- 
thermal radiation. Calculations show 
that this suggestion is reasonable if the 
nonthermal radiation has the same 
spectral index in all the objects and if 
this value of the spectral index persists 
below the Lyman limit. 

Variations of Extragalactic Sources 

A study of the radio flux variations of 
extragalactic radio sources was carried 
out by Rees and Simon. It was shown 
that although the observations suggest 
that the variable flux arises in an ex- 
panding component emitting synchrotron 
radiation, such a model is untenable for 
both 3C 120 (a compact galaxy) and 
3C 273 (a QSS) unless the expansion 
velocities are relativistic. A relativisti- 
cally expanding model would be com- 
patible with the present data in both 
cases, and very long baseline inter- 
ferometers should soon provide a further 
test. It was noted also that the occurrence 
of relativistic expansion would reduce 
the estimated severity of inverse Comp- 
ton losses in the optically variable QSSs. 

The character of the optical variations 
of 3C 273 was studied by Simon and 
T. Manwell of Cornell University. By 
analysis of the spectral density of the 
observed luminosity functions, it was 
shown that there is considerable difficulty 
with the hypothesis that the variations 
are periodic, and it was argued that they 
may be the result of a random process. 
Specific attention was directed to the 
problems of misleading resolution in 
spectral-density estimates of records of 
finite length, and the appropriateness of 
the various statistical tests that had 
been used in past studies of the fluctua- 
tions. 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



51 



GUEST INVESTIGATORS 



Dr. R. d'E. Atkinson of the University 
of Indiana spent some weeks at the Ob- 
servatories scanning and remeasuring a 
number of Humason's early spectro- 
grams of galaxies. He concluded that, 
when small and large redshifts are com- 
pared, there is no significant evidence 
for a change with time in the atomic 
fine-structure constant. 

Dr. Lawrence A. Aller of the Uni- 
versity of California at Los Angeles and 
Dr. Stanley S. Czyzak of Ohio State 
University, with the 100-inch and 60-inch 
telescopes at Mount Wilson, continued 
their observations of planetary nebulae. 
Spectrograms of NGC 6572, 7009, 112165, 
114997, and 115217 were obtained, and 
also plates of suitable comparison stars, 
i.e., 6 Crateris and £ 2 Ceti. 

During the first six months of 1968, 
they were able to secure necessary photo- 
electric data on the following nebulae 
in the blue and visible regions, at the 60- 
inch telescope: blue: NGC 2003, 6886, 
111717, and 115217; visible: NGC 2440, 
3242, 4361, 6210, 6309, and 113568. The 
comparison star for these objects was a 
Leonis. 

Investigations on the following nebu- 
lae, based on Mount Wilson observations, 
have been accepted for publication dur- 
ing the past year: NGC 604, 2440, 6543, 
112149, and 112165. Also, a paper on 
the "Chemical Composition of Planetary 
Nebulae" will appear in the Intern. 
Astron. Union Symp. No. 84, Planetary 
Nebulae, p. 209. Here Aller and Czyak 
have employed certain of their observa- 
tional data with their latest theoretical 
results on forbidden transitions and col- 
lision strengths in the study of the nebu- 
lar chemical composition. The observa- 
tional data in IC 1747, 3568, 5217, and 
NGC 2003, 4361, 6309, 7026 have been 
reduced and work has been completed 
for publication. 

With the recent improvements and de- 
velopments in the calculation of collision 
strengths of forbidden transitions, and in 



particular that of the fine-structure col- 
lision strengths, the study of the physical 
processes in gaseous nebulae demands 
a more careful analysis of the photo- 
electric and photographic photometry. 
A greater stress has to be placed on the 
observational data, i.e., a more detailed 
study of the objects has to be made. 

Dr. Roberto Barbon of the Osservatorio 
Astrophysico, Asiago, used the 48-inch 
schmidt telescope for part of a program 
on the luminosity function of super- 
novae. Light curves of the following four 
supernovae appearing in the Coma 
cluster of galaxies were obtained: SN 
1962a, 1962i, 1963c, and 1963m. 

From the material of the 48-inch 
schmidt supernova search assembled dur- 
ing the last ten years, Barbon investi- 
gated the frequency of occurrence of 
supernovae in clusters of galaxies. The 
value derived — one supernova per gal- 
axy per 316 years — is in satisfactory 
agreement with the previous estimates 
made by Zwicky in 1942 from the ma- 
terial obtained with the 18-inch schmidt 
telescope. 

From May 1966 through November 
1967, the 18-inch schmidt has been used 
by Barbon in surveying spectroscopically 
the regions surrounding the galactic 
poles in a search for bright compact 
galaxies and quasi-stellar objects. The 
limiting magnitude of the survey is 
about 13.5 m pg , and the dispersion is 370 
A/mm at H T . The survey covers approxi- 
mately 2100 square degrees around the 
north galactic pole between Dec +18° 
and +48° and RA 22 h to 16 h , and about 
2500 square degrees in the southern gal- 
actic cap between Dec —12° and +12° 
and RA 22 h to 4 h . The reduction of the 
material is still in progress. While many 
known red stars, Me variables, and 
emission-line galaxies have been identi- 
fied, this spectroscopic survey disclosed 
no bright compact galaxy or quasi- 
stellar obj ect. 

With the 48-inch telescope, Barbon 



52 



CARNEGIE INSTITUTION 



searched 13 fields in the southern galactic 
cap, not covered by the Haro-Luyten 
survey, for faint blue stars by the three- 
color method. Thus far 937 new blue 
stars have been found in the six fields 
already inspected. A catalog giving posi- 
tions, magnitudes, and colors is being 
compiled. 

In collaboration with Drs. A. Braccesi 
and R. Fanti of the University of Bo- 
logna, who provided the radio data, Bar- 
bon found that seven radio sources from 
the Fourth Cambridge Catalogue could 
be optically identified with blue stars. 

Coude spectrograms of y Geminorum 
were measured by Dr. Wallace R. 
Beardsley of the Allegheny Observatory 
in order to obtain accurate radial veloci- 
ties during the cycle 1942-1955. These 
velocities are being analyzed to deter- 
mine whether an apparent shift in the 
velocity of the system, observed for 
other cycles, exists for this cycle also. 
The accurate velocities will also permit 
a definitive orbit to be calculated. At the 
suggestion of Greenstein, several mea- 
sures of coude spectra of Groombridge 
1830 also were made for the period 1948 
to date. These measures are being con- 
tinued in an effort to define a possible ra- 
dial velocity orbit to the visual com- 
panion discovered recently by Worley. 

Dr. Alessandro Braccesi of the Insti- 
tute of Physics, Bologna, obtained multi- 
color plates with the 48-inch schmidt 
telescope for identification of radio-quiet 
quasars by means of the infrared excess 
technique. Initial tests were promising, 
as shown by spectrographic confirmation 
in a number of cases. An iris photometer 
for measurement of the plates has been 
built and put into operation in Bologna. 

Dr. Robert G. Chambers of the Brac- 
kett Observatory, Pomona College, ob- 
tained spectrograms in the near infra- 
red (A~8500 A) of 5 Be stars and 13 
MK standard stars, using the X spectro- 
graph on the 60-inch reflector. The pur- 
pose is to exploit the higher shell trans- 
parency expected at longer wavelengths 
for classification of later Be stars. 



During the first week of August 1967, 
n. Dickel and H. Wendker of the Uni- 
versity of Illinois Observatory completed 
the first half of an Ha survey of the 
emission nebulosities in the Cygnus X 
complex (near the star y Cygni). The 
observational material for the Ha sur- 
vey consists of 28 direct-photographic 
plates taken with an 80-A wide Ha inter- 
ference filter on the 48-inch schmidt 
telescope. Most of the 103aE plates were 
"preflashed" for 2 sec before being ex- 
posed for 75 min at the telescope. Each 
plate has a set of calibration spots made 
with the spot sensitometer. Measure- 
ments of the surface brightness in Ha 
for the four brightest H II regions (IC 
1318a,b,c and HS 191) were made at the 
Lowell Observatory and will be used to 
calibrate the plates. Each plate has a 
usable field of 1.5° X 1.5°. The overlap- 
ping network of 28 plates is centered on 
a(1950.0)~20 h 18 m , 8(1950.0) ~+41.5°, 
or ^ n ~79°, b n ~ +30°. The plates cover 
the region from approximately ^ n = 74° — 
84°and6 n = o -5.5°. 

The Cygnus X region has also been 
observed at centimeter wavelengths by 
Wendker. This complex contains many 
radio sources, of which at least 29 have 
well-determined thermal spectra. All but 
four of these coincide with optically visi- 
ble nebulae on prints of the Palomar Sky 
Atlas. The distribution of Ha emission 
across the nebulae will be obtained from 
the microphotometry of the plates. These 
distributions will be compared with the 
hydrogen continuum emission at the 
radio wavelengths in order to determine 
the amount and distribution of interstel- 
lar absorption in the region. A compari- 
son of the values of the Ha absorption 
with star-reddening data will give an 
independent determination of the dis- 
tances and dimensions of the nebulae. 
Since these H II regions presumably lie 
in the local spiral arm which one views 
tangentially, their distribution in depth 
along the arm may be derived. 

As a guest investigator with the 200- 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



53 



inch telescope and coude spectrograph, 
John Danziger of the Harvard College 
Observatory has completed lithium ob- 
servations in 17 main-sequence stars be- 
tween spectral types F5 and KO in the 
Coma cluster. There is a marked de- 
crease of lithium with decrease in effec- 
tive temperature, and at the later spec- 
tral types the lithium abundance falls 
between those observed in the Pleiades 
and Hyades clusters. This is consistent 
with an inverse main-sequence age versus 
lithium abundance correlation and rein- 
forces the previous observational results 
that lithium is being depleted in the 
main-sequence phase of stars slightly less 
massive than the sun. Lithium observa- 
tions were made also in a number of 
main-sequence field stars with spectral 
types as late as MO. At least two stars 
later than KO have measurable amounts 
of lithium, and this extends the region 
of spectral type in which lithium has 
been found. An interpretation awaits 
further observations. 

With the Cassegrain scanner on the 
100-inch telescope, Danziger has ac- 
cumulated continuum scans for a num- 
ber of short-period variables, either 
dwarf Cepheids or 8 Scuti stars. An in- 
terpretation of their gravities and tem- 
peratures, and hence their evolutionary 
status, depends on the reliable absolute 
calibration of the primary standard. This 
should be forthcoming in the near future. 

Dr. 0. J. Eggen of the Mount Stromlo 
and Siding Spring Observatories observed 
with the 100-inch and 200-inch telescopes 
in February. His program was concerned 
with photometry of faint proper-motion 
stars from Giclas' lists and with some 
faint stars in Selected Area 57. Poor 
weather severely limited the results. 

During three nights in February and 
six nights in March, Dr. Erik Holm- 
berg of Uppsala University used the 48- 
inch schmidt telescope for photometric 
observations of the Virgo cluster of gal- 
axies, the aim being to determine total 
magnitudes in the UBV system for as 



many cluster members as possible. The 
cluster was covered with two series of 
plates in each of the three wavelength 
regions; to calibrate the plates, extrafocal 
exposures were made of two fields of 
standard stars. It is hoped that a detailed 
study of the final magnitudes and colors 
will, among other things, help to solve 
the problem presented by the color ex- 
cesses known to exist for a number of 
the cluster members. 

Coude spectrograms of the blue region 
of Mira variables, taken with the 100- 
inch and 200-inch telescopes by Dr. Philip 
C. Keenan of the Perkins Observatory 
during the summer of 1967 and preceding 
summers, provided estimates of the 
strength of the absorption bands of alu- 
minum oxide. On the average, the A10 
absorption is stronger in Mira variables 
of type Me than in ordinary giant stars 
at the same temperature. These bands 
fluctuate greatly from cycle to cycle, and 
their occasional reversal to emission sug- 
gests that when they appear abnormally 
weak the absorption is reduced by incipi- 
ent emission. This work was done in col- 
laboration with Deutsch and Garrison, 
and was reported by Garrison in a joint 
paper at the April 1968 meeting of the 
American Astronomical Society. 

With the cooperation of Wilson and 
Deutsch, their coude spectrograms of 
giants in the range of types K2 to M6 
were classified by Keenan, using ratios 
of atomic lines. These temperature types 
were found to be consistent with the 
usual types estimated from the strengths 
of TiO bands, and the small group of M 
giants with space velocities > 80 km/sec 
showed no discrepancies in type. The 
temperature scale of M giants is thus 
nearly independent of space motion. 

Dr. E. Khachikian of the Burakan 
Astrophysical Observatory, Armenia, 
U.S.S.R., visited the Mount Wilson and 
Palomar Observatories for several weeks 
in 1968. At Palomar he obtained spectro- 
grams of a number of the Markarian 
"blue galaxies." 



54 



CARNEGIE INSTITUTION 



Dr. F. J. Low of the University of 
Arizona, with the assistance of Mr. A. W. 
Davidson, carried out infrared observa- 
tions at the Palomar Observatory in 
July 1967. Using a newly developed 
helium-cooled germanium photodetector 
operating at an effective wavelength of 
1.5 microns, observations of about 35 
quasars and related objects were made 
with the 200-inch telescope. All objects 
investigated were readily detected. The 
results were consistent with similar ob- 
servations carried out by Neugebauer 
and Becklin at 2.2 microns. A single mea- 
surement of the polarization of 3C 273 at 
1.5 microns yielded the surprisingly 
large value of 38 ±10%. Unfortunately, 
this single observation could not be re- 
peated. Observations at 5 and 10 microns 
were made on a number of different 
objects, including a detailed study of the 
structure of the infrared core in the 
planetary nebulae NGC 7027. It was 
found that the bright infrared core is 
much smaller in diameter than the 
optical nebula. 

During two dark runs in March and 
May 1968, Dr. Willem J. Luyten of the 
University of Minnesota took 102 plates 
in the program of repeating the Palomar 
Observatory-National Geographic So- 
ciety Survey plates for proper motions. 
Sixty-seven more plates were taken by 
Charles Kowal, thus bringing the total 
of new repeat plates to 752. Only 184 
remain to be taken. Some twenty pairs 
of plates were examined and more than 
11,000 proper-motion stars found, among 
which are several degenerate objects of 
exceedingly high velocity. In the sys- 
tematic search for blue stars and quasars, 
three more plates taken by Sandage 
with Haro's three-image method have 
been examined and 1100 more faint blue 
stars found and published. 

Dr. Beverly T. Lynds of the Uni- 
versity of Arizona made two visits to 
the Observatories to undertake a study 
of the distribution of dark nebulae in 
galaxies. Photographs of galaxies in the 
files were examined and notes were made 



of the more promising objects for future 
study. Then measurements were made of 
the size and position of the dark nebulae 
relative to the bright features defining 
the structure of the individual galaxy. 
To date, Mrs. Lynds has measured M 74, 
83, 100, 101, and NGC 5248. These Sc 
spirals were selected for the initial study 
because the dark clouds are relatively 
easily identified and because Ha inter- 
ference photographs were available for 
delineation of H II spiral features. It is 
hoped to continue this work and to ex- 
tend the analysis to both the earlier 
normal and the barred spirals. 

The Snow telescope was used by Dr. 
Walter E. Mitchell, Jr., of Perkins Ob- 
servatory through the month of Septem- 
ber 1967 in conducting trials and mak- 
ing observations with a solar image slicer 
and a signal averager. Opposing limbs of 
the sun were fed in rapid alternation to 
the spectrograph slit and the observation 
repeated in the continuum and in the 
center of a strong Fraunhofer line. The 
photomultiplier signal was amplified and 
recorded in a computer of average 
transients (Technical Measurements, 
Inc.) so that the scans of opposing limbs 
were juxtaposed on the instrument's 
oscilloscope screen. The integrated line 
and continuum scans were subsequently 
rendered on chart paper for ease of mea- 
surement. The method was demonstrated 
to have promise for the determination 
of chromospheric heights relatively free 
from the effects of poor seeing. Mr. Philip 
E. Barnhart assisted in this program. 

Dr. C. R. O'Dell of the Yerkes Ob- 
servatory employed the prime-focus 
scanner of the 200-inch and the Cas- 
segrain-focus scanner on the 100-inch 
telescope for two investigations. Helium 
and Balmer line-ratio changes as a func- 
tion of position in several double-shell 
planetary nebulae were sought in order 
to quantitatively evaluate the role of 
He I triplet self-absorption. Because of 
bad weather, too few observations were 
obtained, so that no statement on the 
results can be made. A search for the 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



55 



A3080 band of OH, expected to be bright 
under several models of OH radio sources, 
was also planned, but again the weather 
intervened. 

Dr. U. W. Steinlin of the University 
of Basel took plates with the 48-inch 
schmidt from January through April 
1968 for the continuation of work on 
galactic structure. Attention was concen- 
trated on two fields toward the anti- 
center for an investigation of the dis- 
tribution of stars of different age across 
a spiral arm and for the determination 
of the exact location of star formation 
within a spiral arm. Plates were taken 
in the RGU system. Also, plates were 
taken in both RGU and UBV systems of 
a field surrounding the globular cluster 
M 5, first for deriving the exact relation 
between the two systems for Population 
II stars, and secondly for investigating 
density and luminosity functions across 
the inner part of the galactic halo. 

Dr. S. Strom of the Smithsonian Astro- 
physical Observatory and Mr. A. Wood, 
a Harvard graduate student, obtained 
spectrum scans of the A stars in the 
Coma cluster using Oke's scanner on the 
60-inch telescope. These scans will be 
used in conjunction with model atmo- 
spheres to provide an accurate choice of 
T ett and gravity for these stars. Abun- 
dance analyses of several of these stars 
are currently under way and will permit 
a discussion of the variation of metal-to- 
hydrogen ratio between this cluster and 
the Hyades. 

Observations with the 48-inch schmidt 
telescope were made in July 1967 by 
Dr. S. van den Bergh of the David 
Dunlap Observatory, University of 
Toronto. The primary purpose was to 
continue his ten-year program on bright 
variable stars in members of the Local 
Group. Multicolor observations were ob- 
tained of the galaxies M 31, M 33, and 
NGC 6822. A second program involved 
observations of absorbing interstellar 
clouds. These observations were used 
to establish a simple classification sys- 
tem for interstellar clouds. On July 10 



a plate was obtained of Nova Delphini 
1967. Comparison of this plate with the 
Sky Survey led to the identification of a 
blue star with 12<F<13 as the prenova 
(van den Bergh and Racine, Intern. 
Astron. Union Information Bull, on 
Variable Stars No. 212) . A plate of the 
field of the first identified pulsating 
radio source (Ryle and Bailey, Nature, 
217, 907, 1968) was obtained on July 5, 
1967. Comparison with a Sky Survey 
plate obtained June 15, 1950, shows 
that the star that Ryle and Bailey 
identify with the radio source has a 
proper motion fx< 0.030 arc-sec per year. 
The brightness of this star on the two 
red plates differs by less than 0.5 mag. 

Dr. G. Wallerstein of the University 
of Washington has obtained 15 A/mm 
spectrograms of the 45-day Cepheid SV 
Vulpeculae from shortly after minimum 
light to a little beyond maximum. Radial 
velocity curves for lines of different ex- 
citation and ionization levels will be 
derived. A displaced component at Ha 
appears near mid-rising light and re- 
mains until after maximum. The same 
phenomenon was found by Kraft in the 
16-day Cepheid X Cygni. 

Wallerstein has obtained spectrograms 
of several double-line binaries in the 
Hyades as part of the cooperative pro- 
gram with Dr. A. Batten at the Dominion 
Astrophysical Observatory. The primary 
purpose of the program is to obtain 
minimum masses that may be compared 
with the rather low masses derived by 
Eggen from the visual binaries and mov- 
ing parallaxes. A byproduct will be the 
derivation of lithium abundances in 
binaries of various periods and the com- 
parison with lithium abundances of single 
stars in the Hyades. 

An 18 A/mm spectrogram of RS 
Ophiuchi taken four months before the 
outburst of October 26, 1967, showed an 
A-type shell spectrum similar to that 
observed by Sanford in 1947 and by 
Wallerstein from 1960 to 1966. Two 
spectra of RS Oph in the red taken in 
February 1968, four months after maxi- 



56 



CARNEGIE INSTITUTION 



mum, show the [Fe X], [Fe XIV], [Ca 
XV] , and other coronal lines. Theoretical 
calculations have been initiated to de- 
scribe the temperature and density struc- 
ture behind an expanding shock wave as 
it penetrates a circumstellar envelope. 
Preliminary print-outs show that strong 
X-ray emission down to 0.1 A may be 
expected during the early phases of the 
outburst. Coronal line strengths will be 
predicted and compared with relative in- 
tensities obtained with a wide slit. 

An unsuccessful search for C 13 was 
made in two stars with strong CN bands, 
e Pegasi (K2 lb) and a Serpentis (K2 
IIIp) , using the (2, 0) CN bands in the 
near infrared. In cooperation with Mr. 
Thomas Greene, Wallerstein found in a 
Ser that the ratio of C 12 /C 13 is greater 
than 10. This is unfortunately below the 
ratio of 30 expected from the nitrogen 
excess predicted by stellar- evolution 
models and observed in a Ser. 

Robert L. Younkin of the Douglas 
Aircraft Company has used the pulse- 



counting equipment now available with 
the scanner and 60-inch reflector to ex- 
tend previous infrared measurements of 
planets and satellites. Radiance measure- 
ments of Io, Ganymede, Uranus, and 
Neptune were carried out to 1.08 /*. 
Special attention was given to accurate 
determination of the radiance of Uranus 
and Neptune in strong methane bands, 
where the planets are approximately of 
the 10th and 12th magnitudes. This will 
permit calculation of the optical depths 
of the scattering layer above the meth- 
ane. This was particularly important 
in the case of Uranus, where comparison 
with similar measurements in about 1986, 
when the planet will be viewed end-on, 
will yield information on the altitude 
variation of methane and other gaseous 
constituents. 

An attempt was made to verify previ- 
ous indications of anomalous limb dark- 
ening in strong methane bands on 
Uranus. The results were inconclusive 
because of unsatisfactory seeing. 



ASTROELECTRONICS LABORATORY 



As a measure of the acceptance of 
photoelectric data systems using digital 
recording, the following table is pro- 
vided to indicate utilization of such sys- 
tems during the 12-month interval end- 
ing on May 1, 1968. 









Total 








Percentage 




Elec- 


Number 


of Nights 




tronic 


of 


Using 


Telescope 


Setup 


Nights 


Data System 


200-inch 


40* 


127 


29% 


100-inch 


63 


248 


68% 


60-inch 


52 


208 


57% 



* In addition, 14 electronic setups were made 
for image tubes. 

200-Inch Data System 

Under the direction of Dennison, the 
principal effort of the laboratory was 



centered on equipment for the 200-inch 
Hale Telescope. The newly installed data 
system was designed to include two re- 
versing counters, an acquisition-interval 
timer, and a timer that could be adjusted 
to work with the various optical chopper 
disks used in front of both the prime- 
focus scanner and the new multichannel 
Cassegrain scanner. In addition, this 
system will accept information from the 
two telescope encoders, a large variety 
of digital alpha-numeric switches, which 
can be used for miscellaneous informa- 
tion, and information taken directly 
from microswitches on the observing 
equipment. As an example of the latter 
type of input, microswitches have been 
placed on the photometer base so that 
the observer automatically records 
whether he is observing a star or the 
sky background. The data from all the 
various channels, which can be put in 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



57 



any order the observer wishes, are re- 
corded on printed paper tape as well as 
on punched cards for computer use at a 
later time. If the observer so desires, 
he may set the data system to restart 
automatically after each recording cycle 
so that it is possible to collect a series of 
accurately timed photometric measure- 
ments over a period of several hours. This 
information can be rapidly put into the 
computer and the time spectrum of 
optical variations in the source can be 
derived easily. 

As reported last year, the system also 
contains the digital control logic for the 
200-inch coude spectrograph scanning 
photometer. The first operation of this 
entire system, including the use of closed- 
circuit television for the purpose of re- 
motely displaying digital panels, was 
successfully used in November 1967. A 
strip-chart recorder was also employed 
to read the digital output from the pulse 
counters, so that the observer could 
monitor his data while the observations 
were in progress. 

The time-consuming and tedious task 
of installing digital data cables, as well 
as many coaxial cables, on the 200-inch 
telescope was virtually completed dur- 
ing the year. The total number of digital 
conductors linking the telescope with 
the data room is approximately 120. 
Fifty-six cables were also installed in 
the 200-inch to permit the multichannel 
spectrum scanner to operate at the Cas- 
segrain focus. A high-voltage power sup- 
ply and temporary high-voltage cables 
have been installed on the telescope to 
permit the operation of the Cassegrain 
image-tube spectrograph. 

Other Data Systems 

A semiportable data system for Oke 
was constructed during the year, and 
initial tests at Palomar indicate that this 
system is successful. This system in- 
cludes a sidereal or civil-time digital 
clock which is accurate to a fraction of 
a second per day, and which can be 



easily set. A high-speed dual counter 
permits the recording of both channels, 
and an internal digital subtraction cir- 
cuit displays the difference between the 
two data channels. The system also con- 
tains an interval timer and a chopper 
disk timer, and the necessary electronics 
to prepare the output of a digital encoder 
for automatic recording. The recording 
is done on printed paper tape, but if 
future requirements arise for summary 
punched-card recording for computer 
analysis, this can be added at a later 
date with relatively minor changes in 
circuitry. The data system includes con- 
trols for starting, stopping, resetting, 
recording, and holding or suspending the 
data-gathering operations. Stepping mo- 
tor controls also permit the observer to 
set the wavelength of the prime-focus 
scanner from a considerable distance. 
In fact, during the operation of this 
data system with the 4-inch telescope 
attached to the prime-focus scanner 
mounted on the 18-inch telescope at 
Palomar, it has been possible to point 
the telescope at a star and then to con- 
trol remotely the entire operation from 
the operating area, which is one floor 
below the observing floor. 

A similar system is nearing completion 
for investigation in the Division of Geo- 
logical Sciences, which will be used for 
astronomical measurements and also in- 
frared measurements at Mount Wilson. 
The outstanding difference between the 
two data systems lies in the fact that 
the "Geology System" will accept either 
ratios between the two numbers in the 
dual-channel counter or differences, as 
in the case of the absolute-calibration 
program. 

Deutsch has developed some design 
considerations for a stellar spectrograph 
that would operate in the mode of a 
two-dimensional pulse-counter. Such an 
instrument would provide important ad- 
vantages over conventional spectro- 
graphs, notably in latitude, speed, and 
photometric precision, while retaining 
high spectral resolution and finesse. In 



58 



CARNEGIE INSTITUTION 



order to find whether existing low-noise 
video cameras could serve in such a 
spectrograph, he has proposed to make 
laboratory tests of their signal-to-noise 
characteristics at very low light levels. 

General Laboratory Projects 

In the planning of digital data sys- 
tems for astronomical telescopes, it is 
necessary to look far ahead to accommo- 
date the probable needs of observers as 
well as to plan for the most efficient and 
flexible use of the possibilities offered 
by a rapidly developing technology. Dur- 
ing the past year, Howard Sachs, senior 
engineer, extended the formulation of 
general concepts to guide such develop- 
ment. These concepts are designed to 
adapt to the current technological change 
from logic elements constructed of indi- 



vidual components to integrated circuit 
modules. The new integrated circuit 
modules are, in general, not only smaller 
and more reliable, but are notably less 
expensive than the assemblies they are 
designed to replace. The logical func- 
tions served by these modules are un- 
changed, however, so that fundamental 
circuit designs remain substantially the 
same. Another important technological 
factor is the introduction of new high- 
capacity, compact, high-performance, 
low-cost computers. Data systems em- 
ploying such computers at the telescope 
offer, for the future, many attractive pos- 
sibilities for efficiency and flexibility. 
They may also permit the data system 
at the telescope to communicate directly 
with larger computers at distant labora- 
tories. 



INSTRUMENTATION 



Optical Design 

The design of the Gascoigne-type cor- 
rector lens for the Cassegrain focus of 
the Palomar 60-inch photoelectric tele- 
scope has been completed by Bowen and 
Vaughan. The lens, both elements of 
which are of fused silica, gives images of 
Y 2 second or smaller over a field 1^4° in 
diameter. 

Bowen has continued his investigations 
of optics for use with image-intensifier 
tubes. These include a spectrograph 
camera of the solid block, concentric- 
mirror Cassegrain-schmidt type, operat- 
ing at a focal ratio of f/1.4, and various 
mirror systems for reimaging the phos- 
phor on the photographic plate. 

Image-Tube Spectrograph 

The design and construction of a nebu- 
lar spectrograph utilizing a Carnegie- 
RCA image tube has been reported in 
preceding year books. The instrument 
was first used on the 200-inch telescope 
in August 1967. The speed gain of the 



instrument over unaided photographic 
spectroscopy is about 15 in the blue part 
of the spectrum. This gain is derived 
from the ratio of exposure times required 
to obtain a well-exposed spectrogram 
with the prime-focus spectrograph at 
400 A/mm and the Cassegrain image- 
tube spectrograph at 200 A/mm. The 
different dispersions used in the com- 
parison reflect the ratio of resolving 
power for the two instruments. 

The instrument has been in general 
use since December 1967, when the large 
instrument adapter with offset guider 
was installed at the Cassegrain focus of 
the 200-inch telescope. This adapter ac- 
commodates both the Cassegrain spectro- 
graph and the new 33-channel spectrom- 
eter; it will be available also for other 
auxiliary instruments. 

Multichannel Spectrometer 

The new spectrometer, built with a 
grant provided by the Advanced Re- 
search Projects Agency, has been com- 
pleted and installed at Palomar on the 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



59 



200-inch Hale telescope. Since only one 
of the two cold boxes is operational, ob- 
servations during the next few months 
will be made with 16 channels covering 
the spectral range from AA5500 to 11,200. 
Within the next two months, the blue 
cold box, with 13 of the 17 possible photo- 
multiplier tubes, will be mounted on the 
instrument. The data system has at pres- 
ent enough counters to accommodate 
20 channels. Advantage will be taken 
of the fact that a small computer is to 
be added to the 200-inch data system 
during the next six months. With the 
computer it will be possible to activate 
all 33 channels without the addition of 
more bulky pulse-counters. Laboratory 
tests have shown that there will be little 
difficulty in operating a large number 
of photomultiplier tubes, amplifiers, and 
counters. High reliability is achieved by 
using pulse-counting techniques and 
solid-state electronics. 

Polarimeter 

The construction of the single-channel 
sky-compensated polarimeter has been 
completed. It was put into operation at 
both the 100-inch and 60-inch telescopes 
by Visvanathan. Observations of twelve 
standard polarized stars from Hiltner's 
list give a mean difference in P (Vis- 
vanathan-Hiltner) of ±0.003 mag, or 
0.15%, and ±1.8° in 6. 

60-Inch Photometric Telescope 

This project for the Palomar Ob- 
servatory is proceeding under the grant 
made by the National Science Founda- 
tion to the Carnegie Institution. The 
60-inch fused silica disk has been fine- 
lapped on the 72-inch grinding machine 
close to the final radius of curvature prior 
to figuring to the final Ritchey-Chretien 
design. 

All major mechanical components of 
the telescope and most subassembly 
items have now been fabricated and 
tested; shop assembly is proceeding un- 



der the direction of Rule. The main 
declination and right ascension gears 
have been made and tested; the drive 
gearing units are complete and ready 
for assembly in proper order with the 
overall schedule. The oil pad and pump- 
ing system are complete and temporarily 
in use for shop assembly and tests. Bear- 
ing tests and gear burnishing are pro- 
ceeding. The upper tube and secondary 
mirror cage are ready for mounting. 

The 60-inch primary cells and mirror- 
support units were completed in 1967 
and all units have now been thoroughly 
inspected, tested, and calibrated in all 
attitudes and for all mirror load condi- 
tions. The 18 axial lever-support units 
under all load conditions show total 
friction of less than 0.1%, as do the 18 
radial compound lever units. The flip- 
cage and secondary focus unit are com- 
plete with mirror cells and are ready 
for drive tests. Limit switches, covers, 
and accessories are being made and fitted 
as the assembly proceeds. 

Since the approval of the telescope 
site at Palomar in 1967, much has been 
done to prepare the site for construction 
of the Oscar G. Mayer Memorial Dome. 
Surveys were completed for location, 
roads, power lines, and utilities. Founda- 
tion investigations and site borings were 
made, resulting in recommendations for 
location of telescope pier and coude 
support on sound bedrock. The site was 
stripped of topsoil to the rock outcrop- 
ping level, ready for building founda- 
tions. An access road 18 feet wide with 
maximum grade of about 8% was com- 
pleted and trenches dug for underground 
water, telephone, and temporary power 
service. All site development was ready 
by May 1, 1968. Building contractors 
are expected to start construction by the 
end of June. 

Two new Palomar cottages were com- 
pleted — one for the night assistant at the 
60-inch telescope, the other as a staff cot- 
tage and lodge annex. 



60 



CARNEGIE INSTITUTION 



Modification of Mount Wilson 60-inch 
Telescope 

With completion of major modifica- 
tions to the dome, considerable progress 
has been made in design and shop fabri- 
cation of components required for im- 
provement and renovation of the tele- 
scope itself. The remounting of secondary 
Cassegrain and coude mirrors into a 
back-to-back turnover mounting has 
been completed. The new declination 
drive gear and mechanism, with remote 
coordinating readout, is nearing comple- 
tion. The right-ascension drive-gear unit 
is under construction, while polar-axis- 
drive modifications are being planned. 
Revision of the tube for the 3-mirror 
coude system and design work on the 
coude room base will be completed 
shortly. 

Solar Instrumentation 

A Littrow spectrograph for the new 
solar equatorial telescope was engineered 
by Boiler and Chivens Division of 
Perkin-Elmer Corporation under the su- 
pervision of Zirin; completion is sched- 
uled for December 1968. The spectro- 
graph has a focal length of 945 inches. 
A special fluorite-BK2 lens designed by 
Dr. Richard Dunn of Sacramento Peak 
Observatory provides essentially achro- 
matic images from AA3000 to 10,000. The 
spectrograph will be mounted vertically 
at the coude focus of the equatorial tele- 
scope. It includes a rapid film transport, 
stepping slit, grating control, and various 
other provisions to permit rapid-sequence 
solar spectroscopy. Two interchangeable 
gratings will be used. 

Instrumental improvements at the 
Hale Solar Laboratory include a photo- 
electric guider built by Julian, a varia- 
ble-speed grating drive, and a tilted slit 
mechanism that permits photography of 
the slit jaws during the making of 
spectrograms. 

Two small photoheliographs designed 
by Mr. George Carroll of Lockheed Solar 
Observatory have been installed, one in 



Tel Aviv, Israel, and one on top of the 
Robinson Laboratory. These 5-inch re- 
fractors are equipped with Halle bire- 
fringent filters for Ha and pulse cameras, 
as well as photoelectric tracking units. 
The two locations, Israel and California, 
present the possibility of obtaining 
nearly continuous records of the sun. Se- 
lected portions of the 50-mm solar image 
are recorded on 35-mm film. The tele- 
scope on Robinson Laboratory has been 
successful, and results are described else- 
where. The Tel Aviv telescope has suf- 
fered from filter problems but is cur- 
rently in operation. 

Robinson Solar Telescope 

Rehabilitation of the Robinson Labo- 
ratory Solar Telescope is rapidly nearing 
completion. This work is funded by a 
NASA contract to Zirin and has been 
directed by J. David Bohlin. The tele- 
scope is an integral part of Robinson 
Laboratory and was originally designed 
and 90% completed over thirty years 
ago (Hale, G. E., Astrophys. J., 82, 111— 
139, 1935). The coelostat consists of 
36-inch and 30-inch diameter flats housed 
under a 23-foot dome on the Robinson 
roof. A free-standing, octagonal shaft 
extends 65 feet deep. The lower 75 feet 
was planned for vertical spectrographs 
(never constructed), while the upper 
50-foot section contains the image-form- 
ing mirror system. The coelostat mirrors 
illuminate a 26-inch off-axis parabola 
located 5 feet above the first basement 
level (the observing level). The primary 
f/16 beam is folded by a Cassegrain 
secondary mirror that may be either a 
flat or any one of three convex mirrors, 
giving final solar-image diameters of 
about 4, 8, 14, or 22 inches. 

The first phase of rehabilitation will 
put the telescope into operation as origi- 
nally designed (using just the 8-inch 
diameter image), followed by a second 
phase of additional instrumentation when 
operational characteristics have been 
evaluated. General overhaul of various 



MOUNT "WILSON AND PALOMAR OBSERVATORIES 



61 



mechanical and electrical systems was 
begun in 1967. A two-level observing 
floor was installed in the shaft at the 
first basement level in January 1968. 
New cells for the 30-inch and 26-inch 
mirrors were fabricated by Boiler and 
Chivens. The mirrors originally con- 
structed for the solar telescope have been 
realuminized and overcoated with sili- 
con-oxide; they were installed in May. 
The 36-inch coelostat flat is a glass-on- 
metal type produced by the Philips 
Lamp Works of Holland prior to 1935. 
It consists of about one inch of glass 
fused onto a ribbed-case iron base hav- 
ing the same coefficient of expansion as 
the glass. The 30-inch flat is the first of 
the ribbed Pyrex blanks cast by Corning 
Glass Works as part of the 200-inch 
project. The 26-inch is solid Pyrex while 
the Cassegrain secondaries are early- 
type fused silica. Final optical and 
mechanical adjustments are in progress 
leading to an operational evaluation. 



Photoheliograph for Apollo 

Howard and Zirin have been conduct- 
ing in coordination with the Jet Pro- 
pulsion Laboratory a design study of a 
proposed high-resolution solar telescope 
for the Apollo applications program of 
the National Aeronautics and Space Ad- 
ministration. The aim is to secure both 
white-light and filter photographs of 
high quality. During this year, funding 
for the project was received, and con- 
siderable progress was made by a group 
working at the Jet Propulsion Labora- 
tory. An optical design of a Gregorian 
system was completed and approved, 
and a primary mirror of internally cooled 
ultra-low-expansion fused silica was de- 
signed. A quarter-scale model has been 
built and work is now under way on a 
full-scale flight-verification unit. It is 
felt that despite uncertainties of the na- 
tional space program and the various 
problems involved with space telescopes, 
this project will eventually be neces- 
sary in the quest for high resolution. 



PHOTOGRAPHIC LABORATORY 



The Photographic Laboratory under 
the supervision of William C. Miller con- 
tinued the program of routine tests of all 
photographic materials received from the 
manufacturers. An improved system of 
reporting results of such tests to the Staff 
is under development; it will provide 
observers with a curve for each emulsion, 
giving speeds at all wavelengths rather 
than at only one, as is the case with the 
existing system. 

Difley has devoted most of his time 
to the production of publication and 
work prints for the Staff and to assisting 
with plate tests. Of particular interest 
are the results that he and Miller ob- 
tained in formulating an improved de- 
veloper for astronomical use with the 
Kodak spectroscopic plates. The new de- 
veloper has been tested extensively both 
in the laboratory and in the dome. After 



submitting samples to several of our 
Staff, samples were sent to interested 
members of other observatories for 
evaluation. All reports to date have been 
favorable. For a density of 0.3 above 
fog, Ha plates developed in the new solu- 
tion require only 60% of the exposure 
needed for plates developed in D-76. At 
a density of 1.0 above fog, the figure is 
80%. With the 103a plates the new de- 
veloper required only 55% as much ex- 
posure to produce a density of 1.0 above 
fog, as with D-19. All other physical 
characteristics of the developed images 
produced by the new solution were at 
least as good as those with the older 
developers. Of particular interest is a 
marked decrease in image-edge effects 
produced by the new formula. The only 
possible criticism so far is that contrast 
cannot be controlled by varying develop- 



62 



CARNEGIE INSTITUTION 



ment time with the new developer. The 
old formulae will gradually be replaced 
in all darkrooms by the new solution. 

Tests of possible reduction of reciproc- 
ity failure at long exposure times by 
refrigerating the photographic plates 
were conducted in cooperation with 
Westphal, who designed the cold box 
and temperature-control units. Optimum 
results with spectroscopic plates were 
achieved at — 20°C, but gains were so 
small that they did not justify the 
mechanical and optical difficulties in- 
volved in providing refrigerated plate 
boxes at the telescopes. Gains through 
other means, such as baking, pre-ex- 
posure, and the new developer formula, 
exceeded those resulting from refrigera- 
tion. 

The demand for astronomical photo- 
graphs as educational and teaching aids 
became so great that steps were taken 
to increase their availability to schools 
and colleges, and simultaneously de- 



crease the time spent by Miller in ad- 
ministering the Observatories' system of 
distribution and sale, by negotiating a 
contract with the Hubbard Scientific 
Company of Northbrook, Illinois, a 
leader in the distribution at wholesale 
of educational material and teaching aids 
to schools and colleges. 

Cooperation has continued with the 
Eastman Kodak Company in mutual 
efforts to improve materials and methods 
for astronomical photography. Informa- 
tion for their data books has been given, 
and Miller is cooperating with their tech- 
nical writers in the preparation of a new 
booklet devoted to technical and practi- 
cal information of use to astronomical 
photographers. 

Miller has actively participated in 
meetings of both the International Astro- 
nomical Union and American Astro- 
nomical Society Working Groups on 
Photographic Materials. 



SITE INVESTIGATION 



Chile 

Work in Chile, under the immediate 
supervision of Buck, has been directed 
mainly toward planning for the develop- 
ment of Cerro Morado. Operation of the 
two astronomical seeing monitors was 
continued there until the end of the re- 
port year. Recording of microthermal 
fluctuations of the air by means of 
thermistors at four levels on each of 
four 100-foot towers, as installed by 
Westphal, was continued. The resulting 
data are definitely useful in the selection 
of specific telescope sites on the summit 
area. The recording of meteorological 
data continues, however, not only on 
Morado but on some other summits in 
the same general area of Chile. The in- 
terest and cooperation offered by the 
Associated Universities of Research in 
Astronomy, the University of Chile, and 
the government of Chile have supported 



and justified the further formulation and 
study of comprehensive plans to develop 
Cerro Morado as a site for a major 
astronomical observatory. From Morado 
field data, and with consideration of re- 
quirements based on experience at Mount 
Wilson and Palomar, Rule has prepared 
a site development plan for telescope lo- 
cations, road access, power, water, and 
other services, providing for optimum 
utilization of the summit area. 

In May 1968, J. W. Boise, Bursar of 
the Carnegie Institution, joined with Bab- 
cock, Rule, and Buck in a visit to Chile 
for site studies and for a meeting with 
representatives of AURA to discuss spe- 
cific plans for boundary locations and 
for land use by the Institution. 

Big Bear Solar Observatory 

A search for a new observing station 
that would permit observation of the 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



63 



sun with good seeing for extended inter- 
vals throughout the day was begun some 
years ago by Leighton, Howard, and 
Zirin. It has now been decided to locate 
at Big Bear Lake in the San Bernardino 
Mountains. The site offers a 6700-foot 
altitude and a stable daytime air mass 
produced by the cool water of the lake; 
sky transparency is good. The California 
Institute of Technology has leased frum 
the Bear Valley Mutual Water Company 
approximately seven acres of land, in- 
cluding 500 feet of lake frontage and a 



1000-foot long strip into the lake. Grants 
from the Max Fleischmann Foundation 
of Nevada, the National Science Founda- 
tion and the National Aeronautics and 
Space Administration will permit the 
construction of a solar observatory with 
a 50-foot concrete tower on a small 
island about 1000 feet from the north 
shore of the lake. The new solar equa- 
torial telescope, now completed, will be 
erected on top of the tower. A labora- 
tory and residence building will be con- 
structed on shore. 



BIBLIOGRAPHY 



Anderson, Jean H., see Luyten, Willem J. 
Arp, Halton, Radio sources and their galaxies 

of origin. Astrofizika (Burakan), 4, 59-91, 

1968. 
Arp, Halton, Analysis of stellar content and 

Cepheid evolution in NGC 1866. Astrophys. 

J., 149, 91-108, 1967. 
Arp, Halton, Investigations of the association 

between radio sources and peculiar galaxies. 

Astrophys. J., 162, 633-637, 1968. 
Arp, Halton, A compact galaxy (III Zw 2) 

and a compact radio galaxy (3C 120) with 

Seyfert-type spectra. Astrophys. J., 162, 1101— 

1107, 1968. 
Arp, Halton, A counter-jet in M 87. Astrophys. 

Letters (England), 1, 1-5, 1967. 
Arp, Halton, Lines of galaxies from radio 

sources. Publ. Astron. Soc. Pacific, 80, 129- 

154, 1968. 
Arp, Halton, Nature of the ring around M 81. 

Sky and Telescope, 85, 11, 1988. 
Arp, Halton, and Jeffrey D. Scargle, A high- 
latitude planetary nebula. Astrophys. J., 150, 

707-710, 1967. 
Arp, Halton, and A. D. Thackeray, The star 

cluster NGC 1866 and its Cepheids. Astro- 
phys. J., 149, 73-89, 1967. 
Arp, Halton, E. Khachikian, C. R. Lynds, and 

D. W. Weedman, Two new Seyfert galaxies. 

Astrophys. J. (Letters), 152, L103-L104, 1968. 
Babcock, Horace W., Frederick Hanley Seares 

(1873-1964), in Am. Phil. Soc. Year Book, 

1967, pp. 145-148. 
Babcock, Horace W., Optical astronomy in 

perspective. Astron. Soc. Pacific Leaflet No. 

462, 8 pp., December 1967. 
Babcock, Horace W., Concentration of elements 

over the surface of a magnetic star, in 

Magnetic and Related Stars, pp. 97-100, 



Robert C. Cameron, ed., Mono Book Corpo- 
ration, Baltimore, Md., 1967. 

Bahcall, John, and Maarten Schmidt, Does the 
fine-structure constant vary with cosmic 
time? Phys. Rev. Letters, 19, 1294-1295, 1967. 

Bahcall, John N., Wallace L. W. Sargent, and 
Maarten Schmidt, An analysis of the ab- 
sorption spectrum of 3C 191. Astrophys. J. 
(Letters), 149, L11-L15, 1967. 

Barbon, R., A. Braccesi, and R. Fanti, Identifi- 
cation of seven 4C sources with blue objects. 
11 Nuovo Cimento, LII-B, No. 1, 262-266, 
1967. 

Barbon, R., see also Zwicky, Fritz. 

Baum, William A., see McGee, J. D. 

Becklin, E. E., see Oke, J. B. 

Bertola, F., see Gates, Howard S. 

Boesgaard, Ann Merchant, Observations of 
beryllium in stars, in Highlights oj Astron- 
omy as Presented at the XHIth General 
Assembly, Intern. Astron. Union, Prague, 
1967, pp. 237-242, L. Perek, ed., D. Reidel 
Publishing Co., Dordrecht-Holland, 1968. 

Bowen, Ira S., Astronomical optics, in Ann. 
Rev. Astron. Astrophys., Vol. 5, pp. 45-66, 
Annual Reviews, Inc., Palo Alto, Calif., 1967. 

Braccesi, A., Lynds, R., and Sandage, A., Spec- 
troscopic and photometric data for a sample 
of quasi-stellar objects identified by their 
infrared excess. Astrophys. J. (Letters), 152, 
L105-L110, 1968. 

Braccesi, A., see also Barbon, R. 

Bumba, V., Robert Howard, and Sara F. Smith, 
The sun as a magnetic star, in Magnetic and 
Related Stars, pp. 131-138, Robert C. 
Cameron, ed., Mono Book Corp., Baltimore, 
Md., 1967. 

Bumba, V., R. Howard, M. J. Martres, and 
I. Soru-Iscovici, Patterns of active-region, 



64 



CARNEGIE INSTITUTION 



magnetic-field development, in Intern. Astron. 
Union Symp. No. 35, pp. 13-24, K. O. Kiepen- 
heuer, ed., D. Reidel Publishing Co., Dor- 
drecht-Holland, 1968. 

Bumba, V., see ako Howard, Robert; Kriiger, 
A. 

Ciatti, F., see Gates, Howard S. 

Cohen, J. G., and G. L. Grasdalen, Limits on 
the C^/C 13 ratio in metal-deficient stars. 
Astrophys. J. (Letters), 151, L41-L44, 1968. 

Conti, Peter S., Metallic-line stars in the 
Pleiades, in Magnetic and Related Stars, pp. 
321-327, Robert C. Cameron, ed., Mono Book 
Corp., Baltimore, Md., 1967. 

Danziger, I. J., and R. J. Dickens, Spectro- 
photometry of new short-period variable 
stars. Astrophys. J., 149, 55-72, 1967. 

Danziger, I. J., see also Kuhi, L. V. 

Davis, Dorothy N., Is there technetium in 
HR 1105? Astrophys. J. (Letters), 152, L13- 
L15, 1968. 

Davis, Dorothy N., see also Schadee, Aert. 

Deutsch, Armin J., Magnetic fields (stellar), in 
Encyclopedia oj Atmospheric Sciences and 
Astrogeology , pp. 534-535, Rhodes W. Fair- 
bridge, ed., Reinhold Publishing Corp., New 
York, 1967. 

Deutsch, Armin J., Inferences from the sta- 
tistics of stellar rotation, in Magnetic and 
Related Stars, pp. 181-207, Robert C. 
Cameron, ed., Mono Book Corp., Baltimore, 
Md., 1967. 

Dickens, R. J., Robert P. Kraft, and W. Krze- 
minski, The effect of rotation on the colors 
and magnitudes of stars in Praesepe. Astron. 
J., 78, 6-13, 1968. 

Dickens, R. J., see also Danziger, I. J. 

Eggen, Olin J., Narrow- and broad-band pho- 
tometry of red stars, I, Northern giants. 
Astrophys. J., Suppl. Ser., 14, No. 131, 307- 
358, 1967. 

Eggen, Olin J., and Jesse L. Greenstein, Ob- 
servations of proper-motion stars, III. Astro- 
phys. J., 150, 927-942, 1967. 

Fanti, R., see, Barbon, R. 

Garrison, R. F., Stellar associations. Astron. 
Soc. Pacific Leaflet No. 465, 8 pp., March 
1968. 

Garrison, R. F., The spectrum of star No. 1 
in NGC 2024. Publ. Astron. Soc. Pacific, 80, 
20-24, 1968. 

Gates, Howard S., F. Zwicky, F. Bertola, F. 
Ciatti, and K. Rudnicki, The supernova 
SN 1966b (Z SN-177) in NGC 4688. Astron. 
J., 72, 912-914, 1967. 

Grasdalen, G. L., see Cohen, J. G. 

Greenstein, Jesse L., Astronomy, in Americana 



Annual, pp. 85-88, The Americana Corpora- 
tion, New York, 1968. 

Greenstein, Jesse L., Horizontal-branch stars 
and interstellar lines at high latitude. Astro- 
phys. J., 152, 431^37, 1968. 

Greenstein, Jesse L., and "Virginia L. Trimble, 
Einstein redshift in white dwarfs. Astrophys. 
J., 149, 283-289, 1967. 

Greenstein, Jesse L., see also Eggen, Olin J. 

Gunn, James E., see Norton, Robert H. 

Herzog, Emil R., On the identification of five 
galaxies in the Virgo cluster. Publ. Astron. 
Soc. Pacific, 79, 627-629, 1967. 

Howard, Robert, Magnetic field of the sun 
(observational), in Ann. Rev. Astron. Astro- 
phys., Vol. 5, pp. 1-24, Annual Reviews, Inc., 
Palo Alto, Calif., 1967. 

Howard, Robert, Solar magnetism, in Encyclo- 
pedia oj Atmospheric Sciences and Astro- 
geology, pp. 880-881, Rhodes W. Fairbridge, 
ed., Reinhold Publishing Corp., New York, 
1967. 

Howard, Robert, Budapest Symposium on solar 
active regions. Sky and Telescope, 84, 296, 
1967. 

Howard, Robert, Velocity fields in the solar 
atmosphere. Solar Physics, 2, 3-33, 1967. 

Howard, Robert, and William C. Livingston, 
Some observations bearing on the problem 
of short-period oscillations. Solar Physics, 8, 
434-438, 1968. 

Howard, Robert, V. Bumba, and Sara F. Smith, 
Atlas oj Solar Magnetic Fields, August 1959- 
June 1966, Carnegie Inst, oj Washington 
Publ. No. 626, Washington, D.C., 1967. 

Howard, Robert, Andrew W. Tanenbaum, and 
John M. Wilcox, A new method of magneto- 
graph observations of the photospheric 
brightness, velocity, and magnetic fields. Solar 
Physics, 4, 286-299, 1968. 

Howard, Robert, see also Bumba, V.; Kriiger, 
A.; Smith, Sara F.; Wilcox, John M. 

Irwin, John B., The need for two large 
Schmidts, in Vistas Astron., Vol. 10, pp. 143- 
147, Arthur Beer, ed., Pergamon Press, Ox- 
ford and New York, 1968. 

Joy, Alfred H., Eclipsing stars, Astron. Soc. 
Pacific Leaflet No. 464, 8 pp., February 1968. 

Joy, Alfred H., Frederick Hanley Seares (1873- 
1964), in Biographical Memoirs, 89, 417-44, 
published for National Academy of Sciences 
of the United States by Columbia University 
Press, New York, 1967. 

Joy, Alfred H., Mount Wilson solar physicist 
dies (Harold D. Babcock). Sky and Tele- 
scope, 85, 350, 1968. 

Jugaku, Jun, and Wallace L. W. Sargent, Stud- 
ies of the peculiar A stars, V, Continuous 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



65 



energy distributions. Astrophys. J., 151, 259- 
268, 1968. 

Khachikian, E., see Arp, Halton. 

Karpowicz, M., On the non-existence of clus- 
ters of clusters of galaxies, III. Z. Astrophys., 
67, 139-142, 1967. 

Karpowicz, M., see also Zwicky, Fritz. 

Khogali, A., see McGee, J. D. 

Kleczek, J., see Kriiger, A. 

Koelbloed, David, Abundances in two ex- 
tremely high-velocity stars. Astrophys. J., 
149, 299-315, 1967. 

Kraft, Robert P., The possible presence of Ap 
stars in young stellar groups, in Magnetic 
and Related Stars, pp. 303-309, Robert C. 
Cameron, ed., Mono Book Corp., Baltimore, 
Md., 1967. 

Kraft, Robert P., On the structure and evolu- 
tion of W Ursae Majoris stars. Publ. Astron. 
Soc. Pacific, 79, 395^13, 1967. 

Kraft, Robert P., see also Dickens, R. J.; 
McGee, J. D. 

Kristian, Jerome, Allan Sandage, and J. A. 
Westphal, Rapid photometric and spectro- 
scopic variations of the X-ray source Cyg 
X-2. Astrophys. J. (Letters), 150, L99-L106, 
1967. 

Kristian, Jerome, and John V. Peach, The op- 
tical image of 3C 287. Astrophys. J. (Letters), 
152, L161-L164, 1968. 

Kriiger, A., V. Bumba, R. Howard, and J. 
Kleczek, The interplanetary sector structure 
and solar radio emission. Bull. Astron. Inst. 
Czechoslovakia, 19, 180-181, 1968. 

Krzeminski, W., see Dickens, R. J. 

Kuhi, L. V., and I. J. Danziger, A spectro- 
photometry study of Delta Scuti and Delta 
Delphini. Astrophys. J., 149, 47-53, 1967. 

Lambert, D. L., and E. A. Mallia, The C^/C 13 
ratio in the solar photosphere. Astrophys. 
Letters (England), 1, 85-87, 1968. 

Lambert, D. L., and B. Warner, The abun- 
dances of the elements in the solar photo- 
sphere, V, The alkaline earths Mg, Ca, Sr, 
Ba. Monthly Notices Roy. Astron. Soc, 140, 
197-221, 1968. 

Lasker, Barry M., The energization of the in- 
terstellar medium by ionization-limited H 
II regions. Astrophys. J., 149, 23-28, 1967; 
and in Intern. Astron. Union Symp. No. 81, 
Radio Astronomy and the Galactic System, 
pp. 125-126, H. van Woerden, ed., Academic 
Press, London, 1967. 

Livingston, William C, see Howard, Robert; 
Norton, Robert H. 

Luyten, Willem J., Jean H. Anderson, and 
Allan Sandage, Faint blue stars in a field 
centered at 9:32+24°. Search for Faint Blue 



Stars, A, Pt. XLVIII, 1 p., published by The 
Observatory, University of Minnesota, Min- 
neapolis, Minn., 1968. 

Luyten, Willem J., Jean H. Anderson, and Al- 
lan Sandage, Faint blue stars in a field cen- 
tered at 12:34+30°. op. cit., Pt. XLIX, 1 p., 
1968. 

Lynds, C. R., see Arp, Halton; Braccesi, A. 

McCammon, D., see Munch, Guido. 

McGee, J. D., A. Khogali, W. A. Baum, and 
R. P. Kraft, Stellar rotations observed with 
a spectracon. Monthly Notices, Roy. Astron. 
Soc, 187, 303-310, 1967. 

Makita, Mitsugu, A study of the green TiO 
band in the sunspot spectrum. Solar Physics, 
3, 557-562, 1968. 

Mallia, E. A., see Lambert, D. L. 

Manwell, Tom, and Michal Simon, Are the 
optical luminosity functions of 3C 273 ran- 
dom? Nature, 217, 931, 1968. 

Marsden, B. G., see Rudnicki, K. 

Martres, M. J., see Bumba, V. 

Moffet, A. T., M. Schmidt, C. H. Slater, and 
A. R. Thompson, The radio source 3C 17. 
Astrophys. J., 148, 283-286, 1967. 

Munch, Guido, The emission in the central 
regions of M 31, in Intern. Astron. Union 
Symp. No. 29, Instability Phenomena in 
Galaxies, pp. 71-74, M. Arakelpan, ed., 
Mezdunarodnaja Kniga, Moscow, 1968. 

Munch, Guido, Internal motions in the plane- 
tary nebula NGC 6543, in Intern. Astron. 
Union Symp. No. 84, Planetary Nebulae, 
pp. 259-266, D. E. Osterbrock and C. R. 
O'Dell, eds., D. Reidel Publishing Co., Dor- 
drecht-Holland, 1968. 

Munch, Guido, Interstellar absorption lines, in 
Stars and Stellar Systems, Vol. VII, Nebulae 
and Interstellar Matter, chap. 7, pp. 365-402, 
Barbara M. Middlehurst and Lawrence H. 
Aller eds., University of Chicago Press, 1968. 

Munch Guido, G. Neugebauer, and D. Mc- 
Cammon, Infrared coronal lines, II, Obser- 
vation of [Si X] XI. 43 n and [Mg VIII] 
X3.03 /i. Astrophys., J., 149, 681-686, 1967. 

Neugebauer, G., see Munch, Guido; Oke, J. B. 

Newkirk, G. A., see Norton, Robert H. 

Norton, Robert H., James E. Gunn, W. C. 
Livingston, G. A. Newkirk, and H. Zirin, 
Surveyor I observations of the solar corona. 
J. Geological Res., 72, 815-817, 1967. 

Oke, J. B., Effective temperatures and gravi- 
ties of X Bootis stars. Astrophys. J., 150, 513- 
520, 1967. 

Oke, J. B., Optical variations in the radio 
galaxy 3C 371. Astrophys. J. (Letters), 150, 
L5-L8, 1967. 



66 



CARNEGIE INSTITUTION 



Oke, J. B., A study of the Seyfert galaxy NGC 
4151. Proc. Astron. Soc. Australia, 1, 7, 1966. 

Oke, J. B., Photoelectric spectrophotometry of 
quasi-stellar sources, in Intern. Astron. Un- 
ion Symp. No. 29, Instability Phenomena 
in Galaxies, pp. 328-341, M. Arakeljan, ed., 
Mezdunarodnaja Kniga, Moscow, 1967. 

Oke, J. B., A study of the nucleus of the Sey- 
fert galaxy NGC 4151 op. cit., pp. 91-96. 

Oke, J. B., Report on absolute spectrophoto- 
metry, in Trans. Intern. Astron. Union, Vol. 
13A, Reports on Astronomy, pp. 637-638, L. 
Perek, ed., D. Reidel Publishing Co., Dord- 
recht-Holland, 1967. 

Oke, J. B., and Wallace L. W. Sargent, Nucleus 
of the Seyfert galaxy, NGC 4151. Astrophys. 
J., 151, 807-824, 1968. 

Oke, J. B., and Rudolph E. Schild, A practical 
multiple-reflection technique for improving 
the quantum efficiency of photomultiplier 
tubes. Applied Optics, 7, 617-622, 1988. 

Oke, J. B., W. L. W. Sargent, G. Neugebauer, 
and E. E. Becklin, A variable radio-quiet 
compact galaxy, I Zw 1727+50. Astrophys. J. 
{Letters), 150, L173-L176, 1967. 

Oke, J. B., see also Visvanathan, N. 

Peach, John V., see Kristian, Jerome. 

Racine, Rene, Preliminary colors of faint ob- 
jects around M 87. Publ. Astron. Soc. Pacific, 
80, 326-329, 1968. 

Rees, M. J., and Michal Simon, Evidence for 
relativistic expansion in variable radio sources. 
Astrophys. J. (Letters), 152, L145-L148, 
1968. I 

Rudnicki, K., Observations of 5 old superno- 
vae. Astron. Nachr., 290, 85-87, 1967. 

Rudnicki, K., Observations of 6 supernovae. 
Astron. Nachr., 290, 135-139, 1967. 

Rudnicki, K., General features of clusters of 
galaxies. Soviet Astron.-A.J., 11, 59-62, 1987. 

Rudnicki, K., and B. G. Marsden, Periodic 

comet Grigg-Skjellerup (1966f). Intern. 

Astron. Union Circular No. 2029, August 9, 
1967. 

Rudnicki, K., see also Gates, Howard S.; 
Zwicky, Fritz. 

Rust, David M., Chromospheric explosions and 
satellite sunspots, in Intern. Astron. Union 
Symp. No. 35, Structure and Development 
of Solar Active Regions, pp. 77-84, K. O. 
Kiepenheuer, ed., D. Reidel Publishing Co., 
Dordrecht-Holland, 1968. 

Sandage, Allan, Additional data on the optical 
variation of 3C 371 and other properties of 
N-type galaxies. Astrophys. J. (Letters), 150, 
L9-L12, 1967. 



Sandage, Allan, Redshifts of nine radio gal- 
axies. Astrophys. J., (Letters), 150, L145- 
L146, 1967. 

Sandage, Allan, Optical variation of the nuclei 
of three compact galaxies together with new 
photometric data for Seyfert galaxies. Astro- 
phys. J., (Letters), 150, L177-L182, 1967. 

Sandage, Allan, A new determination of the 
Hubble constant from globular clusters in 
M 87. Astrophys. J., (Letters), 152, L149- 
L154, 1968. 

Sandage, Allan, The time scale for creation, in 
Galaxies and the Universe, pp. 75-112, Lode- 
wijk Woltjer, ed., Columbia University 
Press, New York, 1968. 

Sandage, Allan, Radio galaxies and quasars, 
II, in Highlights of Astronomy as Presented 
at the Xlllth General Assembly of the 
Intern. Astron. Union, Prague 1967, pp. 45- 
70, L. Perek, ed., D. Reidel Publishing Co., 
Dordrecht-Holland, 1968. 

Sandage, Allan, Three photoelectric measure- 
ments of LP 93-21. Intern. Astron. Union 
Circular No. 2069, April 26, 1968. 

Sandage, Allan, Observational cosmology. Ob- 
servatory, 88, 91-106, 1968. 

Sandage, Allan, Optical data on quasi-stellar 
sources, in Trans. Intern. Astron. Union, 
Vol. 13A, Reports on Astronomy, pp. 579- 
587, L. Perek, ed., D. Reidel Publishing Co., 
Dordrecht-Holland, 1967. 

Sandage, Allan, and G. A. Tammann, A com- 
posite period-luminosity relation for Ce- 
pheids at mean and maximum light. Astro- 
phys. J., 151, 531-546, 1968. 

Sandage, Allan, and Robert Wildey, The 
anomalous color-magnitude diagram of the 
remote globular cluster NGC 7006. Astro- 
phys. J., 150, 469-482, 1967. 

Sandage, Allan, see also Braccesi, A.; Kristian; 
Jerome; Luyten, Willem J.; Tammann, A. 
G. 

Sargent, Wallace L. W., Spectra and masses of 
the blue "horizontal-branch" stars in M 67. 
Astrophys. J., 152, 885-889, 1968. 

Sargent, Wallace L. W., A luminous compact 
galaxy with strong Fe II emission lines. 
Astrophys. J. (Letters), L31-L34, 1968. 

Sargent, Wallace L. W., A new intrinsically 
bright Seyfert galaxy. Publ. Astron. Soc. 
Pacific, 79, 369-371, 1967. 

Sargent, Wallace L. W., and Leonard Searle, 
The interpretation of the helium weakness 
in halo stars. Astrophys. J. (Letters), 150, 
L33-L38, 1967. 

Sargent, Wallace L. W., and Leonard Searle, 
A quantitative description of the spectra 
of the brighter Feige stars. Astrophys. J., 
152, 443-452, 1968. 



MOUNT WILSON AND PALOMAR OBSERVATORIES 



67 



Sargent, Wallace L. W., see also Bahcall, John 
N.; Jugaku, Jun; Oke, J. B. 

Scargle, Jeffrey D., The Crab Nebula— 913 
years after its outburst. Astron. Soc. Pacific 
Leaflet No. 457, 8 pp., July 1967. 

Scargle, Jeffrey D., A note on relativistic mag- 
netohydrodynamics. Astrophys. J., 151, 791- 
796, 1968. 

Scargle, Jeffrey D., see Arp, Halton. 

Schadee, Aert, Molecular band intensities in 
G and K stars. Astrophys. J., 151, 239-258, 
1968. 

Schadee, Aert, and Dorothy N. Davis, Zircon- 
ium isotope ratios in HR 1105. Astrophys. J., 
152, 169-179, 1968. 

Schild, Rudolph E., see Oke, J. B. 

Schmidt, Maarten, Radio galaxies and the ori- 
gin of cosmic rays. Astrophys. J. {Letters), 
149, L39, 1967. 

Schmidt, Maarten, Space distribution and 
luminosity functions of quasi-stellar radio 
sources. Astrophys. J., 151, 393-410, 1968. 

Schmidt, Maarten, see also Bahcall, John N. ; 
Moffet, A. T. 

Searle, Leonard, see Sargent, Wallace L. W. 

Simon, Michal, see Manwell, Tom; Rees, M. J. 

Slater, C. H., see Moffet, A. T. 

Smith, Sara F., and Robert Howard, Magnetic 
classification of active regions. Intern. As- 
tron. Union Symp. No. 35, The Structure and 
Development of Solar Active Regions, pp. 
33^2, K. 0. Kiepenheuer, ed., D. Reidel 
Publishing Co., Dordrecht-Holland, 1968. 

Smith, Sara F., see Bumba, V.; Howard, 
Robert. 

Soru-Iscovici, I., see Bumba V. 

Spiegel, E. A., see Stein, Robert F. 

Stein, Robert F., and E. A. Spiegel, Radiative 
damping of sound waves. J. Acoust. Soc. 
Am., 42, 866-869, 1967. 

Swope, Henrietta H., Progress report on the 
Leo II system. Publ. Astron. Soc. Pacific, 79, 
430-440, 1967. 

Tammann, G. A., and Allan Sandage, The stel- 
lar content and distance of the galaxy NGC 
4203 in the M 81 group. Astrophys. J., 151, 
825-860, 1968. 

Tammann, G. A., see Sandage, Allan. 

Tanenbaum, Andrew W., see Howard, Robert. 

Thackeray, A. D., see Arp, Halton. 

Thompson, A. R., see Moffet, A. T. 

Trimble, Virginia L., see Greenstein, Jesse L. 

Utter, Merwyn G., The heavens in 1968. 
Astron. Soc. Pacific, Ann. Ser., 8 pp., Janu- 
ary 1968. 

Vaughan, Arthur H., Jr., Astronomical inter- 
ference spectroscopy. Ann. Rev. Astron. 



Astrophys., Vol. 5, pp. 139-166, Annual Re- 
views, Inc., Palo Alto, Calif., 1967. 

Vaughan, Arthur H., Jr., Interferometer mea- 
surement of star diameters (a one-hundredth 
anniversary). Astron. Soc. Pacific Leaflet 
No. 458, 8 pp., August 1967. 

Vaughan, Arthur H., Jr., and Harold Zirin, 
The helium line X10830 A in late-type stars. 
Astrophys. J., 152, 123-139, 1968. 

Visvanathan, N., Interstellar polarization. As- 
tron. Soc. Pacific Leaflet No. 4^3, 8 pp., 
January 1968. 

Visvanathan, N., Polarization observations in 
the radio galaxy 3C 371 and X-ray source 
Cyg X-2. Astrophys. J. {Letters), 150, L149- 
L152, 1967. 

Visvanathan, N., and J. B. Oke, Non-thermal 
compound in the continuum of NGC 1068. 
Astrophys. J., 152, L165-L168, 1968. 

Warner, B., see Lambert, David L. 

Weedman, D. W., see Arp, Halton. 

Weistrop, Donna, Equivalent widths of Ha in 
late-type stars. Publ. Astron. Soc. Pacific, 
79, 546-550, 1967. 

Westphal, J. A., see Kristian, Jerome. 

Wilcox, John M., and Robert Howard, Per- 
sistent solar magnetic pattern extending over 
equatorial latitudes. Phys. Rev. Letters, 20, 
1252-1254, 1968. 

Wilcox, John M., see ako Howard, Robert. 

Wildey, Robert, see Sandage, Allan. 

Wilson, Olin C, Radial velocities of dK and 
dM stars. Astron. J., 72, 905-912, 1967. 

Wilson, Olin C, Stellar chromospheres. Ency- 
clopedia oj Atmospheric Sciences and Astro- 
geology, pp. 947-949, Rhodes W. Fairbridge, 
ed., Reinhold Publishing Corp., New York, 
1967. 

Woerden, Hugo van, Structure and motions 
in the interstellar medium, in Intern. Astron. 
Union Symp. No. 31, Radio Astronomy and 
the Galactic System, pp. 3-30, H. van Woer- 
den, ed., Academic Press, London, 1987. 

Zirin, Harold, The excitation of coronal emis- 
sion lines as a measure of electron density. 
Astrophys. J., 151, 383-388, 1968. 

Zirin, Harold, He 3 in several magnetic stars. 
Astrophys. J. {Letters), 152, L177-L178, 1968. 

Zirin, Harold, The solar atmosphere, in The 
Earth in Space, pp. 25-32, Hugh Odishaw, 
ed., Basic Books, New York, 1967; con- 
densed version, Astron. Soc. Pacific Leaflet 
No. 466, 8 pp., April 1968. 

Zirin, Harold, The solar atmosphere, in Proc. 
Intern. School of Physics, Enrico Fermi, 
Course 39, Plasma Astrophysics, pp. 124-141, 
P. A. Sturrock, ed., Academic Press, New 
York and London, 1967. 



68 



CARNEGIE INSTITUTION 



Zirin, Harold, see also Norton, Robert H.; 

Vaughan, Arthur H., Jr. 
Zwicky, Fritz, Compact and dispersed cosmic 

matter, Part 1, in Advances in Astronomy 

and Astrophysics, Vol. 5, pp. 267-343, Z. 

Kopal, ed., Academic Press, New York, 1967. 
Zwicky, Fritz, Neutron-star-studded compact 

galaxies. Astrophysics, Acad. Sci. Armenian 

SSR, 3, 519-524, 1967. 
Zwicky, Fritz, Blue compact galaxies in the 

Leo A and B clusters, op. cit., 525-528. 
Zwicky, Fritz, Decouverte des deux premiers 

amas formes par des galaxies compactes. C. 

R. Acad. Sci. Paris, 266, 103-105, 1968. 
Zwicky, Fritz, Systematische Entdeckung von 

Kosmischen Objekten. Die Sterne, 43, 105- 

112, 1967. 
Zwicky, Fritz, Discovery of two supernovae. 

Intern. Astron. Union Circular No. 2070, 

April 29, 1968. 



Zwicky, Fritz, Morphology of the instability of 
galaxies, in Intern. Astron. Union Symp. 
No. 29, Instability oj Phenomena in Galaxies, 
pp. 46-70, M. Arakeljan, ed., Mezdunarodnaja 
Kniga, Moscow, 1967. 

Zwicky, Fritz, List of compact galaxies and 
compact parts of galaxies, eruptive galaxies 
and post-eruptive galaxies, Sixth and 
Seventh, June 1968, California Institute of 
Technology, Pasadena, Calif. 

Zwicky, Fritz, The 1966 Palomar supernova 
search. Publ. Astron. Soc. Pacific, 79, 456- 
459, 1967. 

Zwicky, Fritz, and R. Barbon, Supernova 
1962a of type I in the Coma cluster of 
galaxies. Astron. J., 72, 1366-1367, 1967. 

Zwicky, Fritz, Maria Karpowicz, and K. Rud- 
nicki, Photographic observation of 3C 273. 
Z. Astrophys., 67, 243-245, 1967. 

Zwicky, Fritz, see also Gates, Howard S. 



STAFF AND ORGANIZATION 



Professor Jan H. Oort, Director of the 
University Observatory, Leiden, The 
Netherlands, held an appointment as 
Research Associate during the year, and 
visited the Observatories during the 
months of December, January, and 
February. 

Dr. Leonard T. Searle became a Staff 
Member of the Observatories on June 21, 
1968. Recently on the staff of the Mount 
Stromlo Observatory, Australian Na- 
tional Observatory, Dr. Searle is well 
known here, having held an appointment 
as research fellow at the Mount Wilson 
and Palomar Observatories from 1960 
to 1963. 

Miss Henrietta Swope, research fellow 
at the Observatories, retired after sixteen 
years during which she made important 
contributions to the study of variable 
stars in dwarf galaxies of the local group 
as well as in the Andromeda Galaxy, 
M 31. Her work, initiated in collabora- 
tion with the late Walter Baade, resulted 
in the most precise calibration of the 
distance to external galaxies. 

Dr. Fritz Zwicky, a Staff Member of 
the Observatories since 1948, retired on 
June 30, 1968. His prolific researches, 
characterized by an independent ap- 



proach and conducted with great energy, 
covered a wide range in the fields of cos- 
mology and astrophysics. He organized 
large-scale collaborative programs for 
the discovery of supernovae and for the 
cataloging of galaxies. 

Research Division 

Distinguished Service Member, Carnegie In- 
stitution of Washington 

Ira S. Bowen 
Research Associate 

Jan H. Oort 
Staff Members 

Halton C. Arp 

Horace W. Babcock, Director 

Edwin W. Dennison 

Armin J. Deutsch 

Jesse L. Greenstein x 

Robert F. Howard 

Robert P. Kraft 2 

Robert B. Leighton 3 

Guido Munch 4 

J. Beverley Oke * 

Bruce H. Rule, Chief Engineer 

Allan R. Sandage 

Wallace L. W. Sargent 5 

Leonard T. Searle 

Maarten Schmidt 4 



MOTJNT WILSON AND PALOMAR OBSERVATORIES 



69 



Arthur H. Vaughan, Jr. 
Olin C. Wilson 
Harold Zirin 6 
Fritz Zwicky 7 

Staff Member Engaged in Post-Retirement 
Studies 

Alfred H. Joy 

Staff Associates 

Bruce C. Murray 8 
Gerry Neugebauer 9 
James A. Westphal 10 

Senior Research Fellow 

Jun Jugaku 11 

Carnegie Fellows 

Robert Dickens 12 
Jerome Kristian 
Wojciech Krzeminski 13 
John V. Peach 14 
Rene Racine 
David M. Rust 15 
Natarajan Visvanathan 

Research Council of Canada Fellowship 

F. David A. Hartwick 16 
NATO Fellow 

Alessandro Braccesi 17 
Research Fellows 

Ann Merchant Boesgaard 18 
J. David Bohlin 
J. W. R. Heintze 2 
Ardon R. Hyland 
William H. Julian 
Keiichi Kodaira 
Dora Russo Lackner 19 
David L. Lambert 
Mitsugu Makita 20 
Martin J. Rees 21 
M. T. Scholz 
Rudolf E. Schild 
Michal Simon 
Robert Stein 22 
Henrietta H. Swope 7 
Takashi Tsuji 

Senior Research Assistant 

Dorothy D. Locanthi 

Research Assistants 

Frank J. Brueckel 
Sylvia Burd 



Thomas A. Cragg 
Robert F. Garrison 16 
Howard Gates 16 
Emil Herzog 16 
Basil Katem 
Margaret Katz 
Charles T. Kowal 
A. Louise Lowen 
Kathleen Reynolds 23 
Malcolm S. Riley 
Annilla Sargent 
Merwyn G. Utter 
Grace D. Vess 

Student Observers 

Saul J. Adelman 
Christopher M. Anderson 
Kurt S. Anderson 
Dennis Baker 
Eric Becklin 
Judith Cohen 
Edward J. Groth 
Theodore Hilgeman 
Thomas B. McCord 
Dennis Matson 
Robert O'Connell 
Patrick S. Osmer 
Arsine V. Peterson 
Bruce A. Peterson 
Virginia L. Trimble 
Donna E. Weistrop 

Photographic Laboratory 

William C. Miller, Photographer 

John A. Difley, Photographic Technician 

Clare Neal, Solar Photographic Assistant 

Librarian 

Marline Gerrity 24 
Charlotte Fournier 

Instrument Design and Construction 

Lawrence E. Blakee, Senior Electronic 

Technician 
Maynard K. Clark, Senior Electronic 

Engineering Assistant 
Floyd E. Day, Head Optician 
Raymond Dreiling, Machinist 
Loyal Elam, Jr., Draftsman 25 
Robert D. Georgen, Machinist 
Fred Idzinga, Electronics Specialist 
Melvin W. Johnson, Optician 
Wilfred H. Leckie, Draftsman 
Ernest 0. Lorenz, Engineering Assistant 
Richard Lucinio, Engineering Assistant 
Frederick O'Neill. Machinist 



70 



CARNEGIE INSTITUTION 



Michael Morrill, Technical Aide 28 
Gerald Preston, Technical Aide 
John D. Raphael, Electronics Specialist 
Rudolf E. Ribbens, Designer and Super- 
intendent of Instrument Shop 
Howard G. Sachs, Engineer 
Benny W. Smith, Electronics Specialist 
Robert G. Stiles, Optician 
David Thompson, Technical Assistant 
Virgal Z. Vaughan, Electronics Specialist 
Ralph W. Wilson, Machinist 
Madeleine Williams, Draftsman 
Felice Woodworth, Draftsman-Illustrator 

Maintenance and Operation 

Mount Wilson Observatory and Offices 

Wilma J. Berkebile, Secretary 
Fern V. Borgen, Stenographer and Recep- 
tionist 
Clyde B. Bornhurst, Mechanic 
Hugh T. Couch, Superintendent of Con- 
struction 
Helen S. Czaplicki, Typist-Editor 
Hazel Fulton, Stewardess 
Eugene L. Hancock, Night Assistant 
Mark D. Henderson, Custodian 7 
Anne Hopper, Accountant 
Rienaldo Jacques, Night Assistant 
Ethel Marzalek, Stewardess 
Frances Maynor, Stewardess 
Alfred H. Olmstead, Laborer 
Glen Sanger, Driver 
William D. St. John, Custodian 
Henry F. Schaefer, Night Assistant 
Elizabeth M. Shuey, Secretary 
Benjamin B. Traxler, Mountain Superin- 
tendent 
Fredrick P. Woodson, Assistant to the 
Director 

1 Professor of Astrophysics and Executive 
Officer for Astronomy, California Institute of 
Technology. 

2 Resigned July 31, 1968. 

s Professor of Physics, California Institute 
of Technology. 

4 Professor of Astronomy, California Institute 
of Technology. 

5 Associate Professor of Astronomy, Cali- 
fornia Institute of Technology. 

"Professor of Astrophysics, California Insti- 
tute of Technology. 

7 Retired June 30, 1968. 

8 Professor of Planetary Science, California 
Institute of Technology. 

"Associate Professor of Physics, California 
Institute of Technology. 

"Senior Research Fellow in Planetary Sci- 
ence, California Institute of Technology. 



Palomar Observatory and Robinson Labora- 
tory 

Fred Anderson, Machinist 27 

Ray L. Ballard, Administrative Assistant 

Jan A. Bruinsma, Custodian 

Maria J. Bruinsma, Stewardess 

Eleanor G. Ellison, Secretary and Librarian 

Beulah Greenlee, Stewardess 

Frank K. Greenlee, Custodian 

Daniel J. Hargraves, Mechanic and Relief 
Night Assistant 

Liselotte M. Hauck, Secretary 

Victor A. Hett, Night Assistant 

Helen D. Holloway, Secretary 

Charles E. Kearns, Assistant Superin- 
tendent 

J. Luz Lara, Mechanic 

Carl D. Palm, Night Assistant 

Catherine T. Paul, Secretary 

Marilynne Rice, Secretary 

Carol Russell, Secretary 28 

Robert T. Snow, Mechanic and Relief 
Night Assistant 

Barrett A. Staples, Mechanic 29 

Gary M. Tuton, Night Assistant 

William C. Van Hook, Mountain Super- 
intendent 

Betty A. Wallace, Secretary 

Ardith Walthers, Secretary 

Warren L. Weaver, Mechanic 

Site-Testing Operations, Chile 

Donald L. Buck, Project Supervisor 

Manuel Blanco, Laborer 

Manuel Casanova, Observer and Utility 

Employee 16 
Roberto Ramos, Laborer 
Tomas Veliz, Trainee Observer 16 
Manfred Wagner, Camp Chief 



Resigned 

1 Resigned 

' Resigned 

Resigned 

' Resigned 

Resigned 

Resigned 

1 Resigned 

1 Resigned 

' Resigned 

Resigned 

Resigned 

Resigned 

Resigned 

Resigned 

Resigned 

Resigned 

1 Resigned 

Resigned 



October 31, 1967. 
September 20, 1967. 
November 30, 1967. 
April 22, 1968. 
February 27, 1968. 
June 30, 1968. 
March 15, 1968. 
August 31, 1967. 
June 13, 1968. 
December 15, 1967. 
May 17, 1968. 
September 1, 1967. 
October 2, 1967. 
May 15, 1968. 
March 1, 1968. 
October 7, 1967. 
February 9, 1968. 
February 2, 1968. 
January 4, 1968. 



