f Si::-

lit'r

^ ^^il

HARVARD UNIVERSITY.

LIBRARY

MUSEUM OF COMPARATIVE ZOOLOGY

^a.A|T

\

PROCEEDINGS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

PROMOTING USEFUL KNOWLEDGE

VOLUME LVIII
1919

PHILADELPHIA
THE AMERICAN PHILOSOPHICAL SOCIETY

1919

PRESS OF

THE NEW ERA PRINTING COMPANY

LANCASTER, PA.

CONTENTS

Page.

The Parasitic Aculeata, A Study in Evolution. By William
Morton Wheeler i

The Relation of the Diet to Pellagra. By E. V. AIcCollum. . 41

Evolution and Mystery in the Discovery of America. By Edwin
Swift Balch 55

Tatar Material in Old Russian. By J. Dyneley Prince 74

The Energy Loss of Young Women during the Muscular Ac-
tivity of Light Household Work. By Francis G. Benedict
and Alice Johnson 89

Alternating-current Planevector Potentiometer Measurements
at Telephonic Frequencies. By A. E. Kennelly and Edy
Velander 97

Some Scientific Aspects of the Meteorological Work of- the
United States Army. By Robert A. Millikan 133

Detecting Ocean Currents by Observing their Hydrogen-ion
Concentration. By Alfred Goldsborough Mayor 150

Recent Discoveries of Fossil Vertebrates in the West Indies and
their Bearing on the Origin of the Antillean Fauna. By W.
D. Matthew 161

The Relative Contribution of the Staple Commodities to the
National Food Consumption. By Raymond Pearl 182

Self-Luminous Night Haze. By E. E. Barnard 223

Graphical Representation of Functions of the nth Degree. By
Francis E. Nipher 236

The Crocker Eclipse Expedition from the Lick Observatory,
June 8, 1918. By W. W. Campbell 241

The Expedition of the Mount Wilson Observatory to the Solar
Eclipse of June 8. 1918. By J. A. Anderson 255

The Lowell Observatory Eclipse Observations, June 8, 1918.
Prominences and Coronal Arches. By C. O. Lampland. . . 259

The Flash Spectrum. By S. A. Mitchell 265

iii

iv CONTENTS.

Page

Photo-electric Photometry of the 1918 Echpse. By Jakob
KuNz and Joel Stebbins 269

The Sproul Observatory Echpse Expedition, June 8, 1918.
Coronal Arches and Streamers. By John A. Miller 272

Results of Observations of the Eclipse by the Expedition from
the Yerkes Observatory. By Edwin B. Frost 282

The Basis of Sex Inheritance in Sphserocarpos. By Charles
E. Allen " 289

The Application of Sanitary Science to the Great War in the
Zone of the Army. By Bailey K. Ashford 317

Star Clusters and Their Contribution to Knowledge of the Uni-
verse. By Harlow Shapley 337

Hydration and Growth. By D. T. MacDougal 346

Some Considerations on the Ballistics of a Gun of Seventy-five-
mile Range. By Arthur Gordon Webster 373

On a New (?) Method in Exterior Ballistics. By Arthur Gor-
don Webster and Mildred Allen 382

Characters and Restoration of the Sauropod Genus Camara-
saurus Cope. By Henry Fairfield Osborn and Charles
Craig Mook 386

Trogloglanis Pattersoni, a New Blind Fish from San Antonio,
Texas. By Carl H. Eigenmann 397

Polarized Light in the Study of Ores and INIetals. By Fred. E.
Wright 401

On the Luciopimelodinae, a New Subfamily of the South Amer-
ican Siluridse. By Chas. E. Driver 448

Minutes iii

Index XV

PROCEEDINGS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

FOR PROMOTING USEFUL KNOWLEDGE
Vol. LVIII. 1919. No. 1.

CONTENTS

PAGE

The Parasitic Aculeata, A Study in Evolution. B}- William Morton

Wheeler .......... i

The Relation of the Diet to Pellagra. By E. V. McCollum . . 41

Evolution and Mystery in the Discovery of America. By Edwix Swift

Balch ........... 55

Tatar Material in Old Russian. By J. Dyneley Prince . . 74

The Energy Loss of Young Women during the Muscular Activity of Light

Household Work. By Francis G. Benedict and Alice Johnson 89

PHILADELPHIA

THE AMERICAN PHILOSOPHICAL SOCIETY
104 South Fifth Street

1919

The American Philosophical Society

anm>w*ces that an award will be made in the year 1921 of the

HENRY M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

The Control of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent, of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after the
award.

Attention is called to the following regulations governing the award
of the Prize:

Competitors for the prize shall affix to their essays some motto or name
(not the proper name of the author, however) and when the essay is forwarded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judges, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essays shall be in their author, subject to
the right of the Society to publish the crowned essay in its " Transactions "
or " Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,
David Jayne Hill,
Simeon E. Baldwin,
John Cadwalader,
W. W. Keen,
William B. Scott,
President ex-officio,

Committee.

The essays must be sent, addressed to the
President of the American Philosophical
Society, No. 104 South Fifth Street, Phila-
delphia, U. S. A.

PROCEEDINGS

OF THE

AMERICAN PHILOSOPHICAL SOCIETY

HELD AT PHILADELPHIA
FOR PROMOTING USEFUL KNOWLEDGE

THE PARASITIC ACULEATA, A STUDY IN EVOLUTION.^

By WILLIAM MORTON WHEELER.
(Read April 25, igig.)

There is undoubtedly much to be said in favor of the opinion
commonly held by entomologists that the fruitfulness of their inves-
tigations is apt to be directly proportional to the intensity of their
specialization, but it is also true that this very specialization may
often preclude an adequate appreciation or even a recognition of
phenomena that would profoundly impress the worker who pos-
sesses more general biological interests. This statement is not
inapplicable to the subject of the following study, which is an at-
tempt to collate the data accumulated in their respective fields by a
number of observers of ants, bees and wasps and relating to certain
types of parasitism w^iich keep recurring in various natural families
of the Aculeata in response to frequently recurring stimuli or situa-
tions in the organic environment. The few who have published
comprehensive accounts of the phenomena have failed to present
then! as clearly and cogently as the facts would seem to warrant.
I am aware that my own treatment of the subject may leave much
to be desired and especially that my account of the bees, a difficult
and extensive group to which I have been able to devote compara-
tively little study, is rather summary, but every attempt to attain

1 Contributions from the Entomological Laboratory of the Bussey Insti-
tution, Harvard University. No. 155.

PROC. AMER. PHIL. SOC, VOL. LVIII. A, JUNE II IQIQ.

2 WHEELER— THE PARASITIC ACULEATA.

broader generalization has its attendant risks and inadequacies. If
I succeed in directing renewed attention to an interesting series of
facts and to some of the problems which they suggest, the purpose
of this article will have been accomplished.

In considering the parasitic Aculeata I shall adhere to the fol-
lowing classification, which though in certain respects artificial and
unsatisfactory from the standpoint of phylogenetic development,
will nevertheless facilitate an understanding of the history of our
knowledge of the subject:

I. Nonsocial Parasites.

1. Solitary Bees.

2. Solitary Wasps.
II. Social Parasites.

1. Ants.

a. Guest Ants.

b. Slave-makers.

c. Temporary Social Parasites.

d. Permanent Social Parasites.

2. Social Bees.
3- Social Wasps.

The solitary bees may justly claim our attention first, because
they comprise such a large number of parasitic forms which have
been objects of study for more than a century. The taxonomy of
the bees, however, notwithstanding the number of able investiga-
tors they have attracted, is still in a very unsatisfactory state. The
very numerous species are often distinguishable only by very minute
or dubious characters, so that many of the genera are large, homo-
geneous and widely distributed. Even the generic characters are
often very feeble as compared with those employed by taxonomists
in other Aculeate families. Hence also the higher groups, such as
the tribes and subfamilies are so poorly characterized that no two
melittologists agree on their limits or number. The wing venation
is extraordinarily uniform throughout the whole family and the
taxonomic use of the mouthparts encounters the usual difficulties
which beset the employment of delicately adaptive structures. All
the members of the group are of comparatively recent phylo-
genetic development and very highly specialized in adaptation to the

WHEELER— THE PARASITIC ACULEATA. 3

pollenation of flowering plants, themselves a group of organisms
of recent origin. As manifold morphological expressions of this
adaptation attention has often been called to the peculiar modifi-
cation of the mouthparts for extracting nectar from flowers, the
singular branched hairs of the body and modification of the hind
legs or hairs on the venter in the female for carrying pollen, and
the highly developed visual and olfactory organs.

On the side of the instincts there are, further, the marvelous
habits of nidification which have aroused the admiration of all stu-
dents since the days of Reatimur. Still the general activities of the
female bee — the male is, of course, an ethological nonentity as in
other Aculeates — are strangely uniform in their general outlines : —
the visiting of flowers, mating, nidification, provisioning the nest
with pollen and honey, oviposition. But different species visit dif-
ferent flowers and build their nests in different places and of
different materials. All this is true of perhaps 80 per cent, of the
thousands of species, but a considerable number — between 15 and
20 per cent., representing fully 70 genera — have become parasitic
and have therefore ceased to collect pollen and nectar or to con-
struct and provision nests, but instead seek out the nests of other
bees and oviposit in their cells, with the result that the larvae reach
maturity by devouring the provisions so carefully stored for their
own offspring by more industrious mothers.

This peculiar habit has profoundly modified the structure of the
parasites. Their mouthparts have not been affected to any extent
because these bees still visit flowers assiduously for food, but the
collecting apparatus has atrophied and the hairs on the body and
appendages have been completely or almost completely lost, so that
the species have sometimes been placed in a group by themselves
called Denudatas. Other peculiarities are also manifested. The
loss of the collecting apparatus, which is one of the most striking
secondary sexual characters of the female bee, excepting in the
Prosopidinse, which swallow the pollen instead of collecting it on
their hind legs or venter, has brought about a close resemblance of
the sexes to one another. In a few cases the female has even taken
on a male secondary sexual character, as in the genus Androgynella,
recently described by Cockerell (1918). It comprises two species.

4 WHEELER— THE PARASITIC ACULEATA.

detcrsa of Australia and subrixafor of the Philippines. In both
species the female, though possessing a well-developed sting, has
13-jointed antennae, a number peculiar to the male in all other bees
and in fact in most other Aculeates. The specimens cannot be
gynandromorphs, because R. E. Turner found the 13 joints in 14
females of detersa, so that Cockerell is justified in regarding it as
" certain that this is a normal condition and must represent an early
stage in the evolution of a parasitic species, like those of Ccclioxys
and Stelis. From the standpoint of genetics, it is an extraordinary
case, since the female seems to have dropped her secondary sexual
characters and thereby assumed those of the male which were pres-
ent in her genetic constitution." He adds that " presumably the
male of A. suhrixator cannot be distinguished from Mcgachile
suhrixator," which is a common species in the Philippines and in all
probability the host.

Another peculiarity of the parasitic bees, to which Friese has
called attention, is their often very vivid coloration. Many of the
species are more or less red or yellow {Sphccodcs, Nomada, Epeo-
his, etc.) or banded and spotted with patches of white or blue
appressed hairs or scales {Epeolus, Crocisa, M electa, etc.), or are
brilliantly metallic (Excerete, Aglac). The red color suggests that
of certain myrmecophilous beetles {Eomcchusa, Hetccrius, Clavigcr,
etc.) and may have a similar meaning, but it is difficult to account
for the spots and bands unless we assume that they are an expres-
sion of peculiarities of metabolism, associated with the active habits
of the parasites, an interpretation which has also been suggested to
account for the more vivid color patterns of the males as contrasted
with the cospecific females of many animals. Perhaps the pecu-
liar odors of certain parasitic bees, c. g., of Nomada, odors which in
some cases at least seem to play a role in the relations to the host,
also point to such peculiarities in metabolism.

From an examination of the brains of two genera of parasitic
bees, Nomada and Psithyrus, von Alten (1910) concluded that
their fungiform bodies, supposed to be the seat of intelligence and
therefore to correspond to our cerebral hemispheres, were more
feebly developed than in the nonparasitic species. He even found
that the fungiform bodies of the male parasites were relatively

WHEELER— THE PARASITIC ACULEATA. 5

larger than in the females. It can hardly be claimed, however,
that the parasitic bee is psychically less endowed than its host, be-
cause the finding and entering of the latter's nest presupposes in-
stinctive activities of a high order.

Our knowledge of the habits of parasitic bees is rather meager
when compared with our knowledge of the species as taxonomic
units. They occur in all parts of the world, but even the hosts of
many of the genera, especially of the nonholarctic forms, have not
yet been ascertained. Within recent years, however, Verhoefif
(1892), Hoppner (1904), and Graenicher (1905) have made some
careful observations on the behavior of a few European and North
American species. Verhoeff studied the activities of the Stclis
minuta larva in the nests of Osjuia Icncomelcrna which are in hollow
blackberry stems. The Osmia makes a row of cells in the cavity,
separating them with partitions of chewed up leaves, provisions each
cell with a ball of honey-soaked pollen, the socalled " bee-bread,"
and lays an egg on it before closing the cell and starting another.
Graenicher summarizes Verhoeff's observations in the following
words :

" I. Stclis minuta deposits its egg earlier than the host-bee, and in the
lower region of the bee-bread. 2. The larva of the parasite hatches a little
earlier than that of the host-bee, whose egg is situated on top of the bee-
bread. 3. Botli larvae, wliich at the beginning are of about the same size,
partake of the bee-bread, the host-larva on top, the parasite below. 4. The
latter gradually' increases in size, and consequently advances towards the
host-larva on top. 5. Finally the parasite, which in the meanwhile has become
twice as large as the host-larva, comes in contact with the latter, kills it and
eats it. Verhoeff informs us that there was a mutual exchange of hostilities
between the two larvae, each trying to grab the other with its mandibles, but
that finally the parasite succeeded in burying its mandibles in the head of the
host-larva. The latter was eaten up within i or 2 days."

Hoppner's observations on the larva of Stelis ornatula in the
cells of Osmia parvida and lencomelccna agree essentially with Ver-
hoefif's, except that he saw no struggle between the parasitic and
host larva. The former bored its way upwards through the bee-
bread, sought out the Osmia larva as soon as possible and plunged
its mandibles into the body of the latter without meeting with any
resistance. Like Verhoefif he found the parasitic to be larger than
the host larva.

6 WHEELER— THE PARASITIC ACULEATA.

Graenicher's observations are more extensive. He studied in
Wisconsin the parasitism of Stelis 6-maciilata on Alcidamia pro-
ducta, of Coclioxys lucrosa on Megachile addenda and of Triepeo-
liis helianthi on Melissodcs trinodis. In all these cases the general
behavior of the parasite is very similar to that of Stelis niinuta and
ornatitla, but he found that the just-hatched larva has sharp, falcate
jaws, which are very large in Calioxys and Triepeolus and are re-
placed by smaller jaws with the next moult, after it has killed the
host larva. The first stage Triepeolus larva, moreover, has pecu-
liar leg-like appendages which enable it to crawl about in the cell.
We are justified, therefore, in speaking of a hypermetamorphosis
in these bees, comparable to that of so many other parasitic insects
(Rhipiphoridse, Strepsiptera, Meloidae, Eucharine Chalcididse,
Chrysididas, Mantispidse, etc.). I quote a portion of Graenicher's
account relating to the Stelis larva.

"July 9, 1903. Nest collected at Milwaukee contains 4 cells. Third cell
(from below) with a parasite. On top of the bee-bread an Alcidamia larva,
about 3 days old. On the side of the bee-bread, about half way up a Stelis
larva feeding on bee-bread. It is smaller than the host larva, and its head is
directed upward, and toward the posterior end of the latter's body.

" July 13. The parasitic larva has grown considerably but is not as large
as the host larva. At i P.M. the parasite moves upward a short distance,
comes in contact with the host larva, and secures a hold on the latter's side
behind the middle of the body. The victim at first makes an effort to free
itself, but offers no serious resistance. The parasite remains in the same
position the whole forenoon, sucking the liquid contents of the host's body.
The latter gradually perishes and shrivels.

" July 14. The parasite has released its hold on the dead host larva, and
is feeding on bee-bread. It has lately increased very much in size. From
now on the parasite does not pay any more attention to the remains of
the host.

" In the cell just considered a single parasite was present, but in a nest
collected at Milwaukee, July 15, 1903, a cell was come across with 3 parasitic
larvae, all of them on the same side of the bee-bread as the head of the
host larva. One of them was sitting above the middle, not far below the
host larva, the second was lower down and directed laterally, and the third
was below the second and quite close to it. In the evening the third parasite,
which throughout the day (July 15) had been partaking of bee-bread and
growing in length, reached the second and killed it. Four days later this same
parasite killed the uppermost one and fed on its contents. Two days after
this (July 21), the surviving parasite killed the host larva. Both were about
equal in size."

WHEELER— THE PARASITIC ACULEATA. 7

These very similar observations of Verhoeff, Hoppner and
Graenicher on three very different genera of parasitic bees cast
some doubt on the older and more meager observations which led
Schmiedeknecht and Sharp to assume that the parasitic bee larva
is merely a commensal that feeds so voraciously and grows so fast
that it compels the host larva to perish from starvation. It was
this assumption which led the earlier writers to call the parasites
"cuckoo bees." It is possible, of course, that some parasites, e. g.,
Noniada, which infests the nests of Andrcna and Halictus, may
conform to this older view, but renewed investigation is certainly
demanded by the results of the authors I have been considering.

Graenicher (1906) has also made some valuable observations
which show that vision as well as odor is an important factor in
the parasitic bee's method of locating the nest of the host. Speak-
ing of Argyrosclenis minima, which is a parasite of Collctcs eiiloplii,
he says :

" It is quite evident that after having discovered the nest this parasitic
bee pursued a course similar to that of a host-bee when constructing a nest.
It started out to make a careful and repeated inspection of the environment
of the nest, gradually covering more territory in different directions, but
often returning to the nest as the main object of its attention. Being pos-
sessed of a good memory for visual impressions it became acquainted with
the locality within 6 minutes, and experienced no difficulty in refinding the
nest at its next visit after an absence of 14 minutes. It gradually acquired
a thorough familiarity with the topography of the region, and on its return
to the nest it was seen to fly towards the opening as directly as the owner
itself.

" Such a parasitic bee when hunting for a nest of a host-bee is not always
flying around in a haphazard way, trusting to its good luck in finding a nest
here today, and one somewhere else tomorrow. When it has come across a
suitable one it is very careful to keep this under observation, and in making
its trips to and away from the nest it is directed by its visual memory in
exactly the same manner as the host-bee itself. It would not be in the inter-
est of such a bee to pursue a different course. The work of the host-bee in
constructing a cell, and provisioning it with the food-supply must have pro-
gressed to a certain point before the parasitic bee may find it suitable for the
reception of the latter's egg. For this reason such a bee has to make re-
peated visits to the nest, in order to be on hand when the right time comes.
If it were in the habit of wandering around until it happened to come across
a host-bee's cell in the proper stage of construction, then it might not get
much chance to deposit an egg within its life-time of a few weeks duration,
especially in rainy seasons. It is even possible that a parasitic bee has more
than one nest under observation during the same period."

8 WHEELER— THE PARASITIC ACULEATA.

Graenicher found the behavior of the female Tricpcohis, Ca:lioxys
and Stelis to be very similar to that of Argyrosclcnis.

There is one genus of bees, Sphecodcs, which must be briefly
considered, because it has been the center of a prolonged contro-
versy. These insects are fairly common in Europe and North Amer-
ica and closely resemble the species of Halictus except in color, as
they have the abdomen wholly or in part vivid red and in the hind
tibiae which are very sparsely pilose and therefore suggest a degen-
erate condition of the pollen-collecting apparatus. More than a
century ago de Walkenaer (1817) maintained that Sphecodcs is a
parasite of Halictus, and the same view was more or less emphat-
ically maintained by Wesmael (1835), Lepeletier (1841) Spinola
(1851), and Taschenberg (1866), but Fred. Smith (1851) and
Sichel (1865) held that it nests independently. The controversy
continued, however. Perkins (1887, 1889) believed that Sphecodcs
might be occasionally parasitic and Friese and von Buttel-Reepen
(1903) regarded it as perhaps incipiently parasitic. Rudow (1902)
repeated the old statement that it nests independently. Marchal
(1890, 1894) and Ferton (1890, 1898) witnessed some serious com-
bats between Sphecodcs and the Halictl, whose nests it was trying
to enter. Ferton (1905) saw a Sphecodcs suhquadratus breaking
into the nest of Halictus malachurus.

" Not being able to seize bj^ the head the sentinel bee which barred her
passage, she tunneled towards the Halictus burrow and succeeded thus in
seizing and kilHng the guardian, which she tossed backward out of the bur-
row. A second and a third Halictus that rose in the burrow in succession to
replace the first, met the same fate."

Morice (1901) contended that such aggressive behavior showed
that the Halictus was not a parasite, because some parasitic bees,
e. g., Notnada, seem to entertain friendly relations with their hosts.
Three investigators, however, have succeeded in breeding Sphe-
codcs from Halictus and Andrcna nests. Breitenbach (1878) long
ago took S. rubicundus from the brood-cells of Halictus 4-cinctus
and Sladen (1895) found pupae of the same species in the nests of
Andrcna nigroccnca and labialis. Finally Nielsen (1903) gave co-
gent reasons for regarding S. gibhus as a parasite of Halictus
4-cinctus. He says :

WHEELER— THE PARASITIC ACULEATA. 9

" When quickly unearthing a nest which I had seen Sphccodcs entering,
I discovered it sitting in a cell nearly filled with honey. Later on I found
several cells containing larvse differing from those of Halictus and which
can hardly be other than those of Sphccodes. Finally I found in the autumn
a cell containing a dead specimen of a fully colored Sphccodcs pupa. It is
therefore proof that Sphccodcs is a cuckoo with Halictus."

Nielsen calls attention to the fact that the parasitic habit of
Sphccodes explains the great variations in size, puncturation, etc.,
which have led taxonomists to multiply species in the genus. He
found that poorly nourished individuals are often only half the
size of well-fed specimens. Perkins had previously noticed that
small forms of Sphccodes live with small Halicti and vice versa.
Sichel, after studying 3,200 specimens of European Sphccodcs,
decided that they represented only three species. He sent 600
other specimens which he referred to two species to Foerster, who
claimed that he could distinguish some 150 species among them,
but wisely refrained from publishing descriptions. Similar varia-
tions are, of course, frequent in other parasitic insects, notably in
Mutillidse and in Ceropalcs.

The aggressive behavior of the female Sphccodcs, which was
also observed by Nielsen, suggests that she may enter Halictus cells
which are already completed and destroy the egg of the host, so
that her own progeny will not have to compete with the lawful
owner of the bee-bread, as in the case of Stclis and the other para-
sites studied by Graenicher. At any rate our knowledge of the
behavior of Sphccodes is in need of further careful investigation.

When commenting on the difficulties encountered by the taxo-
nomic student of the bees, I omitted one of the greatest, viz., that
presented by the numerous parasitic genera. In many cases these
are known to be very closely related to the genera of their hosts, a
fact which was noticed even by the early investigators, although its
full significance became apparent only in the course of time, with
the constant discovery of new species and genera in all parts of the
world and with changes in the interpretation of general biological
phenomena. The whole matter is so interesting that I may be
pardoned for introducing some historical considerations.

The more than a century devoted by entomologists to the study
of bees may be conveniently divided into a pre- and a post-

10 WHEELER— THE PARASITIC ACULEATA.

Darwinian period. Latreille, in a short paper, published in 1802
at the end of his remarkable volume on the habits of ants, and
Kirby in the same year were the first to construct noteworthy classi-
fications of bees. There was a remarkable agreement in their
point of view, both dividing the family into short-tongued and long-
tongued forms, subsequently called Andrenida; and Apidse, the
Andrenatse and Apiaires of Latreille and the supergenera Mclitta
and Apis of Kirby. The parasitic bees that were known in his day
were intercalated by Latreille among the Apiaires in close proximity
to their host genera. Lepeletier de St. Fargeau (1825) divided the
bees into two groups, the " recoltantes," or collecting, and the "para-
sites," and subdivided the former according to the dififerences in
their pollen-collecting apparatus. The views of Latreille and Lepe-
letier have dominated the classification of bees down to the present
time. Certain German melittologists, notably Schmiedeknecht and
Friese, have followed Lepeletier's scheme, whereas Westwood
(1840) and most subsequent workers have agreed with Latreille.
As Westwood's reasons are still interesting and include a good
statement of the pre-Darwinian or special creation conception of
the relations of the parasite to the host, I quote some of his re-
marks :

" Indeed it is to be observed that the variation in the structure of the
species, thus varjang in their habits, does not seem to warrant the establish-
ment of them into separate famihes. This circumstance appears naturally
dependent upon tw^o considerations: ist, it is essential that the parasite in its
perfect state should possess a certain resemblance to the animal in the nest
of which it deposits its eggs, so as to deceive the latter and its associates
(Kirby in a footnote here calls attention to the resemblance of the Dipteron
VoluccUa to Bombus) ; and 2d, the nature of the food of both being similar,
the variation in structure is much less striking than if the parasite were car-
nivorous, as the Ichneumonidae, and the animal attacked (as the caterpillars
of Lepidoptera, etc.) herbivorous. The parasitic connection indeed goes no
further than this, viz., that the larva of the parasite eats up the food of its
fosterer, and so starves it to death ; the larvae of both are therefore pol-
lenivorous, and the dififerences which will naturally be most striking, will con-
sequently be found in those organs which are emploj'ed in the construction
and provisioning of the nest of the working species, and which one may
therefore expect to find in a less developed state than in those species wiiich,
from being parasitic, do not require their full development. Hence it is
that we find the general structure of the parasitic bee closely resembling
that of the bee, at the expense of whose young its own are destined to be

WHEELER— THE PARASITIC ACULEATA. 11

nourished; and hence, if we regard Bojiibus and Psithyrus of St. Fargeau,
Aglae and Euglossa, Mclccta and Anthophora, or Sphccodcs and Halictus,
with reference to their general structure, they will be found most intimately
allied; whilst if, on the other hand, we regard such portion of their economy
is is connected with the formation and provisioning of their nests, it will be
requisite to place them in different divisions. If we observe, however, the
great variation existing among bees in this portion of their economy, it is
evident that this cannot be regarded as a normal or typical character and
that a distribution founded thereupon would necessarily be unnatural."