Geophysical Laboratory 

Washington, District of Columbia 



Philip H. Abelson 
Director 



Carnegie Institution Year Book 67, 1967-1968 



Contents 



Introduction 75 

Silicate Mineralogy and Phase-Equilibria 

Studies 80 

Pyroxenes 80 

Synthesis and stability of iron-free 
pigeonite in the system MgSiOa- 
CaMgSi 2 Oo at high pressures 
(Kushiro) 80 

Electron-probe study of pyroxene 

exsolution (Boyd and Brown) . 83 

Ortho-clino inversion in ferrosilite 

(Lindsley and Munoz) ... 86 

Subsolidus relations on the join 
hedenbergite-ferrosilite at 20 kb 
(Munoz and Lindsley) ... 88 

Unit-cell parameters of clinopyrox- 
enes along the join hedenbergite- 
ferrosilite (Lindsley, Munoz, and 
Finger) . . 91 

Optical properties of synthetic clino- 
pyroxenes on the join hedenber- 
gite-ferrosilite (Myer and Linds- 
ley) 92 

Oxidation of Ca-rich clinopyroxenes 

(Huckenholz) 94 

Stability of omphacite .... 96 

Stability of omphacite in the ab- 
sence of excess silica (Bell and 

Kalb) ■•.. 97 

Stability of omphacite in the pres- 
ence of excess silica (Kushiro) . 98 

The temperature-compression melt- 
ing relation (Gilbert) .... 100 
Feldspars and highly feldspathic rocks . 101 

The melilite-plagioclase incompati- 
bility dilemma in igneous rocks 
(Yoder and Schairer) .... 101 

The join albite-anorthite-akermanite 

(Schairer and Yoder) .... 104 

Anorthite-akermanite and albite- 
soda melilite reaction relations 
(Yoder) _ . . 105 

Experiments bearing on the origin of 
anorthositic intrusions (Emslie 
and Lindsley) 108 

Syenites (Morse) 112 

Feldspars (Morse) 120 

Annealing characteristics of dense 

feldspar glass (Chao and Bell) . 126 
Diamond inclusions, aluminum sili- 
cates, 0-quartz, andradite . . . 130 

Mineral inclusions in diamonds 

(Meyer and Boyd) .... 130 

The andalusite-sillimanite transition 
and the aluminum silicate triple 
point (Gilbert, Bell, and Rich- 
ardson) 135 

X-ray properties and stability rela- 
tions of /3-quartz solid solutions 
along the join LiAlSi20o-Si0 2 
(Munoz) 137 

Synthesis and stability of Ti-andra- 

dite (Huckenholz) 139 



Silicate systems including a vapor 

phase 153 

Melting of forsterite and enstatite 
at high pressures under hydrous 
conditions (Kushiro and Yoder) . 153 

Liquidus relations in the system for- 
sterite-diopside-silica-H 2 at 20 
kb (Kushiro) 158 

Melting of a hydrous phase : phlogo- 

pite (Yoder and Kushiro) . . 161 

Reconnaissance study of the stability 
of amphiboles at high pressure 
(Gilbert) ....... 167 

Control of fugacities in fluorine- 
bearing hydrothermal systems 
(Munoz) 170 

Sulfide Mineralogy and Phase-Equilibria 

Studies 175 

Ore minerals 175 

The Fe-Se system (Kullerud) . . 175 
The Cu-Zn-S system (Craig and 

Kullerud) 177 

Cubic <^ hexagonal inversions in 
some M 3 S 4 -type sulfides (Kul- 
lerud) 179 

Meteorite minerals 182 

The Cr-S and Fe-Cr-S systems (El 

Goresy and Kullerud) . . . 182 
Sulfide assemblages in the Odessa 
meteorite (Kullerud and El Gor- 
esy) .... 187 

Sulfur-isotope fractionation in minerals 

(Puchelt) 192 

Synthesis of /3Cr f FeS« at 35 kb (El 

Goresy, Bell, and England) . . 197 
Melting relations in the Fe-rich por- 
tion of the system Fe-FeS at 30 kb 
pressure (Brett and Bell) . . . 198 

Biogeochemistry 199 

Reactions of the organic matter in a 

Recent marine sediment (Hoer- 

ing) 199 

Branched-chain fatty acids in Recent 

sediments (Hoering) .... 201 
Fattv alcohols in sedimentary rocks 

(Hoering) 202 

Fichtelite hydrocarbons in fossil wood 

(Hoering) . . 203 

Laboratory simulation of amino-acid 

diagenesis in fossils (Hare and 

Mitterer) ........ 205 

Recent amino acids in the Gunflint 

chert (Abelson and Hare) . . . 208 

Computer Reduction of Electron-Probe 

Data (Boyd, Finger, and Chayes) . 210 

Crystallography 215 

A new hexagonal form of carbon from 
the Ries Crater (El Goresy and 

Donnay) 215 

Determination of cation distributions 
by least-squares refinement of 
single-crystal X-ray data (Finger) . 216 



Magnetic susceptibility and exchange 
coupling in ardennite (Senftle, 
Thorpe, and Donnay) .... 218 
"Mckelveyite," a syntactic intergrowth 
of two phases (Donnay and Don- 
nay) 218 

Crystalline heterogeneity .... 219 
Evidence from electron-probe study 
of Brazilian tourmaline (Don- 
nay) 219 

Evidence from source-image distor- 
tion (Young, Wagner, Pollard, 

and Donnay) 220 

Structure refinement of elbaite (Don- 
nay and Barton) 221 

Improvements of crystallographic 

equipment (Donnay) .... 222 
Sonoraite (Gaines, Donnay, Hey, and 

Zemann) 223 

The crystal structure of sonoraite, Fe 2 - 
Te 2 05(OH) 4 -H 2 (Donnay, Stew- 
art, and Zemann) 223 

Geochronology and Geochemistry (Da- 
vis, Krogh, Hart, Aldrich, Morse, 

and Ishizaka) 224 

Geochronology of the Grenville Prov- 
ince (Krogh and Davis) . . . 224 
Fractionation of potassium and rubi- 
dium in a layered intrusion (Morse 
and Davis) 231 

Petrology 233 

Statistical petrography 233 

Identity of expected interdependence 
between pairs of Niggli numbers 
and pairs of analogous remain- 
ing-space variables (Chayes) . 235 
Transformations designed to elimi- 
nate negative elements from 
sample estimates of the vector 
of open variances (Chayes) . . 236 



Curve-fitting in the ternary diagram 

(Chayes) . . 236 

A last look at Gl-Wl (Chayes) . . 239 

Materials balance in igneous rock 

suites (Bryan) 241 

A least-squares approximation for 
estimating the composition of a 
mixture (Bryan, Finger, and 

Chayes) . 243 

Chemical and optical petrography 

(Bryan) 244 

Volcanic rocks from the Bunya 
Mountains, Queensland, Austra- 
lia 244 

An olivine gabbro-microsyenite in- 
trusion from the Carnarvon 
Range, Queensland, Australia . 247 

Structural Geology 251 

Experiments in flow deformation 

(Scott) 251 

Movement directions and the axial- 
plane fabrics of flexural folds 
(Scott and Hansen) 254 

On the "drag folds" of Van Hise and 

Leith (1911) (Hansen and Scott) . 258 

Slip folds in planes of unsystematic 

orientation (Hansen) .... 263 

Staff Activities 265 

Aluminum Silicate Conference . . . 265 

Journal of Petrology 266 

Lectures 266 

Penologists' Club 267 

Bibliography 268 

References Cited 269 

Personnel 276 



\1L1- 



INTRODUCTION 



A creative enterprise depends for its 
success first on gifted people. Over the 
years the Geophysical Laboratory has 
had its share of them. If a laboratory is 
to be more than just a collection of indi- 
viduals, however, a symbiotic relation- 
ship must exist among the staff. Together 
they must create an atmosphere that is 
both stimulating and sustaining of intel- 
lectual endeavor. This atmosphere is 
likely to occur if certain factors are pres- 
ent and others are suppressed. Positive 
factors include a common allegiance to 
scholarship as one of the highest forms 
of human endeavor. They include mutual 
respect among the staff. They include 
freedom for the individual to choose his 
own program within reasonable but 
broad limits, with the choice based solely 
on intellectual criteria. Negative factors 
that must be suppressed are excessive 
bureaucracy and red tape and the domi- 
nance of criteria for choice of program 
other than purely intellectual. 

For nearly two decades scientific re- 
search in this country was carried out 
under circumstances that permitted great 
creativity. Generous support was widely 
available, with few strings attached. Of 
late, the climate has changed drastically. 
Budget cuts, though substantial, are of 
small consequence in comparison to other 
factors. Government regulations are plac- 
ing on universities and their staffs an 
irritating burden of paper work. More 
serious is a deterioration in the intel- 
lectual atmosphere of the university. The 
cries of students and of some faculty for 
"relevance" are basically a manifesta- 
tion of anti-intellectualism. 

A further, related negative influence 
on the creative enterprise is the adverse 
effect resulting from a changed attitude 
on the part of government toward the 
value of basic research. Maynard Hutch- 
ins outlined the scope of the problem 
when discussing the huge modern uni- 
versity. 



"The multiversity, which will do for 
the society anything the society will pay 
for, exists to flatter the spirit of the age. 
One trouble with flattering the spirit of 
the age is that all of a sudden it may 
turn and bite you. Something of the sort 
appears to be happening in California. 
The popular desire, which was formerly, 
for reasons never made clear, to have a 
famous multiversity, is now the desire, 
for reasons equally obscure, to have a 
cheap one, with clean-shaven students, 
and relatively few of them. What are you 
to say to people whose immediate needs 
you are striving to meet, and even to an- 
ticipate, when they tell you they've 
changed their minds and do not need you 
anymore?" 

Today there is widespread continuing 
deterioration in the environment for basic 
research in this country. These trends 
make it essential that at least a few 
places be maintained where creative 
scholarship is a way of life. 

At the Geophysical Laboratory we are 
fortunate in functioning under circum- 
stances that lend themselves to scholarly 
effort. We enjoy freedom from red tape 
and flexibility of programs, and we seek 
to maximize the proportion of independ- 
ently creative minds among us. In such 
circumstances men are creative. High- 
lights of some of the accomplishments of 
the past year follow. 

Meyer and Boyd have used the elec- 
tron probe to make the first chemical 
study of olivine, garnet, chromite, and 
diopside inclusions from natural dia- 
monds. Because of the high pressures and 
temperatures required for stable crystal- 
lization of diamond in the earth it is very 
probable that natural diamonds re- 
covered from kimberlite have come from 
depths greater than 100 km. The minerals 
found as inclusions are similar in a gross 
way to those from the ultramafic nodules 
in kimberlite but there are a number of 
remarkable differences. It has been found 



75 



76 



CARNEGIE INSTITUTION 



that garnet and chromite inclusions are 
unusually rich in chromium, whereas a 
diopside inclusion contains unusually 
little chromium. An important feature 
not yet explained is the monomineralic 
nature of the inclusions, no polymineralic 
inclusion having yet been found. 

Exsolution features are especially well 
developed in pyroxenes from large 
layered intrusives, and their study pro- 
vides insight into the solid-state response 
of these minerals to conditions of slow 
cooling. Boyd and Brown have analyzed 
both host and exsolved lamella in a grain 
of inverted pigeonite and in a coexisting 
grain of augite from the Bushveld gab- 
bro. Their analyses show that the minor 
elements Al, Cr, Mn, Ti, and Na have 
fractionated in the exsolution process. 
The composition of an augite lamella in 
inverted pigeonite is almost identical 
with that of the separate augite host, and 
the composition of a clinohypersthene 
lamella in augite is very similar to that 
of the inverted pigeonite host. These re- 
lations indicate that equilibrium has been 
maintained for a considerable period 
during cooling below the solidus. 

The phase relations of Ca-Fe pyrox- 
enes have continued to be of major inter- 
est to Lindsley. With Munoz he has 
investigated the subsolidus relations of 
pyroxenes on the join hedenbergite-fer- 
rosilite at 20 kb. Finger joined them in 
determining the unit-cell parameters of 
these pyroxenes, and Myer determined 
their optical properties. The results of 
this study provide information concern- 
ing the subsolidus phase relations in the 
pyroxene quadrilateral and offer valuable 
insights concerning the crystallization 
history of layered intrusive igneous 
rocks. Munoz also made a determination 
of the unit-cell parameters and stability 
relations of /?-quartz solid solutions on 
the join LiAlSi 2 6 -Si0 2 at high pressures. 

Huckenholz carried out a detailed 
study of the join andradite-Ti-garnet 
and its relationships in the join wol- 
lastonite-perovskite-hematite. He con- 
cluded that the appearance of Ti-bearing 



garnets and ferri-diopsides is a natural 
consequence of crystallization at or near 
atmospheric conditions and is not neces- 
sarily the result of special deep-seated 
processes. Further experiments on natu- 
ral clinopyroxenes show that the asso- 
ciation of andraditic garnets and ferri- 
diopsides is favored in magmas crystal- 
lizing under oxidizing conditions. 

The pyroxene omphacite (NaAlSi 2 6 - 
CaMgSi 2 O ) is probably formed under 
high pressures, occurring at great depths 
in the earth. Bell, Kalb, and Kushiro 
have been studying the stability of 
omphacite, with particular attention to 
the effect of exchanging diopside for 
jadeite in the solid solution. 

Kushiro and Yoder studied the melt- 
ing relations for MgSi0 3 and Mg 2 Si0 4 
compositions under hydrous conditions 
to determine the effect of water on the 
genesis of basalt magmas. They found 
that enstatite melts incongruently to 
forsterite and liquid in the presence of 
water at pressures up to at least 30 kb, 
compared with about 5 kb under an- 
hydrous conditions. Kushiro studied the 
melting relations in the system forsterite- 
diopside-silica-water at high pressures 
and found that the field of silica relative 
to those of diopside and enstatite solid 
solutions is greatly reduced at 20 kb 
water pressure. The results suggest that 
silica-rich magmas such as andesitic and 
dacitic magmas can be formed by frac- 
tional crystallization of basaltic magmas 
at high water pressures. The process 
may be related to the origin of calcalkali 
rocks in the orogenic zones. 

Kushiro reinvestigated the system 
diopside-enstatite at 20 kb under an- 
hydrous conditions and found a narrow 
field of pigeonitic clinopyroxene in the 
subsolidus region of this system, results 
significant for an understanding of the 
crystallization of pigeonite-bearing rocks. 
The synthesis of "iron-free pigeonite" 
was made in this and adjoining fields. 

Plagioclase and melilite do not coexist 
in igneous rocks, yet they crystallize to- 
gether in the laboratory. This dilemma 



GEOPHYSICAL LABORATORY 



77 



has been the focus of an intensive effort 
by Yoder and Schairer. The association 
of plagioclase and melilite persists up to 
relatively moderate pressures (Yoder), 
in the presence of water (Yoder and 
Schairer), and over a wide range of 
compositions (Schairer and Yoder). The 
presence of ferric iron and titanium may 
greatly restrict the association but other 
factors, not yet discovered, are being 
sought to account for the incompatibility 
of these two minerals in natural lavas. 

Efforts by Yoder and Kushiro to in- 
vestigate the system involving the prin- 
cipal minerals of kimberlite, forsterite- 
calcite-phlogopite, resulted in many diffi- 
culties in the interpretation of the 
experimental products. Two of the prin- 
cipal minerals, forsterite and phlogopite, 
were successfully studied individually 
in the presence of water. Liquids with 
which forsterite and phlogopite are 
stable to depths of about 60 km are rich 
in alkalies and lie outside the range of 
compositions of the principal minerals 
of kimberlite. At greater depths, where 
phlogopite may melt congruently, the 
existence of a kimberlite magma may 
be possible. 

Field deductions suggest that natural 
magmas are water-undersaturated under 
the conditions of generation. Yoder and 
Kushiro have demonstrated with the 
hydrous mineral phlogopite that the 
undersaturation exists in the partial 
melting of mantle materials consisting of 
an assemblage of hydrous and anhydrous 
phases in the absence of a free gas phase. 
The new data help to resolve the ap- 
parent discrepancy between the large 
amounts of water soluble in silicate 
liquids in the laboratory and the small 
amounts of water believed to be con- 
tained in natural magmas. 

Schairer and Yoder have been study- 
ing phase-equilibrium relations in simpli- 
fied systems related to the origin of 
basaltic and alkaline rocks. In the join 
albite-anorthite-akermanite, they have 
found the presence of four crystalline 
phases with a liquid over a considerable 



range of temperature, indicating the 
complex nature of the solid solution. It 
is evident that, not a quaternary, but 
at least a quinary equilibrium is in- 
volved. 

The manner in which H 2 is contained 
in the mantle plays a role in theories of 
the genesis of magmas. Amphiboles, be- 
cause they are hydrous compounds stable 
at temperatures up to 1000 °C at low 
pressure, have been proposed as phases 
that could hold water at depth. In a 
reconnaissance study of the pressure 
effect on amphibole stability, Gilbert 
has shown that calcic amphiboles are 
not likely to be stable at depths greater 
than 70 to 100 km, thus placing one 
more constraint on the phase assem- 
blages present in the upper mantle. 

Petrogenetic grids for metamorphic 
rocks rely heavily on the polymorphs 
of Al 2 Si0 5 — kyanite, andalusite, and 
sillimanite — for their pressure-tempera- 
ture coordinates. Gilbert, Bell, and 
Richardson have extended the work re- 
ported last year to a study of the an- 
dalusite-sillimanite transition. During 
the andalusite-sillimanite grade of meta- 
morphism, temperatures were in the 
range 625° to 850°C. Determination of 
this transition has also made possible a 
better definition of the aluminum silicate 
triple point. It is now clear that pres- 
sures during kyanite-sillimanite meta- 
morphism must have been above 5 kb. 

When meteorites strike the earth at 
high velocity, a high-pressure shock 
wave results. One of the effects observed 
in impact zones on the surface of the 
earth is the conversion of minerals from 
the crystalline state to an extraordinarily 
dense glass. Bell and Chao have experi- 
mentally investigated for the first time 
the formation and annealing of synthetic 
dense glasses of feldspar composition. 
An understanding of impact products as 
compared with other products of natural 
explosive events, such as volcanic erup- 
tions, will be useful in the analysis of 
samples from the moon and from other 
planets as they become available. 



78 



CARNEGIE INSTITUTION 



Morse has shown that the alkali feld- 
spars are related by a "eutectic," rather 
than a "minimum," at Ph 2 o = 5 kb, and 
that new positions of the solvus limbs 
were required. These positions were lo- 
cated by a bracketing technique. Morse 
expanded the studies into the nepheline 
syenite portion of petrogeny's residua 
system, demonstrating the persistence of 
the feldspar join as a thermal barrier 
between nepheline syenites and granites, 
alone and with diopside, up to P H o = 
10 kb. 

Emslie and Lindsley continued their 
investigations into the origin of anortho- 
site, studying the high-pressure melting 
relations of compositions in the join 
forsterite-albite-anorthite and of the 
chilled margin of the Michikamau 
anorthosite complex. 

Hoering has constructed and put into 
operation a mass spectrometer for the 
analysis of the molecular structure of 
organic compounds found in sedimentary 
rocks. A combination of gas-liquid 
chromatography by mass spectrometry 
gives unique information on molecular 
structure with microgram quantities of 
pure samples. Using this tool Hoering 
has identified a large number of fatty 
acids occurring in a Recent marine sedi- 
ment. Surprisingly, he found a relatively 
large amount of highly branched acids. 
Mass spectrometry has been used to 
identify the homologous series of por- 
phyrins generated by mild thermal treat- 
ment of a Recent marine sediment. The 
distribution of molecular weights ob- 
tained this way is very similar to that 
found in material separated from ancient 
sediments. 

In the laboratory Hare and Mitterer 
have simulated processes requiring mil- 
lions of years in fossils. The incubation 
of shells at temperatures from 185° to 
90 °C for 1 day to 3 months produces 
changes in the amino content almost 
identical with those seen in fossils. A 
partial conversion of isoleucine to allo- 
isoleucine is of particular potential use- 
fulness — in dating when the temperature 



is known or as a thermometer when the 
age of the specimens is known accurately. 

Abelson and Hare have examined 
amino acids present in the 1900-million- 
year-old Gunflint chert. A number of 
lines of evidence converge toward the 
conclusion that these amino acids are of 
recent origin. 

Using the methods of dating rocks 
based on the accumulation of radiogenic 
daughter isotopes and the accompanying 
changes in the ratio of radiogenic to non- 
radiogenic isotopes, Krogh and Davis 
have been able to prove the widespread 
occurrence of rocks and sediments as old 
as 1800 m.y. in the lOOO-m.y.-old Gren- 
ville province of the Canadian shield. It 
has been possible to set limits in time 
for periods of regional metamorphism 
and deformation in areas that later were 
remetamorphosed. The interface between 
the old Superior rocks and the younger 
rocks to the south, the Grenville Front, 
has been an active zone for the past 
1800 m.y. 

Kullerud has shown that spinel-type 
sulfide minerals containing metal to 
sulfur in the 3:4 atomic ratio all have 
high-pressure polymorphs. 

Applications of pertinent phase dia- 
grams, produced by experimental studies 
on synthetic systems, have shown that 
lack of equilibrium between minerals in 
meteorites is not a rare occurrence and 
that equilibrium between assemblages, 
often only a few millimeters apart, is 
usually not achieved. Kullerud and El 
Goresy have shown that some of the 
nonequilibrium assemblages found in 
meteorites may be explained as effects of 
shock. 

The mineral daubreelite is very com- 
mon in iron meteorites, enstatite chon- 
drites, and achondrites. For this reason 
the systems Cr-S and Cr-Fe-S have been 
studied over a large temperature range 
by El Goresy and Kullerud, who have 
shown that daubreelite occurs in two 
polymorphic forms and that some sulfides 
in the Cr-S system occur as minerals. 

Brett and Bell have examined the 



GEOPHYSICAL LABORATORY 



79 



system Fe-FeS to determine what effect 
the addition of sulfur to iron would have 
on the melting point of iron at high 
pressure. The effect at 30 kb is several 
hundred degrees, suggesting that even a 
small amount of sulfur present in the 
earth's core would lower the melting 
temperature. 

Puchelt has studied isotope effects in 
inorganic reactions of sulfides. He has 
shown that atoms in sulfur vapor readily 
exchange with sulfur in a number of 
ore-forming sulfides and has applied this 
result to sulfide minerals from the 
Bodenmais, Germany, ore deposit to 
show how the original heterogeneous 
distribution of sulfur-isotope ratios has 
been altered during the metamorphism 
of the original sedimentary ore deposit. 

El Goresy has discovered a new allo- 
tropic form of carbon, which he and 
Donnay have studied by X rays. Three 
other forms are known — graphite, dia- 
mond, and lonsdaleite. The new hexa- 
gonal phase of carbon is present in 
polished sections of shocked graphite 
gneisses from the Ries Crater in Ger- 
many. So far, the mineral (to be called 
chaoite) has been found only in powdered 
form. The unit cell is large (a = 8.948, 
c = 14.978 A) ; a single crystal is needed 
for the structure determination. Chaoite 
is slightly harder than graphite. 

In another study El Goresy, Bell, and 
England have synthesized /3Cr 2 FeS 4 at 
high pressures. This new phase may be 
stable in meteorites at the time of their 
formation and later invert to the alpha 
form, which is commonly observed. 

Two other new minerals were studied 
by Donnay. Sonoraite was found in the 
field by R. V. Gaines. Its chemical for- 
mula, fully established only with the help 
of the crystal-structure determination, is 
Fe 2 3+ Te 2 4+ 5 (OH) 4 -H 2 0. The structure 
was solved by Donnay in cooperation 
with Stewart and Zemann. It consists of 
pairs of edge-sharing oxygen-hydroxyl 
octahedra about Fe 3+ , Fe 2 OH 4 5 groups 
that share an OH corner. The Fe-Fe dis- 
tance within a pair is only about 3.1 A. 



Additional cross-linking of these pairs is 
performed by oxygen triangles bonded to 
tellurium. 

Ewaldite, a complex carbonate occur- 
ring in three-dimensional intergrowth 
with mckelveyite from the Green River 
formation, was discovered by Donnay 
by single-crystal X-ray diffraction study. 

Finger worked on theoretical and com- 
putational aspects of the refinement of 
crystal structures of minerals where the 
average occupancy of a site is to be de- 
termined by means of least-squares 
analyses of X-ray data. 

From a detailed examination of pub- 
lished analyses of two silicate reference 
materials, Gl and Wl, Chayes con- 
cludes that interlaboratory biases must 
account for a very considerable propor- 
tion of the observed dispersion. Sampling 
variance and random error are con- 
founded in the experimental design, and 
no separate estimate of either can be 
obtained from the data. In 50 paired 
analyses, i.e., analyses of both materials 
by the same analyst, however, correla- 
tions between results for the same con- 
stituents are consistently far too strong 
to be dismissed as random. 

In his continuing study of petrographic 
correlation, Chayes has found a simple 
linearizing transformation for use with 
ternary closed data and has been able 
to show that expected correlations be- 
tween what he terms "remaining-space" 
variables (see Year Book 66) are in fact 
identical with those calculated for analo- 
gous pairs of the widely used Niggli 
numbers. Extending preliminary work 
described last year, Chayes has found 
a reasonable linear transformation that 
yields open variance vectors free of 
negative elements for all arrays so far 
examined; he points out that in such 
instances it is both impossible and un- 
necessary to recast the (transformed) 
data either as remaining-space variables 
or as Niggli numbers. 

Many basalts from eastern Australia 
project in the normative composition 
field Ol'-Di'-Hy', resembling in this re- 



80 



CARNEGIE INSTITUTION 



spect basalts of the deep ocean ridges, 
as previously discussed by Chayes (Year 
Book 64). Bryan's study of the petro- 
graphic and chemical details of these 
basalts and of associated andesitic rocks 
has provided further insight into genetic 
relations and problems of classification. 
In a related study of chemically equiva- 
lent intrusive gabbro intimately asso- 
ciated with a microsyenite ring-dike, he 
has shown that crystal fractionation of 
a parent liquid lying close to the norma- 
tive Ol'-Di' join may lead ultimately to 
an oversaturated residual liquid enriched 
in potash relative to soda. 

Genetic hypotheses involving crystal- 
melt equilibria, assimilation, or mixing 
of two or more magmas, have been fre- 
quently proposed. Bryan has now written 
a computer program which generates an 
estimate of the inferred parent liquid 
or of the inferred residual liquid at any 
arbitrarily chosen stage of magmatic 
evolution. This program makes it pos- 
sible to test rapidly many alternative 
hypotheses, using any reasonable num- 
ber of components simultaneously, or to 
test the ability of an inferred mechanism 
to simulate observed variation trends. A 
direct numerical analogue of current 
graphical procedures, his method shares 
with them the restriction that the ob- 
server is required to select from among a 
number of possible solutions, each based 
on only part of the data. In a late and 
not yet extensively tested development, 
Bryan, Finger, and Chayes have formu- 



lated the problem as one in least-squares 
approximation, a procedure which of 
course yields only one solution for a 
given set of data, and should also lead 
to more satisfactory hypothesis testing. 
During the report year, Hansen has 
spent much of his time writing a book 
called Strain Fades, which presents new 
approaches and concepts in the descrip- 
tion and analysis of folds in rocks. He 
has continued testing and calibrating the 
10 kb-1000°C gas apparatus for rock 
deformation, in part with the collabora- 
tion of J. L. England. During the final 
months of the current year, Scott com- 
pleted construction of a flow-deforma- 
tion apparatus for the study of folding 
induced in low-viscosity fluids. Initial 
experiments on finely layered stitching 
wax in Vaseline, within a flow environ- 
ment that involved constriction of the 
layers, as well as simple shear parallel 
with the layers (compound and velocity 
gradient flow), yielded sets of asym- 
metric flexural-slip folds with planar dis- 
tributions of fold axes. The relationships 
between the imposed flow environment 
and the induced fabric of the folds con- 
firmed part of the theory of the "separa- 
tion angle," which is used to deduce the 
movement directions in which natural 
folds develop within the earth. Both 
Scott and Hansen have continued the 
structural analysis of natural folds and 
report relationships between axial-plane 
fabrics, height-width ratios, and the 
movement histories of folded rocks. 



SILICATE MINERALOGY AND PHASE-EQUILIBRIA 

STUDIES 



PYROXENES 
Synthesis and Stability of Iron-Free 

PlGEONITE IN THE SYSTEM MgSi0 3 - 

CaMgSi 2 6 at High Pressures 

/. Kushiro 

Pigeonite occurs in tholeiitic basalts, 
dolerites, and andesites. Many tholeiitic 



gabbros also contain hypersthene with 
augite lamellae, now often interpreted 
as an inversion product of pigeonite 
due to slow cooling and referred to 
below as "inverted pigeonite." The 
stability fields of pigeonite and ferro- 
pigeonite have been considered to be at 
higher temperatures than that of ortho- 
pyroxene of the same Mg/Fe 2+ ratio. 



GEOPHYSICAL LABORATORY 



81 



This conclusion is based on their natural 
occurrence and the clinopyroxene-ortho- 
pyroxene relations in the system MgSi0 3 - 
FeSi0 3 at 1 atm, given by Bo wen and 
Schairer (1935), with the assumption 
that the presence of a small amount of 
Ca does not essentially change the clino- 
pyroxene-orthopyroxene stability rela- 
tions. Foster (1951) and Atlas (1952) 
have shown that clinoenstatite is not 
stable but protoenstatite is stable at 
high temperatures (>985°±10°C) for 
MgSi0 3 composition. Boyd and Schairer 
(1964) suggested that clinoenstatite is 
formed from protoenstatite during 
quenching, that protoenstatite has a sta- 
bility field only in the Mg-rich region, 
and that pigeonite may have its own 
stability field above the orthopyroxene 
stability field in the relatively iron-rich 
region of the join MgSi0 3 -FeSi0 3 . Kuno 
(1966) has suggested similar stability 
relations. Yoder, Tilley, and Schairer 
(Year Book 62, pp. 84-95,), on the basis 
of experiments on natural pigeonites, 
suggested two possibilities: pigeonite 
may be an inversion product of proto- 
hypersthene or a quenched stable prod- 
uct. Thus, notwithstanding many in- 
vestigations, the stability field of pi- 
geonite is still controversial, and even 
synthesis of pigeonite has not been 
achieved. 

During the course of studies on the 
join diopside-enstatite at 20 kb, clino- 
pyroxene showing the 231 reflection in its 
X-ray powder pattern was observed in 
the subsolidus region (Kushiro, Year 
Book 63, pp. 104-105). The 231 reflec- 
tion, a strong reflection with h + k odd, 
is not derived from the diopside structure 
(C2/c) but from the clinoenstatite (or 
pigeonite) structure (P2 1 /c). This evi- 
dence suggests that there may be a field 
of pigeonitic clinopyroxene in the sub- 
solidus region of the join diopside-ensta- 
tite. In the present investigation, the 
subsolidus region of this join has been 
restudied more carefully at 20 kb in the 
compositional range from Di 10 En 90 to 
Di 56 En 44 (wt %). The runs were made 



with a piston-cylinder, solid-media pres- 
sure apparatus similar to that designed 
by Boyd and England (1960). Starting 
materials were glass prepared by Boyd 
and Schairer (1964) and glass crystal- 
lized to diopside and clinoenstatite solid 
solutions at 1 atm. Mechanical mixtures 
of pure diopside and pure orthoenstatite 
were also used as starting material. The 
runs lasted % to 1 hour for temperatures 
above 1600° C and % to 3 hours for 
temperatures between 1400° and 1600°C. 
The results are shown in Fig. 1. Most 
of the runs above and some below the 
solidus have been reported previously 
(Kushiro, Year Book 63, pp. 103-105). 

For compositions between diopside and 
Di 53 En 47 , diopside solid solution is stable 
at 1600°C. Between the compositions 
Di 5 oEn 5 o and Di 20 En 80 , however, two dif- 
ferent clinopyroxenes appear to coexist. 
Their X-ray powder diffraction patterns 
show a split in the 311 reflection (Fig. 2) 
and also the 220 reflection, although the 
split of the latter reflection is not so clear 
as that of the former. Other peaks 
change so little with composition that 
they will not show a split for the compo- 
sition range shown here. A mechanical 
mixture of pure diopside (40 wt %) and 
pure orthoenstatite (60 wt %) held at 
1600°C for 1 hour also recrystallized 
into clinopyroxene (s) showing a split 
311 peak without orthoenstatite. The 
clinopyroxene formed from a glass of 
composition Di 40 En 60 at 1600°C showed 
no split in the 311 and 220 reflections; 
when this clinopyroxene was held at 
1620°C for 30 minutes, however, both 
reflections were clearly separated. The 
possible reflections derived from diopside 
solid solution are not in accord with one 
of the split peaks. The relative intensi- 
ties, but not the position, of the two 
separate peaks for the 311 reflection ap- 
pear to change with the bulk composi- 
tion of the starting material (Fig. 2). 
The intensity of the 231 reflection in- 
creases with increase of MgSi0 3 in the 
bulk composition. 

The clinopyroxene (s) crystallized from 



82 



CARNEGIE INSTITUTION 



1800 



1700 



O 



E 
|2 



1600 



1500 



1400 



1300 



Erw+ L 



■ss 



Nil 



X)px+ 



Eric 



J \ ^Cpx a 

i a i ji b a a 

|En ss + i i Cpx + Di ss 
Cpx 1 

l 
I 



Di ss +L 



a a 



a a 



\ 
\ 
\ 

\ E 



B Die 



— !_ B a v., 

D El 



En ss +Di ss 



MgSiO^ 



20 



90 CaMgSi 2 6 



Weight per cent 

Fig. 1. Phase-equilibrium diagram of the system diopside (CaMgSi 2 06)-enstatite (MgSi0 3 ) 
at 20 kb under anhydrous conditions. Cpx, pigeonitic clinopyroxene; Di 88 , diopside solid solu- 
tion; En s8 , orthoenstatite solid solution; L, liquid. 



the composition Di 20 En 80 at 1600 °C 
shows a split 311 peak, one part of which 
is strong and the other weak; whereas 
at 1630°C, a single clinopyroxene, show- 
ing a single 311 peak with a sharp 231 
peak, was obtained from the same com- 
position. The X-ray powder diffraction 
pattern is very similar to those of natural 



Di 20 En 80 



Di 25 En 75 Di 4C En 60 Di 50 En 5 o 




39° 40° 39° 40° 39° 40° 39° 40° 



Fig. 2. X-ray powder diffraction patterns of 
311 reflections (CuKa) of clinopyroxenes crys- 
tallized from mixtures of compositions Di^Enso, 
DisaEnTs, DiioEneo, and Di.-»oEnso (wt %) at 
1600°, 1600°, 1650°, and 1600°C, respectively. 



pigeonites (Fig. 3) . This pigeonitic clino- 
pyroxene may be called "iron-free pi- 
geonite" on the basis of appropriate Ca 
content as well as structure. Lindsley 
and Munoz (this report) synthesized 
"iron pigeonite" on the join hedenbergite- 
ferrosilite. 

In the compositions between Dii 5 En 85 
and Dii En 90 , orthoenstatite and pi- 
geonitic clinopyroxene coexist at tem- 
peratures between 1600 °C and 1650 °C, 
and for the composition Di 7 . 5 En92.5 only 
orthoenstatite was obtained at 1600 °C. 
The field of pigeonitic clinopyroxene is 
limited, therefore, to a region near the 
composition Di 20 En 80 , as shown in Fig. 1. 
At 1400 °C, diopside and orthoenstatite 
solid solutions were obtained from a 
glass (Di 40 En 60 ) and a mechanical mix- 
ture of pure diopside and orthoenstatite 
(Di 36 . 3 En 6 3. 7 ). At 1450°C, diopside solid 
solution and pigeonitic clinopyroxene 
crystallized from a glass of composition 
Di 40 En 60 , and diopside and orthoenstatite 
solid solution crystallized from composi- 



GEOPHYSICAL LABORATORY 

90 



83 




Fig. 3. X-ray powder diffraction pattern (CuKa) of a single pigeonitic clinopyroxene synthesized 
from a mixture of composition Di^Enso at 1630°C and 20 kb. 



tion Di 2 oEn 80 . The field of pigeonitic 
clinopyroxene is limited, therefore, to 
temperatures at least above 1450 °C, as 
shown in Fig. 1. 

The Ca/(Ca + Mg) ratio of the pi- 
geonitic clinopyroxene field, 0.08 to 0.10, 
is very similar to the Ca/(Ca + Mg + 
Fe 2+ ) ratio of natural pigeonites. The 
limited Ca/(Ca+Mg + Fe 2+ ) ratio of 
natural pigeonites may be explained by 
the narrow field suggested by the present 
experiments. On the basis of the present 
experimental results, it is suggested that 
pigeonite has more Ca than clino- 
enstatite, protoenstatite, or orthoenstatite 
(orthopyroxene) and that pigeonite is 
stable only near the solidus temperatures 
and inverts to orthopyroxene with ex- 
solution of diopside solid solution 
(augite) at lower temperatures. These 
suggestions are also consistent with the 
natural evidence. 

The pigeonitic clinopyroxene field 
shown in Fig. 1 would change with pres- 
sure and iron content. It may be expected 
from natural evidence that the tempera- 
ture of the pigeonitic clinopyroxene field 



is lowered with decreasing pressure or 
increasing iron content, or both, and for 
the Fe 2+ /(Mg + Fe 2+ ) ratios and the con- 
ditions of crystallization of natural ba- 
saltic magmas, pigeonites with Fe 2+ / 
(Mg + Fe 2+ ) ratio greater than 0.3 would 
crystallize. It is also likely that at 1 atm 
there is a narrow field of pigeonite near 
the Ca/(Ca + Mg + Fe 2+ ) ratio of 0.10 in 
the system MgSi0 3 -FeSi0 3 -CaSi0 3 . 

Electron-Probe Study of Pyroxene 
Exsoltjtion 

F. R. Boyd and G. M. Brown * 

A wealth of exsolution features in 
pyroxenes and other minerals has been 
opened up for chemical study by the 
advent of the electron probe. These 
features are of interest because they con- 
tain information on the solid-state re- 
sponse of various igneous and meta- 
morphic rocks and ores to conditions of 
slow cooling. Exsolution features in py- 
roxenes are particularly well developed 

* Department of Geology, University of Dur- 
ham, Durham, England. 



84 



CARNEGIE INSTITUTION 



in rocks from large, basic intrusions such 
as the Bushveld, Skaergaard, and Still- 
water. A qualitative survey {Year Book 
66) of exsolved pyroxenes in a variety of 
gabbroic rocks from these intrusions 
showed remarkably regular but intricate 
patterns of exsolution. Coarse exsolu- 
tion lamellae sometimes exsolve fine 
lamellae on further cooling, and the 
patterns are particularly complex in the 
inverted pigeonites where multiple sets 
of lamellae are common. 

Quantitative analyses of lamella-host 
pairs have been undertaken to determine 
whether or not all cations in these py- 
roxenes participate in the exsolution pro- 
cess and to determine whether equi- 
librium is maintained between exsolving 
pyroxene phases as cooling proceeds. 

A gabbro from the Bushveld intrusion 
(SA 1019) was chosen for the quantita- 
tive study. This specimen contains augite 
with fine, even lamellae of clinohyper- 
sthene ranging in width from 1 ^m up to 
about 15 |im. Associated with the augite 
are inverted pigeonite grains with blebs 
of exsolved augite ranging up to several 
hundred micrometers in width. The grain 
of augite and the grain of pigeonite that 
were selected for analysis lie several 
centimeters apart in the same thin 
section. 

Quantitative analyses were made of a 



single clinohypersthene lamella in augite, 
as well as of adjacent areas of the augite 
host. The lamella is doubly terminated 
within the augite crystal and is about 
12 to 15 /xm thick. The original shape 
of the lamella was evidently that of a flat 
lens, and the thin section has cut a 
segment from the edge of this lens. Con- 
sistently low Ca counts were found only 
in the middle portion of the lamella, 
indicating that the beam was penetrating 
into underlying augite toward the ends. 
A scan across the augite grain monitoring 
CaKa X rays is shown in Fig. 4. This 
augite crystal contains many fine lamel- 
lae 1 to 3 /mi thick as well as a more 
coarse set. As can be seen in Fig. 4, there 
are areas of augite host immediately ad- 
jacent to the analyzed clinohypersthene 
lamella that are free of the fine lamellae, 
and the augite host analyses were made 
here. These areas are restricted, and it 
was necessary to repolish the section 
when the analyses were about half com- 
pleted to remove contamination spots 
left by the electron beam. 

The augite lamella in pigeonite and 
the adjacent areas of pigeonite host 
which were selected for analysis are rela- 
tively free of secondary exsolution fea- 
tures. Nevertheless other pigeonite grains 
in this gabbro are exotic examples of the 
development of multiple sets of lamellae. 




10 MICROMETERS 



Fig. 4. Scan across an augite crystal containing exsolution lamellae of clinohypersthene; 
Bushveld gabbro, specimen no. SA 1019. The clinohypersthene analyses were made in the 
vicinity of point A, and the analyses for the augite host were made in the vicinities of points 
B and B'. 



GEOPHYSICAL LABORATORY 



85 



The analyzed augite lamella is an 
elongate bleb, ranging up to about 50 /*m 
in width, and the area available for 
analysis was much greater than needed. 

Analyses were made at 20 kV for Fe, 
Cr, and Mn, and at 15 kV for other ele- 
ments. The specimen current was 0.03 to 
0.04 ;u,a with an X-ray spot size of 2 to 
3 /mi. The principal standard used was a 
glass prepared by H. G. Huckenholz 
containing Ca, Mg, Fe, and Si. Pure Cr 
and Mn were used along with other 
standards described in Year Book 66. 
The data were reduced with the use of 
the computer programs and correction 
procedures described elsewhere in this 
report. 

Results of the analyses given in 
Table 1 show that all the cations, 



except probably silicon, have partici- 
pated in the unmixing process. This is 
true of the minor elements as well as 
Ca, Mg, and Fe. Even Cr, which is pres- 
ent at a concentration level on the order 
of 100 ppm, is fractionated between 
lamellae and host crystals. Ti, Na, Cr, 
and Al are concentrated in the Ca-rich 
pyroxene phases, and Mn is concen- 
trated in the Ca-poor phases. The compo- 
sition of the clinohypersthene lamella in 
augite is similar to the composition of 
the inverted pigeonite host, and the 
composition of the augite lamella in pi- 
geonite is almost identical with the augite 
host. 

These analyses are plotted in the py- 
roxene quadrilateral in Fig. 5 along 
with trend lines for pyroxenes from the 



TABLE 1. Electron-Probe Analyses of Coexisting Pyroxenes from the Bushveld Intrusion 





Inverted Pigeonite 








Augite 




Host 


Lamella 


Host 




Lamella 




(hypersthene) 


(augite) 




(augite) 




(clinohypersthene) 


Si0 2 


53.0 2 


52.4 


2 


51.8 


S 


52.5 6 


Ti0 2 


0.3 


0.5 




0.5 




0.2 


AlaOs 


0.73 1 


1.34 


h 


1.52 


8 


0.85 2 


Cr 2 3 


<0.01 


0.02 




0.02 




<0.01 


FeO* 


23.6 2 


11.2 


i9 


10.8 


h 


25.6 8 


MnO 


0.5 


0.3 




0.3 




0.6 


CaO 


1.36 15 


21.0 


21 


212 


'h 


0.77 8 


MgO 


20.7 2 


13.6 


8 


13.3 


1 


19.6 8 


Na 2 


0.05 


0.2 




0.2 




0.05 


K 2 


<0.005 




<0.01 
100.5 




<0.005 
995 




<0.01 


Totals 


100.3 


99.9 






Atomic Proportions for 6 


Oxygen Atoms 






Si 


1.98 1 


1.96 1 




1.95 ] 




1.98 1 


Ti 


0.008 V 2.000 


0.01 \ 2.000 




0.01 V 2.000 




0.006 y 2.000 


Al 


0.012J 


0.03 J 




0.04 J 




0.014 J 


Al 


0.020' 




0.029 






0.028' 






0.024" 




Fe 


0.740 




0.349 






0.340 






0.812 




Mn 
Ca 


0.02 
0.055 


-1.998 


0.009 
0.842 


- 2.009 




0.01 
0.857 


-2.000 




0.02 
0.031 


-2.001 


Mg 


1.16 


, 


0.760 






0.748 






1.11 




Na 


0.003J 




0.02 J 






0.017. 






0.004. 








Atomic % 








Ca 


2.8 


43.2 




44.1 




1.6 


Mg 


59.2 


38.9 




38.4 




56.8 


Fe 


38.0 


17.9 




17.5 




41.6 



* Total Fe as FeO. _ 

Note : Numbers in italics are for the ratio a I vN where a is the standard deviation and N is the 
mean count. 



86 



CARNEGIE INSTITUTION 



I/2CoMgSi 2 6 



l/2CaFeSi 2 6 



MgSiO 




40 

Atomic per cent 



FeSiO, 



Fig. 5. A plot of the analytical data from Table 1 in the pyroxene quadrilateral. Trend lines 
established by bulk analyses of pyroxene pairs from the Bushveld intrusion (Wager and 
Brown, 1968) are also shown. 



Bushveld intrusion established by wet- 
chemical analyses of bulk pyroxene sepa- 
rates (Wager and Brown, 1968). The 
pyroxenes from SA 1019 have not been 
separated and analyzed by wet chemical 
methods, but to a close approximation 
the bulk compositions of the augite and 
the inverted pigeonite can be taken as 
the points of intersection of the host- 
lamellae tie lines with their respective 
trend lines. A tie line established by 
joining these points of intersection has 
a slope compatible with those established 
by bulk analyses of other augite-inverted 
pigeonite pairs. The exsolution process 
shifts the compositions of the host and 
lamellae phases to more Ca-rich compo- 
sitions on the augite side and to more 
Ca-poor compositions on the pigeonite 
side, as would be expected by slow cool- 
ing through an expanding two-pyroxene 
field. 

The similarity in composition of the 
two Ca-rich phases and to a lesser extent 
the two Ca-poor phases implies that 
equilibrium was maintained for a period 
of time as the cooling proceeded. Values 
for the ratio a-/ Vn suggest that there are 
inhomogeneities in these phases but that 
they are not grossly inhomogeneous. 



Ortho-Clino Inversion in 
Ferrosilite 

D. H. Lindsley and J. L. Munoz 

In the few years since ferrosilite was 
first synthesized at high pressures the 
apparent relations between orthoferro- 
silite (OFs; FeSi0 3 with space group 
Pbca) and clinoferrosilite (ClFs, FeSi0 3 
with space group P2 1 /c) have been a 
source of confusion. Lindsley, Davis, 
and MacGregor (1964) reported that 
they synthesized OFs at higher tempera- 
tures than ClFs, but they did not claim 
that this was an equilibrium relation. At 
the same time, Akimoto, Fujisawa, and 
Katsura (1964) reported the high-pres- 
sure synthesis of two polymorphs of fer- 
rosilite; they identified one as OFs and 
the other as protoferrosilite (which was 
later shown to be ClFs). Lindsley, Mac- 
Gregor, and Davis (Year Book 63, p. 
175) next presented a phase diagram 
in which fields of ClFs and and ferro- 
silite III (Fs III; a pyroxenoid form 
of FeSi0 3 ) occur just below the solidus 
and a field of OFs appears at lower 
temperatures. The diagram was con- 
firmed by reversed reaction between each 
pair of the three polymorphs, and it ap- 
peared that the true equilibrium rela- 
tions had been determined. But some 



GEOPHYSICAL LABORATORY 



87 



anomalies remained. Although both OFs 
and ClFs readily invert to Fs III when 
held for a few hours in the pressure and 
temperature conditions of the reported 
Fs III field, experiments lasting 24 hours 
sometimes yielded ClFs at the same 
pressures and temperatures. Further- 
more, synthesis experiments on oxide 
mixes within the Fs III field occasionally 
quenched to an orthorhombic form, 
which then rapidly inverted to ClFs at 
ambient pressure and temperature. The 
inversion, which could be observed under 
the petrographic microscope immediately 
after quenching, was too rapid to permit 
X-ray identification of the orthorhombic 
material ; but this behavior was strongly 
reminiscent of protoenstatite. It ap- 
peared, then, that the field of Fs III 
given by Lindsley, MacGregor, and 
Davis might actually represent the sta- 
bility field of yet another nonquenchable 
polymorph. 

Spurred by the report of a low-tem- 
perature field for clinoenstatite (Sclar, 
Carrison, and Schwartz, 1964) , Lindsley 
{Year Book 64, pp. 148-149) discovered 
a field of ClFs lying below about 800° C 
over a wide pressure range. A field 
boundary between OFs (above 800° C) 
and ClFs (below 800°C) was confirmed 
by reversed reaction. (Akimoto et al., 
1965, presented a somewhat different 
boundary between these fields, but they 
too reported ClFs as the low-temperature 
polymorph.) It now appeared that there 
were two distinct fields of ClFs, one at 
high temperatures and one at low tem- 
peratures, separated by the field of OFs, 
and with each field boundary confirmed 
by reversed reaction. Although such a 
relation is theoretically possible, Lindsley 
suggested that high-temperature ClFs as 
well as Fs III might form upon cooling 
of a nonquenchable polymorph, perhaps 
protoferrosilite. Experiments at 20 kb on 
the composition Feo. 4 Mgo.6Si0 3 indicated 
that the monoclinic polymorph was the 
stable low-temperature polymorph of 
intermediate Fe-Mg pyroxenes as well 
as of the end members (Year Book 64, 



pp. 149-150). These results seemed in- 
consistent with data from natural pyrox- 
enes, for pigeonites evidently form at 
higher temperatures than hypersthenes. 
Furthermore, hypersthene rather than 
clinohypersthene is the common Ca-poor 
pyroxene in charnockites and pyroxene 
granulites. Kuno (1966) proposed a 
diagram for the Fe-Mg pyroxenes, which 
had a field of high-temperature clinopy- 
roxene — to account for pigeonite — as well 
as a field of low-temperature clinopyrox- 
ene, extrapolated from the high-pressure 
studies, but it was still not clear why so 
few, if any, natural hypersthenes in- 
verted to the low-temperature mono- 
clinic form on cooling. 

In a study of the join hedenbergite- 
ferrosilite {Year Book 65, pp. 230-234), 
Lindsley reported fields of clinopyroxene 
(space group presumed to be C2/c) and 
of a pyroxenoid at 5, 7.5, 10, and 15 kb. 
Preliminary data obtained above 15 kb 
were not published, but the relations be- 
tween Cpx and pyroxenoid for the com- 
positions Fs 95 Wo 5 and Fs 90 Wo 10 were 
grossly consistent with the fields of ClFs 
and Fs III reported for pure FeSi0 3 
(Year Book 63, p. 175), both fields lying 
immediately below the solidus, with ClFs 
occurring at higher pressures. No further 
evidence was found regarding the pos- 
sible existence of a protoferrosilite field 
immediately below the solidus. 

Riecker and Rooney (1967) reported 
the conversion of clinoenstatite from 
rhombic enstatite under shearing condi- 
tions at temperatures up to 1000 °C — 
well within the rhombic enstatite field as 
reported by Sclar, Carrison, and Schwartz 
(1964), and Boyd and England (Year 
Book 64, pp. 117-120). Munoz (Year 
Book 66, pp. 269-370) confirmed the 
findings of Riecker and Rooney. Evi- 
dently shearing stress increases the sta- 
bility field of clinoenstatite. Inasmuch 
as there is inevitably a component of 
shear in dry experiments performed in 
solid-media pressure apparatus, it ap- 
peared that the entire low-clinopyroxene 
field might owe its existence to shearing 



88 



CARNEGIE INSTITUTION 



stress. Munoz sought to test this hy- 
pothesis by experiments on enstatite 
polymorphs, employing a hydrostatic cell 
at 20 kb. No reaction was observed in 
these experiments; clearly, shearing 
stress influences the kinetics of the ortho- 
clino inversion in enstatite, regardless 
of whether it also displaces the equilib- 
rium. Water added to the starting ma- 
terial might increase the reaction rates, 
but it would also have resulted in the 
formation of hydrous phases when added 
to MgSi0 3 . Hydrous ferrous silicates 
form much less readily than do their 
magnesian counterparts, however, and it 
appeared that high-pressure hydrother- 
mal experiments on ferrosilite might re- 
solve the ortho-clino inversion problem. 

We have successfully performed high- 
pressure hydrothermal experiments on 
FeSi0 3 using silver capsules in piston- 
and-cylinder pressure apparatus; water 
was retained in the capsules, no redox re- 
action took place (indicating that the ex- 
periments were effectively closed to H 2 
and H 2 ) , and no obvious hydrous phases 
were formed. The presence of a vapor 
phase at pressure and temperature is 
demonstrated by the leaching of small 
amounts of silica from the charge and 
the subsequent deposition of this silica 
as small globules with low refractive 
index. We converted ClFs to OFs at 20 kb 
and at temperatures as low as 650 °C — 
in the ClFs field and 160° below the OFs- 
ClFs inversion curve {Year Book 64, 
p. 149). In duplicate dry (and therefore 
shearing) experiments, OFs inverted to 
ClFs. 

There are only two logical explana- 
tions for these results: either shearing 
stress stabilizes clinoferrosilite at low 
temperatures; or small, thus far unde- 
tected, amounts of water enter the ortho- 
ferrosilite structure and stabilize it. We 
have seen no difference in the optical 
properties or X-ray diffraction patterns 
of orthoferrosilite synthesized under wet 
and dry conditions, and we suspect that 
the former explanation — stabilization of 
ClFs by shearing stress — is correct. 



However, particularly in view of the 
hydrous magnesium orthosilicates (with 
apparent pyroxene structures) synthe- 
sized at 35 to 65 kb by Sclar, Carrison, 
and Stewart (1968b), we cannot cate- 
gorically reject the second possibility. 

If further work should demonstrate 
that clinopyroxene is stable at low pres- 
sure only under shearing stress, the pres- 
ence of hypersthenes (rather than clino- 
hypersthenes) in many rocks is no longer 
enigmatic. It is still curious that clino- 
hypersthenes are rare in alpine-type 
peridotites, many of which may have 
been intruded essentially in the solid 
state and presumably under conditions of 
shearing stress; evidently much of the 
shearing is confined to restricted zones, 
with the bulk of the rock acting as inert 
blocks. 

Even if the field of low-temperature 
clinopyroxene disappears under condi- 
tions of low shearing stress, the actual 
stability of high-temperature clino- 
pyroxenes is still under question; it is 
not yet clear whether pigeonites are 
stable at high temperatures or whether 
they form upon cooling of some non- 
quenchable polymorph. The solution to 
this problem seems to require X-ray 
diffraction data obtained at high tem- 
peratures (and, for pure FeSi0 3 , at high 
pressure as well) . 

Subsolidus Relations on the Join 
Hedenbergite-Ferrosilite at 20 KB 

J. L. Munoz and D. H. Lindsley 

Previous investigation of the join 
hedenbergite (Cao. 5 Feo.5Si0 3 , Hd) -ferro- 
silite (FeSi0 3 , Fs) at high pressures in- 
dicated the existence of a broad field 
of clinopyroxene (Cpx) at temperatures 
just below the solidus (Year Book 65, 
pp. 230-234). Reaction rates are slow 
in dry experiments along this join at 
temperatures below 1000°C, however, 
and it was impossible to determine 
whether a two-pyroxene field existed 
below that temperature. During the past 
year we have extended our investigation 



GEOPHYSICAL LABORATORY 



89 



of the Hd-Fs join to the range 700° to 
1010°C using hydrothermal experiments 
at 20 kb. We have found a two-pyroxene 
field: Cpx plus OFs ss , orthoferrosilite 
with approximately 5 mole % CaSi0 3 
(Fig.6A). 

Three types of starting materials were 
employed: single-phase clinopyroxenes of 
intermediate compositions, mechanical 
mixtures of Hd and OFs, and mechanical 
mixtures of two clinopyroxenes (Fs 90 Woi 
and Fs 60 Wo 4 o). From 10 to 15 mg of 
appropriate starting material, plus 0.5 
to 2 mg water, and approximately 0.5 mg 
silica glass (to saturate the vapor phase 
with Si0 2 ) was packed into a silver 
capsule, which was then capped with a 
tightly fitting silver lid. The loaded 
capsules were reacted at the desired 
pressure and temperature in piston-and- 
cylinder apparatus for durations rang- 
ing from 16 hours to 27 days. The pres- 
ence of water at the conclusion of each 
experiment in all but one or two charges 
and the absence of magnetite (which 
would have indicated oxidation of the 
charge by preferential escape of hydro- 
gen) show that the mechanically sealed 
capsules were effectively closed to water 
and hydrogen, and thus no external 
redox buffer was required. Compositions 
of OFs ss were determined by electron- 
microprobe analysis; compositions of 
Cpx were determined by X-ray methods 
and confirmed by microprobe analysis. 

Clinopyroxene starting materials that 
broke down to two pyroxenes invariably 
produced discrete grains of the two 
phases and never lamellae of OFs ss in a 
Cpx host. 

The shape of the curve for the compo- 
sition of Cpx in equilibrium with OFs S3 
(Fig. 6A) is similar to that given by 
Davis and Boyd (1966) for Cpx on the 
join diopside-enstatite at 30 kb. In view 
of Kushiro's discovery of a two-Cpx 
field for the same join at 20 kb (discussed 
elsewhere in this report), we have 
searched carefully for a two-Cpx field 
on the join Hd-Fs, such as is shown 



hypothetically in Fig. 6(B), but have 
not found one. 

A clinopyroxene synthesized by us has 
been shown, on the basis of single-crystal 
X-ray photographs, to have the pigeonite 
(P2 x /c) structure (C. W. Burnham, per- 
sonal communication, 1968; the single 
crystal was shown to have the nominal 
composition Fs 85 Wo 15 by its unit-cell 
parameters). Thus, any two-Cpx field 
must lie at compositions more Ca rich 
than Fs 85 Wo 15 . We have homogenized 
mixtures of two clinopyroxenes (Fs 90 Woi 
and Fs 6 oWo 40 ) with bulk compositions 
Fs 75 Wo 25 and Fs 82 Wo 18 (see Fig. 6A). 
The homogenization is indicated by 
powder X-ray studies of the run products 
and confirmed by electron-microprobe 
examination. We conclude that at 1000°C 
and 20 kb there is continuous solid solu- 
tion between Hd and the composition 
Fs 92 Wo 8 . We cannot categorically elim- 
inate a narrow (less than 10 mole % 
wide) two-clinopyroxene field that might 
indicate a first-order transition between 
clinopyroxenes of the pigeonite structure 
(space group P2 1 /c) and the diopside 
structure (space group C2/c).We have 
tacitly assumed this conclusion in our 
interpretation of Fig. 6(A) as resulting 
from a two-Cpx field that has penetrated 
the lower but not the upper boundary 
of a Cpx-Opx transition loop (Fig. 6C) . 
Under this interpretation the narrow 
two-pyroxene field above about 950°C 
in Fig. 6(A) corresponds to a pigeonite- 
hypersthene transition loop in Mg-bear- 
ing pyroxenes. We have been unable to 
extend this field to its presumed inter- 
section at pure FeSi0 3 because a liquid 
appears in hydrothermal experiments 
above 1010°C and a pyroxenoid replaces 
both pyroxenes above 1085°±15°C in 
dry experiments. 

Assuming (1) that the Cpx series is a 
regular solution, (2) that the presumed 
miscibility ~gap in that series is sym- 
metrical, and (3) that the interpretation 
shown in Fig. 6(C) is correct, we can 
extrapolate the critical temperature {Tc, 
about 870 °C, in Fig. 6C) to low pres- 



90 



CARNEGIE INSTITUTION 







Mole per cent FeSi0 3 






50 60 70 80 90 


100 


1100 


i i i i 
P=20kb 


k A 


1000 


w> 


ft \ - 




■ ■„ / 




Cpx X,/ 








y 


+ 


O 900 


m^A 


> 


- 



•CD 


>£^<1 




Opx 


j3 

"a 


y^ 







a> 

|- 800 

|2 


- *A^ 


h_ 


j - 


j^^n] Cpx +0px 


1 


700 


-Ji 


i i _i_ 




?" 



Cq 5 Fe 5 Si0 3 40 30 20 10 FeSi0 3 

Mole per cent CaSi0 3 



o 



CD 
CL 

£ 
|2 




/C2/o\ #& 



C2/c+0px 



Cpx+Opx— 




(/ Two Clinopyroxenes ^ 
^A ( metastable)- 



Hd 



B 



Fs Hd 

c 



Fs 



Fig. 6. Phase relations for the join hedenbergite-ferrosilite at 20 kb. (A) Data points and 
probable phase boundaries; (B, C) schematic interpretations of data. (A) Open triangles, 
compositions of pyroxenes formed by breakdown of single-phase Cpx. Solid triangles, compositions 
of pyroxenes formed by reaction of Hd -f- OFs pairs. Solid squares, compositions of Cpx 
homogenized from mixtures of FsmWoio and FsooWoio, both Cpx. Horizontal extent of each 
symbol indicates uncertainty in composition; vertical extent indicates uncertainty in temperature. 
X indicates the synthesis conditions of a Cpx shown (by C. W. Burnham) to have the 
pigeonite (P2i/c) structure. (B) Interpretation of relations in A as resulting from the intersection 
of three two-pyroxene regions along an isothermal three-phase line. We have not observed 
a two-clinopyroxene field (C2/c -f- P2i/c), but it might escape detection by X-ray diffraction if less 
than about 10 mole % in width. (C) Interpretation of relations in A as resulting from the 
intersection of a two-Cpx field with the lower but not the upper boundary of a Cpx-Opx 
transition loop. T e is the critical temperature of the assumed miscibility gap. This is our 
preferred interpretation. 



GEOPHYSICAL LABORATORY 



91 



sures. Using the excess volume of mixing 
for the series — 1.4 ±0.5 A 3 at the compo- 
sition Fs 75 Wo 2 5 (Lindsley, Munoz, and 
Finger, this report) — and the method of 
Bell and Davis (Year Book 64, p. 122), 
we calculate that the critical temperature 
lies at approximately 820 °C at 1 atm. 
This result provides a metastable (the 
stable assemblage is Cpx + fayalite + 
silica) terminus for the crest of the two- 
pyroxene field in the pyroxene quadri- 
lateral diopside-hedenbergite-enstatite- 
ferrosilite. It appears that the two- 
pyroxene field might not intersect the 
solidus in the Mg-poor portion of that 
quadrilateral. 



Unit-Cell Parameters of Clinopyrox- 

enes along the Join Hedenbergite- 

Ferrosilite 

D. H. Lindsley, J. L. Munoz, and L. W. Finger 

Unit-cell parameters for clinopyrox- 
enes synthesized along the join heden- 
bergite (Ca .5Feo.5Si0 3 )-ferrosilite (Fe 
Si0 3 ) have been determined through 
least-squares refinement of powder X-ray 
diffraction data (Fig. 7) . The data were 
collected at a scanning rate of %°/ 
minute in the range 19° to 65° 20, with 
CuKa and CuKccx radiation. CaF 2 was 
used as an internal standard in all but 
one powder mount, for which (Fs 65 Wo 3 5) 
a correction term for 20 was calculated 



mole% FeSi0 3 

1 '' ' 60 70 80 90 



Fs 




mole % Co. Si0 3 



mole % Fe Si0 3 

Hd 60 70 80 90 Fs 



450 


. 1 1 1 i 1 


i i i i 


446 


^A 


x 


442 


~ V,A 3 


X- 


438 










. 




40 30 20 

mole % Co. Si0 3 



Fig. 7. Unit-cell parameters for clinopyroxenes synthesized on the join hedenbergite-ferrosilite. 
Values for hedenbergite are from powder data (J. Nolan, personal communication), and for 
ferrosilite from single-crystal data (Burnham, Year Book 64). All cell dimensions for intermediate 
pyroxenes were determined by least-squares refinement of powder diffractometer data. Condi- 
tions of synthesis: FsaoWoio, 25 kb, 1210°C, dry; FsmWobb, 20 kb, 1210°C, dry; FsrsWoas, 20 kb, 
940°C, hydrothermal; FssoWo*,, 25 kb, 1230°C, dry; Fs9 B Wo 5 , 37 kb, 1325°C, dry. Molar volumes 
(7) calculated on the basis of Z = 8(Ca,Fe)Si0 3 per unit cell. 



92 



CARNEGIE INSTITUTION 



as a variable in the least-squares refine- 
ment, with the use of a program adapted 
from Burnham (Year Book 64, PP- 200- 
202). This term was insignificant, result- 
ing in a change of 0.006° 29 for each 
observation. 

The variation of the unit-cell param- 
eters with composition is presented in 
Fig. 7. The data were fitted by using 
weighted regression analysis, and the 
solution having the smallest standard 
error of estimate was plotted. The data 
for a and ft require quadratic solutions; 
the data for the other parameters are 
best described by cubic equations, al- 
though only for b is there a large differ- 
ence between the quadratic and cubic 
solutions. 

The data in Fig. 7 should be useful in 
constructing grids of unit-cell parameters 
for the pyroxene quadrilateral. In addi- 
tion, the small positive volume of mixing 
indicates that the size of the two-pyrox- 
ene field at 20 kb (Munoz and Lindsley, 
this report) should decrease at lower 
pressures. 

Optical Properties of Synthetic 

Clinopyroxenes on the Join 

Hedenbergite-Ferrosilite 

G. H. Myer* and D. H. Lindsley 

Accurate data on the optical properties 
of synthetic clinopyroxenes along the 
join hedenbergite (Cao. 5 Feo.5Si0 3 ; Hd)- 
ferrosilite (FeSi0 3 ; Fs) have long been 
needed to provide one set of limiting con- 
ditions to the optical properties of clino- 
pyroxenes lying in the quadrilateral di- 
opside - hedenbergite - enstatite - f errosilite. 
Unsuitable morphology and fine grain 
size prevented Bowen, Schairer, and 
Posnjak (1933, pp. 260-261) from de- 
termining n$ and 2V, and limited the 
precision of their measurements of n a 
and n y to ±0.003. We have measured 
the optical properties of clinopyroxenes 
synthesized at pressures of 20 kb or 

* Department of Geological Sciences, Univer- 
sity of Maine, Orono, Maine. 



above, where ferrosilite and FeSi0 3 -rich 
solid solutions are stable. 

The clinopyroxenes were first synthe- 
sized in iron capsules in a large-volume 
(%-inch bore) solid-media, piston-and- 
cylinder pressure device. Products that 
appeared homogeneous (with the excep- 
tion of 1 to 2% fayalite + quartz in a few 
samples) upon X-ray and cursory optical 
examination, were then recrystallized in 
iron capsules for several hours at tem- 
peratures just below the solidus. Ex- 
amination of these recrystallized samples 
in thin section reveals a porphyroblastic 
texture, the larger grains being about 
100 fx in length. The demonstration of a 
two-clinopyroxene field intersecting the 
solidus at 20 kb for the join diopside- 
enstatite (Kushiro, this report) suggests 
that this bimodal distribution of grain 
size might reflect two clinopyroxenes of 
different composition. However, the 
smooth variation of both optical proper- 
ties and unit-cell parameters with bulk 
composition of the samples is inconsistent 
with the existence of a miscibility gap, 
10 mole % or more in width, on the join 
Hd-Fs. Inasmuch as the porphyroblastic 
texture is found over a wide range of 
compositions, it cannot have resulted 
from a narrow miscibility gap; instead, 
it probably reflects a variable growth 
rate among crystals of the same compo- 
sition. 

Clinopyroxenes of the compositions 
Fs G5 Wo 3 5, Fs 75 Wo 25 , and Fs 85 Wo 15 were 
too fine-grained when treated as above 
to permit good optical measurements. 
Additional samples of these compositions 
were therefore recrystallized hydro- 
thermally at 20 kb by techniques de- 
scribed by Munoz and Lindsley (this 
report) . 

The clinopyroxenes were studied as 
grains in immersion oils — free on glass 
slides, mounted on epoxy resin sub- 
strates or mounted on glass spindles — 
and in thin sections. Refractive-index 
measurements were made in sodium 
vapor light; oils matching an index were 
transferred by micropipet to a Leitz- 



GEOPHYSICAL LABORATORY 



93 



Jelley microrefractometer for immediate 
calibration. Precision of the refractive- 
index determinations is considered to be 
±0.001. The optic angle, 2V Z , was mea- 
sured in sodium light by the interfer- 
ence-figure method of Tobi (1956) and 
by orthoscopic universal-stage tech- 
niques; both optic axes were observed 
as recommended by Munro (1963). 

The refractive indices are plotted in 
Fig. 8 along with the data of Bowen, 
Schairer, and Posnjak (1933) for n a 
and n y . The curves for n a and np must 
intersect between Fs 95 Wo 5 and Fsi 00 Wo . 
Composition Fs 98 Wo 2 is predicted to 
have n a — rip and thus to be pseudo- 
uniaxial. 

The /3 refractive index of clinoferro- 
silite given by Lindsley, MacGregor, and 



Davis (Year Book 63, p. 176) is 1.767 ± 
0.002. The new measurements give 
1.765 ±0.001, which is at their low ex- 
treme. The data of Bowen, Schairer, and 
Posnjak (1933) for n a and n y of heden- 
bergite and intermediate compositions lie 
consistently above the new curves. The 
small amount of fay alite + quartz in 
some of our samples would bias our 
results in the right direction but to an 
insufficient extent to account for this 
difference; we cannot explain it. 

Between Fs 50 Wo 5 o and Fs 95 Wo 5 the 
optic plane is parallel to (010), hence 
Y = b. The monoclinic inclined dispersion 
with V>r is weak for Fs 50 Wo 5 o and in- 
creases to strong for Fs 95 Wo 5 . Interfer- 
ence figures, as observed in white light, 
exhibit one black isogyre and the other 



Hd 
Ca 5 Fe 5 Si0 3 




60 70 80 90 

Mole per cent FeSi0 3 



60 70 80 SO 

Mole per cent FeSi0 3 



Fig. 8. Variation of optical properties with composition for synthetic clinopyroxenes along 
the join hedenbergite-ferrosilite. Crosses are data points: horizontal bar gives preferred value 
of parameter; extent of vertical bar gives estimated uncertainty. Curves are visual estimates. 
Open circles give values of n a and n y reported by Bowen, Schairer, and Posnjak (1933, pp. 260- 
261). The optic angle 2Vz must go to zero between Fs 9 5Wo 5 and Fsioo because the orientation of 
the optic plane changes from parallel to (010) to perpendicular to (010) in that interval. 



94 



CAENEGIE INSTITUTION 



strongly colored. Clinoferrosilite, on the 
other hand, has the optic plane perpen- 
dicular to (010) and X — b. The mono- 
clinic horizontal dispersion is strong: 
both isogyres are colored with v>r. 2V Z 
must decrease to zero between Fs 95 Wo 5 
and Fs 100 Woo, rather than at Fs 94 Wo 6 
as predicted by Hess (1949) and by 
Muir (1951). 

Extinction angle Z /\c (Fig. 8) could 
not be measured to better than ±2° 
because of imperfect morphology. Twin- 
ning is developed often on (100) and 
occasionally on (001). Extinction angles 
for twinned and untwinned grains agreed 
to within 3°. 

Oxidation of Ca-RiCH Clinopyroxenes 

H. G. Huckenholz 

The activities of Si0 2 in magmas are 
believed to be strongly controlled by 
oxidation, according to Osborn (1959). 
The change from ferrous to ferric iron 
causes silica enrichment, and magmas 
originally Ne normative may change to 
Hy- or even Qz-normative bulk compo- 
sition. Yoder and Tilley (1962) discussed 
and illustrated the effect of oxidation on 
olivines and suggested a similar reaction 
for Ca-rich clinopyroxenes: 

6CaFe 2+ Si 2 6 + 2 ?± 2Fe 3 0,+ 

hedenbergite magnetite 

6CaSi0 3 + 6Si0 2 (1) 

wollastonite quartz 

In the event that ferric iron is retained in 
the clinopyroxene structure, the following 
reaction should take place: 

4CaFe 2+ Si 2 6 + 2 ^± 

hedenbergite 

2CaFe 3+ (Fe 3+ Si) 6 + 2CaSi0 3 + 



i erri -Tschermak' s 
molecule 



wollastonite 



4Si0 2 (2) 

quartz 

The nature of the incorporation of ferric 
iron in Ca-rich clinopyroxene, particu- 
larly in diopside, is now known (Hucken- 
holz, Schairer, and Yoder, Year Book 
66). Diopside may take up in solid 
solution 33% by weight of the ferri- 



Tschermak's molecule (FTs), and as 
much as 19% Fe 2 3 may be retained 
in the diopside structure in the absence 
of soda at 1157°C and 1 atm pressure. 
Additional substitution of the ferri- 
Tschermak's molecule in a clinopyroxene 
solid solution is not stable as a single 
phase, and phase assemblages of clino- 
pyroxene ss + hematite, clinopyroxene ss + 
hematite + andradite, and hematite -t-an- 
dradite are formed. Above 1157°C poly- 
phase assemblages occur in which andra- 
dite is replaced by wollastonite, pseudo- 
wollastonite, and additional hematite. 
Therefore, reaction 2, outlined by Yoder 
and Tilley (1962), should be modified 
with regard to andradite as follows: 

[4CaFe 2+ Si 2 6 + 2CaMgSi 2 6 ] +0 2 ?± 

clinopyroxeness 

[CaFe s+ (Fe 3+ Si) 6 +2CaMgSi 2 6 ] + 

ferri-diopsidess 

Ca 3 Fe 2 3+ Si 3 12 +4Si0 2 (3) 

andradite quartz 

In order to test reaction 3, a series of 
natural Ca-rich clinopyroxenes (Table 1) 
were treated at 1125°C and 1 atm over a 
period of 3 weeks. The envelopes con- 
taining the samples were perforated to 
achieve better circulation of air. It is 
believed that treatment like this con- 
verts most of the Fe 2+ into Fe 3+ , and the 
analyses of the clinopyroxenes have been 
recalculated on an Fe 3+ basis. As plotted 
on the wollastonite-diopside-hematite 
join of Fig. 9 (neglecting the excess of 
free silica), oxidation of augites and 
hedenbergites yields a phase assemblage 
of clinopyroxene ss + hematite ( + cristo- 
balite) in sample 1, whereas in samples 
2, 3, and 4 an assemblage of clinopyrox- 
eness + hematite + andradite ss ( + cristo- 
balite) is obtained. One would expect 
from equation 3 that no hematite should 
be formed. Its appearance is caused by 
the ferrosilite content of the natural 
clinopyroxenes, however, as can be seen 
in the norm of the clinopyroxenes. 

In all four samples the reaction pro- 
duces fine-grained intergrowths of the 
phases, sometimes with oriented hematite 



GEOPHYSICAL LABORATORY 



95 



hem (Feg0 3 ) 



FT S 
(Co Fe +3 2 Si0 6 ) 




(Ca 3 Fe|3 S i 3 |2 ) ( 



(CaSi03T pwo+WOss 



(CaMgSi 2 6 ) 



Fig. 9. Wollastonite-diopside-hematite join below 1137°C and the composition of natural 
clinopyroxenes after "oxidation," neglecting cristobalite. Phases encountered: FTs, ferri- 
Tschermak's molecule; andr, andradite; wo, wollastonite ; pwo, pseudowollastonite ; di, diopside; 
cpx, clinopyroxene ; hem, hematite. Numbers refer to samples, as indicated in Table 2. 



and cristobalite parallel to (001) of the 
new clinopyroxene. The composition of 
the light yellow to yellow-brown ferri- 
diopside may be traced by using the 
angular separation of 20 (240) -20 (041) 
(Huckenholz, Schairer, and Yoder) : 
about di 95 hem 5 in sample 1, and about 
di 67 hem 17 wOie in samples 2, 3, and 4. The 
garnet formed in sample 2 is almost a 
pure andradite with minor amounts of 
a grossularite component, whereas the 
garnet of samples 3 and 4 contains both 
a grossularite and a Ti-garnet com- 
ponent, as revealed by the size of their 
cell edges, and by the Al and Ti contents 
of the samples (Table 2, samples 2, 3, 
and 4) . 
From the data at hand it is evident 



that the formation of ferri-diopside and 
andradite is favored in addition to Fe 3+ - 
bearing oxides if a magma crystallizes 
under oxidizing conditions. However, 
ferri-diopside is very often associated 
with those igneous rocks that have un- 
dergone oxidation and that contain small 
amounts of alkalies, especially sodium. 
Large amounts of sodium lead to the 
formation of the acmite molecule, 
whereas potassium enters the diopside 
structure only in very small amounts. 
If the sodium content of a magma is 
high and insufficient A1 2 3 is available, 
a preferred formation of acmite molecules 
or even of acmite itself may take place 
under oxidizing conditions. 



96 



CARNEGIE INSTITUTION 



TABLE 2. Analyses of Natural Clinopyroxenes Oxidized in this Study 





1 


la 


2 


2a 


3 


3a 


4 


4a 








Analyses 










Si0 2 


50.58 


49.8 


48.34 


47.4 


49.44 


48.3 


47.09 


45.9 


AI2O3 


2.20 


2.1 


0.30 


0.3 


1.31 


13 


0.80 


0.8 


Fe=0 3 


1.57 


18.5 


1.50 


26.5 


0.88 


24.3 


1.61 


32.3 


FeO 


15.53 




22.94 




21.64 




28.41 




MnO 


0.28 


0.3 


3.70 


3.6 


0.42 


0.4 


0.64 


0.6 


MgO 


12.60 


12.4 


1.06 


1.1 


6.92 


6.8 


1.19 


1.1 


CaO 


16.40 


16.1 


21.30 


20.9 


18.23 


17.8 


18.99 


18.5 


Na 2 


0.24 


02 


0.14 


0.1 


0.29 


0.3 






K 2 


0.03 




0.03 




0.03 




0.05 




Ti0 2 


0.61 


0.6 


0.08 


0.1 


0.83 


0.8 


0.83 


0.8 


Cr 2 3 


<0.01 












<0.02 




H 2 






0.46 


100.0 




100.0 


0.14 




Totals 


100.04 


100.0 


99.85 


99.99 


99.97 


100.0 








Norms 










Q 






1.29 








1.27 




Or 


0.18 




0.18 




o.is 




0.30 




Ab 


2.03 




1.18 




2.45 








An 


4.84 




0.10 




2.18 




2.64 




Di 


62.98 




93.33 




75.00 




81.46 




Hy 


25.20 




1.02 




16.79 




10.66 




01 


1.38 








0.54 








Mt 


2.28 




2.17 




128 




2.33 




11 


1.16 




0.15 




1.58 




1.58 




Molecular Fs/En 


















ratio in Hy 


0.39 




0.93 




0.63 




0.93 






Composition in 


di, wo, hem, 


and cr 


after Oxidation, wt % 






di 




66.3 




19.2 




40.3 




7.9 


wo 








34.3 




16.5 




342 


hem 




20.7 




26.2 




25.6 




33.9 


cr 




13.0 




20.3 




17.6 




24.0 








Results of Oxidation 








Ferri-diopside 




* 




* 




* 




* 


Andraditesa 




t 




* 




* 




* 








(12.037 : (V;m±QM " ) 


(12.055 


±0.005 A) 


Hematite 




* 




* 




* 




* 


Cristobalite 




* 




* 




* 




* 



* Present. 
t Absent. 

1. Augite (4330), recalculated as "oxidized" augite (la), from ferrogabbro MZ, Skaergaard 
intrusion, east Greenland (Brown, 1957). Donated by G. M. Brown. 

2. Hedenbergite, recalculated as "oxidized" hedenbergite (2a), from a skarn rock close to 
a diorite intrusion, Herault, Shasta County, California. (Wyckoff, Merwin, and Washington, 1925). 
Collection of H. S. Yoder, Jr. 

3. Ferroaugite (EG 4316), recalculated as "oxidized" ferroaugite (3a), from ferrodiorite 
UZb, Skaergaard intrusion, east Greenland (Brown and Vincent, 1963). Donated by G. M. Brown. 

4. Ferrohedenbergite (K 3381), recalculated as "oxidized" ferrohedenbergite (4a), from ferro- 
syenite, Kiglapait intrusion, Labrador, Canada. (C. O. Ingamells, analyst; contains 0.02 Cr 2 Os, 
020 NiO, 0.005 SrO, and 0.02 P 2 Oe.) Donated by S. A. Morse. 



Stability of Omphacite 

Omphacite, a clinopyroxene solid solu- 
tion consisting mainly of diopside 
(CaMgSi 2 6 ) and jadeite (NaAlSi 2 O e ) 
components, is one of the major constitu- 



ent minerals of eclogites. It also occurs 
in some glaucophane schists and associ- 
ated metamorphic rocks. In the absence 
of free silica, omphacite breaks down 
to nepheline, albite, and a more diopside- 



GEOPHYSICAL LABORATORY 



97 



rich oraphacite with less jadeite compon- 
ent at lower pressures or at higher tem- 
peratures. In the presence of free silica, 
omphacite reacts with silica to form 
albite and a more diopside-rich ompha- 
cite with less jadeite component at 
lower pressures or at higher temperatures. 
In the present experiments the stabil- 
ity field of omphacite has been deter- 
mined in the joins diopside-jadeite (free 
silica absent) and diopside-albite (free 
silica present) . 

Stability of Omphacite in the Absence of 
Excess Silica 

P. M. Bell and J. Kalb 

The pure j adeite end member has been 
studied experimentally in the subsolidus 
by Robertson, Birch, and MacDonald 
(1957) and by Bell {Year Book 63, pp. 
171-174). Pure jadeite breaks down by 
means of the univariant reaction: 

2 jadeite (NaAlSi 2 6 ) <=± albite 

(NaAlSi 3 8 )+nepheline (NaAlSi0 4 ) 

In the present study the primary aim has 
been to evaluate the effect on the uni- 
variant reaction of the addition of di- 
opside (CaMgSi 2 6 ). The breakdown of 
jadeite is highly pressure sensitive and 
has provided a useful geobarometer for 



eclogitic rocks even though most natural 
jadeites are omphacitic since they con- 
tain an average of at least 10 to 15% 
diopside (see Coleman et al., 1965). It is 
essential to know how the presence of the 
diopside component affects the jadeite 
geobarometer. 