The publication of the " Origin of Species " could not fail to
have its efifect on the students of bees. In the light of evolution
the parasitic species acquired a new meaning, for it was at once
apparent that their resemblance to their hosts might have a genetic
significance. One of the first to fall under the spell of the new
conception was Hermann Miiller (1871). He believed that the
genus PsifJiynis was of rather recent descent from its host genus
Bomhus, that Melecta and Crocisa were less recently descended
from Anthophora, and that the phylogenetic origin of Stelis, Cocli-
oxys, Epeolus and Nomada was still more remote, although the
derivation of Stelis and Ca^Iioxys from gastrilegid genera seemed
clear. He was guided to these conclusions by a study of the an-
tennae. Referring to a table of the genera of bees he says: "An
examination of this table shows that in all nonparasitic bees, with-
out exception, the males have a shorter scape but a longer flagel-
lum than the females, but that in some pronounced cuckoo-bees the
very reverse is the case." Smell not only guides the males to the
females, but also the parasites to their hosts and hence the olfac-
tory organs of the female parasites must be highly developed. " A
glance at the development of the male and female olfactory organs
of the cuckoo-bees clearly supports the conclusion that in the an-
tennae of the females the adaptations for working in the brood-
chambers have been lost pari passu with an increase in the olfactory
organs and that these developments correspond in degree to the
period of time at which the transition to a parasitic life took place."
I am not aware that any study of the antennae and their sense-
organs has since been undertaken with a view to testing the correct-
ness of Miiller's conclusions. Allusion has already been made to
Graenicher's discovery that the vision of the parasitic bees is an
important factor in locating the nests of the host.

12 WHEELER— THE PARASITIC ACULEATA.

In 1883 Perez published an important study of the parasitic
bees and republished his general conclusions in a separate article in
1884. After careful morphological investigation he concluded that
the parasitic genera must have evolved from the host genera and
was able in a few instances to point out the very species from which
the parasitic genus had originated. He recognized four distinct
lines of development from as many host genera: Pskhyrus from
Bombus, Sfelis from Anthidium, Ccclioxys and Dioxys from Mega-
chile and Sphecodes from Halictiis. The series of genera known
as the Nomadinae and comprising Epeolus, Melecta, Crocisa, Am-
mohatcs, Pasitcs, Philcrcmus, Biastcs and Nomada, he derived from
Ccclioxys on the supposition that this genus had given rise to a
whole series of parasitic forms which had acquired new hosts among
the various genera of recoltant bees. He contended that Latreille's
example in placing the parasitic genera next to their host genera
should be followed in any attempt at a natural classification of the
Apidae. The truth of his contention has since been conceded and
is clearly expressed in the classifications of Ashmead ( 1899) , Robert-
son (1899) and Cockerell (1910). Dalla Torre (1896) and Friese
in his work on the African bees (1909), however, adopt a com-
promise between the views of Latreille and Lepeletier, dividing the
bees into podilegid, gastrilegid and social sections and appending
to each a series of parasitic genera. Of the classifications I have
seen Robertson's seems to be the most natural, but he is dealing
with a limited fauna, in which the relations of the parasitic genera
are few and fairly well known, whereas Dalla Torre and Friese,
in an attempt to deal with the bees of remote regions or of the
whole world and with dozens of imperfectly known parasitic gen-
era, have some justification for the course they adopt. It is evi-
dent, nevertheless, that no satisfactory classification can be con-
structed till the precise affinities of all the parasitic genera to one
another and to the host genera have been thoroughly elucidated.

The phylogenetic relationships even among the European and
North American parasitic bees are still in part very problematical.
Probably all agree that Psithyrus must be derived from Bombus,
Stelis from Anthidium (scnsu lafo), and Ccclioxys from Mcgachilc,
or some closely related, now extinct, genus. But the origin of the

WHEELER— THE PARASITIC ACULEATA. 13

Nomadine series is by no means clear. Friese in his earlier work
(1889) could not decide whether it was to be derived from Cocli-
oxys, as Perez suggested, or from a form like Eucera. To-day
even such an alternative seems too simple, for in all probability the
long series of " Nomadine " genera now known consists of several
heterophyletic groups. M electa and Epeoliis are derived from A)i-
tJwphora by Robertson and others, and Saunders and Robertson
would derive Nomada from Andrena, whereas such genera as Am-
mobates, Biastes, Pasites and Phiariis are now supposed by Friese
( 1916) to be connected with Megachilc through genera like Cccsarea
and Paraccclioxys, the last being also the source of Ccclioxys and
of Dioxys and Paradioxys through the genus Prodioxys. Among
the exotic parasitic genera it seems clear that some have arisen from
host genera very dififerent from those above mentioned. Thus there
is every reason to suppose that Thalestria has arisen from Oxcea,
Agla'e and Exccrete from Euglossa, Eucondylops from Allodape,
Peresia from Osmia.

It will be seen, therefore, that even if we make all due allow-
ance for dubious cases there still remain a number in which the
closest morphological affinity of the parasitic is with its host genus.
This is evident from the accompanying table (Table I.) in which
the most clearly established cases (fully 50 per cent.) are marked
with an asterisk. In constructing this table I have profited by a
number of valuable suggestions kindly communicated by Professor
Cockerell. We must assume, I believe, that in some cases the
primitive host genera are now extinct, that in some cases, there-
fore, the parasites have come to infest species of genera to which
they have no morphological affinity, that many parasites are directly
derived from other parasitic genera and that in some cases the
phenomena of parasitic convergence are so pronounced and oblit-
erate or obscure the generic affinities to such a degree that they can
be elucidated only by the most painstaking study. For my imme-
diate purposes, however, the present results will suffice, since they
agree with the conditions in other groups of Aculeata, as will be
shown in the sequel.

In marked contrast with the bees, the solitary wasps comprise
few parasitic species, if we exclude the Mutillidae, which I am not

14 WHEELER— THE PARASITIC ACULEATA.

TABLE I.
Genera of Parasitic Bees.

Parasites. Hosts. Ancestral Genus.

*Nomada Andrcna, Halictus, Encera, Collctcs,

Panurgus Andrcna.

*Sphecodcs Halictus Halictus.

*Parhalictus ( ^Halictus Halictus.

*Chlerogas {?)Thrinchostoma Thrinchostoma.

*Mclecta Anthophora Anthophora.

*Bombomclccta ( ?) Anthophora Anthophora.

*Ericrocis ( ?) Anthophora Anthophora.

*Crocisa Anthophora Anthophora.

*Protomelissa ( ?) Anthophora Anthophora.

*Melissa ( ?) Anthophora Anthophora.

Epcolus Collctcs Anthophora.

Tricpcolus Melissodcs, Tctralonia Epcolus.

Argyrosclcnis Collctcs Epcolus.

*Epcoloidcs Macropis Macropis.

*Lciopodus Melitoma Melitoma.

Osiris (f) Tetrapedia Tetrapedia.

Rhathymus ( ?)Epicharis Epicharis.

Mesochcira {?) Ccntris Crocisa.

Acanthopus Ccntris {?)Ccntris.

Eurytus Ccntris ( .?) Ccntris.

Mesonychium Melitoma Acanthopus.

*Aglae Euglossa Euglossa.

*Exaerete Euglossa Euglossa.

*Peresia Osmia Osmia.

*Eucondylops -illodapc 41lodape.

*Stelis Anthidiuni, Chalicodoina, Heriadcs,

Osmia, Ceratina, Alcidamia, Chilos-

toma -Inthidium.

Parcvapis Mcgachilc Stclis.

Euaspis Mcgachilc Stclis.

*Thalcstria O.vcca Oxcca.

*Androgynclla Mcgachilc Mcgachilc.

*Coelioxys Mcgachilc, ( ?) Anthophora Mcgachilc.

Dioxys Osmia, Chalicodoma Mcgachilc.

Pasitcs Nomia, Camptopccuni Ammobatcs.

Ammobatcs .Anthophora, Macroccra, Saropoda.. . Cccsarca.

Biastes Systropha Ammobatcs.

Phiarus Mcliturga Ammobatcs.

Holcopasitcs ( ?)Calliopsis Anunobatcs.

Orcopasites Spinoliclla Anunobatcs.

Philcrcmus Rhophitcs, Halict aides Pasitcs.

Herbstiella ( ?)Psanythia Pasitcs.

*Psithyrus B ambus Bombus.

WHEELER— THE PARASITIC ACULEATA. 15

including in my survey. The following are the only cases I have
found in the literature. According to Ferton (1901) the Gorytid
Nysson diiiiidiafits is a parasite of Gorytcs elcgans. The latter digs
its burrow in the sand and provisions it with larval and adult Hem-
iptera; the N\sson finds it and often enters it during the absence
of the Gorvtes. If the latter happens to be at home the Nysson
waits motionless about a dozen centimeters away, with its head
turned towards the nest, till the Gorytcs departs. Adlerz (1910)
observed very similar behavior on the part of Nysson maculatus
towards Gorytcs lunatus. Apparently both species of Nysson de-
stroy the Gorytcs egg attached to the prey and lay their own in its
place. In 1887, at a time when nothing was known of the parasitic
habits of Nysson, Handlirsch called attention to the superficial re-
semblance of some of the species to parasitic bees.

Williams (1913) and the Raus (1918) have described an inter-
esting sporadic case of parasitism in Sticus unicinctns, a wasp be-
longing to a very different family, the Bembicidse. The Stiziis digs
its way into the nest of a Sphecid, Chlorion thomcr, after the latter
has provisioned it with a cricket, oviposited and closed the entrance.
After the Bembicid has entered the chamber it devours the Chlorion
egg and deposits its own so that the larva can have the cricket all
to itself. This case is extraordinary because the other species of
Stisns {S. tridcns and crrans) , whose habits have been studied by
Fabre (1886) and Ferton (1899, 1908, 1910, 191 1), dig their own
burrows in the sand, glue their egg to the bottom of the cell and
feed the hatching larva continuously with Hemiptera after the man-
ner of other Bembicids (Bicyrtcs). Ferton has also observed
similar behavior in S. gazagnairci and fcrtoni. According to the
same observer (1899, 1901, 1908) S. fasciatns feeds its young with
immature crickets. Our American Stizus with the exception of
unicinctns, seem to have similar habits. During the summer of
1 91 7 I saw a flourishing colony of a small undetermined species
near Tempe, Arizona. It comprised thousands of individuals, all
nesting close together in the sand, like Bcmhix.

The remaining parasitic wasps belong to the family Psammo-
charidse (Pompilidse), all of which prey on spiders. In two of his
papers (1890, 1891) Ferton has shown that some individuals of

16 WHEELER— THE PARASITIC ACULEATA.

Pompilus rufipes (now called Psammo chares) have acquired the
habit of robbing other individuals of their prey which they then
bury and furnish with an egg. They even wage fierce battles for
one another's spiders. These observations acquire added interest
in connection with another very closely related species, P. pectinipcs,
which, according to Ferton (1901, 1902, 1905), enters the sealed
nests of P. rufipes, eats its tgg and deposits its own on the spider.
Ferton was thus led to advance the opinion that we hav2 in pecti-
nipes a parasite that has just become detached phylogenetically from
its host species.

" The parasitic habit," he says, " would therefore appear to have been
built up in the following manner : P. rufipes, living in colonies, has acquired
the habit of stealing the prey of its neighbor and even of fighting for the
possession of prey not its own. Some individuals finally learned to steal the
spiders that had been buried, either by driving away the rightful owner while
she was sealing her burrow, or by ferreting through the soil occupied by the
colony in search of sealed burrows. Their descendants, inheriting this habit,
gave up constructing a nest and transporting the stolen prey to it and left it
in the cell where it was discovered, simply substituting their egg for the one
it bore. Thus P. pcctinipes was evolved, scarcely distinct from the maternal
stock in many of its anatomical characters but become a parasite on the spe-
cies from which it arose."

In Sweden Adlerz (1910, 1912) found that P. campestris ex-
hibits a similar parasitism on P. unguicularis and P. acitleatits on
P. rufipes and funiipennis, and Ferton (1891) has shown that P.
viaticus and pulcher resemble rufipes in their habit of appropriating
the prey of other individuals of their own species.

Finally we have among the Psammocharids a distinct and pecu-
liar genus, Ceropalcs, all the species of which are parasites on other
genera of the family. Lepeletier (1827) was the first to regard
Ceropales as a parasite, but Walsh was the first to breed it from
the nest of another Psammocharid. Riley and Walsh (1869), in
their paper on wasps and their habits, state that a male Ceropales,
which they described under the name C. rufiventris, but which is
now known as C. robinsoni Cresson, emerged from a mud cell that
had been constructed and provisioned by Agenia homhycina. That
they were fully aware of the importance of this observation is clear
from the following remark:

WHEELER— THE PARASITIC ACULEATA. 17

" The inference is unavoidable — more especially as we had previously
bred very numerous specimens of the same little mud-dauber from the same
kind of mud-cells obtained in northern Illinois — that this gaily dressed
Spider wasp (Ccropalcs) had sometime in the summer of 1867, laid an egg
in one of the five mud cells found in south Illinois, and thus appropriated to
the use of its future larva the supply of food laid up by the provident care
of the unfortunate, dingy looking little mud-dauber for its own offspring.
Otherwise it is impossible to account for two distinct kinds of wasp hatching
out from the same lot of mud cells."

Perez (1894) and Ferton (1897) made some very interesting
observations in France on the behavior of Ccropales maciilata and
cribrata, showing that these wasps are parasitic on various species
of Psammochares and Aporus, and Adlerz (1902) has succeeded in
giving us a complete accoimt of the behavior of C. maculata as he
observed it in Sweden. This behavior is so interesting, especially
in connection with Graenicher's observations on the parasitic bees,
that I subjoin a translation of the German resume of the paper:

" Ccropalcs has the habit of visiting the breeding grounds of Pompilus
species and there alights on small eminences of the soil in order to spy on the
Pompilids while they are dragging in. their paralyzed spiders. The tense atti-
tude of the wasp, her deflected antennae and her movements as she turns
towards a Pompilid that has just come within the range of her vision, are
indicative of her keen interest. As Perez and Ferton have observed, the
Ccropalcs either alights on the spider while it is being borne along by the
Pompilus, unobserved by the latter, or on a spider that is lying unguarded
in the open or concealed above the ground, while its captor is busy digging
her nest. In either case the Ccropalcs can be seen bending the tip of her
abdomen under the spider for the evident purpose of ovipositing. Ferton
saw a Ccropalcs cribrata follow a Pouipilus chalybeatus into her burrow
while she was dragging in a spider, but although a Ccropalcs larva was
afterwards found on the prey, it is not certain that the egg was laid on this
occasion. As will be seen from what follows, it might have been laid pre-
viously. The only time I saw a Ccropalcs enter a Pompilus burrow was
when a P. niger was still busy excavating. No spider was therefore on hand
and, of course, oviposition could not have occurred. It was merely a sign of
impatience on the part of the parasite, which, after persistently watching the
digger, stole down into the burrow as if to inspect the progress of the work.
I was present also on a second exceptional occasion when a Ccropalcs
pounced down with such violence on a P. cinctellus with its spider that the
two wasps and the spider tumbled about together. The little P. cinctellus
was so dismayed that she flew away in great haste and never returned. On
this occasion the egg which the Ccropalcs probably laid during her subse-
quent tedious manipulation of the spider must have perished, because the
spider was left in the open where it was exposed to ants and other predatory

PROC. AMER. PHIL. SOC. VOL. LVIII, B, JUNE II, I919.

18 WHEELER— THE PARASITIC ACULEATA.

insects. That the Poinpilus suspects the hostile intentions of the Ceropales
is clear from the behavior of a P. viaticus that hid with her spider among
the dense grass-blades of a road-side and would not venture into the open
because she was being watched by two female Ceropales each perched on a
grass-blade, stretching its antennse downward and edging nearer from time
to time. The angry Pompilus finally gave chase to the parasites and only
after they had flown away did she leave her concealment with her prey.

"When a spider on which a Ceropales has just alighted is examined,
the egg cannot be seen at first because it is placed in such an unsuspected
spot. At the base of the ventral surface of the abdomen the spider has two
slit-shaped stigmata which open into the lung-books. The wasp inserts her
egg into one of these. The stigmata look like pockets, with very closely fit-
ting flaps. After the egg is in place the orifice of the pocket sometimes
gapes slightly so that one end of the egg can be seen. This is apt to be the
case in Drassodes, but in the large Lycosids the pockets are so capacious
that they completely conceal the egg. The place is obviously selected be-
cause in it the egg is perfectly protected when the spider is later dragged
into the burrow by its rightful owner, for it is evident that if the egg were
merely attached to the surface, it would be exposed to serious injury while
the spider is being drawn through the narrow burrow. The last abdominal
segment of the female Ceropales, which is constructed like a short, flat,
truncated ovipositor — a structure vmique among the solitary wasps — evidently
represents an adaptation to the narrow, slit-shaped stigmata, since the latter
can be opened by means of such an instrument and the egg inserted. Not
infrequently I have seen an egg in each of the lung-books of the same spider.
Since the Pompilus later attaches its own tgg to the side of the spider's
abdomen, the situation becomes very complicated, as there are then three
rival claimants for the same spider which is sufficient food for only one. A
few successful breeding experiments have revealed the drama that is subse-
quently enacted in the dark burrow.

" After an embrj'onic period of two to three days, the Ceropales larva
hatches. Its anterior portion, as far back as the tenth segment, extends
straight out from the stigma, while its posterior portion remains concealed
in the lung-book. Soon the exposed portion is seen to bend downward till
the head touches the spider's belly and the larva begins to feed. As soon as
the Ceropales in the other lung-book hatches the older larva evidently smells
a rival, for it stops feeding, stretches itself out and moves its anterior end
freely about in the air in the direction of its competitor. The latter is at
first out of reach, but as soon as the older larva has fed and grown suffi-
ciently in length it attacks its younger companion, which is quite unable to
escape its fate. After its cannibal feast the Ceropales larva again bends
down and continues to devour the spider. Not till several days have elapsed
does the Pompilus larva hatch, although the egg was laid not more than a
few hours after the Ceropales egg. When the Pompilus larva begins to
grow the Ceropales larva becomes aware of a new rival and turns in its
direction. When a little later it has grown sufficiently to reach the Pompilus
larva, the latter's fate, too, is scaled. With the elimination of its last com-

WHEELER— THE PARASITIC ACULEATA. 19

petitor the Ccropales larva turns again to the spider and devours it com-
pletely except for a few unassimilable remnants. Then the larva weaves a
network of pale brown threads among which on the following day it spins a
pale brown cocoon. In one case which I observed the feeding period of the
larva extended over a period of 12 days."

It will be seen that the general ovttlines of the behavior of the
solitary parasitic bees and wasps are strikingly similar. Among the
wasps we can recognize two types, that of Nysson, Stisus unicinctus
and Psammochares pectinipes and that of C^ro/'o/^.y, whereas in bees
only a single type, bearing a great resemblance to that of Ccropales,
is known. It is probable, nevertheless, as I have indicated above,
that the Nysson type may be represented among the bees by SpJic-
codcs. The two types are shown in the accompanying diagrams in
which the main activities of the parasite and its host are repre-
sented in parallel series.

Nysson.

Nyssoji Mating Finding Host Nest

Goryfes Mating Nidification, Provisioning, Oviposition.

Destroying Host Egg Ovipositing Larva Appropriating Prey.

Ccropales.

Ccropales Mating Finding Host and Prey, Ovipositing.

Psammochares Mating Provisioning, Nidification

Larva feeding. . . .Killing Host Larva. . . .Appropriating Prey.

Oviposition Larva feeding

Stelis.

Stelis Flower visiting Mating Finding Host Nest, Ovipositing.

Alcidamia Flower visiting Mating Nidification, Provisioning

Larva feeding Killing Host Larva Appropriating Food.

Ovipositing. . . .Larva feeding

In all the cases the parasite takes possession of the food-supply
(prey or bee-bread) by eliminating the egg or young larva to which
it belongs as a result of the activities of the host, but this elimina-
tion may be effected in two ways, either by the adult or by the larval
parasite. In the Nysson type the mother appropriates the prey and
bequeaths it to her offspring, in the Ceropales-Stelis type the larval

20 WHEELER— THE PARASITIC ACULEATA.

parasite seizes the prey or food for itself, or, regarding the situation
merely from the standpoint of the individual life-history of the
parasite, we may say that it is predacious either in its first larval
stage or as an adult on the egg or young larva of the host. The
host egg or larva constitutes an obstacle to the parasite's enjoyment
of the prey or bee-bread, and as the parasite is a true ir>sect bol-
shevik and member of the I. W. W. its life purpose is completely
expressed in the impudent imperative : " Get out, I want your
place." Nor is it surprising that long before the Russian Soviets
the parasitic wasps and bees had learned that the quickest way to
remove a living obstacle is to kill it.

There is some difificulty in deciding which of the two types of
parasitism represented in the diagrams is the more primitive. Prob-
ably the more aggressive Nysson type was the earlier as indicated
by the behavior of Psammochares rufipcs and pcctinipcs. On this
supposition the role of assassin, directed not only against the host
larva, but also against any competing larva of its own species, was
acquired later by the larval parasite as a result of neglect on the
part of the mother to destroy the egg of the host. The same type
of behavior, however, is also seen in many other insects when more
than one egg is laid by the mother in a very limited supply of food,
e. g., among the larval egg-parasites (Proctotrupids) and the cater-
pillars that live in the heads of composite flowers (Rabaud, 1912,
1914). In the larval egg-parasites the large, sickle-shaped jaws
are beautifully adapted for the purpose of killing competing indi-
viduals of the same species, and the similar mandibles described by
Graenicher in larval bees of the genera Stelis, C(rlioxys and Tric-
pcoliis are equally useful in destroying both the competitors of the
same species and the host larva.

The social parasites are most abundantly represented and have
been most extensively studied among the ants. The literature on
the subject is so voluminous that I am unable to deal with it here.
Much of it is cited in my ant book (1910), where the subject is con-
sidered in greater detail, and in the first volume of W'asmann's
" Gesellschaftsleben der Ameisen " (1915). As would be expected,
the conditions become very complex when a social organism such
as a colony of ants becomes parasitic on another colony. Among

WHEELER— THE PARASITIC ACULEATA. 21

the parasitic relationships four types can be recognized. One of
these, corresponding to Wasmann's category of " compound colo-
nies " is represented by a number of small species which live in
little nests that communicate with the nests of the host by tenuous
galleries. The two species bring up their brood separately, but the
workers consort with one another freely and amicably in the gal-
leries and chambers of the host. The relations of parasite and host,
where they have been determined, are much like those exhibited
between certain ants and their myrmecophiles (symphiles). The
most typical of these guest ants is our North American Leptothorax
cmcrsoni, the behavior of which I have elsewhere described in de-
tail. From the accompanying table (Table III.), in which all the
known guest ants and their hosts are listed, it will be seen that none
of the former is congeneric with its host. Emery's statement ( 1909 ) ,
however, that : " The myrmecophilous ants are not derived from
forms allied to their host species, but from other genera or even
from other subfamilies," is not strictly true, though in all proba-
bility the guest-ants have developed, as he contends, from preda-
tory thief-ants, of which quite a number of species are known to
nest in the walls of the nests of termites and larger ants and to prey
on their brood.

The three other types of parasitism, representing Wasmann's
category of " mixed colonies," are the slave-makers, temporary and
permanent social parasites, which agree in living so intimately with
their host in the same nest that the two species bring up their broods
in common. The differences between the three types is, however,
very striking when we follow the development of the parasitic
colony, although it is founded in every case by a single recently
fecundated female, or queen, that succeeds in entering and estab-
lishing herself in the nest of the host species. The queen slave-
maker, at least in species like Formica sangiiinea, breaks into the
host nest and appropriates and fiercely defends a portion of the host
brood till it matures and surrounds her with a number of loyal
workers, which are then able to rear the brood hatching from her
eggs. The workers produced by such a queen have the extraordi-
nary habit of making periodical, organized raids during the summer
months on other colonies of the host species (usually Formica

22

WHEELER-THE PARASITIC ACULEATA.

-; vj to 2

^ t^ t' ~

V V Q ^ Q

ffi y

CJ O O O U (J u

■:p '^ 'jn '-2 'i^ •- -2 '-'

s o o u

l-H

<

w

^

hJ

y

CQ

H

<

tn

H

OS

<

Ph

^ l-) "!

S "^

■^ Q

5^ s « -§
.^ ^ s ?