Jadeite is in equilibrium with albite 
and nepheline along a curve that is very 
steep to the pressure axis: at 1150°C the 
curve lies at 22 kb, and at 1225 °C the 
curve lies at 24 kb. These two tempera- 
tures, 1150° and 1225°C, were chosen for 
experimental isotherms. The reaction 
omphacite! *± plagioclase + nepheline + 
diopsidic pyroxene (omphacite 2 ) was in- 
vestigated isothermally for the composi- 
tions Jd 5 oDi 50 and Jd 70 Di 30 wt % with 
the piston-cylinder, solid-media appa- 
ratus. The reaction is sensitive to the 
compositions of the phases. Thus in 
selecting starting materials for the right- 
hand side of the reaction (low-pressure 
assemblage), the compositions of the 
phases must be those that are at equi- 
librium on the phase boundary. Theo- 
retically, therefore, the boundary must be 
known before it can be studied. In order 
to avoid this paradox, glasses and crys- 
talline mixes were used to reconnoiter the 
phase boundaries. Boundaries produced 
by using these starting materials were 



ou 


1 1 


1 ' M50°C ' 


1 1 


1 


25 


— 


Omphacitei 




■■ 






D 


n 






"" ■ 


- 






□ y 


■ 




20 


— 


as 




-- 


XI 




A 


Plagioclase 




* 




/ 


+ 




3 15 


— 


/ 


Nepheline 


— 


8? 




/ 


+ 
Omphacite 2 




CL 










10 








~ 


5 


1 1 


I 1 1 1 


1 1 


! 



Diopside 10 
CaMgSi 2 6 6 



20 



30 



70 



40 50 60 

Weight per cent 
Fig. 10. Pressure-composition section for omphacite at 1150°C. 



80 



90 



98 



CARNEGIE INSTITUTION 



30 



25 



20 



5 — 



Omphacite, 



I225°C ' ' I r 

n 
a 
□__ 



Plagioclase 

+ 
Nepheline 

+ 
0mphacite 2 



CaMgSi 2 6 

Weight per cent 

Fig. 11. Pressure-composition section for omphacite at 1225°C. 



Diopside io 20 30 40 50 60 70 80 90 Jadeite 



NaAISuO, 



2^6 



metastable but were useful in giving a 
general location for the equilibria. By 
trial and error, the correct assemblages 
were synthesized close to the boundary 
and used as starting materials. The as- 
semblage plagioclase + nepheline + om- 
phacite 2 was added in equal proportions 
to the omphacite!, and an increase or 
decrease in the amount of reactant or 
products was used as an indication of 
the direction of reaction. The disappear- 
ance of plagioclase was used to fix the 
phase boundary. 

Experimental points determined for 
the isotherms 1150° and 1225°C are 
shown in Figs. 10 and 11. The results sug- 
gest that the breakdown of jadeite is not 
sensitive to the presence of diopside in 
solid solution with it unless the concen- 
tration of diopside is greater than 30 
wt % . These results confirm the theoreti- 
cal result of Bell and Davis {Year Book 
64, pp. 120-123) and the preliminary ex- 
periments with natural samples of 
Newton and Smith (1967). The view of 
Robertson, Birch, and MacDonald 
(1957) that the presence of omphacite in 
rocks suggests that they formed at high 
pressure appears to be valid. 



Stability of Omphacite in the Presence 
of Excess Silica 

I. Kushiro 

The subsolidus region of the join di- 
opside-albite was studied previously in 
the pressure range 10 to 32 kb and the 
temperature range 1050° to 1350 °C, with 
piston-cylinder, solid-media apparatus 
(Year Book 64, pp. 113-115). It was 
shown that the stability field of an om- 
phacite + quartz assemblage relative to 
that of an omphacite + albite-rich plagio- 
clase + quartz assemblage expands toward 
albite with an increase of pressure at 
constant temperature. Starting materials 
used in those experiments were glass of 
compositions on the join diopside (Di)- 
albite (Ab) made by Schairer and Yoder 
(1960). It is possible, however, that 
metastable assemblages may be formed 
from glass even at high temperatures. 
The present experiments have been made 
by the piston-out technique with crystal- 
line starting materials. Experiments were 
performed for the three compositions 
Di 55 Ab 45 , Di 40 Ab 60 , and Di 25 Ab 75 (wt %) 
at 1250 °C in the pressure range 25 to 
34.5 kb. 

For the composition Di 3S Ab 4 5, the om- 
phacite + plagioclase + quartz assemblage 



GEOPHYSICAL LABORATORY 



99 



was converted to omphacite + quartz at 
27.5 kb and the omphacite + quartz as- 
semblage was converted to omphacite + 
plagioclase + quartz at 25.5 kb in 5 to 
6 hours. The boundary between these two 
assemblages for Di 55 Ab 4 5 composition is 
located, therefore, at 26.5 ±1 kb at 
1250°C, about 0.5 kb higher than that 
determined with the use of glass as start- 
ing material (Fig. 12) . For the composi- 
tion Di 40 Ab 60 , the reversal reactions took 
place at 32.5 and 29 kb, respectively. 
The boundary for this composition then, 
is 30.8 ±1.8 kb at 1250°C, about 1.2 kb 
higher than that given in the previous 
experiments. It appears that the reaction 
rate is slower for more albite-rich compo- 
sitions. For the composition Di 25 Ab 75 , 
complete disappearance of plagioclase 
was not observed in the runs at pres- 
sures up to 34.5 kb for 5 to 6 hours at 
1250°C, and the boundary could not be 
determined successfully. The amount of 
plagioclase relative to omphacite and 
quartz, estimated from the relative in- 



tensities of X-ray reflections, decreases 
with increasing pressure. This decrease 
could occur in both the omphacite + 
plagioclase + quartz field and the ompha- 
cite + quartz field, however, and cannot 
be used to locate the boundary curve. 

Extrapolation of the boundary to 
albite composition suggests that the 
breakdown of albite into j adeite + quartz 
would take place at 33.5 ±2 kb at 
1250°C. The value is consistent with that 
obtained by extrapolating the break- 
down curves of albite given by Birch 
and LeComte (1960) and Newton and 
Smith (1967). 

Comparison of the previous results 
with those obtained in the present experi- 
ments suggests that the boundary be- 
tween the field of the omphacite + quartz 
assemblage and that of the omphacite + 
plagioclase + quartz assemblage given in 
the previous experiments is nearly correct 
for compositions more diopside rich than 
DisoAbgo at temperatures at least near 
1250°C. For compositions more albite 



Owl 




1 1 

I250°C 


1 1 


1 1 I] 1 i 

E . — 


-- 


30 

.Q 
CO 

V) 

£ 25 


— 




Omph+Qz 


'A 

Omph + Plag+Qz 


- 


a. 


_ 




/ 

/ 

B 


| Omph + PI+Qz -> Omph+Qz 


\ 


20 








E Omph + PI + Qz -> Omph + PI + Qz 


— 




. 






Omph +Qz -*- Omph+PI+Qz 


\ 




" 


1 1 


1 1 


1 1 1 1 


- 



CaMgSi 2 0, 



2^6 



50 

Weight per cent 



NaAISi,0, 



3 U 8 



Fig. 12. Subsolidus phase relations in the join diopside-albite at 1250°C. The dashed curve 
A-B is a boundary between the field of the omphacite + quartz assemblage and that of the 
omphacite + plagioclase + quartz assemblage given in the previous experiments ( Year Book 64, 
p. 113). The range a-b is the uncertainty of pressure for the reaction albite == j adeite + quartz 
at 1250°C in the extrapolation of the data of Birch and LeComte (1960) and Newton and 
Smith (1967). 



100 



CARNEGIE INSTITUTION 



rich than Di 50 Ab 50 , however, the cor- 
rected boundary lies at 1 to 2 kb higher 
pressure than that of the previous ex- 
periments. 

The Temperature-Compression 
Melting Relation 

M . C. Gilbert 

A "new melting law" of the form T M = 
T M ° (1 + CAF/Fo), for application at 
high pressures, has recently been pro- 
posed by Kraut and Kennedy (1966a). 
It was based on the observation that 
plots of melting temperature at pressure 
against isothermal (room-temperature) 
compression of the solid phase appeared 
to give a linear relation. These authors 
originally suggested that "this result 
seems to apply to all substances examined 
which melt with an increase in vol- 
ume . . . ." Shortly thereafter Kraut 
and Kennedy (19666) indicated that 
some solidified gases did not obey this 
relationship. More recent work by 
Luedemann and Kennedy (1968) has 
shown that the linear relation also breaks 
down for the alkali metals as melting 
maxima are approached, and compres- 
sion of the solid reaches 35 to 40%. The 
latter authors speculated that the results 
may indicate that, at equivalent com- 
pressions for silicates (in the deep mantle 
and core), fusion maxima would be en- 
countered. In the light of these argu- 
ments it seemed of interest to examine 
such a plot for acmite, which melts in- 
congruently (Fig. 13). The melting re- 
lations were reported previously in Year 
Book 65 (pp. 241-244). Compressibility 
values are derived from ultrasonic mea- 
surements on a specimen of aegirite 
(Birch, 1966, Tables 7-10 and 7-12; 
Alexandrov and Ryzhova, 1961). 

It is clear that no linear relation be- 
tween aF/Vq and melting temperature 
exists for acmite. In fact, the resulting 
curve is similar in form to the fusion 
curve itself. It was thought possible that 
deviation from linearity might be due to 



incongruent melting. For comparison, 
plots of melting temperature versus com- 
pressibility for other silicates are shown 
in the inset of Fig. 13. Fayalite 
(Fe 2 Si0 4 ) , which melts incongruently up 
to 40 kb (Hsu, 1967), and sanidine 
(KAlSi 3 8 ), which melts incongruently 
up to 20 kb (Lindsley, 1966) show linear 
relationships over that region; thus, in- 
congruent behavior cannot be used to ex- 
plain curvature. In contrast, diopside, 
which melts congruently (Boyd and 
England, 1963), and sanidine, which 
melts congruently above 20 kb (Lindsley, 
1966), do not show linear relations. On 
the whole, the relationships for silicates 
shown in Fig. 13 reflect the form of the 
fusion curves: linear fusion curves result 
in linear T M versus A7/T curves ; fusions 
with marked curvature result in curved 
T M versus AV/V relations, with the ex- 
ception of albite. 

A pressure effect on the emf of thermo- 
couples cannot be used to account for 
the nonlinear T M -AV/V curves presented 
here, because all curves shown are plotted 
without this correction so that differences 
between curves are real. Nothing in the 
experimental techniques used to deter- 
mine the melting curves can be appealed 
to since all these curves were determined 
by the same type of apparatus and tech- 
nique. It is possible that the compressibil- 
ity measurements at lower pressures for 
silicates may have larger uncertainties 
than expected and that extrapolation to 
higher pressure of low-pressure measure- 
ments is unwarranted. In any case, diffi- 
culties in the simple temperature-com- 
pression relation are not limited to sili- 
cates. Some of the elements and simple 
compounds considered by Kennedy and 
co-workers also do not show linear be- 
havior. There would seem to be no a 
'priori reason why a melting law that 
does not take into account properties of 
both solid and liquid phases should work. 
Whatever the reasons, the "new law of 
melting" does not seem to be substan- 
tiated by silicates over the pressure range 
thus far investigated. 



GEOPHYSICAL LABORATORY 

1 : 

1600 

1500- 

1400 



101 




1300- 



1200- 



1000 



2200 


i 1 1 r 

Fo 


1 


- 


2000 


_ ^^^^ 


^-En 


- 




""""^ ^^^~^' 


^^Di 


• 


1800 


^^^ ^^"^ 




- 




/^ \^^^ Jd ^^ 


^Ac 


■ 


1600 


■/^ ^^"^ /^"-"'^ 




-Sa 


1400 






- 


1200 








1000 


1 1 1 1 . 


i I 


- 



.0100 



.0200 



.0300 



.0400 



A V/ 



V n 



Fig. 13. Melting temperature of acmite versus its compression at room temperature. Inset 
shows similar plots for other silicates. Compression data summarized in Birch (1966, Table 7-10). 
Melting curves for Fo, forsterite, Mg 2 Si0 4 (Davis and England, 1964) ; En, enstatite, MgSiOs 
(Boyd, England, and Davis, 1964) ; Di, diopside, CaMgSi 2 0e, and Ab, albite, NaAlSi 3 8 (Boyd 
and England, 1963); Jd, jadeite, NaAlSi 2 9 (Bell, Year Book 63); Ac, acmite; Sa, sanidine, 
KAlSLsOs (Lindsley, 1966); Fa, fayalite, Fe 2 Si04 (Hsu, 1967). The points shown on the curves 
in the inset represent 10-kb steps in pressure. The first point on the jadeite curve represents 
30 kb. 



FELDSPARS AND HIGHLY 
FELDSPATHIC ROCKS 

The Melilite-Plagioclase Incompati- 
bility Dilemma in Igneous Rocks 

H. S. Yoder, Jr., and J. F. Schairer 

Field occurrences clearly indicate the 
incompatibility of plagioclase and meli- 
lite in lavas of the alkali suites. Volcanic 
centers of the same province eject either 
plagioclase-bearing pyroclastics and 
lavas or melilite-bearing lavas, and oc- 
casionally extrude both lava types sepa- 
rated by long intervals of time. The field 
association is close and definitive; the 
genetic relationship, however, is not as 



yet evident from field deductions. Be- 
cause plagioclase and melilite are the 
key minerals to the major trends of 
alkali igneous rock, the system albite 
(Ab)-anorthite (An)-akermanite (Ak) 
at 1 atm was studied by Schairer and 
Yoder (this report). The results, sur- 
prisingly enough, demonstrated that a 
plagioclase and a melilite were com- 
patible with liquid over much of the 
range of plagioclase, with the exception 
of compositions on the join Ab-Ak. Al- 
though the plagioclase compositions 
were not determined, it was obvious that 
plagioclase and melilite were compatible 
over a significant range of plagioclase 



102 



CARNEGIE INSTITUTION 



composition, in direct contradiction to 
the field observations. The simplified 
nature of the laboratory experiments 
was believed to be the cause in part of 
the dilemma, and other factors, such as 
the presence or absence of water, were 
considered relevant to the problem on 
the basis of field observations. Nixon 
and Clark (1967, p. 470) believed that 
magma richer in volatiles and calcium 
produced melilite-bearing lavas rather 
than plagioclase-bearing lavas. Melilite 
was present in the more calcium-rich 
compositions in the simple Ab-An-Ak 
system, however, with plagioclase. The 



effect of volatiles, specifically H 2 0, was 
therefore investigated as well. 

Compositions on or close to the 
Ab 5 oAk 50 -An5 Ak 5 o were studied at tem- 
peratures from 700° to 1175°C at Ph 2 o = 
2 kb. These conditions were chosen be- 
cause of the restrictions of the stability 
of akermanite itself (Yoder, Year Book 
66, p. 473) and the limitations imposed 
by An 50 Ak 50 (Yoder, this report). The 
anhydrous phase relations are shown in 
Fig. 14 by a series of lines, indicating 
the appearance or disappearance of 
phases. The anhydrous composition close 
to Ab 37 An 13 Ak5o marks the change at the 



1400 



1300 



1200 



100 



1000 



900 



800 



700 



600 




. -J-' '— ■ — .77. 7Z~- ~ 

09* 






U>e\ 



Hydrous. S23! 



WxWL-i- 



o^JH 



e\V\nq — - — 




A 






Cpx+Gr+PI+V.'clss+G. 



Ab 50 Ak 50 



40 50 60 

Weight per cent 



70 



AnsnAkfi 



Fig. 14. At the higher temperatures are presented the 1-atm thermal relations of compositions 
on or near the join albitesoakermaniteso-anorthitesoakermaniteso (see Fig. 15, Schairer and Yoder, 
this report). The appearance of each phase and the disappearance of melilite (Mel) is indicated 
by a patterned curve. At the lower temperatures are presented the hydrothermal relations in 
an excess of water at Ph 2 o = 2 kb for the same compositions. The curves mark the boundaries 
between assemblages. The assemblage Cpx + Mel -f- PI + Gr + Wol 8B + G is believed to be 
due to metastable growth and persistence of garnet (Gr). Cpx, clinopyroxene ; PI, plagioclase; 
Wolas, wollastonite solid solution; G, gas; L, liquid. Water content, if any, of the garnet was 
not determinable optically because of the small grain size. 



GEOPHYSICAL LABORATORY 



103 



solidus from the assemblage melilite + 
plagioclase -t-clinopyroxene+wollaston- 
ite to nepheline + plagioclase + clinopy - 
roxene + wollastonite. On the same figure 
are shown the assemblages found under 
■Ph 2 o = 2 kb. The beginning of melting 
is drastically lowered and the change 
of assemblage at the anhydrous solidus 
extends to more calcium-rich composi- 
tions. Additional assemblages involving 
a garnet (mainly grossularite) were en- 
countered at the most calcium-rich com- 
position at lower temperatures, as pre- 
dicted from the study of An 50 Ak 5 o 
(Yoder, this report). It is evident that 
melilite and plagioclase coexist in the 
presence of water at reasonable magma 
temperatures. Preliminary results at 
higher water pressures, e.g., 5 kb, indi- 
cate a drastic temperature rise in the 
stability of the garnet-bearing assem- 
blages, and the melilite assemblages do 
not appear. Appeal to such water pres- 
sures does not seem appropriate for the 
lavas but may account for the absence 
of melilite in the pyroclastics. The 
dilemma remains for the lavas on the 
basis of these experiments. 

A natural melilite (approximately % 
akermanite and y 3 soda melilite) and a 
natural plagioclase (close to An 50 ) were 
mixed in equal proportions and heated 
at 1 atm at a series of temperatures. No 
reaction was observed below 800 °C. At 
and above 850 °C, to the beginning of 
melting and through much of the melt- 
ing region, the assemblage was mostly 
melilite and plagioclase, with some di- 
opside and wollastonite solid solutions. 
Under hydrous conditions (Pn 2 o = 2 kb) 
the natural minerals reacted completely 
to clinopyroxene + plagioclase + wollas- 
tonite + nepheline at 700°, 800°, and 
900 °C. Melting of the synthetic com- 
position AbasAnasAkso begins at about 
970°C under Ph 2 o = 2 kb, and the meli- 
lite-bearing assemblage, whose plagio- 
clase and melilite compositions are be- 
lieved to be in accord with the natural 



minerals chosen (see equations below), 
is stable down to about 860 °C (Fig. 14). 
Apparently the additional constituents 
in the natural materials are the cause 
of the difference in results in the hydrous 
experiments, whereas rate problems pro- 
hibited attainment of equilibrium in the 
anhydrous experiments. 

The two principal assemblages ex- 
hibited in Fig. 14, Cpx + Pl + Ne 88 +Wol ss 
and Cpx + Mel + Pl+Wol ss , may be re- 
lated, with an initial composition of 
lAb:lAn:2Ak for example, as follows: 

2 Ab + 2An + 4Ak -» ( Ab + An) ss + 
(2Ak+Sm) 88 + (2Di+CaTs) 88 +Wo-» 
( 1 / 2 Ab+An) 8s + (4Di+CaTs) 88 + 

iy 2 Ne+4Wo, 

where Sm represents soda melilite; Di, 
diopside; CaTs, calcium Tschermak's 
molecule; Ne, nepheline; and Wo, wol- 
lastonite. All possible solid solutions are 
not adequately represented; however, 
the apparent incompatibility of nephe- 
line and melilite, which are found co- 
existing in other compositions in Ab-An- 
Ak (Fig. 15 of Schairer and Yoder, this 
report), is outlined. (Attempts to test 
these relationships by means of micro- 
probe analyses of the phases failed be- 
cause of the exceedingly small grain 
size.) The reactions postulated depend 
in part on solid solution of soda melilite 
in akermanite and of "plagioclase" in 
nepheline (see Fig. 17 of Yoder, this 
report). Both types of solid solution 
series are reasonable on the basis of 
analysis of natural minerals and phase 
relations in less complex synthetic sys- 
tems. From another viewpoint, desilica- 
tion of albite to nepheline and anorthite 
to calcium Tschermak's molecule brings 
about silication of the akermanite to 
diopside and wollastonite. The complexi- 
ties of the solid solutions do not yield 
an unambiguous solution to the rela- 
tionship of the two principal assemblages 
with the data at hand. 



104 



CARNEGIE INSTITUTION 



The Join Albite-Anorthite- 

Akermanite 

/. F. Schairer and H. S. Yoder, Jr. 

The phase-equilibrium diagram for 
the join albite-anorthite-akermanite is 
given here as Fig. 15. Data for the limit- 
ing system albite-anorthite were given 
by Bowen (1913), and additional data 
were obtained by Schairer (1957, p. 232, 
Fig. 35) in his studies of the system 
nepheline-anorthite-silica, which con- 
firmed exactly the liquidus curve of 
Bowen but slightly modified the solidus 
curve. Data for the limiting system 
anorthite-akermanite were given by 
deWys and Foster (1956). Data ob- 
tained in the present investigation show 
that anorthite and akermanite do not 
show simple eutectic relations, that 



crystals of a slightly aluminous diopside 
solid solution appear in some of these 
compositions, in addition to anorthite 
and a melilite (probably nearly but not 
exactly of akermanite composition), and 
that beginning of melting occurs through 
a range of temperatures below that of 
the "eutectic" composition anorthite- 
akermanite. Data for the limiting sys- 
tem albite-akermanite were obtained 
by Schairer and Yoder (Year Book 68, 
p. 69, Fig. 5; p. 70, Fig. 6) in their study 
of the joins albite-akermanite-diopside 
and nepheline-akermanite-albite. 

Twenty-eight compositions in the 
join albite-anorthite-akermanite (shown 
as black dots in Fig. 15) were prepared 
and studied by the method of quenching. 
Fields of melilite, diopside solid solution, 
and plagioclase appear on the liquidus 



Akermanite 

(2CaOMgO 2Si0 2 
I454±3° 



When all crystalline 
[Compositions to right of line EFGH ] 
I MeUPlag^Dis^Wosg } 

[ Compositions in area IHG 
1 Melt D^ +W 0ss 
[ Compositions in area IGF 
1 Mel+ Diss tWo ss »Ne ss 

Compositions to left of EFand below IF 
Diss+Woss* Ness* Plag 




Albile 10 

Na,0 ALO, 6Si0, 



40 50 60 

Weight per cent 
Fig. 15. Phase equilibrium diagram for the join albite-anorthite-akermanite at 1 atm 



Anorthite 
CaOAI 2 3 -2Si0 2 



GEOPHYSICAL LABORATORY 



105 



surface. Wollastonite solid solution ap- 
pears as one of the solid phases in all 
compositions studied at some tempera- 
ture below the liquidus, and crystals of 
a nepheline solid solution appear at 
some temperature below the liquidus in 
all compositions to the left of the dashed 
line EFG and below the dashed line IG 
(Fig. 15). There is only one piercing 
point in the join, which lies at D, at 
1228° ±3°C, at the composition albite 
15, anorthite 36, akermanite 49 wt %, 
where the univariant curve melilite + 
diopside solid solution + plagioclase + 
liquid pierces the join. 

One significant piece of information 
emerges from the quenching data on 
the pseudoternary join: Four solid 
phases are in equilibrium with a liquid 
over a considerable range of tempera- 
ture. In a quaternary system at 1 atm 
pressure, four solid phases with a liquid 
can coexist only at one temperature. The 
join albite- anorthite-akermanite is ex- 
hibiting quinary equilibrium. Many of 
the compositions studied show the pres- 
ence of the four crystalline phases meli- 
lite + diopside solid solution + wollaston- 
ite solid solution + plagioclase with a 
liquid phase over a range of temperature 
as much as 75°C. Other compositions 
studied show the presence of the four 
crystalline phases diopside solid solu- 
tion + wollastonite solid solution + plagi- 
oclase + nepheline solid solution with a 
liquid over a range of temperature as 
much as 25 °C. 

All the crystalline phases encountered 
in the join albite-anorthite-akermanite 
are complex solid solutions. The melilites 
are ternary solid solutions of akermanite, 
soda melilite, and a little gehlenite. The 
clinopyroxenes are aluminous diopsides. 
The wollastonites are solid solutions of 
CaSi0 3 and diopside with a possible 
small A1 2 3 content. The feldspars are 
plagioclases. The nepheline solid solu- 
tions probably have appreciable con- 
tents of albite or anorthite, or both. 
The compositions of none of the solid 
solutions are known precisely. We were 



not able to determine the compositions 
of the feldspars in equilibrium with 
melilites or with nephelines that coexist 
with diopsides, wollastonites, and liquids 
by means of the quenching method. The 
crystals obtained thus far are not of 
sufficient size to determine by means of 
the electron microprobe. 

Anorthite-Akermanite and Albite- 
Soda Melilite Reaction Relations 

H. S. Yoder, Jr. 

Plagioclase and melilite are stable to- 
gether in some metamorphic rocks but do 
not coexist in alkali igneous rocks. A 
study of a composition containing equal 
proportions by weight of anorthite (An) 
and akermanite (Ak) yielded results that 
bear on the limits of compatibility of 
these phases. In Fig. 16 it is seen that 
where anorthite (An) and melilite (Mel) 
coexist, a clinopyroxene (Cpx) and a 
wollastonite solid solution (Wol ss ) (not 
identified at 1 atm) are also produced 
from the bulk composition An 50 Ak 5 o. The 
reaction may be in principle: 

CaAl 2 Si 2 8 + Ca 2 MgSi 2 7 -» 

An Ak 

(CaMgSi 2 6 + Ca Al 2 Si0 6 ) S8 + 

Di CaTs 



CaSi0 3 

Wol 



(1) 



where Di represents diopside, and CaTs, 
calcium Tschermak's molecule. The sys- 
tem anorthite-akermanite is no longer 
considered to be binary (c/. deWys and 
Foster, 1956; Schairer and Yoder, this 
report). The wollastonite no doubt con- 
tains some clinopyroxene in solid solu- 
tion because the inversion of wollastonite 
to pseudowollastonite is not observed, 
wollastonite being stable some 65° to 
100 °C above the inversion of pure 
CaSi0 3 (Kushiro, Year Book 68, p. 84) . 
On the other hand, the reaction may also 
be, at least in part: 



Ak ->Ak ss + Di 8S + Wol 



(2) 



as described by Schairer, Yoder, and 



106 
10 



CARNEGIE INSTITUTION 



o 

-Q 
O 



700 




X X X X X X 



f-rP — f- 



800 



900 



1000 MOO 1200 

Temperature, °C 



1400 



1500 



Fig. 16. Pressure-temperature diagram for the bulk composition anorthite 50 akermanite 50 
by weight. Cpx, clinopyroxene ; Wol„a, wollastonite solid solution; An, anorthite; Mel, melilite; 
Gr, garnet; L, liquid; Xtls, various assemblages of some of the above-named crystalline phases. 
One-atm data from Schairer and Yoder (this report). 



Tilley {Year Book 65, pp. 217-219) 
wherein a complex solid solution, as yet 
undefined, of akermanite was formed. 
With increasing total pressure the 
relative proportions of anorthite and 
melilite (ss?) decrease and clinopyrox- 



ene +wollastonite ss increase, mainly by 
solid solution as indicated in equa- 
tion 1. At about 6^4 kb and higher, 
only clinopyroxene and wollastonite ss 
were formed in the region outlined in 
Fig. 16. The clinopyroxene solid solu- 



GEOPHYSICAL LABORATORY 



107 



tion (IDirlCaTs) suggested is rea- 
sonable, since calcium Tschermak's 
molecule itself is stable above about 
12 kb (Hays, 1966) and some solid 
solution occurs at 1 atm, although the 
amount is not yet known with certainty 
(Hytonen and Schairer, Year Book 60, 
pp. 125-141; Clark, Schairer, and de 
Neufville, Year Book 61, pp. 59-68). 
Due note was made of the warning of 
Hytonen and Schairer {Year Book 60, 
p. 139) in regard to metastable solid 
solutions: reversal of reactions was ob- 
tained. 

With lowering temperatures a sluggish 
reaction takes place, whereby the bulk 
composition An 50 Ak 50 is converted to a 
clinopyroxene and a grossularite garnet 
(Gross) . The reaction may be written in 
principle as : 

2CaAl 2 Si 2 8 + 2Ca 2 MgSi 2 T -» 

An Ak 

Ca 3 Al 2 Si 3 12 + (2CaMgSi 2 6 + 

Gross Di 

CaAl 2 Si0 6 ) ss (3) 

CaTs 

Solid solution of pyrope in the garnet no 
doubt occurs. At least 5 wt % pyrope 
was found in solid solution in grossularite 
at 10 kb by Yoder and Chinner {Year 
Book 59, p. 79) in the grossularite- 
pyrope-H 2 join. The products of equa- 
tion 1 are related to the products of 
equation 3 by reaction of a type repre- 
sented by 

CaAl 2 Si0 6 + 2CaSi0 3 -> 

CaTs Wol 

Ca 3 Al 2 Si 3 12 , (4) 

Gross 

thereby reducing the aluminous char- 
acter of the clinopyroxene. Similar de- 
crease in alumina content of pyroxene 
has been observed in other reactions in- 
volving the appearance of garnet (see 
Boyd and England, Year Book 63, pp. 
157-161). 

The upper stability limit of the garnet- 
bearing assemblage is close to that for 
pure grossularite itself (Hays, Year 
Book 65, p. 239). The high-pressure 



termination of the coexistence of aker- 
manite and anorthite is about 1 kb 
lower than that for forsterite and anor- 
thite (Kushiro and Yoder, 1966; Yoder, 
Year Book 65, p. 276) . 

A preliminary beginning-of-melting 
curve under hydrous conditions is also 
shown in Fig. 16. Further limitation of 
the field of coexistence of anorthite and 
akermanite is evident. 

These results indicate severe con- 
straints on assemblages containing 
anorthite and akermanite; a broad 
region of coexistence still persists, how- 
ever, within the pressure-temperature 
conditions where lavas and shallow in- 
trusives are to be expected. Factors other 
than elevated pressures or presence of 
water appear to be necessary to account 
for the discrepancy between field stud- 
ies of igneous rocks and laboratory ob- 
servations. The present results are, how- 
ever, in accord with the observation of 
melilite and plagioclase in metamorphic 
aureoles. 

Oxidizing conditions may tend to in- 
crease the clinopyroxene + garnet field 
at the expense of plagioclase and meli- 
lite. Recently, Huckenholz, Schairer, and 
Yoder {Year Book 66, pp. 335-346) 
showed that an andradite garnet and a 
diopside saturated with a ferri-Tscher- 
mak's molecule were stable up to 
1157°C, almost up to the beginning of 
melting. Presence of titanium would 
also increase the stability field of a 
clinopyroxene (titanaugite) and garnet 
(melanite) association (see Huckenholz, 
this report). Both ferric iron and ti- 
tanium are enriched in alkali igneous 
rocks relative to other major magma 
series. Addition of albite to the system 
initially enlarges the field of coexistence 
of plagioclase and melilite, then dimin- 
ishes the field as illustrated in Fig. 14 
(Yoder and Schairer, this report). 

Preliminary experiments on the re- 
action relations of albite-soda melilite, 
soda analogues of anorthite-akermanite, 
indicate that at pressures below about 
6.5 kb the assemblage nepheline + wol- 



108 



CARNEGIE INSTITUTION 




Ab Ne 
Weight per cent 

Fig. 17. Schematic representation of the assemblages immediately below (A) and above (B) 
about 6.5 kb for the system nepheline (Ne)-wollastonite (Wol)-albite (Ab). The 1-atm 
assemblage is nepheline solid solution + wollastonite + plagioclase according to Foster (1942) 
at the piercing point of the system. 



lastonite + plagioclase is stable. Above 
that pressure melilite + plagioclase are 
the stable phases. The relationships are 
illustrated schematically in Fig. 17; the 
solid solutions are complex and cannot 
be adequately defined. These experi- 
ments lead to the conclusion that the 
nepheline-wollastonite-diopside plane in 
the expanded basalt tetrahedron of 
Schairer and Yoder (Year Book 68, p. 
65), which prohibits the coexistence of 
albite and akermanite, may be pene- 
trated by a family of joins connecting 
albite with soda-rich melilites. The pro- 
hibition of the coexistence of an aker- 
manite-rich melilite and albite remains. 
The data now at hand do not yield 
a satisfactory solution to the melilite- 
plagioclase dilemma of the alkali igneous 
rocks. 

Experiments Bearing on the Origin 
of Anorthositic Intrusions 

R. F. Emslie * and D. H. Lindsley 

We report here the results of some ex- 
periments on natural and synthetic sys- 
tems that have a bearing on the origin 
of anorthosites. Although the high con- 
tent of intermediate plagioclase in 
batholithic anorthosites is explainable 

* Geological Survey of Canada. 



in part by plagioclase accumulation, the 
conclusion that the parent liquid was 
enriched in the components of inter- 
mediate plagioclase seems inescapable. 
Consequently, our experiments were di- 
rected toward investigating suitable 
mechanisms by which such enrichment 
might be accomplished. We reported last 
year (Year Book 66, pp. 497-480) work 
that led us to believe that the system 
albite-anorthite-diopside could not pro- 
vide an adequate model for the genera- 
tion of anorthositic magmas at crustal 
pressures. This year we have focused 
attention on compositions in the system 
albite-anorthite-forsterite at pressures 
up to 20 kb. We have also examined 
the P-T relations of a sample of the 
"chilled margin" of the Michikamau in- 
trusion, a large unmetamorphosed body 
in Labrador with characteristics inter- 
mediate between typical anorthosite 
complexes and stratiform mafic intru- 
sions (Emslie, 1965, 1968). The rock 
assemblage of the Michikamau intru- 
sion strongly suggests that basaltic 
magma played a dominant role in its 
development. Recent strontium and oxy- 
gen isotopic evidence (Epstein and 
Taylor, 1967; Heath, 1966) from wide- 
spread anorthositic intrusions also sup- 
ports arguments relating them to basaltic 



GEOPHYSICAL LABORATORY 



109 



magma. The approach adopted, there- 
fore, is to examine conditions under 
which basaltic liquids, or their deriva- 
tives, might be induced to crystallize 
large volumes of intermediate plagio- 
clase. In effect, this reduces to finding 
processes by which the alumina content 
and the Na/Ca ratio of basaltic liquids 
may be increased simultaneously. 

Experiments at pressures between 5 
and 20 kb were done in a solid-media, 
piston-and-cylinder apparatus under dry 
conditions. 

Composition An 52 . 5 Ab22.5Fo25 
[(Ab 70 Ab so ) s Fo 1 ] 

The composition An 5 2.5Ab 2 2.5Fo25 (mole 
ratio) lies in the plagioclase field at 



1 atm (Fig. 18, inset). It was chosen 
to illustrate boundary-curve shifts and 
reactions between forsterite and plagio- 
clase with increasing pressure. Starting 
materials included forsterite plus plagio- 
clase crystals, glass, and a high-pressure 
assemblage synthesized at 20 kb that 
contained clinopyroxene, plagioclase, 
spinel, and some forsterite (presumably 
residual). All runs were made in plati- 
num capsules. 

With increasing pressure, plagioclase 
is replaced by spinel as the liquidus 
phase, demonstrating that the spinel 
field is enlarged by application of pres- 
sure (Fig. 18). Up to about 19 kb, 
plagioclase joins spinel with decreasing 
temperature, but at higher pressure 



1700 



1400 



1000 







Pressure, kb 



Fig. 18. P-T diagram for the composition AnB^sAbs^Foa, (mole %). Inset at upper left is the 
liquidus diagram at 1 atm for the system forsterite-albite-anorthite (converted to mole percentage 
from Yearbook Book 65, p. 206) with the composition An52.5Ab 2 2.5Fo2s indicated by an X. 
One-atmosphere temperatures plotted in the figure are from the same source. Triangular symbols 
indicate that the starting material was a high-pressure assemblage containing cpx, pi, sp, and 
fo. Solid triangles indicate that fo grew and cpx and sp decreased. Open triangle indicates that 
cpx and sp grew and fo decreased. Symbols containing a solid circle are runs in which fo 
was present as residual phase; with repeated grinding and rerunning the fo decreased in amount 
and cpx and sp increased. L, liquid; sp, spinel; pi, plagioclase; cpx, clinopyroxene; opx, ortho- 
pyroxene; fo, forsterite; ab, albite; an, anorthite. 



110 



CARNEGIE INSTITUTION 



spinel is followed first by clinopyroxene 
and then plagioclase. Liquid + forsterite 
+ plagioclase is stable up to a maximum 
pressure of about 13 kb. At subsolidus 
temperatures the reaction forsterite + 
plagioclase <=± spinel + plagioclase + clino- 
pyroxene ( + orthopyroxene?) is indi- 
cated as a stippled zone of some finite 
but unknown width. Because plagioclase 
and the pyroxene are solid solutions, the 
reaction must proceed over a P-T range. 
The presence of orthopyroxene on the 
high-pressure side of the stippled zone is 
suspected but not verified. At higher 
pressures orthopyroxene is believed to 
disappear by entry into solid solution 
in the clinopyroxene. 

Fractional crystallization of spinel at 
high pressures is capable of driving the 
liquid to more plagioclase-rich compo- 
sitions and presumably to compositions 
richer in albite relative to anorthite be- 
cause spinel removes alumina and en- 
riches the liquid in silica. Addition of 
silica to the starting composition to pro- 
duce normative pyroxene would drasti- 
cally decrease the size of the liquid + 
spinel field, however, and it is unlikely 
that natural basic magmas can precipi- 
tate sufficient amounts of spinel to make 
this an effective mechanism in the pro- 
duction of anorthositic liquids. 

Some experiments were performed on 
the composition An 37 . s Ab 37 .5Fo 25 [ (An 50 
Ab 50 ) 3 Fo 1 ]. Plagioclase is the first phase 
to occur at the liquidus up to about 
11 kb, but at higher pressures spinel 
again takes its place. It is clear, how- 
ever, that if spinel crystallization were 
suppressed by a slightly more silica- 
rich starting composition, clinopyroxene 
would be present at the liquidus at pres- 
sures at least as low as 15 kb. We are 
presently determining the A1 2 3 con- 
tents of the high-pressure clinopyroxene. 
At present we assume that the A1 2 3 
contents are less than those in the co- 
existing liquids. Fractional crystalliza- 
tion of such clinopyroxene should be 
capable of increasing both the alumina 



content of the liquid and its Na/Ca 
ratio. 

The stippled subsolidus reaction 
boundary in Fig. 18 may be compared 
to curve A of Kushiro and Yoder (Year 
Book 64, Fig. 14) for the composition 
forsterite 1: anorthite 1 (mole). The 
comparison illustrates that addition of 
albite stabilizes the assemblage forster- 
ite + plagioclase to higher pressure. Pre- 
sumably troctolites, especially those with 
magnesian olivine, are stable through- 
out most or all of the continental crust. 
This conclusion is of importance in as- 
sessing the maximum depth of crystal- 
lization of some of the largest anortho- 
sitic plutons with olivine-plagioclase 
assemblages in Quebec and Labrador. 

Michikamau Chilled Margin 

A recently obtained sample of very 
fine-grained rock taken a few inches 
from the contact of the Michikamau in- 
trusion is interpreted as approximating 
a liquid composition that was chilled 
against the wall at the time of em- 
placement. An analysis and norm of the 
rock (EC66-21A5) are given in Table 3. 
Field evidence suggests that the bulk 
magma (liquid + crystals) was probably 
considerably richer in plagioclase than 
this sample (Emslie, 1965, 1968). The 

TABLE 3. Chemical Analysis and Norm of 
Michikamau Chilled Margin EC66-21A5 



Analysis 




Norm 




Si0 2 


47.56 


Qz 




A1 2 3 


18.33 


Or 


0.56 


Fe 2 3 


0.30 


Ab 


10.49 


FeO 


11.66 


An 


44.24 


MgO 


6.40 


Di 


10.44 


CaO 


12.07 


Hv 


29.11 


MnO 


0.21 


01 


0.86 


K 2 


0.08 


11 


2.28 


Na 2 


125 


Mt 


0.46 


P2O5 


028 


Ap 


0.66 


Ti0 2 


124 


Ct 


0.40 


H 2 


0.09 


Remainder 


0.18 


H 2 


0.09 







C0 2 


0.16 


Total 


99.68 


Total 


99.72 







Note: Analyst, J. L. Bouvier. 



GEOPHYSICAL LABORATORY 



111 



sample is clearly aluminous and could 
be described as a basalt with a high 
alumina content, as might be expected 
from its association with a plagioclase- 
rich rock assemblage. Assuming that 
the liquid was derived from the mantle, 
it is of interest to examine its phase 
relations under conditions similar to 
those at which it formed or to which it 
was subjected during its rise into the 
crust. 

Finely ground rock powder from the 
chilled margin and a glass prepared 
from it were used as starting materials. 
Most of the high-pressure experiments 
were run in platinum capsules. It is well 
known that iron-bearing silicates and 
liquids tend to lose iron to platinum 
containers. 

We evaluated the approximate extent 
of the iron loss by microprobe analysis 
of duplicate samples and their platinum 
capsules: for a sample of the Michi- 
kamau chilled margin run for % hour 



at 1340 °C and 10 kb (well within the 
liquid field) the Fe loss is 3-5% of the 
amount present, or about 0.4-0.6 wt % 
FeO. The loss of iron to the capsule re- 
sults in the release of oxygen, and thus 
the FeO/Fe 2 3 ratio, as well as the total 
iron content is lowered. Our results must 
be interpreted with these facts in mind; 
we believe, however, that though the 
iron loss affects the details of our dia- 
grams, it does not modify the general 
conclusions drawn from them. 

The experimental results are shown 
in Fig. 19. Plagioclase is the first crystal- 
line phase to appear on cooling at pres- 
sures up to 12 kb. From 12 kb to about 
16.5 kb clinopyroxene is the primary 
phase, and at higher pressures garnet 
crystallizes first. With decreasing tem- 
peratures these phases are joined by 
pyroxene (s) or plagioclase, or both. 

Early plagioclase crystallization rap- 
idly depletes the alumina content of the 
liquid, and it is clear that at pressures 



1400 



Liquid 





1000- 



900 



800 



gar+px+pl 



pl + px 



CZDL 

EE3l+ p i 

EEQL+cpx 

|L+gar + cpx 



L.+ pl+px 
L+ gar + px + pi 
gar+px+pl 
pl+px 



Pressure, kb 



20 



Fig. 19. P-T diagram for the Michikamau chilled margin EC66-21A5. One-atmosphere runs 
were made in evacuated silica tubes through the courtesy of J. F. Schairer. L, liquid; pi, 
plagioclase; cpx, clinopyroxene; gar, garnet; px, clinopyroxene + orthopyroxene. 



112 



CAENEGIE INSTITUTION 



below 12 kb fractional crystallization 
cannot enrich the liquid in alumina. At 
higher pressures, fractional crystalliza- 
tion of clinopyroxene should be capable 
of enriching the liquid in alumina and 
soda. Selective removal of garnet at still 
higher pressures, on the other hand, may 
be expected to cause alumina depletion 
of the liquid. At pressures at least up to 
20 kb, however, garnet is closely followed 
by clinopyroxene with decreasing tem- 
perature, and it is probable that frac- 
tional crystallization of garnet + clino- 
pyroxene could lead to enrichment of 
the liquid in components of intermediate 
plagioclase so long as the ratio of garnet 
to clinopyroxene is not high. The extent 
of this enrichment depends, of course, on 
the extent of the garnet + clinopyroxene 
phase volume and on the composition of 
the clinopyroxene. 

Of the crystalline phases that occur at 
and near liquidus temperatures in basaltic 
magmas at moderate to high pressures, 
clinopyroxene appears most likely to 
play an important role in crystal frac- 
tionation processes capable of producing 
anorthositic liquids. There is probably 
a pressure range — say 10-15 kb — at 
which clinopyroxene is enriched in Ca- 
Tschermak's molecule relative to j adeite ; 
its removal would enrich the liquid in 
albite relative to anorthite, as well as 
enriching it in alumina, because the py- 
roxene is lower in A1 2 3 than the liquid. 
Such clinopyroxene coexists with liquid 
or with garnet and liquid only at pres- 
sures expected near and below the base 
of the continental crust. Maximum pres- 
sures under which suitable fractionation 
could occur are limited by (1) the in- 
creasing amount of garnet that crystal- 
lizes early and (2) breakdown of the 
albite component of plagioclase, which 
also goes into solid solution (as j adeite) 
in clinopyroxene. An estimate of the 
upper pressure limit under which frac- 
tional crystallization of clinopyroxene 
or garnet plus clinopyroxene from ba- 
saltic liquids can lead to anorthositic 
compositions is approximately 20 kb, de- 



pending on compositions (see also Green, 
1966, p. 220). 

Syenites 
S. A. Morse 

Syenites and their extrusive equiva- 
lents trachytes appear to occupy a para- 
doxical position among igneous rocks. 
On the one hand, their small volume in 
the crust and frequent well-defined as- 
sociation with logically parental magma 
series render their derivation by frac- 
tionation from basalt quite credible, and 
in some cases the parent-daughter rela- 
tionship is clearly demonstrable. On the 
other hand, syenites and trachytes are 
critically undersaturated rocks in the 
sense of Yoder and Tilley {Year Book 
59, p. 68) and lie in projection on a 
critical undersaturation line (the alkali 
feldspar join) in petrogeny's residua 
system (Schairer, 1950). From all ap- 
parent evidence, these residual liquids 
lie on a thermal maximum relative to 
their neighbors, granite (or rhyolite) 
and nepheline syenite (or phonolite). 
What mechanism can cause a magma to 
remain critically saturated throughout 
its fractionation history and terminate 
on a thermal maximum instead of frac- 
tionating off toward either oversaturated 
or undersaturated compositions? The 
question is fundamental, not only be- 
cause of the economic importance of 
syenites, which often contain unusual 
abundances of valuable metals and rare- 
earth elements, but also because one or 
more basic concepts of magmatic frac- 
tionation have apparently escaped our 
notice. 

Initial investigations during the past 
year have helped to define the problem 
and narrow the working hypotheses to a 
manageable number. Aside from a char- 
acterization of the natural occurrences, a 
reinvestigation has been made of the 
alkali feldspar-water system at 5 kb, 
discussed elsewhere in this report, and 
the syenite-nepheline syenite portion of 
petrogeny's residua system has been 



GEOPHYSICAL LABORATORY 



113 



investigated at 5 kb water pressure. 
Studies in the system diopside-albite- 
orthoclase have been begun at 1 atmos- 
phere and at 5 kb water pressure. 

Characterization of Syenites 

Syenites are rocks dominated by lime- 
poor feldspar and contain little or no 
quartz or feldspathoid ; by addition of 
these minerals they grade into quartz 
syenite or nepheline syenite, respectively. 
In its simplest form, the mafic mineral- 
ogy is essentially that of basalts: augite, 
olivine, and Fe-Ti oxide, although of 
course the first two are richer in ferrous 
iron than those in basalts. Numerous 
natural examples support this character- 
ization of syenite, among them the 
layered intrusions of Labrador (Kigla- 
pait, Michikamau) and southwest Green- 
land (e.g., Kungnat). Trachytes occur 
in which members of the augite- ferro- 
augite series, olivine, and Fe-Ti oxide 
are the dominant mafics, although par- 
tially resorbed biotite phenocrysts are 
usually present as well. 

The Labrador and Greenland syenites, 
and many tr achy tic suites, show a single 
hypersolvus feldspar of the plagioclase- 
anorthoclase series dominating the frac- 
tionation path. A second, K-rich feldspar 
may appear in the very late stages, but 
attainment of the two-feldspar boundary 
would not appear to be critical to syenitic 
fractionation. 

In the plutonic examples just men- 
tioned, the assemblage fayalite + magne- 
tite + alkali feldspar is common, and the 
oxygen fugacity defined by this assem- 
blage is presumably very close to that 
defined by fayalite + magnetite + quartz 
(FMQ). The oxygen fugacity of the 
magmas during fractionation is prob- 
ably well modeled by the FMQ buffer 
curve (Morse and Stoiber, 1966). 

In summary, it appears that a rea- 
sonable model for the fractionation trend 
from basaltic to syenitic liquids can be 
generated by consideration of the four 
solid solution series, feldspar, augite, 



olivine, and magnetite, crystallizing un- 
der conditions of oxidation corresponding 
to the FMQ buffer. Hydrous, peralkaline, 
and peraluminous syenites may be 
treated as special cases requiring further 
explanation. 

Syenitic Trends 

Within the syenitic kindred there exist 
fractionation trends toward both over- 
saturation and undersaturation with 
silica; examples of these can be found 
with any of the following mafic mineral- 
ogies: olivine -f sodic pyroxene, olivine + 
ferroaugite, ferroaugite alone, and sodic 
pyroxene alone, all with Fe-Ti oxides. 
The existence of such trends and their 
apparent indifference to mafic mineral- 
ogy suggest their origin by fractionation 
of feldspar, a process that is evidently 
ineffectual early in the fractionation 
history. 

Most hydrous syenites (those contain- 
ing amphibole or biotite) in the Green- 
land-Labrador province have silica-over- 
saturated trends, which may be ac- 
counted for by oxidation of normative 
olivine during fractionation due to ele- 
vated Ph 2 o- Fractionation of amphibole 
or biotite itself could also have the effect 
of silica enrichment. 

Projection of the normative salic com- 
ponents of chemical analyses into petro- 
geny's residual system (nepheline-kalsil- 
ite-silica) has often served to illustrate 
the correspondence between rock com- 
positions and experimentally determined 
melting minima, as for example with 
granites (Tuttle and Bowen, 1958; Luth, 
Jahns, and Tuttle, 1964) and nepheline 
syenites (Hamilton and MacKenzie, 
1965). As the plots of the latter authors 
show, however, the example fails with 
syenites, a fact that should prove to be 
very instructive. The relationships of 
81 modern analyses of syenitic rocks 
from the Gardar province of Greenland 
(Watt, 1966) are summarized in Fig. 20, 
which shows a strong maximum on the 
feldspar join at Or 40 , with subsidiary 



114 



CAENEGIE INSTITUTION 




Weight per cent 

Fig. 20. The salic components of 81 natural syenites from the Gardar province, Greenland 
(Watt, 1966), projected from An onto the system Ne-Ks-Sil. The diagrams, classified from 
top to bottom by decreasing anorthite content, illustrate the trend of syenites in the ternary- 
feldspar plane toward the maximum near Or«> and then away from the feldspar join in the low-Ca 
residua system (bottom). 



trends toward granite and nepheline 
syenite. By contrast, the minimum or 
eutectic in the pure feldspar join, as 
determined experimentally, varies from 
near Or 35 at 1 bar to less than Or 30 at 
5 kb Ph 2 o- These natural syenites, there- 
fore, are not well modeled by alkali feld- 
spar alone, as should also be apparent 
from their far from modest mafic and 
anorthite content. 

Anorthite in the feldspar should cause 
a shift toward Or in projection, judging 
from the trend of the two- feldspar 
boundary in the ternary feldspar system 
(e.g., Yoder, Stewart, and Smith, Year 
Book 56, pp. 206-214). Addition of 
anorthite alone should not generate a 
strong maximum when projected onto 



the alkali feldspar join but rather a 
continuum of syenite compositions along 
the join. Moreover, for reasons stated 
above, the two-feldspar boundary seems 
of little importance except in the last 
stages of syenitic differentiation. The 
persistent syenitic maximum near 0r 40 
suggests a crystal <=± liquid control, and 
this is probably generated by the conflu- 
ence of the boundary surfaces of plagio- 
clase, ferroaugite, fayalitic olivine, Fe-Ti 
oxide, and liquid. Although the calcic 
pyroxene might be expected to be domi- 
nant in this respect because of its 
stabilization of Ca-bearing plagioclase 
in place of pure albite (Schairer and 
Yoder, 1960), preliminary experiments 
in the system diopside-albite-orthoclase 



GEOPHYSICAL LABORATORY 



115 



(see below) suggest movement of the 
trace of the feldspar minimum in the 
right direction but somewhat short of the 
required magnitude. 

There are thus at least two parts to 
the syenite problem: first, the mainte- 
nance of critical undersaturation during 
fractionation, and second, the generation 
of liquids that project toward Or from 
the alkali feldspar minimum. The two 
may not be unrelated. 

Experimental Evidence 

The most obvious question that might 
be asked about syenitic liquids is 
whether or not they do indeed occupy 
a thermal maximum relative to granite 
and nepheline syenite. A qualified answer 
to this question is given by current 
experimental work in petrogeny's residua 
system at 5 kb Ph 2 o- 

Nepheline-kalsilite- silica at 5 kb ~Ph 2 o- 
The silica-oversaturated portion of this 
system has been reported in detail by 
Luth, Jahns, and Tuttle (1964). The 
alkali feldspar-water join at 5 kb was 
studied by Yoder, Stewart, and Smith 
(Year Book 56, pp. 206-214) and re- 
studied by Morse during the past year 
(discussed elsewhere in this report) . The 
silica-undersaturated portion has been 
studied at 1 kb Ph 2 o by Hamilton and 
MacKenzie (1965), and portions of the 
bounding system Ab-Ne-H 2 involving 
analcite were studied by Peters, Luth, 
and Tuttle (1966). Part of the interest 
in the 5-kb isobaric section reported here 
lies in the determination of the alkali 
feldspar solvus for undersaturated com- 
positions, and part of it lies in the ap- 
pearance of analcite and the assemblage 
analcite-K feldspar at subsolidus tem- 
peratures. 

The liquidus diagram for the residua 
system at 5 kb Ph 2 o as presently known 
is shown in Fig. 21. The region involving 
leucite was not investigated in this study. 
The beginning of melting of nepheline- 
syenitic compositions occurs at 635° ± 
3°C at an isobaric eutectic projecting 



very near the 1-kb minimum and the 
1-atm reaction point, as sketched in 
Fig. 22. This implies that fractiona- 
tion at almost any geologically rea- 
sonable water pressure and, less cer- 
tainly, most reasonable total pressures 
under water-undersaturated conditions 
should lead to nepheline syenites of 
limited compositional variation, as ap- 
pears to be true in nature (Hamilton and 
MacKenzie, 1965, Fig. 4). 

A series of isobaric, isothermal sec- 
tions, shown in Figs. 23-27, illustrates 
the equilibria encountered by a nephe- 
line-syenitic liquid in its crystallization 
and subsolidus history. The tie lines in 
the 665 °C section are not schematic 
but are determined by X-ray methods 
for feldspar and by electron-probe analy- 
sis of nepheline. The probe analyses 
show the amount of silica dissolved in 
nepheline at this pressure to be measur- 
able, although appreciably less than 
found at lower pressures (Hamilton and 
MacKenzie, 1960, p. 61), as expected. 
The analyzed series lies within experi- 
mental error on a straight line from 
Si0 2 = 2.6% on the Ne-Si0 2 join to 
Ks = 40% on the Ne-Ks join. There is 
no evidence for a two-feldspathoid re- 
gion. 

It is to be noted that the liquid field 
extends to the Ne-Ab sideline at 665°C; 
the assemblage Ne ss -Ab ss does not ap- 
pear. On further cooling from 665 °C, 
the liquid field shrinks away from the 
Ne-Ab sideline, but as it does so, the 
assemblage analcite-liquid becomes 
stable, again preventing the assemblage 
Ns S3 + Ab ss , as shown in the 640°C iso- 
thermal section. At some temperature 
lower than 640 °C but greater than 
635°C, the assemblage Ab + L must be- 
come unstable through the reaction Ab + 
L = Anl + Kf (K feldspar) +G, since the 
beginning of melting occurs at a eutectic 
within the Anl + Kf + Ne field, which 
dominates much of the diagram at sub- 
solidus temperatures (see the 600 °C 
isothermal, isobaric section, Fig. 25). 
This necessitates a narrow field of anal- 



116 



CARNEGIE INSTITUTION 



P,, ^=5kb 



830° 

Ne 

NaAISi0 4 




20 



30 



40 50 60 

Weight per cent 



70 



90 



Ks 
KAISi0 4 



Fig. 21. Liquidus diagram of petrogeny's residua system at 5 kb Ph 2 o, omitting relationships 
involving leucite. The granite portion is replotted from Luth, Jahns, and Tuttle (1964), and 
data on Ne-Ab-H 2 are incorporated from Peters, Luth, and Tuttle (1966). Points show 
compositions studied in this work. 



cite on the liquidus diagram and a re- 
action point R where the two-feldspar + 
liquid boundary curve meets the Anl + 
Ab + L boundary curve (see inset in 
Fig. 21) . On cooling below R, liquids pro- 
duce analcite and K feldspar, until the 
eutectic is reached, at which point 
nepheline joins the assemblage and crys- 
tallization is completed isothermally. 
It is noteworthy that the commonly 
observed assemblage Ne S8 + Ab SB has no 
field of stability at 5 kb Ph 2 o5 therefore 
rocks that show this assemblage must 
have formed under some lower water 
pressure. The termination of the Ne S9 + 



Ab ss stability field in the hydrous sys- 
tem occurs at a quaternary invariant 
point involving Ab+Kf+Ne + Anl + 
L + G at some pressure not far below 
5 kb and a temperature not far above 
635°C. Data are as yet insufficient to 
permit accurate delineation of the uni- 
variant curves around this invariant 
point, but experiments between 2 and 
10 kb are underway for this purpose. 

From the data incorporated in Fig. 21, 
a profile can be drawn along the two- 
feldspar + liquid boundary curve from 
the granite eutectic through the feldspar 
join and to the nepheline syenite eutectic. 



GEOPHYSICAL LABORATORY 

SiO. 



117 




Weight per cent 

Fig. 22. Positions of nepheline syenite minima in petrogeny's residua system at 1 atm and 
1 and 5 kb Ph 2 o. The 1-atm boundaries are from Schairer (1950) ; the 1-kb boundaries are from 
Hamilton and MacKenzie (1965); and the 5-kb boundaries are from the present study. The 
1-atm minimum lies very near the leucite reaction point. The dashed circle is the contour 
around the statistical maximum of nepheline syenite analyses from Hamilton and MacKenzie 
(1965). 



This profile, projected parallel to lines 
of equal silica content onto the line 
Ne/Ks=l, is shown in Fig. 26, along 
with comparable projections at lower 
and higher pressures. The profiles at 
1 atm and 1 kb Ph 2 o are those of the 
minimum liquidus temperatures on the 



feldspar + liquid boundary surface. The 
profile at 10 kb in the nepheline syenite 
portion is supported by a few recon- 
naissance experiments made during the 
current study. It is clear from these pro- 
files that haplosyenites represented by 
petrogeny's residua system do indeed lie 




40 50 60 

Weight per cent 



Ks 



Fig. 23. Isothermal, isobaric section, at 665° C and 5 kb, of the nepheline syenite region of 
Ne-Ks-Si0 2 -H 2 0. The tie lines are not schematic but determined directly by X-ray composition 
of feldspars (except Ores) and electron-probe analyses of nephelines and the most potassic 
feldspar. Fs, feldspar; Kf, K feldspar. 



118 



SiO. 



CARNEGIE INSTITUTION 

SiO, 



P H2 o=5kb 
640°C 



Anl+L 




Weight per cent 



Fig. 24. Section at 640°C, 5 kb. Most of the tie lines are schematic (dashed). The section 
depicts the narrow field of analcite + L + G that prohibits the assemblage Ne + Ab at this 
pressure. Kf, K feldspar. 



on a maximum relative to nepheline 
syenite and granite throughout the range 
of geologically reasonable water pres- 
sures, as well as at 1 atm. Water pres- 
sure has a more severe effect on the 
beginning of melting of nepheline syenite 



than on that of granite, but there is no 
reason to believe that a water-under- 
saturated condition would raise the 
melting minimum of either of these 
above that of syenite. It seems clear that 
the syenite problem cannot be resolved 



600°C 




40 50 60 

Weight per cent 



70 



90 



Ks 



Fig. 25. Section at 600°C, 5 kb. Tie lines are schematic where dashed. The extent of the 
analcite field on Ne-Ab is taken from Peters, Luth, and Tuttle (1966). This essential configuration 
persists up to the beginning of melting at 635° ± 3°C, which takes place within the Anl + Ne + 
Kf+G field. Between 635° and 640°C the Anl + Kf tie line is broken by Ab + L; Ne + Ab 
never occurs at 5 kbs. Fs, feldspar; Kf, K feldspar. 



GEOPHYSICAL LABORATORY 



119 



1100 



!000 



900 



800 



700 



600 



Hamilton & 
MacKenzie 1965 



.present work 



1 T 




atm 

I 
Schairer, 1950 




TuttleaBowen 
1958 



Luth, Johns, Tuttle 1964 . 



10 20 

-Ne, Ks 



30 40 50 60 70 

Weight per cent Si0 2 



80 90 

SiOo- 



Fig. 26. Profiles along minima in petrogeny's residua system, projected parallel to lines of 
equal Si0 3 content onto the line Ne/Ks = l. Syenites (on the feldspar join) always appear 
to lie on maxima relative to nepheline syenites and granites. 



without consideration of the mafic com- 
ponents. 

Diopside-albite-orthoclase at atmos- 
pheric pressure and p,t 5 kb Ph 2 o- The 
addition of diopside to the alkali feld- 
spars is a logical first step in assessing 
the effect of mafic components on 
syenites, since thereby calcium is auto- 
matically introduced to the feldspar (the 
plagioclase effect of Bowen, 1945). On 
the other hand, the exchange of Ca and 
Al between feldspar and diopside renders 
this a nonternary join in a higher-order 
system, and so it is difficult to treat. 



The join forms a basis, however, for 
eventual studies including FeO, which 
may be expected to exert a feedback 
effect by consumption of excess silica 
to form olivine. 

Glasses in Di-Ab-Or were prepared 
with J. F. Schairer, who also carried 
out the 1-atm experiments. The composi- 
tions chosen for study so far are shown 
in Fig. 27. On the basis of preliminary 
work, it can be reported that the maxi- 
mum observed by Schairer and Yoder 
(1960) on the Di-Ab join persists well 
into the "syenitic" region of the Di- 



120 



CARNEGIE INSTITUTION 



Ab 




Di 



Di s 




i-eiasppr ^ ^ ^^ = ~- 




—^ Lc 



Weight per cent 



Or 



Fig. 27. A portion of the join Di- Ab-Or, showing compositions selected for study, a 
preliminary field boundary, and isotherms. Data are incorporated for Di-Ab from Schairer and 
Yoder (1960) and for Ab-Or from Schairer (1950). 



Ab-Or plane. The beginning of melting 
at 1 atm is not far from 1020°C, which 
is about the temperature of the nepheline 
syenite minimum at 1 atm and about 
40° below the alkali feldspar minimum 
at 1 atm. This does not reverse the 
thermal relations between nepheline 
syenite and syenite, since a comparable 
lowering of the beginning of melting is 
to be expected with addition of diopside 
to nepheline-syenite mixtures. The be- 
ginning of melting at 5 kb Ph 2 o lies 
above 650°C, still well above the 635° ± 
3°C nepheline syenite eutectic in the 
residua system. 

Feldspars 

iS. A. Morse 

Alkali Feldspar-Water at 5 kb 

The solvus in the alkali feldspar sys- 
tem (NaAlSisOg-KAlSisOg) is of poten- 
tial importance in unraveling the P-T 
history of many rocks and forms the 
basis of our understanding of the even 
more complex and useful ternary feld- 
spar solvus. Repeated investigations at 
elevated water pressures (such as those 
of Bowen and Tuttle, 1950; Orville, Year 
Book 58, pp. 118-121; Yoder, Stewart, 
and Smith, Year Book 56, pp. 206-214; 
and Luth and Tuttle, 1966) have given 
somewhat diverse results. 



In the present study the liquidus was 
determined with 3-day runs to ensure 
nucleation. Repeated runs were made at 
5° intervals over much of the critical 
region from Ab to Or 40 (wt %). The 
shape of the liquidus bears on whether 
the beginning of melting occurs at an 
isobaric minimum or a eutectic, which 
in turn bears on whether the solvus 
straddles that composition (in projec- 
tion) which has no melting interval. Sub- 
stantial melting intervals were found 
(Fig. 28) for all compositions studied, 
and the liquidus slopes imply a eutectic 
rather than a minimum, as found by 
Yoder, Stewart, and Smith {Year Book 
56, p. 208, Fig. 38) . The eutectic falls at 

vJr28.5±0.5- 

During runs of 3 days or even longer, 
equilibrium is not achieved within the 
crystal + liquid + gas loops. The compo- 
sition of feldspar, as determined by the 
20T spacing, is sensitive to the bulk 
composition at a given temperature, 
being near the correct solidus composition 
only for the bulk composition nearest the 
solidus, and progressively nearer the liq- 
uid composition as the bulk composition 
approaches that of the liquidus. The ratio 
of crystals to liquid is simultaneously 
deranged, and the lever rule cannot be 
used reliably. Knowledge of this effect is 
important in interpretation of solvus 



GEOPHYSICAL LABORATORY 

900 r 1 1 1 



800 



758 £ 3 



O 



8. 

E 
|2 



700 



600 




500, 



121 



(876) 



30 40 50 60 

Weight per cent 



90 



100 

Or 



Fig. 28. Revised equilibrium diagram of the water-saturated alkali feldspar join at 5 kb. 
The diagram is a projection from H 2 0. The K-rich liquidus and solidus portions incorporate 
results of Yoder, Stewart, and Smith (Year Book 56). Arrowheads along the Ab solvus limb 
indicate X-ray feldspar compositions of reactions from homogeneous Oris glass (outward) and 
mixtures of Ab + Ores in Or i5 bulk composition (inward) . Squares denote bulk compositions 
of runs outside the solvus that give feldspar compositions indicated by a large X. These runs 
were used as calibrations for X-ray compositions. Kf-limb arrowheads pointing outward are 
from initial glasses of Oreo, Ores, Or 7 o, and Or 7 5 compositions. 



runs as well, since sensitivity of feld- 
spar compositions to bulk compositions 
persists below the solidus. 

In order to confirm the melting in- 
tervals observed initially, it was neces- 
sary to establish accurately the begin- 
ning of melting, with starting materials 
crystallized at 5 kb in the range 600°- 
680 °C. It is possible to detect very small 
traces of glass in well-crystallized feld- 
spar, and these appeared first at 695 °C 
in bulk compositions Or 15 , Or 25 , and Or 30 . 
The compositions Or 20 and Or 35 showed 
no sure signs of melting until 705 °C. 
Compositions that exhibited premature 
beginning of melting all showed quartz 
when crystallized below about 650 °C 



and no quartz above that temperature; 
apparently the initial liquid lay in the 
ternary system Or+Ab + Qz + H 2 0, not 
on the feldspar join. Significantly, at 
least two of the low-melting starting 
materials (Or 15 and Or 30 ) had been 
ground for 1 hour in agate, and this very 
likely accounts for the excess silica. The 
beginning of melting in the alkali-feld- 
spar join at 5 kb P H o is taken as 703° ± 
2°C. 

Melting relations and solvus both 
derive support from concurrent studies 
in the nepheline syenite portion of petro- 
geny's residua system, nepheline-kalsil- 
ite-silica, discussed elsewhere in this re- 
port. The liquidus surface in the residua 



122 



CARNEGIE INSTITUTION 



system must, of course, be continuous 
with that of the feldspar join itself. The 
liquidus curves from the hydrous quater- 
nary system are shown in Fig. 29 as 
serial sections through the silica-under- 
saturated liquidus surface at 5% incre- 
ments of silica content and are com- 
pared with the feldspar liquidus from 
Fig. 28. The slopes are consistent with 
the liquidus slopes suggested for the 
join. 

Determination of the albite-rich limb 
of the solvus is difficult, owing to slow 
reaction rates, metastability, and pos- 
sibly local equilibrium in the form of 
zoning of crystals. Because reaction rates 
are presumably enhanced in the pres- 
ence of liquid, present studies were made 
initially in the feldspar + liquid + gas 
portions of the residua system. These 
results consistently gave a more sodic 
albite-rich limb than previous deter- 
minations, except the alkali-excess run 
of Luth and Tuttle (1966). Moreover, it 
was noted in a detailed study of the 
5 kb, 665 °C section in the residua system 
that bulk compositions nearer the soda 
feldspar termination of the four-phase 
Na feldspar + K feldspar + liquid + gas 
field gave more Ab-rich values than 
those bulk compositions richer in K, 
just as in the melting loop of the feld- 




20 30 

Weight per cent Or 

Fig. 29. Traces of liquidus surface at 5 kb 
Ph 2 o of the alkali feldspar join (top) and the 
nepheline syenite region at various silica levels 
in Ne-Ks-Sil. The latter four curves are 
projected back onto the alkali feldspar join at 
constant Ks content. 



spar join. This result suggested a princi- 
ple, borne out by subsequent experience: 
in determining a solvus by synthesis 
from glass, those bulk compositions 
nearest the limb to be determined give 
the best approximations to the stable 
(binodal) solvus, determined by the 
X-ray composition method. 

This principle was carried over to 
studies in the feldspar join to test 
whether the excess alkali produced dif- 
ferent results because of solid solution 
off the feldspar join — or nonstoichiom- 
etry — or perhaps because of improved re- 
action rates. The problem was ap- 
proached in several ways. Starting ma- 
terials of various bulk compositions were 
made up of mechanical mixtures of pure 
Ab glass and Or 95 glass, on the theory 
that conversion from soda-rich glass to 
crystals should not overstep the stable 
solvus in the direction of K. If this 
assumption were correct, the runs made 
with these mixed glasses should define 
the Na limit of the possible positions of 
the stable solvus. Results of such runs 
are shown in Fig. 28 as arrows, indicat- 
ing reaction of Ab glass + Or 95 glass to 
feldspar. Arrows pointing outward from 
inside the solvus indicate the X-ray 
composition of feldspar (Orville, 1963) 
formed from homogenous glass of the 
composition Or 15 . The difference between 
feldspar composition and bulk composi- 
tion was calibrated at several tempera- 
tures from runs outside the solvus. The 
combined arrows form a synthesis 
bracket, and the resulting solvus limit is 
consistent with both the eutectic-like 
liquidus surface and results in the pres- 
ence of alkali-excess liquids in the 
residua system. It is also consistent with 
the 5-kb alkali-excess experiments of 
Luth and Tuttle (1966) and, by extra- 
polation, with those of Orville (1963) 
at 2 kb. Similar synthesis brackets are 
being obtained for the K-rich limb of 
the solvus. 

It is noteworthy that in certain ex- 
periments the results from mechanical 
mixtures of Ab glass and Or 95 glass over- 



GEOPHYSICAL LABORATORY 



123 



lap those from homogeneous starting 
materials of the same composition, show- 
ing that reaction from end-member 
glasses does not reliably suggest the 
maximum breadth of the solvus. Both 
mechanical mixtures and homogeneous 
glasses having bulk compositions further 
inside the two-feldspar region gave feld- 
spars that also lay further inside. An 
example of this dependence of feldspar 
composition on bulk composition is il- 
lustrated in Fig. 30. This suggests (1) 
that only those bulk compositions lying 
very closely inside a solvus limb will 
give feldspar compositions approaching 
the stable value and (2) that reactions 
from pure Ab glass or Or 95 glass to feld- 
spar cannot be regarded as reversals in 
the rigorous sense. Reversals with the 
use of crystalline starting materials are 
in progress. 

Although the melting relations of feld- 
spar under water-saturated conditions at 
5 kb may rarely be relevant to the crys- 
tallization of natural magmas, the sub- 
solidus relations, barring appreciable ef- 
fects of leaching, should correspond to 
natural water-undersaturated conditions 
at 5 kb. Thus the effect of pressure on 
the solvus can be qualitatively evalu- 



& 



7(24hr.) 



T=665°C, P H2 o=5kb 
t = 2 weeks (except run 7) 
Ab limb 

A bulk comp.; ><fs. comp 
apparent limb composition 



glass 



<ryk 



~p^r 



TV 



10 20 

Weight per cent Or 



Fig. 30. X-ray compositions of feldspars at 
665°C, 5 kb Ph 2 o, produced from different start- 
ing materials, showing effect of bulk compo- 
sition on the apparent solvus limb composition. 



ated from inspection of the present work 
and the various 2-kb studies cited above. 
Estimation of the critical temperature, 
however, has proved difficult at 5 kb, 
since it requires data on the water- 
deficient crystal + liquid region. Runs 
with crystalline starting materials sug- 
gest a critical temperature of 730° ± 
10°C. 

Feldspar Stoichiometry 

Quartz has been observed in runs at 
630° and 600 °C from some starting ma- 
terials in the alkali feldspar-water sys- 
tem at 5 kb. Runs at and above 665 °C 
on the same starting materials show only 
feldspar. It is therefore possible that 
the feldspars from two different tempera- 
ture regions contain different amounts 
of silica and, more specifically, that the 
feldspars found in the presence of quartz 
are saturated with silica but those syn- 
thesized at higher temperatures are not 
demonstrably saturated. This possibility, 
once again, raises serious questions re- 
garding the stoichiometry of feldspars; 
The subject has been discussed recently 
by Luth and Tuttle (1966) , and Carman 
and Tuttle (1968) have offered evidence 
suggesting the presence of excess silica 
in sanidines from natural rhyolite. 

The experimental results are open to a 
number of interpretations. Leaching can- 
not be the sole cause of the quartz, unless 
the absence of quartz in higher-tempera- 
ture runs is due to increased solubility 
of silica in the vapor. This explanation 
is unlikely, since nothing resembling 
quench-quartz or alkali-silicate glass 
balls was found in the charges, and since 
experiments made with purposely intro- 
duced quartz (4%) show quartz at 
680 °C. An alternative explanation is 
that the starting materials are off compo- 
sition, owing either to loss of alkalies 
during preparation or to lengthy grind- 
ing in agate, and therefore lie in the 
ternary system Ab-Or-Q instead of the 
feldspar join. The fact that some glasses, 
ground only for short periods in agate, 



124 



CARNEGIE INSTITUTION 



do not produce quartz at 600 °C strongly 
suggests that this is the correct explana- 
tion and that the quartz seen at low 
temperatures is indeed dissolved in the 
feldspar at higher temperatures. 

If the feldspars in the residua system 
do indeed lie off the feldspar join at sub- 
solidus temperatures, some interesting 
consequences follow. Pure feldspars 
would exist only at the liquidus of the 
Ab-Or system. The solvus reported here 
for the "pure" system would in fact be 
that applicable to silica-saturated gra- 
nitic rocks. The true solvus for ideally 
stoichiometric feldspars would be some- 
what narrower than depicted here. Co- 
existing alkali feldspars in undersatu- 
rated rocks would have compositions 
differing (perhaps only slightly) from 
those in oversaturated rocks at the same 
T and P of formation. Data from this 
study and from the silica-undersaturated 
portion of petrogeny's residua system 
(elsewhere in this report) , as well as the 
two alkali-excess experiments of Luth 
and Tuttle (1966) at 5 kb, provide an 
inconclusive test of these possibilities. 
For virtually every alkali-excess run 
that can be compared with the presently 
determined solvus in the "pure" join, the 
alkali-excess points fall inside the solvus 
rather than outside. The results are 
nearly identical for 600° and 665 °C on 
the Na-rich limb, and for 665° and 
680° C on the K-rich limb. The devia- 
tions at 680 °C for the Na-rich limb and 
600 °C for the K-rich limb may be due 
to bulk composition effects similar to 
those found in the join: in the residua 
system it is not possible to choose com- 
positions in the two-feldspar + liquid 
field with Na:K ratios as near those of 
the feldspars as in the join itself. The 
600 °C equilibria in the undersaturated 
portion of the residua system are further 
complicated by the presence of analcite. 
As a result, the solvus deduced from 
the undersaturated part of the residua 
system and that determined for composi- 
tions ostensibly in the join, but con- 



taining a slight excess of silica, are not 
demonstrably different. 

Ternary Feldspars 

As shown schematically by Yoder, 
Stewart, and Smith (Year Book 56, Fig. 
46, p. 213) , the height and configuration 
of the feldspar solvus change drastically 
with addition of anorthite. The solvus 
crest rises very steeply with increasing 
An until it intersects the ternary feld- 
spar solidus. The shape of the solvus in 
the ternary T-X prism is known as yet 
only inexactly, however, and chiefly from 
study of natural feldspar chemical 
analyses. The experimental study of se- 
lected natural material offers certain ad- 
vantages in assessing the effects of tem- 
perature and water pressure simultane- 
ously. 

During the past year, long runs made 
on a natural mesoperthite at 500 bars 
water pressure showed a solvus crest at 
about 920°C, a temperature decisively 
of interest for the origin of mesoperthite- 
bearing rocks and their hypabyssal 
equivalents, anorthoclase-bearing rocks. 
The sample studied was an analyzed 
bulk feldspar separate from a ferro- 
syenite of the Kiglapait layered intru- 
sion, Labrador. The separate consists 
largely of mesoperthite, with small 
amounts of discrete oligoclase and K 
feldspar; the bulk analysis gives Or 29 . 6 
Ab 62 .2An 8 . 2 . The mesoperthite consists of 
sharply bounded, subparallel-to-taper- 
ing lamellae of oligoclase and K feld- 
spar, about 2 n wide. Textures in the 
rock indicate that this feldspar crystal- 
lized as a subhedral, homogeneous single 
phase, which subsequently exsolved. The 
slow cooling of this relatively deep- 
seated layered intrusion permitted a 
striking degree of internal equilibration 
within the former single crystals: X-ray 
compositions (projected from An) for 
the present lamellae are Or and Or 89 , 
although of course the Or is not pure 
albite, but around An 12 , if it is assumed 
that the K phase is essentially An free. 



GEOPHYSICAL LABORATORY 



125 



Mesoperthite appears to be a good ex- 
ample of a macroscopically single phase 
that is internally and microscopically 
a mixed-layered type of two-phase inter- 
growth. Further unmixing to discrete 
granules is apparently prevented by the 
large activation energy needed to break 
bonds across the phase interface region. 
When the material was heated for 1 
month with excess water at 500 bars, 
910° C, the X-ray compositions con- 
verged to Or 23 and 0r 40 ; essentially no 
further change was observed after a 
second month. At 930°C, small networks 
of glass were observed locally, indicat- 
ing beginning of melting at 925°±5°C. 
At 935°C, a broadened peak indicated 
the persistence of two feldspars, again 
in the presence of a small amount of 
glass, which increased to only about 5% 



at 950°C, in the presence of two feld- 
spars, Or 2 i and Or 54 . The results are 
summarized in Fig. 31, which is a pro- 
jection from anorthite. The figure shows 
the most likely interpretation of the 
data, i.e., that the solvus crest lies very 
slightly below the solidus in projection. 
The beginning of melting of 925°±5°C 
is in agreement with the alkali feldspar 
diagram of Tuttle and Bowen (1958, 
Figs. 17 and 26) but some 30° higher 
than implied by their text (p. 41) and 
run data (Bowen and Tuttle, 1950, 
Table 1). 

Because the solvus crest rises with 
total pressure and the solidus is lowered 
with water pressure, the implication is 
clear that the Kiglapait rock from which 
this initially homogeneous feldspar was 
taken formed at an effective P H o at least 



1000 



o 

o 



CD 



O 

CD 
Q. 



c900 



£ 



800 



700 



Kl 3001 fs 



"H 1 i'Run 

P H o = 500 bars I length. 
H 2Y I days _ 

I 3 " 

- I 3 




Ab 




40 60 

Weight per cent 



80 



J 



61,30 



25 



30 - 



28 - 



100 
Or 



Fig. 31. Projection from anorthite of X-ray composition of mesoperthite from Kiglapait 
intrusion sample KI 3001 at 500 bars Ph 2 o. 



126 



CARNEGIE INSTITUTION 



as low as 500 bars, and probably much 
lower. If the magma had been saturated 
with water vapor at the pressure of 
formation (perhaps as much as 5 kb), 
a two-feldspar rock would surely have 
resulted. These observations are in ac- 
cord with the essentially anhydrous 
aspect of the rocks as observed in thin 
section. Hydrous rocks that contain 
mesoperthite would be expected to show 
a less calcic (or more sodic) feldspar 
than the Kiglapait sample studied; 
otherwise the origin of the hypersolvus 
feldspar would be difficult to explain. 

The occurrence of mesoperthite in 
granulite facies rocks of the Adirondack 
region of New York has been used by 
de Waard (1967) as a lower limiting 
condition of the temperature of granulite 
facies metamorphism, based on the be- 
havior of the alkali feldspar solvus with 
pressure, deduced by Orville (1963). 
Extrapolation of the present results for 
calcic mesoperthite would lead to the 
very high temperature of 1000 °C at a 
pressure between 4 and 7 kb, allowing 
for uncertainty in the effect of anor- 
thite on the dT/dP of the solvus crest. 
Clearly, therefore, the curve shown by 
de Waard (1967, Fig. 3) is even more 
stringently a minimum condition for 
granulite temperatures than he suggests, 
and again the An content of granulite 
mesoperthites would be expected to be 
low. 

Many more experimental data on the 
solvus and solidus of the ternary feld- 
spars are needed, since a number of un- 
expectedly rigorous bounding conditions 
appear to be imposed on some rocks as a 
result of the An content of their hyper- 
solvus feldspars. 

Annealing Characteristics of Dense 
Feldspar Glass 

E. C. T. Chao * and P. M. Bell 

Coesite, stishovite, and dense feldspar 
glasses have been found in ejecta from 
terrestrial meteorite craters. Their pres- 

*U. S. Geological Survey. 



ence is evidence of a history that in- 
volves shock waves with associated pres- 
sures far greater than those seen in tec- 
tonic or volcanic events. Collisions be- 
tween comets or meteorites, and between 
them and either the earth or the moon, 
could generate shock pressures ranging 
up to several megabars. The study of 
impact products, in the form of abundant 
dense feldspar glasses, should be help- 
ful in investigating the origin of ter- 
restrial and lunar structures that have 
the morphological characteristics of im- 
pact or explosive craters. Such a study 
is timely, particularly in light of the 
need for information to aid in the in- 
terpretation of lunar samples scheduled 
to be returned to the earth in the forth- 
coming Apollo project in the fall of 1969. 