V S

V V C ^

a « 03 2

S ? -

I' i* V* v^ '** ^ ^

^ « « i 2 E -E

r; V V V V V V.

2 o o <^ ^ ^ ^

55 ■* "S S ■£ "S S

< iS, S, ^ fi. f^ o

kJ k4 kJ kJ k4 -^

ffi _o .O

(J o tJ _o _o

w
<

O O <-S "^ ^- ft"^ K* N

.-5^ <o -S •-

2 -i ^ o

to <o <^ rt ^

c « e ■— ' -S

2 •« •« -^

a

■«

5:

g

<-»

ri

•-*

s

s

S:

«

Q

s

<o

^

^

55

a

o

<A.

•^

ti

<-)

>^

rt

w

o

!^

Co

?-> ^. <o V> t-j

:2; :^ o, o, a.

U

;^ :^

rt 03 ^_ _- ^ --• •--
(U W o3 o3 rt 03 03

;z; ;z; CL, Cl, cu Ph Ph

Oi "

w K to ^ "^
^ 2 I c

fc; e « e

a H

•^ s

o ~

S H S 2 ^ 2 i^

X

:§ ^ ti.

^ Q^..

^ w

=: "42 fi »- b
^ s V "-« ^ .„

<o s =: =

S H

kJ I^) ^

s

s

-S

S
H

s

o

o

M

W

':::

O

>.

a

a

o

^

li.

(i.

Cin

u.

WHEELER— THE PARASITIC ACULEATA. 23

fusca or one of its varieties), and of carrying their larvse and pupae
home and permitting a certain number of them to hatch as " slaves,"
so that the colony is maintained as an intimate mixture of two spe-
cies, at least for a considerable period. The queen Polycrgus,
however, kills the queen of the host colony whose nest she enters
and is adopted by the workers, and the slave-making, or dulotic
raids of her offspring are even more perfectly organized than in
sangninca, since Polycrgus in all its phases depends absolutely on
the slaves, or host workers for its food, the rearing of its young
and the construction of the common nest. It will be noticed from
the table (Table IV.) that all the slave-making, or dulotic parasites
belong either to the same genera as their hosts or to closely allied
genera, though the latter represent two different subfamilies.

The recently fecundated queen of the temporary social para-
sites belonging to Formica species of the rufa, microgyna and ex-
secta groups, Bothriomyrmex, Lasius umhratiis and fuliginosus or
some species of Aphccnogaster, enters the nest of the host in a con-
ciliatory or at any rate non-aggressive manner, and after being
adopted by the workers, supplants the host queen, when she is killed
either by her own workers or by the parasite, which then proceeds
to produce her own brood to be reared by the host workers. The
offspring of the parasite, however, are not slave-makers, so that the
host workers gradually die off, leaving a pure and eventually flour-
ishing colony of the intrusive species. As shown in the table (Ta-
ble v.), all the temporary social parasites belong to the same genera
as their hosts, although these genera represent at least three differ-
ent subfamilies.

The queen of the permanent social parasites enters the host col-
ony in the same insinuating and conciliatory manner as the tem-
porary social parasite and is definitively adopted in the same man-
ner after the host queen has been eliminated, but the rate of devel-
opment of the parasitic brood is very rapid, so that adult males and
females are produced within the lifetimes of the host workers.
This development of the sexual forms is the more accelerated be-
cause the worker caste has disappeared among the permanent social
parasites, which represent the culmination, or, more properly
speaking, the level of the greatest "degeneration" (specialization)

24

WHEELER— THE PARASITIC ACULEATA.

o i

V; H

Q o o o '5 'S '5 'S "S
-h-c^-c^-c^C V V 5^ V

K '-P '-w (J 'V

. y u -^ .^

<

<

Oh

1-^ b

w

ID
Q

i

s

g

2

2

k4
<

o

o

S

•1

1

iz;

^j

Vj

s

^

e

«o

<-3

3

•-i

<;

*-

ji

v»

H

%

_s

is.

2

g

1- 7?.

^
■«

S :S •S 5 .2 2 .2

S o Ci o ~ 5 S

•b X ^ ^ i; hI. s^

'■J '^j '^ '^ <5 . O

K, ^ ^ ^ ti, Cl. t.

[i, li.

Oh <-)

Q. ^

.a

5^

"P

e « r ?

S V) a o .- .;i

o S S <, J) J>

I B B i i ^

o *^ *^ ^

*^■ <2 ? H e «

c> "2l o

o U, )<

V 13

~ S K ^ ^ "^

'-I '-^ ^->

5 S r r '-- €
u s c; -c; "t; "^

=: :^ ^

^ ^ 2

rF^'sS'-J'^'o'

fj -- ~

i? ^ --^ V

C ^ -^ *«?

s e e t

CX<^^L)UF~H>^'^

a y _o _rt

c 'c .y .y

rt

O

■^

'^

u

u

o

Q

;-.

V.

O

o

03

03

OJ

(U

OJ

• - i: u o ••-.

O ni o3";=^JiiJi; 03^^
iU(U<U':;no3o3o3«Jn!o34;

:2;:2;:z;uaHPHCiH:z;eua^:z;

<

>

CIh

►J

w

<

o

<

H

» s

* *- « fe b h

V t5 -a 5 2 "^

^ 2, § §> 5> 2

^ o^« ^ '^ s

^ ~ .e "S 'S O

■^ 5^2 2 _s

S "5 ^ ^ C ■*--

"^ jSt- ^- s^ V c

(^ ^ w u to f^

2 -S

^ -J

t.

5- .~a Z .«

« i^ S b

A, Q s a

"2 2

s: '^ ^ si *: a

Si ?: s -Q ni.

^ s^ • —

^ ^ =

I g 1 g 2 -g

to

s ^ »- '^ »* '■^

^ -^ ~^ '^ 'o -^

t?

t? <? ~

lo ^i. <^ V V O

e

««,"='

(i, '^ -^ U L) cq

-J

^ ^ ^

k.

WHEELER— THE PARASITIC ACULEATA.

25

1— (

?^

>

<

w

hJ

w

m

H

<

W

H

<
04

<

lin

Ph

o

s s s s

s

^;^_'T3^_"«_"Q o o o O Ci o ~ ^f-H

" 1 ! i I • I ! I ! ! ! ! ! I

Sh <j o t« '^j rt '-4^ '-m '-ti; o '-0 '-H '-M '-^ '-^

■jn "jn rt ^J -^ <J u o -r; o o o u u

s s s -

■■-J ^
-a: -«

0, Q,

^ « « ^ ^ ■« S

S .2 .s .2 .s .g .g ^ .s

'£! 'C C '51 'v 'v e "S!

n "^^^i "^x "^ ^ '^ ^ '^^ r r

~ S <^ -^

5 -5 ^ ^

>. "e
B

^ . ~ s

s • ~ v! '^ ^
Co tq a, Co ^

"Q .«a _«u _'a

~ -L' '^

? S "o *^

~ <o e

" 2 ^

'^ e ?■ '-^

C^ 2 O Ci

^^

S-i ^ '^

•Ci. JR, -Ci.

tq tej tq ^ n:;

"^

26 WHEELER— THE PARASITIC ACULEATA.

among the parasitic ants. The Hst of the species in the table (Table
VI.) shows that they all belong to a single subfamily, the Myrmi-
cinae, and, with the exception of Hagioxcniis, to genera differing
from though allied to their hosts.

The tables IV. to VI. are much more striking as illustrations of
the natural affinities of the parasites to their respective hosts than
the table of the bees (Table I.). This is undoubtedly due partly
to the fact that the ant-parasites are structurally much less sharply
distinguishable from their hosts and partly to the different views of
myrmecologists and melittologists concerning the scope and dignity
of the genus as a taxonomic category. The myrmecologist is being
so constantly impressed with the great structural variations that
may exist in the same colony of ants and often therefore among the
offspring of the same mother, that he is apt to be a " lumper" with
a vengeance, whereas the melittologist, especially on our radical and
progressive American continent, seems to develop a veritable pas-
sion for erecting new genera or even subfamilies on very minute
morphological characters. Thus Ashmead created a family Psithy-
ridse and Cockerell a subfamily Psithyrinse for the bees of the single
genus Psithyrus, although no one doubts that these insects are very
closely related to Bomb us, whereas no myrmecologist has dreamed
of placing the a.herrsint,worker\esspsLva.s\te Anergatcs atratulus even
in a distinct subfamily, although it differs much more profoundly
from Tetramorium and other Myrmicine genera than Psithyrus
from Bombus. No melittologist, moreover, ever thinks of placing
a new parasitic bee in one of the known genera of recoltant bees,
because the absence of the collecting apparatus is tacitly assumed
to have decided generic or even subfamily value, but among the
ants there are several genera ( Formica , Lasius, Aphccnogastcr, Crc-
matogaster, Lcptothorax) which are made to include both parasitic
and nonparasitic species, because there are no morphological char-
acters by which they can be satisfactorily distinguished. -

Leaving out of consideration the guest ants, the origin of which,
as we have seen, can be accounted for in the same way as the myr-

2 Forel has, indeed, placed Formica sangtiinca in a separate subgenus,
Raptiformica, and its slave in another sugenus, Servifonnica, but in my
opinion without sufficient justification.

WHEELER— THE PARASITIC ACULEATA. 27

mecophiles, there remains the interesting problem as to the phylo-
genetic relations of the slave-makers, temporary and permanent
social parasites. Obviously the permanent parasites can be readily
conceived as developing either from temporary parasites or from
dulotic species. The fact that there are among the social bees and
wasps, as will be shown in the sequel, certain forms which agree in
all essential particulars with the permanent social parasites among
ants, although for obvious reasons they cannot have arisen from
dulotic forms, would seem to point to the origin of the permanent
from the temporary type of social parasitism. On the other hand,
Polycrgiis seems clearly to be in a stage transitional from dulotic
to permanent social parasitism, and a more advanced stage appears
to be represented by Strongylognathus testaceus which lives with
Tctramorium caspitum and produces workers which are few in
number and endowed with very feeble slave-making proclivities.

It is more difficult to determine the phylogenetic relations of the
dulotic to the temporary parasites. Wasmann (1905), Emery
(1909), Viehmeyer (1909, 191 1), Brun (191 2) and I have dis-
cussed this matter in several papers. Wasmann holds that tem-
porary social parasitism, which I first discovered in various North
American species of Formica, arose from the pleometrosis of such
forms as Formica ritfa. In a single colony of this and other acer-
vicolous species the females may be very numerous and new nests
may be formed by daughter queens departing from the maternal
nest with contingents of workers, or recently fecundated queens
may secure adoption in other nests of their own species. At first
I was inclined to derive both dulosis and temporary social para-
sitism from such conditions, but Wasmann insisted on deriving
dulosis from temporary social parasitism, a view which Emery,
Viehmeyer, Brun and I have rejected as unsound on the principle
that parasitism may readily arise from predatism, but that the re-
verse development is biologically highly improbable. I am now
inclined to agree with Emery that pleometrosis and the adoption
of queen ants by workers of their own species are probably phe-
nomena sui generis which did not lead to social parasitism, that we
must assume a predatory stage not unlike that of F. sanguinea as
the starting point for dulosis and that temporary social parasitism

28 WHEELER— THE PARASITIC ACULEATA.

was a subsequent development. Probably the predatory sanguinea
queen originally entered fusca nests for the purpose of devouring
the brood, but later came to care for the larvse and pupae till they
hatched. We may conceive that the number of appropriated fusca
young was more than was needed as food and that the queen ac-
quired a fostering relation towards the remainder by coming in
contact with the buccal secretions or fat-exudates of the larvae. In
other words, trophallactic relations were established between the
queen and the alien brood and led to a rearing of the latter (Wheeler,
1918). This might offer a simple explanation of dulosis, a phe-
nomenon which has always seemed unique and difficult of explana-
tion. At any rate it furnishes an hypothesis to be tested by a closer
study of the relation of F. sanguinea to the larval fusca.

Among the social bees there is only one parasitic genus, Psithy-
rus, to which I have repeatedly alluded. Kirby was the first to dis-
tinguish these bees from their hosts, the species of Bomhus, as long
ago as 1802, but a genus was first established for them by Lepeletier
in 1841. The habits of Psithyriis, as described by Hoffer (1881,
1888), Wagner (1907) and Sladen (1912), show that it is to be
regarded as a permanent social parasite. Like the ants of this
type, it lacks the worker caste. The female hibernates alone like
the queen Bomhus and enters and secures adoption in a young col-
ony of the latter, usually after the first batch of workers has
emerged. The Pslthyrus female has a tougher integument and a
stouter, more curved sting than Bomhus, and though she visits flow-
ers, she does not collect pollen or nectar. After entering the Bom-
hus nest Sladen says :

" Her first care is to ingratiate herself with the inhabitants, and in this
she succeeds so well that the workers soon cease to show any hostility to-
wards her. Even the queen grows accustomed to the presence of the stranger
and her alarm disappears, but it is succeeded by a kind of despondency. Her
interest and pleasure in her brood seem less, and so depressed is she that one
can fancy that she has a presentiment of the fate that awaits her. It is by
no means a cheerful family, and the gloom of impending disaster seems to
hang over it. But while the queen grows more dejected, the Psithynis grows
more lively, and takes an increasing interest in the comb, crawling about
over it with unwonted alacrity and examining it minutely."

The queen is eventually killed by the parasite, which then begins to

WHEELER— THE PARASITIC ACULEATA. 29

lay her eggs. She is at first very prolific, " but she ages and fails
more quickly than the Bomhus queen. . . . The Psithyrus kills the
Bombiis queen before she has laid the full number of worker eggs,'
consequently nests containing Psithyri are not very populous, the
number of workers seldom exceeding eighty." Neither queens nor
males of Bomhus are reared in such infested nests, but the workers
take to ovipositing. Their eggs would, of course, produce males,
but the Psithyrus devours them. She " pays close attention to her
new-laid eggs for several hours, giving the workers no chance to
molest them, but the workers soon get reconciled to them and
henceforth they feed and tend the Psithyrus brood with as much
devotion as if it were their own species; indeed, they seem some-
times to show a greater fondness for it." Sladen's concluding re-
marks are very interesting in connection with the case of Psammo-
chares rufipes and pectinipcs. He says:

" The origin of Psithyrus, more especially of its peculiar parasitical in-
stincts, is an interesting question. If a specimen of Psithyrus be compared
with a specimen of Bombus it is seen that the resemblance is not merely-
superficial but extends to nearly all the important details of structure, so
that it is impossible to avoid the conclusion that Psithyrus has sprung from
Bombus, and this at quite a recent period in the history of life. Moreover,
the Bombi — and this is particularly interesting — show parasitical tendencies
leading to the parasitism of Psithyrus. We have seen (pages 55-58) how
the Bombus queens may enter the nests of their own species and kill one
another, and how, in the case of the twin species, B. terrestris and lucorum,
tcrrestris has extended this habit so as to prey on lucorum, killing the
lucorum queen and getting the lucorum workers to rear her young in practi-
cally the same manner as the Psithyri prey on the Bombi. It is a remarkable
fact that the sting of the tcrrestris queen differs from that of the lucorum
queen and approaches that of Psithyrus in being somewhat stouter and more
curved, and having its thickened basal portion more parallel-sided when
viewed sideways than in lucorum. There is, however, no evidence to show
that any species of Psithyrus has sprung from the particular species of
Bombus on which it preys, such resemblances as it may show to it in coat-
colour, etc., being pretty clearly attributable to mimicry or exposure to the
same conditions of life, and not to ancestry."

Among the social wasps only two parasitic species are known,
Vespa arctica and V. austriaca. The former belongs to our Cana-
dian faunal zone and infests the nests of V. diahoUca, as Fletcher
(1908) has shown ; the latter has long been known in Europe where
it occurs in the nests of V. rufa. Recently Bequaert (1916) has

30 WHEELER— THE PARASITIC ACULEATA.

succeeded in finding austriaca in the United States and surmises
that it may here be a parasite of V. consohrlna, " which, although
very different in coloration, is very probably the American race or
subspecies of Vespa rufa L." A good account of what is known of
the habits and distribution of austriaca may be found in the papers
of Robson (1898), Carpenter and Pack-Beresford (1903) and Be-
quaert. This wasp is so closely related to V. rufa that Carpenter
and Pack-Beresford regard them as a single species, and the aus-
triaca queen as producing both rufa and austriaca offspring. Their
reasons for this assumption are, however, too weak to invalidate
the view of the great majority of authors who hold that austriaca
bears the same relation to rufa that Psithyrus does to Bomhus.
Both arctica and austriaca lack the worker caste and eliminate the
queens of the colonies which they enter. Males of the host species
sometimes develop in colonies infested by austriaca, so that, unlike
Psithyrus and the workerless ants, this parasite seems not to destroy
the eggs laid by the host workers.

In conclusion I would record a few reflections that have been
suggested by the foregoing survey of the various Aculeate parasites.
The tables show in a rather imposing manner that many of these
parasites have arisen from their respective host genera, but apart
from such forms as Psamuiocharcs pcctinipcs, Vcspa austriaca and
some of the parasitic bees like Percsia and Eucondylops there is
little evidence among existing parasites of a direct derivation from
their host species. This is what we might expect, for in the first
place the origin of most of the parasites is so remote that even if
they had remained permanently associated with the species from
which they arose, both host and parasite would by this time have
diverged in structure to such a degree that their genetic afilinities
would no longer be clearly discernible, and in the second place,
many parasites are probably no longer associated with their original
hosts, which have become nearly or quite extinct, so that their para-
sites have been compelled to adapt themselves to new hosts or cease
to exist. Under such circumstances a parasite would naturally
attach itself to a species more or less closely allied to its primitive
host.

That this has been the course of parasitic evolution seems to be

WHEELER— THE PARASITIC ACULEATA. 31

indicated by the fact, which has not, I believe, been emphasized by
other students of the subject, that most of the Aculeate hosts belong
to dominant genera. By dominant genera I mean those that are
represented by a considerable number of species, some of which are
very abundant in individuals and widely distributed as distinguished
from genera that are monotypic or represented by few species of
rare or sporadic occurrence. Such genera are Andrena, Halictus,
Anthophora, Megachilc, Osm'ia and Bomhus among the bees, Vcspa,
and PsainnwcJiarcs among the wasps, and among the ants Formica,
Lasius, Tapinoma, Pheidole, Crematogastcr, Aphccnogastcr, Tetra-
morhim, Monomorium and Leptothorax. It is, nevertheless, sur-
prising that no hosts of parasitic ants are known to occur in genera
like Camponotiis and Polyrhachis, which comprise hundreds of spe-
cies and are widely distributed, the former in all parts of the world,
the latter in the old world tropics. Probably the dominant genera,
owing to their abundance in individuals and the wide distribution
of their species, would act like great nets set to capture any para-
sites that have overstepped the bounds of good parasitic manners
by bringing their original host species to the verge of extinction.
This would account for the close generic affinities which we have
seen to be so evident between parasite and host, for a parasite that
had endangered or destroyed its original host species would itself
more readily escape extinction if the host were already a member
of a dominant genus containing many closely allied species, be-
cause this would permit a comparatively easy re-adaptation of the
parasite to a new host species. Although the parasites would prob-
ably differ in their powers of adaptation, the very similar habits of
species in the same genus, especially among the bees, would greatly
facilitate such a transfer of the parasitic relation.

Still even if we grant that the Aculeate parasite has arisen from
its original host species, we are confronted with two troublesome
questions, for it will be asked : What induced certain individuals to
become parasites on other individuals of their own species ? And
if a parasite originated in this way, what is to prevent it inbreeding
with its host and thus losing its peculiar tendencies by swamping,
or " panmixia " ? One of the difficulties involved in the first ques-
tion lies in deciding on the stimuli that would so afifect some of the

32 WHEELER— THE PARASITIC ACULEATA.

individuals of a species as to compel them to give up the industrious
and nonparasitic habits that have become elaborated and fixed as
an integral part of their genetic constitution. I believe, however,
that such stimuli exist and that they are frequent and compara-
tively simple. If we take such a constantly recurring external
stimulus as temporary scarcity of prey or food, we can understand
how some individuals of a common species that has outrun its food
supply or has emerged in seasons or places of scarcity, might find
it as easy as advantageous to steal the provisions of other indi-
viduals. This is, in fact, a common practice among normally non-
parasitic Aculeata, e. g., in Bcmbix and Psammo chares, in bumble-
bees, and honey-bees. And if this external is reinforced by an
internal stimulus, such as the urgent need for oviposition, we can
see how a parasitic group of individuals might readily arise within
the confines of a single species. This urgency of oviposition is
very apparent in many parasitic Aculeata, especially among the
parasitic bees, which often lay several eggs in a single cell of the
host, though only one larva is able to develop. PsitJiyrus and some
of the parasitic ants seem to reveal the same urgency. It is even
probable that this internal stimulus is more fundamental than the
external stimulus above mentioned and that it may incite the insect
directly to appropriate the provisions collected by other individuals
whose ovarian eggs mature more slowly or in smaller numbers.
When the parasitic habit is once started it tends necessarily, owing
to the saving of energy which would otherwise be expended in
work, to accelerate the maturation of the ova and thus to become
more and more confirmed by one of the circular processes so fa-
miliar to the physiologist.

Urgency of oviposition will, I believe, account also for the many
cases in which Aculeates have been observed in the act of appro-
priating valuable nesting materials or partly constructed nests of
other individuals. Fabre (1890) saw wall-bees (Chalicodoma inu-
raria) take possession of the partly constructed masonry nests be-
longing to other females of the colony and destroy their eggs, and
Adlerz (1904) observed other bees (Trachusa scrratnlcc) enter
each others' burrows and steal the pine-pitch with which they glue
together the pieces of leaves for their nests. Drory (1872) saw

WHEELER— THE PARASITIC ACULEATA. 33

South American stingless bees (McUpona) overpower other indi-
viduals of their species and bite away the propoHs with which their
hind legs were charged.

The difference of sexual maturity between parasite and host
suggests an answer to our second question, for the time of mating
would of course depend on the time of sexual maturity and one
group of individuals may be effectually isolated so far as its further
phylogenetic development is concerned from another group of the
same species, if the mating periods fail to coincide in the two
groups. Thus interbreeding of the parasite with the host might be
avoided in a very simple manner, and parasite and host, though
reared on the same food and in the same environment would never-
theless tend to pursue divergent paths in their subsequent history.
It would be interesting, therefore, to collect accurate data on the
rate of larval development, and the time of emergence and mating
of parasitic Aculeates and their hosts with a view to testing the
strength of the hypothesis here suggested. •

That certain Aculeates respond so readily and in such a uni-
form manner to simple stimuli like the urgency of oviposition and
dearth of food by becoming parasitic on other Aculeates may be
attributed to a peculiar modification of their constitution during
their long phylogenetic history, some of the main outlines of which
have been clearly revealed by morphological studies. Hymenop-
terists agree that the higher Aculeates are descended from primitive
wasps whose modern representatives constitute the families Scoliidse,
Thynnidse and Mutillidse and that their ancestors in turn are to be
sought among groups like the Ichneumonidge. The latter have been
called parasitic, but it is clear that their larva?, which feed on the
tissues of other insects and eventually kill their hosts are really
practicing a refined, protracted and very economical predatism. They
may be more properly designated as parasitoids, as Renter (1913)
has suggested. The Scoliidse, Thynnidse and Mutillidse, which
seek out the larvse of other insects in concealed places, i. c, in the
soil or in nest-cavities, immobilize or kill them by stinging and then
deposit their eggs on them, therefore occupy a position, ethologically
as well as structurally, midway between the higher solitary wasps
and the Ichneumonidae. The higher wasps in constructing nests

PROC. AMER. PHIL. SOC, VOL. LVIII, C. JUNE 25, IQIQ.

34 WHEELER— THE PARASITIC ACULEATA.

and provisioning them with paralyzed insects merely elaborate the
same fundamental behavioristic theme or pattern, the main features
of which were also retained by the solitary bees even after they had
ceased to capture insect prey and had become pollenivorous and
nectarivorous. The social wasps and bees have merely modified
certain details in the behavior of the solitary species. Among the
ants the modifications are more profound, but the most primitive
subfamily, the Ponerinse, still exhibits many Sphecid traits. We
may assume, therefore, that the ancient parasitoid habits of the Ich-
neumonid ancestry still abides as a latent, phylogenetic memory, or
mneme, in the constitution of the whole Aculeate group. Hence it
is not surprising that this mneme can be revived in response to such
recurrent external and internal stimuli as- dearth of food and ur-
gency of oviposition. Under these conditions the solitary Aculeate
readily becomes parasitic and reverts to a type of behavior essen-
tially like that of the Mutillidae, Thynnidse and Chrysididse. In the
parasitic social Aculeates new behavioristic modifications have de-
veloped as the result of the complex and peculiar living environ-
ment presented by the social habit of the host, the trophallactic rela-
tions of the mother insect and her ofl^spring and the existence of a
worker caste.

The general conclusions that may be drawn from the foregoing
survey of the parasitic Aculeata in particular and of insect parasites
in general may be stated as follows :

1. We may distinguish two intergrading types of parasitism
among insects. One of these is true parasitism and is represented
by the lice, fleas, Mallophaga. many Diptera (CEstridse, Pupipara)
and some Hemiptera, which live on mammals and birds and do not
destroy their hosts. The other is parasitoidism, which is really a
refinement of predatism and is eminently characteristic of large
sections of the Hymenoptera and Diptera (Tachinidas). It leads
sooner or later to the death of the host. The difiterence between
the two types is largely due to dififerences in the size and vigor of
the hosts.