At present the identification of dense 
feldspar glass is based on measurement 
of index of refraction and the change to 
a lower index as a result of annealing 
at atmospheric conditions. It is necessary 
to correlate the density of the glasses 
with their pressure of formation. In the 
present work, several feldspar glasses 
were formed from the analyzed natural 
crystals by exposure to elevated tem- 
peratures and pressures in the range 
1-45 kb, in the solid-media, piston- 
cylinder apparatus. The natural crystals 
used for starting materials in the high- 
pressure experiments were orthoclase 
(Or S6 , Benson Mines, New York) , oligo- 
clase (An 20 , Muskwa Lake, Ontario), 
and calcic labradorite (An 68 , Lake 
County, Oregon). 

The synthetic glasses produced from 
the high-pressure experiments were then 
studied in a series of annealing experi- 
ments in which relaxation of density and 
change in compressibility were followed 
by monitoring the index of refraction. 

In Fig. 32 is seen a linear increase of 
refractive index of the three feldspar 
glasses as a function of the pressure at 
which they were run. The precision of 
measurement was ±0.0002 with an ac- 
curacy of ±0.001. The indices were 
monitored on the experimental products 



GEOPHYSICAL LABORATORY 



127 




30 

Pressure , kb 

Fig. 32. Index of refraction of three feldspar glasses (Anas, Ana>, Or 90 ) at the pressures at which 
they were quenched from liquid. 

for 6 months after culmination of the sure and temperature. The quenched 



experiments, and no index change was 
observed. Therefore, the index of refrac- 
tion of a glass rapidly quenched from 
high pressures and temperatures could be 
used as a secondary calibration for pres- 



glasses are also useful in annealing 
studies. 

After a few exploratory experiments it 
became evident that there are three 
distinct parts of a synthetic annealment: 



.580 



1.570- 



.560 — 



1.550 — 



.540, 



L I I ' i 

P"* 4SS°C«428°C 

@ 52 7-C ~^ 77 ° C 


— 


^p630 o C 

64?°rT*" — 

"""""""—» ^. 654°C 
663°C"- 




— 


— 


732°C 
@^732«C 

73l=COO o73| o C 

7^ 723X 
_ u q 73I?$q732' > C 

737>C^_^ 7320C 


— 


_ » 




v q800°C 

8I0°C^^- O8I0-C 

~~ ' ■ — ■ —.0816^0 
834°C 

1 1 1 1 







10 



20 30 

T, seconds 



40 



50 



Fig. 33. Annealing experiments on Anos glass that had been prepared by quenching at 
29.6 kb. Timing began as the glass reached the temperature indicated at each point. No measure- 
ments were made during the approach to annealing temperature. 



128 



CARNEGIE INSTITUTION 



change of index during heating, change 
of index during a period of constant 
temperature, and change of index during 
cooling. It was not feasible for us to 
duplicate the initial stage of the natural 
annealment process ; however, it was pos- 
sible to conduct meaningful studies of the 
second and third stages of the process. 
Experiments lasting up to 90 seconds 
were performed, and changes in refrac- 
tive index were followed during the 
various parts of the annealing cycle. 
Fig. 33 shows the lowering of index at 
various temperatures and heating pe- 
riods. For example, there is a negligible 
change as a result of heating at 400°- 
500 °C, whereas a drop in index to the 
1-atm value occurs at 850 °C in 30 sec- 
onds. In these experiments the dense 
glass starting material (An 68 , synthe- 
sized at 29.6 kb) was raised to the 
temperature indicated and rapidly 
quenched; the total time and index 
measured after the experiment are 



plotted in Fig. 33 ; that is, the three parts 
of the annealing cycle are not dis- 
tinguished. 

In order to evaluate the annealing ef- 
fects during heating, namely, the first 
part of the annealing cycle, rapid 
quenching (less than 2 seconds) was 
applied to two glasses of An 68 composi- 
tion, one previously synthesized at 
29.6 kb and the other previously syn- 
thesized at 15 kb. The results are 
plotted in Fig. 34, each point indicating 
the maximum temperature achieved. A 
significant part of the annealing cycle 
occurs during the first few seconds of 
heating, and the glass of higher density 
anneals approximately twice as fast as 
the glass of lower density, even though 
the heating rate is the same for both. 

A different type of annealing experi- 
ment from those previously mentioned 
was made at high pressures. The An 68 
glass was treated by the following pro- 
cedure. (1) The glass was quenched 



.585 r 



T~ 

1.580- 
1.575- 
1.570- 
N 1565-p- 
1.560' 
1.555 
1.550- 
1.545 



\ 



" n 550°C 

S o600 e C 

S d650°C 

N O700°C 
\ 
\ 
\ 
\ 
- Knor V,750°C 

-2g° c 650°C \ 

"°- o 750°C\ 

750°CA>775°C 
\\ 

\\ 

\ N o800°C 

\ \ 

775°C □•<— < 



\ 



^N 



,^-^— D 850°C 
""• 7~O850°C 



0! 23456789 

T, seconds 

Fig. 34. Annealing experiments with two samples of Anes glass, one prepared at 15 kb 
and the other at 29.6 kb. The annealing furnace was set at 850°C. Temperatures indicated are 
the highest temperatures reached before quenching. Note that the glass prepared at 29.6 kb 
(open squares) anneals approximately twice as fast as the one prepared at 15 kb (open circles). 



GEOPHYSICAL LABORATORY 



129 



from above the liquidus temperature at 
10 kb; (2) the pressure was raised to 
20 kb in one set of experiments and to 
40 kb in another set of experiments ; (3) 
the temperature was raised to various 
values below the liquidus up to 1100°C 
and held for 10 minutes; (4) the experi- 
ment was quenched. Here the purpose 
was to investigate the densification pro- 
cess under static conditions. The results 
for the two sets of experiments, at 20 kb 
and 40 kb, are given in Fig. 35. A num- 
ber of distinct features can be noted. 
The glasses do not undergo densification 
below 200 °C, even if the pressure is 
40 kb. There is a steep rise in density at 
both pressures up to 800 °C, at which 
point the maximum index of refraction 
seems to decrease slightly with further 



increase in temperature. This decrease 
is especially marked for the 20-kb ex- 
periments. 

Our purpose has been to model the 
various parts of the Shockwave process 
for careful observation by using static 
techniques. It is not possible to evaluate 
the actual rate process, since natural 
shock events probably last only a few 
microseconds to seconds and the actual 
characteristics of meteoritic impact 
ejecta will be influenced by many fac- 
tors, for example the characteristics of 
the shock-wave, the release adiabat, and 
the particle size. It is evident from the 
present experiments that glasses pro- 
duced at high pressures have a density 
reflecting the formation pressure. During 
annealing the high-pressure density will 



N 




200 



400 



600 800 

Temperature, °C 



1000 



Fig. 35. Samples of An 6 8 were melted and quenched at 10 kb, then raised to 20 or 40 kb 
(upper curve — 40 kb, open squares; lower curve — 20 kb, open circles) and held at the indicated 
temperatures for 10 minutes, then quenched. 



130 



CARNEGIE INSTITUTION 



be lowered at a rate that is a function 
of temperature. The curves plotted in 
Fig. 34 have a reverse S shape, which 
should prove useful in projecting back to 
the conditions of formation of high-dens- 
ity glasses. These curves cannot be used 
to describe the high-pressure, high-tem- 
perature history of shock-produced 
glasses with great accuracy, but will be 
useful in providing clues to the condi- 
tions of their formation. By analyzing 
dense glasses extracted from crater ejecta 
it should be possible to distinguish 
craters that have been produced by 
meteorite impact from those produced 
by volcanic or tectonic processes. 

DIAMOND INCLUSIONS, ALUMINUM 
SILICATES, 0-QUARTZ, ANDRADITE 

Mineral Inclusions in Diamonds 

H. 0. A. Meyer and F. R. Boyd 

A unique method to obtain geochemi- 
cal and mineralogical data on the phases 
that form the rocks of the upper mantle 
is to study the mineral inclusions found 
in natural diamonds. The conditions of 
high temperature and pressure required 
for the stable formation of diamond are 
such that coarse-grained natural dia- 
monds (as distinguished from those 
formed in meteorite impacts) have most 
likely come from depths below 100 km, 
far below the base of the earth's crust. 

It is of primary interest to determine 
the chemical compositions of these in- 
clusions. In materials thus far made 
available for study, the inclusions are 
very small, the largest being only sev- 
eral hundred micrometers in longest di- 
mension. This size is far too small to 
permit study by the classical methods 
of analytical chemistry, but the composi- 
tions can be determined accurately with 
an electron probe. The quantitative 
analyses reported here are the first to 
have been made of inclusions in dia- 
mond. 

Most primary inclusions thus far 
found in natural diamonds are euhedral, 
and they are invariably single crystals. 



The olivine inclusions especially may 
show particular crystallographic orienta- 
tions with respect to their diamond hosts 
(Mitchell and Giardini, 1953; Meyer, 
Year Book 66) . In some diamonds more 
than one inclusion is present, and oc- 
casionally more than one phase is found 
in a group of individually isolated in- 
clusions. Nevertheless, in such cases each 
individual inclusion is monomineralic. 
This fact is curious, and its interpreta- 
tion is at present not entirely clear but 
it suggests that the diamonds have 
formed by a solid-state process rather 
than by crystallization from a magma. 
In the latter event, the diamonds might 
be expected to have incorporated drop- 
lets of magma that would have crystal- 
lized as polymineralic inclusions. 

Because of the monomineralic nature 
of the inclusions, it is impossible to be 
certain whether they are relics of one or 
more phase assemblages. The fact that 
two markedly different types of garnet 
have been found as inclusions suggests 
the latter possibility. 

In a gross way the minerals that are 
found as inclusions are similar to the 
minerals of the nodules in kimberlite and 
to the individual primary crystals in the 
matrix of kimberlite. The electron-probe 
analyses, however, show that in detail 
there are some striking differences. Four 
of the five garnets analyzed are ex- 
tremely rich in chrome pyrope, so much 
so that they contain about 30% of the 
Mg 3 Cr 2 (SiOJ 3 end member. In this re- 
spect they are very distinctive garnets. 
The chromites analyzed are also un- 
usually chrome rich. In contrast, the 
single diopside inclusion is notably low 
in R 2 3 , particularly in Cr 2 R . In this 
regard, its composition contrasts with 
the chrome diopsides of the peridotite 
nodules and the omphacites of the eclog- 
ite nodules. 

It is obviously very speculative at this 
stage to conclude that these facts are 
related. But it is known that in the 
system enstatite-pyrope high pressure 
favors the crystallization of pyrope 



GEOPHYSICAL LABORATORY 



131 



rather than aluminous enstatite. Possibly 
an analogous reaction exists in the sys- 
tem diopside-Mg 3 Cr 2 (Si0 4 )3. If so, it 
may develop that the inclusions reflect 
different, perhaps more deep-seated, 
pressure-temperature conditions than 
those under which the minerals of the 
nodules equilibrated. It is well estab- 
lished that diamonds have been found as 
primary phases in peridotite and eclogite 
nodules. It is possible that the mono- 
mineralic inclusions in diamond are 
armored relics of pressure-temperature 
events earlier than those reflected by the 
present phase assemblages of the nodules. 
Mercy and O'Hara (1967) have noted 
that garnets from peridotites are rich in 
Cr and poor in Mn and Ti relative to 
garnets from eclogites. On this basis, the 
analytical data in Table 5 suggest that 
the chrome pyropes have a peridotite 
affinity whereas the pyrope-almandine 
could have been derived from an eclogite 
assemblage. 

Method of Study 
H. 0. A. Meyer 

The inclusions must be removed from 
their diamond hosts for study with the 
electron probe, and their removal is most 
easily and reliably accomplished by 
slowly burning the diamonds in air at 
about 800°C. At this temperature sev- 
eral hours are usually required to elimi- 
nate the diamond by oxidation. This 
treatment is destructive to sulfide in- 
clusions, but on the silicates and oxides 
thus far recovered it appears to have 
little effect beyond slight oxidation of 
the olivines. Once removed, the inclu- 
sions are mounted on glass slides and 
lapped to produce flat surfaces suitable 
for electron-probe analysis. Because of 
the small size of the inclusions, areas 
available for analysis were usually re- 
stricted, and it was necessary to re- 
polish an inclusion one or more times 
during analysis to remove contamina- 
tion spots. 

The analyses were made with a Ma- 



terials Analysis Company Model 400 
electron probe.* Standards used were 
predominantly those described by Boyd 
in Year Book 66 (p. 329). The intensity 
ratios were reduced to chemical compo- 
sitions and corrected for instrumental 
and matrix effects with the use of the 
computer programs described elsewhere 
in this report (Boyd, Finger, and 
Chayes). Results for the major elements 
are believed to be accurate to ±2 rela- 
tive %. Results for minor elements are 
somewhat less accurate, particularly 
where line interference (e.g., MnKa and 
CrK/3) makes determination of back- 
ground values uncertain. The ratio 
o-VJv, where a is the standard deviation 
and N is the mean count, was calculated 
for the major elements in the analyses in 
Tables 4-7. This ratio is a useful indi- 
cator of inhomogeneity ; a value greater 
than 3 is taken as evidence of in- 
homogeneity. 

Olivines 
H. 0. A. Meyer 

Five olivine inclusions were selected 
for analysis, and a sixth specimen, an 
olivine from the Stockdale kimberlite 
pipe, Kansas, supplied by D. G. Brook- 
ins, was analyzed for comparison. The 
results are given in Table 4. 

The inclusions are markedly similar to 
olivines from ultramafic rocks (O'Hara 
and Mercy, 1963; Nixon, von Knorring, 
and Rooke, 1963; Ross, Foster, and 
Myers, 1954). Comparison of the inclu- 
sion analyses with that of the Stockdale 
olivine furthers this observation, al- 
though admittedly the latter is from a 
kimberlite and not an ultramafic nodule. 
The minor-element content of the olivine 
inclusions is identical with that of other 
forsterite-rich olivines. No analysis has 
yet been made for NiO, but qualitative 
results indicate it to be present in the 
range 0.2-0.3 wt %. Cobalt has been 

* The assistance of the National Science 
Foundation in the purchase of this instrument 
under grant GP 4384 is gratefully acknowledged. 



132 



CARNEGIE INSTITUTION 



TABLE 4. Analyses of Olivine Inclusions 





9b 


10c 




11a 


13c 


23b 




Stockdale 


Si0 2 


40.7 


40.8 




40.4 


41.0 


4 


40.7 


4 


40.8 8 


TiO* 


0.00 


0.00 




0.00 


0.02 




0.00 




0.02 . . . 


ALA, 


0.07 


0.02 




0.02 


0.06 




0.02 




0.02 . . . 


Cr 2 3 


0.17 


0.06 




0.08 


0.06 




0.04 




0.00 . . . 


FeO* 


7.79 


7.09 




5.52 


7.11 


"h 


7.60 


"2 


7.88 1 


MgO 


49.6 


51.6 




52.9 


50.4 


3 


51.1 


s 


51.7 1 


CaO 


0.18 


0.07 




0.01 


0.09 


2 


0.07 




0.01 . . . 


MnO 


0.17 


0.12 




0.09 


0.12 


1 


0.12 


i 


0.10 1 


Totals 


98.7 


99.8 


99.0 


98.9 


99.7 


100.5 






Number of Ions 


on the Basis of 4 Oxygens 






Si 


1.001 1 


0.991] 




0.982] 




1.003] 


0.992] 




0.987] 


Ti 


o.ooo y 1.003 


0.000 y 


0.992 


0.000 


•0.983 


0.000 } 


1.005 0.000 > 


0.993 


0.000^0.988 


Al 


0.002 J 


0.001 J 




0.001 J 




0.002 J 


0.001 J 




0.001 J 


Cr 


0.003-, 


0.001 1 




0.001-| 




0.001-, 


0.001 1 




0.000> 


Fe 2+ 


0.161 


0.144 




0.112 




0.145 


0.155 




0.159 


Mg 


1.821 y 1.993 


1.868 y 


2.018 


1.918 } 


2.034 


1.839 y 


1.990 1.887 } 


2.048 


1.863 1 2.025 


Ca 


0.005 


0.002 




0.001 




0.002 


0.002 




0.001 


Mn 


0.003 J 


0.003 J 




0.002 J 




0.003 J 


0.003 J 




0.002 J 










End Members, mole % 








Forsterite 


91.5 


92.6 




94.3 


92.5 




92.2 




92.0 


Fayalite 


8.1 


7.1 




5.5 


7.3 




7.6 




7.9 


Larnite 


0.3 


0.1 




0.0 


0.1 




0.1 




0.0 


Tephroite 


0.1 


02 




0.1 


0.1 




0.1 




0.1 



* Total Fe as FeO. 

Note : Numbers in italics are for the ratio <r/vN. 



detected in two out of three inclusions 
in which it was sought ; the concentration 
is of the order of 10 ppm. 

Garnets 
H. O. A. Meyer 

Chrome-rich pyrope and pyrope-al- 
mandine occur as inclusions in the dia- 
monds examined (Table 5). The chrome 
pyropes (Meyer, 1968) are the most 
magnesium-rich yet found in nature. 

The large chromium content of the 
chrome pyropes is expressed in the end- 
member composition Mg3Cr 2 (Si0 4 )3. Be- 
cause of the small calcium content in 
these inclusions it makes little difference 
whether the chromium is first assigned to 
uvarovite or to Mg 3 Cr 2 (Si0 4 )3. Previ- 
ously, garnets containing Mg 3 Cr 2 (Si0 4 )3 
in minor amounts were known only from 
olivine-bearing ultramafic rocks (Nixon, 
von Knorring, and Rooke, 1963; Fiala, 
1965; O'Hara and Mercy, 1963). Re- 



cently, Nixon and Hornung (1968) have 
discovered chrome-pyrope garnet, simi- 
lar to the inclusions herein reported, in 
a heavy mineral concentrate from a 
kimberlite pipe in Lesotho (Basutoland) . 
The stability field of Mg 3 Cr 2 (Si0 4 ) 3 
garnet is unknown, although its synthesis 
at high pressure and temperature has 
been reported (Coes, 1955) . 

The pyrope-almandine inclusion (20f) 
is more iron rich than most garnets from 
kimberlite pipes (cf. Nixon, von Knor- 
ring, and Rooke, 1963; Bobrievich et al., 
1959) . The small size (10 X 18 /mi) of this 
inclusion and the lack of repolishing be- 
tween analyses undoubtedly contributed 
to the low total sum of the oxide weight 
percentages. 

Chromites 

H. O. A. Meyer 

The three chromite inclusions analyzed 
(Table 6) are among the most chrome 



GEOPHYSICAL LABORATORY 



133 



TABLE 5. Analyses of Garnet Inclusions 





1 




6 




15e 




15h 




20f 


SiG, 


42.3 


8 


42.8 


2 


41.8 


8 


422 


2 


37.8 4 


Ti0 2 


0.02 




0.00 




0.02 




0.02 




0.25 7 


AI2O3 


17.2 


"l 


182 


"k 


15.7 


"k 


15.7 


"2 


203 8 


Cr 2 3 


8.93 


1 


7.9 


1 


10.9 


1 


10.7 


1 


0.06 . . . 


FeO* 


5.36 


1 


4.75 


2 


5.71 


1 


5.57 


1 


29.5 1 


MgO 


25.3 


1 


25.5 


2 


242 


1 


24.5 


1 


7.35 1 


CaO 


1.09 


1 


1.35 


1 


2.19 


1 


2.22 


2 


127 9 


MnO 


0.21 


1 


0.17 


1 


020 


1 


0.19 


1 


0.39 2 


Totals 


100.4 


100.7 


100.7 


101.1 


96.9 t 






Number of Ions on 


the Basis of 12 Oxygens 






Si 


2.9911 




3.0021 




2.9821 




2.9941 




3.040 3.040 


Ti 


0.001 y 


3.000 


0.000 


> 3.002 


0.001 


•3.000 


0.001 


-3.000 


0.0161 


Al 


0.008 J 




...J 




0.017 J 




0.005 J 




1.942 


Al 


1.4241 




1.5061 




1.3001 




1.3041 




1.922 } 


Cr 


0.499 > 


2.000 


0.438 


-1.997 


0.617 


-2.000 


0.601 


•2.000 


0.004 


Fe s+ t 


0.077 J 




0.053 J 




0.083 J 




0.095 J 




J 


Fe 2+ 


0.2401 
2.668 I 
0.083 [ 




0.2261 




0.2581 




02351 




i.9821 


Mg 

Ca 


3.003 


2.663 1 
0.101 I 


> 3.000 


2.570 I 
0.167 


•3.007 


2.587 1 
0.168 1 


-3.001 


jgg 3.000 


Mn 


0.012 J 




0.010 J 




0.012 J 




0.011 J 




0.026 J 








End Members, mole % 








Spessartite 


0.4 




0.3 




0.4 




0.4 




0.9 


Andradite t 


2.8 




2.7 




4.1 




45 






Skiagite t 


1.1 


















Uvarovite 






0.7 




1.4 




0.8 




0.2 


Grossularite 


















3.6 


Mg3Cr 2 (Si0 4 ) 


3 24.9 




21.2 




29.5 




29.2 






Pyrope 


63.9 




67.7 




56.2 




57.0 




30.3 


Almandine 


6.9 




7.4 




8.4 




7.8 




65.0 



* Total Fe as FeO. 

t Fe 3+ calculated to satisfy charge requirements. 

t See text. 

Note: Numbers in italics are for the ratio cr/ViV. 



rich so far recorded (Irvine, 1967; Ross, 
Foster, and Myers, 1954; Nixon, von 
Knorring, and Rooke, 1963). 

In two of the three inclusions ex- 
amined (2 and 3b), exsolution lamellae 
were observed. Chromite inclusion 3b 
exhibited two different exsolution struc- 
tures, believed to be possibly magnetite 
and ilmenite (A. El Goresy, personal 
communication). Similar but coarser il- 
menite exsolution lamellae were present 
in chromite inclusion 2. Unfortunately, 
these lamellae could not be resolved by 
the microscope on the probe. In chrom- 
ite 2, local increase in the FeKa count 
rate was believed to indicate the presence 
of an exsolution lamella. The analyses 
presented in Table 6 do not differentiate 
between the host or lamellae and are 



thus bulk analyses approximating more 
closely the composition of the host than 
that of the lamellae. 

The cell size of a third chromite from 
the same diamond as inclusions 3 and 
3b was determined by X-ray diffraction 
to be 8.35 ±0.02 A. A second crystal 
structure evident in the X-ray photo- 
graph had an axis (repeat distance 8.5 ± 
0.1 A) parallel to [110] of the chromite 
host. This cell size and orientation are 
suggestive of magnetite and are in keep- 
ing with the above exsolution observa- 
tions. 

Diopside 
F. R. Boyd 
Both diopside and enstatite have been 
found as inclusions in diamond, but they 



134 



CARNEGIE INSTITUTION 



TABLE 6. Analyses of Chromites 





2 




3 




3b 




SiO a 


0.13 


3 


0.29 


7 


0.43 


8 


Ti0 2 


0.12 


1 


0.09 




0.09 




A1 2 3 


5.12 


3 


3.20 


"k 


3.26 


19 


Cr 2 3 


67.2 


2 


61.4 


2 


62.1 


2 


FeO* 


14.5 


6 


31.5 


1 


31.7 


1 


MgO 


14.2 


2 


0.54 


1 


0.48 


3 


CaO 


0.02 




0.02 




0.04 




MnO 


0.00 




0.42 


k 


0.45 


2 


ZnOt 


0.04 




1.93 


1 


220 


2 


Totals 


101.3 


99.4 


100.8 






Number of Ions 


on the Basis of 4 Oxygens 






Si 


0.004-1 




0.011-1 




0.016-. 




Ti 


0.003 




0.003 




0.003 




Al 


0.195 } 


1.998 


0.138 \ 


1.995 


0.139 


•1.995 


Cr 


1.716 




1.781 




1.776 




Fe 8+ t 


0.080 J 




0.062J 




0.061 J 




Fe 2+ 


0.3131 




0.905-1 




0.900-1 




Mg 


0.685 




0.029 




0.026 




Ca 


0.001 > 


1.000 


0.001 \ 


1.000 


0.001 


1.000 


Mn 


0.000 




0.013 




0.014 




Zn 


0.001 J 




0.052 J 




0.059 J 








End Members, mole % 






MnALO* 






1.3 




1.4 




ZnAl*0« 


o.i 




52 




5.9 




MgCr 2 0, 


69.7 




3.0 




2.7 




FeCr 2 0* 


17.5 




862 




86.3 




FeAl 2 4 


8.7 




1.1 




0.6 




FeFe 2 4 


4.0 




32 




3.1 





* Total Fe as FeO. 

t Matrix corrections calculated without aid of computer program. 

% Fe 3+ calculated from total iron as Fe 2+ to satisfy charge requirements. 

Note: Numbers in italics are for the ratio a/vN. 



are uncommon. A diamond crystal in the 
collection of the Department of Chem- 
istry, University College London, was 
discovered to contain two separate in- 
clusions of diopside. They have been 
removed by fracturing the diamond, and 
one of them provided for electron-probe 
analysis through the courtesy of H. 
Judith Milledge. 

The inclusion is tabular, parallel to 
(100), and measures about 200 /j.m in its 
greatest dimension. There has been some 
alteration to kaolinite, presumably 
through cracks in the diamond host. 
When attempts were made to polish a 
flat surface on the crystal, it was found 
that only restricted, irregular areas 
would take a good polish. It appeared 
that the kaolinite alteration was more 
than surficial. Although the inclusion is 



large in relation to the minimum-sized 
crystal that can be analyzed with a 
microprobe, the analysis was compli- 
cated because of the restricted areas of 
good polish. 

Results of the analysis (Table 7) show 
that the inclusion differs in a number of 
ways from other diopsides from kimber- 
lite nodules (Boyd, Year Book 65 and 
Year Book 66) . It is markedly inhomoge- 
neous, particularly for Ti, Al, Fe, and 
Mn, as shown by high values for the 
ratio o- VW ■ However, no regular pattern 
of zoning was found. The contents of 
ALO3 and particularly Cr 2 3 are not- 
ably low in comparison to other kimber- 
lite diopsides. A1 2 3 and Cr 2 3 show wide 
ranges in these diopsides, but mean 
values of 42 analyses by Boyd (1968) 
are 0.9 wt % Cr 2 3 and 1.6 wt % A1 2 3 . 



GEOPHYSICAL LABORATORY 



135 



TABLE 7. Analysis of Diopside Inclusion 









] 


Slumber of Ions 










on the Basis of 




Weight % 




6 Oxygens 


Si0 2 


52.8 


4 


Si 


1.97 1 


Ti0 2 


0.43 


29 


Ti 


0.012^2.000 


AI2O3 


0.86 


74 


Al 


0.018J 


O2O3 


0.09 




Al 


0.020^ 




FeO* 


5.89 


"u 


Cr 


0.003 




MnO 


0.71 


57 


Fe 


0.183 




CaO 


20.9 


7 


Mn 


0.022 




MgO 


16.1 


5 


Ca 


0.834 


^2.057 


Na 2 


1.38 


10 


Mg 


0.895 




K 2 


<0.004 




Na 
K 


0.100 




Total 


99.2 




Atomic °/o 








Ca 




43.6 






Mg 




46.8 






Fe 




9.6 






Ca/(Ca 


+ Mg) 


48.2 







* Total Fe as FeO. 

Note: Numbers in italics are for the ratio 
<r/V¥. 



On the other hand, FeO and particu- 
larly MnO are notably high in the dia- 
mond inclusion. 

Solid solution toward enstatite is very 
restricted in this inclusion; its Ca/(Ca + 
Mg) ratio is 0.48. This is not outside the 
range found for other kimberlite diop- 
sides, but it would imply a temperature 
of equilibration of about 950°C if the 
inclusion were to have crystallized in 
equilibrium with enstatite. This tem- 
perature is low in comparison to the 
range in which diamonds have been syn- 
thesized in the laboratory. 

The Andalusite-Sillimanite Transi- 
tion and the Aluminum Silicate 
Triple Point 

M. C. Gilbert, P. M. Bell, and 
S. W. Richardson* 

Our study of the stability relations 
of the three ALSi0 5 polymorphs, kyanite, 
sillimanite, and andalusite, is nearly 
completed. Experiments determining the 
kyanite-sillimanite and kyanite-andalu- 

* Grant Institute of Geology, University of 
Edinburgh. 



site univariant reactions were reported 
in Year Book 66 (Richardson, Bell, and 
Gilbert, pp. 392-397) . Subsequent analy- 
sis of those results yields the equation 
for the kyanite-sillimanite transition 
above zero pressure and 396°C: 

P(kb)=0.02437TC)-9.63 

Similarly, the equation for the kyanite- 
andalusite transition, based on our work 
plus that of Newton (1966), is, above 
zero pressure and 110°C, 

P(kb) =0.01077TC) -1.173 

The intersection of these curves at 
5.5 kb and 622 °C defines the triple point 
for the three polymorphs. Because of the 
low angle of intersection, a considerable 
region of uncertainty existed around this 
point. In order to reduce the uncertainty, 
we have studied the andalusite-silliman- 
ite transition hydrothermally in cold- 
seal pressure vessels at 2, 3, and 4 kb. 

Starting materials were equal mixtures 
of analyzed sillimanite from Brandywine 
Springs, Delaware (electron-microprobe 
analysis gave total Fe as Fe 2 3 as 0.03 ± 
0.015 wt %), and andalusite from Stand- 
ish, Maine (total Fe as Fe 2 3 is 0.4 ± 
0.1 wt %). The sillimanite is fibrolitic 
and contains quartz. Experimental pro- 
cedures and analysis of results are similar 
to those used in our kyanite-sillimanite 
study. An X-ray peak-height ratio, 
determined from diffractometer charts 
for andalusite (220) /sillimanite (120), 
was used to monitor the reaction. Aver- 
age values of this ratio for many mounts 
of the starting material, determined by 
scanning each mount four times, ranged 
between 0.53 and 0.70, with values of all 
individual scans falling between 0.47 
and 0.76. Reaction, if it has occurred, 
cannot be detected if mean values of the 
ratio fall in this range. But if the mea- 
sured ratio of peak heights for a par- 
ticular run product is less than 0.40, the 
change in ratio is ascribed to growth of 
andalusite; if 0.90 or more, to growth of 
sillimanite. 

This reaction has proved particularly 



136 



CARNEGIE INSTITUTION 



sluggish, and run times of 2% to 4 
months were necessary to demonstrate 
reaction. The runs were interrupted two 
or three times for measurement of the 
peak-height ratio and for regrinding. 
Runs meeting the direction-of-reaction 
criterion set forth above are shown in 
Fig. 36 and fix the transition within fair 
limits. 

From the slopes of the kyanite-silli- 
manite and kyanite-andalusite transi- 
tions given by the equations above, and 
from application of the molar-volume 
data of Skinner, Clark, and Appleman 
(1961), the slope of the andalusite-silli- 
manite transition radiating from the 
triple point can be calculated by means 
of the equation 



which satisfies our run data very well 
(Fig. 36). 

Figure 37 shows the limits of uncer- 
tainty around the triple point, based on 
the experimental data and determined by 
application of maximum and minimum 
slopes possible for each of the three 
curves. 

The andalusite-sillimanite transition 
can be used as a fairly good geothermom- 
eter for a metamorphic facies series cor- 
responding to relatively low-pressure 
conditions. Combination of information 
from this transition with that of de- 
hydration reactions and/or with mini- 
mum melting curves of rock systems 
(Fig. 37) may also allow quantitative 
estimates of fugacity of H 2 during 



(ML) = 



©„ xAf "-(S« xAf » 



where P is pressure in kilobars, T is 
temperature in degrees Celsius, and AF 
is the change in volume across the ap- 
propriate Al 2 Si0 5 transition. This results 
in an equation for the andalusite-sil- 
limanite transition above zero pressure 
and 847 °C of 

P(kb) = -0.0243T( o C) +20.59 



A7 A -s 
metamorphism. Finally, the maximum 
pressure stability of andalusite (at the 
triple point) at about 5.5 kb provides 
a working minimum limit of pressure 



■ Growth of sillimonite 
□ Growth of andalusite 





500 600 700 800 

Temperature, C C 

Fig. 36. Hydrothermal runs bracketing the 
andalusite-sillimanite transition. The kyanite- 
sillimanite, kyanite-andalusite, and andalusite- 
sillimanite curves are given by equations in 
the text. 



500 600 700 800 

Temperature, C C 

Fig. 37. Pressure-temperature uncertainty 
around the triple point determined by taking 
the maximum and minimum slopes permitted 
by the experimental data only. The Ab-Or-Q- 
H 2 (NaAlSiaOs-KAlSiaOs-SiO-HsO) curve 
modeling the minimum melting of granitic 
systems is taken from Luth, Jahns, and Tuttle 
(1964). 



GEOPHYSICAL LABORATORY 



137 



during kyanite-sillimanite metamorph- 
ism. 

X-ray Properties and Stability Rela- 
tions op /?-Quartz Solid Solutions 
along the Join LiAlSi 2 6 -Si0 2 

J. L. Munoz 

The most common polymorph of Si0 2 
(a-quartz) is one of the purest natural 
compounds known. However, /?-quartz, 
a nonquenchable, high-temperature poly- 
morph of Si0 2 , can accommodate foreign 
ions with relative ease because of large 
holes in the structure adjacent to Si0 4 
tetrahedra. One common example of this 
tendency is the addition of Li in tetra- 
hedral sites, coupled by substitution of 
Al for Si in the tetrahedron (Li + Al^± 
Si). A suitable Li + Al end member for 
this substitution is eucryptite, LiAlSi0 4 , 
which crystallizes at high temperatures 
in a structure of the /?-quartz type, ex- 
cept that in /3-eucryptite the c axis is 
doubled compared with the c axis of 
/?-quartz (Winkler, 1948). At low pres- 
sures, however, extensive solid solution 
of eucryptite in Si0 2 is interrupted by 
the existence of a broad band of tetra- 
gonal solid solutions (/?-spodumene solid 
solution) . This report describes the effect 
of pressure on greatly expanding the field 
of /?-quartz solid solution. 

Above 10 kb, /?-spodumene (LiAl 
Si 2 6 ) becomes unstable and is replaced 
by a hexagonal phase {Year Book 66, 
p. 370). On the basis of powder diffrac- 
tometer data only, this high-pressure 
phase was called "/?-eucryptite ss ." Re- 
cently, however, X-ray precession pho- 
tographs taken of a single crystal of this 
hexagonal phase, which had been 
quenched from 10 kb and 1350°C, 
demonstrated that the c axis of this 
phase is not the doubled c axis of /?-eu- 
cryptite but rather the 5. 5- A c axis of 
/3-quartz, as previously indicated by Li 
(1967). Thus, the high-pressure, high- 
temperature polymorph of LiAlSi 2 6 is 
isomorphous with /3-quartz. 

In an attempt to find out whether the 



/?-quartz solid solution that formed on 
LiAlSi 2 6 composition would extend con- 
tinuously toward Si0 2 composition at 
high pressures, a number of glasses on 
the join LiAlSi 2 6 -Si0 2 (generously 
donated by David B. Stewart) were 
crystallized at 15 kb and 1350°C. Cell- 
dimension and molar-volume data (Fig. 
38) were obtained from quenched 
charges crystallized in this way. Note 
that values obtained for a, c, and V m 
from composition Spod 9 . 5 Q 9 o.5 lie very 
far from the extrapolated trends and 
have not been considered in curve-plot- 
ting. These points are probably the re- 
sult of one of two effects: either the 
molar volume curve itself may be 
strongly inflected near Si0 2 composition, 
or phases that closely approach Si0 2 
composition cannot be quenched to room 
temperature in the /?-quartz structure, 
as is the case with pure Si0 2 . Inasmuch 
as we cannot measure the molar volume 
of Si0 2 in the ^-quartz structure at 25°C, 
no end points for these curves are avail- 
able, and hence we cannot decide be- 
tween the two possibilities at this time. 
Direct measurements of the cell param- 
eters of these phases in their own stabil- 
ity fields with a high-pressure, high- 
temperature X-ray camera would resolve 
the ambiguity. 

The ^-eucryptite solid solutions along 
part of the join LiAlSi0 4 -Si0 2 readily 
crystallize metastably at low pressures. 
Accordingly, it was necessary to per- 
form a number of experiments on 
/?-quartz solid solutions to determine 
whether they are stable at 15 kb or 
whether they might be metastable — 
either with respect to a two-phase as- 
semblage (e.g., two coexisting /3-quartz 
solid solutions) or to a completely dif- 
ferent single phase (e.g., /?-spodumene 
solid solution). The present interpreta- 
tion of these experiments is that ^-quartz 
solid solutions are stable at 15 kb and 
1350°C for all compositions on the join 
LiAlSi 2 6 -Si0 2 , but data sufficient for 
construction of T-X diagrams, including 
temperatures above and below the sta- 



138 



CARNEGIE INSTITUTION 



Mole percent LiAISi?0, 



2 W 6 




Mole per cent (3Si0 2 ) 



LiAISi 2 6 



Fig. 38. Cell dimensions and molar volume for /3-quartz solid solutions on the join LiAlSi 2 06- 
SiO a synthesized at 15 kb, 1350°C. Reflections were indexed according to the unit cell of /3-quartz 
and were measured from diffractometer powder patterns, with CaF 2 as an internal standard. 
Data were refined with the use of the lattice-constant least-squares refinement program of 
Burnham (Year Book 61, p. 132). 



bility field of these /?-quartz solid solu- 
tions, are not yet available. Results of 
the experiments may be summarized as 
follows: (1) At 10, 15, and 20 kb, single- 
phase /?-quartz solid solutions show no 
evidence of breaking down when run for 
6 hours at 1500°C. (Contamination of 
the thermocouple precludes longer runs 
at this temperature.) (2) Two-phase 
mixtures consisting of Si0 2 + LiAlSi 2 6 
(/3-quartz ss ) can be homogenized to 
single-phase hexagonal solid solutions of 
compositions Spod 39 Q 6 i at 15 kb and 
Spod 3 oQ 7 o at 20 kb; the homogenization 



reaction is extremely sluggish, however. 
(3) At 15 kb and very high temperatures 
(1650°-1800°C) compositions in the 
range SpodsoQso-SpodsoQro crystallize in 
part as two hexagonal phases; in these 
experiments glass is invariably present 
(varying from about 5 to 30%) as an 
additional run product. The presence of 
glass — obviously representing a quenched 
liquid coexisting with two crystalline 
phases at temperature and pressure — 
may indicate the onset of nonbinary 
phase relations at these high tempera- 
tures. (4) At all pressures above 10 kb, 



GEOPHYSICAL LABORATORY 



139 



and for all compositions along the join, 
runs initially containing /3-spodumene or 
a /?-spodumene solid solution produced 
a single hexagonal phase in the run 
product. 

The interpretation of these results pro- 
vides a model for the phase relations 
along the join LiAlSi 2 6 -Si0 2 as a func- 
tion of pressure. At low pressures, exten- 
sive substitution of Li + Al for 2Si in 
/?-quartz is prevented by the presence 
of a wide field of tetragonal /3-spodumene 
solid solutions. With increasing pressure 
this field contracts until, at about 10 kb, 
/3-spodumene solid solutions disappear 
completely and are replaced by a con- 
tinuous series of hexagonal phases. At 
still higher pressures (about 25 kb) this 
continuous series will ultimately be in- 
terrupted by the rapidly increasing sta- 
bility of spodumene (c/. Year Book 66, 
p. 372). 

Synthesis and Stability of 
Ti-Andradite 

H. G. Huckenholz 

Andradite, Ca 3 Fe2 3+ Si 3 0i2, and its ti- 
tanium-rich varieties, melanite, schor- 
lomite, and iiaavarite, are common 
garnets in alkaline igneous rocks and in 
thermally metamorphosed, impure lime- 
stone and skarn deposits. The origin of 
Ti-bearing andradites has been the sub- 
ject of much discussion because the 
structural positions of Ti and its valency 
states remain problematical. Zedlitz 
(1933) suggested that a part of Ti is 
present in the trivalent state, but optical 
absorption spectra of a melanite speci- 
men from San Benito County, Cali- 
fornia, obtained by Manning (1967) do 
not appear to adequately prove this as- 
sumption. Suitable ion distribution may 
be achieved in Ti-rich andradites by 
substituting Ti 4+ for Fe 3+ and Si. Infra- 
red spectra of titaniferous garnets 
(Tarte, 1960) and the synthesis of ger- 
manate, rare-earth iron, and rare-earth 
gallium garnets (Espinosa, 1964; Ito 
and Frondel, 1967; Geller, 1967) have 



demonstrated that Ti*+ prefers a six- 
fold position in the garnet structure but 
may also enter a four-fold position. 
Most analyses of Ti-rich grandite gar- 
nets show an excess of tetravalent and 
divalent cations but a deficiency of the 
trivalent group. This fact indicates that 
2R 3+ cations (Fe 3+ , Al) may also be re- 
placed by R 4+ R 2+ (Ti 4+ and Mg,Fe 2+ ,Mn) , 
as demonstrated by Geller, Miller, and 
Treuting (1960) for germanate garnets 
having Ti 4+ Mg, Ti 4+ Ni, and Ti 4+ Co for 
2R 3+ . Some of the Ti-garnet analyses re- 
port alkalies, and if they are not at- 
tributable to impurities, a minor sub- 
stitution of NaTi 4+ for CaFe 3+ may be 
possible too. 

In dealing with garnet analyses, one 
has to keep in mind the great analytical 
and separatory difficulties, particularly 
if minute inclusions are involved and if 
the ferric iron content is high compared 
with the ferrous iron. However, chemical 
analyses of Ti-garnets may be balanced 
ionically within the limit of error. (In 
some older analyses that show an ex- 
cess of R 4+ cations, a conversion of Ti 4+ 
to Ti 3+ would have been necessary but 
a change of Fe 3+ to Fe 2+ would have 
the same balancing effect. Such analyses 
have been neglected here.) The general- 
ized formula of Ti-rich grandite garnets 
as considered in this study may be ex- 
pressed on the basis of 24 oxygens as 

(Ca,Na,Fe 2+ ,Mg) 6 (Fe 3+ ,Ti 4+ ,Al,Mg, 

Fe 2+ ,Mn) 4 (Si,Al,Fe 3+ ,Ti 4+ ) 6 24 

As a first approximation, reliable 
chemical analyses of Ti-rich grandites 
from both igneous (48) and metamorphic 
rocks (12) plot with their molecular 
norm in the enstatite (MgOSi0 2 )-wol- 
lastonite (CaO-Si0 2 ) -hematite (Fe 2 3 )- 
perovskite (CaO-Ti0 2 ) system on or 
below the plane wollastonite-hematite- 
perovskite (Fig. 39) and close to the 
join andradite (3CaO-Fe 2 3 -3Si0 2 ) -Ti- 
garnet (3CaOFe 2 3 -3Ti0 2 ), correcting 
for a minor content of alkalies as a 
Na 2 • CaO • 2Ti0 2 • 3Si0 2 component. Hy- 
pothetical garnet components may be 



140 



CARNEGIE INSTITUTION 



hematite 




andradifei^®® 



Ti -garnet 



wollastonite 



perovskite 



Fig. 39. The quaternary system en-wo-pv-hem, which includes the join wollastonite-perovskite- 
hematite with a plot of analyses of Ti-bearing garnets from metamorphic (solid circles), alkaline 
plutonic (open circles), alkaline volcanic (half -shaded circles), and alkaline pegmatitic rocks 
(circles with crosses). Data were taken from 60 published analyses; complete citations will 
be given in a forthcoming paper. Abbreviations and compositions of phases encountered: hem, 
Fe 2 3 ; TFTs, CaO-Fe 2 3 -Ti0 2 ; FTs, CaO-Fe 2 O s -Si0 2 ; pv, CaO-Ti0 2 ; wo, CaO-Si0 2 ; di, 
CaOMgO2Si0 2 ; en, MgOSi0 2 ; khoharite, 3MgO-Fe 2 3 -3Si0 2 ; andradite, 3CaO • Fe 2 3 • 3Si0 2 ; 
Ti-garnet, 3CaO-Fe 2 3 -3Ti0 2 ; and Mg-melanite, 3CaO-MgO. Ti0 2 -3Si0 2 ; as well as fs, FeOSiO a 
and Na-melanite, Na 2 0-CaO-2Ti0 2 -3Si0 2 . 



calculated on the basis of en, wo, hem, 
and pv. After deduction of Na 2 0-CaO 
2Ti0 2 -3Si0 2 (Na-melanite) the remain- 
ing sum of wo-f-en + fs + pv is equal to 
or larger than 3(hem + c) in acceptable 
analyses, and may be expressed as a 
mixture of 

1 en + 2 wo + 1 pv = Mg-melanite 

(3CaO-MgOTi0 2 -3Si0 2 ) 

1 fs + 2 wo-f-1 pv = Fe-melanite 

(3CaOFeOTi0 2 -3Si0 2 ) 



if wo-r-en-|-fs + pv>3(hemH-c). If wo + 
en + fs + pv = 3(hem + c) the following 
additional components may be formed: 

3 pv + 1 hem = melanite 

(3CaO-Fe 2 3 -3Ti0 2 ) 

3 en + 1 hem = khoharite 

(3MgO-Fe 2 3 -3Si0 2 ) 

3 f s + 1 hem == skiagite 

(3FeO-Fe 2 3 -3Si0 2 ) 



GEOPHYSICAL LABORATORY 



141 



3 wo + 1 hem = andradite 

(3CaO-Fe 2 3 -3Si0 2 ) 



grandite 



3 wo + 1 c = grossularite 

(3CaOAl 2 3 -3Si0 2 ). 

Ti-rich grandites from the Kaiserstuhl 
volcanic area, Germany, and from Mag- 
net Cove, Arkansas, have the composi- 
tion given in Table 8. 

The Join Andradite-Ti-Garnet 

In order to obtain information about 
the nature of the incorporation of ti- 
tanium (calculated as tetravalent) in 
andradite, a series of nineteen composi- 
tions in the plane wollastonite-perovsk- 
ite-hematite was prepared along the 
join andradite-Ti-garnet and for four 
additional compositions in this plane. 
DeVries, Roy, and Osborn (1956) ex- 
amined the join wollastonite-perovskite 
at 1 atm. Ito and Frondel (1967) studied 
parts of the join andradite-Ti-garnet 
from andr 100 Ti-gar to andr 31 Ti-gar 69 at 
1050 °C and 1 atm, using gel techniques. 
The latter authors reported a maximum 
heating time of 20 hours for the forma- 
tion of a garnet solid solution, far short 
of the time required to attain equilibrium 



in the present study, in which glasses 
and crystallized glasses were used. At 
least 50 days at 1050 °C are neces- 
sary with ten thorough grindings in 
the crystallizing period for a com- 
plete solution of all metastably formed 
wollastonite in the garnet ss , even when 
the starting material is a very fine 
powdered glass (Fig. 40) . No wollaston- 
ite and pseudowollastonite (or only 
traces) were formed metastably from a 
starting material that was prepared in 
this way when treated under subsolidus 
conditions at temperatures between 
1000° and 1268°C. The data in the pres- 
ent investigation were obtained at 1 atm 
pressure from quenching mixtures that 
had been held between 1000° and 1400°C. 
Results are presented in a T-X diagram 
(Fig. 41) and in seven critical and unique 
isothermal sections of the wollastonite 
perovskite-hematite plane. 

The stable phases crystallizing on the 
join andr-Ti-gar (Fig. 41) are garnet 
solid solution (gar ss ) , pseudowollastonite 
(pwo), hematite (hem), and perovskite 
solution (pv ss ). Garnet solid solutions 
have compositions that correspond to the 
join andr-Ti-gar; perovskite solid solu- 



TABLE 8. Chemical Composition, Molecular Norm, and Hypothetical 
Garnet Components in Mole % from Two Ti-Bearing Garnets 





1 


2 




1 


2 


Chemical Composition 


Garnet Components 




SiO a 


27.94 


27.89 


Na-melanite 


3.4 




Ti0 2 


12.10 


15.51 


Mg-melanite 


3.6 


16.6 


Al 2 O s 


5.17 


2.12 


Fe-melanite 


12.4 


15.3 


Fe 2 3 


17.47 


18.32 


Melanite 


16.9 


22.7 


FeO 


3.26 


2.91 


Khoharite 






MnO 


0.27 


0.57 


Skiagite 


3.6 


3.1 


MgO 


0.31 


1.22 


Andradite 


36.0 


32.5 


CaO 


31.90 


31.79 


Grossularite 


24.1 


10.4 


Na 2 


0.52 










K 2 


0.14 




Molecular Norm 




H 2 


0.59 
99.93 




Na-melanite 
hem + c 


3.4 
19.2 




Totals 


100.33 


17.3 








pv 


17.6 


24.8 








en 


0.9 


4.0 








fs 


6.1 


6.1 








wo 


52.8 


47.8 



1. Melanite from phono lite, Oberrotweil, Kaiserstuhl volcanic area, Ger- 
many (Zedlitz, 1933). 

2. Melanite from biotite-garnet ijolite, alkaline igneous complex, Magnet 
Cove, Arkansas (Erickson and Blade, 1963) . 



142 



CARNEGIE INSTITUTION 



0.5 



0.3 



0.1 



1 1 1 1 — 1 1 I I 1 



1 — 1 — T 



I- 




_ j_ wollastonite (310) 
garnet ss (400) 



days I 



5 10 20 

Log time at 1050° C 



70 90 



Fig. 40. The solution of metastably formed wollastonite in the garnet solid solution of 
andradite S oTi-garaetio as a function of time, based on relative intensities of the (310) and (400) 
reflections, respectively. 



tion, pseudowollastonite, hematite and, 
above 1268 °C, even the liquid are on 
the ternary join wo-pv-hem. 

The maximum degree of solid solution 
of 3CaO-Fe 2 3 -3Ti0 2 in the stable gar- 
net is 54.5 wt % at a temperature of 
1137°C. A composition of 3CaO-Fe 2 3 - 
1.5Ti0 2 -1.5Si0 2 is obtained close to 
1000°C. Above 1137°C pure andradite 
is no longer stable as a single phase 
(Huckenholz, Schairer, and Yoder, Year 
Book 66), and garnet ss of compositions 
ranging from andri 00 Ti-gar to andr 80 
Ti-gar 20 yield the assemblage pwo + 
gar ss + hem and pwo + hem before they 
begin to melt at 1268° ±2°C. Ti-gar- 
net ss melts incongruently to perov- 
skite S8 + liquid at 1315° ±2°C and at a 
composition greater than 46.5 but less 
than 50 wt % 3CaO-Fe 2 3 -3Si0 2 . The 
composition with the maximum thermal 
stability is believed to be close to 48%, 
as depicted in Fig. 40. It was not pos- 
sible to synthesize an andradite ss con- 
taining 69% of the Ti-garnet component, 
believed by Ito and Frondel (1967) to 
have been synthesized by them. The 
Ti-garnet component in the garnet S8 does 



not exceed 55 wt %, and at composi- 
tions greater than 55% the polyphase as- 
semblages of pv ss + gar ss + hem or pv ss + 
hem are formed. The latter assemblage is 
restricted to a composition less than 
andr 7 Ti-gar 93 . 

In order to set up suitable determina- 
tive procedures for Ti-andradite ss , the 
unit-cell parameter a was measured on 
garnets along the join andr-Ti-gar with 
the use of material that had been held at 
temperatures of 1100°, 1175°, 1225°, 
1280°, and 1300°C for periods of time 
required to obtain equilibrium (Fig. 42). 
The high-angle reflections (640) and 
(642) of the garnet ss were measured 
against (300) and (024) reflections of 
KBr0 3 , which are located at 52.723° and 
57.391° 26 CuKa for the entire range 
of the Ti-andradite ss obtained in this 
study. Two patterns were run for a 
single garnet composition, and the dif- 
fraction chart was read ten times for 
every peak. The readings are reproduci- 
ble to better than 0.01° 26, and the dif- 
ference between a calculated from (642) 
and (640) is rarely larger than 0.003 A. 

The cell parameter for pure andradite 



GEOPHYSICAL LABORATORY 



143 



o 



1400° 


1 | 1 I 


1 1 1 1 J 1 ' 


1 
• 


1 1 ' 1 


1 1 


343i2°< 


\ LIQUID 


/ P v ss + L J 




pv^+hemt L 


/ 

/ 




- pwotLV 
\\^ ■— -g. / 


1 I3I512' / 
Jo -Z— O— — Qrri^^J 




o ,310 -' 2 * 


129312°' 


^*f 9 ar ss + L jT J 


i 






\ hem 
\ 




/pwo+hem+L /" 


££0? — pwo+gar ss +L 








\ 




/ pwotgar ss / 










\ 




/ .hem J 




9 






\ 


1200° 


I pwo+ / 
J hem / 










1 - 




/ 


gar ss « 




pv ss +gar ss +hem 




1 157*5°' 


, / 










' _ 


1 137±5°' 




o~ 










1100° 




c\i 

If. £ 






cS 1 

CO 


1 


inAno 


1 , ! 


<3 ? 

i 1 ,? 1 i 1 •? 


1 


1 


o 

o 
o 

1 .1 1 1 


1 > 



Andradite 



10 



20 



30 



40 



50 



60 



70 



80 



3CaO-Fe 2 3 -3SiO. 



Weight percent 3CaOFe 2 3 -3Ti0 2 



90 

Ti -garnet 
3CoO-Fe 2 3 -3Ti0 2 



Fig. 41. Temperature-versus-composition plot of data obtained on the join andradite-Ti-garnet 
at 1 atm pressure. Abbreviations for phases encountered: gar,,, garnet solid solution; pwo, 
pseudowollastonite ; hem, hematite; pv,,, perovskite solid solution. Open circles represent data 
obtained by optical and X-ray determinations; half-shaded circles represent data obtained by 
the determination of the unit-cell parameters of garnet,, (illustrated in Fig. 42). 



synthesized at 1 atm and at 1137°C was 
recently determined by Huckenholz, 
Schairer, and Yoder (Year Book 66) as 
12.053 ±0.003 A, using reflections (642), 
(640), (611), (521), (510), (422), (420), 
and (400), with silicon as an internal 
standard. The increase of the cell 
parameter from andri 00 Ti-gar to andr 45 
Ti-gar 55 is 0.038 ±0.002 A per 10 wt % 
and 0.040 ±0.002 A per mole %. The cell 
dimensions reported by Ito and Frondel 
(1967) for five Ti-andradites synthesized 
along the join andr-Ti-gar at 1050 °C 
lie close to the 1100°C unit-cell curve 
of this study (Fig. 42) within the limits 
of error. Their cell dimensions indicate, 
on the basis of the present interpretation, 
a termination of the garnet solid solu- 
tion between 56 and 59 wt % ( = 84 to 



88 mole % of Ito and Frondel's scale). 

Determination of cell parameters is a 
very helpful means of obtaining garnet 
compositions in polyphase assemblages, 
and the cell parameter a has been used 
in mapping field boundaries on the bi- 
nary join andr-Ti-gar as well as in the 
ternary system wo-pv-hem. At compo- 
sitions more Ti-rich than andr. 15 Ti-gar5 5 
and in the assemblage pv ss + gar ss + hem 
in which hematite is present in larger 
amounts, however, the (640) reflection 
of the garnet cannot be read with great 
precision. The hematite reflection (1126) 
at 53.88° 26 coincides with that of the 
garnet ss (640). In these determinations 
the (642) reflection of the garnet has 
been measured four times. 

The liquidus phases on the join andr- 



144 



CARNEGIE INSTITUTION 



12.250 



12.200 



-t 1 r 




° u-i300°]: 



12.250 



-0-1280°!; 



Q o-|225°-0 



-H75°i:|2.250 



2.250 



12.250 



1 100*02.250 



(2.200 



2.150 



2.100 



3CaOFe 2 3 -3Si0 2 



90 



Weight per cent 3CaOFe 2 3 3Ti0 2 3CaO-Fe 2 3 -3Ti0 2 



Fig. 42. Unit-cell parameters of garnets from compositions along the join andradite-Ti-garnet 
treated at temperatures as indicated in the diagram. Starting materials used are garnets crystallized 
at 1050°C (open circles) or finely powdered glasses (solid circles). Abbreviations as in Fig. 41. 



Ti-gar are pseudowollastonite, garnet ss , 
and perovskite ss . The composition of the 
garnet that appears on the liquidus is 
restricted between andr 86 Ti-gar 14 and 
andr 71 Ti-gar 29 . Garnet ss reacts at tem- 
peratures higher than 1315° ±2°C to 
pv ss + liquid when the time is sufficient 
(>2 hours) to obtain equilibrium. The 
primary perovskite contains small 
amounts of silicon and traces of iron, 
as determined qualitatively by micro- 
probe analyses, indicating a limited solid 
solution of CaOSi0 2 in the perovskite. 
Above 1315°C the amount of perovskite ss 
in the glass is small ; below that tempera- 
ture garnet gs is very abundant, and 
the amount of glass becomes consider- 
ably smaller approaching the solidus of 
the garnets S . A quaternary univariant 



line pierces the join andr-Ti-gar at 1310° 
±2°C and 55 wt %. The piercing point 
may be considered as a ternary invariant 
point because perovskite ss , garnet ss , 
hematite, and liquid are in equilibrium, 
assuming that all compositions lie on the 
join wo-pv-hem. 

Isothermal Sections 

In order to elucidate the phase rela- 
tionships in the ternary system wol- 
lastonite (CaO-Si0 2 ) -perovskite (CaO 
Ti0 2 ) -hematite (Fe 2 3 ), a series of iso- 
thermal sections at 1320°, 1312°, 1300°, 
1280°, 1225°, 1175°, and 1100°C are 
presented. At 1320°C (Fig. 43) the stable 
phase assemblages on the join wo-pv- 
hem are L, hem + L, hem+pVas+I/, 



GEOPHYSICAL LABORATORY 



145 



Fe 9 



1320° C 



FTs 
(Ca0-Fe 2 3 -Si0 2 ) 



(Ca0-Fe 2 3 -Ti0 2 ) 




(3CaO-Fe 2 3 -3Si0 2 ) 



Ti- garnet 
5CaOFe 2 3 -3Ti0 2 ) 



CaOSiO, 



Ca0-Ti0 2 



Fig. 43. Isothermal section of the wollastonite (CaOSiCW-perovskite (CaOTiOz) -hematite 
(Fe 2 3 ) join at 1320° C. 



hem+pv gs , pv ss +L, pwo + pv SB +L, and 
pwo + L. Garnet ss appears on the join 
by the reaction perovskite ss + L— > gar- 
net ss at 1315° ±2°C. Figure 44 depicts 
a condition slightly below this tempera- 
ture. As can be seen, perovskite S3 is in 
equilibrium with a liquid either rich or 
poor in Fe 2 3 . At 1300°C (Fig. 45) phase 
assemblages occur of gar ss with either 
hem or pv ss , or both, because perovskite ss 
is no longer in coexistence with the liquid 
rich in Fe 2 3 . The temperature for this 
reaction is 1310° ±2°C, and at this tem- 
perature the liquid, garnet ss , and perov- 
skitess lie on or very close to a straight 
line. The three solid phases gar ss , pv ss , 
and hem, as well as the liquid, are in 
equilibrium, and a ternary invariant 
point appears in Fig. 41. The assem- 



blages pv BS + L (Fe 2 3 -poor) and pv ss + 
gar 8s +L (Fe 2 3 -poor) , still present at 
1300° and 1290°C, have disappeared be- 
fore the liquid reaches the binary join 
andr-Ti-gar at 1283° ±2°C. Pwo + gar ss 
+ pvss and pwo + gar ss become the stable 
phase assemblages, and the liquid region 
lies wholly on the Fe 2 3 -rich side of the 
binary join andr-Ti-gar (Fig. 46). The 
liquid disappears from the ternary join 
wollastonite-perovskite-hematite be- 
tween 1280° and 1225 °C. The tempera- 
ture was fixed in Fig. 41 at 1268° ±2°C. 
From 1225°C (Fig. 47) there is a sub- 
stantial increase of the garnet ss series, 
which almost reaches the join wo-hem 
in the 1175°C section (Fig. 48). The 
garnet S s has closed the gap at 1100°C 



146 



CARNEGIE INSTITUTION 



Fe 2 3 



I3I2°C 



(CaOFe 2 3 -Si0 2 ), 



(CaOFe 2 3 -Ti0 2 ) 




(3Ca0-Fe 2 3 -3Si0 o )^ 



Ti-garnet 
,(3CaOFe 2 3 -3Ti0 2 ) 



CaO-SiO, 



CaOTiQ, 



Fig. 44. Isothermal section of the wollastonite (CaOSiCW -perovskite (CaO'TiOa) -hematite 
(Fe 2 3 ) join at 1312°C. 



(Fig. 49). The temperature for this 
closure is the upper stability of pure 
andradite (Huckenholz, Schairer, and 
Yoder, Year Book 66) . 

The Join Diopside-Titanium Ferri- 
Tschermak's Molecule 

In addition to the ternary join wol- 
lastonite-perovskite-hematite, a series of 
compositions were studied along the 
binary join diopside (CaO-MgO-2Si0 2 )- 
titanium ferri-Tschermak's molecule 
(CaO-Ti02-Fe 2 3 ) on the plane diop- 
side-perovskite-hematite, within the en- 
statite - wollastonite - perovskite - hematite 
system (Fig. 39), in order to obtain 
information about the phase relation- 
ships of clinopyroxene and garnet. The 



stable phases along the join diopside 
(di) -titanium ferri-Tschermak's mole- 
cule (TFTs) are clinopyroxene solid 
solution (cpx ss ), garnet solid solution 
(gar sg ), hematite (hem), and perovskite 
solid solution (pv ss ) (Fig. 50) . From the 
results obtained, clinopyroxene ss and 
magnetite ss appear on the liquidus in 
the di-rich part of the join. There is a 
very limited range of solid solution of 
TFTs in the diopside. The solid solu- 
tion is less than 5 wt % TFTs and is 
believed to be of the order of 2% at 
temperatures between 1000° and 1150°C. 
Beyond 2% TFTs, clinopyroxene ss is no 
longer stable as a single phase, and as- 
semblages of cpx ss + gar ss , cpx ss + gar ss + 
hem, cpx 8S + gar ss + hem-|-pv ss , gar ss + 



GEOPHYSICAL LABORATORY 



147 



!300°C 



(CaOFe 2 3 -Si0 2 ) 




(CaOFe 2 3 -Ti0 2 ) 



(3CaOFe 2 3 -3SiQ 



Ti-garnet 
^CaOFe^-STiCy 



CaOSiO, 



CaO-TiO„ 



Fig. 45. Isothermal section of the wollastonite (CaO-Si0 2 )-perovskite (CaOTi0 2 ) -hematite 
(Fe 2 3 ) join at 1300°C. 



hem+pv ss , and hem + pv ss are intersected 
bythejoindi-TFTs. 

The cpx ss in the polyphase assem- 
blages cannot lie on the plane di-pv-hem, 
because tie lines from garnet solid solu- 
tions on or close to the plane wo-pv-hem 
in the en-wo-pv-hem system to diopside 
do not intersect the plane di-pv-hem. 
For this reason the cpx ss composition 
must contain the enstatite component. 
A least-squares refinement of X-ray dif- 
fraction measurements was carried out 
on the composition dis TFTs 2 o, which 
was treated at 1150°C for 8 days. The 
cell parameters (a = 9.78 ±0.02 A, b = 
8.94 ±0.01 A, c = 5.29 ±0.02 A, 7 = 443.7 
±1.4 A 3 ) agree with those obtained for 
ferri-diopside S8 (Huckenholz, Schairer, 



and Yoder, Year Book 66) within the 
limits of error. The angle /3 (106.08° ± 
0.16°) indicates, however, that the clino- 
pyroxene is not simply ferri-diopside but 
is also a member of the diopside-enstatite 
solid solution series (Clark, Schairer, 
and de Neufville, Year Book 61). The 
substitution of Mg for Ca in eight-fold 
positions is indicated by an increase of 
/? compared with pure diopside and ferri- 
diopside. Therefore, the clinopyroxene ss 
in coexistence with garnet ss is considered 
to be of diopside composition, containing 
both ferri-Tschermak's and enstatite 
molecules in solid solution. 

The cell edges of the garnet along the 
join di-TFTs are 12.170 and 12.180 A in 
the two-phase volume cpx 88 -l-gar sa ; and 



148 



CARNEGIE INSTITUTION 



I280°C 



FTs 
(CaOFe,0,-SiOJ 




(CaOFe 2 3 -Ti0 2 ) 



pwo+hem 



(3Ca0Fe 2 3 -3Si0 o l 



Ti -garnet 
'5CaOFe 2 3 -3Ti0 2 ) 



CaOSiO, 



CaOTiOo 



Fig. 46. Isothermal section of the wollastonite (CaO-SiC^-perovskite (CaOTiOa) -hematite 
(Fe 2 3 ) join at 1280°C. 



12.207, 12.208, and 12.201 A in the three- 
phase volume cpx ss -t-gar ss -r-hem, re- 
spectively. These cell edges refer to the 
solid solution series of andradite-Ti- 
garnet. The relationship will probably 
change, however, if Mg enters the andra- 
dite structure. There is a very limited 
substitution for andradite (3CaOFe 2 3 - 
3Si0 2 ) of khoharite (3MgOFe 2 3 - 
3Si0 2 ) , as demonstrated by Huckenholz, 
Schairer, and Yoder (Year Book 66). 
The substitution of MgTi 4+ for 2Fe 3+ was 
tested for this investigation along the 
join andradite (3CaO-Fe 2 3 -3Si0 2 )- 
Mg-melanite (3CaO • MgO • Ti0 2 • 3Si0 2 ) 
(Fig. 39). Compositions of andr 66 . 7 Mg- 
mela 33 . 3 and andr Mg-melai 00 were 
treated at temperatures of 1050°, 1150°, 
and 1200°C. Polyphase assemblages of 



garnet ss + wollastonite + rnelilite ss and of 
clinopyroxene ss + perovskite ss + melilite S8 , 
respectively, were obtained under the 
applied conditions. The garnet of the 
andr 66 . 7 Mg-mela 33 . 3 composition treated 
at 1150°C has a cell edge of 12.129 + 
0.003 A. It cannot be unambiguously as- 
certained whether this garnet ss belongs 
to the solid solution series of andradite- 
Ti-garnet or of andradite-Ti-garnet-Mg- 
melanite. The substitution of MgTi 4+ for 
2Fe 3+ will not substantially change the 
cell edge of the garnet ss because the inter- 
atomic distances of O to Fe 3+ (Fe 3+ -0 = 
2.01 A), Mg (Mg-O = 2.07 A), and Ti 4+ 
(Ti 4+ -O = 2.01 A) for six-fold coordina- 
tion are very similar. At present, the 
garnets stable on the join andradite-Mg- 



GEOPHYSICAL LABORATORY 



149 



FTs 
(Ca0-Fe 2 3 -Si0 2 y 



pwo+hem 



Andr 
(3CaOFe 2 3 -3Si0 2 ) 



225°C 




,(CaO-Fe 2 3 Ti0 2 ) 



Ti- garnet 
— ^h(3CaO'Fe 2 3 -3TiCy 



CaOSiO, 



CaOTiO, 



Fig. 47. Isothermal section of the wollastonite (CaO-Si02)-perovskite (CaOTiOa) -hematite 
(Fe 2 3 ) join at 1225°C. 



melanite and on the join di-TFTs are 
considered to be members of an andra- 
dite-Ti-garnet-Mg-melanite solid solu- 
tion series containing less than 33.3 wt % 
of an Mg-melanite component. However, 
the Mg-melanite component does not ap- 
pear to exceed 20 wt % in natural Ti- 
rich garnets. 

Geologic Discussion 

The most common rock types in which 
Ti-bearing garnets are abundant min- 
erals belong to the nepheline syenite and 
ijolite families. They are often directly 
related to carbonatites and their altera- 
tion products, the fenites. The main 
feature of these rocks is the presence 



of the mafic mineral association of clino- 
pyroxenes and of Ti-bearing garnets (re- 
ported as melanite, schorlomite, and 
iiaavarite) . Wollastonite, perovskite, and 
iron ores (magnetite, titanomagnetite, 
ilmenite, and occasionally hematite) are 
sometimes abundant phases too. Biotite, 
amphibole, and sphene may also occur. 
The composition of the clinopyroxene 
varies from aegirine or aegirine-augite 
to diopside, sometimes rich in ferri- 
Tschermak's molecule. Even pigeonitic 
augites that contain melanite are re- 
ported from a pyroxenite and foyaite- 
porphyrite dike, from Alno, Sweden 
(von Eckermann, 1948, 61 and 91). The 
associated garnet is andraditic in com- 
position in most cases and contains in 



150 



CARNEGIE INSTITUTION 



II75°C 



hem+pv ss 



(CaOFe 2 3 -Si0 2 ) 




(CaOFe 2 3 -Ti0 2 ) 



pwo + hem 



(3CaOFe 2 0,-3SiO,), 



Ti-garnet 
;3CaO-Fe 2 3 -3Ti0 2 ) 



CaOSiO, 



CaO-TiO c 



Fig. 48. Isothermal section of the wollastonite (CaO-Si0 2 )-perovskite (CaO-TiC^) -hematite 
(Fe 2 3 ) join at 1175°C. 



addition other components, as illustrated 
in Table 8. In 60 analyses of Ti-bearing 
garnets a relatively high Ti content in 
addition to Fe 3+ is always related to a 
low Al content. This is in direct contrast 
to the associated pyroxenes: their Ti- 
content increases with increasing Al, a 
behavior outlined in general for clino- 
pyroxenes by Kushiro (1960). Unfortun- 
ately, data about coexisting clinopyrox- 
enes and garnets from alkaline igneous 
rock are rare. Examples are known from 
ijolites from Iron Hill, Colorado (Lar- 



sen, 1941) , the Napak volcanoes, Uganda 
(King, 1949), and the alkalic igneous 
complex at Magnet Cove, Arkansas 
(Erickson and Blade, 1963), and from 
a wollastonite-melanite melteigite from 
the alkaline complex at Oka, Quebec 
(Gold, 1966). The coexisting minerals 
or minerals from different hand speci- 
mens of similar rocks were plotted on an 
atomic Ti-Fe 3+ -Al basis in Fig. 51. It 
can be seen that the Napak garnet, rela- 
tively high in Al, has a very low Ti 
content. The Magnet Cove garnet is low 



GEOPHYSICAL LABORATORY 



151 



IOO°C 



(CaOFe 2 3 -Si0 2 ), 




v (CaOFe 2 3 Ti0 2 ) 



(3CaO-Fe 2 3 -3Si0 2 \ 



Ti-garnet 
3CaOFe 2 3 -3Ti0 2 ) 



CaOSiO. 



CaO-TiO, 



Fig. 49. Isothermal section of the wollastonite (CaO-Si0 2 )-perovskite (CaO-TiCW -hematite 
(Fe 2 3 ) join at 1100°C. 



in Al but is very high in Ti. An inter- 
mediate position is held by the garnet 
from Iron Hill. The Oka garnet is al- 
most a pure andradite, having only traces 
of Al and a low Ti content. It can be 
concluded, therefore, that Al-rich grand- 
ites are not expected to be rich in Ti. 

The phase relationships in the ternary 
joins diopside-wollastonite-hematite 
(Huckenholz, Schairer, and Yoder, Year 
Book 66, p. 344) , wollastonite-perovskite- 
hematite, and diopside-perovskite-hema- 
tite, and the binary join andradite-Mg- 



melanite, as well as the behavior of 
natural clinopyroxenes under oxidizing 
conditions (Huckenholz, elsewhere in this 
report), bear directly on the formation 
of andradite and its Ti-rich varieties 
under magmatic conditions. The results 
of these studies suggest the concept that 
the appearance of Ti-bearing garnets and 
ferri-diopsides is the natural consequence 
of crystallization at or near atmospheric 
conditions and is not necessarily the re- 
sult of unique deep-seated processes. 



152 



CARNEGIE INSTITUTION 



^3- 



-3- 



-D> 







,E ~~- — — -J§_____ - 







to 

o 

CM 



o aj 



CM 



£ E =» o 
3 <5 u i~ 
e -c .2 h" 

°^ p9 

•,— i— c o 



O 



CVJ 

O 



O 
o 
o 



g» 



o •ajnpjaduuaL 



a s 

o > 

•43 ft 

'£ d 

.Q ... 

<! .o 
■s~ 

C3 o 



S3 . S 

CU -f^ 
IS 

CQ A 



& a 

oo <u a 
EH {So 

-a J § 

.a a-3 

.t* -~ cu 
■f 

4.2 a 

-d 43 bfi 

'S -2 be 
^ o d 
.2 ™ "43 
-p d xj .2 

.9 o ^ 

O CQ O 

-t-3 

CD 0) 0) 

- 1- 

§ M o 

CU 

.9 bb - 

* -1 

■8 9 8 

O O o3 

^ o'oj 



- CU 
c3 " 



**"' PT-I += 

O ^ "A 

-=; d 

43 O 3 
O CQ 

ft CU S 



u o 



d 
o 

'43 O sh 

CQ 5.«4-l 

ft o 

S.S 2 

8-i.S 
i :| 

CU fti 2 ! 

d 

o 



CU 



d 



d 

+= H 

o3 CU 

CD d CQ 

alS 

o 



CU 

cq 42 



"5 co jb£| 

. Oj CQ 

faD-G !> 
- ft O 

«M CU 

O ft 



s 



GEOPHYSICAL LABORATORY 



153 




Fe +3 



Fig. 51. Coexisting clinopyroxenes and Ti-bearing garnets from alkaline igneous complexes: 
(1) Iron Hill, Colorado; sodian augite and melanite from ijolite IH-129 (Larsen, 1941). (2) Napak, 
Uganda; aegirine augite and melanite from ijolite K.352 and K.372, respectively (King, 1949). 
(3) Magnet Cove, Arkansas; diopside from biotite-garnet ijolite L-123-6 and L-123-7, garnet from 
biotite-garnet ijolite MC-216-8 (Erickson and Blade, 1963). (4) Oka, Quebec; sodian augite and 
melanite from wollastonite-melanite melteigite DDH G 15, 330 feet (Gold, 1966). 



SILICATE SYSTEMS INCLUDING A 
VAPOR PHASE 

Melting of Forsterite and Enstatite 

at High Pressures under 

Hydrous Conditions 

I. Kushiro and H. S. Yoder, Jr. 

The effects of water on the melting 
behavior of forsterite (Mg 2 Si0 4 ) and 
enstatite (MgSi0 3 ), the main constitu- 
ents of peridotites, have been studied in 
the pressure range of 2 to 30 kb. The 
modified gas-media apparatus designed 
by Yoder (1950) was used for runs con- 
ducted at and below 10 kb, and a piston- 
cylinder, solid-media apparatus similar 
to that designed by Boyd and England 
(1960) was used for runs above 10 kb. 
Sealed platinum capsules with an out- 
side diameter of 3.0 mm were used for 
the gas-media apparatus, and sealed 
platinum capsules with an outside diam- 
eter of 1.8 mm were used for the piston- 



cylinder apparatus. The water content 
ranged from 3.8 to about 50 wt %. In 
the piston-cylinder apparatus, the vacant 
space between the capsule and the 
ceramic sleeve was filled with powdered 
crushable alumina in order to reduce the 
deformation of the capsule. Starting ma- 
terials were MgSi0 3 glass made by the 
late Dr. N. L. Bowen, orthoenstatite 
crystallized from the MgSi0 3 glass at 
10 kb, and pure forsterite synthesized 
with the help of Dr. J. F. Schairer. 

Melting of Forsterite 

The beginning of melting of forsterite 
under hydrous conditions is considerably 
lower than under anhydrous conditions 
at the same pressure. At 10 kb, clear 
granular crystals of forsterite (0.03- 
0.15 mm in diameter) were obtained 
from finely ground forsterite exposed to 
H 2 at temperatures between 1350° and 



154 



CARNEGIE INSTITUTION 



1500°C. On the other hand, opaque- 
looking, irregular crystals of forsterite 
containing numerous fine bubblelike in- 
clusions, sometimes arranged regularly 
to form a herringbonelike texture, were 
obtained when the same starting ma- 
terials were held at 1525 °C. The inclu- 
sions are glass and liquid with or without 
gas bubbles and are probably formed 
from liquid and vapor that existed at the 
pressure and temperature of the runs 
and was trapped in rapidly grown 
forsterite crystals during the quenching. 
The relatively opaque-looking forsterite 
crystals with numerous inclusions are, 
therefore, interpreted to be quench crys- 
tals. In some liquid inclusions, a bubble 
exists and is observed moving rapidly, 
probably due to the Brownian action 
of the liquid. Because of the distinct dif- 
ference between the melted and the un- 
melted textures of forsterite, the tem- 
perature of the beginning of melting of 
forsterite can be located with precision 
between 1500° and 1525°C at 10 kb. 

At 20 kb the same distinct difference 
in texture was observed between the runs 
made at 1400° and 1425°C for H 2 
contents greater than at least 23 wt %. 
In runs with H 2 contents of about 12 
and 15 wt %, however, clear granular 
forsterite crystals coexist with quench 
forsterite crystals at 1425° and 1450°. 
Theoretically, complete melting occurs in 
the vapor-present region (excess water 
region), whereas forsterite crystals can 
exist stably at temperatures above the 
beginning of melting in the vapor-absent 
region (water-deficient region), although 
the temperatures of the beginning of 
melting are the same for both the water- 
deficient and excess-water regions. The 
experiments with 12 and 15 wt % H 2 
are interpreted, therefore, as being in 
the water-deficient region. The maximum 
content of water that dissolves in a 
forsterite melt is, therefore, between 15 
and 23 wt % at 20 kb. At 10 kb the run 
with 10.0 wt °/o H 2 does not contain 
primary forsterite at a temperature just 
above the melting temperature, and thus 



the maximum water content in the liquid 
is less than 10 wt %. 

Fig. 52 shows the beginning of melting 
of forsterite determined on the basis of 
the runs in the excess-water region. As 
shown in the figure, the melting tempera- 
ture drops greatly with increasing pres- 
sure under hydrous conditions. The tem- 
perature of melting of forsterite at 20 kb 
under hydrous conditions, determined 
by Sclar, Carrison, and Stewart (1968a) 
by a belt apparatus, is about 80°C lower 
than the present data. The temperature 
difference between the melting curve 
under anhydrous conditions determined 
by Davis and England (1964) and that 
under hydrous conditions is about 450 °C 
at 10 kb and about 670°C at 30 kb. The 
dT/dP of the melting curve probably 
becomes smaller with further increase 
of pressure. This results from the smaller 
volume difference between forsterite + 
vapor and liquid with increasing pressure 
due to the larger compression of vapor 
relative to liquid at high pressures. 

Melting of Enstatite 

Sixty runs have been carried out on 
the MgSi0 3 composition with various 
amounts of water in the pressure range 
2 to 30 kb. The detailed results are given 
in a separate paper (Kushiro, Yoder, and 
Nishikawa, 1968), and a brief sum- 
mary is reported here. 

The temperatures of the beginning of 
melting of enstatite are 1360° ±15° at 
10 kb, 1280° ± 15° at 20 kb, and 1270° ± 
15 °C at 30 kb, as shown in Fig. 53. Just 
above these temperatures, euhedral to 
subhedral crystals of forsterite, as well 
as herringbonelike and radially grown 
crystals with or without patches of glass, 
glass globules, and glass coatings on 
very thin needlelike crystals, which may 
be enstatite, were obtained in the pres- 
sure range 10 to 30 kb. The herringbone- 
like and radially grown crystals are con- 
sidered to be quench crystals; that is, 
they were formed from liquid during the 
quenching. The quench crystals consist 



GEOPHYSICAL LABORATORY 



155 



1200 



1000- 



800, 




Fo+V 



10 20 

Pressure, Kb 



Fig. 52. Melting curve of forsterite under water-saturated conditions. Melting curve of 
forsterite under anhydrous conditions is from Davis and England (1964). Fo, forsterite; L, liquid; 
V, vapor. 



of orthoenstatite or orthoenstatite sepa- 
rating forsterite and quartz. The pres- 
ence of forsterite and quartz, although 
not in contact, clearly represents a meta- 
stable assemblage. The glass globules 
and coatings are concentrated more 
along the inner surface of capsules and 
have refractive indices lower than those 
of patches of glass quenched from liquid. 
They are considered to be quenched 
vapor; that is, they were formed directly 
from the materials in the vapor that 
existed during the runs. 

Formation of the forsterite + liquid + 
vapor assemblage from >MgSi0 3 + H 2 
composition indicates the incongruent 
melting of enstatite in the presence of 
water vapor. The temperature interval 
where forsterite and liquid coexist de- 
pends on the water content, and these 
relations are shown on the MgSi0 3 -H 2 
join at 10 kb (Fig. 54) . For H 2 contents 
less than 11 wt %, forsterite coexists 
with enstatite and liquid or with liquid 
only, whereas for H 2 contents greater 



than 11 wt %, forsterite coexists with 
liquid and vapor. The upper tempera- 
ture limit of the presence of forsterite 
ranges from about 1400° to about 1540°C 
in the water-deficient region, and from 
about 1400 °C to somewhat higher tem- 
peratures ( + 10° to ~20°) in the excess- 
water region for H 2 contents of 11 to 
30 wt %. The upper temperature limit 
of the forsterite-)- liquid + vapor region in 
Fig. 53 is for limited H 2 saturation 
(less than 50 wt % H 2 0). Above the 
forsterite + liquid + vapor region, liquid 
and vapor coexist, as shown by the 
evidence of quench crystals and ma- 
terials quenched from vapor. 

At temperatures below the beginning 
of melting and above 1000°C, prismatic 
enstatite (0.01-0.1 mm in length) with 
small amounts of round forsterite and 
glass globules and coatings on very thin 
needlelike crystals is obtained in most 
of the runs. The glass globules and coat- 
ings are considered to be materials 
quenched from vapor. The presence of 



156 



CARNEGIE INSTITUTION 



1800 


I 




1 


1 


I 


Drv^- 


1 


i 


- 


1700 


~~ 






^^"^En 










— 


1600 


















- 


o l500 

o 

|l400 

CD 

£ 

<D 


Pr + Fo + V / 




a 

B 

B 


L+V 

B 


B 

_ B 


B 
B 




B 
B 


- 


H 1300 


/ 




■ 
■ 
■ 




-— JL 
■ 


ffl 


a 


a 
a 


_ 






■ 


a 




1200 


- /■ 




■ 


En + Fo + V 


■ 
■ 


a 


■ 
■ 


a 


- 


1100 








■ 
■ 


■ 


■ 




a 


— 


1000 












■ 




a 


- 


900 


i 




I 


i 


i 




i 


l 


- 



15 
Pressure, Kb 



20 



25 



30 



Fig. 53. Melting relations for MgSi0 3 composition under water-saturated conditions. The 
upper limit of the Fo -\- L -{- V field is a projected liquidus for limited excess water (<50 wt %). 
The melting relations under anhydrous conditions are from Boyd, England, and Davis (1964). 
The inversion point between orthoenstatite and protoenstatite at 1 atm is from Atlas (1952). 
En, orthoenstatite; Pr, protoenstatite; other abbreviations as in Fig. 52. 