2. Parasitoids are of two classes, one of which is best repre-
sented by the so-called Parasitica among the Hymenoptera and the
Tachinidje among the Diptera, which have no genetic relationship

WHEELER— THE PARASITIC ACULEATA. 35

with their hosts. The other class of parasitoids is represented by
the Aculeates which have sprung directly from their host species
(intraspecific parasitoids), though they may subsequently acquire
hosts among other species of the same genus or of other genera and
may in turn be the ancestors of parasitic species.

3. The derivation of all the existing Aculeata from primitive
insectivorous wasp-like ancestors may account for the retention of
a rather uniform pattern of behavior among the parasitic species.
The parasites, both among the solitary wasps and the solitary bees,
behave in a very similar manner, though the former are reared on
insect prey, the latter on pollen and honey. In both groups the
object of the parasite is to secure the provisions accumulated by the
host for its own progeny. This involves a destruction of the egg
or young larva of the host. The social parasites, however, have
passed beyond this destruction of the host brood to a stage involving
the fostering of the host brood as a means of insuring the rearing
and alimentation of their own young. This change may have been
due in the first instance to the formation of trophallactic relations
between the parasite and the host brood.

4. The origin of parasitism among the Aculeata may be attribu-
ted to urgency of oviposition and temporary or local dearth of the
supply of provisions for the offspring.

5. In all the different forms of parasitism among the Aculeata,
there are traces of the primitive predatism or parasitoidism from
which it arose, although in some of the social parasites this is repre-
sented only by the aggressive or conciliatory intrusion of the re-
cently fecundated female into the host colony. Even the more ex-
treme forms of behavior, such as those of the temporary and per-
manent social parasites, were derived from predatory behavior like
that manifested by Formica sanguinca and its various subspecies
and varieties.

6. Although many cases of parasitism are known to occur among
the Aculeata, and although many others will doubtless be discovered
in the future, nevertheless the total number must be small in com-
parison with the thousands of nonparasitic species. Contemplation
of such a series as we find among the ants, beginning with Formica
sanguinca, which is an abundant, vigorous and aggressive species

36 WHEELER— THE PARASITIC ACULEATA.

and ending with Ancrgatcs atratiiliis, a small, sporadic, and appar-
ently evanescent species, without workers and with wingless,
nymphoid males, suggests that parasitism among the Aculeates
tends to such extreme specialization ("degeneration") as to lead
to extinction. If we possessed a knowledge of the whole evolution
of the Aculeate group, we should probably find that the total num-
ber of parasitic species which it produced during the ages was very
great, but that the vast majority of them, after reaching the Ancr-
gates or a similarly specialized, or degenerate stage, lingered on
precariously for a time and then disappeared.

LITERATURE.

1902. Adlerz, G. Ceropales maculata Fab. en Parasitisk Pompilid. Bih. K.

Svcnsk. Vet. Akad. Hand!., 28, 1902, 20 pp.
1904. Adlerz, G. Om cellbyggnad och tjuvl)in hos Trachusa serratuls Panz.

E>it. Tidskr., 25, 1904, pp. 121-129.
1910-12. Adlerz, G. Lefnadsforhallanden och Instinkter inom Familjerna

Pompilidse och Sphegidae III. and IV. K. Svcnska Vctcnsk. Handl.,

45, No. 12, 1910, pp. 5-75; 47, 1912, No. ID, pp. 5-61.
1910. von Alten, H. Zur Phylogenie des Hymenopterengehirns. Jen. Zcit-

sclir. Naturw., 46, 1910, pp. 511-590, 4 pis., 28 text-figs.
1899. Ashmead, W. H. Classification of the Bees, or the Superfamily

Apoidea. Trans. Amcr. Ent. Soc. 26. 1899, pp. 49-100.
1916. Bequaert, J. On the Occurrence of Vcspa austriaca Panzer in the

Northeastern United States. Bull. Brooklyn Ent. Soc, 11, 1916, pp.

101-107.

1902. Brauns, H. Eucondylops n. g. Apidarum. Zcitschr. Hym. Dipt., 6.

1902, pp. 377-380.
1878. Breitenbach, W. Ueber Halictus 4-cinctiis F. und Sphccodcs gibbiis L.

Stcft. Ent. Zeitg., 1878, pp. 241-243.
1912. Brun, R. Weitere Beitrage zur Frage der Koloniegriindung bei den

Ameisen, mit besonderer Beriicksichtigung der Phylogenese des

sozialen Parasitismus und der Dulosis bei Formica. Biol. Ccntralbh,

32, 1912, pp. 154-180, 216-226.

1903. von Buttel-Reepen, H. Die Stammesgeschichtliche Entstehung des

Bienenstaates. Leipzig, 1903.
1903. Carpenter, G. H., and D. R. Pack-Beresford. The Relationship of Vcspa

austriaca and Vespa rufa. Ent. Month. Mag.. (2), 14, 1903, pp. 230-

242, I pi.
1902. Cockerel!, T. D. A. Coiitril)utions from the New Mexico Biological

Station. — XII. On some Genera of Bees. .];;;;. Mag. Xat. Hist.,

(7), 9, 1902, pp. 230-234.
1906. Cockerell, T. D. A. The Bees of Florissant. Bull. Amcr. Mus. Nat.

Hist., 22, 1906, pp. 419-455, 2 figs.

WHEELER— THE PARASITIC ACULEATA. 37

1908. Cockerell, T. D. A. Notes on the Bee-Genus Exaerete. Psyche, Apr.,

1908, pp. 41-42.
1910. Cockerell, T. D. A. An Introduction to the Study of Rocky Mountain

Bees. Univ. Colo. Studies, 7, 1910, pp. 179-195, 50 figs.
1916. Cockerell, T. D. A. Some Neotropical Parasitic Bees. Eiit. Nczvs, 27,

1916, pp. 208-210.

1918. Cockerell, T. D. A. The MegachiHd Bees of the Philippine Islands.
Philip. Jouru. Sci.. 13, 1918, pp. 127-144.

1919. Cockerell, T. D. A. A New Genus of Bees from Peru. Eiit. Nczvs, 30,

1919, pp. 17-18.

191 5. Crawford, J. C. The Bee Genus Holcopasitcs Ashmead. his. Ins.

Mens., 3, 1915, pp. 123-126.

1916. Crawford, J. C. Some New American Hymenoptera. Ins. Ins. Mens.,

4. 1916, pp. I3.S-I44-

1896. de Dalla Torre, C. G. Catalogus Hymenopterorum 10, Apidse. Leip-

zig, Engelmann, 1896.
1872. Drory, E. Quelques Observations sur la Melipone {Mclipona scutcl-

laris). Bordeaux, 1872.
1906. Ducke, A. Neue Beobachtungen fiber die Bienen der Amazonaslander.

Zcitschr. f. wiss. Insekth., 2, 1906, pp. 51-60.
1912. Ducke, A. Die natiirlichen Bienengenera Siidamerikas. Zool. Jahrb.

Abth. Syst., 34, 1912, pp. 51-116.

1909. Emery, C. Intorno all'origine delle Formiche dulotiche, parassitiche

e mirmecofile. Rend. R. Accad. Sci. Bologna, 1909, pp. 36-51, Germ.
Trans, in Biol. Centralbi, 29, 1909, pp. 352-362.
1890. Fabre, J. H. Souvenirs Entomologiques 3, 2d edition. Delagrave,
Paris, 1890.

1890. Ferton, Ch. L'evolution de I'instinct chez les Hymenopteres. Rev.

Scicnf., 1890, pp. 45, 496-498.

1891. Ferton, Ch. Notes pour servir a I'histoire de I'instinct de Pompilides.

Act. Soc. Linn. Bordcau.v, 44, 1891, pp. 281-294.

1897. Ferton, Ch. Nouvelles Observations sur I'instinct des Pompihdes.

Act. Soc. Linn. Bordcau.v, 52, 1897, pp. 101-132.
1901-1911. Ferton, Ch. Notes detachees sur I'instinct des H3mienopteres

Melliferes et Ravisseurs. Ann. Soc. Ent. France, ist series, 70,

1901, pp. 83-148, 3 pis. ; 2d series, ibid., yi, 1902, pp. 499-529, i pi. ;

3d series, ibid., 74, 1905, pp. 56-104, 2 pis. ; 4th series, ibid., 77, 1909,

PP- 535-586, I fig. ; 5th series, ibid., 78, 1909, pp. 401-422 ; 6th series,

ibid., 79, 1910, pp. 145-178; 7th series, ibid., 80, 1911, pp. 351-412,

7 figs.
1914. Ferton, Ch. Perezia maura Genre d'Apiaires Parasites d'Algerie et

Observations sur ce Genre. Ann. Soc. Ent. Soc. France, 83, 1914,

pp. 22,2,-22,7, I fig.
1908. Fletcher, J. Notice of a paper entitled " Is Vcspa borcalis an Inqui-

line?" Ann. Ent. Soc. America, i, 1908, p. 30.
1899. Friese, H, Die Schmarotzerbienen und ihre Wirte. Zool. Jahrb. Abt.

Syst., 3, 1889, pp. 847-870.

38 WHEELER— THE PARASITIC ACULEATA.

1891. Friese, H. Beitrage zur Biologie der soHtaren Blumenwespen (Apidae).
Zool. Jahrb. Abth. Syst., 5, 1891, pp. 751-860, i pi.

1895. Friese, H. Die Bienen Europas. I. Schmarotzerbienen. Berlin, Fried-
lander, 1895.

1909. Friese, H. Die Bienen Afrikas nach dem Stande unserer heutigen
Kenntnisse. Zool. Anthr. Ergeb. cincr Forschtmgsrcisc tin ijccstl. u.
zcntr. Sildafrika, 2, 1909, pp. 85-475, 2 pis.

1914. Friese, H. Neue Bienenarten der orientalischen Region. Deutsch.
Ent. Zeitschr., 1914, pp. 320-324.

1906. Forel, A. Moeurs des Fourmis Parasites des Genres Whcclcria et
Bothriomyrmex. Rev. Suisse Zool., 14, 1906, pp. 51^69.

1905. Graenicher, S. Some Observations on the Life-History and Habits of

Parasitic Bees. Bull. Wis. Nat. Hist. Sac, 3, 1905, pp. 153-167,
I pi.

1906. Graenicher, S. A Contribution to our Knowledge of the Visual

Memory of Bees. Bull. Wis. Nat. Hist. Soc, 4, 1906, pp. 135-142.

1887. Handlirsch, A. Monographic der mit Nysson und Bembex verwandten

Grabwespen. — I. Sit::b. Akad. Wiss. Wien., 95, 1887, pp. 246-421,

5 pis.
1904. Hoppner, H. Zur Biologie der Rubus-Bewohner II. Osmia parvula

Duf. et Perr., Osmia leucomelcena K. und ihr Schmarotzer Stelis

ornatula Nyl. Allgem. Zeitschr. Ent., 9, 1904, pp. 129-134, 2 figs.
1881. Hoffer, E. Ueber die Lebensweise des Apathus (Psithyrus) campes-

tris. Mitth. naturzv. Vcr. Steicrmark., 1881, pp. 87-92.

1888. Hoffer, E. Die Schmarotzerhummeln Steiermarks. Graz, 1889. Also

in Mitth. naturw. Ver. Steicrmark, Jahrg., 1888, pp. 84-88.
1802. Kirby, W. Monographia Apium Angliae. 2 vols. Ipswich, 1802.
1802. Latreille, P. A. Ordre Naturel des Insectes design-es generalement

sous le nom d'Abeille, Apis Lin. Geoff. In Histoire Naturelle des

Fourmis. Crapelet, Paris, 1802, pp. 401-438.
1810. Latreille, P. A. Considerations Generales sur I'ordre des animaux

composant les classes des Crustaces, des Arachnides et des Insectes,

1810.
1827. Lepeletier de St. Fargeau, A. L. M. Encyclopedie Methodique, 10,

Paris, 1827, p. 183.
1841. Lepeletier de St. Fargeau, A. L. M. Histoire Naturelle des Insectes.

Suites a Buff on. Hjnnenopteres 2, 1841.
1890. Marchal, P. Formation d'une Espece par le parasitisme. fitude sur

le Sphccodes gibbus. Rev. Scieut., 45, 1890, pp. 199-204.
1894. Marchal, P. Le Parasitisme des Sphccodes. Bull. Soc. Ent. France,

63, 1894, p. cxv.
1864. Morawitz, F. Ueber Vespa aiistriaca Panz. und drei neue Bienen.

Bull. Soc. Imper. Nat. Moscou, 37, 1864, p. 439-449.
1901. Morice, R. F. D. Observations on Sphccodes. Ent. Month. Mag., 12,

1 901, pp. 53-58.
1871. Miiller, H. Anwendung der Darwinischen Lehrc auf Bienen. I'crh.

Naturh. Ver. Prcuss. Rhcinl. u. Westphal, 1871. pp. 1-96, 2 pis.

WHEELER— THE PARASITIC ACULEATA. 39

1903. Nielsen, J. C. Om Bislaegten Sphccodcs Latr. Ent. Meddel, (2), 2,

1903, pp. 21-30.
1883. Perez, J. Contribution a la Faune des Apiaires de France, 2^ partie.

Act. Soc. Linn. Bordeaux, 27, 1883, pp. 205-378, 2 pis.
-' 1884. Perez, J. Les Apiaires Parasites au point de Vue de la Theorie de

I'Evolution. Bordeaux, 1884.

1894. Perez, J. Notes Zoologiques. Act. Soc. Linn. Bordcau.v, 47, 1894, pp.

231-331-
1910. Perez, J. Notes sur les Vespides. Act. Soc. Linn. Bordcau.v, 64, 1910,

pp. 1-20.
1887. Perkins, R. C. L. Notes on Some Habits of Sphecodes Latr. and
Noniada Fabr. Ent. Month. Mag., 23, 1887, pp. 271-274.

1889. Perkins, R. C. L. Is Sphecodes parasitic? Ent. Month. Mag., 25, 1889,

pp. 206-208.

1912-14. Rabaud, E. fithologie et Comportement de Diverses Larves Endo-
phytes. I., Olcthreutcs ohlongana. Bull. Scient. France Bclg., (7),
46, 1912, pp. 1-28, I fig.; II., Myclois cribclla, ibid., (7), 48, 1914, pp.
28-106, I fig.

1913. Renter, 0. M. Lebensgewohnheiten und Instinkte der Insekten. Ber-
lin, Friedlander, 1913.

1869. Riley, C. V., and B. J. Walsh. Wasps and their Habits. Amcr. Ent.,
I, 1869, pp. 122-142, 17 figs.

1899. Robertson, Chas. On the Classification of Bees. Canad. Ent., 31, 1899,

pp. 338-343-
1898. Robson, J. E. Vespa austriaca, a Cuckoo Wasp. Science Gossip, (n.
s.), 5, 1898, pp. 69-73.

1906. Santschi, F. A propos des Moeurs parasitiques Temporaires des

Fourmis du Genre Bothrioinynnex. Ann. Soc. Ent. France, 1906,

PP- 363-392.
1913. Santschi, F. Une nouvelle fourmi parasite. Bull. Soc. d'Hist. Nat.
d'Afr. Nord, 5, 1913, 2 pp.

1890. Saunders, E. On the Tongues of the British Anthophila. Joiirn. Linn.

Soc. Lond. Zool., 23, 1890, pp. 410-432, 8 pis.
1903. Saunders, E. On the Relationship of Aculeate Inquilines and their

Hosts. Ent. Month. Mag., (2), 14, 1903, pp. 272-274.
1881. Schmiedeknecht, 0. Ueber einige deutsche Vespa-Arten. Ent. Nach-

richt., 7, 1881, pp. 313-318.
1882-1886. Schmiedeknecht, 0. Apidje Europese (Die Bienen Europas) Bd.

I, u. 2. Berlin, Gumperda, 1882-1886.

1907. Schmiedknecht, 0. Die Hymenopteren Mitteleuropas. Jena, Gustav

Fischer, 1907.
1903. Sharp, D. Vespa rufay, austriaca. Ent. Month. Mag., (2), i, 1903,

pp. 297-299.
1865. Sichel, J. Revision Monographique, Critique et Synonymique du Genre

Mellifere Sphecodes Latr. Ann. Soc. Ent. France, (4), 5, 1865, pp.

397-466.

1895. Sladen, F. W. Observations on Sphecodes rubicundus v. Hag. Ent.

Month. Mag., 31, 1895, p. 256.

40 WHEELER— THE PARASITIC ACULEATA.

1912. Sladen, F. W. L. The Humble-Bee, Its Life-History and how to Do-

mesticate it. Macmillan & Co., London, 1912.
1918. Sladen, F. W. L. The Genus Vcspo in Canada. Ottazva Naturalist,

32, 1918, pp. 71-72.
1892. Verhoeff, C. Zur Kenntnis des biologischen Verhaltnisses zwischen

Wirt- und Parasiten-Bienenlarven. Zoo/. ^4nrc/(7., 15, 1892, pp. 41-43.

1908. Viehmeyer, H. Zur Koloniegriindiing der parasitischen Ameisen.

Biol. Ccntralbl., 28, 1908, pp. 18-32.

1909. Viehmeyer, H. Beobachtungen und Experimente zur Koloniegriindung

von Formica sangiiinca Latr. ZcitscJir. zciss. hisektbiol., 5, 1909, pp.

353-394-

1910. Viehmeyer, H. Ontogenetische u. phylogenetische Beobachtungen iiber

die parasitische Koloniegriindung von Formica sanguinca. Biol.

Ccntralbl., 30, 1910, pp. 569-580.
1907. Wagner, M. Psychobiologische Uhtersuchungen an Hummehi.

Zoologica, 19, 1907, 239 pp., i pi.
1817. de Walckanaer, C. A. Memoires pour servir a I'Histoire naturelle des

Abeilles Solitaires, qui composent le genre Halictc. Paris, Didot,

1817, 90 pp., I pi.

1905. Wasmann, E. Ursprung u. Entwickelung der Sklavrei bei den

Ameisen. Biol. Ccntralbl., 25, 1905, pp. 1 17-127, 129-144. 161-169,

193-216, 256-270, 273-292.
1915. Wasmann, E. Das Gesellschaftsleben der Ameisen. i Bd., ^Miinster i.

W., Aschendorffsche Verlagsbuchh., 191 5.
1835. Wesmael, C. Observations sur les Especes du Genre Sphccodc. Bull.

Acad. Roy. Sc. BruxcUcs, 2, 1835, pp. 279-287.
1840. Westwood, J. Introduction to the Modern Classification of Insects.

2d vol., London, Longmans, etc., 1840.
1904. Wheeler, W. M. A New Type of Social Parasitism among Ants.

Bull. Amcr. Mus. Nat. Hist., 20, 1904, pp. 347-375.

1906. Wheeler, W. M. On the- Founding of Colonies by Queen Ants, with

Special Reference to the Parasitic and Slave-making Species. Bull.
Amcr. Mus. Nat. Hist., 22, 1906, pp. 33-105, 7 pis.

1907. Wheeler, W. M. The Origin of Slavery Among Ants. Pop. Sci.

Month., 71, 1907, pp. 550-559-

1908. Wheeler, W. M. The Ants of Casco Bay, Maine, with Observations on

two Races of Formica sanguinca Latreille. Bull. Amcr. Mus. Nat.
Hist.. 24, 1908, pp. 619-645.

1909. Wheeler, W. M. Observations on Some European Ants. Jouni. A''. Y.

Ent. Soc, 17, 1909, pp. 172-184, 2 figs.

1910. Wheeler, W. M. Ants, their Structure, Development and Behavior.

Columbia Univ. Press, 1910.
1918. Wheeler, W. M. A Study of Some Ant Larvae, with a Consideration of
the Origin and Meaning of the Social Habit Among Insects. Proc.
Amcr. Phil Soc, 57, 1918, pp. 293-343, 12 figs.

1913. Williams, F. X. Notes on the Habits of Some Wasps that Occur in

Kansas, with Description of New Species. Kans. Univ. Sc. Bull., 8,
191 3, pp. 223-230, I pi.

THE RELATION OF THE DIET TO PELLAGRA.

By E. V. McCOLLUM, Ph.D.

(Read April 25, igiQ-)

Pellagra has long been suspected of being caused by faulty
diet, and the eating of maize, particularly moldy maize, has been
considered by some students of the disease to be the specific cause.
The studies of Dr. Goldberger of the Public Health Service have
eliminated corn as a causative agent in the etiology of this disease.
Funk in his enthusiasm over the " vitamine " hypothesis adopted the
view that not only beri-beri but scurvy, rickets, and pellagra were
each due to the lack of a specific "vitamine'' in the diet. He
further assumed, in order to explain the conflicting results in some
of his experimental work, that other " vitamines " necessary for
maintenance and for growth respectively are necessary in the food
supply. We have attempted during the last two years to discover
the exact nature of the deficiencies of such diets as are in common
use among the people of the cotton mill villages in the South where
pellagra is very common. We have employed what may properly
be described as a biological method for the analysis of a food stuii'
or of mixtures of foods. This consists in feeding any foodstuff
which is faulty in one or more respects to a group of animals, and
in other experimental groups the same food supplemented with
single or multiple food additions, such as pure protein, one or more
pure mineral salts, one or more of the still unidentified dietary
factors in the form of suitably prepared preparations. We have
throughout these studies employed as a working hypothesis the as-
sumption that the essential constituents of an adecpate diet are
protein of suitable quality and quantity, an adequate supply of the
necessary inorganic elements in suitable combinations, an adequate
energy supply in the form of protein, carbohydrate, and fat, and
two as yet chemically unidentified dietary essentials which we have

41

42 McCOLLUM— RELATION OF DIET TO PELLAGRA.

designated " fat-soluble A " and " water-soluble B." The lack of
the former leads to the development of a specific eye trouble which
seems to be accurately described as a type of xerophthalmia. The
factor water-soluble B is we believe identical wtih the substance
which prevents or cures the disease beri-beri characterized by gen-
eral paralysis which is common in the Orient.

Our experimental studies have now progressed so far as to
enable us to assert with confidence that a satisfactory diet cannot be
secured from mixtures containing any number of seeds or products
derived from the milling of seeds together with tubers, edible roots,
and meats. The vegetable foods which may be classed as seeds,
tubers, and roots are all functionally storage organs, and their
content of active protoplasm is relatively small in comparison with
their bulk because of the large amount of reserve food material
laid down in them. They may be sharply contrasted with the leaf
of the plant, which except in special cases is not a repository for
reserve proteins, carbohydrates, and fats, but represents, aside from
its skeletal tissues, functionally active protoplasm. The leaf has
very different dietary properties from those possessed by the tissues
which are modified as storage organs, and in many instances at
least represents complete foods for those types of animals whose
digestive tracts are so capacious as to permit them to eat a sufficient
amount of bulky material. We have been able to prepare fairly
satisfactory diets for an omnivorous animal, the rat, from these
two types of vegetable foods together, /. c, leaves and seeds, but
never from the group of vegetable foods which are functionally
storage organs.

From this experience we have been led to differentiate sharply
between two classes of foods which are usually collectively desig-
nated as vegetables. Leaves are constituted so as to correct the
dietary deficiencies of the storage tissues, whereas the seeds, tubers,
and roots fail to supplement mutually each other's deficiencies with
respect to either the inorganic moiety or the fat-soluble A. They
do in some degree mutually enhance the quality of each other's pro-
teins, but to a lesser degree than we had supposed before the com-
pletion of a large amount of experimental work directed toward the

McCOLLUM— RELATION OF DIET TO PELLAGRA. 43

quantitative comparison of the protein mixtures derived from pairs
of seeds in considerable number.

Mixtures of seeds, or of seeds, tubers, and roots, will in all cases
require supplementing with respect to calcium, sodium, and chlorine
among the inorganic elements, and fat-soluble A. In most such
mixtures the quality of the proteins will likewise be sufficiently poor
to require improvement before the optimum well-being can be
secured.

We are now in possession of a considerable amount of knowl-
edge concerning the distribution of the dietary factor, fat-soluble A,
in animal tissues. The body fats of the ruminants will probably
always be found to be richer in this substance than the body fats of
the omnivora because of the greater intake of it in the food. Mus-
cle tissue has been found to be very poor in fat-soluble A, but the
fats from the glandular organs, i. e., intracellular fats, are a good
source of it. It follows, therefore, that muscle tissue such as round
steak should not supplement mixtures of vegetable foods which be-
long to the storage organ group with respect to fat-soluble A, and
in our experience this proves to be the case. The inorganic content
of muscle tissue resembles in a general way that of the storage or-
gans of plants except in its very low content of magnesium. It is
too poor in calcium, and to a lesser degree in sodium and chlorine,
to support the optimum well-being in an animal. Muscle tissue
fails to supplement the seeds, tubers, and roots on the inorganic side.
The protein content of muscle tissue is high and the proteins are
probably of high biological value, and, except as respects palata-
bility, it is only in improving the quality of the protein content of
the ration that the addition of meats of this class enhances the value
of a mixture of products derived from the storage organ group of
plant products.

These considerations indicate the basis for our distinction be-
tween two groups of foodstuffs. One of these, which includes
milk, eggs, and the leafy vegetables, we have designated as "pro-
tective foods," in order to call attention to their special importance
in the diet. They are protective in that they are so constituted with
respect to their inorganic content, content of fat-soluble A, and the
quality of their proteins that they correct in great measure when

44 McCOLLUM— RELATION OF DIET TO PELLAGRA.

used in sufficient amounts the faults of the remainder of the food
mixture irrespective of the extent to which it is derived from either
seed, tuber, or root products. We have been able to plan satisfac-
tory diets of naturally occurring foods only by the inclusion of one
or more of these protective foods. The other group of natural
foodstuffs includes all seeds and seed products, such as the cereal
grains and their milling products (wheat flour, corn-meal, polished
rice, etc.), the legume seeds, tubers, edible roots, nuts, fruits, and
such cuts of meats as come from muscle tissue.

In all cases where we have attempted to correct the dietary defi-
ciencies of a seed mixture by the addition of leaf only we have not
secured results so good as with milk, especially with such amounts
of leaf as would be acceptable in the human diet. The leafy foods
are eaten by Europeans and Americans only in a very water rich
condition, and it is difficult to secure the consumption of enough to
correct the deficiencies in the remainder of the diet. With animals,
when we have fed dry powdered mixtures containing as much as
25 to 40 per cent, of the diet derived from leaf and the remainder
from plant products of the storage organ class, the nutrition has
been very good in some instances, but not all combinations will be
equally valuable. Eggs are decidedly poorer in calcium than are
the leaves or milk, when only the part exclusive of the shell is con-
sidered. The shell serves as a source of lime to the developing
chick. Eggs do not, therefore, supplement food mixtures derived
from storage tissues with respect to calcium to the degree that milk
and leafy vegetables do.

Even in such types of diet as contain one or more of the pro-
tective foods in fairly liberal amounts, it is certain that for such
rapidly growing species of animals as the hog and rat the inorganic
content is not entirely satisfactory, although it may be good enough
to enable the animal to perform all the functions of growth and
reproduction in a way which, in the absence of definite knowledge
of what the species is capable of, we should regard as normal. \\'e
have been accustomed to regard as normal an achievement in vigor
and well being both in man and animals which falls far short of that
seen in exceptional cases. The most important inorganic deficiency
in seed, tuber, and meat mixtures is calcium, and this is so pro-

McCOLLUM— RELATION OF DIET TO PELLAGRA. 45

nounced that we are of the opinion that even in those human die-
taries in which such calcium-rich food as milk is used in fair liber-
ality, the intake of calcium may be still below the optimum, and that
a direct addition of this element in the form of the carbonate or
lactate might be of distinct benefit in human nutrition except per-
haps in those regions where the w-ater is unusually rich in calcium
salts. Since civilized man usually adds sodium chloride to his foods
to suit the taste, the shortage of sodium and chlorine in the diet of
man presents no problem. An addition of calcium could be most
conveniently made to our foods through the use of a mixture of
equal parts of common salt and of calcium carbonate in the kitchen
and on the table.

A question \\hicli has never been answered to the satisfaction of
physiologists is : How nuich protein should the diet contain in order
to maintain physiological well being? At about the time when the
question was being most discussed, the chemistry of the proteins
was developed to a point which made it clear that there were great
differences in the biological values of the proteins from dift'erent
sources, depending on their yields of certain amino-acids. This
makes futile any attempt to establish a particular intake of protein
which may represent the minimum, optimum, or maximum amount
consistent with maintenance of " normal " vitality and health. The
quality of the proteins must be known before anything can be said
about the amount of protein necessary. From biological tests we
now know that the proteins of the pea or navy bean are worth only
about half as much for growth in the rat as are equal amounts of
proteins from one of the cereal grains, and that the latter have about
half the value for the conversion into body proteins which can be
shown for the proteins of milk. The relative values of the proteins
from different sources, as well as the absolute values of certain of
them, are just now becoming appreciated.

There are two opposing views regarding the amount of protein
which will produce the best results. Those who advocate the low
protein diet point to the "specific dynamic action" of protein,
through which it stimulates metabolism. They believe that a high
consumption of protein furnishes pabulum for the development of
an excessive growth of putrefactive bacteria, with the result that

46 M.cCOLLUM— RELATION OF DIET TO PELLAGRA.

toxic or irritating products of the degradation of certain amino-
acids are absorbed in amounts sufficient to cause damage to the tis-
sues. It has been recommended that man should, in adult life, take
only such an amount of protein as will cover the endogenous loss
due to tissue metabolism, together with a not well-defined " margin
of safety." The opponents of this view regard a liberal protein
allowance as essential to vigor and aggressiveness, and point to the
use of liberal amounts of meat by the peoples who have been char-
acterized by greatest achievement. Among all the progressive peo-
ples of the world the food supply is derived to a greater or less
extent from daily products, and this portion rather than the meat
eaten we have come to regard as of peculiar importance in improv-
ing the quality of the diet. In order to test this question we con-
ducted a series of experiments, employing rats which were about
nine months old. or about one fourth through the normal span of
life for this species, and were in excellent nutritive condition. They
were fed diets which were fairly satisfactory in all respects except
that the protein content was not far from the actual amount re-
quired for the maintenance of body weight for a few" weeks. We
observed unmistakable signs that the vitality of the animals was rap-
idly lowered on such a dietary regime. This was shown especially
by the rapid aging and short span of life. Even though the initial
body weight was approximately maintained for a period of three
months or more, distinct signs of aging were always apparent within
five to ten months. Three months in the life of a rat correspond
to about 8.4 per cent, of the average span of life. It can be readily
appreciated that if harmful effects in corresponding degree follow
the adherence by man to such low protein diets they would not be-
come apparent within the time covered by any experiment yet con-
ducted upon a diet squad, few of which have been restricted to any
experimental diet beyond six months. A reputed satisfactory out-
come of such experiments cannot be accepted as evidence that men
on diets which furnish but a small margin of protein over the actual
maintenance requirements are so nourished as best to promote
health. Aging at two to four times the rate observed in the most
satisfactorily nourished would escape observation in any experiment
on man with which we are familiar.

McCOLLUM— RELATION OF DIET TO PELLAGRA. 47

The results of experiments with grown men restricted to experi-
mental diets for a few weeks or months do not form a safe basis for
drawing conclusions as to the quality of the foods employed. Cer-
tain conclusions may be warranted from general observations on
children living on faulty diets, and important deductions may safely
be drawn from the experiences of large groups of people living upon
more or less restricted lists of foodstulTs. Beyond this we must be
guided in human nutrition by the results of animal experimentation,
in which the conditions can be made sufficiently rigid to bring into
stronger contrast the faults of certain types of diets as contrasted
with others. It is certain that the injurious efifects of certain die-
tary practices are very real and yet not promptly apparent. The
debilitating effects of faulty diets may vary in their severity from
such as will produce polyneuritis or xerophthalmia or scurvy within
a few weeks, at one extreme, to such as will cause nervousness and
restlessness in varying degree, susceptibility to disease, and the ac-
quisition of. all those characters such as roughness of the skin, thin-
ness and coarseness of hair, and attenuation of form which accom-
pany the process of aging at a distinctly greater rate than would be
the case were the diet of a highly satisfactory character.

We have much evidence that in case there is a close approxima-
tion of the actual physiological minimum for any factor during
growth, such as one or more of the essential inorganic elements or
one of the unidentified dietary essentials, lack of ability to meet the
more strenuous demands of reproduction and the suckling of young
will be observed, and the tendency will be great for the individual
to be carried off suddenly either by disease, or, as frequently hap-
pens, by causes which are not readily determinable.

All our experience with diets of low protein content have indi-
cated that animals do not remain in a state of optimum well-being
even when the content of protein is sufficiently high to maintain in
certain individuals the initial body weight over as much as lo per
cent, of the normal span of life. We believe that health and vigor
are promoted by a liberal intake of protein of good quality better
than by any diet in which there is a tendency towards parsimony
with respect to this dietary factory. It should not be lost sight of,
however, that there are other factors in nutrition which are of equal

48 McCOLLUM— RELATION OF DIET TO PELLAGRA.

importance with protein, and that if the optimum well-being is to
be attained the diet must be rightly constituted with respect to all
its parts. In addition to this the prompt elimination of the fecal
residues is essential and is a great relief to the tissues of the entire
body.

With an understanding such as we now have of the nature of
the faults of diets of different types, and an appreciation of the
fundamental importance of deriving the constituents of the diet
from the right sources, this being of much greater importance than
composition as revealed by chemical analysis, one is in a position to
interpret the relation of pellagra to diet.

Goldberger has emphasized the fact that the diet of those living
in districts in which pellagra is common is lacking in sufficient
amounts of certain foodstuffs, especially milk, eggs, meats, and the
legume seeds. In many instances bolted wheat flour, degerminated
corn-meal, polished rice, sugar, syrups, or molasses, sweet potatoes,
and meat, principally pork, form almost the entire list of foods
eaten by families during the winter season, at the end of which new
attacks of pellagra are regularly seen. From what has been said it
will be evident that the diet of the pellagrous is deficient in four re-
spects, and that the nature of these is well understood. They are
the deficiencies of the plant products which belong to the storage
organ group, but more pronounced because of the prominent place
which milling products, which represent the endosperm of the seed,
find in such diets. Products such as bolted fiour, degerminated
corn-meal, and polished rice are decidedly poorer in inorganic ele-
ments than are the seeds from which they are derived ; their pro-
teins appear to be of poorer quality than are those of the cell-rich
structures near the periphery, or of the germ, and they are almost
devoid of fat-soluble A and very poor in water-soluble B. Wliereas
diets derived from whole seeds, tubers, and edible roots contain
sufficient phosphorus to meet the requirements of the most rapidly
growing species of animal, such as the rat. and the limiting inor-
ganic elements are calcium, sodium and chlorine, it may be that in
diets in which the degerminated and decorticated cereal products are
employed in liberal amounts, and where in addition starch, sugar
and molasses are regularly used freely, phos^jhorus or iron or both
may likewise become important deficiencies.

McCOLLUM— RELATION OF DIET TO PELLAGRA. 49

Goldberger attempted to solve the problem of whether pellagra
is due to lack of something essential in the typical " pellagrous " diet
by a direct experiment on man. He restricted men to a diet pre-
pared from bolted wheat flour, degerminated corn-meal, polished
rice, starch, sugar, syrup, pork fat, sweet potatoes, cabbage, collards,
turnip greens and coftee, and at the end of five and a half months
five of the eleven men who took this diet were diagnosed as exhib-
iting incipient signs of pellagra. That the disease was actually pro-
duced has been emphatically denied by McNeal.

In another experiment Goldberger and fifteen of his associates
made heroic attempts to infect themselves with material from the
lesions of pellagra, and with excreta from pellagrins, but without
success. The experimenters were, however, taking a diet of good
quality while these attempts were being made.

Still more convincing evidence that the diet is at least an impor-
tant predisposing factor in the etiology of pellagra is furnished by
the experience of Goldberger in improving the diets in institutions
in which the disease was common. These diets were observed to
consist largely of degerminated seed products, tubers or roots, and
fat pork, together with minimal amounts of leafy vegetables, fruits,
eggs, meats, and milk, and the legume seeds. On modifying the
diets of orphanages and of an insane asylum by the addition of lean
meat, milk, eggs, and peas or beans, the condition with respect to
pellagra steadily improved, and the disease promptly disappeared.
New cases were admitted from without and the sick were mingled
with the well, but after the improvement of the diet no new cases
developed.

Those who have had extensive experience with pellagra are in
agreement in the matter of the fundamental importance of dietary
treatment together with any other method of management of pella-
grins, and the assertion has been made by Roussel that without
dietary measures all remedies fail. The results obtained by Gold-
berger point clearly to the belief that the disease develops because
of some one or more faults in the diet. They afiford no basis, how-
ever, for judging as to the nature of these faults, whether they are
in the nature of a lack of a sufficient amount of one or more chem-
ically unidentified dietary essentials of a specific character, as is

PROC. AMER. PHIL. SOC, VOL. LVIII, D, JUNE 25, I9I9.

50 McCOLLUM— RELATION OF DIET TO PELLAGRA.

known to be the case with beri-beri and the xerophthahnia of die-
tary origin, or whether pellagra may be the result of taking a diet
faulty in respect to the quality or quantity of protein, relative short-
age of one or more of the essential inorganic elements, or of the
recognized unidentified dietary essentials as contributing factors.

In his earlier papers Goldberger expressed the view that : " On
the whole, however, the trend of available evidence strongly sug-
gests that pellagra will prove to be a ' deficiency ' disease very closely
related to beri-beri." Chittenden and Underbill reported the pro-
duction in dogs of a condition suggestive of pellagra in man by re-
stricting the animals for periods of from two to eight months to a
■ diet of crackers, peas, and cottonseed oil. They formulated the con-
clusion that : " From the facts enumerated the conclusion seems
tenable that the abnormal state may be referred to a deficiency of
some essential dietary constituent or constituents, presumably be-
longing to the group of hitherto unrecognized but essential compo-
nents of an adequate diet."

We have reported elsewhere the results of a study of the nature
of the dietary faults of a mixture of bolted wheat flour, peas, and
cottonseed oil, and found that it was an incomplete food, but that it
was rendered complete for the support of normal growth in the
young rat by the addition of purified protein, certain inorganic salts
(NaCl and CaCOs) and fat-soluble A (in butter fat). It is of
course not satisfactorily established that the condition produced in
dogs by the diet of Chittenden and Underbill was actually the coun-
terpart of pellagra in man, strikingly similar as the results appear.
We hold the view that if the condition produced in the dogs of these
investigators is actually to be regarded as experimental pellagra, it
cannot be regarded as caused by the lack of an unidentified dietary
essential, since the only one of these necessary for completing the
diet (for the rat) is that contained in butter fat, and the latter sub-
stance is not curative for any condition resembling pellagra, but for
a specific eye disease, xerophthalmia.

In his most recent studies Goldberger and his associates exam-
ined the diets of pellagrous and non-pellagrous families in villages
in South Carolina, and found that the diet of the non-pellagrous
contained more milk, fresh meats, eggs, butter, and cheese than did

McCOLLUM— RELATION OF DIET TO PELLAGRA. 51

the diets of pellagrous families, and that calorific value of the diets
of the former households was somewhat higher than of the latter.
Animal proteins were eaten more liberally and cereal proteins were
eat<=^n less abundantly by the non-pellagrous than by the pellagrous
households. The pellagrous households had a distinctly smaller
supply of fat-soluble A, and a somewhat smaller supply of water-
soluble B than did the non-pellagrous, and the inorganic content of
the diets of the latter were of less satisfactory character than those
of the former households. We do not regard a moderate shortage
of one or another of the chemically unidentified dietary factors as
of greater gravity than faulty character in any other dietary factor.
Our studies of the several foodstufl:s lead us to agree with Gold-
berger's interpretation of the quality of the diets of pellagrous and
non-pellagrous households in all respects.

From the observations which we have made concerning the
chemical factors which the diet must contain in order to be adequate
for the support of growth in the young, or the maintenance of phys-
iological well-being in the adult, together with the results of our
studies of the qualities of each of the more important kinds of nat-
ural foodstuffs, we are not able to account for the etiology of pel-
legra on the assumption that it is a disease which is due to the lack
of a specific substance or substances of unknown chemical nature,
as are without question beri-beri and xerophthalmia. This follows
from the fact that, with the exceptions of certain manufactured food
products which are derived from the endosperm of the decorticated
grains, any natural foodstuffs of the class of seeds, tubers, edible
roots, or leafy parts of the plant, are so constituted that they can be
supplemented by means of three kinds of purified food additions of
known nature : viz., protein, certain salts, and fat-soluble A, so as
to be complete for the nutrition of the young rat throughout the
growing period. This has been demonstrated to be true not only
for each of the ordinary human foods but likewise for such mixtures
as form the monotonous diets of the pellagrous.

It is necessary, therefore, that we choose between two alterna-
tives in arriving at an opinion concerning the etiology of pellagra.
We have the assurances of Goldberger and his associates that a diet
such as that described on page 49, and having the qualities described

52 McCOLLUM— RELATION OF DIET TO PELLAGRA.

in the preceding paragraph, has produced incipient pellagra experi-
mentally in man, but this claim has been disputed by other competent
observers. In our experimental work with the diet of peas, crackers
(wheat flour and fat), and cottonseed oil, which in the experience
of Chittenden and Underbill produced in dogs a condition resem-
bling pellagra in man, produced in rats only general malnutrition,
without the skin changes, diarrhea, or pathological changes in the
mucosa of the alimentary tract. Are we to accept the view that
pellagra is actually produced by a deficiency of something necessary
to the normal nutrition of man but not necessary for the rat ? The
possibility that the dogs of Chittenden and Underbill were infected
is not excluded, and an infectious agent may well have established
itself in animals restricted to a diet so faulty as one derived from
crackers (wheat flour and fat), peas, and cottonseed oil. Gold-
berger seems to have safeguarded his experimental men against in-
fection, and it is unfortunate that a sufficient number of undisputed
authorities were not called into consultation to forestall the possi-
bility of a question arising concerning the accuracy of the diagnosis
of pellagra, such as McNeal has raised.

We are left in the situation which has arisen in the discussion of
the etiology of scurvy. It has been clearly shown that there is <no
difficulty in repeating the experimental work which demonstrated
that a guinea pig will develop severe scurvy (or some syndrome
resembling it) on a number of diets on which the rat will thrive
during the growing period. Does this mean that the guinea pig
requires one or more chemical complexes for its nutrition that are
dispensable to the rat? There is no doubt that the guinea pig nor-
mally takes a diet rich in succulent vegetables, and which produces
bulky, easily eliminable feces. The rat and swine, as well as man,
thrive on certain diets which leave little indigestible residue. Such
special requirements in the guinea pig make it next to impossible to
compare this species with man or the rat in similar dietary studies.
The experimental data obtained with the guinea pig must be used
with caution in reasoning concerning the etiology of human scurvy.

In our attempts to produce in animals a condition analogous to
pellagra in man we have not been successful, but have observed only
a generalized poor condition instead. The evidence is practically

McCOLLUM— RELATION OF DIET TO PELLAGRA. 53

nil that Chittenden and Underhill's dogs suffered all the pathological
changes which they record solely as the result of chemical faults in
the diet. Our experiments with their diet show it to be incapable
of maintaining satisfactory nutrition because of faults in the dietary
factors. The possibility of an infection in their animals is not ex-
cluded, and is indeed rendered probable if we grant that lowered
vitality predisposes to infection. These reasons together with the
lack of positive proof that the men restricted in Goldberger's experi-
ments to a diet similar to that described were actually developing
pellagra, warrant, we believe, our accepting as probably correct the
conclusions of the Thompson-McFadden Commission and of Jobling
and Peterson that pellagra is caused by an infectious agent, and
that unless it has been introduced into a district there -may develop
such a condition of lowered vitality from faulty diet or other debili-
tating influence as would predispose one to an attack, without the
appearance of the disease. The debilitating effects on animals of
diets derived from cereals, tubers, roots, and any food products
formed from the milling of grains together with legume seeds and
meats, are so striking that we believe similar diets would produce
in man a susceptibility to infectious diseases such as tuberculosis or
pellagra. We have come to hold the view, as the result of our
studies of diets of the type common in pellagrous households, that
the predisposing influence for both is in general the same, and the
character of the unsanitary conditions surrounding the individual
may determine which of these two diseases he will develop.

From the studies which we have described elsewhere we are
enabled to point out definitely the relative values of several foods as
correctives in the diet of the pellagrous. Hitherto the legume seeds
and lean meat were classed with milk and eggs in this respect, and
nothing was said about the unique qualities of the leafy vegetables
as supplements for food mixtures derived from plant products of
the storage organ group. It is clear that the most important food
to be recommended for consumption in pellagrous districts is milk,
because of its cheapness as compared with the same protective value
in foods from other sources and its threefold corrective character
as contrasted with meat which enhances the type of diet found in
the pellagrous household only with respect to the protein factor, and

54: McCOLLUM— RELATION OF DIET TO PELLAGRA.

eggs which are not so good as milk because of their lower calcium
content. The legume seeds, notwithstanding their high content of
protein, are without any appreciable value for improving the diets
which predispose to pellagra, because of the poor quality of their
protein and their failure to supplement a diet derived from vegetable
foods of the storage tissue class in other respects.

Both meats and eggs are more expensive sources of protection
against faulty diet than milk. An effective campaign of education
should be conducted in all districts where diets of a character likely
to predispose to pellagra are common, informing the people about
the great benefits to health from regular and very liberal use of leafy
vegetables. This would be a movement toward the establishment
of dietary practices resembling those of the more nearly vegetarian
groups of Chinese and Japanese, and if in addition the inclusion of
a suitable amount of milk in the diet can be secured, not only would
pellagra disappear, but the general health of the people would be
promoted.

The prevalence of pellagra in certain parts of the South rather
than in other sections of the country is probably closely connected
with the development of the modern milling industry. This places
in the grocery store the degerminated and decorticated part of
the grain. The rise of the sugar industry offers for human con-
sumption both sugar and molasses in quantities unheard of until
recent years.

The widespread practice of growing a cash crop (cotton), and
of depending on the retail store for the greater part of the food
supply rather than of engaging in diversified farming appears to be
in great measure responsible for the existence of pellagra. The
food products which can be handled commercially without hazard
are not in general satisfactory foodstuff's unless properly supple-
mented with certain others which correct their deficiencies.

School of Hygiene and Public Health,
Johns Hopkins University,
Baltimore.

EVOLUTION AND MYSTERY IN THE DISCOVERY OF

AMERICA.

By EDWIN SWIFT BALCH.
(Read April 24, 19 IQ-)

It is usually supposed that the American continent was revealed
to Europeans by an incident known as the discovery of America.
This is assumed to have taken place almost with the unexpected-
ness and rapidity of a stroke of lightning or the breaking of an egg.
The reality, however, is quite different. The opening up of the
North and South American continents to the white races of Europe
is due to a long series of events and movements, that is to evolu-
tion, and not to a single occurrence ideally fathered in one indi-
vidual's brain and which his genius brought from an ideal stage
into a concrete matter of fact one. Not only was it a process of
evolution which made known the New World to the Old World,
but it was an evolution so long drawn out that it was entirely un-
noticed at the time and as a result many of its steps are forgotten
and much of it is most hazy. Indeed the most salient point in
regard to the recognition by Europeans of the existence of the
American continent is that it is an evolution veiled in mystery.

All geographical discovery indeed is an evolution and this evolu-
tion proceeds from many causes. One of them is the movement
and migrations of peoples. People slowly filter into lands new to
them, discovering as they go new sites in which to hunt and in
which later to dwell. They are not on the outlook for discoveries
proper, they are not searching to penetrate the secrets of the
unknown : they are merely seeking new spots in which to live, be-
cause for some reason they are driven out of their former homes.
Overpopulation it may be, or the drying up of the water supply and
the destroying of the soil by cutting down the forests, or some other
cause ; but whatever the cause something impels certain peoples

55

56 BALCH— EVOLUTION AND MYSTERY

to move into fresh fields and pastures new and by so doing they
open up new lands and new waters.

The most potent cause acting on the individual to make him a
geographical explorer is the fascination of the unknown. Some
men seem impelled by the spirit of curiosity to pry into things
hidden from their ken. When more highly developed this changes
into the spirit of research and those smitten by it delve into the
unknown with the scientific purpose of adding to the sum of human
knowledge. In some men this spirit manifests itself in trying to
unveil the untouched parts of the earth, seas or lands, even if merely
some small untrodden mountain peak, to which the fascination of
the unknown lures them on with irresistible appeal.

As a sequel to this, though in lesser degree, a good deal of ex-
ploration is due to the desire for adventure or sport which some men
are imbued with. Without any special curiosity to seek new parts
of the earth, they are driven by their nature to hunt or to fish.
And in their wanderings carried out to gratify other tastes, inci-
dentally frequently they make geographical discoveries which later
comers complete and verify.

An extremely active and effective agent in bringing about knowl-
edge of the surface of the earth is commerce. Peoples search for
greater food supplies ; they search for new countries from which to
bring home useful products ; they search for new markets in which
to sell their own productions. Merchants put up the cash and sea
captains and sailors start off to make that cash fructify. Many
voyages have been made in past centuries in which geographical
discovery was entirely secondary to commercial gain and yet in the
pursuit of the latter some rich prizes in exploration have fallen to
traders who cared but little for them. Some men have sought new
fishing grounds ; others have followed fur-bearing animals ; others
again have gone to enslave their fellow men. Sometimes success-
ful, sometimes unsuccessful in their quests, these men, in a purely
accidental way, have revealed many a new land or sea to their
fellows.

What we know of past geographical discovery depends mainly
on two sets of evidences : written records and maps. .V trifiing part

IN THE DISCOVERY OF AMERICA. 57

of it depends on archeological evidence and another on tradition.
The most important evidences are : the narratives of travelers them-
selves or the accounts written of travelers by their contemporaries ;
maps. In many cases a map is the only record. It delineates a
land or a sea of which there is no written account dating back as
far as the date of the map. And when one finds on a map some
earth feature correctly placed and even if one knows nothing of it?
discovery or its discoverer, it is difficult to refuse to recognize that
that particular bit of the world had been seen and reported by some
perhaps forgotten person. Maps in fact are something like por-
traits. A genuine portrait is hard to deny. So valuable as iden-
tifications are photographic portraits, that now they are placed on
passports in some countries. If at all accurate, maps are among the
most uncontrovertible evidences of geographical discovery. And it
is precisely on maps that many of the riddles connected with the
unfolding of the geography of the American continent are
presented.