GEOPHYSICAL LABORATORY 



157 



1600 



1500- 



o 



■fi 1400 



CL 

E 



1300- 



1200 







10 20 


30 


gSi0 3 


Weight percent 


^H 2 



Fig. 54. Phase-equilibrium relations on the 
join MgSi0 3 -H 2 at 10 kb. Abbreviations as in 
Figs. 52 and 53. 

a small amount of forsterite indicates 
that the vapor dissolves excess silica 
(that is, silica in excess of the propor- 
tionate amount in MgSi0 3 ). 

Application of Results 

The most important result of the pres- 
ent experiments is that enstatite melts 
incongruently to forsterite and liquid in 
the presence of water in the pressure 
range of at least 10 to 30 kb. Under 
anhydrous conditions, the incongruent 
melting of enstatite terminates at pres- 
sures higher than about 5 kb (Boyd, 
England, and Davis, 1964). Even if the 
water content is small (in the vapor- 
absent or water-deficient region), enstat- 
ite still melts incongruently to form 
enstatite, forsterite, and liquid. The 
liquids coexisting with forsterite in the 
water-deficient region are silica satu- 
rated or more silica rich than the pro- 
portion for the MgSi0 3 composition. This 



is an important conclusion, which bears 
on the origin of silica-saturated magmas, 
such as quartz-tholeiite magma. 

If partial melting of a rock consisting 
of enstatite and forsterite with a small 
amount of uncombined water takes place 
at pressures up to at least 30 kb, the 
first liquid to form is silica saturated 
and water saturated, containing 11 wt % 
water or more. With increasing tempera- 
ture the amount of liquid increases, the 
water content in the liquid decreases, 
and the degree of silica saturation de- 
creases; however, the liquids are still 
silica saturated. The presence of diopside 
does not change this conclusion (Kushiro, 
elsewhere in this report). Although the 
present system is far simpler than nat- 
ural rocks, it is suggested that silica- 
saturated magmas could be generated by 
partial melting of the upper-mantle ma- 
terial, believed to be olivine-rich lherzol- 
ite, in the presence of a small amount 
of uncombined water at depths at least 
down to 100 km. 

The present results also may be appli- 
cable to the formation of silica-saturated 
magmas by fractional crystallization of 
basaltic magmas containing water at 
high pressures. If crystallization of a 
magma, just silica saturated and con- 
taining a small amount of water, takes 
place at high pressures, forsterite will 
crystallize with enstatite with lowering 
temperature, and residual liquids will be- 
come silica saturated. With further de- 
crease of temperature, the degree of silica 
saturation will increase. 

As shown in Figs. 52 and 53, the 
temperatures of the beginning of melting 
of forsterite and enstatite, which have 
been determined in the presence of ex- 
cess vapor (PH 2 o=-Ptotai), are surpris- 
ingly low at high pressures. Some of 
the possible geotherms (e.g., oceanic geo- 
therm, Clark and Ringwood, 1964) ap- 
proach closely even to the beginning 
of melting of forsterite under hydrous 
conditions (Fig. 52) at pressures higher 
than 30 kb. For a mixture of forsterite 
and enstatite, melting begins to take 



158 



CARNEGIE INSTITUTION 



place at temperatures on the curve shown 
in Fig. 53, when the water pressure is 
equal to total pressure. It is suggested, 
therefore, that if the water pressure is 
equal to or close to total pressure at 
depths near 100 km or deeper in the 
upper mantle, melting of these parts may 
begin to take place. Water pressure in 
the upper mantle may be equal to or 
close to total pressure if water vapor or 
fluid exists as a separate phase in the 
upper mantle, regardless of its amount. 

Liqtjidus Relations in the System 

FORSTERITE-DlOPSIDE-SlLICA-H 2 
AT 20 KB 

I. Kushiro 

The system forsterite-diopside-silica, 
one of the most important systems for 
the understanding of the crystallization 
of basaltic magmas, has been studied 
at 20 kb under hydrous conditions in 
order to determine the effect of water 
on the course of crystallization of basal- 
tic magmas at high pressures. 

All the runs were carried out with a 
piston-cylinder, solid-media apparatus 
similar to that designed by Boyd and 
England (1960). Sealed platinum cap- 
sules with water contents of 25 to 50 
wt % were used, and all the charges 
were wet after the runs. A small vacant 
space between the capsule and the cer- 
amic sleeve was filled with powdered 
crushable alumina to reduce the deforma- 
tion of the capsule. Twenty-two compo- 
sitions were selected in the system for- 
sterite-diopside-silica. The starting ma- 
terials for fifteen compositions were crys- 
tallized glass prepared by Schairer and 
Yoder (Year Book 61, pp. 75-82), Ku- 
shiro and Schairer [Year Book 62, pp. 
95-103), and Boyd and Schairer (1964). 
Mechanical mixtures of cristobalite, 
clinoenstatite, and diopside were used 
for seven compositions. 

The System Forsterite-Silica-H 2 

The results of the runs on the system 
forsterite-silica-H 2 at 20 kb are shown 



in Fig. 55, which is a projection of the 
ternary system onto the join forsterite- 
silica from the H 2 apex. The mixtures 
of compositions more silica rich than 
Fo 40 Qz 60 are mechanical mixtures of 
clinoenstatite and cristobalite. The re- 
sults on the Mg 2 Si0 4 and MgSi0 3 compo- 
sitions are from Kushiro and Yoder 
(elsewhere in this report). As shown in 
Fig. 55, the melting temperature of silica 
(quartz) is greatly lowered as compared 
with that of enstatite; consequently, the 
primary phase field of enstatite is greatly 
expanded toward silica and the primary 
phase field of quartz is greatly reduced, 
as compared with those at 1 atm given 
by Bowen and Andersen (1914) and 
Greig (1927). The forsterite liquidus ex- 
tends beyond enstatite composition to 
about Fo 61 Qz 39 , where forsterite reacts 
with liquid to form enstatite in the pres- 
ence of vapor. The "eutectic point" be- 
tween enstatite and quartz, which is a 
projection of a point on the four-phase 
curve En + Qz + y + L (or an isobaric 
invariant point) exists at about Fo 14 Qz 86 
composition. 

All the runs quenched from tempera- 
tures above the solidus contain glass 
that has numerous bubble inclusions. 
Glass globules, which are probably 
quenched vapor (Kushiro and Yoder, 
this report), were also obtained, except 
for compositions close to silica (>85 wt 
% Si0 2 ), indicating that vapor was 
present during the runs. It is not certain, 
however, whether both vapor and liquid 
or only a supercritical fluid existed above 
the solidus for compositions close to 
silica. In the En + L + F region, ortho- 
enstatite crystals are prismatic and 
euhedral, and are most likely stable. It 
is possible, however, that metastable 
enstatite persists even above the solidus. 
To test this possibility, the mixture of 
composition Fo 2 oQz 8 o was held at 1250° C 
(well above the liquidus) for 45 minutes 
and then at 1100°C (just below the 
liquidus) for 90 minutes. In this run a 
small amount of large crystals of 
euhedral enstatite was obtained. It is 






GEOPHYSICAL LABORATORY 



159 



1900 


1 1 


i i I i i i 
latm 


i 


1800 


\. 


L 

I Two L 


\ 


1700 


Fo+L \ 


i 


V 


1600 


Pr y.YU Cr + L 


} 

Cr + L 


1500 


_ 




_ 


<l> l400 




20 kb Ph 2 




► ■ — -^.^^ 




- 


2 1300 


Fo + L + V^i 

<■ 


k L + V 


- 


CD 


< 




J 1200 
1 — 

1100 


< 

< 
Fo + En + V 


En + L + V e |^ 






< 


► e 


\ ° i 


1000 


1 l 


e e e 


W^rf 


900 


c c 

En + Qz+V 

i > i i i 


Qz+L+V 

i 



Mg 2 Si0 4 



MgSi0 3 50 

Weight per cent 



SiO, 



Fig. 55. Projection of the system forsterite (Mg 2 SiO<) -silica-water at Ph 2 o = 20 kb. Phase 
relations at 1 atm are from Bowen and Andersen (1914) and Greig (1927). Cr, cristobalite; En, 
orthoenstatite ; Fo, forsterite; L, liquid; Pr, protoenstatite ; Qz, quartz; V, vapor. 



concluded that the liquidus of enstatite 
shown in Fig. 55 is not metastable. 

Quartz crystals in the Qz + L + V re- 
gion showed round or square forms and 
appeared to be high quartz, although 
those with double terminated pyramids 
were rarely observed. 

The System Forsterite-Diopside- 
Silica-H 2 

The liquidus boundaries between the 
fields of forsterite, diopside, and enstatite 
solid solutions and quartz at 20 kb 
under hydrous conditions are shown in 
Fig. 56. They have been determined in 
the presence of excess vapor (Ph 2 o = 
Ptotai) and are projections of the bound- 



aries between the primary, solid-phase 
volumes in the quaternary system onto 
the forsterite-diopside-silica plane from 
the H 2 apex. 

As shown in Fig. 56, the liquidus fields 
of both enstatite and diopside solid solu- 
tions are greatly expanded toward silica 
and the liquidus field of quartz is much 
reduced as compared with those at 1 atm. 
The forsterite field covers the enstatite- 
rich part of the diopside-enstatite join, 
as at 1 atm, indicating that forsterite re- 
acts with an array of liquids to form 
enstatite solid solution in the presence 
of water vapor. The forsterite-enstatite ss 
liquidus boundary is more silica rich 
than at 1 atm, whereas the forsterite- 
diopside BS liquidus boundary is, for the 



160 



CARNEGIE INSTITUTION 



CaMgSi 2 6 
I240±I0" 




Mg 2 SiO. 



Weight per cent 



Fig. 56. Projection of the liquidus boundaries in the system forsterite (Mg 2 SiOi)-diopside 
(CaMgSi 2 8 ) -silica-water at Ph 2 o = 20 kb. Liquidus diagram of the system forsterite-diopside-silica 
at 1 atm shown in the upper right is from Bowen (1914), Schairer and Yoder {Year Book 61, 
pp. 75-82), and Kushiro and Schairer {Year Book 62, pp. 95-103). 



most part, more forsterite rich than at 
1 atm. Forsterite crystallizing near the 
liquidus temperature at 20 kb Ph 2 o is 
essentially pure forsterite, as inferred 
from the reflections in the X-ray powder 
diffraction pattern. 

On the forsterite-pyroxene liquidus 
boundary is a projection of a point on 
the five-phase curve Fo + En ss + Di ss + 
L + V (X in Fig. 56) (or an isobaric 
invariant point), which is a reaction 
point where forsterite reacts with liquid 
to form enstatite ss and diopside ss in the 
presence of vapor. The forsterite-diop- 
side ss liquidus boundary probably has a 
temperature maximum between the joins 
diopside-enstatite and forsterite-diop- 
side, and its temperature probably drops 



continuously from this maximum to a 
critically silica-undersaturated composi- 
tion lying on the other side of the forster- 
ite-diopside join, as at 1 atm. The join 
forsterite-diopside is not binary, because 
diopside crystallizing from this join is 
not pure diopside but a solid solution 
containing a small amount of enstatite, 
as inferred from the 20(311)-20(31O) 
value in the X-ray powder diffraction 
pattern (Kushiro and Schairer, Year 
Book 62, p. 99) . 

The projection of another isobaric in- 
variant point, En ss -r-Di ss + Qz + L-r-y, 
exists close to Si0 2 (Y in Fig. 56). In 
the present experiments, it was not de- 
termined whether this is a reaction point 
or a eutectic point. The temperature of 



GEOPHYSICAL LABORATORY 



161 



the diopside ss -enstatite ss liquidus bound- 
ary drops from 1220°±15°C at X to 
960°±20°Cat7 

The system forsterite-diopside-silica- 
H 2 includes the following common rock- 
forming hydrous minerals: serpentine 
[Mg 3 Si 2 5 (OH) 4 ], talc [Mg 6 Si 8 O 20 
(0H) 4 ], anthophyllite [Mg 7 Si 8 22 
(OH) 2 ], and tremolite [Ca 2 Mg 5 Si 8 2 2 
(OH) 2 ] . In the present experiments, con- 
ducted at temperatures above 950 °C, 
however, none of these hydrous minerals 
was encountered. Tremolite is stable up 
to the highest temperature, about 880 °C 
at 2 kb (Boyd, 1959) . The volume change 
(AF) of the reaction tremolite = 2 diop- 
side + 3 enstatite + quartz + H 2 can be 
expressed as AV= — 24.9 + y H2 o (cc) 
and can be calculated by the use of the 
molar volume of water (Fh 2 o) given by 
Rice and Walsh (1957) for a wide pres- 
sure-temperature range. It was found 
that the A 7 has a negative value at pres- 
sures above about 6 kb Ph 2 o, indicating 
that the breakdown curve of tremolite 
has a negative slope at these pressures. 
On the basis of the experimental data at 
low pressures and the calculated volume 
change, it is expected that the upper 
limit of the stability field of tremolite at 
20 kb Fh 2 o is near 850°C, which is below 
the temperatures of the present experi- 
ments. 

The present experimental results may 
have an important bearing on the origin 
of silica-rich magmas such as andesite, 
dacite, and rhyolite, which commonly 
occur in orogenic regions. From the 
liquidus relations given in Fig. 56, it is 
expected that liquid undersaturated with 
silica can produce silica-rich magmas 
by fractional crystallization. By maxi- 
mum fractionation at 20 kb Ph 2 o, the 
residual liquids attain the isobaric in- 
variant point Y, the silica content of 
which is about 80 wt %. It is suggested 
that if fractional crystallization of oli- 
vine tholeiitic magmas containing water 
takes place at pressures up to at least 
20 kb, magmas more silica rich than 
basalts, such as andesite, dacite, and 



rhyolite magmas, will be produced. Even 
if the water content of the original 
magma was small, the same conclusion 
would be obtained in the light of the re- 
sults on the melting of enstatite (Kushiro 
and Yoder, this report). It should be 
noted that high pressure is not neces- 
sarily required for the formation of 
silica-rich magmas, which can be pro- 
duced by the fractional crystallization 
of water-bearing basaltic magmas at 
pressures of 5 kb or even less. In natural 
water-bearing basalt magmas the crys- 
tallization of amphibole may affect the 
course of fractional crystallization. 

The present experimental results may 
also have a bearing on the generation of 
silica-saturated basaltic magmas in an 
upper mantle composed of olivine-rich 
lherzolite. In this model, under anhy- 
drous conditions silica-saturated magmas 
can be formed only at pressures below 
6 and 7 kb either by the partial melting 
of the lherzolite or by the fractional 
crystallization of olivine tholeiitic mag- 
mas (Kushiro, 19686). Under hydrous 
conditions (Ph 2 o — Ptotai) , however, the 
first liquid to form by the partial melt- 
ing of a material consisting of forsterite 
and diopside and enstatite solid solutions 
(simple lherzolite) at 20 kb has the 
composition X (Fig. 56) , which is silica 
saturated. On the basis of the experi- 
ments on the melting of enstatite, it is 
expected that silica-saturated liquids can 
be formed at pressures up to at least 
30 kb in the presence of even small 
amounts of water. It is suggested from 
these considerations that some quartz 
tholeiites and quartz gabbros, particu- 
larly those of the calcalkali rock series, 
may have been formed in the upper 
mantle in the presence of water. 

Melting of a Hydrous Phase: 
Phlogopite 

H. S. Yoder, Jr., and I. Kushiro 

Understanding of the melting char- 
acter of a hydrous phase is essential for 
predicting explosive volcanism, estimat- 



162 



CAENEGIE INSTITUTION 



ing the water content of the mantle, and 
outlining magma trends in which hy- 
drous phases are involved. The begin- 
ning-of-melting curves for some simple 
and complex amphiboles, zeolites, and 
micas have previously been studied in 
the presence of what was believed to be 
an excess of water. The beginning-of- 
melting curves originate at an invariant 
point at elevated pressure where the 
anhydrous breakdown products, the hy- 
drous phase, liquid, and gas are pre- 
sumed to be in equilibrium. Only a few 
experiments have been made in which 
the water content is insufficient to satu- 
rate the liquid with gas, the most likely 
natural condition of a magma at genera- 
tion (Yoder, Year Book 64, pp. 82-89) . 
The principal problem now at hand is to 
account for the large amounts of water 
observed in water-saturated silicate liq- 
uids in laboratory experiments, whereas 
field deductions suggest that most natu- 
ral magmas are low in water and water- 
undersaturated under the conditions of 
generation. Experiments in the gas- 
absent region were therefore undertaken 
in the hope of resolving this apparent 
disagreement in the water content of 
silicate liquids and to ascertain the be- 
havior of hydrous minerals in the gas- 
absent region. 

Phlogopite stability fields have been 
studied by Yoder and Eugster (1954), 
Crowley and Roy (1964), Wones 
(1967a), Luth (1967), and Kushiro, 
Syono, and Akimoto (1967). Luth (1967) 
examined phlogopite in the broader scope 
of the K 2 0-MgO-Al 2 3 -Si0 2 -H 2 sys- 
tem. The invariant point marking the 
onset of melting for phlogopite was ex- 
amined in particular by Luth, and the 
sequence of curves immediately around 
that point was presented. The phases 
involved are phlogopite (Ph), forsterite 
(Fo) , leucite (Lc) , orthorhombic kalsilite 
(Ok), liquid (L), and gas (G), which 
produce six univariant reactions (indi- 
cated below by the absent phase). The 
sequence of univariant curves illustrated 
by Luth is similar to that shown in Fig. 



57, with the exception that (G) and 
(Lc) are reversed in position. His se- 
quence was based on the assumption that 
Ph = Fo + Ok +L + G, which requires that 
the water content of the liquid be less 
than that of those possible liquids that 
lie on a plane passing through the com- 
positions of Fo-Ok-Ph. He obtained no 
data in the water-undersaturated region. 
The phases reported for his several runs 
on the phlogopite-water join (black dots 
of Fig. 58) support for the most part 
the sequence of curves either as drawn 
by him or as illustrated in Fig. 57, where 
the breakdown of phlogopite is Ph = 
Fo + Ok + Lc+L, no gas being involved. 
The issue of singular import is the water 
content of the liquid or liquids involved 
in the melting of phlogopite. 

The first step was to determine the 
melting of phlogopite in the presence of 
an excess of water over a wide range of 
pressures, including those believed to 
exist in the region of magma generation. 
The results up to 37.5 kb are shown in 
Fig. 58. The invariant point and the 
curves extending to pressures lower than 
that of the invariant point are those of 
Luth (1967). The garnet-bearing assem- 
blage of Kushiro, Syono, and Akimoto 
(1967) was not encountered at the high- 
est pressure investigated. A singular 
point arises on the (G) curve of Fig. 57, 
marking the reaction Ph = Ok + Fo + 
Lc+L, which then becomes Ph+Ks + 
Lc = L + Fo and Ph = Fo + L, as shown 
by (6) and (c), respectively, in Figs. 58 
and 60. Two curves and the related 
singular points that describe the con- 
sumption of Lc and Ok in the water- 
saturated melt are not shown ; the curves 
lie below about 2 kb and above the Ph- 
absent univariant curve, however, as il- 
lustrated by Luth (1967, p. 393, Fig. 4). 
Forsterite is the liquidus phase through- 
out the range studied and at 1 atm, 
where Schairer (1954) observed a liq- 
uidus of 1628°C. The appearance of 
spinel at lower temperatures at 1 atm, 
however, indicates that loss of alkalies 
may have occurred at the very high 



GEOPHYSICAL LABORATORY 



163 



CO 



,_. (Ok) (Lc) 
(Fo) \-q 



(G) 




Temperature 

Fig. 57. Sequence of invariant curves immediately about the invariant point involving 
phlogopite (Ph), forsterite ,(Fo), leucite (Lc), orthorhombic kalsilite (Ok), liquid (L), and 
gas (G). Each curve is indicated by the absent phase. Beginning of melting in the presence of gas 
is given by curves (Fo) and (Ph). Heavy lines are those reactions exhibited by compositions on 
the join K 2 0-6MgO-Al203 , 6Si02-H20, which includes phlogopite composition. 



temperatures, and thus the liquidus of 
the ideal bulk composition may be some- 
what lower. 

The second step was to study the K 2 • 
6MgO-Al 2 3 -6Si0 2 -H 2 join at a tem- 
perature above the melting of phlogopite 
in the presence of an excess of water, 
1225°C, and 10 kb. The results of that 



study are shown in Fig. 59 in the broad 
view of the plane Fo-H 2 0-Lc:Ks (1:1 
mole). The region marked Fo+ L + G is 
based on those run products that appeared 
after quenching as : euhedral or subhedral 
forsterite crystals ( = Fo); clear glass 
( = L), usually highly vesiculated; balls 
or coatings of a pinkish glass, having a 



164 



CARNEGIE INSTITUTION 



40 



35 



30 



25 



20 



15 



10 



TT 




XX 



b o 



Ph + Fo + G 



Fo + L+G 



b 



L4-G 



-) y>< /(c) ■'.!,!; 



I™ fO 

/ \ 



\ 



/ 



\ 



^<. /^Fo + Ok + Lc+G 

i J- — --> — — i — i r~— i— -4-— j— . 



\ 



\ 



\ 



■I- . I 



900 1000 1100 1200 1300 1400 

Temperature, °C 



1500 



1600 1700 



Fig. 58. Pressure-temperature diagram for compositions on the join K20-6MgO-Al 2 03-6Si02- 
H 2 0. Lettered points appear in the 10-kb section presented in Fig. 60. The four uni- 
variant curves presented as heavy lines in Fig. 57 appear around the invariant point at 
about 1 kb and 1175°C. Curve on which point (a) lies is the upper stability limit of 
phlogopite in the presence of an excess of gas. Dashed curve on which point (b) lies represents 
the beginning of melting of the assemblage Ph + Fo -f- Lc + Ks. Dashed curve on which point 
(c) lies represents the locus of maximum melting points of phlogopite in the water-deficient 
region. Dashed curve on which the point (d) lies is the water-saturated liquidus. 



GEOPHYSICAL LABORATORY 



165 



P=IOKb 
T=I225 °C 




Lc^Ks 



Fig. 59. Pseudoternary section at P T — 10 kb and 1225°C for the system Fo-H 2 0-Lc:Ks (1:1 
mole). Compositions investigated, marked with an X, lie on the join K 2 O6Mg0 -AloCVeSiOa-ELO, 
and are the basis of this schematic construction. 



variable and very much lower index of 
refraction ( = G in part) ; a milky fluid, 
often exuded when the container was 
punctured ( = G in part) ; and needles of 
mica, considered to have formed during 
the quenching process, not being stable 
during the run. The Ph + Fo+L regions 
exhibited faceted crystals of phlogopite, 
often in books 10 ju, thick, as well, but 
did not have any of the above-named 
products attributed to the gas phase, 
nor was the glass highly vesiculated. 
The anhydrous composition produced 
crystals of forsterite + leucite + kalsilite. 

(The anhydrous composition crystallized 
at 1 atm consisted of f orsterite + leucite 

+ orthorhombic kalsilite.) 



The third step taken was a study of 
the K 2 0-6MgO-Al 2 3 -6Si0 2 -H 2 join 
at a series of temperatures at 10 kb. The 
results are presented in Fig. 60. Atten- 
tion is called to the fact that the system 
is indeed quinary, and only those fields 
cut by part of the join are illustrated. 
The critical observation is that phlogo- 
pite is stable in the water-deficient re- 
gion at temperatures (c) well above the 
beginning of melting (a) of the silicate- 
rich portion of the join. Furthermore, 
the initiation of melting in the gas-free 
assemblage, Ph + Fo + Lc + ks, is at a 
somewhat higher temperature (6) than 
that for the gas-present region (a) . It is 
the former region which is of great im- 



CARNEGIE INSTITUTION 




5 

Phlogopite 
K 2 0'6MgOAI 2 03-6Si02 



10 



15 20 25 

Weight per cent 



30 



35 



40 
►H,0 



Fig. 60. The temperature-composition section at 10 kb for the silicate-rich portion of the 
K 2 0-6MgO-Al 2 03 - 6Si02-H20 join. The lettered points and lines are also indicated in Fig. 58, 
and the data at 1225°C are exhibited in Fig. 59. 



port in the partial melting process of the 
mantle and lower continental crust. 

Water is no doubt stored in the upper 
mantle, not as free gas but in hydrous 
phases, a large number of which have 
been shown to be stable to exceedingly- 



high pressures. In the absence of a gas 
phase, the melting of a mixture of hy- 
drous and anhydrous phases will prob- 
ably take place under conditions illus- 
trated by the undersaturated region of 
Fig. 60. The water contents of the initial 



GEOPHYSICAL LABORATORY 



167 



melt ■will be of the order of a few percent 
(«*4 wt %, Fig. 59), not the large 
amounts of water (*»22 wt %, Fig 59) 
in magmas formed in the presence of an 
excess of water. With increase of partial 
melting, two possible types of behavior 
may occur. If the mantle rock has mostly 
hydrous phases, then the water content 
of the liquid will increase. If the mantle 
rock consists mostly of anhydrous 
phases, the water content of the liquid 
will decrease with further rise in tem- 
perature. In each behavior the magma 
will remain undersaturated, even at tem- 
peratures at which all the hydrous phases 
are consumed. In the event that the 
hydrous mineral breaks down in the solid 
state, that is, before melting begins, a 
gas will be released, and the melting be- 
havior is as outlined by Yoder (Year 
Book 64, pp. 82-89) — a water-under- 
saturated magma is produced on partial 
melting. 

The results obtained are believed to 
apply, in general, to the melting of hy- 
drous phases. The results of Ernst 
(1961), for example, on the melting of 
the amphibole glaucophane support this 
view, even though he chose a compo- 
sition of liquid too low in water in 
determining the sequence of curves 
about the relevant invariant point. The 
water content of liquid (5 ±2 wt %) re- 
ported by him is not in accord with the 
sequence of curves presented (Ernst, 
1961, p. 755, Fig. 6). Glaucophane (Gl) 
would be stable in the gas-absent region 
at temperatures above the melting curve 
illustrated as Gl = Fo + EnH-L + G, which 
was actually observed by him to be G1 + 
G = Fo + En + L. 

Extrapolation of the curve on which 
the point (c) lies to the gas-absent liq- 
uidus in Fig. 60 indicates that phlogo- 
pite, if stable, may melt congruently at 
pressures above the limit of these experi- 
ments. Possible solid solutions of phlogo- 
pite were not investigated. Substitution 
of 2A1 for 3Mg would be possible within 
the system, as well as H + — >K + , leading 
to an alkali-deficient hydrophlogopite. 



In addition to outlining the behavior 
of a hydrous mineral in the gas-absent 
region, the principal conclusions drawn 
from this simple study are the follow- 
ing: (1) Partial melting at high pressures 
of forsterite-rock assemblages containing 
hydrous phases may yield water-under- 
saturated magmas having a low water 
content in the absence of a gas phase. 
(2) Continued melting of a hydrous 
mineral-bearing assemblage at high pres- 
sures yields either an undersaturated 
magma or a water-saturated magma, de- 
pending on whether the hydrous mineral 
melts incongruently or congruently. (3) 
Extremely water-rich magmas may be 
produced at high pressures from the 
melting of hydrous mineral-bearing as- 
semblages when free water is present or 
becomes present on the breakdown of 
the hydrous mineral in the solid state. 

It is thus demonstrated by experiment 
that liquids low in water content can 
be achieved in the laboratory even in 
the presence of a hydrous phase. The 
apparent disagreement between field de- 
duction and previous laboratory studies 
in the presence of excess water is re- 
solved. 

Reconnaissance Study of the 

Stability of Amphiboles at 

High Pressure 

M. C. Gilbert 

It is generally accepted that the hy- 
drosphere has been generated throughout 
the history of the earth by degassing of 
the mantle, principally through vulcan- 
ism (Rubey, 1951). All indications are 
that this process has not been completed, 
and we are faced with the question of 
how H 2 is and has been held in the 
mantle. Answers are crucially important 
to understanding the origin of magmas 
because of the strong effect of H 2 on 
the melting of silicates. 

H 2 would occur in the mantle (1) as 
an H 2 0-rich fluid phase, (2) adsorbed 
on the grain boundaries, or (3) bound 
in hydrous mineral phases. Amphiboles 



168 



CARNEGIE INSTITUTION 



and micas are the most promising of the 
hydrous phases because of their rela- 
tively high upper thermal stability limits 
at low pressure. (See Ernst, 1968, for a 
summary of the stability relations of 
amphibole, and Wones, 19676, for a 
summary of the stability relations of 
mica) . The role of the hydroxylated py- 
roxene reported by Sclar, Carrison, and 
Stewart (1967) is not clear at present. 

The question arises as to what hap- 
pens to these hydrous phases at high 
pressure. Mica can be stable to tem- 
peratures of 1200° to 1300°C near 40 kb 
(Kushiro, Syono, and Akimoto, 1967; 
Yoder and Kushiro, this report) , but ex- 
periments bearing on the upper thermal 
stability limit of amphiboles above 10 kb 
have not previously been reported. Yoder 
and Tilley (1962) and T. H. Green and 
Ringwood (1967) have indicated that 
amphiboles may be stable up to about 
1000 °C at 10 kb in systems of basaltic 
composition plus H 2 0. Ernst (1963) 
studied the polymorphism of alkali 
amphiboles at pressures up to 40 kb 
but did not investigate the decomposition 
boundaries at this pressure. Greenwood 
(1963) calculated that anthophyllite 
would not be stable at pressures much 
higher than 20 kb at any temperature. 
D. H. Green and Ringwood (1967) sug- 
gested that amphibole breakdown curves 
would have negative slopes at pressures 
higher than 10 to 20 kb because the large 
decrease in the specific volume of water 
at high pressure would change the AV 
of the reaction amphibole «=± breakdown 
products from positive to negative. Ku- 
shiro (1968a) has considered in detail 
the breakdown of hydrous minerals at 



high pressures and calculates that prac- 
tically all known amphibole species have 
negative slopes at higher pressure. 

Reconnaissance experiments on a num- 
ber of amphibole compositions in the 
presence of excess H 2 confirm these 
suggestions and calculations. Starting 
materials for the Mg-bearing composi- 
tions were glasses and oxide prepared by 
F. R. Boyd. Amphibole-bearing assem- 
blages synthesized from these materials 
were also used as charges for some of the 
runs. Sealed platinum capsules were run 
in a %-inch bore piston-cylinder ap- 
paratus with a talc pressure medium. 
The uncorrected nominal pressure is re- 
ported. 

The amphibole compositions and no- 
menclature employed are given in Table 
9. Some of the experimental results so 
far obtained are given in Table 10. 
Tremolite 50 tschermakite 50 has been syn- 
thesized at 800°C and 10 kb, confirming 
the earlier report of Boyd (Year Book 
53, p. 110). He attempted the synthesis 
of tschermakite but was not confident 
that the amphibole obtained was on 
composition. An amphibole has been 
synthesized at 800°C and 10 kb from 
a starting material of tschermakite bulk 
composition with about an 85 to 90% 
yield. It is tentatively identified as 
tschermakite. Boyd had originally sug- 
gested that this phase might be stable 
at high pressure. It appears to be stable 
up to at least 20 kb, but significantly, 
tschermakite, tremolite, and tremolite 50 
tschermakite 50 all seem not to be stable 
at 800 °C and 30 kb. The decomposition 
assemblage is dominated by pyroxenes 
at the tremolite end (alumina-free, silica- 



TABLE 9. Amphibole Nomenclature and Compositions Employed 



Abbreviation Name 


Composition 


Tr Tremolite 


°Ca 2 Mg 5 Si,0 22 (OH) 2 


TT Tremolitesotschermakiteso 


°Ca 2 Mg 4 AlSi7A10 22 (OH) 2 


T Tschermakite 


°Ca 2 Mg J Al 2 SioAl 2 2 2(OH) 2 


Ed Edenite 


NaCa 2 Mg 5 Si 7 A10 22 (OH) 2 


Pa Pargasite 


NaCa 2 Mg4 AlSie A1 2 22 ( OH ) 2 


Ha Hastingsite 


NaCa 2 Fe/ + Fe 3+ SieAl 2 22 ( OH ) 2 



GEOPHYSICAL LABORATORY 



169 



TABLE 10. Some Results of High -Pressure Hydro thermal Runs for Amphibole End 

Members Studied 



Condensed Reactants * 


T,°C 


Duration, 
P, kb hours 


Condensed Run Products 




"CaaMgsSisO^COHJa + SiOat 




Oxide mix 
Tr + Qtz 


800 
800 


10 15% 
30 71 


Tr + Qtz 

Tc + Opx + Qtz 




°Ca 2 Mg*AlSi 7 A10 22 (OH) 2 




gl 
gl 
gl 
TT + Cpx + Opx? + G? 

gl 

TT + Cpx + Opx? + G? 


800 
900 
700 
800 
900 
800 


10 65 
10 17 
20 146 
20 75 
20 97 
30 71 


TT + Cpx + Opx? + G? 
TT + Cpx + Opx? + G? 
TT + Tc + Cpx + Opx + G? 
TT+(Cpx + Opx? + G?) 
Cpx + Opx + G 
Cpx + Opx + G 




°Ca 2 Mg 3 Al 2 Si 8 Al 2 22 (OH) 2 




gl 

T + Opx + G + tr Cpx 

gl 

T + Opx + G + tr Cpx 

gl 

T + G 
T + G 


800 
800 
900 
800 
900 
800 
900 


10 65 
10 143 
10 17 

20 75 
20 16 
30 71 

38.7 2iy 2 


T + G 

No change 

T + Opx + G + tr Cpx 

T+(G) 

G + Cpx 

G+ ? 

G+? 




NaCa 2 Mg 5 Si7A10 22 ( OH) 2 




Oxide mix 

Ed + Cpx + Fo + tr gl 

Ed + Cpx + Fo + tr gl 


900 
950 
900 


20 ~25 

282 20y 2 
38.7 2iy 2 


Ed + Cpx + Fo + tr gl 
(Ed)+Cpx + Fo + gl 
Cpx + Fo + gl 




NaCa 2 Mg4AlSieAl 2 22 (OH) a 




Oxide mix 
Pa + Cpx 
Oxide mix 
Pa + Cpx 
Pa + Cpx 
Pa + tr Cpx 


800 
800 
900 
800 
950 
900 


10 143 
20 75 
20 ~25 
30 71 
28.2 20 
38.7 2iy 2 


Pa + Cpx 

Pa + (Cpx) 

Pa + Cpx 

Sheet silicate + G + Cpx + gl 

G + Cpx + Fo + gl 

G + Cpx + Fo + gl 




NaCa 2 Fe4 2+ Fe 8+ Si e Al 2 22 (OH) 3 




Ha + Mt + Cpx + tr G + Pc 
Ha + Mt + Cpx + tr G + Pc 
Ha + G + Mt + Cpx 
Ha + G + Mt + Cpx 


700 
750 
650 
700 


11.8 168 
12 97 

19.9 237 
20.9 93 


G + Mt + Cpx [FMQ] t 
G + Mt + Cpx [FFsM] 
G + Mt + Cpx [FFsM] 
G + Mt + Cpx [FFsM] 



* Amphibole abbreviations are given in Table 9. Qtz, quartz; Tc, talc; Opx, orthopyroxene ; 
Cpx, clinopyroxene ; G, garnet; Fo, forsterite; Mt, magnetite; gl, glass; ?, presence not con- 
firmed, or unidentified phase or phases ; tr, trace ; parentheses indicate metastable phases ; bold- 
face type indicates dominant phase. 

tBulk compositions of the runs expressed in terms of amphibole end-member formulas. Only 
tremolite composition was silica excess. 

t Oxygen fugacity buffers: FMQ denotes fayalite-magnetite-quartz assemblage; FFsM denotes 
fayalite-orthoferrosilite-magnetite assemblage. 



rich end) of the tremolite-tschermakite 
join and by garnet at the tschermakite 
end (alumina-rich, silica-poor end). A 
run with anhydrous tschermakite compo- 
sition at 1200 °C and 30 kb also gave a 
breakdown assemblage consisting almost 
entirely of garnet. These nonsodium- 
bearing calcic amphiboles appear to have 



comparable upper thermal stability 
limits, but their stability limits are lower 
than those of the sodium-bearing calcic 
amphiboles, the hornblendes. 

The amphibole with the highest ther- 
mal stability limit is pargasite, reaching 
1040°C at 1 kb (Boyd, 1959). Illustra- 
tion of the possible stability limits of 



170 



CARNEGIE INSTITUTION 



intermediate members of the pargasite- 
hastingsite group was shown in the 
studies of Yoder and Tilley (1962) and 
T. H. Green and Ringwood (1967), men- 
tioned earlier. From the work reported 
here, edenite and paragasite both appear 
to be unstable at 950 °C and higher, 
near 30 kb, or at 900°C and higher, 
near 40 kb. Sodium-bearing, purely 
magnesian, calcic amphiboles appear to 
continue to melt incongruently, whereas 
those that are nonsodium-bearing break 
down in the solid state. In either case, the 
upper thermal stability limits are lower 
at higher pressures and must have nega- 
tive slopes over all or part of the pressure 
range studied. 

To determine the effect of pressure on 
the stability of one of the purely iron 
end members of the hornblendes, hasting- 
site was studied at 12 and 20 kb under 
conditions of controlled oxygen fugacity 
(/o 2 ). The assemblages fayalite-magne- 
tite-quartz (FMQ) and fayalite-ortho- 
ferrosilite-magnetite (FFsM) were used 
to fix /o 2 in the standard double-capsule 
method. The FFsM assemblage is stable 
to higher oxygen fugacities than FMQ 
(Lindsley, Speidel, and Nafziger, 1968). 

Although the iron-amphibole stability 
field at high pressure has not been de- 
lineated, conditions of instability are 
known and place maximum temperature 
limits on stability. Thus hastingsite is 
not stable at 650 °C and higher at 20 kb, 
or at 700 °C and higher at 12 kb when 
/o a is defined by FFsM; or at 750 °C at 
12 kb when f 0t is defined by FMQ. 

Hastingsite dehydrates at low pressure 
(3 kb) to a plagioclase-bearing assem- 
blage. At 12 and 20 kb, plagioclase dis- 
appears as a breakdown product, giving 
one more indication that the plagio- 
clase +± garnet + pyroxene reaction, in 
chemically complex systems, occurs at 
pressures suggestively close to those to 
be found at the crust-mantle transition. 

If 950 °C is a maximum temperature 
of stability at 30 kb, then depending on 
the geothermal gradient, no amphibole 
would be stable below about 70 to 



100 km depth. Amphiboles may thus 
play a significant role in the petrology 
of the lower crust and uppermost mantle 
but cannot be the water containers at 
greater depths. It may be significant that 
no amphibole has ever been reported 
as a primary inclusion in diamond 
(Meyer, Year Book 66, pp. 446-450). 
It seems possible that some amphibole 
stability fields may be pinched out 
entirely at high pressure between expand- 
ing stability fields for sheet silicate-bear- 
ing assemblages from the low-tempera- 
ture side and garnet- and pyroxene-bear- 
ing assemblages from the high-tempera- 
ture side. The fact that phlogopite is 
stable to about 1300°C near 40 kb and 
pargasite is unstable at 900 °C at 40 kb 
suggests that a depth-facies classification 
could be devised for the upper mantle 
based on amphibole/mica occurrence. 

Control of Fugacities in Fluorine- 
Bearing Hydrothermal Systems 

J. L. Munoz 

Fluorine is a common constituent of 
many rock-forming minerals, especially 
micas and amphiboles. Fluorine substi- 
tutes for the hydroxyl group in these 
minerals, and this substitution consider- 
ably affects their thermal stability. Thus, 
both the compositions and the stability 
relations of these minerals are partly 
dependent upon the magnitudes of the 
H 2 and HF fugacities of their environ- 
ment. In order to calibrate these effects 
in synthetic systems, it is necessary to 
develop methods that permit control of 
the fugacities at elevated pressures and 
temperatures in a quantitative manner. 
The most simple and direct method is to 
vary the bulk composition of the gas 
phase. This approach is unsatisfactory 
for fluorine-bearing systems because of 
the high solubility of silicates in aqueous 
HF at low temperatures, analytical dif- 
ficulties, and inadequate knowledge of 
the P-V-T and fugacity relations in the 
system O-H-F under the conditions of 
experimental interest. Alternatively, the 



GEOPHYSICAL LABORATORY 



171 



composition of the gas phase and its 
fugacities may be buffered by appro- 
priate assemblages of crystalline phases. 
With this technique, the composition and 
fugacities of the equilibrium gases as 
functions of total P and T are calcu- 
lated from the equilibrium constants 
and fugacity coefficients for all the im- 
portant molecular species present (as- 
suming ideal mixing) and from the free- 
energy data for the solid buffer phases. 

Figure 61 shows the experimental ar- 
rangement. The charge is contained in 
gold foil and is surrounded by the fluo- 
rine buffer assemblage and a gas of com- 
position O-H-F or C-O-H-F, depending 
on whether graphite is present as a buffer 
phase. This charge-buffer system is 
sealed in a Pt membrane. Because the 
fluorine buffer fixes only a fugacity ratio 
(e.g., /h 2 o//hp or /o 2 //f 2 ), it is necessary 
to fix independently one additional 
fugacity within the system. This can be 
done by externally fixing the hydrogen 
fugacity, either by using a Shaw bomb, 
a standard oxygen buffer, or the methane 
buffer (see Eugster and Skippen, 1967). 

Calculations have been performed for 
the following fluorine buffers over a 
wide range of P to tai, T, and / Ha : 

CaSi0 3 + 2HF <=± CaF 2 + Si0 2 + H 2 



SEALED 
Au CAPSULE 



Pt 



SEALED 
CAPSULE 

Au FOIL- 



OXYGEN 

BUFFER (X) 

CHARGE 



_ FLUORINE 
BUFFER (Y) 

X,OH(Y,OHF) 
X,OH(Y,COHF) 
X,COH(Y,OHF) 
X,COH(Y,COHF) 



Fig. 61. Diagram of triple capsule arrange- 
ment used for fluorine buffer experiments. 
Notation: The charge-buffer system is en- 
closed by parentheses, together with the ele- 
mental composition of the gas phase. The 
external oxygen buffer and its gas components 
are written to the left of the parentheses. 



wollastonite + 2HF ^± nuorite-r- 

quartz+H 2 (WFQ) (1) 

CaAl 2 Si 2 8 +2HF ±± CaF 2 + 

Al 2 Si0 5 +Si0 2 +H 2 

anorthite + 2HF ?=> fluorite-1- 
sillimanite-f- quartz + H 2 (AFSQ) (2) 



CaC0 3 + 2HF 



CaF 2 + C + 

H 2 + 2 



calcite+2HF <=> fluorite + 

graphite +H 2 + 2 (CFG) (3) 

Buffer notation is given as B , OH or 
COH (F B , OHF or COHF) where B 
represents the external buffer assem- 
blage, F B represents the fluorine buffer, 
and the letters OHF or COHF refer to 
the components present in the gas phase. 
Details regarding the equations used in 
the calculations, and results for a num- 
ber of different sets of conditions will 
appear in a forthcoming paper. As an 
example, Fig. 62 shows the dependence 
of the fugacity of HF on temperature, 
fluorine buffer, and /h 3 for a total pres- 
sure of 2 kb. 

An alternative approach to controlling 
the fugacities of H 2 and HF is to con- 
trol the activities of H + and F~. Assum- 
ing that only monovalent ions are pres- 
ent, these two approaches can be related 
through the equations for the ionic dis- 
sociation of HF and H 2 0: 

hp = — : and K H ,o =~~ 



&HF 



a H„0 



Combining these two equations with 
the restriction that the sum of the moles 
of positive ions must equal the sum of 
the moles of negative ions, we obtain 



„ + / +\ Khf g H f ■ Kh. 
«H + =(yH + ) - 



O flH 2 Q 
OH~ J 



■/f yoH 

where y H + , yF, and y H" represent the 
activity coefficients for the respective 
ions. This equation shows that if the 
activities of the molecular species are 
independently defined, the pH, and hence 
the activities of the other ions in the gas 
phase, are fixed. Conversely, fixing the 



172 



CARNEGIE INSTITUTION 



O 



O 




Fig. 62. Variation of HF fugacity with temperature calculated for a number of fluorine buffers 
at 2 kb. Buffer assemblages (neglecting gas) are HM, hematite + magnetite ; NNO, nickel + 
nickel oxide ; QFI, quartz + fayalite + iron ; WMC, wustite + magnetite + graphite ; AFSQ, 
anorthite + fluorite + sillimanite -)- quartz ; WFQ, wollastonite -f- fluorite + quartz ; CFG, cal- 
cite -f- fluorite + graphite. 



activities of the ions defines the fugacities 
of the molecular species. The molecular 
approach has been chosen only because 
this method creates fewer technical and 



experimental difficulties ; results obtained 
from both methods should, however, be 
identical. 
To demonstrate the feasibility of fluo- 



GEOPHYSICAL LABORATORY 



173 



rine buffers, a number of experiments 
were performed equilibrating phlogopite 
[KMg 3 Si 3 A10 10 (F,OH) 2 ] with gases that 
were equilibrated with different fluorine 
buffer assemblages. Phlogopite was 
chosen because it exhibits complete solid 
solution between the F and OH end 
members; also, it is possible to prepare 
both end members in a pure state. By 
using both pure fluorophlogopite and pure 
hydroxyphlogopite as starting materials, 
the exchange reaction was approached 
from both sides, and equilibrium with 
the buffer was thus demonstrated (Table 
11). Compositions of the phlogopites 
were determined by X rays to ± 2 mole % 
fluorophlogopite on the basis of the 
d(005) value measured against CaF 2 
(220) with copper Ka x radiation. 

The phlogopite exchange reaction 
studied can be represented by the equa- 
tion 

KMg3Si 3 A10 10 (OH) 2 +2HF ^ 

KMg 3 Si 3 A10 10 F 2 +2H 2 

for which we may write an equilibrium 
constant in terms of the fugacities of 
H 2 and HF, K = /h 2 o7/hf 2 . Fluorine 
buffer curves, when plotted in terms of 
the log of this equilibrium constant 
versus reciprocal temperature, yield 
straight lines that show the calculated 
/h 2 o//hf ratios for each buffer (Fig. 63) . 
For a fixed temperature, these four 
curves are arranged in order of increas- 
ing fluorination (from top to bottom). 



The compositions of the phlogopites 
as a function of temperature are shown 
in Fig. 64 (A) . Only one temperature was 
run for the buffer HM, OH (CFG, 
COHF) ; note that the composition of 
this mica is nearly pure fluorophlogopite, 
in contrast with the much more hy- 
droxylated mica obtained by using NNO, 
OH (CFG, COHF) . The large differences 
in mica composition also serve to demon- 
strate communication between the ex- 
ternal oxygen buffer and the fluorine 
buffer, since the fluorine buffer was 
identical in both cases. Moreover, these 
differences in composition are consistent 
with relative fugacity ratios shown in 
Fig. 63 for these two buffers. Note also 
that the composition of the phlogopite 
equilibrated with the NNO, OH (ASFQ, 
OHF) buffer at 700°C is very close to 
that of the extremely fluorinated phlogo- 
pite equilibrated with the HM, OH 
(CFG, COHF) buffer at the same tem- 
perature (Table 11, Fig. 64A). This is 
precisely what the calculations predict, 
inasmuch as the values for the equilib- 
rium constants for these buffers nearly 
overlap (Fig. 63) . 

A calibration problem was discovered 
between the CFG and WFQ buffers, how- 
ever, as shown in Fig. 64(B). At all 
temperatures studied, phlogopites equili- 
brated with the WFQ buffer are more 
fluorine rich than those equilibrated with 
the CFG buffer, whereas calculated 
/h 2 o 2 //hf 2 values predict the opposite 



TABLE 11. Final Compositions of Phlogopite Solid Solutions Equilibrated 
with Various Fluorine Buffers at 700°C, 2 Kb* 



Buffer 



Starting 
Composition 



Final Composition, 

±2% 



NNO, OH (CFG, COHF) 


OHPh 


FPh 5 sOHPh 42 


NNO, OH (CFG, OHF) 


FPh 


FPhsoOHPlu* 


HM, OH (CFG, COHF) 


OHPh 


FPh 96 OHPh 4 


HM, OH (CFG, COHF) 


FPh 


FPh !1E OHPh s 


NNO, OH (WFQ, OHF) 


OHPh + FPh 


FPh 71 OHPh29 


NNO, OH (AFSQ, OHF) 


OHPh 


FPh,cOHPh 5 



* Run times ranged between 20 and 40 days. 



174 



CARNEGIE INSTITUTION 



T,°C 

400 500 600 700 800 



12- 



10- 




4- 



1 1 1 


1 1 


- 


; lv N- 




- 


- s -K % 




- 






- 






- 








\N 


s 






ss 




1 1 1 1 1 1 


1 


^ 



1.6 



.4 1.2 1.0 

(°KX I0 3 ) 



0.8 



T 



Fig. 63. Log K (= /h 2 o 2 //hp 2 ) versus reciprocal temperature for four fluorine buffers. Height 
of solid bars represents the amount of uncertainty in the K that can be attributed to un- 
certainties in the AG values for the solid phases involved. For symbols, see Fig. 62. 



result. This difficulty is most likely re- 
lated to either (1) inadequately known 
AG° data for the buffer phases or (2) 
nonideal mixing in the gas phase. The 
rectangles in Fig. 63 represent the esti- 
mated maximum error in /h 2 o 2 //hp 2 due 
to uncertainties in the available free- 
energy data (Robie, 1962) ; the differ- 
ences are not enough to account for the 
discrepancy. Fugacity coefficients for a 
number of the molecular species (includ- 
ing HF) are approximations and could 
well introduce a considerable source of 
error. Compounding this problem, how- 
ever, the assumption of ideal mixing in 
the gas phase may not apply in the 



presence of such a chemically reactive 
component as fluorine. A solution to the 
problem may be apparent when a number 
of different fluorine buffers are tested 
against one another; ultimately a sys- 
tematic study of the system O-H-F may 
be necessary. In the meantime, these 
phlogopite experiments have demon- 
strated that the fluorine buffers function 
well experimentally even though some 
calibration problems are present. More- 
over, knowledge of the compositions of 
these phlogopite F, OH solid solutions 
now provides a sensitive sliding scale, 
which will be useful in determining / H f/ 
/h o ratios in future experiments. 



GEOPHYSICAL LABORATORY 



175 



O 



h- 



750 
700 
650 
600 
550 



HM,0H 
(CFG.COHF) 

NNO.OH 
(AFSQ.OHF) 



NNO 
(WFQ ; 




OH (CFG.COHF) 



cm x 



O 
O 



FPh 80 60 40 20 OHPh 
Mole per cent "FPh" 

A 



T 1 

2 Kb 
700 °C - 



WFQ^ / 
/ 



I 

,'|CFG 



AFSQ/ 
/ 



i 



CFG 



J 1 



>h 80 60 40 20 OHPh 
Mole percent "FPh" 

B 



Fig. 64. (A) Compositions of phlogopites in equilibrium with four different fluorine buffers 
as a function of temperature at 2 kb. Compositions are precise to ±2 mole % FPh (fluoro- 
phlogopite). Arrows show direction from which equilibrium was approached. (B) Compositions 
of phlogopites equilibrated with four fluorine buffers at 700°C (Fig. 64A) plotted as a function 
of the log of the equilibrium constant. The relative compositions of the micas equilibrated with 
NNO, OH (CFG, COHF) and NNO, OH (WFQ, OHF) buffers are in the opposite direction 
from that predicted by the equilibrium constant. 



SULFIDE MINERALOGY AND 
PHASE-EQUILIBRIA STUDIES 



ORE MINERALS 

The Fe-Se System 

G. Kullerud 

The Fe-Se system has been studied by 
quenching and differential thermal 
analysis (DTA) experiments. The phase 
relations at this point have been clarified 
above 400°C, as shown in Fig. 65. The 
solubility of Se in Fe is slight (less than 
0.5 atomic % even at 900°C). A eutectic 
exists at 940 °C and 5.5 atomic % Se, 
where the four phases Fe + FeSe+L + V 
coexist. DTA experiments on fifteen dif- 
ferent compositions in the Fe-FeSe range 
all show a distinct thermal effect at this 
temperature. 

Monotectic conditions exist at 962 °C 
and 46.0 atomic % Se. At this point the 
four phases FeSe+L (containing 6.5 
atomic °/o Se) +L (containing 46 atomic 



°/o Se) +V coexist. Thus a wide field of 
liquid immiscibility exists above 962 °C 
in the Fe-FeSe portion of this system. 
The high-temperature boundary curves 
of this field were not detected in DTA 
experiments. 

The melting relations of the mono- 
selenide phase are very similar to those 
of a number of monosulfides, such as 
Fe 1 _ a; S, Ni^S, and Co^S. The maximum 
temperature on the monoselenide liq- 
uidus curve is 1065° ±5°C. The composi- 
tion of the only monoselenide that melts 
directly to a liquid is about Fe 47 Se 33 
(corresponding to x = 0.113 in the Fe^Se 
formula). Beyond this composition, the 
liquidus curve slopes steeply toward 
lower temperatures, as shown in Fig. 65, 
and reaches 790°C at 71.5 atomic % Se. 
A distinct thermal effect was recorded 
at this temperature in all DTA experi- 
ments containing 57.5 atomic °/o Se or 



176 



CARNEGIE INSTITUTION 



1200 




/ 2 Liquids 



/7I.5 790 



Fei-jfSe + Liquid 
727 



Fe3Se4 + Liquid 



580 



FeSe2 + Liquid 



J L 



80 



90 Se 



Fig. 65. Phase relations in the Fe-Se system above 400 C C. Three compounds occur: Fei_,Se, 
Fe 3 Se4, and FeSe 2 . 



more. It was detected even in an experi- 
ment containing 98 atomic % Se. These 
results show that a liquid immiscibility 
field exists above 790°C. It spans the 
compositional region from a monotectic 
point (at 71.5 atomic % Se) to more 
than 98 atomic % Se. The presence of 
this monotectic, where the four phases 
Fe^Se+L (with 71.5 atomic % Se) +L 
(of essentially Se composition) +V co- 
exist, was demonstrated by an experi- 
ment containing 71.5 atomic % Se. When 
viewed by eye at 800 °C a homogeneous 
liquid is observed; on cooling below 
790°C growth of Fe^Se and generation 
of brownish liquid Se are clearly noted 
to take place simultaneously. In the 
DTA experiments containing from 56 
to 68 atomic % Se, distinct thermal ef- 
fects were recorded at 835 °C. The solidus 



curve on the iron-deficient side of the 
Fe x _ a Se solid solution intersects the 
835 °C horizontal line at about 57.5 
atomic % Se. Experiments with 49 
atomic °/o Se give a thermal effect at 
872 °C, and those with 56 atomic % Se 
give thermal effects at 835° and 845 °C. 
Combined, these data indicate that an 
inversion takes place in the Fe^Se 
phase. The inversion temperature of stoi- 
chiometric FeSe is 872 °C. Fe^Se with 
maximum Fe deficiency (7.5 atomic % 
Fe) inverts at 835°. We shall refer to 
this high-temperature phase as cxFe^Se 
to distinguish it from previously reported 
polymorphs stable at lower temperatures. 
Below this inversion, Fe^Se has hexa- 
gonal NiAs-type structure down to 335 °C 
for stoichiometric FeSe and to about 
380°C for Fe^Se (with x = 0.13), as 



GEOPHYSICAL LABORATORY 



177 



indicated by DTA and quenching ex- 
periments. Below this second inversion, 
Fe^Se is reported to have tetragonal 
structure. 

In DTA experiments containing 58 to 
98 atomic % Se, distinct thermal effects 
were recorded at 727° C. Below this tem- 
perature the Fe 3 Se 4 compound is stable. 
The temperature of 727°C is taken as 
its incongruent melting point, where the 
four phases Fe 3 Se 4 + Fe 1 _J3+.L + F co- 
exist. When it first crystallizes the Fe 3 Se 4 
compound contains about 58 atomic °/o 
Se, which is 1% more than indicated by 
the stoichiometric formula. The Fe 3 Se 4 
compound forms solid solution on both 
sides of the stoichiometric composition. 
This solid solution apparently reaches a 
maximum at 580°C, where the Fe 3 Se 4 
phase spans the region from about 56.0 
to 60.0 atomic % Se. 

At high temperatures, Fe 3 Se 4 has a 
hexagonal crystal structure, which can- 
not be quenched. During cooling to some 
temperature below 600 °C, an inversion 
takes place that results in prominent 
twinning. The low-temperature form is 
monoclinic. The structure and cell di- 
mensions were determined by Morimoto 
and Kullerud {Year Book 58, pp. 199- 
201, 1959). The solvi defining" coexisting 
compositions of the Fe^Se and Fe 3 Se 4 
phases have not been determined ac- 
curately. Phase relations at elevated 
temperatures in the 50-60 atomic % 
composition range cannot be studied suc- 
cessfully by quench-type experimenta- 
tion because even the fastest chilling will 
not freeze the equilibria achieved during 
the heating period. Constant heating and 
cooling rates are not readily achieved 
in DTA experiments at temperatures be- 
low 400°C, and the magnitude of most 
low-temperature heat effects in this 
critical region of the system appears to 
be near the limit of detection by our 
method. 

The FeSe 2 phase melts incongruently 
at 580°C, where the four phases Fe 3 Se 4 + 
FeSe 2 +L + 7 coexist. Distinct thermal 
effects were recorded at this temperature 



in all DTA experiments containing 61 to 
98 atomic % Se. This compound, which 
is analogous with the mineral ferrosilite, 
is orthorhombic (Kullerud and Donnay, 
1958). 

The Cu-Zn-S System 
J. R. Craig and G. Kullerud 

Although copper and zinc sulfides are 
common, mineral assemblages containing 
coexisting copper sulfides and zinc sulfide 
are not common. Under natural condi- 
tions the almost ubiquitous presence of 
elements such as iron in addition to 
copper and zinc complicates the mineral- 
ogy. Knowledge of phase relations in 
the pure Cu-Zn-S system, however, 
serves to outline the stability limits of 
coexistence of the pure Cu and Zn sul- 
fides and is a necessary prerequisite to 
investigation of complex systems like Cu- 
Fe-Zn-S, Cu-Pb-Zn-S, and Cu-Fe-Pb- 
Zn-S. 

The phase relations in the Cu-Zn-S 
system were studied with the use of 
silica-tube reaction vessels for quenching 
and DTA experiments, with principal 
efforts on the 800° and 500°C isotherms. 
As shown in Fig. 66, tie lines exist be- 
tween ZnS (sphalerite) and the Cu 2 S- 
Cu 9 S 5 solid solution at 800°C. ZnS is 
also noted to coexist with a CuZn alloy 
(a), which is stable over a composition 
range extending from nearly pure Cu to 
about 35 wt % Zn. Two additional CuZn 
alloys, referred to as /2 and y, coexist with 
the ZnS phase. Pure Zn melts at about 
419 °C, and at 800 °C liquid is present 
from the Zn composition corner to about 
30 wt °/o Cu. As seen in Fig. 66, ZnS also 
coexists with this liquid phase. The phase 
relations on the Cu-Zn join were taken 
from Hansen and Anderko (1958). It is 
noted that in the phase diagram of Fig. 
66 sphalerite and metallic copper form 
a stable assemblage, which occasionally 
is represented in ores, whereas coexist- 
ence of chalcocite-digenite solid solution 
with metallic zinc is prohibited by three 
different sets of tie lines. At 800 °C the 



178 



CARNEGIE INSTITUTION 



800°C 



k ZnS 



Cu 9 S 5> 
Cu 2 S/ 






Cu 



Zn 



Fig. 66. Phase relations in the Cu-Zn-S system at 800°C. All phases and phase assemblages 
coexist with vapor. 



solid solution of ZnS in Cu 2 S ss is 7± 
1 wt %. The solid solution of Cu 2 S ss in 
ZnS is much less than 1 wt % at the 
same temperature but is sufficient to alter 
the color from the snow white of pure 
ZnS to a resinous yellowish brown. The 
ZnS in all experiments at 800 °C had the 
cubic sphalerite structure. The unit-cell 
dimension of ZnS, whether present as 
the only phase or in equilibrium with any 
of the other phases shown in Fig. 66, 
is always within the limits of experi- 
mental error of a = 5.409 ±0.003 A. 

At 500 °C the phase relations (shown 
in Fig. 67) are basically the same as 
those at 800°C. The solid solution of ZnS 
and Cu 2 S ss has decreased to 3 wt % ; 
the solid solution of Zn in the aCuZn al- 
loy has increased to almost 40 wt % ; the 
solubility of Cu in the Zn liquid has de- 
creased to about 5 wt °/o ; and a third 
CuZn alloy (e) has appeared. The covel- 
lite (CuS) phase becomes stable in the 
pure Cu-S system at 507°C (Kullerud, 



1965). Tie lines between ZnS and CuS 
are already established at 503°±2°C. 
The ZnS in all the experiments conducted 
at 500 °C was found to have the sphaler- 
ite structure. Measurements of the cell 
edge were performed on ZnS from all 
stable assemblages containing this phase. 
As in the 800 °C experiments, it was al- 
ways found to lie within the limits of 
a = 5.409 ±0.003 A. 

Differential thermal analyses on Cu 2 S 
ss saturated in ZnS indicate that the 
temperature of inversion is influenced 
by ZnS. Pure Cu 2 S inverts on heating 
from hexagonal to cubic symmetry at 
430 °C, whereas the inversion takes place 
at 419° ±3°C when maximum solid solu- 
tion of ZnS exists. Pure Cu S 5 (digenite) 
inverts at 73° ±3°C (Morimoto and Kul- 
lerud, 1963), but digenite with maximum 
ZnS in solid solution inverts at 79° ±2°C. 

Moh (1960) reported the presence of 
two or possibly three ternary phases in 
the Cu-Zn-S system at 400°C. We have 



GEOPHYSICAL LABORATORY 



179 



500°C 




Liquid 



Fig. 67. Phase relations in the Cu-Zn-S system at 500°C. All phases and phase assemblages 
coexist with vapor. 



not been able to synthesize any of the 
reported ternary compounds by heating, 
at 400° and 500°C, mixtures of the 
elements in some experiments, and mix- 
tures of presynthesized ZnS and copper 
sulfides in others, for periods up to 8 
months. Experiments recently performed 
by Moh at our request, using the same 
starting materials he employed earlier 
(Moh, 1960), produce small amounts of 
idaite, bornite, chalcopyrite, and copper 
oxides in addition to our reported binary 
phases. The compounds reported by Moh 
(1960) are not ternary phases in the 
Cu-Zn-S system. They are probably due 
to impurities, mostly Fe, in his starting 
materials. 

Cubic <=± Hexagonal Inversions in 
Some M 3 S 4 -Type Sulfides 

G. Kullerud 

A number of sulfide-type compounds 
have the spinel structure. These com- 
pounds, which contain metals of the 
transition elements and sulfur, selenium, 



or tellurium in the 3 : 4 atomic ratio, are 
of considerable interest, not only because 
many of them are minerals, but also be- 
cause they display a large variation in 
electrical and magnetic properties. Some 
of these compounds are semiconductors, 
and others are metallic conductors. In 
contrast, the corresponding oxyspinels 
are all semiconductors. 

Studies of the thermal stabilities of 
the M 3 S 4 -type compounds show that 
most of them decompose to a mixture of 
monosulfides and disulfides ; for instance, 
Ni 3 S 4 produces Ni 1 -J3 + NiS 2 when de- 
composing. Some, such as FeCr 2 S 4 dau- 
breelite), are reported to melt con- 
gruently. 

During an investigation of the Ni-S 
system, Kullerud and Yund (1962) de- 
termined the temperature (356°C) of the 
invariant point where Ni 3 S 4 breaks down 
to Nix-JS + NiS 2 in the presence of vapor. 
In another study (Year Book 60, pp. 
176-178) they determined the upper sta- 
bility curve of the Ni 3 S 4 compound up to 
2 kb confining pressure. This curve has a 



180 



CARNEGIE INSTITUTION 



positive slope of about 6°C/kb. The unit- 
cell dimensions of cubic spinel-type Ni 3 S 4 
as well as of the breakdown products, 
hexagonal Ni^S (with £ = 0.05) and 
cubic NiS 2 , were determined, and the AF 
for the balanced equation was calculated. 
The calculation showed that the volume 
of the breakdown products is actually 
smaller than that of the cubic Ni 3 S 4 . 
The curve describing the reaction (1 — 
2a;)Ni 3 S 4 (cubic) <f± 2Ni 1 _J3 (hexagonal) 
+ (1 — 4:r)NiS 2 (cubic) accordingly 
should have negative slope. 

A number of pressure experiments on 
cubic Ni 3 S 4 , which was synthesized in 
silica tubes below 350°C, showed that 
an inversion takes place in the Ni 3 S 4 
compound at low temperatures and low 
pressures. The schematic P-T diagram 
illustrating the behavior of the Ni 3 S 4 
compound is shown in Fig. 68. At the in- 
variant point c (at 356 °C) the four 
phases Ni 3 S 4 (cubic) +Ni 1 _ iC S (hexa- 
gonal) + NiS 2 (cubic) +7 coexist. The 
univariant curve Ni 3 S 4 (cubic) + Ni 1 _ a ,S 
(hexagonal) +NiS 2 (cubic) has a small 
negative slope between point c and its 
termination at about 400 bars and 
353°C. At pressures exceeding 400 bars, 
cubic Ni 3 S 4 is no longer stable; a high- 
pressure nonquenchable polymorph oc- 
curs instead. The inversion from the 
high-pressure to the low-pressure cubic 
symmetry is accompanied by twinning, 
which can be observed in appropriately 
etched polished sections. The approxi- 
mate position of the P-T curve of the 
high-low inversion in the Ni 3 S 4 com- 
pound is shown in Fig. 68. Appearance or 
lack of twinning in cubic Ni 3 S 4 was used 
to bracket the curve in successive ex- 
periments. 

This study shows that the field of 
stability of cubic Ni 3 S 4 , which is repre- 
sented in nature by the mineral poly- 
dymite, is very limited. In ores this 
mineral can have formed only at tem- 
peratures below 356°C, and in addition, 
if untwinned, it cannot have been ex- 
posed to pressures exceeding 400 to 500 
bars. Since polydymite is cubic and 



opaque, twinning is not readily observed, 
even in the best of polished sections, un- 
less they are carefully etched. Etching 
shows that twinning is very common 
in this mineral in many localities; in 
others, twinning is entirely lacking. Its 
typical twinning gave this mineral its 
name polydymite, meaning many twins. 
The new experimental data provide a 
tool that can be used to obtain important 
information, not only on the maximum 
temperature prevailing during and after 
formation of this mineral, but on the 
pressure conditions as well. 

The Co 3 S 4 compound, which also has 
the spinel structure, is stable below 
664°C in the condensed Co-S system. 
At this temperature it breaks down to 
hexagonal Co^S and cubic CoS 2 . 

Pressure experiments show that the 
stability curve of Co 3 S 4 has a negative 
slope, which is in agreement with the 
calculated AV of the balanced reaction 
(l-2z)Co 3 S 4 ^ 2Co!^S+ (l-4z)CoS 2 . 
At about 2.5 kb and 640 *C the Co 3 S 4 
compound goes through an inversion 
similar to that observed for Ni 3 S 4 . Albers 
and Rooymans (1965) have shown that 
the FeCr 2 S 4 compound occurs in two 
polymorphs: a low-temperature, low- 
pressure, cubic spinel-type structure 
analogous to the mineral daubreelite, 
and a high-temperature, high-pressure 
form with a NiAs-derivative type struc- 
ture. In DTA experiments the inversion 
in this compound is observed at about 
1060 °C in the presence of vapor. The 
morphology of single crystals grown in 
silica tubes above 1060°C is preserved 
on quenching and indicates a hexagonal 
symmetry for the high-temperature poly- 
morph. With increasing pressure, the 
temperature of inversion decreases mark- 
edly. A series of synthetic compounds 
have also been found to invert from 
cubic spinel to NiAs-derivative struc- 
tures when exposed to high pressures — 
examples are CuCr 2 Se 4 , CuCr 2 Te 4 , and 
CoCr 2 S 4 . It is assumed by analogy that 
the high-pressure forms of spinel-type 
minerals, such as polydymite and lin- 



GEOPHYSICAL LABORATORY 



181 



1800 


1 1 1 


III 1 1 


1600 




1 


1400 




I 


1200 




1, — Hexagonal Ni3S4 breakdown curve 


1000 


- 


1 


800 


[\li3S4 with hexagonal structure 


1 


600 


~~"— -— ^ _ 


if 
if 
ij 


400 






200 


Polydymite 
1 1 1 


11 
11 

1 1 

' \^" Polydymite breakdown curve 

p Nj 3 s 4 =^ a Ni|_jrS+NiS 2 
1 L C 1 1 1 



250 



275 



300 



325 350 375 

Temperature, °C 



400 



425 



450 



Fig. 68. The stability field of polydymite is limited to 356°C along the temperature axis and 
to about 500 bars along the pressure axis. Twinning, responsible for the mineral name polydymite 
(many twins), takes place when N13S4 with hexagonal structure inverts to the cubic polydymite 
form. 



naeite, also have structures of the NiAs 
type. 

The Fe 3 S 4 compound is stable at low 
temperatures. On heating it decomposes 
below 100 °C to monoclinic pyrrhotite 
of about Fe 7 S 8 composition and to pyrite 
(FeS 2 ). This compound occurs in two 



polymorphic forms: greigite, which has 
cubic spinel-type structure, and smythite, 
which has a NiAs-derivative type struc- 
ture. The densities of the two polymorphs 
are nearly equal — Erd, Evans, and 
Richter (1957) gave 4.06 for smythite; 
and Skinner, Erd, and Grimaldi (1964) 



182 



CARNEGIE INSTITUTION 



give 4.08 for greigite. The effect of pres- 
sure on the greigite-smythite inversion, 
therefore, appears to be very small. Evi- 
dently, temperature is the main factor in 
deciding whether smythite or greigite 
forms. By analogy with the behavior of 
such minerals as polydymite, linnaeite, 
and daubreelite, smythite should be the 
high-temperature Fe 3 S 4 polymorph, and 
greigite should be stable at low tempera- 
tures. 

Preliminary studies on violarite (Fe 
Ni 2 S 4 ), carrollite (CuCo 2 S 4 ), siegenite 
[(Co,Ni) 3 S 4 ], bornhardtite (Co 3 Se 4 ), 
and triistedtite (Ni 3 Se 4 ) indicate that 
these minerals, which all have cubic 
spinel-type structures, invert to poly- 
morphs with NiAs-derivative hexagonal 
structures when exposed to elevated tem- 
peratures and pressures. A number of 
other sulfides, such as bornite (Cu 5 FeS 4 ) , 
chalcocite (Cu 2 S), and digenite (Cu 9 S 5 ), 
have Fd3m (spinel) -type structures at 
elevated temperatures. Bornite has this 
structure above 248 °C, chalcocite above 
430°C, and digenite above 73°C. In ac- 
cordance with the observations discussed 
above, it is to be expected that under 
pressure these spinel-type structures will 
become unstable and invert to a NiAs- 
derivative symmetry. 

The NiAs-type structure is stable to 
very high pressures, as indicated by 
experiments of Kullerud, Bell, and 
England (Year Book 64, pp. 197-199) 
on FeS and Fe^S, which have this 
structure. 

METEORITE MINERALS 

The Cr-S and Fe-Cr-S Systems 

A. El Goresy and G. Kullerud 

The purpose of the present study was 
to determine the phase relations in the 
Cr-S and Fe-Cr-S systems. Two new 
minerals have recently been reported to 
occur in the Cr-S system: Cr 2 S 3 (mineral 
A; Ramdohr, 1964) and CrS (El Goresy, 
1965). The study of the Cr-S system 
delineates the pressure-temperature fields 
of formation of the new minerals and 



is a prerequisite to investigation of the 
complex Fe-Cr-S system. Daubreelite 
(FeCr 2 S 4 ) is the only ternary phase. It is 
a common constituent of iron meteorites, 
enstatite chondrites, and achondrites, 
but is absent in all other types of meteor- 
ites. Troilite (FeS), the most common 
mineral in the ternary system, occurs 
in meteorites of all types. 

The Cr-S System 

Figure 69 shows the Cr-S phase dia- 
gram, based on data by previous in- 
vestigators and on our studies at 500°, 
600 o ,700°,and800°C. 

Cr^osS. This compound, which ap- 
parently occurs only in one crystallo- 
graphic modification, was observed in all 
experiments conducted with less than 
39.61 wt % S. In polished sections it is 
seen as lenticular lamellae occurring 
parallel to at least four crystallographic 
directions in /JCr^S. This type of inter- 
growth indicates an exsolution during 
quenching from a high-temperature ho- 
mogeneous /SCr^S phase. The aCr^sS 
phase is strongly reflecting (reflectivity 
similar to arsenopyrite) with weak bire- 
fringence and whitish color. The X-ray 
powder diffraction pattern of aCr^sS 
was indexed by Jellinek (1957) as prob- 
ably monoclinic. The X-ray powder dif- 
fraction pattern of this phase obtained 
by us agrees closely with that of Jellinek. 

Cr t . x S. This compound occurs in three 
crystallographic forms: a NiAs-type 
hexagonal (/?) form, a hexagonal form 
with NiAs supercell, and a monoclinic, 
low-temperature modification. 

/JCr^S was observed in all experi- 
ments containing less than 41.74 wt % S 
(Cr . 86 S). In quench products of experi- 
ments containing 37.62 wt % S or less, 
it occurs together with aCr^sS and Cr, 
and in quench products of experiments 
containing 39.61 wt % S or less it occurs 
with ceCri.osS only. In polished sections, 
/JCr^S is brownish in color, with reflec- 
tivity similar to that of pyrrhotite. The 
X-ray powder diffraction pattern of 



GEOPHYSICAL LABORATORY 



183 



80 



70 



60 



Atomic per cent 

50 40 30 



20 




Weight per cent 

Fig. 69. A partially schematic T-X diagram of the system Cr-S. Dashed lines represent earlier 
work or relations which were not accurately determined. 



/JCrx-sS is identical with that of Cr 7 S 8 , as 
given by Jellinek (1957) , but its stability- 
field is not so narrow as reported by 
Jellinek. The approximate compositional 
stability limits of ^Cr^S are shown in 
Fig. 69. 

X-ray studies on /JCr^S mix- crystals 
containing 41.74 to 39.37 wt % S (Cr .8 6 S 
to Cr . 93 S), synthesized at 700°C and 
quenched, revealed that the <i 102 value 
may vary considerably. Fig. 70 illustrates 
the variation of d 102 with composition, 
indicating that the d 1Q2 curve slopes 
steeply between Cr . 93 S and Cr . 89 S com- 
positions but its slope is gentle in the 
O0.89S to Cro.seS composition range. The 
curve offers a convenient method for 



determining the composition of /JCr^S 
mix-crystals. 

Hexagonal Cr^S with superstructure 
was observed only in experiments with 
sulfur ranging from 42.04 to 43.53 wt % 
S. In polished section this phase cannot 
be distinguished from /jCr^S. The X-ray 
powder diffraction patterns are distinctly 
different, however, with the superstruc- 
ture, hexagonal Cr^S phase displaying 
two characteristic reflections at low 20 
angles (d = 5.7759 A and d = 4.7374 A). 

The monoclinic Cri_J3 phase was ob- 
served in experiments containing 44.79 to 
43.53 wt % S. It occurs with rhombo- 
hedral yCr 2 .iS 3 ss. In polished sections 
it can be recognized because it contains 



184 



CARNEGIE INSTITUTION 

Weight per cent sulfur 



2.070 




39.36 


40.66 


42.05 


43.5 




1 


1 


1 I 




2.080 


- 








- 


2.090 
o tr\r\ 


- 


1 


1 1 


l ! I 


- 



Cr 0.95 s Cr 0.90 S Cr 0.8 5 S Cr 0.80 s 

52.50 55.00 57.50 60.00 

Atomic per cent sulfur 

Fig. 70. Plot of dio2 versus sulfur content for compounds of ^Cti- x S ss series. 



inversion twin lamellae similar to those 
produced when chalcopyrite inverts from 
a high-temperature cubic to a low-tem- 
perature tetragonal form. The X-ray 
powder diffraction pattern has two re- 
flections in the region of the 102 reflec- 
tion of the hexagonal (3 pattern and an 
additional reflection at d — 5.2430 A. The 
X-ray powder diffraction pattern given 
by Jellinek for monoclinic Cr 3 S 4 (Cr . 79 S 
to Cr . 76 S) is identical with that of a 
composite of reflections produced from 
a mixture of monoclinic Cr^S and 

yOaaSa SS. 

Cr s S s . Jellinek (1957) reported two 
forms of Cr 2 S 3 , a rhombohedral (y) form 
with a narrow compositional range 
(Cr .67S) and a trigonal form with Cr . 69 S 
composition. In the present investiga- 
tion, only the rhombohedral form was 
observed in runs with sulfur content 
ranging from 46.84 to 45.12 wt %, indi- 
cating a wide range of compositional sta- 
bility. In polished sections Cr 2 .iS 3 ss 
cannot easily be distinguished from 



other chromium sulfides. This compound 
does have a low-temperature form, as 
evidenced by twinning displayed in some 
polished sections. X-ray powder diffrac- 
tion patterns of the twinned material 
show that the low-temperature form 
contains two reflections in the general 
region of the 114 reflection of the original 
rhombohedral yCr 2 .iS 3 ss. This pattern, 
however, does not coincide with that of 
the trigonal Cr 2 S 3 reported by Jellinek. 

The Cr-Fe-S System 

Experiments were performed with 
compositions lying in the monosulfide 
solid solution field at 700° and 600 °C 
with mixtures of previously synthesized 
stoichiometric FeS and CrS as starting 
material, and in the sulfur-rich portion 
of the system with previously synthesized 
stoichiometric Cr 2 FeS 4 and elemental 
sulfur. Fig. 71 shows the phase relations 
in the Cr-Fe-S system at 700 °C. Above 
this temperature, complete solid solu- 



GEOPHYSICAL LABORATORY 



185 






700 °C 



lFe,Cr)S 2 +L+V 



2 ,|S 3 ;tL+V 



(Fe,Cr)S 2 +V 




(Fe,Cr)S 2 +{Fe,Cr),_ J1 S+V 
(Fe,Cr),_,S+V 



- S 3.t V 
Cr 2 FeS 4 +Cr 2 jS 3 +(Cr,Fe)|_ /r S+V 

Cr 21 S 3 +(Cr,Fe!|_,S+V 



Fig. 71. Phase relations in the Cr-Fe-S system at 700° C. All assemblages are in equilibrium 
with vapor. 



tion exists between Fe^S and Cr^JS. 
At 700° C, two monosulfide solid solutions 
produced through breakdown of the Mss 
phase coexist: (Fe,Cr)!_J3 (Mss^, ex- 
tending from the Fe-S boundary, and 
(Cr,Fe)i_J3 (Mssn), extending from the 
Cr-S boundary. These monosulfide solid 
solutions are separated by a divariant 
region containing Mssi and Mssn + F. 
At 700 °C the ternary system contains 
eight univariant fields. The composition 
of aFeCr ss coexisting with Mssi and 
Mssn was determined with the electron 
microprobe and found to be 93.8-95.7% 
Fe and 4.3-6.2% Cr. Fig. 72 shows the 
phase relations at 600°C. Figs. 71 and 72 
show that extensive changes take place 



in the system between 700° and 600°C. 
The solid solution of Cr in Mssi decreases 
markedly, whereas the solid solution of 
Fe in Mssn decreases only slightly. In 
the metal-rich portion of the system, the 
solubilities of Cr in aFeCr ss and of Fe 
in CrFe ss decrease considerably and the 
sizes of the divariant regions CrFe ss + 
(Cr^eh-aS+y and aFeCr ss+ (Cr, 
Fe)i_J3 + V narrow markedly. Our ex- 
periments demonstrate that at 650° ± 
50°C an invariant reaction takes place: 
(Fe,Cr)i-*S+ (Cr^eh^S (Mssx+Mssn) 
+ y^±Cr 2 FeS 4 + a:FeCr ss. As a result, 
two new univariant fields appear in the 
system: Cr 2 FeS 4 + (Fe,Cr) 1 _^S + aFeCr 
ss + 7 and Cr 2 FeS 4 + (Cn J Fe) 1 - - S+aFe 



186 



CARNEGIE INSTITUTION 



600 °C 



<Fe,Cr)S a +L+ 



•2.1S3+L+V 




(Fe,Cr)S 2 +V 
(Fe,Cr)S 2 +{Fe.Cr) l _ x S+ 
(Fe,Cr)^..S+V 



2.|S 3 + V 

Cr 2l S3+(Cr,Fe),.,S+V 
GjFeS^+C^ .S^+tCr.Fel^S+V 
(Cr,Fe)|_,S+V 



Fe),_,S+CrFe„+V 



Fig. 72. Phase relations in the Cr-Fe-S system at 600° C. All assemblages are in equilibrium 
with vapor. 



Cr ss + V. The composition of the FeCr ss 
phase, as synthesized in equilibrium with 
daubreelite at 600 °C, was determined 
with the electron microprobe and found 
to be 98.9 to 99.4 wt % Fe and 0.6 to 
1.1 wt °/o Cr. The cell parameter of 
daubreelite was measured with the use 
of its 440 reflection, with Lake Toxaway 
quartz as internal standard. The cell edge 
of Cr 2 FeS 4 coexisting with aFeCr ss was 
found to be 9.996 ±0.004 A. At 600 °C, 
as at 700 °C, eight univariant fields exist 
in the system (Fig. 72) . 

In the sulfur-rich portion of the sys- 
tem Cr-Fe-S, tie lines exist between Cr 2 
FeS 4 and liquid sulfur at 600 °C. At lower 
temperatures (probably near 500 °C) 



an invariant reaction takes place, Cr 2 
FeS 4 +L s + 7 ? ±Cr 2 . 1 Ss ss+ (Fe,Cr)S 2 , 
establishing the assemblage Cr 2 iS 3 ss + 
(Fe,Cr)S 2 . 

Applications of Phase Diagrams to 
Meteorites 

Typical sulfide nodules in iron meteor- 
ites contain troilite and daubreelite as 
major constituents. In enstatite chon- 
drites and achondrites these minerals 
also occur together and form mutual 
boundaries with kamacite. The daubree- 
lite occurs as lamellae of variable sizes, 
oriented parallel to the 0001 of troilite 
(Ramdohr, 1963; El Goresy, 1965, 1967). 
The daubreelite-to-troilite ratios ob- 



GEOPHYSICAL LABORATORY 



187 



served vary considerably from one 
meteorite to another, and the daubreelite 
content may even vary from s=«5% to 
more than 60% by volume in different 
parts of the same meteorite. This varia- 
tion in mineral concentration corresponds 
to a range of 1 to 25% Cr, 20 to 30% S, 
and 60 to 70% Fe. Our experiments 
demonstrate that bulk compositions 
ranging from 50FeS + 50CrS to 25FeS + 
75CrS (mole %) are accounted for by 
Mssi + Mssn + aFeCr ss assemblages at 
700°C. Compositions with 50 to 75 mole 
% FeS will form one homogeneous phase 
(Mssi), whereas compositions with more 
than 75 mole % CrS will form another 
homogeneous phase (Mss n ) . In a cooling 
meteorite parent body containing (Fe, 
Cr)!_J3 mix-crystals with compositions 
in the range 50 to 75 mole % Cr^S, 
the invariant reaction Mssi + Mss n + 
V ;=± db + aFeCr ss takes place at 650° ± 
50°C. Mix-crystals with more than 50% 
CrS will give the assemblage daubreelite 
+ Mss n + aFeCr ss-f-F, whereas those 
with less than 50% CrS result in the as- 
semblage daubreelite + Mss! + o!FeCr ss 
+ V. With decreasing temperature, the 
solubility of Cr in Mss r diminishes con- 
siderably, the solubility of Fe in Mss n 
decreases only slightly, and the uni- 
variant field Cr 2 FeS 4 + Mssi + aFeCr ss 
+ V widens markedly. The crystallization 
of daubreelite and aFeCr ss from Mssi 
mix-crystals with ~25 mole % CrS takes 
place below 600°C, and the aFeCr ss 
separating from these mix-crystals con- 
tains less than 0.5 wt % Cr. In iron 
meteorites the aFeCr ss tends to react 
with kamacite during the long cooling 
periods. Thus Cr is distributed in the 
major metallic phase. Under these condi- 
tions, minor amounts of Cr in the metal- 
lic phase of iron meteorites would not 
be detectable with microprobe tech- 
niques. Enstatite chondrites and achond- 
rites contain much less kamacite than 
iron meteorites, and hence Cr might be 
detected in the reequilibrated kamacite 
of the latter. 
The present study demonstrates that 



the assemblage daubreelite H-Mssi + aFe 
Cr ss originates at relatively low tem- 
peratures from Cr-poor monosulfide mix- 
crystals. The presence of Mssi + Mss n in 
the 40:60% ratio — as in the Bethany, 
Obernkirchen, and Russel Gulch meteor- 
ites — may indicate rapid cooling to tem- 
peratures below 700 °C. The presence of 
Ni in solid solution, forming (Cr,Fe, 
Ni)i_*S mix-crystals, however, may pre- 
vent the establishment of the daubree- 
lite -l-aFeCr ss assemblage, since Fd-^S- 
Nii-^S solid solution is complete down to 
300°C (Naldrett, Craig, and Kullerud, 
1967). 

Sulfide Assemblages in the Odessa 
Meteorite 

G. Kullerud and A. El Goresy 

Applications of pertinent phase dia- 
grams, produced by experimental studies 
on synthetic systems, have shown that 
lack of equilibrium between minerals in 
one and the same assemblage is not a 
rare occurrence in meteorites and that 
equilibrium between different assem- 
blages, often only a few millimeters 
apart, is usually not achieved. It appears, 
as we shall see below, that some of the 
nonequilibrium assemblages found in 
meteorites may be explained as effects of 
shock. 

To investigate this phenomenon we 
have chosen to make a detailed examina- 
tion of mineral associations occurring 
in the Odessa meteorite. This meteorite, 
which was found near Odessa, Texas, in 
1922, weighed more than 1 ton and is 
classified as an iron meteorite. In the 
Odessa iron meteorite small troilite 
nodules, usually rimmed by graphite, are 
distributed throughout an FeNi matrix. 
Studies of polished sections of these 
nodules reveal that they are polymin- 
eralic, although troilite is always by far 
the most abundant phase. Some of the 
troilite spherules in the Odessa meteorite 
contain assemblages of heazlewoodite 
(Ni 3 S 2 ), daubreelite (FeCr 2 S 4 ), and 
sphalerite [(Zn,Fe)S], in addition to 



188 



CARNEGIE INSTITUTION 



troilite (FeS). Others contain assem- 
blages of awaruite (Ni 3 Fe), daubreelite, 
sphalerite, and troilite. 

We have some knowledge of the be- 
havior of the FeCr 2 S 4 and (Zn,Fe) S com- 
pounds and have detailed information on 
the Fe-Ni-S system. Neither troilite and 
heazlewoodite nor troilite and awaruite 
coexist stably at any temperature in the 
condensed Fe-Ni-S system. At elevated 
pressures, troilite becomes stable with 
heazlewoodite over a relatively large 
temperature range and with awaruite 
over a much more restricted range. To 
provide a background for interpreting 
observations on minerals in meteorites, 
we present a discussion of some of the 
physical chemistry of the pertinent con- 
stituents. 

Troilite (FeS) has a simple NiAs-type 
structure above 139 °C and forms a 
superstructure below this temperature. 
The high-temperature form cannot be 
quenched, but the inversion can readily 
be detected by DTA experiments. Pres- 
sure decreases the inversion temperature 
to about 90°C at 20 kb (Kullerud, Bell, 
and England, Year Book 64, pp. 197- 
199) . In other words, pressure lowers the 
inversion temperature by about 2.5 °C/ 
kb. 

The Ni 3 S 2 compound occurs in two 
crystalline modifications. The low-tem- 



perature form is hexagonal and is repre- 
sented in nature by the mineral heazle- 
woodite. This form is stoichiometric 
(Kullerud and Yund, 1962) and takes 
less than 1 wt % Fe in solid solution. 
At 556 °C it inverts to a high-temperature 
pseudocubic (a) modification that ex- 
hibits extensive solid solution toward 
both Ni and S (Kullerud and Yund, 
1962) and may also take as much as 
5 wt % Fe in solid solution, lowering 
the inversion temperature to about 
525 °C. Kullerud, Bell, and England 
{Year Book 64, pp. 197-199) found that 
pressure increases the inversion tempera- 
ture of pure stoichiometric Ni 3 S 2 by 
about 1.7°C/kb and that of iron-satu- 
rated a(Ni,Fe) 3± *S 2 by about 0.3°C/kb. 
Taenite (yFe,Ni) is a high-tempera- 
ture cubic Fe-Ni solid solution. Awaruite 
is a cubic Ni 3 Fe phase, which is stable 
below 503 °C in the condensed Fe-Ni-S 
system. The phase relations at 500 °C in 
the pertinent portion of the condensed 
Fe-Ni-S system are shown in Fig. 73. 
The cubic pentlandite (Fe,Ni) 9 S 8 phase 
in the presence of vapor is stable below 
610°C (Kullerud, 1963). Above this 
temperature it decomposes to the hexa- 
gonal (Fe,Ni) i-J3-type phase and a(Ni, 
Feh-aSo. Bell, England, and Kullerud 
(Year Book 63, pp. 206-207) showed 
that the temperature at which pentland- 




Fe aFeNi 



Fig. 73. Phase relations at 500°C in the metal-rich portion of the condensed Fe-Ni-S system. 
On cooling to about 450°C the 7FeNi ss-Ni 3 S 2 tie lines are replaced by awaruite-pentlandite tie 
lines. 



GEOPHYSICAL LABOEATOEY 



189 



ite decomposes decreases markedly with 
increased pressures; for instance, at 
25 kb pentlandite decomposed at 425 °C. 
The pentlandite-taenite tie lines shown 
in Fig. 73 exist below about 580°C, pro- 
hibiting stable coexistence of troilite and 
heazlewoodite. It is shown in this figure 
that at 500 °C tie lines exist between 
taenite and heazlewoodite; thus troilite- 
awaruite coexistence is prohibited by two 
sets of tie lines (taenite-pentlandite and 
taenite-heazlewoodite) . At about 450°C, 
however, the taenite-heazlewoodite as- 
semblage becomes unstable in the con- 
densed system, and below this tempera- 
ture awaruite is stable with pentlandite. 
The reaction rates in the central por- 
tion of the system are extremely rapid. 
For instance, it is not possible to pre- 
serve troilite + a: (Ni,Fe) 3 S 2 assemblages 
in the condensed system, even by the 
most rapid chilling. On cooling, peni- 
landite always forms at 610 °C, taenite 
appears at 580°C, and the ce(Ni,Fe) 3 S 2 
phase inverts to the heazlewoodite form 
at about 525 °C. 

Troilite-Heazlewoodite-Daubreelite- 
Sphalerite Assemblages 

Some nodules of the Odessa meteorite 
contain troilite, heazlewoodite, daubree- 
lite, and sphalerite in physical contact 
(El Goresy, 1967). Examination of fine 
textures displayed in the heazlewoodite 
reveals that originally an a(Fe,Ni) 3 S 2 
phase formed. It later broke down to 
heazlewoodite and a few percent of 
troilite, which form exceedingly fine 
myrmekitic intergrowths. 

By electron probe analyses the heazle- 
woodite was found to contain 65.3 Ni, 
4.6 Fe, 0.2 Co, and 27.4 S (wt %). The 
composition corresponds closely to that 
of the a. (Ni,Fe) 3± J3 2 phase when coexist- 
ing with troilite above 610°C (Kullerud, 
Year Book 62, pp. 175-189) . Apparently, 
the heazlewoodite is the product of 
breakdown of a homogeneous a(Ni,Fe, 
Co) 8±a ,S 2 phase, which had a metal-to- 
sulfur ratio of 2.80:2, as calculated from 



the analysis given above. The P-T curve 
(I) for this breakdown is shown in 
Fig. 74. The a(Ni,Fe,Co) 2 . 8 S 2 phase 
formed in the P-T field on the right side 
of curve I. Since it coexisted with 
troilite, and pentlandite did not form, 
the P-T conditions must have exceeded 
those of curve II. Above this curve, troi- 
lite + a (Ni,Fe,Co) 3±2! S2 coexist stably; 
below this curve pentlandite is stable. 
The a (Ni,Fe,Co) 2 . 8 S 2 +± heazlewoodite 
breakdown must have taken place at 
pressures exceeding 14 kb; otherwise 
heazlewoodite + troilite would have re- 
acted to produce pentlandite. 

The reactions on curves I and II are 
so rapid that Bell, England, and Kullerud 
(Year Book 63) were able to determine 
these curves by DTA experiments. Even 
at 425 °C the reaction is sufficiently 
rapid to be detected by DTA. 

Preservation of the troilite-heazle- 
woodite mineral assemblage in a meteor- 
ite, therefore, necessitates that cooling 
take place under pressures above curve 
II, to low temperatures, at which the 
reaction to form pentlandite is very 
sluggish. Reaction-rate studies conducted 
in silica tubes at 200°, 250°, and 300°C 
on synthetic mixtures of FeS and Ni 3 S 2 , 
showed no reaction even after 6 months 
at 200°C, beginning reaction at 250°C, 
and a high percentage of pentlandite at 
300°C. 

Although reaction rates in silica tubes 
are not directly comparable with those 
attained under high confining pressures, 
these experiments indicate that preserva- 
tion of the troilite-heazlewoodite as- 
semblage in meteorites requires mainte- 
nance of pressure in excess of 40 kb 
down to a temperature of about 250°C. 
The study of polished sections also re- 
veals that the FeCr 2 S 4 compound was 
not cubic when originally formed, but 
apparently hexagonal. It inverted to 
daubreelite at a later stage in the history 
of the meteorite. Albers and Rooymans 
(1965) found that FeCr 2 S 4 exists in two 
polymorphs, daubreelite and a hexagonal 
form with NiAs-type structure. They 



190 



CARNEGIE INSTITUTION 




500 600 700 

Temperature, °C 



100 



Fig. 74. Stability fields of the assemblages troilite + hexagonal FeCr 2 S4 + awaruite +sphalerite 
and troilite -f- hexagonal FeC^ + a(Ni,Fe) 3 ± a ,S 2 + wurtzite. 



determined the P-T curve of this in- 
version (curve IV in Fig. 74). 

Fine textures in polished sections also 
show that the zinc sulfide when originally 
deposited did not have the cubic sphaler- 
ite structure, as evidenced by its 
morphology. Pure ZnS, which is cubic at 
low temperatures, inverts to hexagonal 



wurtzite when heated to about 1020°C. 
Zinc sulfide forms extensive solid solu- 
tions with FeS and MnS (Kullerud, 
1953), both of which depress the tem- 
perature of the cubic ?=± hexagonal inver- 
sion. Electron probe analyses of the 
sphalerite in the Odessa meteorite show 
that it consists of 50.8 ZnS, 38.2 FeS, and 



GEOPHYSICAL LABORATORY 



191 



11.0 MnS (mole %). The P-T curve for 
the sphalerite-wurtzite transition for 
this composition is not known. It is esti- 
mated that the inversion takes place at 
700°±50°C at 1 atm, from the work 
on the FeS-ZnS and MnS-ZnS joins by 
Kullerud (1953), and that the slope of 
the P-T curve is 5° to 10°C/kb. This 
curve is labeled V in Fig. 74. 

The original assemblage troilite + 
a(NiFe) s-^ + hexagonal FeCr 2 S 4 + 
wurtzite must have formed at a tempera- 
ture limited downward by the sphalerite- 
wurtzite curve and upward by the melt- 
ing curve of a:(Ni,Fe)3S 2 . This curve 
(labeled VI in Fig. 74) starts at 862 °C 
in the condensed system (Kullerud, 
1963) and is expected to have a steep 
slope (==* + 3°C/kb). The pressure when 
the above assemblage formed exceeded 
the FeCr 2 S 4 inversion curve. On cooling, 
the a(Ni,Fe) 3 S 2 phase decomposed to 
heazlewoodite and troilite at about 
550 °C under pressure exceeding about 
20 kb. Troilite and heazlewoodite did 
not react to form pentlandite. The pres- 
sure, therefore, must have been main- 
tained above curve III to at least 250°C. 
Reaction rates in the Fe-Ni-S system 
are sluggish below this temperature, and 
a release of pressure would not result 
in pentlandite formation during subse- 
quent cooling. The inversion from hexa- 
gonal FeCr 2 S 4 to daubreelite apparently 
took place when pressure was released 
below about 250°C. 

Troilite- Awaruite-Daubreelite- 
Sphalerite Assemblages 

Studies of the fine textures displayed 
in some troilite nodules in polished sec- 
tions reveal that hexagonal FeCr 2 S 4 , 
cubic sphalerite, troilite, and awaruite 
originally coexisted (El Goresy, 1967). 
At the time of formation of this assem- 
blage the pressure must have exceeded 
that of the FeCr 2 S 4 hexagonal *± cubic 
inversion curve. The temperature must 
have been below that of the sphalerite ?± 
wurtzite inversion curve. The awaruite 



phase is stable below 503 °C in the con- 
densed Fe-Ni system. Pressure will influ- 
ence its stability at the rate of 1°-3°C/ 
kb. The temperature could not have ex- 
ceeded the awaruite stability curve (VII 
in Fig. 74). Coexistence of troilite and 
awaruite requires changes in two sets 
of tie lines, those connecting taenite and 
heazlewoodite and those connecting tae- 
nite and pentlandite. In the condensed 
Fe-Ni-S system the reaction taenite + 
heazlewoodite + V ?±awaruite + pentland- 
ite takes place at about 450° C, awaruite 
+ pentlandite being stable together be- 
low this temperature. Under high confin- 
ing pressures the balanced reaction can 
be written 

10.5yFeNi+4Ni 8 S 2 ?± 6Ni 3 Fe+ 

Fe 4 . 5 Ni4. 5 S 8 

Calculations of the volumes involved 
demonstrate that the awaruite-pentland- 
ite assemblage is favored by pressure. 
The awaruite -(-pentlandite mineral pair 
is stable at increasing temperatures (be- 
yond 450 °C) with increasing pressure. 
Kullerud (Year Book 62) showed that 
pentlandite-taenite tie lines exist below 
about 580 °C in the condensed Fe-Ni-S 
system. The balanced equation for the 
upper stability curve for the pentlandite- 
taenite assemblage can be written 

5Fe 4 5 Ni 4 5 S 8 +1.5 7 FeNi <=± 

24FeS + 8Ni 3 S 2 

Calculations of the volumes of the 
phases involved demonstrate that the 
assemblage of troilite + heazlewoodite is 
favored by pressure, and the upper stabil- 
ity curve of the pentlandite-taenite as- 
semblage has a negative slope. This 
reaction is similar to that involving 
breakdown of pentlandite to heazlewood- 
ite + a (Fe,Nih_sS. The P-T curve of 
pentlandite was shown by Bell, England, 
and Kullerud {Year Book 63) to have a 
pronounced negative slope. In Fig. 74 the 
curves pentlandite + taenite <=± heazle- 
woodite + troilite (III) and pentlandite 
?± heazlewoodite + a(Fe,Ni)i_ ;r S have 
been assumed to be essentially parallel. 



192 



CARNEGIE INSTITUTION 



It is seen from this discussion and 
from Fig. 74 that troilite-awaruite as- 
semblages are stable in the P-T region 
bounded upward in temperature by the 
upper stability curve of the awaruite 
phase and downward in pressure by the 
troilite + heazlewoodite ^ pentlandite + 
taenite curve. 

Formation and preservation of the as- 
semblages troilite + hexagonal FeCr 2 S 4 + 
a(Ni,Fe) 3 S 2 + wurtzite and troilite + 
hexagonal FeCr 2 S 4 + awaruite + sphaler- 
ite require pressures of at least 40 kb. 
The first assemblage, however, requires 
an initial temperature of at least 750 °C, 
whereas the second assemblage is not 
stable beyond about 550°C. If these as- 
semblages formed more or less simul- 
taneously in adjacent spherules, perhaps 
only a few millimeters apart, a very 
strong temperature gradient must have 
existed. Such a gradient could not have 
persisted for long in an iron-nickel mass, 
which is an excellent heat conductor. 

Graphite is associated with all troilite 
nodules. The graphite-diamond P-T 
curve is shown in Fig. 74. It is noted 
that preservation to low temperatures 
of the mineral assemblages discussed 
above requires pressures in the diamond 
stability field. Maintenance of such pres- 
sures over an extended period of time 
would have converted the graphite to 
diamond. 

The metal phase of the Odessa meteor- 
ite contains small amounts of almost 
pure albite and orthoclase (El Goresy, 
1967). Under the P-T conditions indi- 
cated by the sulfide assemblages, albite 
under equilibrium conditions reacts to 
give jadeite + Si0 2 . The P-T curve for 
albite <=» jadeite + Si0 2 , therefore, could 
not have been exceeded for very long. 

The mineral assemblages discussed 
above collectively present strong evi- 
dence of shock. Because of the rapid 
reaction rates displayed by sulfides, these 
minerals responded to the shock, whereas 
slowly reacting minerals, such as albite, 
did not. 

The shock evidently produced very 



inhomogeneous temperatures. The maxi- 
mum temperature did not result in melt- 
ing of the a (Ni,Fe) 3 -tfS 2 phase and there- 
fore did not exceed about 900 °C. The 
pressure of the shock must have exceeded 
40 kb but was not sufficient to produce 
maskelynite, since the orthoclase phase 
remains. 

SULFUR-ISOTOPE FRACTIONATION 
IN MINERALS 

H. Puchelt 

The S 34 /S 32 ratios of minerals from 
different geological settings vary con- 
siderably. A number of geochemical pro- 
cesses are believed to have caused this 
phenomenon. Isotope fractionation in 
biogeochemical processes is substantial 
and has received much attention, but 
comparatively few efforts have been 
made to determine sulfur-isotope frac- 
tionation accompanying inorganic reac- 
tions of metallic sulfides under controlled 
conditions. During the past year, such 
experiments have been carried out in 
well-known, simple sulfide systems. 

Analytical Procedure 

The method of Sabels and Hoering 
(Year Book 62, pp. 238-239) for the 
preparation of sulfur hexafluoride and 
its subsequent introduction into the mass 
spectrometer for isotope analysis was 
improved and developed into a routine 
procedure. The main simplifications were 
the dispensing of the ffuorinating agent, 
BrF 3 , into the reaction vessels as a liquid 
and the quantitative recovery of the re- 
action product. A fiftyfold excess of BrF 3 
was reacted with sulfides at 230 °C for 
at least 1 hour. A quantitative yield of 
SF 6 was obtained from pyrite, galena, 
sphalerite, covellite, chalcopyrite, cinna- 
bar, cadmium sulfide, and elemental sul- 
fur. Pyrrhotite, troilite, and digenite do 
not react completely but could be con- 
verted to cadmium sulfide first and then 
fluorinated. 

The method has been used for over 
350 samples. An interlaboratory check 



GEOPHYSICAL LABORATORY 



193 



was carried out with the Zentrallabora- 
torium fur die Geochemie der stabilen 
Isotope, Gottingen, Germany, which uses 
a completely different system of sample 
preparation, sulfur dioxide being the gas 
employed for mass spectrometry. The 
results of both laboratories agreed to 
within 0.2 % for a suite of four sulfide 
samples. 

Results are expressed in parts per 
thousand difference of the S 34 /S 32 ratios 
as compared with a standard sulfur 



8S 34 = 



(S 3 7S 82 )*-(S 8 7S 32 ). 



X1000 



(S 34 /s 32 ) s 

where the subscript x refers to the sample 
under investigation and the subscript s 
refers to the standard sulfur, which in 
this study was troilite from the Canyon 
Diablo meteorite. 

Sulfur-Isotope Fractionation in the 
Reactions of Metallic Sulfides 

Formation of sulfides from the ele- 
ments. Fifty mg of lead and a 50% 
excess of sulfur were sealed in evacuated 
quartz tubes and heated to 300 °C. Like- 
wise, copper, lead, and excess sulfur were 
reacted, with the metal powders kept 
separated in open gold tubes. Metallic 
sulfides formed rapidly. The runs were 
quenched, and the excess sulfur was ex- 
tracted with trichloroethylene. The rela- 
tive abundance of the sulfur isotopes in 
the separated phases was measured. 

Table 12 shows the difference in SS 34 
of the lead sulfide and sulfur as a 
function of reaction time. During the 

TABLE 12. Isotopic Composition of the Sulfur 

in Galena, and Sulfur during the Reaction 

of Lead and Sulfur at 300°C 



initial reaction, the light isotope concen- 
trated in the lead sulfide. Then, a slow 
exchange of sulfur atoms between sulfur 
vapor and the sulfide took place: 



PbS 34 + S 3 



PbS 32 + S 34 



(1) 



Extrapolation of the data to long periods 
of time indicates that the equilibrium 
constant for reaction 1 may be 1.002 at 
300 °C. Table 13 shows the results for the 
combined copper-lead sulfur system. 
Digenite, an intermediate phase in the 
reaction, was found in both runs, the 
larger proportion being found in the 
shorter run. During the initial sulfide 
formation, S 34 concentrated in the copper 
sulfides and S 32 in the lead sulfide. Fol- 
lowing this, isotopic exchange between 
the sulfides and sulfur vapor took place, 
tending toward the equilibrium state. 

In contrast to the rapid exchanges of 
sulfides and sulfur, no measurable change 
in sulfur-isotope ratios occurred when 
BaS0 4 and galena were heated to 500 °C 
for 240 hours. 

Sulfur-isotope exchange between sul- 
fides and sulfur. Fifty-mg samples of a 
natural galena were heated with an 
equivalent amount of sulfur at 400° and 
500 °C for periods up to 96 hours. The 
initial difference in the S 34 content was 
18 %o. Isotopic exchange took place, re- 
ducing the difference in SS 34 to 0.8 % 
in 24 hours at 500 °C and to 4.1 % in 
96 hours at 400 °C (Fig. 75). Extrapola- 
tion of the data in Table 12 indicates 
that at 300 °C an initial difference of 
10.8 % decreased to 4.6 % in 16 hours. 
The sulfide prepared from pure materials 

TABLE 13. Sulfur-Isotope Fractionation be- 
tween Copper Sulfide, Lead Sulfide, and 



5S S4 (sulfur)-SS 34 (galena), 

«/ 

/CO 


Reaction Time, 
hours 




Sulfur 


Vapor 


at 300 °C 




Phase 




5S 34 , %o 


Reaction Time, 
hours 


4.8 
7.6 
10.8 
7.5 
6.1 
4.6 
3.9 
2.8 


0.08 3 

0.25 

0.42 

1 

5 

16 

96 
240 


Sulfur 

Covellite 

Galena 

Sulfur 

Covellite 

Galena 




2.6 
4.3 

—1.4 
6.5 
1.6 

—1.1 




24 
192 



194 

40 



CARNEGIE INSTITUTION 



D—D Galena - sulfur 400°C 
o o Galena- sulfur 500°C 
*■— k Pyrrhotite- sulfur 800=0 



Q 



0, 



I <L 

Reaction time, in days 

Fig. 75. Sulfur-isotope exchange between sulfides and sulfur. 



seems to exchange more rapidly than 
the natural one. 

The same reaction was brought about 
with pyrrhotite and sulfur at 800 °C 
(above the pyrite invariant point) . Here, 
the initial difference in 8S 34 of 38.4 % 
decreased to 1.5 % in 16 hours. 

Some geological consequences are im- 
plied by these results. A sulfur vapor 
phase, generated by the breakdown of 
incongruently melting sulfides or by sol- 
fataric volcanic activity, may react with 
existing sulfides, causing exchange of 
their sulfur isotopes. A complete altera- 
tion of the original sulfur-isotope ratio 
in the sulfides may take place. A change 
in the isotope ratio of the sulfur vapor 
may also occur. 

Isotope homogenization between sul- 
fides. In a series of experiments, two 
samples of a mineral of natural origin 



with different sulfur-isotope ratios were 
placed separately in gold tubes contained 
in a quartz vessel. The evacuated, sealed 
vessel was heated, the reaction was 
quenched, and the isotopic composition 
of the minerals was redetermined. 

Two galena samples (50 mg each; dif- 
fering by 27.9 % ) were heated at 500° 
and 600 °C. The difference in sulfur-iso- 
tope ratio as a function of time is plotted 
in Fig. 76. After 264 hours at 500 °C, the 
samples had exchanged and had a dif- 
ference of 2.4 % . At 600 °C the same de- 
gree of homogenization was reached in 
only 12 hours. The minerals, which were 
initially fine powders (less than 150 
mesh), were considerably recrystallized 
at the end of the run. Equal weights of 
two sphalerites showed a much lower 
rate of exchange. Heating for 96 hours 
at 600 °C produced no appreciable change 



GEOPHYSICAL LABORATORY 



195 



30 



D— O Sphalerites 600°C 

H H Sphalerites 700°C 

■ ■ Sphalentes 800°C 



"•crw^wwrnrnm m mm mm m^Trm it. 



Galenas 500°C 
O— -O Galenas 600°C 
• • Galenas 650°C 




ir-.TT.rr.-r.Ta. 



1' 






2 3 4 5 6 7 8 

Reaction time, in days 
Fig. 76. Sulfur-isotope homogenization between galenas and sphalerites. 






in isotopic ratios; 240 hours at 800 °C 
caused the initial difference of 19.2 % to 
change to 16.4 % . Pyrites reacted very 
slowly at 500°C. The data at 650°C in 
Fig. 77 show that after 39 days the initial 
difference in SS 34 values had decreased 
from 56.7 % to 45.6 % . The experiments 
establish a trend in the rates of reac- 
tions of isotope homogenization in min- 
eral sulfides: galena is more rapid than 
pyrite, which is more rapid than sphaler- 
ite. 

Breakdown Experiments 

A sulfur-isotope fractionation has been 
reported during the incongruent melting 
of sulfides such as pyrite and covellite 
(Thode, 1963). This has been investi- 
gated further by analysis of S generated 
by heating pyrite, marcasite, and covel- 
lite for 15 minutes above their invariant 
points. The pyrite and marcasite showed 
little or no isotope effect during break- 
down under these conditions. Covellite 



showed a slight isotope effect; the sulfur 
was +2.9 % different from the starting 
material. 

The breakdown of pyrite below the 
invariant point was studied in a tube 
with a temperature gradient. The sample 
was kept at 600°C, and the sulfur dis- 
tilled to the cold end. Table 14 shows 
the results of such an experiment where 
the sulfur has a relatively large isotopic 
difference from the pyrite. Presumably 
this is due to the lack of exchange of 
sulfur vapor with the remaining pyrite. 
In a natural, open system, such a process 
might cause considerable isotope varia- 
tions. 

Sulfur-Isotope Ratios in Minerals of the 
Bodenmais, Germany, Deposit 

The results discussed above suggest 
that homogenization of sulfur-isotope 
ratios in sulfide minerals may take place 
at elevated temperatures in nature. This 
process may utilize a vapor of elemental 



196 



CARNEGIE INSTITUTION 




"0 20 

Fig. 77. Sulfur-isotope homogenization between pyrites 



40 60 

Reaction time, in days 



100 



sulfur. Conditions of this kind occur in 
sulfide ore deposits that have undergone 
thermal metamorphism. The conditions 
of metamorphism of the Bodenmais, Ger- 
many, ore deposit have been studied by 
Schreyer, Kullerud, and Ramdohr 
(1964), who showed that temperatures 
as high as 730 °C and pressures of as 
much as 3000 bars prevailed. 

The sulfides from 23 hand specimens 
of this deposit have been analyzed for 
their sulfur-isotope compositions, and 
the distribution of the results is shown 
in Fig. 78. The results fall in a relatively 
narrow range of ratios ( + 10.3 to 
14.7 %o), with a mean value for SS 34 of 
+ 12.2 %o. The samples were taken from 

TABLE 14. Isotope Effect in the Breakdown of 
Pyrite 



Heating 
Time, 
days 



% S Released 
from the 
Possible 
Amount 



A %o 
5S 34 (sulfuri- 
cs 34 (initial pyrite) 



1 

2 

3 

10 



5 

7 

7 

81 



—9.2 
—3.4 
—1.2 
—0.6 



a number of levels in the mine, over a 
vertical distance of several hundred 
meters. 

Two or more coexisting sulfides in 18 
hand specimens were separated and 
analyzed for sulfur-isotope ratios. A 
trend in the enrichment of sulfur isotopes 
between coexisting minerals was ob- 
served. In 7 out of 10 coexisting pyrrho- 
tite-pyrite pairs, SS 34 (pyrite) was equal 
to or greater than SS 34 (pyrrhotite) . In 
7 out of 8 examples, SS 34 (sphalerite) 
was less than SS 34 (pyrrhotite) . In 4 out 
of 5 pairs SS 34 (sphalerite) was less than 
SS 34 (pyrite). Thus the minerals seem 
to show a tendency for concentrating S 32 
in the order sphalerite > pyrrhotite > py- 
rite. 

Hegemann and Maucher (1933) have 
shown that the ore in Bodenmais was 
probably originally sedimentary. One of 
the characteristics of many sedimentary 
sulfide ore deposits is their wide range of 
sulfur-isotope ratios. Ore from a mag- 
matic or magmatic-hydrothermal source, 
on the other hand, is characterized by a 
narrow range of sulfur isotope, with 
SS 34 commonly near zero on our scale. 



GEOPHYSICAL LABORATORY 



197 



13 
12 
il 

CD 

\?\0 

•S 8 
o 

<D 

.E 7 
to 
a> 

o 
in 

"b 5 

-Q 4 

e 

2 3 
2 








- 












































- 






































































n 















+10 



►12 +13 +14 



£S 34 7« 



Fig. 78. Sulfur-isotope composition in Boden- 
mais, Germany, sulfide samples. 

The isotopic composition of the Boden- 
mais ores indicates that the sulfur had 
at least one process that fractionated 
isotopes. A bacteriogenic sulfate reduc- 
tion, with an isotope effect, may have 
taken place but the metamorphic event 
that followed destroyed the hetero- 
geneous distribution. 

The results of sulfur-isotope measure- 
ments on Bodenmais may be most easily 
explained in terms of an originally sedi- 
mentary sulfide deposit that suffered 
thermal metamorphism. A sulfur vapor 
phase, generated from the breakdown of 
pyrite, caused the homogenization of the 
isotope ratios between the minerals. 
These conclusions are consistent with the 
geological situation, petrography, and 
microscopic studies of the ore minerals. 

SYNTHESIS OF ^C^FeS* AT 35 KB 

A. El Goresy, P. M. Bell, and J. L. England 

In the condensed Cr-Fe-S system, 
Cr 2 FeS 4 is a stable compound below 



1350 °C at 1 atm. It occurs naturally as 
a low-temperature spinel-type mineral 
known as daubreelite (a:Cr 2 FeS 4 ) , a com- 
mon accessory in enstatite chondrites, 
achondrites, and iron meteorites. 

The textural relations of daubreelite- 
bearing meteorites suggest that Cr 2 FeS 4 
formed as a high-pressure (3 phase that 
inverted to the a form on cooling. A high- 
pressure hexagonal polymorph with a 
structure analogous to that of NiAs, re- 
ported by Rooymans and Albers (1967), 
is stable at 1 atm above 1060°C. These 
investigators studied the stability rela- 
tions of Cr 2 FeS 4 and showed that the 
P-T curve for the inversion has a 
strongly negative slope. It was not pos- 
sible for them to recover the stoichio- 
metric high-pressure form or produce its 
X-ray powder diffraction pattern be- 
cause they used a squeezer apparatus of 
limited temperature range. 

In the present study, synthesis of the 
high-pressure form of daubreelite was 
accomplished with a solid-media, piston- 
cylinder apparatus. Pure synthetic dau- 
breelite (a = 9.996 ±0.002 A) was used 
as starting material. 

Syntheses were conducted at 300° and 
750°C, for 48 hours at 35 kb. These runs 
were quenched to 100°C in 10 to 15 
seconds under pressure. From the 300° C 
run, only cubic aCr 2 FeS 4 was recovered. 
At 750°C, however, complete conversion 
of aCr 2 FeS 4 to a high-pressure phase 
(/3Cr 2 FeS 4 ) took place. This is the first 
time that the (3 form has been synthe- 
sized. Electron-microprobe analyses were 
carried out on the recovered high-pres- 
sure phase with stoichiometric Cr 2 FeS 4 
as a standard. No compositional differ- 
ences between the standard and most of 
the high-pressure phase were observed. 
Only one grain of the /?Cr 2 FeS 4 was 
found to contain small domains (2 to 3 /* 
wide) with higher Cr and lower Fe and 
S contents. The density of the high-pres- 
sure form was 4.15. 

/?Cr 2 FeS 4 is considerably harder to 
abrade and more strongly reflecting than 
cubic aCr 2 FeS 4 . In crossed nicols all 



198 



CARNEGIE INSTITUTION 



medium- and large-sized grains (50 to 
100 /a) show one set of twin lamellae 
parallel in one direction (presumably 
basal). 

The material was subjected to high- 
pressure differential thermal analysis 
(Year Book 63, p. 176) at 9 kb, 367° to 
818°C; 11 kb, 579° to 818°C; 18 kb, 
509° to 697°C; and 20 kb, 509° to 649°C. 
No thermal effects were observed. 

The X-ray powder diffraction pattern 
(Table 15) of /?Cr 2 FeS 4 consists of 28 
reflections of varying quality. It is char- 
acterized by two reflections (d = 2.041 
and d= 1.993 A) in the region of 102 of 
the hexagonal NiAs type. /3Cr 2 FeS 4 may 
be either monoclinic or hexagonal with 
a structure related to that of NiAs. 



TABLE 15. d values of j3Cr 2 FeSi* 



No. 



d value 



1 

2 

3 

4 

5 

6 

7 

8 

9 

10 

11 

12 

13 

14 

15 

16 

17 

18 

19 

20 

21 

22 

23 

24 

25 

26 

27 

28 



6.695 

5.648 

5.205 

4.120 

3.711 

3.293 

2.944 

2.612 

2.532 

2.041 

1.993 

1.930 

1.7100 

1.589 

1.432 

1.418 

1.303 

1.182 

1.158 

1.100 

1.093 

1.085 

1.046 

0.9917 

0.9039 

0.8765 

0.8405 

05164 



v.w. 
m. 
m. 
st. 
m. 
m. 
m. 

St. 

w. 

v.st. 

m. 

m. 

St. 

m. 
w. 
w. 
m. 
w. 
w. 
m. 
m. 
m. 
st.t 

St. 

m. 

st.t 
st.t 
st.t 



* Ni-filtered CuKa radiation. 
t Broad line. 

Note: v.st., very strong; st., strong; m., 
medium; w., weak; v.w., very weak. 



MELTING RELATIONS IN THE Fe-RICH 

PORTION OF THE SYSTEM Fe-FeS 

AT 30 KB PRESSURE 

P.R. Brett* and P. M. Bell 

The Fe-rich portion of the system 
Fe-FeS has been thoroughly studied at 
1 atm, and the main features of the 
phase relations have been reviewed by 
Hansen and Anderko (1958). The sys- 
tem has now been studied at 30 kb since 
(1) sulfur may occur with iron as a 
constituent of the earth's core and mantle 
and (2) iron sulfides commonly occur 
in meteorites. If sulfur is a constituent, 
its effect on the melting point of iron 
may prove to be a significant geophysical 
parameter (Clark, 1963; Birch, 1964; 
Mason, 1966). 

Mixtures of Fe and FeS were loaded 
into small capsules of MgO and sub- 
jected to high-pressure, high-temperature 
conditions in the piston-cylinder, solid- 
media apparatus. The presence of liquid 
was determined on the basis of quench 
textures. Compositions were checked by 
the phase appearance-disappearance 
method, by electron probe, and by X-ray 
diffraction. The results of the probe 
analyses checked with each other to 

* U. S. Geological Survey. 



1600 


^ 


1 1 

I 


i I 1 

30 kb 


1 


1500 


— \ 


V 




— 


O 


T 


V* I 








<d 1400 




1 £ N 


Liquid 















3 




x 
















O 






\ 














g_ 1300 






\l 




£ 






l\ 




^ 1200 


— 


Fe+ Liquid 


\ 


i7 

/i 


1100 


"""■ 




^ / 


'Liquid 


1000 




1 I 


! 1 l \> i 


+ 
| FeS 



100 95 90 85 SO £ 75 l 70 65 



Fe 



Weight per cent Fe 



FeS 



Fig. 79. Temperature-composition section for 
the system Fe-FeS at 30 kb. 



GEOPHYSICAL LABORATORY 



199 



better than ±2 wt % of the amount 
present. 

A binary eutectic similar to that ob- 
served at 1 atm was discovered at 30 kb, 
and is shown in Fig. 79. The eutectic 
temperature at 30 kb is 990° ±5°C. This 
temperature is nearly identical with that 
for the 1 atm results (988°C). There is 
no possible confusion in the interpreta- 
tion of the present results, however, be- 



cause all experiments were quenched at 
30 kb pressure. The eutectic composition 
has moved from 31 wt % sulfur at 1 atm 
to 26.5 wt % sulfur at 30 kb. 

The results shown in Fig. 79 suggest 
that, even if only a small percentage of 
sulfur is present in the earth's core, the 
effect on the melting of iron may be 
significant. 



BIO GEOCHEMISTRY 



Reactions of the Organic Matter in 
a Recent Marine Sediment 

T. C. Hoering 

Some of the chemical reactions pro- 
duced by elevated temperatures in a 
short time in the organic matter in a 
Recent marine sediment from the San 
Nicholas Basin are the same as those 
that occur slowly at lower temperatures 
in nature. Thus the nature of organic 
geochemical processes can be investi- 
gated in the laboratory. A first report of 
our experiments was given last year 
(Year Book 66), and the experiments 
have been continued. Several new results 
are reported here. 

If a dried sediment is exposed to an 
excess of deuterium oxide, redried, and 
then heated to 225 °C for 1 week under 
an inert atmosphere, new hydrocarbons 
are produced with deuterium atoms sub- 
stituted for hydrogen. A number of such 
hydrocarbons have been separated by 
preparative gas-liquid chromatography, 
and the amount of deuterium substitu- 
tion has been determined by mass spec- 
trometry. Table 16 shows a result de- 
termined for n-C 2 5H 52 . Molecules with at 
least eight deuterium atoms are present. 
The results for a number of other normal 
and isoprenoid hydrocarbons are similar. 

The extent of the deuterium exchange 
of the normal fatty acids that exist in the 
sediment is shown in Table 16. Much less 
deuterium is present in them than in 
the newly formed alkanes. The results 



for the acids are most easily explained 
in terms of exchange of the hydrogens on 
the (3 carbons of the acid with water. 
Simple homogeneous exchange of water 
with the hydrogens on long, saturated 
alkyl groups is not likely. 

The multiple deuteration of the al- 
kanes is unexpected and indicates that 
the reaction mechanism of hydrocarbon 
production is complex. It is known that 
surface-catalyzed reductions of organic 
compounds in the presence of deuterated 

TABLE 16. Relative Concentrations of Deu- 
terated Species of n-Pentacosane and Hexa- 
decanoic Acid from Deuterated Tanner 
Basin Sediment Heated to 225 °C 
for 1 Week 



Normalized 


Species Abundance 


n-Pentacosane 




C25H52 


100 


C25H51D 


28 


U25H50D2 


16 


C25H.11JD3 


12 


C/25-H.4.slJl 


10 


C25-H.47D5 


7 


C25H46D8 


4 


C_/25-H.4bDt 


2 


C25H44D8 


1 


n-Hexadecanoic Acid 




Kyl(sHs2\J2 


85 


CluHsiDC^ 


100 


C1UH30D2O2 


37 


L/K1XI211D3O2 


5 


C1C1H2SD4O2 


1 



Note: The data from the mass spectra have 
been corrected for contributions of C 13 -contain- 
ing species. The concentration of deuterium- 
substituted compounds arising from the natural 
abundance of deuterium is negligible. 



200 



CAKNEGIE INSTITUTION 



substances often lead to extensive deu- 
terium replacement at positions away 
from the immediate reaction site. In- 
vestigations with the use of this tracer 
technique may give some information on 
the role of mineral surfaces during the 
diagenesis of organic matter. 

When the Recent marine sediment is 
heated to 225 °C for a few days in an 
inert atmosphere, porphyrins are gener- 
ated from the chlorophyll and chloro- 
phyll residues present in it. The mass 
spectrum of the demetallated etiopor- 
phyrins and deoxophylloerythroetiopor- 
phyrins generated by such thermal treat- 
ment is shown in Fig. 80. The porphyrins 
are a complex mixture of homologues of 
the two classes. The spectrum is similar 
to that reported for the prtroporphyrins 
in ancient sediments. 

The presence of a regular series of 



porphyrins with molecular weight 310 + 
14w, where n is an integer, indicates that 
alkyl-substituted etioporphyrins are 
present, since 310 is the molecular weight 
of the unsubstituted tetrapyrrole ring 
and 14 is the molecular weight of the 
added methylene substituents. Likewise, 
a series of porphyrins with molecular 
weights 308 + 14n must indicate a ho- 
mologous series of alkyl-substituted de- 
oxophylloerythroetioporphyrins with an 
isocyclic ring fused to the tetrapyrrole 
ring. 

Chlorophyll is the major pigment pres- 
ent in the system, and it is the most 
likely precursor of the porphyrins. If so, 
there must have been alkylation and 
transalkylation of peripheral positions of 
the tetrapyrrole ring during the heat- 
ing experiments to account for the mass 
spectra with molecular weight appear- 



ed 



Low Voltage Mass 


Spectrum of 












Porphyrins from San Nicolas Basin 


Sediment 225 °C for 3 days 


42 


6 

'"448 


462 
^50 


47 


6 

\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 
\ 


422 


/ 
/ 
/ 
/ 
/ 
/ 


"'•■ i 


164 


\ 
\ 
\ 
\ 


490 




43 

s 


V 








•478 


s 

X 
N 
S 
S 
S 
S 

V ^504 


i 




| 








1 


1_ 


1 



410 



420 



430 



440 450 



m /< 



460 

e 3 



470 



480 



490 



500 



510 



Fig. 80. Parent ion mass spectrum of porphyrins from heated San Nicolas Basin sediment. 
The spectrum was measured by West Coast Analytical Service Company, San Gabriel, California, 
on a Hitachi RMU-6D mass spectrometer. The porphyrins were volatilized into the ion source 
from a direct insertion probe at a temperature of 170°C. Electrons with an energy of 10 ev 
were used for ionization. Virtually no fragment peaks appeared in the mass spectrum. The 
dashed lines define the envelope of parent peaks due to deoxophylloerythroetioporphyrins. The 
dotted lines connect the parent peaks of etioporphyrins. 



GEOPHYSICAL LABORATORY 



201 



ing at increments of 14 units. Migration 
and transfer of alkyl groups may be an 
important organic reaction that occurs 
in sedimentary rocks, in addition to hy- 
drogenation, dehydrogenation, cycliza- 
tion, and hydrolysis. 

Our laboratory heating experiments 
indicate that a study of the thermal re- 
action of sedimentary rocks gives im- 
portant clues to the nature of the organic 
matter in them and to the organic reac- 



tions that occur over geological lengths of 
time. 

Branched-Chain Fatty Acids in 
Recent Sediments 

T. C. Hoering 

Normal fatty acids predominate in 
Recent sediments, but there is an ap- 
preciable concentration of acids with 
branched carbon chains. The mass spec- 



Fatty Acid Esters 
Tanner Basin Sediment 




100° 



Column Temperature - 
47min. 



320° 



Fig. 81. Gas-liquid chromatogram of esters of free fatty acids from Tanner Basin sediment. 
Ten-foot by ^-inch column with 5% Apiezon L. Helium flow 40 cc/minute. Temperature pro- 
grammed from 100° to 320°C at rate of 4°/minute. Identification of separated peaks by mass 
spectrometry. Peak symbols: (A) Methyl 3,7,11 trimethyl dodecanoate. (B) Methyl isotetra- 
decanoate and methyl anteisotetradecanoate. (C) Methyl n-tetradecanoate. (D) Methyl 4,8,12 
trimethyl tridecanoate. (E) Methyl isopentadecanoate and methyl anteisopentadecanoate. 
(F) Methyl -n-pentadecanoate. (G) Methyl 5,9,13 trimethyl tetradecanoate plus a compound, 
not a methyl ester of fatty acid. (H) Methyl isohexadecanoate, methyl anteisohexadecanoate, 
methyl palmitoleate, and nonmethyl ester. (/) Methyl n-hexadecanoate. (J) Methyl 2,6,10,14 
tetramethyl pentadecanoate. (K) Methyl isoheptadecanoate, methyl anteisoheptadecanoate 
and nonmethyl ester. (L) Methyl n-heptadecanoate. (M) Methyl 3,7,11,15 tetramethyl hexa- 
decanoate. (N) Complex unresolvable mixture of esters. (O) Methyl n-octadecanoate. (P) Methyl 
ester of cyclic acid. (Q) Methyl n-nonadecanoate. (P) Methyl ester of cyclic acid. (S) Methyl 
n-eicosanate. 



202 



CARNEGIE INSTITUTION 



trometric facilities put into operation 
during the last year make it possible to 
identify them. The molecular structure 
of pertinent species is as follows. 

Normal acid: CH 3 -(CH 2 ) m COOH 

H 

Iso acids: CH 3 -C-(CH 2 ) m COOH 



CH 3 



Anteiso acids : 



H 



CH 3 -CH 2 -C- (CH 2 ) p COOH 

CH 3 

H 

Isoprenoid acids: CH 3 -C-CH 2 -CH 2 - 



CH 3 
H 
CHQ-C-CH9-CH2 

I 
■ CH 3 



-COOH 



Figure 81 shows a gas-liquid chromato- 
gram of the methyl esters of fatty acids 
extracted from a surface sediment of the 
Tanner Basin in the continental shelf off 
southern California. The normal fatty 
acids form the major peaks, but between 
them, especially in the region of C14 to 
C19, are peaks due to the branched- 
chain acids. 

Individual pure esters were isolated 
by multiple, preparative gas-liquid 
chromatography with the use of polar 
and nonpolar substrates. The individual 
peaks were complex, some having as 
many as four components. The mass 
spectrum of each pure compound was 
measured; the components are identified 
in Fig. 81. 

The molecular structure of branched- 
chain fatty acid esters can be determined 
readily from the fragmentation pattern 
obtained by electron bombardment in the 
mass spectrometer ion source. The mass 
spectra of a number of such compounds 
have been correlated with their molecular 
structure. 

The presence of iso and anteiso acids 



is not surprising, for they are known 
constituents of the lipides of living or- 
ganisms, especially marine bacteria. The 
high concentration and variety of iso- 
prenoid acids are somewhat unexpected. 
Such acids are known in lipides but are 
usually considered to be relatively rare. 
Whether the isoprenoid acids are pri- 
mary products of the living biota or the 
products of diagenesis of chlorophyll 
and its degradation products is not 
known. Chlorophyll is a major organic 
constituent of the sediment and has a 
phytol side-chain with an isoprenoid 
structure. 

Fatty Alcohols in Sedimentary Rocks 
T. C. Hoering 

Living organisms synthesize many 
lipides with long, normal chains of car- 
bon atoms. Molecules of this structure 
are important constitutents of sedimen- 
tary rocks. Last year's report described 
the unexpected occurrence in Recent 
marine sediments of normal fatty acids 
having more than twenty carbon atoms. 
It was shown that the chemical oxida- 
tion of the non-acidic "polar molecule" 
fraction of an extract of the Recent 
marine sediment produced normal fatty 
acids. It has been shown that mild ther- 
mal treatment of unextracted Recent 
sediment produced a good yield of the 
normal alkanes n-C 22 H 46 and n-C 24 H 50 
and that under the same conditions Re- 
cent sediment will reduce a fatty alcohol 
to an alkane. 

These considerations suggest that fatty 
alcohols in the form of wax esters may be 
present in sedimentary rocks. Plant 
waxes are esters of fatty acids (I) and 
fatty alcohols (II) 

CH 3 (CH,) m COOH CH 3 (CH 2 ) n CH 2 OH 
(I) (ID 

where m and n are integers that typically 
run from 12 to 30. 

Waxes have been identified in lignite 
coals and peats. Although the ester link- 
age is hydrolyzed readily, waxes because 



GEOPHYSICAL LABORATORY 



203 



of their hydrophobic nature, can persist 
in relatively young sedimentary rocks. 
This report describes their occurrence 
in a Recent marine sediment from the 
Tanner Basin in the continental shelf 
off southern California and in the Green 
River shale of Colorado. 

The alcohols were isolated as their 
acetates from benzene-methanol extracts 
of the sediments. The standard methods 
of lipide chemistry were modified to 
take into account the interfering sub- 
stances found in rocks. A tracer of C 14 - 
tagged n-octadecanol was added to the 
initial extract. A yield of about one- 
fourth of the added radioactivity was ob- 
tained in the final acetate preparation. 

The acetates of the fatty alcohols were 
separated and analyzed by gas-liquid 
chromatography. Identification of the 
compounds is based on a comparison of 
retention times with those of pure stand- 
ard compounds and on the mass spectra 
of individual pure compounds separated 
by preparative gas-liquid chromatog- 
raphy. Table 17 gives a quantitative 
analysis of normal fatty alcohols in the 
two rocks. 

Fatty alcohols, presumably present in 
the form of wax esters, make a small but 

TABLE 17. Fatty Alcohols in Extracts of 
Sedimentary Rocks * 



Carbon No., 


Tanner 


Green River 


Alcohol 


Basin 


Shale 


16 


20 


2.0 


17 


0.8 


0.8 


18 


8.2 


1.5 


19 


0.5 


0.8 


20 


12 


1.2 


21 


0.5 


0.9 


22 


22 


2.4 


23 


2 


0.7 


24 


12 


0.5 


25 


1 


0.3 


26 


5 


0.1 


27 


0.7 




28 


4 




Totals 


88.7 


11.3 



* The values are in parts per million of the 
individual alcohols per gram of organic carbon 
in the sample. Values have been corrected for 
yield of octadecanol tracer added to initial 
extract. 



significant contribution to the inventory 
of normal alkyl groups found in sedi- 
mentary rocks. Branched-chain alcohols 
are present too, but their mass spectra are 
not well suited for exact structure de- 
termination. Analogy with structures of 
the branched-chain fatty acids given in 
another part of this report suggests that 
they may be isoprenoid alcohols, but firm 
identification will require preparation of 
derivatives more suitable for mass spec- 
trometry. There was no indication of the 
unsaturated alcohol phytol, which could 
be derived from the hydrolysis of chloro- 
phyll initially present in the sediment. 

The normal fatty acids in living or- 
ganisms and in Recent sediments have 
predominantly an even number of carbon 
atoms. In ancient sediments, the ratio of 
acids with an even number of carbon 
atoms to those with an odd number ap- 
proaches 1. On the other hand, the nor- 
mal alkanes of living organisms and 
Recent sediments have predominantly 
an odd number of carbon atoms, and in 
ancient sediments the ratio of even to 
odd approaches 1. Table 17 indicates 
that the same thing is happening to the 
normal fatty alcohols, which have pre- 
dominantly an even number of carbon 
atoms in cells and the Recent sediment 
from the Tanner Basin but a much higher 
relative abundance of odd-carbon-num- 
bered molecules in the older Green River 
shale. 

It may be speculated that the alcohols 
observed in the Tanner Basin sediment 
originated as waxes from land-based, 
higher plants, in which they form pro- 
tective coatings. Waxes, containing alco- 
hols in this molecular weight range, are 
not conspicuous constituents of marine 
organisms. 

Fichtelite Hydrocarbons in Fossil 
Wood 

T. C. Hoering 

The mineral called fichtelite is one of 
the few naturally occurring crystalline 
organic substances found in sedimentary 



204 



CARNEGIE INSTITUTION 



formations. It occurs in fossil conifer 
wood in the Fichtelgebirge area of south- 
ern Germany and is of Quaternary age. 
Fichtelite has been known for 140 years, 
but only in the past 30 years has the 
molecular structure of its major com- 
ponents been rigorously established. The 
saturated hydrocarbon 1.4a-dimethyl-7- 
isopropyl perhydrophenanthrene (given 
the chemical name fichtelite) (I) and the 
aromatic hydrocarbon l-methyl-7-iso- 
propyl phenanthrene (retene) (II) are 
present. Abietic acid (III) is one of the 
important resin acids of living conifers 
(Fig. 82). 

The similarities in the carbon skeletons 
of the first three compounds is obvious. 
Apparently, abietic acid is readily satu- 
rated with hydrogen and decarboxylated 
in the sedimentary environment to give 
fichtelite. It is well known that elemental 
sulfur converts both abietic acid and 
fichtelite to retene. 

In view of the rarity of pure crystalline 
organic substances in sediments, two 
specimens of fichtelite from the type lo- 
cality were obtained from the Division of 
Mineralogy, U. S. National Museum 
(Museum Nos. C-3897 and R-7270) . The 
specimens were tough, compact pieces of 
wood with no visible crystalline ma- 
terial. The wood was carved into small 
pieces and extracted in a Soxhlet ex- 
tractor with benzene-methanol. Sulfur 
was removed with metallic copper. The 
mahogany-colored extract was separated 
by silica-gel chromatograph. Sample 
R-7270 yielded 1.2 wt % of a colorless, 



saturated hydrocarbon oil and 4.3 wt % 
of a crystalline aromatic hydrocarbon. 
Gas-liquid chromatography, infrared 
spectrometry, and mass spectrometry 
showed that the aromatic hydrocarbon 
was pure retene. A gas-liquid chromato- 
graph of the saturated hydrocarbons con- 
tained three closely spaced peaks, in the 
ratio of 30:100:28. The major component 
was identified as fichtelite (I) . The other 
peaks are tentatively identified as iosine 
(IV) and a methyl-substituted iosine. 
Iosine is a known component of some 
lignite coals. 

There were at least two stages in the 
diagenesis of the fossil wood. One was in 
a highly reducing environment, produc- 
ing saturation of the resin acid to yield 
fichtelite and causing cyclization to give 
iosine. The second was in a sulfur-rich, 
dehydrogenating environment, producing 
retene. 

Studies of fichtelite present an oppor- 
tunity to investigate geochemical hy- 
drogenation and dehydrogenation of 
organic substances. Such processes are 
important in generating hydrocarbons 
in bulk sediments, but the sequence 
of events there is difficult to unravel. 
In fossil wood the problem is less com- 
plex; the starting compounds seem to 
be known, the products are simpler, 
and the types of organic reactions 
are more straightforward. A search 
was made for fichtelite and retene in ex- 
tracts of the Green River shale of Colo- 
rado, Chattanooga shale of Tennessee, 
tasmanite oil shale from Tasmania, and 




C COOH 

i n nr 

Fichtelite Retene Abietic 

Acid 



c c 



Iosene 



Fig. 82. Structures of hydrocarbons in fossil wood. 



GEOPHYSICAL LABORATORY 



205 



a lignite coal from North Dakota. The 
results were negative. Fichtelite seems to 
be limited to fossil wood. 

Laboratory Simulation of Amino-Acid 

dlagenesis in fossils 

P. E. Hare and R. M. Mitterer 

Changes in the amino-acid content 
of fossil shells with passage of time have 
been found to be consistent with deduc- 
tions based on the order of stability of 
free amino acids heated in dilute aqueous 
solutions, as shown first by Abelson 
(Year Book 53, pp. 97-101; and 1959) 
and more recently by Vallentyne (1964). 

Amino acids found stable in such heat- 
ing tests included alanine, glycine, glu- 
tamic acid, valine, and the leucines 
(Abelson, 1959) ; these are the amino 
acids that are primarily found in fossils 
older than Pleistocene. Amino acids un- 
stable to heating, such as serine and 
threonine, and therefore not expected in 
older fossils, have been reported (e.g. 
Degens, 1965, pp. 202-280) in older sedi- 
ments. Amino-acid stabilization by sur- 
rounding matrix has been suggested 
(Prashnowsky and Schidlowski, 1967) 
to explain the apparent anomaly of un- 
stable amino acids in ancient deposits. 

In addition, experiments at this lab- 
oratory on heating free amino acids in 
solutions indicate that pH can be a sig- 
nificant factor in the stability of some 
amino acids. Samples of a standard mix- 
ture of amino acids buffered at pH 2.0, 
4.6, and 9.4 (as measured at 25 °C) were 
heated in sealed tubes under nitrogen 
for 4 days at 140°C. Arginine disap- 
peared completely at pH 9.4 to form 
ornithine, whereas at pH 2.0 and 4.6 
no appreciable amount of arginine was 
converted to ornithine. Cystine was 
stable only at pH 2.0, virtually disap- 
pearing at both other pH values. The al- 
loisoleucine-to-isoleucine ratio observed 
after incubation was 0.05 at pH 2.0 but 
0.2 at pH 9.4. Aspartic acid was less 
stable at pH 4.6 than at pH 2.0 or 9.4. 
At pH 9.4 the serine-to-threonine ratio 



was 10 after heating. At pH 4.6 the 
serine-to-threonine ratio was 1.5, and at 
pH 2.0 it was reversed at 0.3. 

In view of this pH effect and the pos- 
sibility of a matrix effect on amino-acid 
stability, we have attempted to simulate 
more closely the diagenesis of amino 
acids in fossils by subjecting fragments 
of modern shells to incubation under a 
variety of temperatures and environ- 
'ments. A first series of experiments tested 
the role of water in the environment, al- 
lowing for the fact that in nature water 
is usually present. Modern shell frag- 
ments of Mercenaria were heated to 
160°C for 10 hours in varying amounts 
of water or water vapor. The sample 
heated dry showed little or no reaction, 
in contrast to the samples heated in the 
presence of water or water vapor. 

In subsequent experiments shell frag- 
ments of about 100 mg were heated in 
1 ml of water sealed in a tube under 
nitrogen. The changes observed upon 
heating Recent shell fragments in the 
presence of water are analogous to and 
duplicate those changes found in the 
natural series of fossils. With continued 
heating the unstable amino acids virtu- 
ally disappeared, the remaining amino 
acids racemized, and the mixture of 
amino acids approached that found in 
the oldest fossils examined. 

Evidence that we are dealing with the 
same mechanisms and changes that take 
place in the natural environment is found 
in the data from a series of Recent and 
fossil shells heated for 1 day at 185°C. 
The results are shown in Table 18. 

In the oldest shell, a sample from the 
Miocene St. Mary's formation, relatively 
few changes took place. Serine and 
threonine were already absent; the ratio 
of alloisoleucine to isoleucine was already 
the equilibrium ratio, so that heating 
for 1 day at 185 °C in the presence of 
water changed the pattern very little. 
Similar treatment of a specimen from 
Wailes Bluff, of Upper Pleistocene age, 
caused much more extensive changes, but 
the final mixture resembled that ob- 



206 



CARNEGIE INSTITUTION 



TABLE 18. Amino-Acid Composition for Inner Layer of Mercenaria Shells Before and After 
Heating to 185°C for 24 Hours, nM/g shell 





Recent 


Upper Pleistocene 
Unheated Heated 


Miocene 




Unheated 


Heated 


Unheated 


Heated 


Aspartic acid 


1960 


80 


865 


78 


35 


12 


Threonine 


1170 





306 











Serine 


1430 





247 











Glutamic acid 


1400 


575 


709 


863 


140 


140 


Proline 


1790 


570 


767 


375 


120 


120 


Glycine 


1500 


735 


572 


325 


104 


99 


Alanine 


1210 


850 


715 


523 


166 


165 


Cystine 


403 





26 











Valine 


740 


290 


423 


198 


127 


114 


Methionine 


225 


77 


85 


35 


10 


10 


Alloisoleucine 





100 


52 


69 


42 


42 


Isoleucine 


370 


80 


163 


55 


34 


34 


Leucine 


515 


190 


280 


115 


78 


74 


Tyrosine 


515 


134 


286 


94 


38 


38 


Phenylalanine 


354 


147 


202 


84 


52 


47 


Ornithine 





97 


52 


62 


9 


5 


Lysine 


1143 


155 


442 


81 


46 


20 


Histidine 


338 





65 











Arginine 


612 





241 











Ratio allo/iso 





1.25 


0.32 


1.25 


1.25 


1.25 



tained in the previous experiment. When 
a Recent shell was heated, profound 
changes were observed, the final pattern 
again approaching that obtained in the 
other two heated samples. 

To study the kinetics of the reactions 
in more detail, fragments of modern 
shells were heated at temperatures of 
165°, 140°, 125°, 105°, and 90°C for 
periods of time ranging from 1 hour to 
over 3 months. The shell fragments were 
heated in water in tubes sealed under 
nitrogen. The total amount of amino 
acids decreased with time of heating. 
Amino-acid ratios changed systemati- 
cally with time of heating, in a fashion 
analogous to that found in fossil shells. 

Comparison of data on free amino 
acids and shell fragments showed that 
free amino acids reacted more slowly. 
For example, in 4 days at 140°C, serine 
decreased only to 80% of its original 
value (at pH 9.4), whereas in the shell 
fragments it decreased to less than 10%. 
This apparent decrease in stability in 
the shell matrix seemed to be true for 
all the amino acids. 

Isoleucine racemizes to alloisoleucine 
by a first-order reaction. This reaction is 



reversible, and by heating the two sepa- 
rately we have shown that the isoleucine- 
to-alloisoleucine reaction is approxi- 
mately 25% faster than the reverse re- 
action, so that at equilibrium the ratio 
of alloisoleucine to isoleucine is about 
1.25. 

The Arrhenius equation used in Fig. 83 
relates the rate constant of a reaction 
to the absolute temperature. A plot of 
the time necessary for a given amount 
of isoleucine to diminish to 1/e of its 
initial value at a particular temperature 
is shown. It is seen that extrapolation 
of data based on the laboratory treat- 
ment checks well with data from a series 
of radiocarbon-dated fossil-shell ma- 
terial from areas with average tempera- 
tures of 23° to 24° C. These data were 
obtained from the middle shell layer of 
Mercenaria. There is evidence that other 
shell structures in Mercenaria react at 
somewhat different rates. 

Point A represents data from Foram- 
inifera tests from an Antarctic deep-sea 
core (core 15-16 of Goodell and Watkins, 
1968) supplied by Florida State Uni- 
versity. Sedimentation rates are based 
on recognized magnetic reversals. Agree- 



GEOPHYSICAL LABOEATOEY 



207 









IU 

I0 7 


1 i 1 


— 1 — 1 1 1 1 ■ . 


10 s 
I0 5 
I0 4 


A 

- x a s 


- 


000 




- 


100 


^v> 


- 


10 


^\ 


- 


. I 




\. 


iooo 


- 


^\ 


100 


- 


^v 


10 




^" - 


1 


1 t 1 


II II 1 



1 



11 



24 



90 105 125 140 165 



Temperature, °C 
Fig. 83. Arrhenius plot of time (1/e of reaction) as a function of temperature" 1 . 



ment with extrapolated values is good 
and gives us confidence that the reac- 
tions occurring in the laboratory treat- 
ment are similar to those occurring in 
fossils. 

Point B represents data from the 
Upper Pleistocene Wailes Bluff locality. 
Radiocarbon dating shows that this de- 
posit is older than 40,000 years. Amino- 
acid ratios suggest that it is probably 
around 60,000 years old if the tempera- 
ture has been around 11° to 12°C. 

The agreement between the values 
found in these Pleistocene samples and 
the Arrhenius plot suggests potential uses 
of amino-acid ratios in fossils. If the 
temperature is known precisely, an ap- 
proximate age can be calculated. Al- 
ternatively, effective incubation tem- 
perature can be inferred on well-dated 
material. Still another potential use is 
for stratigraphic correlation, which 
should be particularly useful in deep-sea 
cores, since the temperatures are uniform 
and constant. 

In addition to providing quantitative 
information on reaction rates, the heat- 



ing experiments afforded an opportunity 
to study contamination of the shell and 
loss of amino acids from it. Amino acids 
were found in the water surrounding the 
heated shell fragment. The concentra- 
tion of amino acids in the water varied 
directly with time of heating and tem- 
perature. With prolonged heating at 
higher temperatures, the amounts of 
amino acids outside the shell were ap- 
proximately equal to those within the 
shell. 

In another experiment, to test for the 
possibility of amino acids migrating into 
the shell, two shell fragments were placed 
in a 1-ml solution containing 10 ju,M 
of the nonprotein amino acid norleucine, 
which is not found in modern or fossil 
shells. The sealed tube containing the 
shell fragments was heated for 4% days 
at 165 °C under the usual conditions. 
After the tube was opened, one fragment 
was removed, thoroughly washed, and 
prepared for amino-acid analysis. The 
other fragment was washed and dried, 
and surface material was removed with 
a dental tool. The fragment was then 



208 



CARNEGIE INSTITUTION 



acid-washed to remove more of the outer 
surface and analyzed. In the water- 
washed sample, norleucine amounted to 
1.3% of the total amino acids. In the 
sample that was abraded and acid- 
washed, norleucine was about 0.5% of 
the total acids. 

The results suggest that young intact 
fossil shells are probably not contami- 
nated to a significant degree by amino 
acids in groundwater and sediment. On 
the other hand, older fossil specimens, 
having lost much of their amino acids 
and having been continuously subjected 
to contamination over a long period of 
time, might contain some adventitious 
amino acids. 

The results of this study indicate that 
it is possible to simulate the diagenetic 
reactions occurring in nature by simple 
heating of shell fragments in water. Fur- 
thermore, when the data from calcareous 
fossils were compared with those from 
free amino acids heated in aqueous solu- 
tions, it was found that there is indeed a 
matrix effect. The matrix effect resulted 
in faster reaction in the shell than for 
free amino acids, however, in contrast to 
a stabilizing matrix effect that has been 
postulated for sediments. 

Recent Amino Acids in the Gunflint 
Chert 

P. H. Abelson and P. E. Hare 

Organic chemicals surviving for bil- 
lions of years in sedimentary rocks are 
potential sources of evidence concerning 
the nature of ancient life. The quantities 
of original chemicals now present in 
many old rocks, however, are not much 
larger than those ubiquitously present 
as adventitious contaminants. Conse- 
quently, in organic geochemistry there 
is a substantial possibility that many 
interesting but illusory discoveries will 
be made. Rarely can a second investiga- 
tor obtain specimens identical with those 
on which the original work was per- 
formed, and exact duplication of earlier 
work is usually not feasible. Moreover, 



once a fictive result appears in the geo- 
chemical literature, the error is difficult 
to erase. 

Work performed this year in examin- 
ing the possible amino-acid content of 
the Gunflint chert illustrates an approach 
that often might be usefully employed. 
In the 1.9-billion-year-old Gunflint iron 
formation near the town of Schreiber, 
Ontario, Tyler and Barghoorn (1954) 
discovered a rich fauna of microfossils 
embedded in a flinty chert. Thin sections 
of this chert have been examined by a 
number of paleontologists, and there is 
general agreement that the forms noted 
are of biological origin. 

Recently, Schopf, Kvenvolden, and 
Barghoorn (1968) reported finding amino 
acids in the chert that they believe have 
remained there 1.9 billion years. We 
questioned the finding, however, since 
some of the amino acids reported as 
present, notably serine and threonine, 
are not very stable chemically. Conse- 
quently, we have conducted a series of 
tests, which show that although speci- 
mens of the Gunflint chert examined 
by us contained small amounts of recent 
amino acids, they did not contain sig- 
nificant amounts of ancient amino acids. 

Since it could be argued that the chert 
matrix might have a stabilizing effect 
on amino acids, we have conducted ther- 
mal tests on the chert similar to those 
employed with shells and described pre- 
viously in this report. Samples of chert 
thought to contain ancient amino acids 
were heated at 165°C for 1 to 3 days. 
If the chert matrix confers special sta- 
bility on serine, that amino acid and 
others should be affected little if at all 
by the treatment. 

The specimens employed came from 
the type locality at Schreiber. Chemi- 
cally they were about 99% Si0 2 . They 
also contained organic carbon, pyrite, 
and CaC0 3 . The setting and the appear- 
ance of the specimens testify to the pos- 
sibility of contamination. The type lo- 
cality is on the shore line of Lake 
Superior, and lichens abound on the sur- 



GEOPHYSICAL LABORATORY 



209 



face of the chert. When a cobble is bi- 
sected and etched much structure and 
some cracks may be seen. Schopf, Kven- 
volden, and Barghoorn employed pro- 
cedures well designed to guard against 
adventitious contamination. However, 
the amounts of amino acids that they 
found are tiny in comparison with the 
amounts present in the environment and 
available as possible contaminants. 

Method 

Special efforts were made to avoid 
contamination, and all reagents were 
tested repeatedly. Glassware was rou- 
tinely heated to 500 °C before use to 
destroy possible contaminants. Water 
was deionized, filtered through 0.2-fi 
membrane niters, and distilled. We have 
found that distillation without nitration 
does not eliminate dust contamination. 
HC1 was redistilled and yielded a very 
low level of amino acids. HF was re- 
distilled in a platinum still. 

Cobbles of Gunflint chert were treated 
with boiling 6 N HC1 and chromic acid 
to remove surface contamination of 
amino acids. HF was used to remove the 
outer 10% of the sample before size re- 
duction. 

An aliquot (a) of powdered Gunflint 
chert was compared to other aliquots 
treated as follows: One aliquot of pow- 
dered sample (aliquot b) was heated to 
400 °C for 2 hours to destroy any amino 
acids present. Analysis of this sample 
provides a measure of reagent and lab- 
oratory contamination. Another aliquot 
(c) was heated in the presence of water 
at 165 °C for 3 days. To check recovery 
of amino acids in the procedure a stan- 
dard amino-acid mixture was added to 
still another aliquot (d) , which had pre- 
viously been heated to 400 °C to destroy 
the contained amino acids. 

After the above pretreatments, all 
aliquots were extracted in Soxhlet ex- 
tractors with water for 1 day, followed 
by 6 N HC1 for 1 day. The residue from 
this treatment was reacted with HF to 



volatilize off Si0 2 . Later HC1 was added 
to decompose resistant materials. 

Desalting was accomplished by pre- 
cipitating the interfering ions. HF was 
used to remove calcium. Iron was re- 
moved by adding NaOH to make the 
solution alkaline, thus precipitating iron 
hydroxides. Sodium ion was removed 
from a saturated HC1 solution made by 
passing in HC1 gas in the cold. The su- 
pernatant solution was dried in vacuo. 

Results 

Water extraction of the chert alone 
was not very effective. When standard 
amino-acid mixtures were added to 
finely ground chert, recovery of the hy- 
drophilic acids was fair (~50%) and 
that of the more hydrophobic amino 
acids was poor (~20%). Recovery with 
HC1 extraction, however, was excellent 
(90 to 100%). Recoveries of added stan- 
dard amino acids were also poor follow- 
ing the treatment of chert with HF. The 
residue from this treatment included or- 
ganic matter and pyrite. Even when this 
residue was cautiously taken to dryness, 
irreversible loss of hydrophobic amino 
acids was noted. 

The use of tracer amino acids, includ- 
ing norleucine, gave us assurance that 
we were isolating most of the amino 
acids present in the samples of chert. 
Nevertheless, the total amounts isolated 
from unheated chert by H 2 0+HC1 ex- 
traction were less than those reported 
by Schopf et al. For example, they found 
the following (expressed in 10^ 9 M/g 
chert) : glycine, 11.6; alanine, 2.8; serine, 
1.5. In one of our cobbles we found, 
correspondingly, glycine, 1.2; alanine, 
0.79; and serine, 0.80. In another cobble 
we found glycine, 1.08; alanine, 0.48; 
and serine, 0.47. Other specimens yielded 
even smaller amounts. The amino-acid 
content of one chert specimen was in- 
distinguishable from the reagent back- 
ground, which for that particular experi- 
ment was equivalent to 10 -11 M/g chert. 

In contrast, HC1 treatment of the 



210 



CARNEGIE INSTITUTION 



outer surface of the chert yielded sub- 
stantial amounts of amino acids. From 
the surface of a 400-g cobble was iso- 
lated glycine, 1030 nM; alanine, 270; 
and serine, 630; and of course, the other 
usual amino acids. 

Treatment of ground chert with water 
and HC1 extracted essentially all the 
amino acids associated with the chert 
while leaving behind more than 99% of 
the weight of the original sample. The 
amounts of amino acids found after HF 
treatment to volatilize Si0 2 were about 
those of our reagent blanks. For example 
in one instance we found glycine, 0.038 
nM/g chert; alanine, 0.014; and serine, 
0.017; and this was indistinguishable 
from the blank. 

It seems significant that the water and 
HC1 treatments collected most, if not all, 
of the amino acid present in the chert 
while leaving behind the main body of 
the chert. Either the main part of the 
chert is readily pervious to extraction or 
it did not contain amino acids. If the 
chert is so permeable that amino acids 
can be completely extracted in 1 day, 
then amino acids from modern lichens 
would probably contaminate the chert 
after a few thousand years. 

The heating experiments showed that 
amino acids associated with the chert 
were degraded in much the same way 
as amino acids in fossil shells. Exposure 
of chert to 165 °C for 3 days resulted in 
a marked degradation of serine and 



threonine and in the conversion of half 
of the isoleucine to alloisoleucine. In 
three unheated specimens the ratios of 
serine to glycine found were 0.90, 0.48, 
and 0.41. After 1 day of heating at 165 °C 
the ratios were 0.38, 0.20, and 0.20. 

Our observations on the stereo-isomers 
alloisoleucine and isoleucine seem par- 
ticularly significant. As pointed out else- 
where in this report, the biologically 
produced isoleucine, in nature, is par- 
tially converted into alloisoleucine, and 
after a few million years an equilibrium 
is established in which 1.25 parts of 
alloisoleucine and 1.0 part isoleucine co- 
exist. We have examined the amino acids 
in eight different specimens of Gunflint 
chert and have found isoleucine in all of 
them. However, in no unheated specimen 
have we found any alloisoleucine. In 
specimens heated to 165°C for 1 day or 
more, alloisoleucine is readily detectable, 
showing that the chert matrix does not 
interfere with isomerization. Thus the 
absence of alloisoleucine in the chert is a 
strong indication that the isoleucine pres- 
ent is of recent origin. 

In another experiment, designed to 
examine the antiquity of the amino acid, 
L-amino acid oxidase was used to de- 
termine the optical configuration of the 
amino acids in the chert. All the amino 
acids that reacted with this enzyme 
were of the L configuration. This result is 
also suggestive of amino-acid material 
of recent origin. 



COMPUTER REDUCTION OF ELECTRON-PROBE DATA 



F. R. Boyd, L. W. Finger, and F. Chayes 



Reducing X-ray intensity measure- 
ments determined with an electron probe 
to estimates of chemical composition re- 
quires performance of a large number 
of simple mathematical operations. In 
addition to averaging groups of counts 
and forming intensity ratios, it is neces- 
sary to correct the observed intensities 
for instrumental factors and matrix ef- 
fects. A complete correction procedure 



includes corrections for drift, back- 
ground, deadtime, absorption, fluores- 
cence, and atomic number effects. Pre- 
liminary estimates of the composition are 
usually the basis for calculating matrix 
effects, and the correction elements are 
refined by iteration. Moreover, the in- 
vestigator usually wishes to calculate 
statistical parameters to evaluate homo- 
geneity, and if the analysis includes trace 



GEOPHYSICAL LABORATORY 



211 



elements, calculations of significance of 
small signals may also be required. To 
make such a complete set of calcula- 
tions with a desk calculator for an analy- 
sis of a natural mineral containing eight 
to twelve elements would require at least 
weeks of work. Fortunately, the correc- 
tion calculations can be programmed for 
a large computer, and this can be done 
in a way that leaves the investigator with 
considerable flexibility in his approach 
to an analytical problem. Other electron- 
probe laboratories, including those at the 
Scripps Institution of Oceanography, the 
U. S. Geological Survey at Menlo Park 
and the Goddard Space Flight Center, 
have developed computer programs for 
reducing probe data, and we have bene- 
fited from their experience in writing 
the programs described in this report. 

The convenience with which a pro- 
gram can be used depends on the number 
of control cards that must be prepared 
for each analysis. These can be greatly 
reduced by providing for automatic 
printout of code numbers that identify 
the elements being analyzed, the stan- 
dards used, and the type of measurement 
being made. Table 19 shows a line of out- 



put from our electron probe. Ten con- 
trol digits are followed by the counting 
interval, the counts obtained from the 
three channels, and the specimen current. 
These control digits are conveniently 
set by switches mounted on the data 
translator. The first digit specifies the 
type of measurement, i.e., whether a 
count is on peak or on background and 
whether it is for a standard or the un- 
known. The remaining digits, which 
specify the elements being analyzed and 
identify the standards, can be set at the 
beginning of an analysis and need not 
be changed. Output is recorded simul- 
taneously on an automatic typewriter 
and a card punch. 

We have found it convenient to set up 
the data-reduction procedure in two pro- 
grams. The first, program CONE, aver- 
ages counts, evaluates statistical param- 
eters, and makes the instrumental cor- 
rections. The output of CONE is the 
initial approximation to the composi- 
tion. A second program, ABFAN, calcu- 
lates the matrix corrections and, if re- 
quired, corrects the initial composition 
by iteration. These programs are written 



TABLE 19. Line of Output from the Electron Probe 




ata categories (see below) 
tandard numbers (spectrometers 1-3) 
tomic numbers (spectrometers 1-3) 
Time (050.00 sec) 



159 14 20 12 05000 



10028 1 



61299 1 



26255 1 



0571 



Observed counts with scale 
factors (1:1,2:10,3:100) 



Specimen current 
(0.0571 pa) 



Data Categories 

1. Standard peak (1) 

2. Standard peak (2) 

3. Standard background 

4. Unknown peak 

5. Unknown background (1) 

6. Unknown background (2) 

Any combination of categories that includes 1, 2, and 4 is permissible. 



212 



CARNEGIE INSTITUTION 



in Fortran IV for a Univac 1108 com- 
puter. 

A sample sheet of output from pro- 
gram CONE is shown in Table 20. CONE 
can be used to calculate simple intensity 
ratios relative to any standard, but in 
the event that data on the compositions 
of the standards used are in the pro- 
gram's block data storage, the output will 
include analyses in weight percentages 
of the elements and oxides. Standard 
compositions are stored in a two-dimen- 
sional array for elements up to atomic 
number 28 (Ni) and with up to nine 
standards for each element. The drift 
calculated from counts on the standards 
taken before and after analysis of the 
unknown is also printed. Printout of 
the peak-to-background ratio provides 
a control on whether the background has 
been adequately evaluated. 

The homogeneity index is the ratio of 
the observed standard deviation to the 
standard deviation predicted from count- 
ing statistics, and CONE calculates these 
indices for each group of counts before 
reducing their average to counts per 
second. A sample is considered inho- 
mogeneous if its index is greater than 3. 
In the example shown, which is an analy- 
sis of a diopside inclusion from diamond, 



it can be seen that Al and Ti have ex- 
tremely inhomogeneous distributions. The 
index loses its sensitivity, however, at 
low peak-to-background ratios. Stan- 
dards will normally be homogeneous, and 
printout of homogeneity indices for them 
is intended only to inform the user when 
counts are erratic for instrumental rea- 
sons. 