The earliest European invaders of the American continent found
it tenanted by native races. Who were these races and when and
where did they come from? While no positive answer can be
given to these questions, yet we know now that our American
natives have been here a long, long time, and that they had been
acted on by the force of evolution for many millenniums before the
white race came to contest with them the ownership of the lands
stretching from Grant Land to Tierra del Fuego. The ancestors of
our historic Indians and of our present Eskimo were, of course,
the real original inhabitants of America, but when and how they
became so is at present unknown to us. Were they autochthones?
We do not know ! Or did they wander in from Asia ? Again we
do not know ! But if they did the latter, and it certainly seems
the most probable, then they spread over the American continent by
migration and, in a certain sense therefore, were its original
discoverers.

The evolution of the European invasion of America, as far as
accessible records show, begins in early historic times. Between
the years looo and 500 B.C., some Phoenicians circumnavigated

58 BALCH— EVOLUTION AND MYSTERY

Africa. They started from Egypt, went down the east coast of
Africa, rounded the Cape of Good Hope, and returned via the west
coast of Africa and the Straits of GibraUar. Their chronicler,
Herodotus, unfortunately gives them a very meager notice. Never-
theless theirs was the first historic voyage on the Atlantic Ocean and
the first recorded step leading to the discovery of America by the
peoples of Europe.^

The next attempts at exploration in the Atlantic of which we
have any record are the journeys of the Carthaginians Hanno and
Himilko. Hanno sailed from Carthage about 500 B.C., passed
through the Straits of Gibraltar and turned south along the African
coast. He reached a country where he found hairy men with some
of whom he had fights in which he killed three whose skins he
brought back to Carthage where they long hung in one of the
temples. These hairy men are by some writers surmised to have
been chimpanzees or gorillas. At about the same time as Hanno,
Himilko is said to have made a journey beyond the columns of
Hercules, to have turned north and to have reached a land where
there was much tin. This probably was Great Britain. The
Phoenicians, however, may have known more about the Atlantic than
has come down to us in any record, for it is said that coins from
Carthage and other towns in North Africa have been found on the
island of Corvo, the most westerly of the Azores.

The fourth recorded journey of exploration in the Atlantic is the
voyage of Pytheas. A Greek, of Massilia (Marseilles), he sailed
through the Straits of Gibraltar about 350 B.C., turned north and
reached the shores of Britain. Along these he sailed far to the
north, until he arrived in a region where a sort of thickening of
the elements, which he said resembled a marine lung, filled all space.
It is surmised that he had reached the rains and fogs of northern
Scotland and perhaps of the Orkneys and Shetlands. a surmise the
more probable as he speaks of the sun dipping under the horizon for
only a short time.

^ A few references are given with this paper in regard to facts only
lately brought into prominence. For the other facts mentioned, references
by the hundred may be found in Edouard Charton's " Voyageurs Anciens et
Modernes"; in Justin Winsor's "Narrative and Critical History of America":
and in the works of A. E. Nordenskjold and Henry Harrisse.

IN THE DISCOVERY OF AMERICA. 59

The Romans, through JuHus Caesar's conquest of Britain, made
a step forward by clearing up the geography of the British Isles.
After their departure, when Britain proper went into anarchy, for
several centuries Ireland remained in a decidedly advanced state
of culture. Ireland also is the most westerly outpost of Europe.
These two factors combined lend authority to the statement in one
of the Norsk Sagas that when the Vikings first reached Iceland in
the ninth century A.D. they found some Irishmen settled there.
The importance of this is obvious. Iceland is much nearer to
America than it is to Europe. From a navigator's point of view it
is an American island. Any people who reached Iceland was bound
sooner or later to reach America. And effectively in another of the
Norsk Sagas, there is a mention of a Great Ireland far in the west,
which may refer to Greenland. But while it is impossible to say
exactly what the Irish did do or where they got to, it is nevertheless
unquestionable that they forged an important link in the evolution
of the discovery of America. Indeed their discovery of Iceland
seems to warrant considering them the first European discoverers
of an American land.

The Norsemen or Vikings it was who first made recorded ex-
plorations beyond Iceland, which they reached by the ninth century
A.D., and where they settled, lived and quarrelled. They left semi-
historic and semi-poetic narratives, the Sagas, of some of their
doings and from these we can glean a good deal about their ex-
plorations which if not historically accurate, is at least too accurate
not to be based on actual facts. It was some time before the year
looo A.D. that one of the Vikings reached Greenland. Did he sight
Greenland from Iceland, which is said to happen occasionally in the
clearest weather? At any rate a certain Gunnbiorn is claimed to
have sighted Greenland before 900 A.D. Or did the first discoverer
have a quarrel and was he driven out of Iceland ? There is a story
to that effect about Eric the Red, who is believed to be the first
Norseman who actually landed on the shores of Greenland. He
was followed by his son Leif Ericson, by Bjarni Herjulfson, by
Thorfinn Karlsefni, and others. Settlements were formed in
Greenland and further voyages were made to Markland, to Hellu-

60 BALCH— EVOLUTION AND MYSTERY

land and to Vinland. All these places were real, were genuinely
reached, but unfortunately their location is uncertain. Some
writers think the Vikings reached Labrador and Newfoundland;
others think they got as far as Cape Cod; and to this view I am
myself inclined. But from the evidence, the most southern locality
reached by the Vikings can not be positively identified.

But in Greenland the Vikings left positive traces of their so-
journ in the shape of ruins. Some of these are the remains of
churches, and it is known that there were missionaries and religious
men among them. And towards the years 1 100-1200 A.D., there
is no doubt that there was intercourse between the Greenland settle-
ments and Iceland. Slowly however, darkness settles over the
Greenland Vikings. Did they all come back : did they die out : or
did they migrate to the West? We do not know. Only a few
years ago, some rather light-haired Eskimo were found near Corona-
tion Gulf and were surmised to be descended partly from the Green-
land Vikings. At any rate these latter were gradually forgotten
in Europe, where Greenland remained on the maps, however, as a
peninsula attached to Northern Norway, for instance in a map of
1368 and in the Ptolemy of i486. All this is hazy, it is mysterious,
but nevertheless it is certain that the Norse Vikings were genuine
European invaders of the New World.

One tale there is that about 1170 A.D. a Welsh chieftain, by
the name of Madoc, reached the coast of North America. This
legend is unsupported by any evidence save that there is a simi-
larity between some half a dozen Welsh and American Indian
words, and this evidence is so perfectly flimsy, that it seems safe to
relegate a Welsh discovery of America to the realm of fairy stories.

It was not, however, the voyages of the northern Vikings, it was
the voyages of the Portuguese, Spaniards and English, which led to
the peopling of the American continent by the European races.
While striving to reach the East Indies by sailing west, the hardy
mariners from southern and western Europe found the way barred
by the West Indies or New World. Already in the fourteenth
century, it is said as early as 1350, the Portuguese began to make
gradually lengthening voyages to the islands off the north-west

IN THE DISCOVERY OF AMERICA. 61

coast of Africa. Why did they do so? It seems as if there were
several causes. Of course there was the fascination of the un-
known, the joy of discovery, but undoubtedly the main cause was
commercial. And this commercial impetus apparently was due to
the Crusades and to the invasion of Europe by the Turks. For
those wars, in which Christians and Moslems smashed each other's
heads for the glory of God as they still only yesterday were doing in
the Balkans, blocked the commerce of Venice with the East and led
the Portuguese and the Spaniards to seek a way around Africa to
obtain the spices of India. Thus although the great Portuguese
explorations were not wholly due to the blocking of the Oriental
trade routes by the Turks, still they must have been so to some
extent. And, in certain respects therefore, the invasion of America
by the Europeans must be considered as a result of the invasion of
Europe by the Turks.

The lost Atlantis also must have helped to arouse the curiosity
of medieval seamen. Eor the account given in Plato that a great
island or continent had been destroyed in past millenniums was cer-
tainly known in the Middle Ages. Atlantis was then supposed to
be in the Atlantic Ocean, whose name not impossibly is connected
with that of the island, and some early navigators may easily have
had it in mind in searching for new lands or islands. It was indeed
not until A.D. 1909, that a real foundation for the Atlantean tale
was offered, when some thoughtful person suggested in an anony-
mous letter to an English newspaper that the lost Atlantis of Plato
was Minoan Crete. And certainly Plato's story answers in many
respects to Minoan Crete and its destruction as revealed by archeol-
ogy, and in an article " Atlantis or Minoan Crete, "^ I tried to ex-
plain Plato's narrative by comparing it with Minoan Cretan and
Greek history and mythology. One point, however, namely the
statement that the sea around Atlantis had become unnavigable, re-
mained very dim. About this, Mr. William H. Babcock, the author
of numerous scholarly and important papers about the discovery
of America, in a letter to me, suggested that this might refer to
the Sargasso Sea, and this certainly seems like a plausible explana-

- 2 The Geographical Review, May, 1917.

62 BALCH— EVOLUTION AND MYSTERY

tion. For if it is true that Phoenician coins have been found on
one of the Azores, it is possible that some knowledge of the Sargasso
Sea may have filtered into Egypt and become mixed up with the
destruction of Minoan Crete. But although the story of Atlantis is
so confused and so jumbled out of discordant elements that we may
never feel quite sure of them all, yet as far as the discovery of
America is concerned, we must remember that the belief that there
had been such an island in the Atlantic and that it was close to other
lands, may well have urged on some of the early Atlantic navigators.
Among the impelling causes of Portuguese discovery also, must
be reckoned one man. This was the Portuguese Prince, known as
Henry the Navigator. He w'as the son of John I., King of Por-
tugal, and his wife Philippa, daughter of John of Gaunt, Duke of
Lancaster, and was thus half English. Born in 1394, he died in
1460, and devoted most of his life to fathering voyages of discovery.
Some English writers laud him to the skies as the finest flower of
chivalry, but one Portuguese wn'iter at least claims that he was a
brutal bigot and slave driver. Leaving his moral character to the
tender mercies of others, history records that in June, 141 5, Henry
the Navigator took part in a predatory attack on the town of Ceuta,
Morocco, which was wrested vi ct arniis from the Moors. Appar-
ently his sojourn with the Moors inspired him with a desire to ex-
plore the African coast of the Atlantic Ocean and after his return
to Portugal in 1418, he settled on the promontory of Sagres, Por-
tugal, a name corrupted from its name Sacrum of Roman days, and
built an observatory and some houses there. Here he spent most
of his life, studying geography, interviewing seamen and preparing
expeditions of discovery. In certain respects, the settlement of
Sagres might be considered the first geographical society of the
world. And although Henry the Navigator did not himself make
any voyages of discovery and although none of the voyages which
he fathered made known to the world the existence of the Ameri-
can continent, still his work aroused the interest and opened the
way which eventually led to the successful voyage of Columbus.
Indeed there is probably no individual who did as nuich to bring
about the European invasion of America as Henry the Navigator.

IN THE DISCOVERY OF AMERICA. 63

Many of the early voyages of Portuguese or Spaniards were
either unrecorded or else the records were destroyed and lost.
Proof of this is shown by the fact that constantly the Portuguese
and Spanish libraries and private archives yield documents revealing
totally unknown circumstances. As an instance of this may be cited
the case of one whole district of the island of Madeira which is
called Machico, and that this was the name of a Portuguese sailor
of 1379 is shown by a unique document found only in 1894.^ Of
one early voyage, however, the strange invasion or conquest of the
Canary Islands by the Sieur Jean de Bethencourt, a Norman French
nobleman, in 1402-1405, we have an elaborate account, one copy of
which in the Bibliotheque Nationale of Paris is illuminated like a
missal with old drawings. But although the written records are
scanty, there is no doubt that during the fifteenth century many
commercial ventures and voyages of exploration into the Atlantic
were made, not only by Portuguese and Spaniards, but also by
Englishmen. Cape Bojador was rounded by Gil Eanes in 1435,
Cape Verde in 1445 and between then and 1448 it is said that more
than thirty ships sailed round it. Long voyages into the Atlantic
are recorded in 1452, 1457, 1460, 1462, 1473, 1475, 1476, 1480-81,
1484, i486. Some of these went at least 150 leagues to the west;
one of them discovered or rediscovered the Sargasso Sea ; all of
them were in search of lands or islands in which the belief seems
to have been very general. There are some rough written records
of some of these voyages and they tell very fairly the evolution of
discovery down the western coast of Africa.

The best records of discoveries in the western Atlantic in the
fourteenth and fifteenth centuries, however, are maps. For as I
have said earlier in this article, the evidence of maps is exceedingly
hard to controvert. When land or sea is charted with some ac-
curacy, even if there is no message in words about the matter, the
chance is that a discovery had been made. And on these early
maps of the Atlantic, the factor of especial importance in regard to
America are the islands and their names.

Some names of islands persist with striking tenacity on these

2 Batalha-Reis, The Geographical Journal, February, 1897.

64 BALCH— EVOLUTION AND MYSTERY

early maps. One of these names is Brazil or Brazir, which is
claimed to be Irish, either a compound of the Irish Gaelic words
"breas" and "ail" or to be descended from the Irish Saint Bresal.
It is applied both to a small island not far west of Cape Saint
Vincent, Portugal, and also to an island and sometimes to two
islands in the same longitude not far west of Ireland. This
northern Brazir is found on Dulcert's map of 1339, on the Catalan
Atlas of 1375, etc. On the Pizigani map of 1367, there are three
islands of Brazir in about the same longitude, and one of them is
placed in the latitude of Brittany with two ships in distress flying
the Breton flag near it. Mr. William H. Babcock,* the best living
authority on the subject, has suggested that this island is intended
to represent Newfoundland and possibly was discovered by the
Irish, and before 1367 reached also by some Breton sailors. There
is one objection to this theory, however, and that is the longitude of
the islands of Brazir. The northern island or islands are not far
from Ireland. If any mariner had genuinely reached Newfound-
land before 1367, it would seem as if he would have reported it as
an island far away from Ireland. There is of course no island in
the charted position, and if the latitude is correct it could be only
Newfoundland. And it must be recognized that even though the
longitude is incorrect yet that the latitude and the bigness of the
island are decided evidence in favor of Brazir being Newfoundland.
While the theory must be considered as non proven, still the evi-
dence is sufificiently strong to prevent it from being lightly thrown
aside.

There is one island or group of islands which is especially note-
worthy, Antillia, from which we have made the Antilles. Some
writers have sought to derive the name from Atlantis, but Mr.
W'illiam H. Babcock has shown that the proper derivation comes
from Anti-Ilha or Ante-Ilha, meaning the island in front of, that is
to the west of the others. Originally Antillia seems to have been
considered as an island to which many Portuguese are reported to
have fled as a refuge when the Arabs invaded the Iberian peninsula
in the beginning of the eighth century. At least there is a legend

* " Indications of visits of White Men to America before Columbus,"
Proc. Nineteenth International Congress of Americanists, Washington, 1915.

IN THE DISCOVERY OF AMERICA. 65

to this effect on Martin Behaim's globe of 1492, on which it is also
stated that a Spanish ship visited Antillia in 1414. In the four-
teenth century the name Antillia was applied to the most westerly
of the Azores. In the fifteenth century, however, several cartog-
raphers, Andrea Biancho in 1436 among others, drew maps which
jump Antillia far, far to the west of the Azores. It is figured as a
great big island, in about the latitude of Portugal; with another
big island, Salvagio or Satanaxio, to the north of it. The main
extension of Antillia and Satanaxio is about north and south, their
latitude is somewheres between Florida and Massachusetts, and
considering their distance beyond the Azores, their longitude would
coincide roughly with the eastern coast of the United States.

Although the matter is hazy and there is no verbal evidence
beyond the few words on Behaim's globe, yet the cartographical
evidence shows positively that map makers of the fifteenth century
had a distinct belief in big lands far west of the Azores, in about
the latitude and longitude of the eastern coast of the United States.
Several of these maps certainly point to extensive coasts way
beyond the Azores having been reached before 1435. The name
Antillia would validate rather than invalidate the reality of such
a discovery. And where there is so much smoke it seems as if
there must be some fire. A visit to the Antilles or to the coast of
the United States before 1435 seems much less improbable than that
map-makers should have correctly guessed the existence of big land
masses so far to the west, which in their extension north and
south and in their latitude and longitude, so closely approximate to
the eastern coast of the United States. Personally I am strongly
inclined to believe in the genuineness of this discovery, which may
perchance have been made by the Spanish ship whereof Behaim
speaks.

The most interesting and important pre-Columbian map of the
Atlantic, however, without question is the Venetian cartographer
Andrea Biancho's map of 1448. The problems presented by this
map were broached by Dr. Yule Oldham'' and were admirably de-

5 The Geographical Journal, March, 1895.

PROC. AMER. PHIL. SOC, VOL. LVIII, E, JULY I I, I9I9.

66 BALCH— EVOLUTION AND MYSTERY

veloped by Senor Batalha-Reis.^ Here I wish to merely recapitulate
briefly the chief arguments of these scholars.

Andrea Biancho's map is on vellum 86 centimeters by 63 cen-
timeters. It is now in the Ambrosian Library at Milan and its
authenticity is generally accepted. It is one of the portolani, that
is one of the maps made by seaman for seamen for practical pur-
poses and free from the fanciful geographical conceptions and
fabulous zoological monstrosities of the maps drawn by convent
monks. The map takes in France, Spain and the African coast
down to Cape Verde, and, as far as known, is the first record of the
coast line of Cape Verde.

The remarkable feature of the map, however, is that at the ex-
treme lower left hand corner is represented an extensive coast line
southwest of Cape Verde. On this an inscription in the old Vene-
tian dialect seems to read : " ixola otinticha xe longa a ponente
1500 mia," which may be translated: "Authentic island distant to
the south west 1500 miles." The position of the coast line and the
legend together seem to show that somebody, before the year 1448,
had seen a land southwest of Cape Verde and some 1,500 miles
distant. But there is no island in this position, while at 1,520 miles
southwest of Cape Verde is the northeast promontory of South
America.

There is no known account recording any voyage making such a
discovery, therefore it is well to see whether there is any indirect
evidence that such a discovery might have been made. In the first
place there were a great number of voyages in search of lands
during the fifteenth century. Then we must remember moreover
that some of these voyages were not recorded and were sometimes
entirely forgotten. It is, therefore, perfectly possible that such a
discovery may have been made and left unrecorded if, for instance,
it was made by some merchant vessel or fishing boat blown out to
sea. Indeed this very thing happened in this very place in the year
1500, when Cabral's expedition, en route for India round the Cape
of Good Hope, was blown out to sea and strayed over in a casual

^ " The Supposed Discovery of South America before 1448 and the
Critical Methods of Historians of Geographical Discovery," The Geograph-
ical Journal, February, 1897.

IN THE DISCOVERY OF AMERICA. 67

sort of way to Brazil. That such a discovery was not followed up
before 1492 would be simply because no one at that time could
have any idea of what such a discovery meant. The land would
have been looked on as another island and one too far away to be
of much use.

On Martin Behaim's globe of 1492, there is a large island in the
position of part of the coast of South America, in fact in the exact
position given by the legend on Biancho's map of 1448.

The Treaty of Tordesillas of 1494 insured to Portugal all the
eastern part of South America. Why this treaty should have been
ratified unless the Portuguese already had knowledge of these lands
it is hard to see. Spain claimed the lands discovered by Columbus
and Portugal claimed the lands to the south probably reported by
Portuguese navigators.

On the 1st of May, A. D. 1500, Master Joao, physician to D.
Manuel of Portugal, wrote to the King about the land (Brazil)
just found by the fleet of Cabral on board of which he was, "that
those lands might the King see represented on the mappamundi
which Pero Vaz Bisagudo had." He adds that the mappamundi
does not mention if the land were inhabited, while he could certify
it to be well peopled. He also says that the said mappamundi was
ancient. This last statement is strong evidence that somebody must
have seen the land before 1492.

Las Casas, writing between 1552 and 1561 about the third
voyage of Columbus in 1498, says that Columbus wanted to go
south because he believed he could find lands and islands, and also
because he wanted to see what King D. Joao of Portugal meant
when he said that there was '"' terra firma " to the south. The
grounds for the belief of Columbus have never been explained.

The map itself and the side evidence do not make this landfall
a certainty. But they certainly make it a probability. Personally
the weight of evidence seems to me to lean very decidedly in favor
of South America having been reached by some unknown Portu-
guese navigator half a century before Columbus reached American
shores.

The first southern European voyage to America, however, of

68 BALCH— EVOLUTION AND MYSTERY •

which we have absolute definite historic record, is, of course, the
first voyage of Columbus. This voyage it was which made known
generally to Medieval Europe the existence of lands of great extent
far away to the west in the Atlantic. And although the Irish
reached Iceland and the Norsemen Greenland and some points of
the American coast, and although some Portuguese or Spaniards
probably also did the latter before Columbus, yet, to the world of
Europe, Columbus was the man who kindled the torch which dis-
pelled the darkness enwrapping the West, and in this matter at least,
the popular verdict of history is correct in calling Columbus the
discoverer of America. But, although we have records of his four
great voyages, although we know much about Columbus, yet of
many things connected with Columbus we are still ignorant.

Among the things we do not know with any certainty is why
Columbus came to start on his journey. There has been much talk
that he corresponded with an Italian philosopher named Toscanelli,
who is said to have proved to him that by sailing westward he would
find the coast of Asia. Some have claimed that a study of Marco
Polo was the influencing motive of Columbus. There is a story
that he met on some isle near the African coast a pilot who died in
his arms and who assured him that he had sailed far across the
Atlantic to distant shores. There is another tale which shows Co-
lumbus arriving at Iceland and thus hearing of the lands, already
nearly forgotten in Europe, which had been the goal of the North-
men several centuries before. None of this, however, is in any
respect authenticated : none of it indeed rises much beyond the
status of a legend. What we do know, however, is that Columbus
had the spark of exploration fever burning brightly in his heart
and that it spurred him on for years until he finally won the greatest
prize in the history of discovery.

Again, who actually sighted first an American land on Columbus'
first voyage? Was it Columbus or one of his companions? And
on what day did this happen ? It is usually assumed that Columbus
himself made the landfall on the 12th of October, 1492. But there
is no certainty of this. The records of the voyage show that Co-
lumbus himself claimed sighting land before dawn on October 12,

IN THE DISCOVERY OF AMERICA. 69

but one of his sailors, Rodriguez of Tryana, put in a claim that he
had seen the land in the early morning before Columbus ; and
another sailor on Pinzon's ship, Juan Rodriguez Bermejo, there-
upon put in a claim that he had seen a sand beach reflecting the
moonlight in the evening of October ii. It also happens that
as the old calendar was still in use, the correct dates according to
our present chronology should be some ten days later, either October
20 or October 21. And there the matter stands ! We know neither
who it was of Columbus's expedition who first sighted an Ameri-
can shore, nor do we know exactly on what day this took place.
And this perforce must always remain a mystery, as the data are
insufficient to clear it up.

Perhaps the queerest adjunct of Columbus's great discovery,
however, is the fact that Columbus himself never knew that he
had discovered a continent. Columbus till his dying breath believed
that he had reached outlying parts of India. And there is one
proof of this which can not be gainsaid and that is that he called
the copper-colored natives whom he met Indians. To this day, we
keep on repeating this error and call the original inhabitants of our
continent Indians, with the result that sometimes it is hard to
distinguish as to whether we are speaking of the natives of America
or of the swarthy races of Hindustan. But it does seem hard that
Columbus should have revealed to Europeans the existence of the
vast territory extending from Bradley Land to Tierra del Fuego
and yet have gone to his rest believing he had been on the shores
of Hindustan.

But the very ignorance of Columbus in regard to the American
continent is evidence in favor of earlier discoverers. The early
maps always show islands. No one dreamed of a new continent.
A mariner sighted a coast and told of it and the map-makers charted
it as an island. Maps before Columbus indicated numerous islands
in the western ocean corresponding roughly with parts of America.
It certainly seems far more likely that these islands were based on
some foundation of genuine exploration and of actual perception
by early mariners than that they should have been invented cor-
rectly in cartographic workshops in Europe. Especially probable

70 BALCH— EVOLUTION AND MYSTERY

does this seem, when one remembers that for some years after
parts of the continent had been visited and portions of its coast
had been charted, the explorers themselves still believed that they
had merely revealed the existence of some big islands belonging to
the land mass of Asia.

Mysteries thicken rather than lessen after Colmnbus had gone
around the turning point in American discovery. The explorer who
first announced that there were lands, continental in size, in the
west, apparently was Amerigo Vespucci. He did so in several little
tracts which are tangled up and hard to unravel, but whose publica-
tion in his lifetime tended to make him better known than other
more silent explorers. It was in his letter to Lorenzo di Pier-
francesco di Medici especially that he spoke of the new world he
had visited on the caravels of the King of Portugal. He also said
that the land which they — that is the expedition of which he was a
member — had discovered, seemed to them, not an island, but a
continent. Furthermore he mentioned that they had sighted land
90 degrees of latitude distant from Lisbon — referring almost cer-
tainly to South Georgia in 52° south latitude whilst Lisbon is in.
38° north latitude — and thus had measured the fourth part of the
globe.

Martin Waldseemiiller, an armchair geographer residing at
Saint Die in the Vosges mountains, became impressed with these
statements of Amerigo. In 1507 he published a booklet in which
he suggested that since Amerigo had discovered a Novus Mundus
or fourth part of the world, therefore it should be called after
him. With the book, he brought out a world map, on which he
placed the name America on the lands south of the Equator visited
by Amerigo.'^ Undoubtedly he took an exaggerated view of
Amerigo's reports. He passed over Columbus and Amerigo's even
now unknown commander and looked on Amerigo as the discoverer
of this New World. But barring this error he located the name
America correctly on his map on what is now Brazil.

There is no evidence that Amerigo ever heard of the name
America. Nevertheless because of it he has been abused like a pick-

^ Basil H. Soulsby, " The First Map Containing the Name America,"
The Geographical Journal, Vol. XIX., Feb., 1902.

IN THE DISCOVERY OF AMERICA. 71

pocket and has been accused of stealing the glory which should
belong to Columbus. But all the evidence shows Amerigo guiltless
of aught except of being a member of several important voyages
of exploration and of writing somewhat hazily about them. And
equally, barring the error in regard to Amerigo's commander, there
was every reason why Waldseemuller should name the new land
America. For remember he did not impinge on the discoveries of
Columbus : he applied the name America only to the coast of Brazil,
where it should still be. It was only years after all the actors in
the matter were dead, that the name gradually became attached to
the entire continent. Having myself named the two halves of the
Antarctic Continent, West Antarctica and East Antarctica, and also
some parts of it such as Charcot Land and Nordenskjold Land,
personally I have the deepest fellow feeling for Waldseemuller.

Occasionally some mystery is cleared up. For instance, in 1894
entirely private documents were discovered showing that Joao
Fernandes Lavrador was one of the commanders of one or more
expeditions sent to discover lands to the northwest of Europe be-
tween 1491 and 1496. This was probably the origin, formerly
unknown, of the name of the part of North America which we
call Labrador.^

More often, mysteries are not cleared up. For instance the
map of Juan de la Cosa of 1500 shows a coast line extending from
Newfoundland to Florida which must be the coast of the United
States. The maps of Cantino and of Canerio of 1502 show the
southern part of the coast between Florida and about New York,
then a blank, then Newfoundland. Apparently two navigators had
explored the coast roughly before 1502 but who were they? John
Cabot possibly was the first and Amerigo is sometimes guessed to
be the second culprit. But no one knows !

There are some curious mysteries connected with Magellan's
voyage. One is the statement of Antonio Pigafetta that when
Magellan's expedition arrived at the north of the straits so well
named after him : " all the men were so assured that these straits
had no opening to the west, that no one would have thought of

8 Batalha-Reis, The Geographical Journal, February, 1897.

72 BALCH— EVOLUTION AND MYSTERY

seeking it, except for the great knowledge of the Captain-general
[Magellan], This man, as clever as he was brave, knew that one
must pass through a strait very hidden, but which he had seen
marked on a map made by Martin Behaim, a most excellent cos-
mographer, which the King of Portugal kept in his treasury." The
other strange occurrence is that the same year that Magellan sailed
through the straits, Johannes Schoner made his famous globe, now
in Niiremberg, showing South America much in its real shape,
although extending only to 40 degrees south latitude. Pigafetta's
statement and Schoner's map are not satisfactorily explained as yet.
Perhaps they never will be !

One more unexplained cartographical mystery and I have done.
Straits under the name of Fretum Anian appear as a water separa-
tion between Asia and America on Zaltieri's chart of the year 1566,
almost exactly in the position of Bering Straits explored by Vitus
Bering in 1728. And from 1566 on they hold their own on charts
and atlases, varying their location a little each time until after the
year 1700. If one looks at one of these old charts it is hard to
believe that some one had not explored Bering Straits before 1566.
And yet not only is there no record of any such voyage, but all
the evidence which there is goes to show that there was no such
voyage.® The only solution of this mystery would seem to be
native reports of these straits filtering through to the Chinese and
being communicated by them to some early European traveler. But
there is no record of anything of the kind. Nevertheless although
the known evidence seems to prove that Bering was the first to
sail through Bering Straits, the fact remains that more than a cen-
tury and a half before Bering, some cartographer appears to have
found out that wide straits separated the Asiatic and American
continents.

There are many other unexplained occurrences connected with
the evolution of American discovery which are not touched on in
this paper. But this paper is not an attempt to give a complete
history of the subject. It is rather an attempt to get at the phi-
losophy of the matter ; of the causes which led to the discovery ;

^ F. A. Colder, " Russian Expansion on the Pacific," 1914.

IN THE DISCOVERY OF AMERICA. 73

of its long, slow and gradual development ; and of our lack of
knowledge about any number of its incidents. And I hope I have
carried out my main purpose, whcih is to show that the discovery
of America is not in the least the single occurrence it is popularly
supposed to be, but that it is an evolution, and an evolution tinged
throughout with mystery.

TATAR MATERIAL IN OLD RUSSIAN.

By J. DYNELEY PRINCE.
(Read April 25, 19 19.)

It seems to have been a characteristic of Russia from the earhest
times until the present moment to take a morbid pleasure in her own
failures. Whatever one may think in general of Stephen Graham's^
opinions regarding Russia, he was certainly correct in emphasizing
the prevalence of what may be termed the gospel of incompetency
among the Russians of to-day. Public sympathy has been at all
times in Russian history with the unsuccessful, rather than with the
triumphant hero, and nowhere is this disconcerting trait more co-
gently evident than in some of the literature of the old Russian
period, best exemplified by the " Epic of Igor," or, more fully, the
"Tale of the Armament of Igor" (1185 A.D.).^ This poem relates
in grandiloquent style, often verging upon that of a Scandinavian
Saga, the defeat of the ancient Russian Prince Igorj^ Svjatoslavic
by the well-disciplined Tatar hordes of the Polovtsy in southern
Russia. The epic abounds with words and other traces of the influ-
ence of this and perhaps of other Tatar civilizations, a fact which is
all the more interesting, because this literature antedates by about
two generations the advent of the Golden Horde under the succes-

1 Stephen Graham, " The Way of Martha and the Way of Mary," Lon-
don, 1915.

2 The name Igor (Igorj) Hke so many other princely names of this
period is pure Norse (=Ingvar); cf. Rjurik=: Hrorekr ; Truvor = Thor-
vardr; 01eg = Helgi; Rogvolod = Rognvaldr, etc. For the poem, cf. L. A.
Magnus, "The Tale of the Armament of Igor," Oxford, 1915.

3 The phonetic system of transcription in the present article is essen-
tially the Serbo-Croatian. Note, however, that the apostrophe is used to
denote the Russian hard sign ^ stop or short vowel (Scliwund) and that the
;■ after a consonant = palatalization (Russian soft sign). The Russian
vowel yery (=i in English lid) is represented by y. As regards the abbre-
viations, C. = Cumanian ; CC. = " Codex Cumanicus"; OR. = Old Russian;
OS. = Old Slavonic, and R. = Russian ; ZDMG. = Zcitschrift der Deutschen
M orgenldndischen Gesellschaft.

74

PRINCE— TATAR MATERIAL IN OLD RUSSIAN. 75

sors of Jenghis Khan. These rulers held the Russians in well or-
ganized tributary thralldom for nearly two centuries (from 1223
A.D.).

As attention to the Oriental material in Russian has been called
quite recently, perhaps for the first time in English, by Mr. Magnus,
I have in the present paper ventured to advance some of my own
views as to this subject and to emphasize the points regarding which
I am at variance with the work of that scholar, as well as to set
forth some of the facts now established with tolerable certainty con-
cerning this early period of Slavo-Turkic intercourse. The infor-
mation herein gathered is not intended to be exhaustive and may
be supplemented on the lexicographical side from such works as"
Berneker's " Slavisch-Etymologisches Worterbuch " and Radlofif's
" Worterbuch der Tiirk-Dialekte."

The first question confronting the student of this Tatar* influ-
fluence in Igor is that of the identity of the Polovtsy, who appear
throughout the Epic as the successful and often not unchivalrous
foes of the adventurous hero and his company.

We have direct and convincing evidence in the Chronicle of
Nestor ( 1096) as follows : " And Ismael begat twelve sons, whence
come the Turks, Pecenegs^ (White Huns), Torks (remnants of the
Pecenegs) and Kumans, that is to say the Polovtsy who 'came out
of the desert.' " In other words, the Turkic tribes known to us as
Cumanians were identical with the Polovtsy. It is highly probable
that the word Kiiman is a popular etymology from qum, " sand,"
indicating that these tribes originated in the sandy steppe ; i. e.,
"came out of the desert," but that the original of the word was

Our chief source of information as to the idiom of these Kumans

* Tatar is a name generally applied to all Turkic, Mongolian and Hunnic
tribes; in short, to every Oriental non-Russian people in the former Russian
Empire. See below, note 9. Turkish of practically every variety is more or
less intelligible in essentials to all the Turkic tribes. Hunnic (Finno-Ugric),
however, differs very much in its various dialects.

5 The Pecenegs, or White Huns, were also called Bisseni, Bysseni,
Yla.l^i.va.KiTa.1 in Arabic Badzak, etc. Cf. Anna Comnena, Bonn Ed. p. 404) :
TrpSffeuri ToTi Ka/idvois ws oixwyXuxraon " they are very close lingiiistically to the
Cumanians."

76 PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

or Cumanians is the " Codex Cumanicus,"'^ edited by the Hungarian
Count Geza Kuun, and, in spite of many errors, a most vakiable
record of the speech of the Cumanians, giving a sketch of the gram-
mar, word-lists, and texts with late Latin-Persian-Cumanian in the
first part, and Cumanian-Old-German in the second part/ Besides
this, mention should be made of the brief " Interpreter of the Lan-
guage of the Polovtsy," found in a Russian manuscript of the six-
teenth century,^ which gives a small number of so-called Polovtsian
words with Russian translation. As to the term " Polovtsy " itself,
it would seem to be a cognate with the race-term " Palocz," found
in the Hungarian Chronicle, lised interchangeably with Kun=Hun
= Kuman.® In the Chronicle of Nestor, the word Polovtsy was

^ Comes Geza Kuun, " Codex Cumanicus Bibliothecae ad Templum Divi
Venetiarum," Budapest, 1880; " Additamenta ad Codicem Cumanicum, Nova
Series," Budapest, 1883; W. Radloff, "Das tiirkische Sprachmaterial des
Codex Cumanicus," St. Petersburg, 1887 (Academie Imperiale des Sciences),
criticized by W. Bang in the following works : " Beitrage zur Erklarung des
Komanischen Hymnus," in Nachrichten der kon. Ges. der JVisscnsch. su
Gottingen, philologisch-historische Klasse, 1910, pp. 61-78; Uber einen
komanischen Kommunionshymnus," in Academie royale de Belgique,
Bruxelles, 1910, pp. 1-12; "Zur Kritik des Codex Cumanicus, Librairie uni-
versitaire des trois rois." Louvain, 1910, pp. 1-17; "Beitrage zur Kritik des
Codex Cumanicus," in Academie royale de Belgique, Bruxelles, 191 1, pp.
13-40.

'''As Bang has pointed out ("Beitrage," pp. 32 ff.), the first part of the
"Codex" was probably written by Italians and the second half by Germans,
both parts having been composed under Franciscan influence, as is evident
from the prominence accorded to St. Francis. The scope of the work was
undoubtedly missionary and not commercial, as the chief stress in the
vocabulary and texts is laid on religious material. The " Codex " in both parts
belonged to the library of the poet Petrarch, 1350-1370. Before that date,
the documents were in the possession of one Antonius de Finale ("Codex,"
p. 218). Both parts were probably brought from the Black Sea missions to
Italy, where the manuscript was compiled and edited by Genoese and Vene-
tians. It seems clear that this " Codex " had nothing to do with the Cu-
manians settled in Hungary, who kept their idiom as late as 1744.

8 P. K. Simoni in Proceedings of the Department of the Russian Lan-
guage and Literature of the Imperial Academy of Sciences, 8, 17^191 ; 185-
197, St. Petersburg, 1909.

* See Friedrich Hirth, " Die historisch-g^eographischen Beweise der
Hiung-nu =: Hun Identitat," Budapest, 1910, and cf. also his " Ancient His-
tory of China," pp. 31-35, New York, 1908.

PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

77

plainly associated with Slavonic polje, " field " ; hence " desert," but
polje is a soft noun and would have produced the derivative poljevec
and never polovec.

A brief examination of the material found in the " Interpreter "
mentioned above and a comparison with the Cumanian of the Codex
and with modern Osmanli will satisfy the most cursory reader as
to the true Turkic character of the Cumanian-Polovtsy language.

POLOVTSIAN

I NTERPRETER.

Cumanian.

Osmanli.

teiigri'^^

tengri

tangri

" God "

kok

kok

kjok (gjok)

" heaven "

kujas

kujas

giiiies

" sun "

(probably error

for kunjas)

ilttduz^^

juldus

jildis

" star "

aan

ay

ay

" moon, month "

kar

kar

k(j}ar

" snow "

amgur

yamgur

jaghmur

" rain"

sunk

suk, saok, saogh

souk (soghuk)

" cold "

WJ13

isi hot

ysyq

ekmek (original

" light " {sydzak
hot)

etmenk

ettnac

etmek)

" bread "

The grammatical structure of the Cumanian was also strikingly
similar to that of Sart and Osmanli, as may be noted from the fol-
lowing few examples of the pronouns, nouns and tenses of the verb:

^° In the " Interpreter," the first vowel is the 39th letter of the OS.
alpliabet, often wrongly transcribed ja in Russian. Its real value was a
nasal e, as in eng (= Polish nasal e), but the vowel frequently corresponds
to Russian ja.

For a similar comparison between Cumanian and Tatar, see the work
cited above note 8, and note the incorrect vocalization in tjagri, op. cit., p.
igi. This universal Turkish word is very probably connected with the
ancient Sumerian dingir "God" (soft form dimmer); cf. Prince, Materials
for a Sumerian Lexicon, p. viii, Leipzig, 1909.

1^ Scribal error for iuldus = juldus.

12 Written aan; evidently scribal error for aai.

^^Isi^=ysy, with obscure vowel y; not is (Radloff, op. cit., p. 120).
Rndloff's readings of the Codex are not always trustworthy.

78

PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

PRONOUNS.

CUMAXIAN.

Sart.

OSMANLI.

men; man

vi'dn

ben

"I"

mening

mening

benim

"of me"

manga; man ga

manga

bana (banga)

" to me"

meni; menj

meni

beni

" me " (ace.)

mendan

mendin

benden

" from me"

his

biz

biz

" we "

bizing

bising

bizim

" of us"

bisga

bizga

bize

" to us "

hizni

bizni

bizi

"us" (ace.)

bizdan

bizdin

bizden

" from us "

CUMANIAN.

Sart.

Osmanli.

sen; son

san

sen

" thou "

sening

sening

senin

" of thee "

sanga

sanga

Sana

(saha; saa)

(sanga)

" to thee "

seni

seni

seni

"thee" (ace.)

sendan

sendin

senden

" from thee "'

siz

siz

siz

" you "

sizing

sizing

sizin

" of you "

sizga

sizga

size

" to you "

sizni; sisnj

sizni

sisi

" you " (ace.)

siz dan

sizdin

sizden
NOUN.

" from you "

CUMANIAN.

Sart.

Osmanli.

su

su

su

" water "

suning

suning

sunyn

" of water"

suga

suga

suja

" to water "

suni

suni

suju

'^acc.)

Sudan

sudin

Sudan

" from water "

CUMANIAN.

Sart.

Osmanli.

sular

sidar

sular

" waters "

sularning

sularning

sularnyn

" of waters"

sularga

sularga

sulara

" to waters "

sularni

sularni

sular y

(ace.)

sulardan

sulardin

sulardan

" from waters '

PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

79

VERB.

Present Tense.

CUMANIAN.

Sart.

Osmanli.

anglarmen

anglamcn

anglarym

" I understand "

anglarsen

anglasen

anglarsen

" thou — "

anglar

angladur

anglar

" he — "

anglarbiz

anglamiz

anglarys

" we — "

anglarsic; -sis

anglasis

anglarsynys

"you — "

anglarlar

angladtirlar

anglarlar

" they — "

Present

Tense Negative

anglannan

aiiglamaimdn

anglamam

" I do not under-

(-men)

stand "

anglarmasen

aiiglamaisen

anglamassen

" thou — "

anglamas

anglamaidur

anglamas

"he — "

anglamashis

anglamaimis

anglamajys

" we — "

anglamassis

anglamaibiz

anglamassynys

"you — "

aiiglamaslar

anglamaidurlar

anglamazlar
Future.

" they — "

anglagaymen

anglarmen

anylajadzaghym

" I shall under-
stand "

anglagaysen

anglarsen

anglajadzaqsyn

" thou — "

an g lag ay

anglar

anglajadzaq

"he — "

anglagaybis

anglar mis

anglajadzaghys-

" we — "

anglagaysiz

anglar sis

a nglaja dza qsynyz

" you — "

anglagaylar

anglarlar

anglajadzaqlar

" they — "

As will be observed, the Sart Tatar of Eastern Russia is even
more similar to Cumanian than is Osmanli, as the m-form of the
pronoun of the first person man-men constantly appears instead of
the Osmanli ben. The inserted n before the nominal-pronominal
genitive-ending -ing (->'«), which remains in Osmanli only in words
ending in a vowel, is still common in Sart, as it was in Cumanian.

In 1338 A.D., the Franciscan Friar Pascal of Vittoria wrote that
he learned the lingua Chamanica and the Uigur letters, " which are
used commonly throughout these kingdoms ; " that is. throughout
the empires of the Tatars, Persians, Chaldaeans, Medes and Cathay.^*
In other words, Pascal states that Cumanian was the idiom in com-
mon use as a vernacular throughout Central Asia as far as China

1* Cited Bang, Beitrdge, p. 33.

80 PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

and that it was written with Uigur characters. Cumanian was evi-
dently a term applicable to Tatar in general, including Uigur. ^^
There can be no doubt that the material of the " Codex Cumanicus "
is of great value, therefore, in fixing the philological status of all
prse-mediseval and mediaeval Tatar and especially of the Polovtsian
idiom, with which it was practically identical.

I am particularly indebted to Mr. Feliciu Vexler, Assistant in
Slavonic in my Department in Columbia University, for his able
assistance in collecting most of the following Tatar material, bear-
ing directly on the language of the Epic of Igor.

Tatar Material in Igor.^®

Bl'van (Igor 112)= modern R. holvan, "block, blockhead,
statue, idol" (Berneker, p. 41), C. halahan, " falcon," possibly owing
to the statue-like attitude of the bird when perched. In Mag>'ar,
hdlvdny = " idol of any sort " ; Rumanian holovan, " cobble- stone,"
formerly "idol" (Slavonic loanword). There may be two words
here, the first referring to a bird of some sort ; cf . Turkish biilbiU,
"nightingale" (in CC. rosignolus) ; and the second actually mean-
ing " block " or " idol." The word is clearly of Tatar origin.

Bojan (Igor, passim). For full discussion, cf. "Prince, Troyan
and Boyan," in Proc. of the Amer. Philos. Soc, 56, 152-160 and
see below, s. v. kur. Vyazemski (Magnus, p. xlvii-xlviii) has already
connected this word with Salvonic bajati, bojati " speak, relate."
The meaning of bojan may therefore, be "singer"; cf. R. Gypsy
bagan,^'' "to sing," and note Slav, bajan, "enchanter," in Dan. 5, ii
(OR. version). It is possible, however, that the word may be of
Tatar origin (cf. Magnus, p. xlvii). Note that Turkish boj = " per-
son"; Mongol &o; = " clever archer" or "person"; Altai pajana =
" God " ; Cuvas pojan = " rich " and bajan is a tribal name of the
Altai. See below, s. v. bojarin. The word bajan appears also as a
proper name, Vajanos (but note the Greek v), son of Kubra.
Dubious words of this type are often the result of a compound

15 Cited Bang, Beitr'dge, p. 2>3-

16 Arranged in the order of the Russian alphabet.

1^ Cf. P. Istomin (Patkanoff), "The Gypsy Language (in Russian),"
Moscow, 1900.

PRINCE— TATAR MATERIAL IN OLD RUSSIAN. 81

derivation, possibly originally Tatar with a superimposition of a
later Slavonic folks-etymology, based on resemblance of sound (see
^. z'. huj-tiir, jar-tur, below).

Bojarin (Igor, passim) ; the common OS. word for "magnate"
( Berneker, p. 72), usually employed for Slavonic boj, "fight," fol-
lowing the idea that the boyars were essentially warriors. It may
however be connected as a loan-word with the above mentioned
Turkish baj-, boj- " rich." since the probably cognate R. bariii,
" gentleman " does not seem to be from a Slavonic stem boj-,
" fight." The words bariii, bojarin, therefore, are possibly Tatar.
In OS. and Bulgarian, boljarin, the / is probably due to the influ-
ence of the Slavonic bolj-, "great." See the Tatar material cited
above. .?. v. Bojan.

Biij-fitr (Igor, 80) varies with bitj (Berneker, p. 98) and is an
epithet of Prince Vsevolod. Here again is a word of possible double
etymology. The Slavonic elements appear to be buj, " bull," and
tur, also "bull," meaning "aurochs" in modern R. A similar
popular combination is buj-vol, " buffalo." from buj, " bull " and
vol, " ox."

The buj-ionw is apparently cognate with Greek, ^vui "to sprout,
be born " ; cf . Rumanian buiac, " lustful." The word buj alone
appears in Igor, 465; Buj Rjurice "O hero (bull) Rurik " ; the
genitive is bii-j-cgo. This &My-element can have no connection with
C. boga, biiga, Turkish bugha, Cagatai buka, etc.

All through the Tatar idioms we find variants of the word
bahadiir, " noble, mighty," now a common word in Hindustani dia-
lects borrowed through the Mogul (Mongol) ; cf. C. bahadur,
Mongol batur, Manchu baturu, Nogait matur, beautiful, Kazanj
mater, etc.

Note CC. 145: bahadur soi dcgclim, " te potentem esse dica-
mus" ; CC. 116: bagaf^" probus." In spite of Magnus, Igor, p.
51,1 believe that buj-tur is a Slavonic popular etymology from Tatar
bahadur, or its cognate ; cf. s. v. jar-tur, below.

Bus: busovi zrani "the crows of Bus" (Igor, 375), altered by
Magnus from bosuvi, but better = " steel gray crows," a variety
common in Russia to-day. Magnus, Igor, p. 50, associates it with

PROC. AMER. PHIL. SOC, VOL. LVIII, F, JULY II, IQIQ.

82 PRINCE— TATAR MATERIAL IN OLD RUSSIAN. .

Booz = Bluz (Balus), a leader of the Polovsty in 1054, during the
first invasion ; cf. pojut vremja hiisovo, "they sing the times of Bus"
(in this passage, plainly a proper name). The word husovi (Igor,
375) is more likely connected with hosy (Igor, 685) : bosym volkom,
" like a gray wolf," not " bare-footed " and hence " swift-footed "
(Magnus). For the idea of color, see C. huz^=" caernlens" (CC.
145); note also hoxag {=^hozag) "fiisciis" and cf. Turkish hoz,
" steel-gray," and Osmanli bus, " ice." These Slavonic expressions
are all certainly loan-words from Tatar bus, "blue; gray."

Bjes: djeti bjesovi (Igor, 186) ; translated by Magnus "children
of Baal," 7. e., "devilish children" (cf. Berneker, p. 56). Magnus
thinks bjes is a variant of bus-, but this is probably incorrect,
although it suggests the Cummanian bus, bos, seen in busov=^
" ruina" (CC. 195). The phrase bjesovi djeti must mean "chil-
dren of the devil," from the Slavonic stem bjes (bes) "rage."

Zencjug, "a collection of pearls" (Igor, 371) ; an older form
than the present zcmaig. Prof. Friedrich Hirth states that this is
an international word, known also in China; cf. Lithuanian loan-
word semczugas. This same stem is seen in Magyar gyongy and in
Osmanli inhi, pronounced indzi. Note that the change of / or / to
the palatalized dz is not unusual in Turkish; cf. C. ingcu (CC. 109),
Orkhonski Tatar jdnci'i, etc. This word does not appear in the non-
Russian Slavonic languages.

Kaninu (Igor, 225) : na Kaninu zelenu papolomu polstla "and
bedded in him in the Kanina with a garment " ; thus Magnus. Note
that papoloma = Greek ircTrXuifm . Magnus, p. 74. rejects a Tatar
derivation, but C. kan and Turkish ^an = " blood." This word
kanina is probably a hybrid adjective meaning "bloody" and the
phrase should be translated: "and bedded him on a blood-stained
green garment " ; viz., in the earth. I question as to whether
Kanina in Igor is a place-name =Kajala; cf. 229: .y toj-ze kajoly
ought probably to be read: 5 toj-ze kaniny "from that place of
blood." The hero's father-in-law ordered his body to be carried
to Kiev.

Kogan (Igor, 746): na kogana "against the Khan " = Tatar
kaghan (Orkhonski inscriptions; cf. Berneker, 468). This title

PRINCE— TATAR MATERIAL IN OLD RUSSIAN. 83

was given to Vladimir in Chron. 1171: kagan and kan in 1191.
The gutturalized khan is a later form; Greek x^-^V^ > X"-y°-^°^' ol^
Mongol and Avar khaganns (-z<i-=Latin ending), Osmanli khan
and C. han = " God." In Cumanian the h represented a guttural.
The Tunguz of Nercinsk say kan with hard k.