Program CONE permits much flexibil- 
ity with regard to the measurement of 
background. It can be determined on the 
standard, on the unknown, on both, or 
not at all. Background may be measured 
on the unknown both before and after 
analysis, and if this is done the printout 
will contain a calculation of background 
drift. If in a given analysis, background 
is measured on either the standard or 
the unknown, but not both, calculations 
are made on the assumption that the 
backgrounds of standard and unknown 
are equal. In the analysis of trace con- 
stituents, we have adopted the practice 
of scanning a peak and forming a ratio 
between background measured at some 
arbitrary wavelength (usually 0.050 A 
above the peak) and the true background 
beneath the peak. During analysis the 
background is measured at this arbitrary 
wavelength and then corrected by the 



TABLE 20. Example of CONE Printout 



Weight % 



Weight % 



Intensity 
Ratio U/S 



Standard Peak to 
Drift, % Background 



AI2O3 

K 2 

Ti0 2 



K 



0.93 
0.00 
0.43 



Al 
K 
Ti 



Al 
K 
Ti 



0.49 
0.00 
0.26 



0.1852-00 
0.1395-03 
0.2140-00 



1.4 
2.6 
2.3 



Homogeneity Indices 



Standard (1) 
0.3 
0.9 
0.4 



Standard (2) 
0.8 
0.7 
12 



Unknown 
73.7 
2.3 
29.3 



Backgrounds 



Unknown cps Drift, % 
Al 23.12 0.0 

K 20.12 0.0 

Ti 75.55 0.0 



Standard cps 
23.12 
20.02 
75.55 



8.53 
1.03 
2.89 



Concentration, 

ppm 

17 



Student's 

t 

1.935 



Number of 

Counts 

10 



Detectability 
Limit, ppm 
22 



GEOPHYSICAL LABORATORY 



213 



predetermined ratio. If such background 
factors are entered on a control card, 
program CONE will apply them. 

Calculation of Student's t with a de- 
tectability limit provides a convenient 
test of the significance of peaks that are 
very small in relation to background. 
If an observed concentration is less than 
0.1%, program CONE prints the result 
in ppm and a detectability limit for a 
confidence level of 99%. The detectabil- 
ity limit (Ziebold, 1966) is a function of 
the variance of a particular analysis, the 
number of counts taken, and the chosen 
confidence level. In the example shown 
in Table 20 the detectability limit for 
potassium is slightly greater than the 
concentration found. Hence, the presence 
of potassium in this sample has not been 
established at a confidence level of 99%. 

Theoretical and empirical development 
of matrix corrections have advanced to 
the point where they can be used very 
successfully for silicates, provided the 
absorption corrections are kept below 
about 10 relative %. The absorption cor- 
rection of Philibert (1963) with the over- 
voltage modification of Duncumb and 
Shields (1963) has been used in program 
ABFAN. K. F. J. Heinrich (personal 
communication) has recently suggested a 
change in the constants for the Leonard 
coefficient (o-).* Heinrich's constants are 
used in ABFAN, and their use makes a 
substantial improvement in the ana- 
lytical totals. ABFAN corrects for fluo- 
rescence of K lines by K lines with 

*With Heinrich's constants o- = 4.25Xl0 5 / 



Reed's (1965) modification of the 
Castaing equation. Reed's simplifications 
for calculating the term (r A — l)/r A and 
for the fluorescence yield, W k (B) , have 
been followed. 

Atomic number effects are a combina- 
tion of differences between standard and 
unknown in electron backscatter and 
stopping power. A recent treatment of 
these effects by Duncumb and Reed (un- 
published manuscript) is promising, at 
least for heavy elements in a light matrix. 
Their treatment is used in ABFAN, in- 
cluding an equation adapted for com- 
puter calculation of the backscatter 
term made available by Duncumb. 

Table 21 shows analyses that were 
made to test the Duncumb and Reed 
atomic number corrections. The ma- 
terials analysed are themselves stan- 
dards, either synthetic compositions or 
natural minerals whose compositions 
have been well established by wet-chemi- 
cal analyses. These materials have been 
analysed with the probe relative to pure 
elements, and the individual corrections 
as well as the results are shown. Atomic 
number effects are large, with stopping- 
power corrections in the range 12-24 
relative %. Results for Fe and Cr agree 
with expected compositions to ±2 rela- 
tive %, but the Ti analyses appear to be 
overcorrected by 3 to 5 relative % . These 
results are encouraging, particularly in- 
asmuch as synthetic standards contain- 
ing the transition elements are very 
difficult to synthesize. 

ABFAN stores Heinrich's (1966) con- 
stants for calculation of mass absorp- 



TABLE 21. Analyses of Various Materials, with Pure Elements as Standards 











Atomic Number Effects 


Weight % Element 




Stopping 




Corrected 




Material 




Absorption 


Fluorescence 


Power 


Backscatter 


Probe 


Expected 


Analyzed 


Element 


Correction 


Correction 


Correction 


Correction 


Result 


Composition 


Chromite 


Fe 


1.055 


1.000 


1.131 


0.965 


17.0 


17.17 


Glass 


Fe 


1.021 


1.000 


1.189 


0.957 


13.7 


13.99 


Olivine 


Fe 


0.997 


1.000 


1.235 


0.944 


5.76 


5.66 


Chromite 


Cr 


1.001 


0.967 


1.125 


0.966 


30.7 


30.45 


Glass 


Cr 


1.033 


1.000 


1.222 


0.949 


1.28 


1.3 


Glass 


Ti 


1.052 


0.987 


1.125 


0.975 


11.4 


10.89 


Glass 


Ti 


1.051 


1.000 


1.198 


0.965 


1.24 


1.20 



214 



CARNEGIE INSTITUTION 



tion coefficients. Other constants, such as 
wavelengths for Ka lines and K edges, 
atomic weights, etc., are also stored. It 
would be possible to compute matrix ele- 
ments for the standards as well as the 
unknown in each problem. But our 
analytical method uses relatively few 
standards, and it is simpler to store them 
in a data block. The matrix elements are 
dependent on kv, and handling them in 
this way requires the user to make analy- 
ses at the same kv values as those used 
in calculation of the standard constants. 
Normally this is not a severe restriction, 
but in the event a user wishes to employ 
some kv-standard combination different 
from those used in calculation of the 
values in the program, the required 
matrix constants can be read in with 
the data cards. 

Program ABFAN can be used in two 
ways. It will calculate and print out 
matrix elements for an input composi- 
tion. Alternatively it will correct an in- 
put composition, refining the matrix cor- 
rections by iteration with reference to 
standards. Table 22 shows an example 
of the first application. Use of ABFAN 
in this way provides constants for stan- 



dards, which are then stored in data 
statements. It may also be used, in 
setting up an analytical program, to run 
trial compositions so that standards can 
be selected that will minimize absorption 
corrections. 

Table 23 shows an example of a prob- 
lem in which ABFAN is used to correct 
an input composition by iteration. 
ABFAN prints out the number of itera- 
tions required to produce a composition 
that differs from the composition on the 
previous cycle by less than 0.1 relative °/o 
of any of the elements present. Normally 
three or four iterations are sufficient and 
more than five have never been required. 

An array of values of Reed's gamma 
function (Reed, 1965) is also printed 
(Table 22), a feature useful in locating 
fluorescence effects. In the example 
shown, only the fluorescence of KKa by 
CaKa is significant. 

Oxygen contents of silicates and other 
oxides cannot readily be determined with 
an electron probe. ABFAN estimates the 
oxygen content of an unknown by sum- 
ming the weight fractions of the ana- 
lyzed elements, subtracting the sum 



TABLE 22. Output of Program ABFAN for E-ll Diopside, No Iteration 



Weight % 



Kv 



F(x) 



Fluorescence 
Factor * 



Stopping Power 
Factor 



Backscatter 
Factor 



Ca 


13.95 


15 


0.9194 


0.9985 


2.0675 


0.9303 


Mg 


10.43 


15 


0.5585 


0.9953 


2.0675 


0.9303 


Al 


1.60 


15 


0.6063 


0.9913 


2.0675 


0.9303 


Fe 


2.02 


20 


0.9533 


1.0000 


2.1988 


0.9264 


Cr 


0.99 


20 


0.9268 


0.9950 


2.1988 


0.9264 


Si 


25.33 


15 


0.7012 


0.9988 


2.0675 


0.9303 


Na 


1.34 


15 


0.4246 


0.9961 


2.0675 


0.9303 


Ti 


0.25 


15 


0.9275 


0.9969 


2.0675 


0.9303 


Mn 


0.07 


20 


0.9428 


1.0000 


2.1988 


0.9264 


K 


0.01 


15 


0.8917 


0.9794 


2.0675 


0.9303 



Note: Mean atomic weight = 21.70 ; mean atomic no. = 10.78; H = 0.224. 

Values of Reed's Gamma Function * 



Ca 



Fe 



Cr 



Ca 


0.0000 


0.0000 


0.0000 


0.0006 


0.0006 


0.0000 


0.0000 


0.0002 


0.0000 


0.0000 


0.0000 


Mg 


0.0003 


0.0000 


0.0003 


0.0000 


0.0000 


0.0041 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


Al 


0.0005 


0.0000 


0.0000 


0.0000 


0.0000 


0.0082 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


Fpi 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


Cr 


0.0000 


0.0000 


0.0000 


0.0050 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


Si 


0.0011 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


Na 


0.0001 


0.0017 


0.0002 


0.0000 


0.0000 


0.0019 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


Ti 


0.0000 


0.0000 


0.0000 


0.0016 


0.0014 


0.0000 


0.0000 


0.0000 


0.0001 


0.0000 


0.0000 


Mn 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0.0000 


0000 


0.0000 


0.0000 


0.0000 


K 


0.0202 


0.0000 


0.0000 


0.0004 


0.0003 


0.0000 


0.0000 


0.0O02 


0.0000 


0.0000 


0.0000 


Note : Row 


headings 


are analyzed elements; column head 


ings are : 


fluorescing 


elements. 







* Fluorescence factor = 1/(1 + 71 + 72 + 73 + 



7«). 



GEOPHYSICAL LABORATORY 



215 



TABLE 23. Output of Program ABFAN for E-ll Diopside Iterated with Reference to Standards 













Stopping 














Absorption 


Fluorescence 


Backscatter 


Power 


Initial 


Corrected 








Kv 


Correction 


Correction 


Correction 


Correction 


Composition 


Composition 






Ca 


15 


1.0024 


0.9985 


0.9997 


0.9970 


13.98 


13.95 


CaO 


19.51 


Mg 


15 


1.0328 


1.0001 


0.9997 


0.9970 


10.13 


10.43 


MgO 


17.29 


Al 


15 


0.9397 


0.9990 


1.0060 


0.9809 


1.73 


1.60 


AW B 


3.03 


Fe 


20 


1.0184 


1.0000 


1.0049 


0.9898 


1.99 


2.02 


FeO 


2.59 


Cr 


20 


1.0013 


0.9950 


1.0014 


0.9948 


1.00 


0.99 


Cr 2 3 


1.45 


Si 


15 


1.0241 


1.0004 


0.9997 


0.9970 


24.81 


25.33 


Si0 2 


54.18 


Na 


15 


0.9848 


0.9987 


1.0017 


0.9936 


1.37 


1.34 


Na 2 


1.81 


Ti 


15 


0.9931 


0.9969 


0.9993 


0.9988 


0.25 


0.25 


Ti0 2 


0.41 


Mn 


20 


1.0200 


1.0000 


1.0053 


0.9916 


0.07 


0.07 


MnO 


0.09 


K 


15 

tals 


0.9950 


0.9794 


1.0016 


0.9971 


0.01 


0.01 


K 2 


0.01 


To 


55.34 


55.99 


100.38 



Note : Number of iterations = 3. 

from 1.0, and assigning the difference 
to oxygen. ABFAN will correct for the 
presence of any two elements by differ- 
ence if their atomic numbers and atomic 
proportions are read in with the data 
cards. This feature is most useful in 



analysis of carbonates and nitrates, for 
the concentrations of C and N cannot 
be determined with an accuracy com- 
parable to that possible with heavier ele- 
ments. It can also be used for silicates 
if the ratio of Si to O is well established. 



CRYSTALLOGRAPHY 



A New Hexagonal Form op Carbon 

FROM THE RlES CRATER 

A. El Goresy and G. Donnay 

The mineralogy and textural relations 
of shocked graphite gneisses from the 
Ries Crater were studied in polished sec- 
tions. Several graphite grains contain a 
new phase. X-ray diffraction and elec- 
tron-microprobe investigations of these 
grains reveal the presence of a new hexa- 
gonal form of carbon, which occurs in 
relatively thin lamellae (3 to 15 /* wide) 
alternating with graphite and perpen- 
dicular to the 0001 face of graphite. The 
new phase is much more strongly reflect- 
ing (R^4Q%) than hexagonal graphite 
and shows no sign of bireflection. It is 
slightly harder than graphite, and its 
reflection color is metallic gray to white. 
No anisotropism was observed, probably 
on account of the extremely small grain 
size. Electron-microprobe analysis of the 
new phase indicated carbon as the only 
major element; small amounts of CI and 
Si (<0.5%) were also detected. They 



appear to be due to contamination by the 
glass matrix and the araldite mounting 
material in which the sample is em- 
bedded. 

A grain containing more than 60% of 
the new phase, as estimated under the 
reflecting microscope, was carefully iso- 
lated mechanically from the polished 
section and mounted without crushing in 
a Debye-Scherrer camera (114.6 mm 
radius). The patterns made with FeKa 
and CuKa radiations (AFe = 1.9373, 
ACu = 1.51418 A), consisted of 22 sharp 
lines (Table 24) in addition to the eleven 
strongest lines of hexagonal graphite. 
There is no evidence of preferred orienta- 
tion. No rhombohedral graphite lines 
were observed. 

A primitive hexagonal cell with cell 
dimensions a = 8.948 ±0.009, c = 14.078 ± 
0.017 A, permits the indexing of all ob- 
served lines (Table 24). The least- 
squares program LCLSQ of C. W. Burn- 
ham {Year Book 61, pp. 131-135) was 
used to obtain the above refined cell di- 
mensions. Note that although no 000Z 



216 



CARNEGIE INSTITUTION 



TABLE 24. X-ray Powder Pattern of Chaoite 



No. 


Clots. 


dc.le. 


I 


hk-l 


1 


4.47 


4.465 


v.v.st. 


11-0 


2 


4.26 


4.256 


v.v.st. 


11-1 


3 


4.12 


4.014 


v.st. 


10-3* 


4 


3.71 


3.728 


m. 


20-1 


5 


3.22 


3.206 


m. 


10-4 


6 


3.03 


2.985 


St. 


20-3 


7 


2.94 


2.923 


w. 


21-0 


8 


2.55 


2.536 


St. 


30-1 


9 


2.46 


2.482 


m. 


21-3 


10 


2.28 


2.277 


St. 


20-5 


11 


2.24 


2.232 


m. 


22-0 


12 


2.10 


2.080 


m. 


30-4 


13 


1.983 


2.007 


w. 


20-6 


14 


1.910 


1.915 


w. 


40-1 


15 


1.496 


1.495 


w. 


22-7 


16 


1.370 


1.370 


w. 


41-6,22-8 


17 


1.289 


1.289 


w. 


60-0,50-6 


18 


1.26 


1.257 


w. 


33-6 


19 


1.197 


1.197 


m. 


33-7,52-3 


20 


1.184 


1.183 


m. 


42-7 


21 


1.080 


1.0815 


w. 


41-10 


22 


0.8642 


0.8643 


w. 


41-14 



* Possibly the strongest line of an impurity, 
since the disagreement of observed and cal- 
culated d values, 2.7%, is considerably greater 
than for the next poorest reflection, 20-3, where 
it is 1.6%. 

Note: v.v.st., very very strong; v.st., very 
strong; st., strong; m., medium; w., weak. 

reflections are observed, so that a layer 
structure is excluded, the nature of the 
principal axis of symmetry is undeter- 
mined. If the holohedral space group 
P6/mmm, which has a 24-fold general 
position, is assumed, a reasonable cal- 
culated density of 3.43 g/cm 3 is obtained 
by filling the general position seven 
times, leading to 168 atoms per cell. Not 
enough of the pure, new phase has yet 
been isolated for density measurement. 
The name chaoite is proposed for this 
mineral to honor Dr. Edward C. T. Chao, 
of the U. S. Geological Survey, who con- 
tributed greatly to the recognition and 
discovery of meteorite craters and the 
shock-induced metamorphic processes 
that accompany meteoritic impacts. 

Determination of Cation Distribu- 
tions by Least-Squares Refinement 
of Single-Crystal X-Ray Data 

L. W. Finger 

The ordering of cations in the crystal 
structure of a material belonging to a 



solid solution series greatly affects its 
thermochemical properties and its sta- 
bility. Substitutions, such as of aluminum 
for silicon, in tetrahedra in which the 
variation in site chemistry changes the 
average cation-anion bond distance have 
been successfully rationalized by many 
investigators in terms of ordering param- 
eters. For many octahedrally coordinated 
cations there is no reliable correlation 
between site chemistry and bond dis- 
tance, and the difference in X-ray scat- 
tering power between cations has been 
used to allocate the different elements 
among the possible sites. Ghose and 
Hellner (1959) used this difference im- 
plicitly by noting that if the Fe/Mg 
ratio for a site in their model of grunerite 
were wrong, the least-squares treatment 
would change the scattering power in 
each site by adjusting the temperature 
factor. Their procedure assumes that the 
temperature factors for all the octahedral 
sites are equal, but one of the amphibole 
sites in fact has a much different en- 
vironment than the others, and will 
therefore probably have a different tem- 
perature factor. Accordingly, a solution 
not involving this assumption was used 
by Fischer (1966) in refining the cation 
distributions in a cummingtonite by a 
direct least-squares treatment, applying 
no physical or chemical constraints to 
the solutions. Occupancy constraints are 
developed here, and their application is 
shown. 

The first constraint is purely physical 
and limits the maximum occupancy of 
the site. If we define a mn as the fractional 
occupancy of the mth site by the nth 
chemical species, then the equation 

2a m „^l (1) 

n 

must be satisfied. If there are no vacan- 
cies, then the equality in (1) holds. An- 
other requirement of the solution for 
the a mn 's is that the site chemistry must 
agree with the bulk chemistry. Mathe- 
matically this may be written as 



-H O m d r , 



(2) 



GEOPHYSICAL LABORATORY 



217 



where c n is the total number of atoms 
of species n per unit cell and b m is the 
multiplicity of the position in which the 
mth atom is located. This latter con- 
straint is usually handled implicitly by 
subjecting each atom in the asymmetric 
unit to the first condition while refining 
the coefficients and then checking that 
the chemical formula is confirmed. How- 
ever, in at least one structure refine- 
ment, that of a grunerite by Finger and 
Zoltai (1967), this procedure did not 
work. This Mg-Fe amphibole had a 
value for the ratio Mg/(Fe + Mg) of 
0.12, according to the chemical analysis 
of Klein (1964), and a total of 1.68 Mg 
atoms per cell, but the least-squares re- 
sult led to a ratio of 0.33. The chemical 
equivalence of the material studied by 
Klein and the single crystal used in this 
work is reasonably certain, since both 
investigators determined the same value 
for the unit-cell volume, a parameter 
shown to be very sensitive to composi- 
tion by Klein and Waldbaum (1967). 
The discrepancy in the chemical compo- 
sition was initially resolved by a very 
complicated series of refinement steps. 
The final solution was checked with a 
difference Fourier synthesis, which 
showed very little discrepancy in the 
electron density at the locations of the 
cations. When the above refinement de- 
scription was presented, C. T. Prewitt 
(personal communication) suggested that 
the constraint of (2) should be applied 
explicitly to this structure. If mag- 
nesium is chosen as species 1 and the 
amphibole cation positions Mi to M 4 , as 
positions 1 to 4, respectively, the second 
constraint may be written as 

4a 11 + 4a 21 + 2a 3 i+4a 4 i = 1.68 (3) 

for this structure and composition. Note 
that the similar equation relating the 
iron contents of the sites is not inde- 
pendent, since it may be found from a 
linear combination of (3) and (1) be- 
cause, in this case, (1) is an equality. 
Thus, for this structure, (2) contributes 
one equation to the eight needed to de- 



termine the eight coefficients of occu- 
pancy. Four equations are supplied by 
(1), and there are three parameters to 
be fitted by least-squares analysis. 

When the occupancy coefficients are 
fitted by refinement, the derivatives of 
the structure factors with respect to the 
parameters being varied must be com- 
puted. If the changes in the model are 
to be calculated properly, however, the 
constraints must be used in the calcula- 
tion of the derivatives. The nature of the 
relationships introduced by (1) is well 
known and will not be covered here. The 
relationships for (2), although just as 
obvious, have not been described. In 
general, if parameter s is dependent on 
parameter r, this dependency must be 
applied to the least-squares equations as 
a modification to the derivatives of the 
structure factor with respect to the re- 
fined parameter. In addition, s must not 
be varied as a least-squares parameter. 
The derivatives are modified with the 
use of the chain rule to yield equations 
of the form 



(4) 



\dr) \drji + \ds)t\dr) 

where F is the calculated structure fac- 
tor for the observation in question, the 
subscript i refers to the derivatives com- 
puted as if the parameters were inde- 
pendent, the quantity on the left being 
the derivative needed for the least- 
squares analysis. If the dependent oc- 
cupancy parameter is a pn , then (4) may 
be rewritten as 

(°L) = (°L) _ A{ aq (5) 

\ oa mn ) \ ca mn )i o m \ aa pn ) t 

with one equation of this form for each 
of the three refined occupancies. 

When the data on the grunerite men- 
tioned above were subjected to least- 
squares analysis with the use of (5), the 
occupancy parameters converged in two 
cycles to the values obtained by the very 
laborious procedure referred to above. 



218 



CARNEGIE INSTITUTION 



Magnetic Susceptibility and 

Exchange Coupling in 

Ardennite * 

F. E. Sen] tie, j A. Thorpe, ,t and G. Donnay 

The highly distorted oxygen octahedra 
about Mn 2+ and (Mn 2+ ,Ca) in ardennite 
were suspected to result in abnormal mag- 
netic behavior {Year Book 66, p. 485). 
Magnetic susceptibility measurements 
were made along different crystallo- 
graphic directions from room temperature 
down to 5°K. They showed anisotropic 
behavior along the short b axis (5.81 A) 
and isotropic behavior in the (010) plane. 
The variation of the magnetic suscepti- 
bility with temperature can be described, 
according to Earnshaw and Lewis 
(1958, 1960), as follows: 

X T =±[A+B-G(6)]+N(a) 

where A and B are constants, G{6) is 
an exponential term, and N(a) is a tem- 
perature-independent term. Computer 
calculations give the best fit for .4 = 17.5, 
-8 = 4.5 emu/°K/g, and for the exponen- 
tial term, = 1.35 to 1.60 depending on 
the direction with respect to the b axis. 
The data so obtained can be interpreted 
if, in addition to the normal paramag- 
netic contribution of manganese, there 
exists an antiferromagnetic contribution 
due to exchange coupling between ad- 
jacent manganese atoms through an 
intervening oxygen atom. Manganese 
atoms in both crystallographic positions 
participate in this exchange coupling. 
The effect appears to be due to the short 
Mn-Mn distance in the structure, rather 
than the distortion of the oxygen octa- 
hedra about the manganese atoms. 

"MCKELVEYITE," A SYNTACTIC INTER- 
GROWTH OF TWO PHASES 

G. Donnay and J. D. H. Donnay 

The first publication on mckelveyite 
(Milton et al., 1965) attracted the atten- 

* Publication authorized by the Director, U. S. 
Geological Survey, Washington, D. C. 
t U. S. Geological Survey, Washington, D. C. 



tion of one of us because the reported 
space group PS and the development of 
the crystals shown on photographs (Mil- 
ton et al, 1965, Figs. 16 and 3) did not 
jibe. A second paper (Desautels, 1967), 
specifically treating the morphology of 
mckelveyite, assigned it to space group 
P3ml or P3, still not in agreement with 
the symmetry observed on inspection 
(Desautels, 1967, Fig. 1). We were suf- 
ficiently intrigued by then to ask the 
authors for type material that we might 
study. We wish to thank Dr. Charles 
Milton for a most informative discussion 
and for his generous cooperation in pre- 
senting us with all the specimens of his 
collection. 

The samples represent an example of 
"polycrystals" (Donnay, 1953) strik- 
ingly similar to the ones formed by the 
bastnaesite, parisite, roentgenite, synchi- 
site series, mCeFC0 3 -nCaC0 3 (Donnay 
and Donnay, 1953). Two distinct phases 
are intergrown, one of which has the cell 
dimensions reported by Milton et al. 
(1965) but with space group P31m, the 
absence of the center of symmetry being 
indicated by morphology. The other 
phase is also hexagonal; it controls the 
morphology of the polycrystals studied 
by Desautels. We propose for it the name 
ewaldite in honor of Professor P. P. 
Ewald (pronounced avald, a as in able, 
a as in act). Its space group is P6 3 rac, 
again taking into account the morpho- 
logical evidence for the absence of center 
and horizontal mirror on crystals of 
the second phase. The a axes of the two 
phases are turned 30° with respect to 
each other about the common c axis. 
The cell dimensions of ewaldite (a e , c e ) 
are related to those of mckelveyite (a m , 
c m ) as follows: 

a e =5.291 ±0.006 A=a m /V3 = 

9.164/ V3± 0.010 A, 

c e = 12.751 ±0.012 A = 2c m /3 = 

2x19.126/3 ±0.018 A, 

7 e =27 m /9 = 2 x 1391/9 ±4 A 3 



GEOPHYSICAL LABORATORY 



219 



The syncrystallization of the two species 
is thus explained by the syntactic rela- 
tions: 3c e = 2c m , a e = a m /VS (in the xyu 
plane, the ewaldite net "//'-centers" the 
mckelveyite net). These data were ob- 
tained from back-reflection Weissenberg 
photographs taken at 21 °C with CuKa 
radiation (Aa 1 = 1.54051 A, Xa 2 = 1.54434 
A). 

X-ray studies of some twelve speci- 
mens showed that the shiny black crys- 
tals consist exclusively of mckelveyite, 
whereas the dull light bluish-green to 
dark green individuals are polycrystals 
of ewaldite and mckelveyite. Only one 
crystal of pure ewaldite has so far been 
discovered. Its color is deep green. The 
quickest method of determining the rela- 
tive amounts of the two phases present 
in a specimen consists in taking a cone- 
axis precession photograph with the c 
direction along the X-ray beam (Plate 
1). Since there is no difficulty in identi- 
fying the c direction and adjusting the 
(0001) face in prism position on the 
two-circle goniometer, it takes only a 
few hours to obtain the desired infor- 
mation. 

As would be expected, the black mckel- 
veyite crystals frequently show twinning 
(by merohedry) on 21.0 (Milton et al., 
1965, Fig. 3) , with twin symmetry to be 
written 6'm'm in black-white symbolism. 
Ewaldite, contrary to expectation, has 
not been observed to twin. Another 
feature characteristic of mckelveyite 
consists of continuous and near-uniform 
diffraction streaks in reciprocal space 
running parallel to c* for all those hk-l 
values that obey ky^Sp + h, l^Zr — h, 
where p and r are any integers. Reflec- 
tions of the type h- (3p + h) • (3r — h) give 
the usual sharp Bragg reflections. About 
250 of the latter have been recorded for 
a pure mckelveyite crystal, whose struc- 
ture determination is now under way. 

Ewaldite shows only sharp reflections. 
In addition to the space-group absences, 
we observe structural extinctions hk'l 
absent for l — 2n + l and h — k = 3n. This 



is an absence criterion for position 2(6) 
in P6 3 mc and thus indicates that the 
heavy atoms, barium, rare earths, uran- 
ium, must occupy this position with site 
symmetry 3m. 

Electron-probe studies of four poly- 
crystals showed no evidence of chemical 
differences in heavy atom and calcium 
content between mckelveyite and ewald- 
ite. They could be different hydrates, of 
course. Dr. Max Hey is performing 
microchemical analyses on crystals stud- 
ied by X rays, to give further informa- 
tion on the compositions. The Fe 2 3 
content reported by Milton et al. (1965) 
can be ascribed to iron-rich inclusions, 
which show up strikingly on an X-ray 
image of FeKa radiation (Plate 2) taken 
by Mr. C. Hadidiacos. The inclusion 
was identified under the reflecting micro- 
scope by Dr. A. El Goresy as hematite 
pseudomorphous after magnetite. For the 
time being we are using, for mckelveyite, 
an idealized cell content based on the 
published analysis (Milton et al., 1965) : 

2(Na 2 Ca x ) (Ba 4 R.E.x. 7 Uo. 3 ) (C0 3 ) 9 

The density measured for several pure 
mckelveyite crystals on the Berman bal- 
ance at 24°C is 3.25 ±0.05 g/cm 3 , con- 
siderably less than the calculated value 
of 3.54 g/cm 3 obtained for the above 
cell content. 

Crystalline Heterogeneity 

Evidence from Electron-Probe Study 
of Brazilian Tourmaline 

G. Donnay 

The concept of "single crystal" de- 
fined in the conventional way as a 
"homogeneous edifice" is in need of criti- 
cal reexamination when applied to min- 
erals. We began this study with thin 
plates showing drastic color changes, cut 
from tourmaline crystals of Brazilian 
origin. Mr. Earl Williams of The Johns 
Hopkins University kindly supplied the 
crystals and cut the plates. The cuts 
were placed nearly perpendicular to the 



220 



CARNEGIE INSTITUTION 



boundary zone of the differently colored 
regions. Some crystals show color varia- 
tion along the c axis; they were cut 
parallel to (1210). Others show color 
changes perpendicular to the c axis ; they 
were cut parallel to (0001). An elbaite 
crystal (U. S. National Museum 
R- 17011) that looks uniformly pink to 
the naked eye was also studied. Its 
manganese content, which is mainly re- 
sponsible for the pink color, varies con- 
tinuously through a range down to 43% 
of the maximum over a distance of 
0.45 mm. This was an unexpected find- 
ing. The sharp changes that we observed 
in Mg, Fe, and Al concentrations at the 
color boundaries of the multicolored 
plates were expected, but they too were 
preceded and followed by gradual, con- 
tinuous changes in concentrations. The 
electron-probe results by themselves thus 
invalidate the assumption that a one- 
color chip picked from such a tourmaline 
crystal can be considered a "single crys- 
tal"; it is not chemically homogeneous. 
The same thin tourmaline plates were 
oriented normal to the X-ray beam on 
the precession camera. We obtained dif- 
fraction patterns of differently colored 
regions on the same film by raising the 
cassette slightly when, with the help of 
the translation movement of the goniom- 
eter head, a differently colored region 
was brought into the beam. On such 
doubly and triply exposed films, very 
small changes in cell dimensions and 
axial orientations can be observed. The 
maximum differences that we measured 
were of the order of 0.7% in a, 0.3% 
in c, and a 0°10' rotation of the a axes 
about a constant c-axis direction. From 
three chips from the one elbaite crystal 
R-17011 cell dimensions were refined by 
a least-squares treatment of back-reflec- 
tion Weissenberg data (CuKa^ 1.54051 
A) , with the following results : a x = 15.836 
±0.002, d= 7.0986 ±0.0005; a 2 = 15.846 
±0.004, c 2 = 7.1020 ±0.0008; a 3 = 15.838 
±0.001, c 3 = 7.1032 ±0.0002 A. This 
agreement would be considered good 
enough to warrant the conventional as- 



sumption of constant chemical composi- 
tion throughout the crystal if the elec- 
tron-probe results had not warned us. 
Bragg reflections give the combined dif- 
fraction contribution of millions of unit 
cells and do not tell us whether or not 
the solid solutions present show short- or 
long-range order. If then we use the in- 
tensities from such a chip to refine the 
elbaite crystal structure, i.e. the atomic 
coordinates, temperature factors, and 
chemical occupancies, implicitly letting 
the chemical composition be constant 
throughout the crystal, we are misin- 
terpreting the diffraction data. We shall, 
probably with the help of least-squares 
refinement, arrive at a structure that 
gives a small residual R, but the refined 
structure may have limited physical sig- 
nificance. 

Careful testing of the crystal to be 
used for X-ray structure determination 
appears to be a necessity. Assurance of 
constant concentration of the major 
chemical constituents in a mineral crys- 
tal should be a prerequisite for using 
the crystal in a structure refinement. 

Evidence from Source-Image Distortion 

R. A. Young* C. E. Wagner* C. O. Pollard* 
and G. Donnay 

Seven of the multicolored Brazilian 
tourmaline plates mentioned above were 
examined for variations in macromosaic 
texture by an X-ray diffraction tech- 
nique described by Young and Wagner 
(1966). Five of the plates came from 
different crystals, were 0.7 mm thick, 
and were cut parallel to (0001). Two 
thinner ones, 0.2 mm across, came from 
the same crystal; one was cut parallel 
to the basal plane, and the other was 
cut parallel to (1210). 

In the arrangement used for SID 
studies, a diverging CuKcx X-ray beam 
from a spot source (angle of divergence 
equals 13' of arc) impinges on the crystal 
plate at a distance of 134 cm from the 

* Physical Sciences Division, Georgia Insti- 
tute of Technology. 



GEOPHYSICAL LABORATORY 



221 






X-ray source; a 5-mm-wide slit is placed 
in the incident beam, 4 cm from the 
sample. The plate is oriented so that one 
of the Bragg reflections is in diffracting 
position. The diffracted beam is recorded 
for about 10 minutes on Ultra-Speed 
dental film, placed perpendicular to the 
diffracted beam, about 3 cm from the 
crystal (Plate 3) . 

An X-ray radiograph of the exposed 
section of the same crystal plate (Plate 
4) shows a less-absorbing ring on the 
outside; it is colorless. A more-absorbing 
inside field is blue. Cell dimensions ob- 
tained from X-ray precession films in- 
dicate that the composition of the plate 
is on the dravite-schorl solid solution 
range, near the halfway point. The dif- 
ferences in a and c axes for the differ- 
ently colored regions are both close to 
0.2%. 

On all samples studied, we observe a 
striking, abrupt change in macromosaic 
texture (Plate 3). It coincides with the 
color change but is more sharply de- 
lineated. The outside region of all the 
plates studied so far is light pink or 
colorless. It is always composed of small 
elongated grains in which the direction 
of elongation serves as the line about 
which lattice misalignments of the order 
of several minutes of arc occur. The cen- 
tral, darker, more iron-rich regions con- 
sist of larger grains, with only small mis- 
alignments in lattice directions between 
grains. Texturally, the central region 
would be described as a nearly perfect 
single crystal, the outside region as an 
exceedingly imperfect crystal. 

We are led to the questions: Are the 
changes in texture a result of the chemi- 
cal change, or are they due to drastic 
changes in the conditions of crystalliza- 
tion, such as temperature and pressure, 
that may be accompanying the chemical 
changes? Do other mineral crystals also 
show more than one texture type? What 
will synthetic crystals show? There is 
need for further extensive work along 
these lines. 



Structure Refinement of Elbaite 
G. Donnay and R. Barton* 

The refinement of buergerite, NaFe 3 3+ 
Al 6 B 3 Si 6 O 30 F (Barton, 1967), shows no 
significant changes in fractional coordi- 
nates from those determined by Buerger, 
Burnham, and Peacor (1962) for dravite, 
the magnesium tourmaline variety. This 
was a surprising find, because chemical 
as well as physical properties of the two 
varieties differ markedly. We must now 
investigate the Li,Al end member, el- 
baite, to see whether its fractional co- 
ordinates are also the same. There ap- 
pears to be no natural solid solution be- 
tween elbaite and dravite, and the two 
minerals therefore may be expected to 
show some significant differences in 
crystal structure. 

If, as results so far obtained suggest, 
elbaite proves to have the same fractional 
coordinates as buergerite and dravite, 
we could, of course, calculate interatomic 
distances and angles in all tourmalines 
from knowledge of cell dimensions alone. 
This seems highly unlikely. 

Elbaite was claimed to have been the 
tourmaline studied by Ito and Sadanaga 
(1951). When plotting the given cell 
dimensions, a=16.0, c = 7.17 A on the 
diagram of cell dimensions c versus a 
(Mason, Donnay, and Hardie, 1964, 
Fig. 1), however, we were impressed by 
the fact that these dimensions placed 
the composition on the schorl-dravite 
solid solution join about one third of 
the way from schorl. The "probable 
atomic ratios Na:Ca = 7:l, Li:Al = l:2, 
Al:Fe:Mn = 6:0.1:0.1" (given by Ito and 
Sadanaga, 1951) are identical with those 
given in Dana's System of Mineralogy 
(sixth edition, p. 555) for Brazilian 
rubellite. We asked Professor Sadanaga 
for the crystal he had used and were told 
in his answer of December 9, 1967, "In 
fact, I did not even see the specimen, 
only being informed that it was a Brazil- 
ian rubellite. I therefore carried out a 

* Carothers Laboratory, E. I. du Pont de 
Nemours and Company, Wilmington, Delaware. 



222 



CARNEGIE INSTITUTION 



structure determination by assuming the 
composition as identical with those de- 
scribed in Dana's System." Unfortun- 
ately, the crystal itself has been lost. 

We therefore decided to carry out an 
elbaite study on a single crystal large 
enough so that the chemical analysis, 
electron-probe tests, crystal optics, and 
infrared and ultraviolet spectroscopy (to 
be done by Dr. R. W. T. Wilkins at 
Harvard University), as well as the 
structure refinement, could be performed 
on material from the one crystal. Dr. G. 
Switzer, of the U. S. National Museum, 
kindly supplied a large, gem-quality, 
pink crystal (USNM R-17011) from 
San Diego County, California. Dr. B. 
Wiik has performed a wet-chemical 
analysis, which shows it to be a nearly 
pure Li,Al composition, with only 0.24 
wt % FeO and 0.02 wt % MgO. As 
mentioned above in the section on "Crys- 
talline Heterogeneity, Electron-Probe 
Evidence," we tested several chips of 
this crystal for constancy of chemical 
composition and found especially that 
the Mn content was variable. Wiik found 
2.36 wt fo of MnO, an unusually high 
value. One circular traverse with the 
electron probe gave nearly constant con- 
centrations of Si, Al, Mn, and Ca. This 
encircled region was chosen for X-ray 
study. Although there is the danger of 
heterogeneity in the specimen, we do not 
dare probe it until the structure refine- 
ment reaches a final stage so we can be 
sure that no further intensity measure- 
ments will be needed. This precaution is 
necessary because electron-probe scan- 
ning alters the chemical composition; 
especially, it removes alkali-metal ions 
from the specimen. 

A complete set of intensities has been 
collected with MoKa radiation on the 
Supper-Pace automated diffractometer, 
and refinement is under way. 



Improvements of Crystallographic 
Equipment 

G. Donnay 

The Optical Analyzer 

The optical analyzer (Donnay and 
Donnay, 1957, 1966) is mounted on the 
stage of the microscope. When making 
the necessary adjustments of the goniom- 
eter head on which the crystal is 
mounted, one would like to have the 
analyzer stationary, not sliding about. 
Transfer tape has been used but has 
proved unsatisfactory. Instead, three 
magnets have now been mounted in the 
base of the analyzer, and whenever the 
analyzer is being used, the glass plate 
of the microscope is replaced by a metal 
plate in which a circular hole has been 
cut in the center, the same size as the 
hole in the optical analyzer, permitting 
light from below to pass through the 
specimen. The magnets hold the analyzer 
with sufficient strength to prevent slid- 
ing yet permit easy removal of the in- 
strument. The arrangement has proved 
highly satisfactory. 

Precision Film-Measuring Device 

Back-reflection Weissenberg as well as 
powder films should be measured with 
the highest possible precision whenever 
a least-squares refinement of the data 
they furnish is to be carried out. A re- 
cently developed film-measuring device 
of Picker Corporation (catalog no. 684- 
813) permits readings to be made to 
0.01 mm with a dial indicator, a pre- 
cision greater than is warranted by the 
size of most diffraction spots and lines. 
The instrument does not provide, how- 
ever, for movement of the film in a di- 
rection normal to the reading scale, so 
that a powder film cannot be readily 
centered before the measurements are 
begun and the Weissenberg film, which 
has to be translated parallel with itself 
during measurements, cannot be read at 
all. 



GEOPHYSICAL LABORATORY 



223 



The instrument has been rebuilt so 
that the frame holding the reading scale 
can be moved up and down over a dis- 
tance of iy 2 inches in carefully machined 
grooves. The film remains stationary 
during the measurement, while an ad- 
justing screw moves the scale parallel to 
itself. 

SONORAITE 

Richard V. Gaines* G. Donnay, Max H. Heyj 
and J. Zemann t 

A new iron tellurite mineral, Fe 2 8+ 
Te 2 4+ 7 -zH 2 0, has been discovered by 
the senior author in the oxidized zone 
of the Moctezuma tellurium-gold mine 
in the state of Sonora, Mexico. It con- 
sists of highly lustrous, yellowish-green 
rosettes, usually associated with emmon- 
site, Fe 2 3+ Te 3 4+ 9 -2H 2 0, and has been 
found so far only in very small amounts. 
The chemical analysis was carried out 
(by Hey) on only 890 fig of material, 
and the water content is therefore not 
firmly established. 

The space group P2 1 /c is uniquely 
given by the observed systematic ab- 
sences: 0/cO absent with k odd, hOl ab- 
sent with I odd. The cell dimensions at 
21 °C obtained from a least-squares re- 
finement (Program LCLSQ of C. W. 
Burnham, Year Book 61, pp. 132-135) 
of back-reflection Weissenberg photo- 
graphs with CuKa radiation (X.a 1 = 
1.54051, Aa 2 = 1.54434 A) are a = 10.984 ± 
0.002, 6 = 10.268 ±0.001, c = 7.917±0.002 
A, £ = 108.49 + 0.02°, V = 846.8 ±0.5 A 3 . 
For 4(Fe 2 Te 2 7 -2H 2 0) per cell, the 
calculated density is 4.04 g/cm 3 , in good 
agreement with the measured value of 
3.95 ±0.01 g/cm 3 obtained by the flota- 
tion method in Clerici solution. 

The thin platy habit of sonoraite cry- 
stals shows little variation. The dominant 
form is the pinacoid (100) ; rhombic 
prisms (110) and (011) border the plate 
in the order of importance predicted by 

* Universidad Nacional Autonomo de Mexico, 
t British Museum (Natural History). 
t University of Vienna. 



the Generalized Law of Bravais (Don- 
nay and Harker, 1937) . Striations paral- 
lel to c appear on (100) and (110). 

The crystals are soft (H~ 3). They 
show biaxial negative optical character, 
with a = 2.018 ±0.003, /? = 2.023 ± 0.003 
and y = 2.025 ±0.003, determined in Car- 
gille oils. 2V is in the range 20° to 25°. 
No pleochroism is observed. 

The mineral has been approved by the 
Commission on New Minerals and Min- 
eral Names of the International Min- 
eralogical Association. 

The Crystal Structure of Sonoraite, 
Fe 2 Te 2 5 (OH) 4 -H 2 

G. Donnay, J. Stewart* and J. Zemann 

We recorded 1895 observable, sym- 
metry-independent reflections on the 
automated Supper-Pace diffractometer 
with MoKa radiation. The "X-ray 67 
phase program" (Technical Report 67-58 
of the Computer Science Center, Uni- 
versity of Maryland) led to the place- 
ment of Te and Fe atoms, which were 
then used for sign calculations for a 
first three-dimensional electron density, 
which gave ten oxygen locations (R = 
11.9%). All atoms are in general 4-fold 
position. Refinement by the least-squares 
method is in progress. 

Of the ten oxygens present, one or two 
belong to water molecules. The presence 
of at least two hydroxyl ions is indicated 
by the Pauling bond strength distribu- 
tion. We may picture the structure as 
composed of stepped chains of anion 
octahedra about ferric ions extending 
along [101]. The average iron-anion 
distance is 2.05 A. There are two such 
chains per cell, related to each other by 
the c glide plane. The four octahedra 
per cell in each chain form two edge- 
sharing pairs with a center of symmetry 
bisecting the shared edge; these pairs are 
joined by one shared corner. The Fe-Fe 
distances within the pairs are 3.12 and 
3.28 A, whereas the distance between 

* University of Maryland. 



224 



CARNEGIE INSTITUTION 



symmetry-independent iron atoms in 
the chain is 3.70 A. Both tellurium ions 
are tetrahedrally coordinated by three 
oxygens and the unshared electron pair 
of Te 4+ . The average Te-0 distance is 



1.92 A, the average 0-0 distance in the 
triangles bonded to Te is 2.75 A. The 
oxygen triangles connect the chains of 
iron octahedra into a three-dimensional 
framework. 



GEOCHRONOLOGY AND GEOCHEMISTRY 

G. L. Davis, T. E. Krogh, S. R. Hart* and L. T. Aldrich* with S. A. Morse and Kyoichi Ishizaka * 



The time of formation of the rocks 
in the Grenville structural and lithologic 
province, an area of high-grade meta- 
morphic rocks in the Canadian shield, 
has been a subject of controversy since 
the earliest geological investigations. The 
vastness of the area and its obvious 
boundary with the older Superior prov- 
ince rocks to the northwest make it a 
key to theories regarding the continental 
development of North America. About 
a decade ago several workers on geo- 
chronology concluded that these rocks 
were formed shortly before 900 m.y. 
ago, the age value obtained by the K-Ar 
method on micas in the area. Those char- 
acteristics of the rocks that indicated in- 
tense metamorphism and plastic defor- 
mation were assumed to have developed 
during the time of last heating, 900 m.y. 
ago. Our current results demonstrate 
that these conclusions are largely in- 
correct. 

Earlier studies by Krogh showed that 
intrusion and metamorphism occurred 
in the Grenville in southeastern Ontario 
between 1300 and 900 m.y. ago. Our 
recent studies in the northwest Grenville 
area of Ontario show that certain rocks 
in this area were metamorphosed and 
intruded between 1500 and 1800 m.y. 
ago. The whole-rock Rb-Sr method made 
it possible to look behind the younger 
event that reset all mineral Rb-Sr or 
K-Ar ages. We thus found a bimodal 
whole-rock age pattern. In the northwest 
Grenville area of Ontario extensive plu- 
tonism and regional metamorphism oc- 
curred between 1500 and 1800 m.y. ago 

* Department of Terrestrial Magnetism. 



and again between 1300 and 900 m.y. 
ago. Furthermore, it is probable that the 
Grenville Front existed as an active 
structural entity as early as 1700 m.y. 
ago, and was again the locus of move- 
ment between adjacent continental 
blocks about 1000 m.y. ago. 

In this report the results of our cur- 
rent studies are combined with some of 
those from the past 2 years. Last year's 
study of a single outcrop of paragneiss 
demonstrated that the chemical inter- 
actions that accompanied the formation 
of the gneiss took place between 1500 
and 1800 m.y. ago. We have verified this 
conclusion and extended the area to 
which it applies by examining paragneiss 
in other parts of the region. These recent 
studies, although less detailed, are more 
reliable because of the large variation in 
the Rb/Sr ratios discovered in adjacent 
paragneiss layers. 

Specific times of emplacement of 
granite at 1500 and 1725 m.y. were estab- 
lished by detailed whole-rock isochron 
studies presented last year. Other gran- 
ites in the region are now known to have 
been formed at these same times. 

Geochronology of the Grenville 
Province 

T. E. Krogh and G. L. Davis 

Paragneiss in the Northwest 
Grenville Area 

The geochronology report in Year 
Book 66 included a detailed discussion of 
Rb-Sr isotopic studies of a single block 
from a paragneiss outcrop in the French 
River area, location 10. Closely spaced 



GEOPHYSICAL LABOKATOKY 



225 



samples of this compositionally layered 
rock were not isotopically similar during 
the 1000-m.y. event that caused isotopic 
mixing among certain coexisting minerals 
analyzed. Considering the large amount 
of metamorphic chemical interaction be- 
tween certain layers, it seemed likely 
that the paragneiss formed during an 
earlier metamorphic event, a conclusion 
supported by the fact that an apparently 
syntectonic and synmetamorphic granite 
sill nearby yielded an isochron age of 
1730 my. 

To determine the regional extent of 
the metamorphism of paragneisses, we 
collected samples at locations 11, 12, 13, 
and 14 on Fig. 84. The diagrams on Figs. 
85 and 86 show the variation in the con- 
centrations of rubidium and strontium 
in sections normal to the layering. The 
dashed and crosshatched portions repre- 



sent felsic gneiss and biotite amphibolite, 
respectively. Concentrations estimated 
by X-ray fluorescence differ from isotope 
dilution values by less than 5%. As noted 
previously in Year Book 66 (p. 531), 
plagioclase-quartz layers commonly oc- 
cur between quartz-microcline-plagio- 
clase gneiss and plagioclase-hornblende- 
biotite amphibolite. The variations in 
plagioclase and microcline content, and 
concomitant Sr and Rb concentration, 
cause large variations in the Rb/Sr ratio. 
Limited isotopic data for small, single 
whole-rock slices taken near steep grad- 
ients in the Rb/Sr ratio generally can- 
not give precise ages. They are, however, 
sensitive indicators of isotopic mixing. 
For example, the two samples from sec- 
tion 12 subjected to isotope analyses 
were separated by 3 cm of rock. If the 
isotopic composition became similar over 




Fig. 84. Map of northwest Grenville area in Ontario, showing sample locations. Sample location 
numbers are used to identify analyzed specimens on the isochron diagrams. 



226 



CAKNEGIE INSTITUTION 




Fig. 85. (A) Paragneiss, location 14; analyzed samples 14-7 and 14-10. (B) Paragneiss section, 
location 12; analyzed samples 12-1 and 12-3. 







Fig. 86. (A) Paragneiss, location 13; analyzed samples 13-3 and 13-10. (B) Paragneiss, location 
11 ; analyzed samples 11-2, 11-9, and 11-10, the latter collected 2 meters to the right of the section. 



GEOPHYSICAL LABORATORY 



227 



these limited distances during the 1000- 
m.y. thermal event, a 1000-m.y. isochron 
slope would have resulted. Data for sec- 
tion 12 on Fig. 87 indicate an apparent 
age of 1560 m.y., demonstrating that iso- 



topic equilibration did not take place at 
1000 m.y. 

Data for three samples from section 11 
are colinear on the isochron diagram 
(Fig.- 87) and probably specify a time 



1.060 — 



1.020- 



.980- 



.940 



CD 
00 

v. 

.CO 



fc_^ .900 
CO 



.860 — 



.820- 



.780 



.740- 



.700 



1 1 


1 


1 1 


- 


_ 






_ 


— 




1 no y 


— 


— 




/ ® 12-3 


•" _ 


— 11-9/ 


4// 




__ 


— Jy 


13-3 




— 


_ 12-1 JW 








//I3-I0 

„ // + 14-10 

. +<u? 14.7 

1 1 


1 


1 1 


— 



1.0 



2.0 



3.0 



4.0 



5.0 



Rb A 



Sr 



Fig. 87. Isochron plot for paragneisa samples. Sample numbers specify locations on Fig. 84. 
Projected initial strontium ratios for related samples are 0.704 for location 11, 0.715 for location 
12, 0.716 for location 13, and 0.703 for location 14. 



228 



CAENEGIE INSTITUTION 



1840 m.y. ago when the isotopic compo- 
sition of strontium was identical in all 
parts of the section. These samples and 
those in section 12 are especially sig- 
nificant because their unusually high 
Rb/Sr ratios allow a more accurate de- 
termination of the slope of the isochron. 
The paragneisses in sections 13 and 14 
yield apparent isochrons for 1170 and 
1290 m.y., respectively, suggesting that 
these gneisses either formed or underwent 
isotopic mixing at these times. 

These studies confirm and extend the 
earlier conclusions, based on the French 
River paragneiss, that much of the para- 
gneiss in the northwest Grenville area 
formed from its sedimentary parent be- 
tween 1500 and 1900 m.y. ago. A re- 
gional metamorphism in this time inter- 
val is indicated. Other paragneisses in 
this region either formed or were iso- 
topically equilibrated about 1300 and 
1100 m.y. ago. 

Granitic Rocks in the Northwest 
Grenville Area 

French River granite and coeval 
granitic rocks. An isochron obtained for a 
granitic layer in a major northwest- 
trending syncline in the French River 
area (location 4, Fig. 84) was presented 
last year (Year Book 66). This isochron, 
with a slope of 1725 m.y., is shown with- 
out the data points on Fig. 88. Isotopic 
data for biotite and apatite in one sample 
of granite indicated isotopic mixing 
among minerals in the rock about 1000 
m.y. ago. Two sphene samples from this 
granite, analyzed by Tilton and Grunen- 
felder (1968), gave concordant U-Pb 
ages very close to 1100 m.y., indicating a 
younger thermal event. 

Data for whole-rock samples of other 
foliated leucogranites at locations 2 and 
3 (Fig. 84) are shown on Fig. 88. Incom- 
plete isochrons for 1730 and 1820 m.y. 
are apparent. The country rocks in these 
widely spaced localities must be at least 
as old as these intrusives. 

Lake Muskoka granite and similar 



granitic rocks. In Year Book 66 we pre- 
sented a Rb-Sr isochron study of an in- 
tensely metamorphosed and deformed 
granitic body situated approximately 
150 km southeast of the Grenville Front 
(location 1, Fig. 84). The isochron ob- 
tained from ten analyzed samples is 
shown as a reference isochron without 
data points on Fig. 88. Isotopic mixing 
about 800 m.y. ago is indicated by min- 
eral analyses. Our previous study also 
included the analysis of small, intensely 
recrystallized whole-rock volumes situ- 
ated 10 cm apart on the outcrop, with 
very different Rb/Sr ratios. These 
samples showed no evidence of gain or 
loss of the dating elements during the 
past 1500 m.y.; accordingly, the meta- 
morphism recorded in the rocks very 
likely occurred at about this time. 

Data for two samples from a similarly 
deformed granitic body at location 18 
(Fig. 84) lie on the reference 1500-m.y. 
isochron. In the same way, single whole- 
rock samples from locations 15, 16, 17, 
and 19 lie on or above this isochron, 
inferring that they too are at least 1500 
m.y. old. Certainly country rocks older 
than about 1500 m.y. must occur over a 
very large portion of the northwest 
Grenville area (Fig. 84) . 

Because the Muskoka granite is in- 
tensely recrystallized and deformed, the 
possibility exists that the isochron age 
indicates only the time since the rock 
underwent extensive metamorphism. A 
short pre-1500-m.y. history is implied, 
however, by the low initial Sr 87 /Sr 86 
ratio and the high degree of fit of the iso- 
topic data to a single isochron. We hope 
to elucidate this problem by a zircon 
study now in progress. To date analyses 
of zircons from three whole-rock stron- 
tium samples have been completed. These 
define a chord on the concordia diagram 
with intercepts at 1440 and 260 m.y. 
Significantly, the chord does not yield 
an intercept at the 800-1000 m.y. thermal 
event that produced isotopic mixing of 
strontium between coexisting mineral 
phases. Indeed, the three zircon points 



GEOPHYSICAL LABORATORY 



229 



in 



.960 



.940 



.920 



.900 



.880 



.860 



.840 



.820 



.800 



.780 



.740 



.720 



.700 




Muskoka granite, loc. I, 1500 my 
reference isochron 



Fig. 88. Isochron plot for granitic rocks. Reference isochrons shown here without data points 
are from previous report. French River granite occurs at location 4; Muskoka granite, at 
location 1. Sample numbers specify locations on Fig. 84. 



lie on a continuous diffusion curve drawn 
from 1440 m.y. 

North Bay trondhjemite and younger 
granites. As examples of the younger in- 
trusives we have studied garnetiferous 
trondhjemites in the North Bay area. 
Several samples collected at locations 6 
and 7 (Fig. 84) yield an isochron for 
1330 m.y. (see Year Book 65). Field 
work by S. B. Lumbers, of the Ontario 
Department of Mines, indicates that 
these bodies are generally large, 15 to 
30 km across, intensely deformed on their 
margins, and recrystallized but not 
strongly lineated or foliated in their 



cores. These bodies of metamorphosed 
igneous rocks may be useful in determin- 
ing the post-1300-m.y. metamorphism 
and deformation in the region. Two 
samples from a foliated granite body at 
location 8 (Fig. 84) yielded an approxi- 
mate maximum whole-rock age of 1200 
m.y. 

Age Relationships along the Grenville 
Front 

Sudbury area, Grenville Front granite. 
A previous report {Year Book 65) con- 
tained data for analyzed whole-rock 



230 



CARNEGIE INSTITUTION 



samples from a granite situated along 
the Grenville Front south and west of 
Sudbury, Ontario. A tentative isochron 
for 1750 m.y. was shown, but several 
of the analysed samples plotted below 
the isochron. Since that time, further 
samples and encouragement to renew 
our investigation were provided by Jack 
Henderson, of McMaster University, 
who recently completed a Ph.D. study on 
the structure and petrology of this por- 
tion of the Grenville Front. 

Data for four of the five whole-rock 
samples provided by Henderson lie on 
our previous 1750-m.y. isochron. The 
pre-Grenville age of this granite was 
confirmed by a muscovite Rb-Sr age of 
about 1600 m.y. and a zircon Pb-Pb 
age of 1655 m.y. These minerals are from 
a small northeast-trending granite dike 
that intrudes the Huronian section along 
the northwest margin of the granite. 

A dominant characteristic of this gran- 
ite is the pervasive evidence of both 
plastic and brittle deformation. Hender- 
son has shown that the structural ele- 
ments in the granite, in the plastically 
deformed migmatites on the southeast 
margin, and in the brittlely deformed 
Huronian rocks to the northwest, formed 
in response to a single stress pattern. A 
single Rb-Sr age of 1450 m.y. on musco- 
vite from a strongly foliated granite 
sample indicates that some of the folia- 
tion developed prior to the 1000-m.y. 
event. 

This granite was formerly considered 
to be a part of the Grenville structural 
province and to have been emplaced and 
deformed during the formation of the 
Grenville Front, 1000 m.y. ago. Later, 
after we discovered evidence for a 
1750-m.y. age, all the deformation fea- 
tures were believed to have formed 1000 
m.y. ago. Considering our isotopic data, 
it is probable that the northeast-trending 
granite mass with its parallel satellite 
dikes was intruded into the "Grenville 
Front" 1700 m.y. ago and that it under- 
went plastic and brittle deformation then 



or at some time earlier than 1450 m.y. 
ago. 

North Bay area, Grenville Front gran- 
ite. Data for several samples from a 
granitic body situated 14 km south of 
the biotite isograd along the Grenville 
Front, north of North Bay, Ontario, 
yield an apparent isochron age of 2600 
m.y. (Fig. 88) . The isotopic composition 
of apatite from two of the analyzed 
samples is identical with that present in 
each host rock 860 m.y. ago (Fig. 88). 
A single biotite sample from this granite 
has a Rb-Sr isotopic age of 960 m.y. 

To date, this granite comprises the 
maximum southward extension of the 
Superior Province into the region af- 
fected by the so-called Grenville orogeny 
about 1000 m.y. ago. 

Chibougamou area, Grenville Front at 
Surprise Lake, Quebec. A metamorphic 
transition situated along the boundary 
between the Superior and Grenville prov- 
inces has been mapped and studied in de- 
tail by Deland (1956). In this area fine- 
grained metasediments and basic vol- 
canics can be observed in various stages 
of recrystallization at progressively 
higher temperatures as one proceeds 
along the regional strike from west to 
east. Migmatites and gneisses typical of 
the Grenville province occur beyond this 
transition to the south and east. 

In our investigation of age relation- 
ships along the Grenville Front, we 
selected muscovite samples from the edge 
of the transition zone, as well as 6 and 
12 km to the southeast. A Rb-Sr isotopic 
age between 2600 and 2700 m.y. at each 
location demonstrates that this meta- 
morphic transition developed about 2700 
m.y. ago. A coexisting biotite from the 
sample at the edge of the transition has 
a Rb-Sr age of 2100 m.y., and another 
biotite collected 15 km to the south- 
west has an age of 880 m.y. These ages 
on biotites, which are more responsive 
to thermal events than muscovite, dem- 
onstrate the existence of the younger 
metamorphism in part of this area. 



GEOPHYSICAL LABORATORY 



231 



Fractionation of Potassium and 
Rubidium in a Layered Intrusion 

S. A. Morse and G. L. Davis 

The ratio of potassium to rubidium in 
igneous rocks is known to vary through 
approximately two orders of magnitude, 
i.e., from --'30 to ~3000. It may be in- 
ferred from this that the K/Rb ratio is 
potentially an indicator of igneous frac- 
tionation or of the source regions of 
magmas. Such an inference has underlain 
main interpretations of K/Rb variation, 
although detailed examination of the 
partitioning of these two alkali elements 
between crystals and liquids is lacking. 
The evidence, largely from volcanic 
rocks, points to decrease of K/Rb with 
increased fractionation, particularly in 
late stages (e.g., Lessing, Decker, and 
Reynolds, 1963; Gast, 1965). 

We have found in the Kiglapait 
layered intrusion, Labrador, an excep- 
tion to the supposed normal trend 
(Morse and Davis, 1968). According to 
Morse (1968), the intrusion provides an 
excellent natural example of in situ frac- 
tionation of basaltic magma under plu- 
tonic conditions, several lines of evidence 
indicating that the magma was virtually 
a closed system throughout its fractiona- 
tion history. We have analyzed K and 
Rb by stable-isotope dilution techniques 
for 19 whole rocks and for several sepa- 
rated feldspars and mafic minerals. 
Figure 89 shows a K/Rb versus K plot 
for the whole-rock data, along with the 
approximate fields of abyssal tholeiites, 
some meteorites, and ultramafics. Many 
of the Kiglapait whole-rock ratios in 
Fig. 89 fall far from the previously de- 
termined fields, and although a valid 
trend cannot be established from this 
figure, a gentle increase in K/Rb with 
fractionation is suggested by study of a 
K versus Rb plot and by Fig. 90. 

In Fig. 90 the whole-rock values of K 
and Rb are plotted against the fraction 
of liquid remaining, which is computed 
on a geometrical basis from the form and 
internal structure of the intrusion 



3000 


i i — r 

Key- 




1 




O Kiglopoit rocks 






2500 


_: + Inferred Kiglapait magma 




<fe 


- 


2000 




Abyssal tholeiites — l~\ ° 
(Gast, 1965) V \ C 






1500 




o 


- 


1000 


<\ \ 




- 


500 


- \ > 

-^QrChondrites ' — ' 

x^TJItramafic rocks r'\ \ T*"-. 




- 





Q (Steuber + Murthy i ra66J^arb,"Chondrites 




i 


0.01 O.I 1.0 




10.0 




Per cent K 







Fig. 89. K/Rb versus K plot for Kiglapait 
intrusion whole rocks and the inferred initial 
magma compared with ultramafic rocks, abyssal 
tholeiites, and chondritic meteorites. The 
Kiglapait magma is derived by a numerical 
summation of K and Rb over the volume of 
the intrusion. 

(Morse, 1968). The scatter of these 
points is due largely to variation in the 
modal amount of mafics in the rocks. 
Data for separated feldspars probably 
define a smoother trend and are indicated 
by special symbols. The central portion 
of Fig. 90 shows that K/Rb surely does 
not decrease with fractionation and 
probably increases slightly. 

Both K and Rb increase with frac- 
tionation. Both are concentrated in feld- 
spar relative to the whole rock, but the 
early plagioclases have consistently 
higher K/Rb ratios than their host rocks 
and their calculated parent liquids. There 
is little doubt that plagioclase discrim- 
inates against Rb relative to K in slow 
magmatic processes. It would be ex- 
pected, therefore, that Rb would be con- 
centrated in the liquid and that the K/Rb 
ratio of the liquid and later crystals 
would tend to fall. The role of the mafics 
may be to suppress this tendency. Oli- 
vine and augite in the Kiglapait clearly 
have low K/Rb ratios as well as low 
alkali contents, if not for crystal-chemi- 
cal reasons, then simply because they 
reflect the ratio of the nearby liquid. 
It is possible to generate numerical 
closed-system models of Kiglapait frac- 



232 



CARNEGIE INSTITUTION 



E 



CD 
Cl 



£ 









1 




1 


1 


1 

□ 
















o ° ° 


o 4 










□ o 
O x 

o 
o 


o 


X _ 


03 


a 
o O 

o 
o 

A 


□ 

o 


oo x 
o 


o 


K 




- 




- o 








o 


o 


o o I 




-o 
o 

o 


o 


o 

o 
o 


o 


° 
K/Rb 




o 
















□ 
















o 


10 










o 
□ 
° o 


o 


o ° 




□ 
o 

o o 

a 


o 
□ 


o 
o 

o 


o 


Rb 






1 


□ 


KEY 
Feldspar 














O, X 


Whole rock : ID, X-ray 








1 




1 


A 

1 


Olivine 

1 



.0 



-2500 

1500 
K/Rb 

500 



0.1 0.01 0.001 

Fraction of liquid remaining 



o.oooi 



Fig. 90. Values of K and Rb for whole rocks and separated minerals as well as K/Rb for 
whole rocks of the Kiglapait layered intrusion, plotted against the fraction of liquid remaining, 
derived from geometrical analysis of the intrusion structure. 



tionation that mimic the observed effects 
and culminate in high K/Rb ratios. 

Assuming a closed system, a numerical 
summation of K and Rb for the Kig- 
lapait yields K = 1530, Rb = 1.2 ppm, 



K/Rb = 1270 for the whole intrusion and 
thus the initial magma. In this respect, 
the Kiglapait belongs to a group of high 
K/Rb rocks, typified by the abyssal 
tholeiites and including anorthosite 



GEOPHYSICAL LABOKATOKY 



233 



(Reynolds, Whitney, and Isachsen, 1968; 
Gill, 1968), amphibolite (Hart and 
Aldrich, 1967), and parts of the Bush- 
veld complex (Erlank, Danchin, and 
Fullard,1968). 

It is difficult to ascribe such high 
K/Rb ratios to mineralogical control in 
the source regions of basalt, since the 
only two major rock-forming minerals 
known to have high ratios are plagio- 
clase and amphibole, and they probably 
have limited stability in the upper 
mantle. It may be profitable to consider 
prior stripping of alkalies, especially Rb, 
by limited partial fusion in the upper 
mantle. A more extensive later fusion 
might then produce basic magmas show- 
ing high K/Rb, provided some mineral 
with high K/Rb remained from the first 
event. Such stripping processes, if they 
did occur, would have important impli- 
cations for models of heat flow, crustal 



evolution, and the geochemistry of K and 
Rb and their radiogenic daughters. 

The' evidence from the Kiglapait in- 
trusion emphasizes our need to under- 
stand the fundamental controls on ele- 
ment partitioning between crystals and 
liquid in magmatic processes. It is clear 
that we could have neither predicted the 
Kiglapait trend nor made intelligent geo- 
chemical guesses about the source of the 
Kiglapait magma from a knowledge of 
the initial K and Rb values alone, using 
volcanic rocks as a model. Yet many 
volcanic rocks presumably owe some por- 
tion of their evolution to fractionation 
in subjacent chambers that resemble 
layered intrusions in principle, and an 
understanding of plutonic-element par- 
titioning is therefore essential to correct 
geochemical interpretation of many vol- 
canic suites. 



PETROLOGY 



STATISTICAL PETROGRAPHY 

The work in statistical petrography 
has ranged over such a broad subject- 
matter spectrum that it is difficult to pre- 
sent a coherent resume or single out par- 
ticular projects for detailed review. A 
long manuscript on petrographic ratio 
correlation, integrating results obtained 
over nearly 20 years of intermittent 
study of the percentage data that form 
the core of chemical petrography, has 
been completed. As might almost have 
been guessed, the attempt to present a 
unified, consistent development of old 
material has led to discoveries of previ- 
ously unsuspected relationships; two of 
these are described below. In the first it 
is shown that, for the particular null 
model in question, the correlations ex- 
pected between the variables of the re- 
cently proposed and so far unused "re- 
maining-space" transformation are iden- 
tical, to the limits of approximation, with 
those between the "Niggli numbers" so 
well known and widely used in petrog- 



raphy. As shown in last year's report, 
however, it will usually be impossible to 
calculate either set of expectations, for 
the reason that the hypothesized "open" 
parent is numerically impossible, the 
sample estimates of certain parent vari- 
ances proving consistently negative. Year 
Book 66 reported preliminary work on 
transformations aimed at eliminating 
the negative elements of the open vari- 
ance vector. The transformations de- 
scribed there performed rather unevenly, 
however, often either introducing new 
negative elements in the course of elim- 
inating the original ones or failing to 
eliminate the latter. The transformation 
described below operates satisfactorily 
on all data so far tested. 

As an outgrowth of work on the "re- 
maining-space" transformation (see Fear 
Book 65, p. 376) attention was directed 
to a transformation that may often 
linearize ternary closed data that appear 
to follow some gently nonlinear trend. 
The potential importance of this and 



234 



CARNEGIE INSTITUTION 



similar transformations arises from the 
fact that the variables of a closed array 
are interdependent, so that conventional 
regression analysis, which requires iden- 
tification of one of the variables as inde- 
pendent and the other as dependent, is 
inapplicable. Where linear fitting is suffi- 
cient, the "reduced major axis" or "line 
of organic correlation" may be used, 
but this is not defined for nonlinear inter- 
dependence. Accordingly, a transforma- 
tion eliminating systematic departure 
from linearity in ternary data might 
prove very useful. The "binary ratio" 
transformation described below appears 
to have this effect on the data for the 
calciferous pyroxenes of the layered 
series of Skaergaard. 

A much simpler and more conven- 
tional examination of interdependence 
seems to establish beyond any reason- 
able doubt the existence of marked inter- 
laboratory bias in the analysis of sili- 
cates. Paired analyses of the well-known 
granite (Gl) and basalt (Wl) refer- 
ence materials distributed by the U. S. 
Geological Survey are the subject of this 
study. If there were no systematic dif- 
ferences between laboratories one would 
expect no correlation between results ob- 
tained by an individual analyst for the 
same oxide (s) in the two rocks. In fact, 
there is very strong positive correla- 
tion between paired values of Si0 2 , MgO, 
Na 2 0, K 2 0, ALO3, and FeO, the ob- 
served values of four of these correla- 
tions being significant at the 1% -level 
and the other two at the 5% -level. 

A carefully designed test (Flanagan, 
in Stevens et al., 1960) has failed to 
detect inhomogeneity with regard to Pb 
content. Despite this evidence, com- 
plaints about the alleged heterogeneity 
of the two (powdered) materials are not 
infrequent. The present examination of 
already published data strongly supports 
Flanagan's position, in the sense that 
the combined effect of random analytical 
error with whatever sample inhomo- 
geneity exists is obviously insufficient to 
obscure interlaboratory bias. The popu- 



larity of powdered reference materials 
among geochemists is now such that al- 
legations of sample inhomogeneity should 
be neither casually brought nor casually 
dismissed. But the burden of proof is 
surely on the plaintiffs in the present 
instance. With regard to major constitu- 
ents, what evidence there is strongly 
suggests a considerable interlaboratory 
bias. Unless this bias is eliminated, or 
data are obtained in a pattern that per- 
mits its isolation and estimation, no col- 
lection of analyses will throw much light 
on the vexed and vexing question of 
sample inhomogeneity. 

Turning now from the criticism of 
chemical analyses to their use and ap- 
preciation, speculation on the penologi- 
cal significance of crystal-melt fractiona- 
tion often relies heavily on demonstra- 
tions of "gradual transitions" as il- 
lustrated by "trend lines" in variation 
diagrams. Studies of this kind are all 
too rarely accompanied by demonstra- 
tion that a proposed fractionation scheme 
could in fact produce the observed or 
inferred variation trends. Graphical 
methods have been employed to test the 
effects of adding or subtracting various 
combinations of mineral phases or xeno- 
liths, an excellent example being the 
study by Wilcox (1954) of the variation 
in the lavas of Paricutin volcano. Al- 
though feasible where the number of 
variables is small, the graphical method 
is slow, cumbersome, and subject to plot- 
ting and reading errors; with more than 
three variables it usually requires a 
series of trial and error approximations 
and often yields no useful result at all. 
In an effort to improve on the present 
state of the art, part of the work during 
the report year was devoted to the de- 
velopment of a computational procedure 
that requires no intermediate approxi- 
mations and processes up to ten variables 
simultaneously. In principle, any peno- 
logical problem that can be restated as a 
problem in linear combination can be 
formulated for this type of calculation. 
The method is illustrated here in three 



GEOPHYSICAL LABORATORY 



235 






calculations based on the Paricutin data 
of Wilcox. 

During the last part of the report 
year the Laboratory took delivery on 
a medium-speed data terminal which, 
when it is operating routinely, will trans- 
mit information at a rate making it prac- 
tical to resume pilot operations on stor- 
age and retrieval of rock analyses. This 
extremely productive work, results of 
which have been presented in numerous 
journal articles and summarized in Year 
Books 61-64, was brought to a tempor- 

cov (V if V,) 



the remaining space occupied by some 
one of the others. The null value against 
which- the observed correlation is to be 
tested, following the general approach 
outlined in Year Book 65, is the correla- 
tion expected between V 1 =Y t and Vj= 
Yj/il — Y-i) on the assumption that each 
Y has been formed by closure of 
a randomly drawn sample of X, whose 
elements are by definition open and un- 
corrected. The required approximate 
expectation, not available at the time of 
the earlier writing, is 



ViQfiTu —Vi<rf 



Vvar (70 -var (7,-) V[?iW+ d-2p 1 )a 1 2 ] [fcW + (l-2g j )a/] ' 



(1) 



ary halt when we lost ready access to a 
high-speed computation facility. The 
limited capacity of the new terminal 
forces us to a difficult decision that should 
have been made long ago. The entire 
library of analyses and all the longer 
programs must be held on tape at the 
central processor, reserving the terminal, 
as far as possible, for the transmission 
of commands and the printing of output. 
Our first major task will be preparation 
of programs for the generation and main- 
tenance of the tape containing the li- 
brary of rock analyses. A preliminary 
tape generator is already operating, but 
at the present writing the circuits that 
permit modification of an existing tape — 
i.e., the editing and updating functions 
that will be in frequent demand during 
reorganization and expansion of the 
punch-card libary — have not been fully 
tested. 

Identity of Expected Interdependence 

between Pairs of Niggli Numbers and 

Pairs of Analogous Remaining-Space 

Variables 

F. Chayes 

In the remaining-space transforma- 
tion (see Year Book 65, p. 376) atten- 
tion is directed to detecting nonrandom 
association of some one of a set of pro- 
portions, say Y 1} with the proportion of 



where o- fc 2 is the variance of the fcth ele- 
ment Of X, a t z = %{v k 2 ) , (rJ = (Tt 2 — o-i 2 , 
p k = X k /$(X), q k =Pk/(l-Pi), and ;>1. 
If the elements of Y are molar pro- 
portions and we denote Si0 2 by subscript 
1, the well-known Niggli numbers are 
sums of quantities of the form Nj—Yjf 
S(Ffc), k>l. For ;>1 any N t is thus a 
remaining-space variable, so that var 
(JV ; -)=var (Vj). N 1} however, is not 
simply y x under a new name, as was the 
case with V x , for in the Niggli transfor- 
mation the molar proportion of Si0 2 is 
also divided by the sum of the other 
molar proportions. Under this condition, 
and after considerable algebraic manipu- 
lation, the approximation procedure re- 
viewed in Year Book 65 leads to 



var (iVO = (1 -pO- 4 var (70, (2) 



and 



cov (A^O-U-pJ-'cov (V lf Vj). (3) 

Since (3) appears in the numerator and 
the square root of (2) in the denominator 
of pN-^Np it follows at once that pn 1 n s = 
P v x y. for ;>1, and thus that pn-n^pv-Vk 
for all j^=k. 

Given the null model basic to this 
discussion, the effects of the remaining- 
space and Niggli-number transforma- 
tions on interdependence generated by 
closure are identical, to the limits of ap- 
proximation. 



236 



CARNEGIE INSTITUTION 



Transformations Designed to Elimi- 
nate Negative Elements from Sample 
Estimates of the Vector of 
Open Variances 

F. Chayes 

The results of the preceding section 
would find wide application if, in situa- 
tions of practical petrographic interest, 
realistic estimates of the parameters of 
the hypothetical open parent could be 
obtained from the data. For one very 
common type of data this will nearly al- 
ways be impossible. In every Harker 
array so far examined, one or two of the 
calculated variances of the open varia- 
bles prove to be negative; the variance 
of the open equivalent of A1 2 3 is always 
one of these, and where there are two 
the second is that for the open equivalent 
of Na 2 0. As suggested earlier, this may 
reflect unsatisfactory or inadequate 
sampling, may indicate that the under- 
lying statistical model is unrealistic, or 
may be the consequence of an unwise 
choice of variables. In Year Book 66 the 
latter possibility was discussed at con- 
siderable length. It was argued that the 
oxides as reported by the analyst may 
indeed be an unwise "choice" of var- 
iables, and reasons were given for pre- 
ferring a number of sets of synthetic 
variables obtained as linear combina- 
tions of the original oxides. Seven such 
combinations were proposed, of which 
none completely eliminated negative ele- 
ments from the open variance vectors of 
the trial arrays. The transformation of 
Na 2 to ab was the most nearly success- 
ful, eliminating both negative elements 
in the open variance vectors of five of 
the six test arrays, but failing to elimi- 
nate a negative variance for the open 
equivalent of Na 2 in the sixth. Trans- 
formations of K 2 to or or of Na 2 
and K 2 to ab and or were ineffective. 
It has since been found that the com- 
bined transformation of Na 2 to ab 
and K 2 to KA10 2 is completely success- 
ful in the original test data (for identifi- 
cation of which see Year Book 66), as 



well as in a large number of other suites 
of analyses to which it has been applied. 
At the present writing the bearing of 
this finding on the work of the preceding 
section is not clear. At first glance, no 
further manipulation of the data would 
appear to be either necessary or de- 
sirable if, as appears to be the case, a 
reasonable choice of variables makes 
possible a direct test of the announced 
null hypothesis. 

Curve-Fitting in the Ternary 
Diagram 

F. Chayes 

The relation between the variables in 
a ternary diagram is necessarily one of 
mutual interdependence rather than the 
dependence-independence essential to a 
conceptually sound regression analysis. 
It is always possible to compute two 
(linear) regression lines from the data 
of a ternary array, but since no variable 
is independent there is no objective rea- 
son for preferring one of these lines to 
the other. This is not likely to be a 
serious handicap if the correlation is very 
strong. If the residual dispersion is more 
than trifling, however, the regression 
lines will be separated by a considerable 
angle, and even as a sample description 
the conventional regression calculation 
will be of little use. 

When the relation between two var- 
iables is one of interdependence the "re- 
duced major axis" (Kermack and Hal- 
dane, 1950) — also called the "diagonal 
line of organic correlation" (Kruskal, 
1953) — should be fitted in preference to 
the conventional regression line(s). This 
is a line passing through the mean of the 
distribution with a slope whose absolute 
value is the ratio of the standard devia- 
tions and whose sign is that of the co- 
variance. For any set of data there is 
only one line of organic correlation; it 
can be shown that it always lies in the 
acute angle between the regression lines, 
that it bisects this angle when the vari- 
ances are equal, and that it is the line 



GEOPHYSICAL LABORATORY 



237 



toward which the regression lines rotate 
with increase in correlation, means and 
variances remaining unchanged. (From 
these properties it is clear that when 
p 2 = l the line of organic correlation co- 
incides with the regression lines.) 

The probability of correctly predicting 
the value of either of two normally dis- 
tributed interdependent variables from 
a given value of the other is maximized 
if the predicting line is the diagonal line 
of organic correlation (Kruskal, 1953, 
p. 55). As a descriptor, this line has the 
intuitively appealing property of mini- 
mizing the sum of squares of deviations 
taken normal to it, a characteristic 
which, like the analogous minimizing 
property of the conventional regression 
line, is distribution free. 

In the absence of a tendency toward 
systematic departure from linearity in 
the data, the diagonal line of organic 
correlation would seem to be admirably 
suited for use in the ternary diagram. 
When there is a strong tendency toward 
curvilinear variation, however, the lack 
of an independent variable again makes 
trouble, for the only additional variables 
available are logs, roots, powers, prod- 
ucts or other transformations of those 
already given. There is no systematic 
way in which to decide which variable 
in a ternary array is to be chosen for 
this special treatment. 

A linearizing transformation might 
sometimes provide an escape from the 
dilemma; if desired, the line of organic 
correlation computed from the trans- 
formed data could be drawn in the 
original ternary. The rather gentle cur- 
vature characteristic of the lines in many 
published ternary diagrams may some- 
times be materially reduced by trans- 
formation from the original ternary co- 
ordinates, X 1} X 2 , and X 3 — l — X 1 — X 2 , 
to Y^X./iX.+X,) and Y i =X 1 /(X 1 + 
X s ), referred to below as the "binary 
ratios." The diagonal line of organic cor- 
relation is then 

Yj-yj^sgn. M—iYi-yi), (1) 



where sgn (o- i7 ) is the sign of the co- 
variance, y denotes a mean, and o- a 
standard deviation. A straight line in 
rectangular coordinates, equation 1 is in 
general nonlinear in the trilinear co- 
ordinate system.* Depending on the 
values of o-j, o-,-, 1/* and Vh m ^ ac t, its 
curvature in trilinear coordinates may 
be quite pronounced. 

The transformation to binary ratios is 
not to be regarded as a panacea which 
will automatically eliminate systematic 
departure from linearity in ternary 
closed data. Indeed, if the relation be- 
tween any two ternary variables is al- 
ready linear, then in general its trans- 
formation to binary ratios will not be 
linear in rectangular coordinates. The 
only exceptions to this rule are equations 
that plot as straight lines either parallel 
to an edge or terminating in an apex 
of the triangle, i.e., relations in which 
some one of the ternary variables or the 
ratio of some pair of them is constant. 
The need for a transformation only 
arises, however, when there seems to be 
a systematic departure from linearity 
in the ternary diagram. When non- 
linearity is apparent in the ternary dia- 
gram it is worth inquiring whether it 
can be reduced by transformation to 
binary ratios, for, in general, linear rela- 
tions between binary ratios will plot as 
curves in trilinear coordinates. 

In principle, a point may be located 
as readily in the ternary diagram by its 
binary ratios as by its trilinear co- 
ordinates. For any pair of values (Yi,Yj) 
the desired point is the intersection of 
the lines drawn from each marginal Y 
to the opposite apex of the triangle. If 
a desk calculator is available, however, 
it may be more convenient, and will 
usually be more accurate, to revert to the 
original coordinate system before plot- 
ting. 

*This is readily shown by substituting the 
trilinear definitions of Yi,Y } into (1), solving 
explicitly for Xi, and then differentiating; in 
general the derivatives of Xi with regard to X 2 
and X3 are not constants. 



238 



CARNEGIE INSTITUTION 



From the definitions of (Yi,Yj) it is 
evident that 



X 



<^h 



andX s 



Wh 



(2) 



Recalling that X 1 +X 2 +X a = l, we have 
at once that 



or 



X 1 =Y i Y j /(Y i +Y j -Y i Y j ). (3) 

Any X 1 found from Yi and Y, satisfying 
(1) is one of the trilinear coordinates of 
a point on the desired curve. For any 



such point a second coordinate may then 
be found by substitution in either of the 
definitions given in (2) . 

Figure 91(A) is a plot of Y t =Ca/ 
(Ca + Fe) against 7 ; - = Ca/(Ca + Mg) 
for the CaO-rich pyroxenes of the layered 
series of Skaergaard (data from Brown, 
1957, and Brown and Vincent, 1963). 
There is little doubt of the basically 
linear character of the variation; r y is 
-0.94 for the diopside "trend" and -0.99 
for the hedenbergite "trend." The equa- 
tions of the "diagonal lines of organic 
correlation" are: 

di; Y j= 51.88 -0.061 7* (4) 

hd) Yi= 64.21 -0.022 Y, (5) 




45 50 55 60 65 70 75 80 85 

Yi 

(A) 

Fig. 91. Calciferous pyroxenes of the layered series of Skaergaard. Circles, data points used 
in computing equation (4) ; squares, data points used in computing equation (5) ; triangle, 
datum point not used in either set of calculations. The coordinates are the binary ratios 
K« = (Ca/ + Fe), Y, = Ca/(Ca + Mg) in (A), and the ternary ratios Ca/(Ca + Mg + Fe), 
Fe/(Ca + Mg -f- Fe) in (B). The lines are traces of equations (4) and (5) in the two reference 
systems. 



GEOPHYSICAL LABORATORY 



239 



In Fig. 91(B) the data and these 
equations are plotted in the original 
ternary coordinates (Ca,Mg,Fe). In this 
reference system equations (4) and (5) 
are not straight lines; the curvature of 
(4) is trifling, but (5) closely follows 
the rather marked departure from 
linearity shown by the data. 

Except for the implication of a central 
discontinuity, the plots of equations (4) 
and (5) in Fig. 91(B) are virtually in- 
distinguishable from the line shown in 
one of the source references (Brown and 
Vincent, 1963, Fig. 3), so that the effect 
of the calculations in this case is essen- 
tially to "numericalize" a relationship 
already found by other means. This is a 
useful but not particularly profound re- 
sult and there are other, and perhaps 
simpler, ways in which it could have been 
reached. These other methods, however, 
are scarcely appropriate unless discrep- 
ancies from the proposed relations are 
trifling, for they provide no measure of 
the adequacy, as sample descriptions, 
of either the original graph or the equa- 
tion derived directly from it. If the pro- 
posed (or any) transformation does in- 
deed eliminate a tendency toward 
systematic departure from linearity, on 
the other hand, the quantity r i; - 2 has its 
usual meaning as a measure of the 
strength of the interdependence between 
the (transposed) sample statistics. 

A Last Look at Gl-Wl 

F. Chayes 

After a detailed initial test whose re- 
sults were summarized in U.S. Geological 
Survey Bulletin 980 (Fairbairn et al., 
1951), the U. S. Geological Survey gen- 
erously distributed, upon request, sam- 
ples of two now well-known reference 
materials, a granite from Westerly, 
Rhode Island (Gl), and a diabase from 
Centerville, Virginia (Wl). Bulletin 980 
reported 30 paired analyses of these ma- 
terials (i.e., analyses of both materials 
by the same analyst) , and an additional 
20 pairs were given in Bulletin 1113 



(Stevens et al., 1960) . Distribution of the 
samples continued for some time after 
the appearance of Bulletin 1113 but 
the supply of Gl is now exhausted; hence 
the title of this note. 

The ready accessibility of reference 
materials presumed to be reasonably 
uniform in composition proved so useful 
that similar distribution schemes were 
established elsewhere — notably in Can- 
ada and France — and our own Survey 
was led to materially expand its activity 
in this field.* From the beginning of 
the program there have nevertheless 
been persistent suggestions that the very 
property which made the reference ma- 
terials so useful, their alleged uniformity, 
was open to serious doubt. These doubts 
were thoroughly aired in the delibera- 
tions of the Standards Committee of 
the Geochemical Society, charged with 
obtaining a successor to Gl. There is no 
published record of a soundly designed 
experiment substantiating the charge of 
heterogeneity in either Gl or Wl, and 
in fact a small but well-designed experi- 
ment (c/. Flanagan in Stevens et al., 
1960) failed to establish significant het- 
erogeneity in the distribution of Pb. But 
Pb is present only in trace amounts, the 
experiment led to a sharp revision of 
the previously "preferred" values, and 
the residual fear of heterogeneity was 
sufficient to persuade the committee to 
overrule a suggestion that the successor 
of Gl should be coarse enough to permit 
preparation of reasonably pure mono- 
mineralic separates. (No one, to my 
knowledge, has successfully separated 
pure mineral fractions from Gl in more 
than micro amounts.) 

Now, of course, the possibility of het- 
erogeneity in any powdered material 
cannot be denied, and the temptation to 
ascribe embarrassingly large discrepan- 

* The number and variety of Survey speci- 
mens available for general distribution is con- 
tinually increasing. Interested analysts should 
address requests to the Liaison Officer, Analy- 
tical Laboratories, U. S. Geological Survey, 
Washington, D. C. 



240 



CAENEGIE INSTITUTION 



cies to this source is sometimes difficult 
to resist. Indeed, in connection with the 
work on Gl and Wl, the only alternative 
appears to be to accept what seem 
utterly unrealistic estimates of interlab- 
oratory differences. The structure of the 
published work on Gl and Wl is such 
that if one considers the results for 
either rock separately, the decision be- 
tween these alternatives must be made 
quite independently of the evidence. Con- 
sidering the results for Gl and Wl 
jointly, however, it is easy to show that 
there is indeed a very strong interlabora- 
tory effect, an effect so strong as to be 
readily perceptible despite the masking 
countereffects of random analytical error 
and the supposed sample heterogeneity. 
If there were no consistent interlabora- 
tory bias, one would expect only in- 
significant correlation between the result 
an analyst obtained for a particular 
oxide in Gl and the result he obtained 
for the same oxide in Wl. The correla- 
tions computed from the paired analyses 
published in Bulletins 980 and 1113 are 
shown in Table 25. In the data from 
Bulletin 980 the correlations for Si0 2 , 
MgO, Na 2 0, and K 2 are easily sig- 
nificant at the 0.01 point and those for 
A1 2 3 and FeO at the 0.05 point. The 
correlations are by no means strong 
enough so that an analyst's result for a 
particular oxide in one rock can be pre- 
cisely predicted from his result for the 

TABLE 25. Correlation between Results for 
Essential Constituents of Gl and Wl t 



Oxide 


A 


B 


C 


Si0 2 


0.724** 


0.725** 


0.545 


A1 2 3 


0.444* 


0.579** 


0.304 


Fe 2 O s 


0.250 


0.243 


0.352 


FeO 


0.430* 


0.428* 


0.515 


MgO 


0.708** 


0.720** 


0.198 


CaO 


0.136 


0.506** 


0.156 


Na B 


0.685** 


0.684** 


0.459 


K 2 


0.568** 


0.706** 


0.655 



t A, for 30 pairs from U. S. Geol. Surv. Bull. 
980; B, for data of Bull. 980 less pair no. 26; 
C, for 20 pairs from Bull. 1113. Against the 
alternative that p = 0, * indicates significance 
at the 0.05 and ** at the 0.01 level. 



same oxide in the other rock. In general, 
however, a man reporting appreciably 
high (or low) values for Si0 2 , MgO, 
Na 2 0, K 2 0, FeO, or A1 2 3 in either rock 
usually reported high (or low) values for 
these oxides in the other rock as well; 
indeed, something between a quarter and 
a half of the total variation exhibited 
by these oxides may be "accounted for" 
in this fashion. 

Even in a sample of 30 pairs, the effect 
of occasional extreme outliers on calcu- 
lations of this sort may be considerable. 
It is therefore of some interest that dele- 
tion of one pair rendered suspect by an 
A1 2 3 value differing from the group 
mean by more than 4a materially 
strengthens the correlations for A1 2 3 , 
CaO, and K 2 and does not reduce the 
other correlations by more than trifling 
amounts; the results are shown in 
column B of the table. Clearly the inter- 
dependence of the sheep is not to be at- 
tributed to the presence of an occasional 
goat. Example may be more persuasive 
than argument, but the result of this re- 
calculation is not to be considered par- 
ticularly unexpected. Bona fide incompe- 
tence should not in general lead to sys- 
tematic differences of the kind in ques- 
tion; we are almost certainly concerned 
here with unsuspected differences be- 
tween reasonably competent analysts 
whom the uninitiate, at least, presume to 
be using reasonably standardized tech- 
niques. 

Does the final column of the table sug- 
gest substantial improvement in the 
course of the next 10 years? The shock- 
ingly large correlations are no longer in 
evidence; against the hypothesis that 
P = 0, only one correlation in column C 
exceeds the 0.01 point and only three 
others exceed the 0.05 point. But it hardly 
seems legitimate to test the entries in 
column C against this hypothesis. The 
work summarized in column C was done 
in the shadow of that summarized in 
column A. The meaningful query is not 
whether an entry in column C differs 



GEOPHYSICAL LABORATORY 



241 



significantly from zero but whether it 
differs significantly from the analogous 
entry in column A or B. The only entry 
in column C that is significantly smaller 
than its analogue in column A is that for 
MgO. For Si0 2 , A1 2 3 , FeO, Na 2 0, and 
K 2 no significant improvement is in- 
dicated. The combination of random 
analytical error and sampling variation 
is insufficient to obscure a powerful inter- 
laboratory bias affecting these con- 
stituents. 

The bearing of this discussion on the 
question of sample inhomogeneity is evi- 
dent. The 50 paired analyses so far 
published suggest a quite different ex- 
planation for the observed dispersion 
and in fact contain no direct or useful 
information about sample inhomogeneity. 
If we want to know whether powdered 
rock samples are sufficiently uniform in 
composition for use in interlaboratory 
calibration, we shall have to conduct an 
experiment designed to answer this ques- 
tion. In such an experiment each of a 
group of analysts would analyze each — 
or a considerable number — of a series of 
(supposedly identical) powders. Given 
appropriate prior agreement about a 
common set of analytical procedures to 
be followed throughout the experiment, 
it would then be possible to isolate a sum 
of squares associated with differences be- 
tween analysts and another generated by 
differences between powders. The mean 
square for discrepance would provide a 
sound estimate of random analytical 
error — now quite unknown — against 
which the significance of the nonrandom 
effects could be tested. 

Although it is too late to conduct this 
experiment with Gl, the work on Gl 
and Wl makes it abundantly clear that 
much could be learned in this fashion. 
The increasing popularity of powdered 
reference materials makes a careful par- 
tition of analytical variance into sample, 
systematic, and random components 
virtually indispensable. 



Materials Balance in Igneous Rock 

Suites 

W. B. Bryan 

In order to improve upon the tradi- 
tional graphical methods of testing ef- 
fects of addition and subtraction of 
mineral phases in igneous rock suites, it 
is reasonable to seek a computational 
procedure which is a numerical equiva- 
lent of the graphical method. One thinks 
naturally of expressing the composition 
vectors as linear equations, in which the 
known values are the weight percentages 
of the individual cation oxides in each 
mineral or rock analysis, and the un- 
knowns are the weight fractions of each 
mineral or rock to be combined in order 
to obtain some given composition. 

An appropriate working model can be 
formulated from consideration of the 
nature of the data that usually emerge 
from the study of an intrusive igneous 
rock suite. Field and petrographic studies 
represent a search for all the physically 
independent phases that may have been 
added to or subtracted from the original 
parent magma, or that remain as liquid 
residua after successive modifications 
of the parent magma. The resulting col- 
lection of phases represents a complete 
system of variables that is closed with 
respect to composition. An igneous intru- 
sion that has solidified without reaction 
with the enclosing rock is the simplest 
example of such a closed system, its bulk 
composition being identical to that of 
the parent magma. Individual minerals 
and rocks contained within this intrusion 
will differ appreciably in composition 
from the parent liquid, but the chemical 
components of the parent magma must 
be distributed in such a way that the 
sum of the products of the weight frac- 
tion of each phase and its composition 
equals the composition of the parent 
magma. This will seem intuitively ob- 
vious to petrographers who have calcu- 
lated chemical compositions of rocks 
from their modes. If the magma has as- 
similated a portion of the enclosing coun- 



242 



CAKNEGIE INSTITUTION 



try rock, it will be necessary to extend 
the system to include the country rock, 
and the weight fraction of country rock 
assimilated must be subtracted from the 
summation to arrive at the parent 
magma composition. 

It is evident then that problems of this 
sort have associated with them an im- 
plicit materials balance that may be 
used as a check on the genetic model in- 



inferred from the text discussion, it is 
possible to estimate the weight fractions 
from the compositional data. 

The procedure is best summarized in 
matrix notation, where a y , for i= 1, k and 
; = 1, n, is the weight percentage of oxide 
i in phase ;', Xj is the unknown weight 
fraction of each phase, and p f is the 
weight percentage of oxide i in the as- 
sumed parent composition. Then, 



^11 ^12 ^1 

0*21 ^22 ^2 



dm 

a 3n 



a kl 



die* 



ttfc3 



Gfcn 





x x 




Pi 




%2 




V2 




%3 


= 


Ps 




x n 




Vk 

_ — 1 



(1) 



ferred from field and petrographic data. 
Thus, one might proceed by multiplying 
the composition of each rock or mineral 
phase by its measured weight fraction 
with appropriate regard to sign, summing 
the products, and comparing the result 
with the parent-magma composition. A 
close correspondence between the two 
would be expected if all phases have 
been accounted for, if their compositions 
are accurately determined, if the rock 
representing the parent magma has been 
correctly specified, and if the correct 
amount of each phase has been added 
into or subtracted from the system ac- 
cording to a properly conceived peno- 
logical hypothesis. But it may be much 
easier to determine the composition of a 
rock or mineral than to measure its abso- 
lute quantity in volume or weight per- 
cent. The exact size and shape of rock 
bodies are often obscured by vagaries 
of erosion and exposure, and modal 
analysis of thin sections becomes im- 
practical for rocks that are either very 
coarse or very fine in grain size. The 
density measurements necessary to con- 
vert volume to weight fractions are also 
frequently subject to serious difficulties. 
Where the composition of the presumed 
parent is known or can reasonably be 



Ordinarily the subscript k will be not 
more than 10, the usual number of essen- 
tial elements reported in a standard 
chemical analysis, and n is much less 
than 10, as one would indeed expect from 
phase-rule considerations. Barring singu- 
larity, which should be very uncommon, 
any n equations selected from (1) will 
yield a solution for the re's. The final 
step in the calculation is to sum for 
each unused row, i, the products (a^-rc,), 
to yield an estimate, say p t . It is then a 
matter of judgment whether the differ- 
ences [pi — Pi] are small enough to justify 
belief that all the positively signed 
phases were in fact derived from the 
parent. One would also expect the sign 
and magnitude of the coefficients to agree 
with field or petrographic data such as 
estimated rock volumes or modal pro- 
portions, with a negative sign implying 
removal of a phase to obtain the desired 
parent composition. 

Table 26 gives the results of three cal- 
culations based on the data for Paricutin 
(Wilcox, 1954). Andesite P-10 was used 
by Wilcox as a starting composition 
from which to test other members of the 
series, and Si0 2 , MgO, and Al 2 O s were 
the "oxide controls" he selected for the 
graphical solution. These and FeO were 



GEOPHYSICAL LABOKATORY 



243 



TABLE 26. Paricutin Volcano, Mexico: Calculated and Observed 
Compositions, and Fractions of Inferred Components 





P-l 




P-15 


P-21 




1 


2 


1 2 


1 2 


SiO* 


55.00* 


55.00 


58.63* 58.63 


60.50* 60.50 


A1 2 3 


18.81* 


18.81 


17.85* 17.85 


17.30* 17.30 


FeO 


7.41* 


7.41 


6.22* 6.22 


5.66* 5.66 


MgO 


5.68* 


5.68 


4.05* 4.05 


3.60* 3.60 


CaO 


8.04 


7.16 


6.76 6.78 


6.05 6.17 


Na 2 


4.03 


3.88 


4.06 3.88 


4.02 3.90 


K 2 


0.80 


0.85 


1.42 1.31 


1.73 1.69 


MnO 


0.13 


0.07 


0.12 0.12 


0.11 0.10 


Ti0 2 


1.12 


0.94 


0.98 0.86 


0.89 0.80 


P 2 Os 


0.36 


021 


0.32 0.30 


028 0.28 


Totals 


101.38 


100.01 


100.41 100.00 


100.14 100.00 


Andesite P-10 


1.2538 


1.0480 


0.9066 


Average xenolith — 0.1954 


0.0399 


0.1726 


Plagioclase An 


70 —0.0149 


—0.0387 


—0.0384 


Olivine Foso 


— 0.02< 


)7 
18 


—0.0452 


—0.0394 


Totals 


1.01J 


1.0040 


1.0014 



* Oxide used as control. 

1. Calculated composition. 

2. Observed composition, recalculated to 100% minus CO2, H 2 0, all iron 
as FeO. Data from Wilcox (1954, Table 2, columns 1, 15, and 21). 



also used for the numerical solution, 
which requires four equations. Calcula- 
tions are presented only for analyses 1, 
15, and 21, which adequately represent 
the full range of compositions erupted 
at Paricutin. However, by repeating the 
calculation for each analysis in turn 
entered into the final column of the aug- 
mented matrix, it is possible to simulate 
the complete set of variation curves. As 
in the original paper, all data were re- 
calculated anhydrous, with total Fe ex- 
pressed as FeO. 

Using only the major cation oxides, 
Wilcox (1954) was able to show that 
simultaneous addition of xenolith ma- 
terial and removal of plagioclase and 
olivine appearing as phenocrysts could 
duplicate the bulk variation trends of the 
Paricutin lavas. The computer-generated 
compositions (Table 26) based on this 
hypothesis are in excellent agreement 
with the bulk analyses of the Paricutin 
lavas, even when extended to alkalies and 
minor elements not included in the 
original graphical solution. The some- 
what less satisfactory agreement at the 



low-silica end of the series, represented 
by analysis 1, was also noted by Wilcox 
(1954) and is illustrated in his Fig. 106. 

A Least-Squares Approximation for 

Estimating the Composition of a 

Mixture 

W. B. Bryan, L. W. Finger, and F. Chayes 

The procedure described in the pre- 
ceding section is an exact numerical 
analogue of graphical devices widely em- 
ployed by petrographers. Much quicker 
and more precise than the graphical 
procedures, it is nevertheless subject to 
the same general restrictions. Chief 
among these, in the notation used on 
page 242, is the circumstance that if fc<n 
there is no solution at all, whereas if 
&>n there are in principle k\/n\(k — n) ! 
solutions, each of which uses only n/k 
of the data. If, for instance, k = 9 and 
n = 5, each of the 126 different sets of 
n simultaneous equations that may be 
drawn from (1) of page 242 yields either 
a solution or a singular matrix. If some 
of these sets are soluble and others are 



244 



CARNEGIE INSTITUTION 



not, there would appear to be no rational 
interpretation of the result, but even in 
the absence of singularity the situation 
may sometimes be little better. If the 
entire body of data is worth having in 
the first place, one feels intuitively that 
a single solution using all the informa- 
tion would be preferable to a large num- 
ber of solutions each of which ignores 
(1 — n/k) of the information. 

Now such a solution is in fact avail- 
able if, in the fashion long adopted in 
other fields of inquiry (see, for instance, 
Legendre, 1805), one is prepared to ac- 
cept as preferable that result which 
minimizes the sum of squares of the 
residuals. Letting A and X denote, re- 
spectively, the matrix and vector on the 
left side, and P the vector on the right 
side, of (1), the equations of condition 
may be restated as 



AX = P. 



(2) 



Premultiplication of (2) by the trans- 
pose of A then gives the normal equations 



A'AX=A'P, 



(3) 



and, finally, premultiplication of (3) by 
the inverse of [A'A], yields the desired 
solution, 



X=[A'A]- 1 A'P. 



(4) 



There is only one such solution, and all 
of the available information contributes 
to it. 

If we now substitute X from (4) into 
the left side of (2) and perform the 
indicated multiplication, we obtain a 
vector, say, P, each element of which is 
the least-squares estimate of the analo- 
gous element of P, i.e., the solution ob- 
tained from (4) is the one which mini- 
mizes the quantity %{]?§— VjY- This solu- 

;" - 
tion exists when k>n, the inequality 

denoting the common situation, in which, 
incidentally, graphical procedures and 
their numerical analogues fail to yield 
unique solutions. 



CHEMICAL AND OPTICAL 
PETROGRAPHY 

W. B. Bryan 

During the report year, work has 
centered on the petrography and min- 
eralogy of basalts and associated vol- 
canic and intrusive rocks from the 
Bunya Mountains and Carnarvon Range, 
Australia. The Bunya Mountains lavas, 
for which chemical and mineralogical 
data are being prepared for later publica- 
tion, illustrate well the compositional 
characteristics and resulting problems 
of classification that appear typical of 
many basaltic lavas from eastern 
Australia. Although intermediate compo- 
sitions appear to be as rare among the 
volcanic rocks of eastern Australia as 
among the lavas of oceanic islands sum- 
marized by Chayes (1963), the six 
analyzed samples from the Bunya Moun- 
tains discussed below include one sample 
of trachyandesite (benmoreite) , indicat- 
ing that liquids of this composition are 
not entirely lacking among the volcanic 
rocks of southeast Queensland. 

The extrusive rocks from the Carnar- 
von Range are discussed only briefly 
here, as the data are being prepared for 
publication in the near future. In spite 
of an apparently extreme compositional 
hiatus between basalt and trachyte in this 
area, these two compositions are inti- 
mately associated in a composite olivine 
gabbro-microsyenite intrusion, suggest- 
ing a close genetic relation between 
basaltic and trachytic magmas. The oc- 
currence provides a rare opportunity to 
examine mineral paragenesis and crystal- 
melt equilibria relevant to the basalt- 
trachyte problem. 

Volcanic Rocks from the Bunya 
Mountains, Queensland, Australia 

The Bunya Mountains form a high, 
domelike plateau located about 80 miles 
northwest of Brisbane, in the southeast 
corner of Queensland, Australia. Samples 
from the Bunya Mountains were not 
included in the pioneering work of 



GEOPHYSICAL LABORATORY 



245 



Richards (1916) on the volcanic rocks 
of southeast Queensland, and although 
the mountains are readily accessible they 
have continued to escape the attention 
of petrographers. The six analyzed sam- 
ples discussed here are the first to be 
described from this area, which appears 
to be a distinct volcanic center, prob- 
ably the remnant of a once more ex- 
tensive shield volcano. The rocks are of 
special interest because the basalts fall 
in the composition field termed olivine 
tholeiite by Yoder and Tilley (1962, p. 
352) and lie close to the quadratic dis- 
criminant function proposed by Chayes 



(1966, pp. 137-144) for distinguishing 
between alkaline and subalkaline basalts 
in the normative diopside-olivine-hy- 
persthene ternary (Fig. 92). As demon- 
strated previously by Chayes (Year 
Book 64, p. 155), basalt analyses yield- 
ing neither normative quartz nor norma- 
tive nepheline are relatively uncommon, 
although the majority of those recently 
dredged from the submarine ridges are 
in this category. Basalts of this type are 
evidently also common in eastern Aus- 
tralia. Similar rocks from the Mt. Warn- 
ing shield volcano south of Brisbane were 
called "subtholeiitic" by Green (1964) 




Basalt, Bunya Mountains 

Trachyandesite, Bunya Mountains 

Trachybasalt, Bunya Mountains 

Basalts from the Carnarvon Range 

A, Average Alkali Basalt |,_ , _ 

-r A — , .. • n i KEngel, Engel, and Havens, 1965) 

T, Average Tholentic Basalt i ' 

G, Intrusive Gabbro, Carnarvon Range 

D-D' Second-degree discriminant function (Chayes, 1966) 



Fig. 92. Lavas from the Bunya Mountains, Carnarvon Range, and averages of oceanic alkaline 
and tholeiitic basalt, projected in the ternary fields ne-ol-di, ol-di-hy, and hy-di-Q. Two strongly 
oxidized Carnarvon basalts, included in Figs. 93 and 94. are excluded from this figure. 



246 



CARNEGIE INSTITUTION 



and more recently, "transitional" (Green, 
1968) ; and McDougall and Wilkinson 
(1967, pp. 229-230) have described 
others from northern New South Wales 
as "mildly alkaline," "transitional," 
and "mildly tholeiitic." It will be evident 
from Fig. 92 that basaltic rocks from 
the Carnarvon Range, Queensland 
(Bryan, 1968), are also in this category. 
Other graphical plots that have been 
used to distinguish alkaline and tholeiitic 
basalts confirm the ambiguous nature of 
these rocks. Figure 93 is essentially that 
used by Yoder and Tilley (Year Book 
64, p. 75), except that total iron is ex- 
pressed as FeO. The Bunya Mountains 
lavas show the initial iron-enrichment 
trend regarded as typical of tholeiitic 
suites, but the more siliceous rocks revert 
to the iron-depletion trend typical of 
alkaline lavas. The Carnarvon lavas 
show the same initial iron-enrichment 
trend, but the data do not extend over 
a sufficient composition range to indicate 
whether reversal to an alkaline trend is 
to be expected. In the familiar alkali- 
silica plot (Fig. 94), with the alkali 



1 


i — r 


— 1 1 1 ' 1 1 






• 


- 


T ® 


G ® . 


- 






"®A 










/ 










»■— 


^' 


1 


1 1 


i i — i — i — i — 



10 
9 
8 

7 

O 6 

en 

2 5 

4 
3 
2 



6 7 8 9 10 II 12 13 14 15 

Total Fe as FeO 

Fig. 93. Variation of MgO versus total iron 
as FeO. All symbols as in Fig. 92. Dashed line 
connects lavas of Bunya Mountains. 

basalt-tholeiite boundary as defined by 
Macdonald and Katsura (1964), the 
Queensland basalts tend to fall within 
the alkaline field between the averages 




Fig. 94. Variation of (Na 2 + K 2 0) versus Si0 2 . All symbols as in Fig. 92; dashed line H-H' 
is boundary between alkali basalt field (upper) and tholeiite field (lower) as defined by 
Macdonald and Katsura (1964). 



GEOPHYSICAL LABORATORY 



247 



of oceanic alkaline and tholeiitic basalt, 
with the trachybasalt and trachyandesite 
also in the alkaline field at higher silica 
percentages. 

The basalts contain phenocrysts of 
olivine (Fo 65 _ 70 ) and, in the groundmass, 
pink augite with optical properties sug- 
gesting a composition approximately Ca 35 
Mg 3 oFe 3 5, minerals generally regarded 
as typical of alkaline basalts. The glassy 
groundmass of the trachybasalt contains 
a few granules of olivine but no pyrox- 
ene. The trachyandesite, the most alka- 
line rock in Figs. 93 and 94, contains two 
pyroxenes, a hypersthene pleochroic in 
pale pink and green, and a pale green 
augite zoned outward to a rim of dark 
green. Optical properties of the latter 
mineral suggest ferroaugite zoned to 
ferrohedenbergite rather than a sodic 
pyroxene. These pyroxenes are suggestive 
of a tholeiitic rather than an alkaline 
affinity, although ferrohedenbergite is 
common in supposedly alkaline trachytes 
and rhyolites. 

Analyses projecting well toward the 
diopside-hypersthene boundary or into 
the quartz field of Fig. 92 are perhaps 
best regarded as siliceous derivatives of 
the more mafic basalts. One of the olivine 
basalts falls in this region, owing to a 
relatively high content of ferric iron and 
low magnesia. The trachybasalt is simi- 
lar in composition to the type tholeiite 
from the Tholey sill (Jung, 1958), as 
are equivalent rocks from New South 
Wales described as "potassic tholeiitic 
andesite" by McDougall and Wilkinson 
(1967, p. 230). The trachyandesite 
analysis resembles closely the sodic in- 
termediate rocks called benmoreite by 
Tilley and Muir (1964). In view of the 
interest in analyzed rocks of intermediate 
composition since their scarcity in the 
ocean basins was publicized by Chayes 
(1963), it must be emphasized that this 
specimen is the most leucocratic of the 
nineteen basaltic rocks collected in a 
traverse of the Bunya Mountains, during 
which special effort was made to include 
rocks thought to be of intermediate com- 



position. Equivalent rocks were not dis- 
covered on a traverse of the southern 
end of the Carnarvon Range, nor were 
they found by Green (1964) in a 
thorough sampling of part of the Mt. 
Warning shield volcano south of Bris- 
bane. Rocks called mugearite were in- 
cluded in a summary of volcanic rocks 
from the Main Range, southwest of Bris- 
bane (Stevens, 1965), where they are ap- 
parently common, though volumetrically 
subordinate to the basalts (Stevens, 1966, 
personal communication) . 

An Olivine Gabbro-Microsyenite In- 
trusion from the Carnarvon Range, 
Queensland, Australia 

An unusual volcanic neck, located 
just north of the divide between the 
headwaters of Marlong Creek and Me- 
teor Creek in the southern part of the 
Carnarvon Range, provides good geo- 
logic evidence of a genetic relation be- 
tween basaltic and trachytic magma. 
The neck is circular in plan, about % 
mile in diameter, and consists of a 
coarsely crystallized gabbro core sur- 
rounded by a ring-dike of microsyenite. 
Geologic and topographic relations in- 
dicate that the neck was intrusive into 
at least the lower half of the sheet of 
basaltic lavas capping the adjacent 
peaks. The gabbro is similar in bulk 
chemical composition to the uppermost 
lavas of the sheet, and it seems likely 
that the neck originally served as a con- 
duit for these lavas. The gabbro evi- 
dently pulled away from the steeply 
dipping sides of the conduit shortly after 
consolidation, and settled downward 
into the underlying partly consolidated 
magma, allowing residual trachytic liq- 
uid generated at depth to rise into the 
circular void thus created. 

The gabbro is composed of plagio- 
clase (An 59 ) zoned outward through nar- 
row rims of oligoclase to overgrowths of 
alkali feldspar, the crystals being 3 to 
6 mm in length. Olivine (Fo 53 ) occurs 
as rounded grains, 1 to 2 mm in diameter; 



248 



CARNEGIE INSTITUTION 



both olivine and plagioclase are ophiti- 
cally enclosed by large patches of opti- 
cally continuous pink augite. Minor 
constituents include ilmenite, apatite, 
and rare flakes of red-brown biotite. 
Quartz appears in small areas of micro- 
pegmatite interstitial to the alkali feld- 
spar overgrowths. An iddingsitelike 
brown alteration product penetrates 
fractures in many olivine crystals. The 
microsyenite is composed primarily of 
alkali feldspar, with conspicuous rounded 
phenocrysts of anorthoclase, 1 to 2 mm 
in diameter, and scattered micropheno- 
crysts of ferrohedenbergite. The ground- 
mass consists of flow-oriented laths of 
alkali feldspar with interstitial granules 
of magnetite, ferrohedenbergite, and 
cryptocrystalline feldspar. There are 
rounded brown iddingsitelike pseudo- 
morphs, 0.5 to 1.0 mm in diameter, which 
may originally have been olivine. Many 
of the anorthoclase phenocrysts show 
patchy extinction, traces of perthitic 
lamellae, and grid twinning, and have 
a few tiny inclusions of ferrohedenberg- 
ite. The chemical compositions of the 
rocks and of some major mineral phases 
are given in Tables 27 and 28. The esti- 

TABLE 27. Chemical Composition of Gabbro 

and Microsyenite, Carnarvon Range, 

Queensland 





1 


2 




la 


2a 


Si0 2 


49.19 


63.94 


Q 




8.07 


Ti0 2 


1.87 


020 


Or 


6.86 


31.50 


A1 2 3 


16.83 


16.74 


Ab 


29.87 


45.10 


Fe 2 3 


1.72 


2.67 


An 


26.65 


6.01 


FeO 


8.66 


2.41 


Di 


10.97 


2.17 


MnO 


0.14 


0.12 


Hy 


3.79 


1.59 


MgO 


6.30 


054 


01 


13.29 




CaO 


8.45 


1.78 


Mt 


2.49 


3.87 


Na 2 


3.53 


5.33 


11 


3.55 


0.38 


K 2 


1.16 


5.33 


Ap 


0.70 


0.11 


P 2 O s 


0.32 


0.05 








H 2 + 


1.45 


0.26 








H 2 0" 


0.51 


1.18 








Totals 


100.13 


10025 




D 


2.899 


2.657 









1. Olivine gabbro C16-64; J. R. Dickson, 
analyst. 

la. Normative composition of olivine gabbro. 

2. Microsyenite C17-64; J. R. Dickson, 
analyst. 

2a. Normative composition of microsyenite. 



mated composition of olivine is Fo 53 
(D = 3.737, a = 1.725-1.727, d 130 = 2.802). 
Ilmenite was identified in polished sec- 
tion and by X-ray diffraction; no mag- 
netite was detected optically or in the 
diffraction pattern. Apatite is evidently 
a fluorapatite (o> = 1.638-1 .645). 

Thin sections of the gabbro provide 
the usual, not entirely unambiguous evi- 
dence for the sequence of crystallization. 
Olivine may be enclosed by all minerals, 
although it tends to be crowded into 
areas between the large plagioclase laths 
rather than to be enclosed by them, sug- 
gesting that the crystallization of plagio- 
clase began later than that of olivine. 
Ilmenite is interstitial to plagioclase and 
olivine, but tends to be intergrown with 
augite, indicating essentially simultane- 
ous crystallization of these minerals. 
Apatite is imbedded in the outer zones of 
plagioclase but is especially conspicuous 
in the alkali feldspar overgrowths, with 
which it is evidently in part contempor- 
ary. Such overgrowths are present only 
on those portions of plagioclase laths 
not ophitically enclosed by augite. 
Quartz occurs only as intergrowths with 
alkali feldspar interstitial to plagioclase. 
Carbonate, traces of biotite, and margi- 
nal replacement of olivine by hypers- 
thene or "iddingsite" are also restricted 
to these areas. 

The calculated mineral proportions of 
the microsyenite are given in Table 29. 
Table 30, column A, shows the measured 
modal proportions of the gabbro with the 
chemical analysis recalculated anhy- 
drous, and column B shows the modal 
minerals expressed as weight percentages, 
with the corresponding calculated chemi- 
cal composition as a check on the modal 
mineral data. In column C, Table 30, 
alkali feldspar is replaced by the micro- 
syenite, with the equivalent amount of 
plagioclase, augite, and iron-titanium 
oxide in the microsyenite subtracted 
from the modally determined weight 
fractions. The corresponding chemical 
composition and density calculated from 
this adjusted mode agree well with those 



GEOPHYSICAL LABORATORY 



249 



TABLE 28. Chemical Composition of Minerals from Gabbro and Microsyenite 





1 


2 


3 


4 


5 


6 


7 


Si0 2 


51.12 


48.00 


66.45 


66.36 


54.70 






Ti0 2 


0.72 


0.31 


nil 






48.86 


1422 


AI2O3 


3.04 


126 


19.06 


19.37 


28.8i 






Fe 2 3 


0.44 


2.97 


0.52 






3.79 


41.40 


FeO 


8.66 


25.59 


0.43 






47.36 


44.30 


MnO 


0.22 


0.59 


nil 










MgO 


1522 


1.40 


0.10 










CaO 


2026 


18.16 


0.55 


0.50 


11.66 






Na 2 


0.18 


1.17 


6.80 


6.35 


5.08 






K 2 


tr. 


0.34 


5.77 


7.41 


0.40 






P 2 5 


0.06 


n.d. 


n.d. 










H 2 + 
H 2 0" 


0.15 
0.10 


020 


0.10 










Totals 


100.17 


99.99 


99.78 


99.99 


99.99 


100.01 


99.92 


a 


1.689 


1.738 


1.525 




1.558-1.560 






j8 


1.701 


1.746 


1.530 










7 


1.720 


1.765 


1.533 










27 


(+)49° 


(+)55°±5 


(-)57° 










D 


3.276 


3.639 


2.572 


<2.570 


>2.682 


4.736 


5.03 



1. Augite, olivine gabbro C16-64; J. R. Dickson, analyst. 

2. Ferrohedenbergite, microsyenite C17-64; L. J. Sutherland, analyst . 

3. Anorthoclase, microsyenite C17-64; L. J. Sutherland, analyst. 

4. Alkali feldspar, olivine gabbro C16-64, calculated from Na,K,Ca analysis by electron micro- 
probe; Si0 2 and ALO3 by difference, assuming ideal formula. 

5. Plagioclase feldspar (cores), olivine gabbro C16-64, calculated from Na,K,Ca analysis by 
electron microprobe; Si0 2 and A1 2 3 by difference, assuming ideal formula. 

6. Ilmenite, olivine gabbro C16-64, calculated from Fe,Ti analysis by electron microprobe, 
assuming ideal ilmenite-hematite solid solution. 

7. Titaniferous magnetite, microsyenite C17-64, calculated from Fe,Ti analysis by electron micro- 
probe, assuming ideal magnetite-ulvospinel solid solution. Density calculated for 60% magnetite, 
D = 520; 40% ulvospinel, D = 4.78. 



TABLE 29. Calculated Modal Proportions and Chemical Analysis of 
Microsyenite 



Volume % 


Weight % 




A 


B 


Anorthoclase* 91.08 


88.15 


SiO a 


64.71 


64.71 


Ferrohedenbergite * 4.97 


6.80 


ALOs 


16.94 


16.94 


Quartz t 2.96 


2.68 


Fe 2 O s 


2.17 


2.70 


Apatite t 0.10 


0.12 


FeO 


2.68 


2.44 


Magnetite t 0.67 


127 


MgO 


0.19 


024 


Hematite § 0.50 


0.98 


CaO 


1.79 


150 






Na 2 


6.09 


5.39 






K 2 


5.12 


5.39 






Ti0 2 


020 


020 






P 2 5 


0.05 


0.05 






MnO 
Totals 


0.04 


0.12 




99.98 


100.00 






D 


2.658 


2.657 



* Weight fraction computed from simultaneous solution for Al 2 Os and 
CaO, after subtracting CaO in apatite, and using microsyenite analysis 
recalculated anhydrous. 

t Excess SiOa after subtracting anorthoclase and ferrohedenbergite ; ideal 
formula and density assumed. 

J Weight fractions estimated from P 2 Os and Ti0 2 in bulk analysis; ideal 
formula and density assumed. 

§ Weight fraction is that required to bring sum of weight fractions to 
1.000; ideal formula and density assumed. 

A, Composition and density calculated from mode. 

B, Actual analysis recalculated anhydrous, and measured density. 



250 



CARNEGIE INSTITUTION 



TABLE 30. Measured and Adjusted Mode of Gabbro C16-64, with 

Corresponding Observed and Calculated Chemical Composition 

and Absolute Density * 





A 


B 


C 
Adjusted 




Modal 


Modal 


Modal 




Volume % 


Weight % 


Weight % 


Alkali feldspar and 








micropegmatite 


17.07 


14.92 




Microsyenite 






18.49 


Plagioclase 


52.34 


47.76 


45.84 


Augite 


15.01 


16.93 


15.68 


Olivine 


10.72 


13.79 


13.79 


Ilmenite 


2.15 


3.51 


3.11 


Iddingsite t 


2.70 


2.32 


2.32 


Apatite $ 




0.76 


0.72 


SiO a 


50.11 


50.23 


50.61 


AU03 


17.14 


17.34 


16.99 


Fe 2 3 


1.75 


1.05 


1.53 


FeO 


8.82 


8.72 


8.87 


MgO 


6.42 


6.46 


6.32 


CaO 


8.61 


9.23 


9.02 


Na 2 


3.60 


3.41 


3.36 


K 2 


1.18 


1.30 


1.19 


TiOa 


1.90 


1.S5 


1.69 


P2O5 


0.33 


0.34 


0.35 


MnO 


0.14 


0.04 


0.06 


Totals 


100.00 


99.97 


99.99 


Absolute density 


2.899 


2.904§ 


2.913§ 



* All chemical analyses used in calculations have been recalculated to 
100%, minus H 2 0. 

t Composition from Gay and LeMaitre (1961, p. 106, Table 3, column 5). 

% Ideal formula and density of 3.21 assumed, weight fraction inferred 
from P2O5 in bulk rock analysis, and an equivalent weight subtracted from 
the feldspars. 

§ Calculated. 

A, Measured modal volume % (average of three thin sections), chemical 
analysis of gabbro recalculated to 100% anhydrous, and measured absolute 
density. 

B, Mineral weight percentages and bulk chemical composition calculated 
from modal data. 

C, Adjusted weight percentages, substituting microsyenite for alkali feld- 
spar, and calculated bulk chemical composition. 



of the gabbro. The compositions and 
modal proportions of the minerals of the 
gabbro are thus compatible with geo- 
logical evidence suggesting derivation of 
the microsyenite liquid from the gabbro. 
The residual liquid would apparently 
have had the appropriate composition 
when the gabbro was 81,5 wt % crystal- 
lized. It is interesting that the liquid 
remaining after more than 80% crystal- 
lization of the gabbro is only slightly 
oversaturated with respect to silica 
(Table 27, column 2a). K slight further 



crystallization of remaining plagioclase, 
augite, and oxide components would 
leave a residual liquid having essentially 
the composition of the interstitial alkali 
feldspar of the gabbro (Table 28, column 
4) with a small amount of quartz. In 
this example at least, which is repre- 
sentative of low-pressure crystal-melt 
equilibria, a basaltic liquid lying near 
the ol-di join in the ternary ol-di-hy 
(G, Fig. 92) yields a slightly oversatu- 
rated residual liquid, which is ultimately 
enriched in potash relative to soda. 



GEOPHYSICAL LABORATORY 

STRUCTURAL GEOLOGY 



251 



Experiments in Flow Deformation 

William H. Scott 

In most geologic experiments designed 
to reproduce natural folds, particular 
attention is given to scaling the physical 
properties of rock bodies and the time 
and size of the natural process, so that 
the experiment more closely reproduces 
geologic deformation. It is probable that 
the three-dimensional conditions of flow 
— whether convergent, divergent, or par- 
allel ; whether velocity gradient or not — 
are factors as important as the physical 
properties of the rock body in controlling 
the geometry and style of the folds pro- 
duced (Hansen, 1963, 1968) . The experi- 
ments described below are the first in a 
series designed to test the relationships 
between the conditions of flow and the 
resultant plastic strain seen as folds in 
anisotropic fluids. 

Design of the Apparatus 

Fig. 95 is a schematic diagram of the 
experimental apparatus. A charge of 
low-viscosity fluids is forced through a 
chamber under conditions of controlled 
flow. The charge, measuring 5 X 5 X 12 
inches, is composed of alternating layers 
of two petroleum-based materials of dif- 
ferent viscosities. It is built in a staging 
chamber, open at each end. Confining 
the charge at the ends are two expand- 
able plates, the driving and receiving 
plates, which are connected by thrust 
rods and velocity-gradient assemblies to 
a common drive shaft. This shaft, driven 
at 2.9 rpm, translates the driving and 
receiving plates at 18 inches/hour. The 
charge is thus forced to flow through the 
staging and flow chambers at that rate, 
confined at all times by the chamber 
walls and the expandable plates. The 
flow chambers are 36 inches long; they 
are detachable and can be replaced by 
others that provide for any one of a 
variety of flow conditions. In order to 



maintain constant volume in chambers 
with progressively changing cross sec- 
tions, the rear velocity-gradient plate is 
spring-loaded on a spline (not shown 
in Fig. 95), as well as attached to the 
driveshaft, allowing the receiving plate 
to move at a variable rate that will just 
maintain equal-volume flow. Motor- 
driven gear trains on the velocity-gradi- 
ent plates rotate the driving and receiv- 
ing plates 15°/hour. Besides translating 
the charge through the chamber, there- 
fore, the velocity-gradient plates pro- 
duce a velocity gradient in the vertical 
plane by forcing the top of the charge 
to move faster than the bottom of the 
charge. Thus, at constant volume, the 
apparatus can impose variable condi- 
tions of convergent, divergent, parallel, 
and velocity-gradient flow on charges of 
anisotropic fluids. 

Experimental Conditions 

A series of experiments has been run 
with interlay ered stitching wax (^esIO 8 
poises at 25°C, layers 0.075 to 0.100 inch 
thick) and Vaseline (t^10 3 poises at 
25°C, layers 0.200 to 0.300 inch thick). 
Charges were made to flow through a 
compound flow environment, converg- 
ing 8° in the horizontal plane and di- 
verging 16° in the vertical plane, during 
36 inches of translation in 2 hours (see 
Fig. 95). In addition, 30° rotation of the 
driving and receiving plates imposed a 
vertical velocity gradient by cumula- 
tive relative slip of 4 inches between the 
layers, d and e 3 for the total strain of 
the charges are located in the vertical 
plane containing the direction of transla- 
tion; ei, vertical at the beginning of flow, 
plunges 60° back toward the driving 
plate at the end of flow. e 2 is horizontal 
and perpendicular to the direction of 
translation. Strain rates for total strain 
are of the order of 10~ 5 /second. 

Translation of a charge during an ex- 
periment was horizontal and parallel to 



252 



CARNEGIE INSTITUTION 



UJ 
> 

< 




<A CC 

.o 



n CC 

o o 



5 3 

_] < 
< x 

2l 

UJ U- 



> 

UJ 
Q 



W 



fH 



GEOPHYSICAL LABORATORY 



253 



the imposed velocity gradient. The major 
component of slip between the layers, 
therefore, was parallel to the velocity 
gradient and to the direction of transla- 
tion. The layers were also constricted in 
the horizontal plane, shortening both 
parallel and perpendicular to translation. 
Wall friction during flow was con- 
siderable, noticeably deflecting the wax 
layers within 2 cm of the walls. No struc- 
tural data are reported from this region, 
but it is not known how far the shear 



from this friction penetrated into the 
charge. Succeeding experiments are ex- 
pected to include modifications that ef- 
fectively eliminate wall friction. 

Initial Results 

1. The layers of stitching wax were 
deformed into folds during flow. The 
folds are concentric in profile and dis- 
harmonic, and display curved hinge 
lines. 



S.L. 




Fig. 96. Orientation of, and asymmetry about, fold axes formed during experimental flow 
(8° horizontal convergence perpendicular to the direction of transport, 16° vertical divergence, 
and velocity-gradient flow). The separation angle (bold arc of horizontal great circle) includes 
the slip line, SI/. 



254 



CARNEGIE INSTITUTION 



2. In all layers, the folds display a 
planar preferred orientation of fold axes. 
The planes so described are horizontal 
and parallel to the wax layers ("bed- 
ding"). 

3. Most of the folds are asymmetrical. 
The sense of overturning is consistently 
top-over-bottom in the sense and direc- 
tion of the shear couple imposed by the 
velocity gradient. 

4. The orientation of the fold axes 
and the asymmetry of the folds in a 
single, deformed charge are shown in 
spherical projection in Fig. 96; folds in 
the eastern portions of the net are coun- 
terclockwise, and in the western portions, 
clockwise. The 18° separation angle, 
which separates the clockwise axes from 
the counterclockwise axes (Hansen et al., 
1961; Hansen, Year Book 65, pp. 390- 
405), includes the line of slip (S.L.) 
produced by the velocity gradient be- 
tween the layers. Hansen (Year Book 
65, pp. 390-397) showed by geometric 
analysis of folds in a tundra landslide 
that the separation angle contains the 
slip line for folding. These experiments 
reproduce nearly identical fold fabrics, 
independently confirming his conclusions. 

Movement Directions and the Axial- 
Plane Fabrics of Flexural Folds 

William H. Scott and Edward Hansen 

It was predicted from theoretical con- 
siderations in an earlier report that the 
axial planes of a group of flexures or 
flexural-slip folds would be found to 
share a common axis (Hansen, Year 
Book 65, pp. 393-398). Subsequently a 
common axis, or zone axis, has been ob- 
served in the axial-plane fabrics of these 
types of folds in several localities. Two 
field examples are described here that 
illustrate the relationship between the 
zone axis and the movement history of 
the folds. These examples indicate that 
the zone axis can be used as another 
independent means of deducing move- 
ment directions, or slip-line orientations, 



that obtained during flexural and flexural- 
slip folding. 

"Slip-line orientations" of flexures and 
flexural-slip folds, as meant here, does 
not refer to the directions of slip between 
layers within the limbs of flexural-slip 
folds. Instead it refers to the orientations 
of the lines of relative slip between the 
layer (s) that undergoes folding and the 
adjacent layers that do not fold. 

Flexures in Lava 

Flexural folds displaying a zone axis 
of axial planes occur on a small volcanic 
spattercone near Reykjavik, Iceland. The 
cone is the subject of a study of the 
relationships between folds and the flow 
in lava (Scott, 1968). In the study, fabric 
data were collected from nine small areas 
on the sides of the cone. Data from 
one representative area serve here to il- 
lustrate the relationship between the 
axial-plane fabric and the movement 
direction of the lava. 

Ejected lava bombs, after falling onto 
the cone surface, flattened into plates 
and flowed several inches downslope. The 
lava plates deformed plastically into 
concentric, disharmonic folds, consist- 
ently overturned in the downslope di- 
rection. Fold axes of 14 folds are shown 
in spherical projection in Fig. 97(A). 
Semicircular arrows modifying the axes 
indicate the asymmetry of the folds 
(Hansen et al., 1961). The plane that 
best fits the distribution of axes has been 
calculated as the plane whose pole is axis 
1 of the Dimroth ellipsoid for the distri- 
bution (Dimroth, 1962a, 19626, 1963); 
it is shown as the dashed great circle, S, 
which approximates the slope of the spat- 
tercone. Flow of the lava plates under 
gravity was downslope. Orientation of 
the downslope direction — the slip-line 
orientation for the area — is shown as a 
square in Fig. 97(A). It is oriented be- 
tween the group of six clockwise axes 
and the group of eight counterclockwise 
axes, within the separation angle (solid 



GEOPHYSICAL LABORATORY 



255 




Fig. 97. Orientation data of 14 flexures in lava on a small spattercone, Iceland. (A) Fabric 
data: Dots represent fold axes, semicircular arrows show asymmetry patterns, S is the best- 
fitting plane to the fold axes, and P is the pole to S. Slip-line solutions: The square represents 
the downslope direction, the solid arc represents the separation angle, and the triangle is the 
solution from the zone axis in diagram C. (B) Axial planes of folds in diagram (A). Z represents 
the zone axis. (C) Contour diagram of the 91 intersections of the axial planes in diagram 
(B). S.L. is the slip-line solution from the zone axis. Contours: 1.1, 3.3, 5.5, 9.9% per 1% area; 
maximum, 51%. 



arc in Fig. 97A; Hansen et al., 1961; 
Hansen, Year Book 65, pp. 390-405). 

Axial planes of the 14 folds are shown 
in Fig. 97(B). They intersect in a strong 
point maximum in the southwest quad- 
rant of the diagram. The pole to the 
plane that best fits the poles to the 14 
axial planes is labeled Z ; it approximates 
the center of gravity of the maximum. 
The maximum is emphasized by con- 



touring of the 91 intersections in Fig. 
97(C). 

The slip-line orientation of flexural 
folds is contained in the plane of layer- 
ing (*S in Fig. 97A) . It follows from the 
discussion of Fig. 108 (p. 395) in Year 
Book 65 that the pole to layering (P, Fig. 
97A) and the zone axis (Z, Fig. 97B) de- 
fine a plane that also contains the slip-line 
orientation. The plane is shown as a solid 



256 



CARNEGIE INSTITUTION 



great circle in Fig. 97(C) . Its intersection 
with S, therefore, should parallel the slip 
line {S.L., Fig. 97C) . The orientation of 
the intersection, which is a solution for 
the slip-line orientation, is plotted in Fig. 
97(A) as a triangle. It is included in the 
separation angle and oriented 5° from 
parallelism with the downslope direction. 
All three solutions are compatible. 

Flexural-Slip Folds in Migmatite 

The best set of flexural-slip folds found 
during a reconnaissance of several moun- 
tain chains occurs in granite migmatite 
in Lilledalen, southern Norway, deep 
within the metamorphic core of the Cale- 
donides (Hansen, Scott, and Stanley, 
Year Book 65, p. 407, Fig. 117, area 6). 
Fold axes and the asymmetry patterns of 
45 of these concentric, disharmonic folds 
are shown in projection in Fig. 98(A). 
The plane that best fits the axes is shown 
as the dashed great circle, S, which par- 
allels the general attitude of composi- 
tional layering in the area. The solid arc 
within S is the separation angle between 
the groups of clockwise and counterclock- 
wise axes. It is a solution for the slip- 
line orientation that obtained during the 
development of the folds. 

Axial planes of 11 folds are shown in 
Fig. 98(B). Their intersections cluster 
in the eastern part of the diagram; 
the point of highest density in the cluster 
is labeled Z. In Fig. 98(C), which con- 
tains the contoured intersections of the 
axial planes, the planes PZ (solid) and 
S (dashed) are shown. Their intersec- 
tion (S.L.) is a solution for the slip- 
line orientation. It is plotted as a triangle 
in Fig. 98(A) for comparison with the 
separation angle. The two solutions are 
compatible. 

Use of the Axial-Plane Fabrics to 
Deduce Slip-Line Orientations 

These examples indicate that the zone 
axes of the axial planes of flexural folds 
can be used to deduce slip-line orienta- 
tions. To do so, it is necessary (1) to 



measure the fold axes, find the plane (<S) 
that best fits the axes, and find its pole 
(P) ; (2) to measure the axial planes 
and find their zone axis (Z) ; and (3) 
to construct the plane PZ. The inter- 
section of PZ with S is a solution for the 
slip-line orientation. It should be noted, 
however, that any solution obtained by 
this method is based upon the same as- 
sumption of the intermediate principal 
stress axis (o- 2 ) being parallel to S as any 
solution obtained from the separation 
angle (c/. Hansen, Year Book 65, pp. 
396-397). 

For any group of folds under con- 
sideration, the zone axis can be taken 
as either (1) the pole (tt) to the plane 
that best fits the poles to the axial planes 
or (2) the highest density (/J) of inter- 
sections of the axial planes (Turner and 
Weiss, 1963, pp. 83, 154). Although ■* 
was used for the zone axis in the lava 
plates (Fig. 97C) and ft for the zone 
axis in the migmatite (Fig. 98C), it is 
not clear which of the two will prove 
more accurate or more useful. 

Perhaps it should be made explicit that 
the kind of zone axis used here to deduce 
slip-line orientations is found in origi- 
nally nonparallel axial planes of a 
single generation of flexural folds that 
display a planar distribution of fold axes. 
This kind of zone axis should not be con- 
fused with other kinds, such as those 
produced by the subsequent rotation of 
originally parallel axial planes in areas 
of superposed folding. 

An Explanation of the Spread of Zone 
Axes 

The contours around the major point 
maximum that represents the zone axis 
in Fig. 97(C) describe a V pattern. The 
V points away from the plane of axes 
(S) and opens toward that plane. Al- 
though a V pattern is not obvious in 
Fig. 98(C), it has been observed in many 
diagrams of this type and appears to 
be characteristic. It is interpreted to 
mean that the folds formed consecutively 



GEOPHYSICAL LABORATORY 



257 




Fig. 98. Orientation data of nexural-slip folds in magmatite, Lilledalen, Norway. (A) Fabric 
data: Dots represent fold axes (n = 45), semicircular arrows show asymmetry patterns, S is 
the best-fitting plane to the fold axes, and P is the pole to S. Slip-line solutions: The solid 
arc represents the separation angle, and the triangle is the solution from the zone axis in 
diagram (C). (B) Axial planes of 11 folds in diagram (A). Z represents the zone axis. (C) Contour 
diagram of the 55 intersections of the axial planes in diagram (B). S.L. represents the slip-line 
solution from the zone axis. Contours: 1.8, 5.5, 9.1, 12.7% per 1% area; maximum, 34.6%. 



under rotation due to a shear couple 
(i.e., under drag) . This interpretation is 
supported by the following geometrical 
demonstration that such a pattern can be 
produced in this manner. 

Suppose that four flexural-slip folds 
with fold axes b' 1} b' 2 , b' s , b' \ (Fig. 99A) 
develop in bedding S under compression 
(a 2 //S; cf. Hansen, Year Book 65, pp. 
393-396) ; their axial planes (S'x, S' 2} S' 3 , 
S'i) intersect at Z. Under the influence 
of the maximum resolved shear stress on 



S (open circle), the axial planes are 
rotated toward parallelism with S, and 
the zone axis is rotated toward parallel- 
ism with the line of relative slip, or drag, 
between beds. If we assume that slip 
can occur with equal ease in any direc- 
tion within S, the slip line parallels the 
maximum resolved shear stress (open 
circle). New orientations of the axial 
planes (S' lr , $' 2r , S' 3r , S r ir ) and the zone 
axis (Z r ) after rotation are shown in the 
figure. Suppose now that four additional 



258 



CARNEGIE INSTITUTION 




Fig. 99. Hypothetical development of a set of twelve fiexural folds during rotation by drag. 
(A) Orientation of bedding (S), fold axes (£>'), axial planes (S')> and zone axes (Z) ; the open 
circle represents the slip-line orientation. (B) Contour diagram of 66 intersections of the 
twelve axial planes in diagram (A). Contours: 1.5, 4.5, 7.6, 10.6% per 1% area; maximum, 25.8%. 



folds develop in the same manner as the 
first four folds and, for simplicity, in 
the same orientations as the first folds 
before rotation (i.e., b' ' u b' 2) b' 3 , b' \, S' lf 
S' 2 , S' s , S'j„ Z, Fig. 99A) . Under the same 
shear that rotated the first set of folds, 
the axial planes of the second set are 
rotated in the same direction as the first, 
so that their common axis at Z is rotated 
to Z r . At the same time, the common 
axis of the first set at Z r is rotated to Z r >. 
Finally, let us suppose that a third set of 
four folds develops and, again for sim- 
plicity, that they have the initial orien- 
tations (b' 1} b' 2 , b's, b\, S't, S' t , S's, S',„ 
Z) . At this point, the stress is dissipated, 
and folding and rotation stop. The fold 
axes b' t, b' 2 , b' s , b' \ are shared by all 
three fold sets, but the orientations of 
the axial planes are unique ; the zone axis 
of the first set is at Z r >, the zone axis of 
the second set is at Z r , and the zone axis 
of the third set is at Z (Fig. 99A) . 

Mutual intersections of the twelve 
axial planes in Fig. 99(A) are contoured 
in Fig. 99(B). The higher contours 
around the most important maximum 
show an irregular V pattern, which points 



away from S. Therefore, the V pattern 
in this diagram is similar to that in 
Fig. 97(C). 

It is expected that in the natural for- 
mation of a group of folds their develop- 
ment and the rotation of their axial 
planes would occur as one continuous 
process. Hence, the development of 
multiple, discrete zone axes, as illustrated 
in Fig. 99(A), probably would not occur 
in nature. The unusual coincidence of 
some of the contours of different levels 
in Fig. 99(B) is the product of restrict- 
ing the folds to four axial orientations 
and the zone axes to three discrete 
orientations. Nevertheless, the spread of 
the zone axis into a V pattern as con- 
toured in Fig. 99(B) is the product of 
the rotation of some of the axial planes 
with respect to the others; it is not the 
product of the simplifications and re- 
strictions. 

On the "Drag Folds" of Van Hise 
and Leith (1911) 

Edward Hansen and William H. Scott 

In his recent text, Whitten drew atten- 
tion to the apparent introduction into the 



GEOPHYSICAL LABORATORY 



259 



structure literature of the term "drag 
fold" by Van Hise and Leith (1911) for 
a distinct type of fold seen in the Soudan 
formation, Vermillion iron district, Min- 
nesota. Three descriptive properties of 
this fold type were given (Van Hise and 
Leith, 1911, p. 123; Whitten, 1966, pp. 
164-165) : (1) The folds display a planar 
distribution of fold axes parallel with 
bedding. (2) The folds, though locally 
complex, do not alter the general trend 
of the beds. (3) The profile geometry is 
concentric but somewhat flattened (ob- 
served in Fig. 12 of Van Hise and Leith) . 
In addition, the authors mentioned that 
the folding involved buckling of the beds. 
Since the time of Van Hise and Leith, 
the name "drag fold" has been used by 
many geologists for most of the minor 
or small-scale folds encountered in the 
field, whether or not they shared the 
descriptive properties of the original 
"drag folds." The planar distribution of 
fold axes was forgotten, and even the 
concentric profile section, flattened or 
not, was not characteristic of all so-called 
drag folds. Rather, the basis for using the 
name was apparently genetic — the idea 
that small folds are produced by the 
drag generated between beds on the 
limbs of larger folds undergoing flexural 
slip (c/. Knopf and Ingerson, 1938, pp. 
159-160). However, the difficulty of as- 
certaining that any natural fold was 
actually produced or even modified by 
drag, as well as the recognition that 
many so-called drag folds occur in hinge 

I areas of larger folds where no drag is 
generated during flexural slip, has placed 
the term "drag fold" in general disre- 
pute (de Sitter, 1956, p. 226; Whitten, 
1966, pp. 164-168; Ramsay, 1967, pp. 
396-397). 
Nevertheless, the folds that Van Hise 
and Leith described are of interest be- 
cause they are the "type" folds for 
which one of the important dynamic con- 
cepts of folding was introduced — folding 
by drag. In particular, these folds are of 
interest here because of their similarity 
on all three counts, above, to the folds 



that the present writers and Stanley 
studied rather extensively in reconnais- 
sance (Year Book 65, pp. 406^10). 
Consequently, a trip was made to the 
area where the folds are reportedly well 
developed (sees. 13 and 14, T.62N., 
R.13W., Minn.) in the vicinity of 
Mitchell Lake and Twin Lakes, between 
Tower and Ely. The relationships noted 
during the trip are reported here. 

Three sets of folds are described for 
the Soudan formation (Van Hise and 
Leith, 1911, pp. 123-124). From the 
original descriptions, all three are easily 
recognizable in the outcrops near Mitchell 
and Twin Lakes: an early set of iso- 
clinal "longitudinal folds" with nearly 
horizontal fold axes, a later set of more 
open "cross folds" with axes that com- 
monly plunge 50° to 60°, and a set of 
"drag folds" with axes in all orienta- 
tions within the compositional layering. 
The relative age of the "drag folds" is 
unclear, but they seem to be intermediate 
between the "longitudinal folds" and the 
"cross folds." The geometry of the rocks 
is complex. 

The best group of "drag folds" un- 
complicated by folds of the other genera- 
tions was found on the southern shore of 
the eastern Twin Lake, in a frost-heaved 
block (2 mx2 m) that had rotated 
about 20° from its original orientation in 
outcrop. However, because our interest 
was not in the relationship of the folds 
to the regional geology but in the nature 
of the folds themselves, this group was 
satisfactory. The rock types are inter- 
layered greenstone and jasper. The folds 
are contained in an incompetent layer 
3 cm thick with internal competent layers 
approximately 1 mm thick. The folds 
are seen better in the competent layers, 
where most of the orientation data were 
collected. The portion of the layer in 
which the folds occur is about 40 cm 
long; it is part of a long limb of a "cross 
fold" that is nearly two orders of magni- 
tude larger than the "drag folds." 



260 



CARNEGIE INSTITUTION 



Fold- Axis and Axial-Plane Fabrics 

The folds are concentric in profile 
geometry, but somewhat flattened, very 
much like the one illustrated by Van 
Hise and Leith (1911, Fig. 12). Fold 
axes and asymmetry patterns of eight 
folds are shown in Fig. 100 (A) . The best- 
fitting plane (S) to the fold axes parallels 
the compositional layering. A separation 
angle of 39° is defined in *S by the re- 



versal in asymmetry in the southwestern 
quadrant of the diagram. Therefore, the 
folds are compatible with a single shear 
couple acting parallel with S and within 
the separation angle, in the sense of top 
downward (Hansen, et al., 1961; Han- 
sen, Year Book 65, pp. 390^05). The 
concentric profile geometry, the planar 
distribution of fold axes, and the dis- 
tribution of asymmetry patterns corn- 




Fig. 100. Orientation data of eight "drag folds" in the Soudan formation, Twin Lakes, Minnesota. 
(A) Dots represent fold axes, semicircular arrows show asymmetry patterns, S is the best- 
fitting plane to the fold axes, and P is the pole to S. The arc with the open circle within S 
represents the separation angle. (B) Axial planes of the folds in diagram (A). Z represents the 
zone axis. (C) Contour diagram of the 28 intersections of the axial planes in diagram (B). 
S.L. is the slip-line solution from the zone axis. Contours: 3.6, 10.7, 17.8, 25% per 1% area; 
maximum, 32%. 



GEOPHYSICAL LABORATORY 



261 



patible with a single shear couple in- 
dicate that these "drag folds" are the 
same kind of fold as the so-called drag 
folds described in Year Book 65 (pp. 406- 
410). 

Axial planes of the eight folds in dia- 
gram (A) are shown in diagram (B). 
The intersections of the axial planes are 
contoured in diagram (C). They define 
two major point maxima 70° apart, the 
most important of which is located in 
the southwestern quadrant. The highest 
density of intersections is labeled Z (Fig. 
100B). Z can be considered a zone axis 
of the axial planes and can be used to 
get a solution for the slip-line orientation 
that obtained during folding (c/. pre- 
ceding report on axial-plane fabrics of 
flexural folds) . The intersection of S with 
the plane containing Z and the pole (P) 
to S is the slip-line solution, labeled S.L. 
in Fig. 100(C). This solution is com- 
patible with the separation angle, which 
it bisects. Therefore, the axial planes 
of the "drag folds," though less regular, 
display the same relationships of fabric 
and kinematics as the axial planes of 
other groups of folds of the same type 
(Figs. 97 and 98, this report). 

Height-Width and Depth-Width Ratios 

Where folds are asymmetric, two ad- 
jacent hinges are commonly coupled by 
an intervening short limb. The form of 
such a coupled fold can be described 



by various ratios of absolute measure- 
ments. The measurements used in this 
study are the height (Matthews, 1958), 
width, and depth (Hansen, 1963, pp. 17- 
20; 1968). All three distances are mea- 
sured in the profile section of a fold (Fig. 
101). The height {H) is the distance be- 
tween the two hinges coupled by the 
short limb, measured parallel to the trace 
of the axial surfaces {AT, Fig. 101) . Two 
lines can be constructed perpendicular 
to the axial traces, each tangent to the 
fold at a hinge and intersecting a long 
limb; the width {W) is the projected 
distance between the points of intersec- 
tion of these lines with the long limbs, 
measured perpendicular to the axial 
traces. The width is measured at maxi- 
mum height. The depth (D) is the length 
of the axial surface in profile (Fig. 101). 
These distances were measured on 
seven of the eight folds of Fig. 100. (The 
eighth fold was not exposed well enough 
for this purpose.) The heights range 
from 2 to 7 mm; the mean value is 
4.4 mm. The widths vary from 3 to 
6 mm; the mean is 4.3 mm. And the 
depths vary from 2.5 to 5 cm; the mean 
is 3.9 cm. Height-width ratios range 
from 0.67 to 1.33, and their mean is 1.0; 
depth-width ratios range from 6.3 to 
13.3, and their mean is 9.7. These values 
show that an average "drag fold" in 
this group is near the boundary between 
open and isoclinal (H/W = 1.0), and that 




Fig. 101. Profile section of a hypothetical, coupled similar fold. AT are traces of the axial 
surfaces. H is the height, W the width, and D the depth. 



262 



CARNEGIE INSTITUTION 



it is an order of magnitude deeper than 
it is high. 

To see if the form of the folds varies 
as a function of the angle between the 
fold axis and the slip-line orientation 
(kinematic a), the height-width ratios 
are plotted in Fig. 102(A) against the 
angle measured in S, between the fold 
axes and the nearest end of the separa- 
tion angle (c/. Fig. 100A). A line has 
been fitted to the points by the method of 
least squares, and a correlation coeffi- 
cient of 0.92 indicates that, for (n — 2) 
degrees of freedom, the correlation is 
significant at the 0.005 level. It is ap- 
parent from the diagram that the height 
of the folds increases relative to the 
width as the fold axis diverges from par- 
allelism with the slip-line orientation. 
Thus the folds in which the fold axes 
are nearly parallel to the slip-line orien- 
tation are relatively open, and the folds 
in which the axes are nearly perpendicu- 
lar to the slip-line orientation are rela- 
tively isoclinal. 



_H_ 
W 




O 10 20 30 40 50 60 70 
fold axis A separation angle 



_D_ 
W 



B 



20-i 



10 -0 




fold axis A separation angle 



Fig. 102. Height-width ratios (A) and depth- 
width ratios (B) of seven "drag folds" in 
Fig. 100 plotted against the angle, measured 
in S, between the fold axes and the nearer end 
of the separation angle. 



Ratios of depth to width are plotted 
in a similar fashion in Fig. 102(B). A 
line has been fitted to the points by the 
method of least squares. There appears 
to be no significant correlation between 
depth-width ratio and divergence be- 
tween fold axis and slip-line orientation. 

Dynamic Models 

The relationships observed in the 
"drag folds" under consideration appear 
to be compatible with several modes of 
development, of which drag is one pos- 
sibility. Three models are outlined in 
the following paragraphs. Primarily they 
seek to explain the relationship between 
height-width ratio and divergence be- 
tween fold axis and slip-line orientation. 
The zone axis of axial planes and the 
separation angle defined by the patterns 
of asymmetry have been discussed previ- 
ously (Hansen, Year Book 65, pp. 393- 
397) . It is assumed that the "drag folds" 
developed by flexural slip, as inferred 
from their concentric profile sections and 
their nonparallel axial planes sharing a 
common axis. 

1. Let us suppose that the folds de- 
velop only under drag — the traction pro- 
duced on an active layer by another layer 
moving past it. This physical situation 
can be described dynamically by a single 
shear couple, or kinematically by a 
single slip-line orientation, acting par- 
allel to the plane of layering. Depending 
upon the disposition of inhomogeneities 
that can serve as triggers within the 
layer, potential flexural-slip folds could 
develop with axes in any orientation 
within the plane of layering. However, 
the layer can only be dragged into folds 
about the axial orientations that the 
shear couple acts across; thus, potential 
folds with axes parallel to the shear 
couple (or slip-line orientation) do not 
develop. If the layer were dragged about 
all the fold axes the same distance par- 
allel with the shear couple, the distance 
that the layer would become folded over 
itself, which is measured as the height 



GEOPHYSICAL LABORATORY 



263 



of a fold, would be directly proportional 
to the sine of the angle between the fold 
axis and the line of action of the shear 
couple. Therefore, the height would at- 
tain a maximum value in folds with axes 
perpendicular to the shear couple, and 
it would grade down to a minimum value 
in folds with axes nearly parallel to the 
couple. The height-width ratios would 
probably show a similar relationship 
(Fig. 102A) because the width of a fold 
appears to be more a function of the 
strength of the layer relative to its sur- 
roundings, which remains constant, than 
a function of the relative orientation of 
the fold to the shear couple. Of course, 
in nature, the layer would not be dragged 
exactly the same amount in every fold, 
and some "noise" in the relationship 
should be expected. 

2. Suppose that the folds develop in 
response to a triaxial stress field (o t > 
o- 2 >o- 3 ) such that o 2 parallels the plane 
of layering, a 1 makes a small angle with 
the layering, and the layer to be folded 
is under constriction. In response to the 
constriction, flexural-slip folds develop 
with fold axes in all orientations within 
the layering. Let us stipulate that the 
two principal stress directions within the 
layering (o 2 and ai- 3 , which is the inter- 
section of the 0-1-0-3 plane with the layer- 
ing) are different such that o!-3>o 2 . If 
the folds develop in equal numbers in 
all orientations within the layering and 
if the strain manifested as folds is pro- 
portional to stress, it is expected that the 
height of the folds would attain a maxi- 
mum value where fold axes are perpen- 
dicular to the direction of major com- 
pression (0-1-3) and would grade down 
to a minimum value where fold axes are 
perpendicular to the direction of minor 
compression (o 2 ). If the width of the 
folds did not respond in this manner but 
remained fairly constant as a function 
of the relative strength of the layer, the 
height-width ratios would also attain a 
maximum value where fold axes are 
perpendicular to o-i- 3 and grade down 
to a minimum value where fold axes are 



perpendicular to o- 2 . The maximum re- 
solved shear stress on the plane of layer- 
ing parallels o-i_ 3 ; on the assumption 
that slip can occur with equal ease in 
all directions within the layering, the 
maximum resolved shear stress also par- 
allels the slip-line orientation for the 
whole domain. Thus, the height-width 
ratios should bear the same relationship 
to the maximum resolved shear stress 
and the slip-line orientation as sug- 
gested for CT1-3. 

3. A combination of the two models 
just outlined could also be proposed in 
which the folds develop as compressional 
folds under the influence of a stress field 
similar to the one described for model 2, 
but were subsequently dragged and/or 
flattened into the present forms under 
the influence of the resolved shear stresses 
on the layer (s). This model has been 
described elsewhere in detail (Hansen, 
Year Book 65, pp. 393-395). 

Slip Folds in Planes of Unsystematic 
Orientation 

Edward Hansen 

Part of the past field season was spent 
working on the compound problem of 
the strain involved in, and the directions 
of flow that obtained during, the devel- 
opment of the earliest generation of folds 
in the metamorphic core of the Nor- 
wegian Caledonides. These folds are im- 
portant to an understanding of orogeny 
because, as in most orogenic belts, they 
appear to involve the greatest amount 
of strain of all orogenic folds. Neverthe- 
less, they are difficult to interpret be- 
cause they have formed in a geosynclinal 
sequence practically devoid of preexisting 
penetrative structures providing refer- 
ence to the original unstrained state and 
because they are superimposed by two 
or more subsequent generations of folds. 
In an attempt to minimize the former 
difficulty, these folds are being studied 
in a pre-Caledonide metamorphic terrane 
between Grotli and Geiranger, southern 
Norway (Strand, 1966; Brueckner, 



264 



CARNEGIE INSTITUTION 



1968), where they themselves may be 
observed superimposed upon structures 
of an earlier orogeny. The rocks in this 
area are predominantly gray gneisses and 
schists that are locally cut by granite 
veins in virtually all orientations. The 
most obvious folds are seen in the granite 
veins and are apparently slip folds. The 
initial problem in this study is whether 
the asymmetry patterns of a single sys- 
tem of slip folds in planes of unsyste- 
matic orientation can be used to deduce 
part of the movement history of the 
folds. 

The problem can be solved by using 
the kind of analysis of asymmetry that 
has been reported earlier (Year Book 65, 
pp. 393-405). For this purpose, let us 
consider the hypothetical situation sum- 
marized in Fig. 103. Six planes (Si, 
S 2 , . . ., S 6 ) are intersected by a family 
of slip planes (£') ; the orientation of 
the slip lines within S' is indicated by 
an open circle. Slip along S' is such as to 



produce folds within the six planes. The 
axes (6'i, 6' 2 , . . ., 6' 6 ) of the resulting 
folds parallel the intersections of S' with 
the six planes. The dashed arrows drawn 
from the poles of S lt S 2 , . . ., S 6 to the 
pole of the slip planes (£') indicate the 
asymmetry patterns of the folds, shown 
by the semicircular arrows modifying the 
fold axes (c/. Year Book 65, p. 394, Fig. 
107). It is clear from the resulting dia- 
gram that the asymmetry patterns of the 
six folds are mixed, both clockwise and 
counterclockwise, within the slip planes 
and bear no systematic relationship to 
the slip-line orientation. In conclusion, 
the asymmetry of slip folds in unsys- 
tematically oriented planes has no kine- 
matic significance.* 

* We can also regard the six planes (Si, S 2 , 
. . ., S 6 ) as different attitudes of a single surface 
folded about several axes, represented in the 
diagram by some of the mutual intersections of 
the planes. The conclusion also applies, there- 
fore, to the asymmetry of slip folds in non- 
coaxially folded surfaces. 




Fig. 103. Lower-hemisphere, equal-area projection of structural elements of six planes super- 
imposed by a system of slip folds. 



GEOPHYSICAL LABORATORY 



265 



Nevertheless, the preceding section of 
this report suggests that the height- width 
ratios of folds might be used instead to 
determine the slip-line orientation that 
obtained during folding. The height (H) 
of a slip fold is proportional to the sine 
of the angle a between the fold axis and 
the slip-line orientation {Year Book 66, 
p. 539, Fig. 104*). The spacing (S) be- 
tween the axial planes of slip folds 
coupled by a short limb is completely 
independent of the angle a, and the 
width (W) is practically independent. 
Therefore, we can expect the height- 
spacing and height-width ratios of slip 

* With regard to Fig. 104 in Year Book 66, 
the height was loosely termed "amplitude (H) ." 
Although the amplitude (A) of a fold may, in 
the usage of a few workers, be equal to the 
short-limb height (H) of a fold (Matthews, 
1958), A is best considered roughly equal to 
H/2 (cf. Ramsay, 1967; Hansen, 1968). 



folds to be proportional to sin a, attain- 
ing their maximum values where the slip 
lines are perpendicular to the fold axes 
(a = 90°) and their minimum values 
where the slip lines approach parallelism 
to the fold axes (a = 0°). It follows that, 
in a group of slip folds with a planar 
distribution of fold axes, we can measure 
a set of either of these ratios, plot the 
ratios against the orientation of the fold 
axes, and take the minimum in the 
curve as the orientation of the slip lines 
by which the folds developed {cf. Fig. 
102A). 

This relationship applies to the folds 
in planes of random orientation in the 
Grotli-Geiranger area, which are the im- 
mediate concern here, and should apply 
also to groups of slip folds with planar 
distributions of fold axes in any geo- 
logical situation. 



STAFF ACTIVITIES 



Aluminum Silicate Conference 

A conference on the experimental rela- 
tionships and field occurrences of the 
aluminum silicate minerals, kyanite, an- 
dalusite, and sillimanite was held jointly 
by the Geophysical Laboratory and Yale 
University at the Klein Geology Labora- 
tory, New Haven, Conn., May 20-22, 
1968. P. M. Bell, of the Geophysical 
Laboratory, and S. P. Clark, Jr., and 
B. J. Skinner, of Yale University, planned 
the conference for the purpose of analyz- 
ing recent experimental and field ob- 
servations that are highly significant 
in interpreting the pressures and tem- 
peratures of the earth's crust. 

Papers were presented by the follow- 
ing participants: A. Albee (California 
Institute of Technology), E. Althaus 
(Yale University), P. M. Bell (Geo- 
physical Laboratory), C. W. Burnham 
(Harvard University), G. A. Chinner 
(Cambridge University), M. C. Gilbert 
(Geophysical Laboratory), H. J. Green- 
wood (University of British Columbia), 



S. S. Hafner (University of Chicago), 
A. Hietanen (United States Geological 
Survey), M. J. Holdaway (Southern 
Methodist University), L. S. Hollister 
(University of California at Los 
Angeles), G. C. Kennedy (University of 
California at Los Angeles) , 0. J. Kleppa 
(University of Chicago), R. C. Newton 
(University of Chicago) , S. W. Richard- 
son (University of Edinburgh), P. Rob- 
inson (University of Massachusetts), 
J. L. Rosenfeld (University of Cali- 
fornia at Los Angeles), W. Schreyer 
(University of Kiel), R. D. Schuilling 
(University of The Utrecht), J. B. 
Thompson (Harvard University), E. Zen 
(United States Geological Survey). The 
last day of the conference included a 
field trip to the unusual occurrences of 
aluminum silicate minerals in Connecti- 
cut, led by John Rodgers of Yale Uni- 
versity. Several papers presented at the 
conference will be published in the fall 
of 1968 in a special issue of The Ameri- 
can Journal of Science, devoted to the 
aluminum silicates. 



266 



CARNEGIE INSTITUTION 



Journal of Petrology 

The Journal of Petrology, aided by the 
careful work of an internationally re- 
nowned group of reviewers and a co- 
operative printer, continues to offer an 
outlet for definitive researches based on 
extensive new data bearing on major 
petrological problems. Volume 8 for 1967 
consisted of 531 pages with contributions, 
by, among others, one Staff Member, 
one Fellow, three alumni, and one guest 
investigator. Professor C. E. Tilley, Re- 
search Associate, and Dr. H. S. Yoder, 
Jr., Staff Member, continue to serve as 
Editors. Professor G. M. Brown, a recent 
Fellow, is the Senior Managing Editor. 

Lectures 

During the report year staff members 
and fellows were invited to present lec- 
tures as follows : 

P. H. Abelson made a total of 20 in- 
vited public appearances. He made 
speeches in the following capacities: As 
opening speaker ("How Man Shapes 
His Environment") at a special session on 
man-made environmental hazards at the 
Annual Meeting of the American Public 
Health Association at Miami Beach; as 
Distinguished Lecturer ("Chemical 
Events on the Primitive Earth") at the 
Department of Geology and Dana Club, 
Yale University; as a participant in a 
Panel on Environmental Pollution at 
the annual meeting of the American As- 
sociation for the Advancement of Sci- 
ence, New York City; as the keynote 
speaker ("The Future of Science and 
Its Effects on Society") at the Southern 
Regional Science Seminar for University 
Information Officers at the College of 
Journalism and Communications, Uni- 
versity of Florida, Gainesville; as 1968 
Henry Darwin Rogers Lecturer ("Or- 
ganic Chemicals in Rocks and Fossils") 
at the Department of Geology, Uni- 
versity of Pennsylvania, Philadelphia; 
as a banquet speaker ("The Changing 
Role of Science in Our Society") at the 
49th Annual Meeting of the American 



Geophysical Union, Washington, D. C; 
and as a speaker ("Can Man Learn to 
Live with Technology" at the dedica- 
tion of a new science complex at the 
University of Puget Sound, Tacoma, 
Washington, where he received an Hon- 
orary D