Koscej (Igor, 360) : v. sjedlo kdscievo "in a captive's saddle"
(not "slave's," with Magnus). Cf. Berneker 585. The word is
clearly a Tatar element from kos, " military camp," from which
comes R. kos, " camp of the Zaporozhian^^ Cossacks " ; hence, the
word used so often in Gogol's " Taras Bulba," koshevoj^" chief of
the Cossack camp." The word kosci must originally have meant
" prisoner, servant, groom." There can fee no connection here with
C. cue and ciicermen, " coerce," as some have suggested. In the
R. ballads koscej meant " magician, giant." It is possible that the
modern R. koscej, " skinflint, miser " may be the same word mis-
applied under the popular etymological influence of kostj, " bone."
The stem kos may be the same as that seen in Osmanli qawas ( ?)

Komonj, "horse" (Igor, passim) is probably not a Tatar word.
It has been connected with a supposed kobmonj, the same stem as
that seen in R. kobyla and English-Celtic " cob " = thick-set horse
(cob in Celtic = " tuft, abundance"), but Berneker, 555, rightly
rejects this as a doubtful derivation. The usual R. word for
" horse " is losadj, q. v., below.

Kur (Igor, 595) : doriskasc do kur Tjmutarakanja. There is
no reason to alter to cur with Magnus ^ " he raced to the precincts
of Tjmutarakanj." Magnus's emendation would refer to Cur, a
deity ( ?) of boundaries. The word kur is Tatar qur, "enclosure,"
with which kiirgan (see the following word) is probably connected.

Kurgan " tumulus, grave-mound," a common modern R. word
(Berneker, 648) appears also in Rumanian gorgan and is clearly
Tatar. Note C. gurgan, " burgh " and gurgatmen, " strengthen,"
and cf. Osmanli kurkhane. See the preceding word in this list.

Losadj, "horse" (not in Igor, which always uses komonj, q. v.).
The word losadj (Berneker 734) appears in Nestor's Chronicle, 1103

^8 The Zaporozhian Cossacks were the " Backfallsmen " of the Dnieper
who played so important a part in Polish mediaeval history (cf. Gogol's Taras
Bulba, etc.).

84. PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

and mi, used by Vladimir Monomakh in the council regarding the
Polovtsian expedition. The term was unknown to the Polovtsy
and was of southern Russian origin, passing into Russian, perhaps,
by way of the Viatici tribe (cf. Sakhmatov, Introd. to the History of
the Russian Language, I. 8i). The word appears in OR. as losa;
gen. losate (f-stem) and has had the form losadj since the thirteenth
century; cf. losdk, "mule," Pol. lossak, " horse,'' etc. It is unques-
tionably a Tatar loan-word; cf. Turkish alasa, "gelding," and
Magyar lo, " horse."

There were wild horses on the Asiatic steppes, as Vladimir
Monomakh speaks of catching and taming ten or twenty of them
at Cernigov.

Nogata (Igor, 460) : to byla by caga po nogafjc a kosccj po
rcsanje: "then a female slave would be worth twelve pence and a
groom for five pence."

This is a loan-word through the Tatar from the Arabic naqd,
"small coin." The intermediate form seems to have been nagd.
For the values in furs, one gr iz'j en j =^= twenty nogaty, or fifty
resany, see Magnus, p. 113. See below s. v. caga.

Ovhir (Igor, 675) is a proper name ; probably the same as Lavor
in Nestor's Chron. 1185. This appears to contain the same elements
as are seen in the Turkish oghlan, "servant, lad"; we have the
record in Nestor of the Tatar servants of David Igorevic, named
Oulan, Kolca, etc. The form Lavor is certainly not as correct as
Ovlur. The final r in both forms is difficult to explain, unless it is
a variant of the -n in oghlan, oulan.

Oljher (Igor, loi) is clearly not Tatar olybyr, "weak, ill" (re-
jected also by Magnus, p. 102). Magnus is probably right in
attributing this name to the series of geographical terms referring to
the Tatar territory, npw in Czecho-Slovakia. Note that there is a
Polish village Olbierzowice, not far from Warsaw (Magnus, lac.
cit.). Vexler derives this from Pol. olbrzym, "giant," applied to
the Avars. Cf. .y. v. scljbir.

Or'tama (Igor, 142 ; only once) : or'tmami i japoncicami koziikhi,
"with the mantles, cloaks and coats" (they bridged the mire, etc.).
This is plainly the same as C. ortma = " mantica" ; cf. art, "back,

PRINCE— TATAR MATERIAL IN OLD RUSSIAN. 85

top" (CC. 146), artarmen, "I excel" (CC. 54). In Osmanli,
drtJuck=^"to cover," and we find in Persian the noun ortme,
" covering " from a plainly Turkish formation which, however, does
not occur in modern Osmanli. This is undoubtedly our or'tma =
Osmanli i'lrtil, "covering."

Saltany (Igor, 489) : "thou shootest from the golden throne of
thy father the Saltany who are beyond Russia" (^sa semljami).
Every authority but Magnus regards this as the Arabo-Tatar saltan,
sultan, a reference to the chiefs of the Tatars. Magnus, however,
considers, that it alludes to the men of Salatyn on the lower Tatra
mountains in Hungary, whence came the barbarian auxiliaries of
Igor, such as the Topchaks. It must refer to an attack on the
Tatar foe, but the term saltan (sultan) is not commonly used to
denote the Tatar khans.

Tl'kovin (Igor, 369): pogauykh tl'kovin, "of the heathen
tl'koviny" ; perhaps the Ta\/xaT|oiof Ptolemy. The term is very diffi-
cult. It is usually rendered "nomads," from R. tolcdk, from tolkatj,
" roam," as the form occurs in Nestor, 907, alluding to the Varjags,.
Slovenes and Tivercy. A. Weseloffsky (ZDMG. 1877, p. 301)
refers the term to the Torki, the remnants of the Pecenegs. This
is not possible, since the Varjags (Norsemen) and Slovenes
(southern Slavs) were certainly not Torki. The derivation of the
word tl'koviny is uncertain. The proper pronunciation is tlkoviny
or tolkoviny, as the hard sign in Igor tl'koviny is a mere stop. Sakh-
matov thinks it means " bi-lingual," comparing it with tolinac " in-
terpreter," from tolkovdti {op. cit., p. 98).

Tjmutarakan (Igor, 384) was the last outpost of the Russo-
Hellenic influence and had heathen temples even in Strabo's day.
It was on the Taman peninsula, bordering on the Sea of Azov and
the Black Sea. Constantine Prophyrogenitus calls it Ta/narapxa. ,

Topcak (Igor, 432) alludes to the barbarian allies. Magnus
states " this word has an unmistakable Turanian form " and refers
to C. toprak, "corn" (CC. 208). In Osmanli topraq^=" soW, terri-
tory," and also " clay." It may refer to the nature of the soil of a
certain territory. Magnus identifies it geographically with Top-
czewo, a village in the province of Grodno, twenty versts from

86 PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

Bielsk, or with Topczykaly, seven miles from Grodno. There can
be little doubt that these people were Tatars.

Seljhiry (Igor, 432) may be cognate with Kalmuck silbyr, "long
whip," but the term seems to accord with the rest of the geograph-
ical series; cf. .?. v. Oljber, and Magnus, p. loi. It is probably
another reference to the barbarian allies of the Russians from the
Tatra. Note that Pol. szalhiers means " rogue."

Seresiry (Igor, 462; only once) :

ty by mozesi po sukhu Thou canst on dry land

(zivymi) s seresiry streljati shoot with bold seresiry

udalymi^^ syny Gljebovi the sons of Gljeb.

The sons of Gljeb were princes of Rjazanj. The passage is very
obscure and it is apparent that the copyist himself did not understand
it. Cf. Magnus, pp. 106 jf., for seven views. I believe that seresiry
must have been an implement. The Persian fire-hurling machines
were known as tir-car, an iron pipe filled with explosive powder
and employed very early in the East. Magnus, p. 107, suggests that
seresir may be cognate with Magyar seres, " worry, trouble," but
this seems improbable. Vexler suggests that the initial s may be a
scribal error for t, as the letters are not dissimilar in Cyrillic, but
this is not necessary, as a f palatalized before the f-vowel might be-
come s. The word seresiry suggests a Polovtsian word ciricar and
seems in this passage of Igor to be a synonym of the plamenny rog
"flaming horn," of Igor, 312; note also smaga, "fierce heat" (Igor,
311), a Little-Russian word. "Live seresiry" must mean "loaded
implements."

Kharalusny (Igor, 194) : meci kharaluznymi , "with steel
swords" (Berneker, 385; lOO) is undoubtedly connected with C.
karalic, " blackness," used for atramentum, " ink," in CC. 94, but
referring in Igor to the dark color of tempered steel. It is interest-
ing to note that in modern Osmanli, qarsilyq is used for the steel of
a flint-lock gun, but this really means " the opposite thing," i. e., the
thing opposite (qarsi) the flint." On the other hand, qarsilvq may

1* It is not necessary with Magnus to separate -mi from udaly and to
regard mi as the ist personal possessive "my brave sons of Gljeb"; udalymi
is instrumental plural agreeing with seresiry.

PRINCE— TATAR MATERIAL IN OLD RUSSIAN. 87

be a popular etymology containing an original q'am-stem = " black
steel (?)." Perhaps qarsilyq stands for qar-celik, "black steel," as
celik = " steel " in Osmanli. In the Russian ballads, bulatny means
the same as kharalu2ny = Turkish hulat, which is from Persian
pulad; thus, in Zadonscina; kopija kharalusnymi, meci hulatnyja,
topory legkie, "steel spears, steel swords (and) light battle-axes."
Note that k and q often become kh, especially in Azerbaijan and
Mariopol Tatar (cf. Blau, "Ueber Volksthum u. Sprache d. Ku-
manen," ZDMC. 29, 1876 [55^567]. PP- 569 ^■

Khinovy (Igor, 403). This original form Magnus has need-
lessly altered to khinju; khinovy is probably an adjective and means
simply " Hunnish " ( thus Sobolevsky, A. S. P., xxx, p. 474). It is
derived by Magnus from Tatar khan and taken to mean " belonging
to the Tatars" (khans), a theory based on the change of 0 to i in
Little Russian, seen, for example, in Little Russian pid for R. pod,
"under"; zvikno for okno, "window," etc. But this change of 0
to i is a very late phenomenon in the Ukraine. It is not likely that
this word has any connection with C. kinov, "crooked" (CC. 138),
kingir, "curved" (CC. 140).

Khorjugov (Igor, 146) ; cf. Berneker, 398. This word has been
derived from Mongol orongo " standard " and also from Gothic
hrugga " shaft," pronounced hrunga, which is not even identical in
meaning. The word occurs in Old Bulgarian khorangv, " pennant "
and in modern khorugv, " church banner," Pol. choragiew, etc. It
is more probable that this is a Tartar loan-word and not Gothic
hrugga which is the same word as English " rung " of a ladder.
The Mongol orongo may be a modification of an original khorongo.

Japoncica (Igor, 142) "Capuchin cloak" (Berneker, p. 445).
Magnus has wrongly japoncica (p. 115). This is identical with
OR. cpanca and Turkish iapanca, or iaponca; in Polish oponcza
means " rain-coat." Note Cagatai japonci " cloak." See s. v. or'tma,
below.

Jaruga (Igor, 92) "rill" (Berneker, 445) is clearly the same as
in Cagatai jarugh, "left, split." The ;ar-stem appears in OR. jar,
" cliff, ravine " ; Old Bulgarian jar " steep shore," Rumanian eriiga.
In modern R., we have jarug (Tula dialect), and eruk; jaruska
Little-Russian.

88 PRINCE— TATAR MATERIAL IN OLD RUSSIAN.

Jar-tur (Igor, 190; Magnus, p. 117; Berneker, 447). This is an
epithet applied to heroes ; conventionally = " fierce bull." See huj-
tur, above. There is a modern word jary " fierce, grim," which
probably does not belong in this connection. As biij-titr seems to be
a popular etymology of Slavonic elements suggested by a primitive
Tatar form, it is highly likely that the same is true of Jar-tur, with
which the Indo-Germanic elements jar, " fierce " and tur, " bull "
have been associated. Note that R. jary appears in a number of
Slavonic proper names, as Jarosldv; Jaromir, etc.

This Indo-Germanic jar is usually connected with Greek ^S)po<i,
" fiery " ; possibly it has the same stem as the Latin ira " wrath " ( ?).
According to Berneker, p. 448, this jar has no connection with the
Tatar iar, "light, bright," which occurs CC. 254. The question is
confusing, as jary, "bright" is also a Slavonic stem, R., etc. It is
conceivable that jar-tur might readily be a variant of Tatar iardur,
"he is (dur) splendid," a form which subsequently might have
been confounded with Slavonic jary (?).

As to the possible connection between Slavonic jary and Tatar
iar {iariklich = " lumen ," CC. 154; jarictc, " illuminavit," CC. 159;
iarkin, "splendor," CC. 193), this opens up the whole question as
to the primitive common origin of the Indo-Germanic and Ugro-
Turkic idioms which cannot be discussed here.

Caga (Igor, 460) "female slave" (Nestor, Chron. 1018) is
undoubtedly Tatar and should not be rendered "potentate"
(Magnus, p. 113). See above s. v. nogate.

THE ENERGY LOSS OF YOUNG WOMEN DURING THE
MUSCULAR ACTIVITY OF LIGHT HOUSEHOLD

WORK.

By FRANCIS G. BENEDICT and ALICE JOHNSON.
(Read April 24, 19 i9-)

In the computation of the food requirements of women and chil-
dren, certain factors have been applied to the standards for men —
factors which have long been established and which are based solely
upon differences in body-weight. Since the body-weight of the
average woman is but 80 per cent, of the body-weight of the average
man, it has been assumed that the food requirements of women are
in like proportion. The average energy requirement of the resting
woman has been found to be actually 17 per cent, less than that
of the average man under the same conditions of muscular repose.
Furthermore, studies in the Nutrition Laboratory upon the differ-
ences in the need for energy of quiet men and women have shown
that with complete muscular repose and absence of food in the
stomach, men on an average produce 6.2 per cent, more heat than
women of the same age, height, and weight.

But food requirements are made up of two factors : (a) the food
required for the maintenance of the quiet body doing no external
work; {h) the excess food needed for the accomplishment of the
various daily tasks. Dietary allotments, therefore, deal not only
with the basal requirements but also with the very considerable ex-
cess food required for the multitudinous daily tasks. The variety of
mechanical operations and muscular movements incidental thereto
which are engaged in by men is very great ; ultimately the energy
requirements for these different operations should be established.
It is popularly considered that the muscular activity of the average
woman is much less than that of the average man. This is probably
true, but as household work in some form or other still remains a

89

90 BENEDICT AND JOHNSON— ENERGY LOSS OF

not inconsiderable factor in the muscular activity of most women,
exact information as to the energy needs for the performance of the
domestic duties of the household is essential for an intelligent com-
putation of the daily requirements of the average woman. These
domestic activities consist for the most part of sitting, standing,
walking, sewing, sweeping, dusting, washing clothes and dishes, iron-
ing, cooking, bed making, etc. What, then, are the requirements for
energy above the basal for these and other activities?

The Nutrition Laboratory, in establishing its long series of basal
values for new-born infants, children, and adults, has adhered to the
time-consuming method of studying individuals, and a decade has
been needed to secure this basal information. In the belief that with
definitely established basal values, the excess energy required for
muscular activity could be studied to greater advantage with groups,
and thus general average values be obtained more rapidly and pos-
sibly with a greater degree of accuracy, a respiration chamber of
sufficient size to seat 30 to 40 persons was constructed and its ac-
curacy tested.^

This respiration chamber, 5.18 meters long, 3.81 meters wide,
and 2.29 meters high, is well ventilated by forcing outdoor air in
at one end and withdrawing the chamber air at the other. The
special feature of this apparatus is the sampling device. A rotary
air impeller discharges air from the chamber into a copper box or
wind chest having three circular openings in it, two of these being
10 millimeters in diameter and one 29 millimeters in diameter. The
amount of air leaving the wind chest through these three openings
is roughly proportional to the area of each opening and is obviously
alike for the lo-millimeter openings. Over each of these two open-
ings is placed a can covered with a rubber bathing cap to enclose the
discharged air. By means of a supplementary blower, air is with-
drawn from each of the sampling cans at exactly the rate at which
it is delivered into the can. In actual practice, therefore, the air
enters the cans at atmospheric pressure, is immediately withdrawn,
forced through containers holding sulphuric acid and soda lime,
respectively, and then discharged. By weighing the soda-lime con-

1 For a detailed description of this chamber, see Benedict, Miles, Roth,
and Smith, Carnegie Inst. Wash. Pub. No. 280, 1919, pp. 92 to 119, inclusive.

YOUNG WOMEN DOING LIGHT HOUSEHOLD WORK. 91

tainer, the carbon dioxide absorbed from the air passing through
the lo-millimeter openings may be quantitatively determined.

The exact proportion of air passing through the lO-milHmeter
openings was found by admitting a known weight of carbon dioxide
to the chamber from a steel bottle of the liquefied gas, and determin-
ing the amount of this gas which passed through each opening. In
the present apparatus this happens to be exactly lo per cent, of the
total, L e., 10 per cent, of the air is delivered through each of the
lo-millimeter openings and 80 per cent, through the 29-millimeter
aperture. The weight of carbon dioxide absorbed in the purifying
vessel, corrected for the amount of carbon dioxide contained in the
incoming air (a correction obtained by a simple calculation), is
therefore exactly one tenth of the total amount produced in the
chamber during the experiment, provided no change has taken place
in the percentage of carbon dioxide residual in the chamber air.
Analyses of the residual air, made with a Haldane apparatus at the
beginning and end of the period, give this correction, and the true
carbon-dioxide production is thus rapidly determined.

With this respiration chamber, experiments with periods no
longer than twenty minutes are perfectly practicable. The de-
termination of the carbon-dioxide production has of itself but little
value, but is used for computing the heat production by means of the
calorific value of carbon dioxide at an assumed respiratory quotient.
Such a method is admittedly not so accurate as the direct measure-
ment of the heat production, or the computation of the heat produc-
tion from measurements of the oxygen consumption. It is, how-
ever, a rapid method and not too costly for determining the approxi-
mate heat output of a group of people.

Through the personal interest of Professor Alice F. Blood,
several groups of Simmons College undergraduates were interested
in the research and volunteered to act as subjects for this series of
experiments. The studies were made with over two hundred
women. In this research the standard values used for comparison
were not obtained under the usual experimental conditions for de-
termining basal metabolism, that is, twelve hours after the last meal
and with the subject lying down in complete muscular repose, but to

92 BENEDICT AND JOHNSON— ENERGY LOSS OF

give practical living conditions, they were determined two hours
after a very light breakfast and with the degree of repose possible
with the subjects sitting quietly, reading. Hence, this standard
value or base line should not be confused with true basal minimum
metabolism measurements. From past experience, the standard
value here used may be estimated to be not far from lo per cent,
higher than the true basal or minimum value.

All of our experiments were made with one or more periods at
the beginning, in which the standard value was established for the
day under the conditions outlined. Observations were then made
of the increment due to reading aloud, singing, sewing, standing
quietly, sweeping, dusting, standing up and immediately sitting down,
and with the subjects marching in single file around the chamber.
These observations are now being extended so as to include many
more household activities. The ages, heights, and body-weights of
all of the subjects have been recorded. The calculations of the
metabolism have been made on the basis of uniform body-weight, as
Harris- has recently shown that the metabolism per kilogram of
body-weight is, with homogeneous material, a satisfactory method
of comparison.

Since every experiment contained periods in which the standard
metabolism was determined for use as a basal value, a considerable
amount of data is available showing the metabolism of the subjects
while they were sitting quietly, reading. The results of these quiet
periods are given in Table I. as heat output per kilogram of body-
weight per hour. In computing these values, we have assumed a
respiratory quotient of 0.90, since all of the experiments were made
two hours after a light breakfast. The calorific value of carbon
dioxide at this respiratory quotient has therefore been used in all of
our calculations, namely, 2.785 calories per gram of carbon dioxide.
From this series of experiments on twelve different days, embody-
ing twenty-three periods, in which over two hundred women i)ar-
ticipated, it is seen that the average heat production per kilogram
per hour was 1.12 calories. This average figure of 1.12 calories has
a specific interest in that it indicates the probable heat production of

2 Harris and Benedict, Carnegie Inst. Wash. Pub. No. 279, 1919, p. 160.

YOUNG WOMEN DOING LIGHT HOUSEHOLD WORK.

93

women sitting quietly under ordinary living conditions with a
moderate amount of food in the stomach.

TABLE I.

Resting Metabolism of Groups of Young Women.
(Subjects sitting quietly reading, 2 hours after light breakfast.)

Date.

Number of Women,

Number of Periods.

Heat Output per
Kgm. per Hour.

1917
January 15

20

April 7

14

21

23

30

May 5

1918

March 16

April 6

April 27

May 4

Average

20
14
15
IS

23

18

19

25
15

IS

cals.
1. 16
1.07
1.05
1. 12
1. 1 2
1.06
1.05

1. 11
1. 19
I-I5
113
1.06
1.07

1-23

1. 12

1.20
1. 21
1. 18
1. 18
1. 21
I-I3

1. 00
0.97

The percentage increments in the metabolism due to the various
household activities studied are given in Table II. These are
arranged in the order of increment in energy requirement.
Although perhaps of minor practical importance, it is of physio-
logical interest to note the metabolism necessary for reading aloud.
In these experiments the students read in unison either from a
standard college text book or from the book of Psalms. Three ex-
periments wath twenty-three to twenty-five women, with a total of
eight periods, showed increases of 3, i, and 5 per cent., respectively,
with an average of 3 per cent. Frankly, this was rather a surpris-

94

BENEDICT AND JOHNSON— ENERGY LOSS OF

ing observation, for it would be expected that the effort of reading
aloud would be measurably greater than here indicated.

The slight increase in metabolism due to reading aloud is of
even greater interest when compared to the large increments noted
when the subjects were singing. In these experiments the subjects
sang from a book of college songs or a standard hymnal used in
chapel. In the three experiments with fourteen to twenty women,
covering in all four periods, the rather irregular percentages of 17,
34, and 16 were obtained, with an average of 22 per cent., this being

TABLE IL

Percentage Increase in Metabolism of Women Due to Light Muscular

Activity.

Occupation.

Number of
Women.

Number
of Experi-
ments.

Number of
Periods.

Increment.

In Indi-
vidual Ex-
periments.

Average.

Reading aloud . .
Standing quietly
Hemming

Singing

Dusting

Sweeping

23 to 25

3

20

I

15

2

14 to 20

3

15 to 19

3

15

2

p. ct
3
I
5
9

16

10

17

34

16
126
121
156
139
161

p. ct.

3

9

13

22

134
150

more than seven times greater than that noted when the subjects
were reading aloud.

To determine the increment in the metabolism due to the labor
of plain sewing (henmiing), two experiments with fifteen women,
and covering two periods, gave 16 and 10 per cent., respectively, with
an average of 13 per cent. Since in this sitting occupation, the
bodies of the subjects remained relatively quiet and the arms alone
were moved, it is perhaps not surprising that the increment was no
greater.

YOUNG WOMEN DOING LIGHT HOUSEHOLD WORK. 95

To throw some light upon the additional energy requirements
necessary for maintaining the standing position, one experiment
with twenty women with one period was made in which it was found
that the extra metabolism required for standing was 9 per cent,
above the standard sitting resting metabolism. This agrees reason-
ably well with values found in this laboratory for subjects in the
post-absorptive condition for the differences between sitting and
standing.

As a typical household occupation, involving moderate muscular
exercise, the metabolism during sweeping was studied in two ex-
periments with fifteen women and covering in all four periods.
The increments noted were 139 and 161 per cent., respectively, with
an average of 150 per cent. This is somewhat less than the increase
due to moderate walking found in the Nutrition Laboratory, which
was approximately 250 per cent, above the metabolism with the sub-
jects in the standing position.

Another household occupation studied, involving rather moderate
muscular effort, was that of dusting. In these experiments the sub-
jects were required to dust chairs, wiping the rounds and seat, and
at a given signal change to the next chair in rotation. One minute
was allowed for dusting each chair. Three experiments with four
periods with fifteen and nineteen women gave increments above the
basal value of 126, 121, and 156 per cent., respectively, with an
average of 134 per cent., an increase only slightly less than that
noted in the sweeping experiments, namely, 150 per cent.

The increment in energy with two other muscular activities was
studied in this series of experiments but the results are not included
in Table II. The muscular work of standing up and sitting down
enters into a relatively large number of the daily activities and a
special study of this operation was made in that the subjects while
sitting quietly in a chair were, on a given signal, required to stand
up, remain standing one second, and then immediately seat them-
selves. This routine was carried out approximately once every
minute during the period. In this particular series, it seemed best to
compute the increments not as a percentage above basal but as the
actual increase in energy for standing up and sitting down per move-

96 BENEDICT AND JOHNSON— ENERGY LOSS OF

ment and per individual. Two experiments with eighteen and
twenty women, covering a total of three periods, gave values of 0.31
and 0.32 calorie, respectively, averaging 0.32 calorie. It thus ap-
pears that with young women an energy expenditure of approxi-
ately one third calorie is required for the activity of standing up and
sitting down.

On one date, April 6, 191 8, a group of twenty-five young women
marched slowly about the respiration chamber for twenty-five
minutes. The distance walked in going around the chamber once
was forty-five feet ; this was accomplished about fifty-three times
in twenty-five minutes, the exact total distance walked being 2,356
feet. On that particular day the standard value found and used
for the base line was 1.20 calories per kilogram per hour. During
the walking the heat output was 2.44 calories per kilogram per hour,
or an increment over the standard value of 1.24 calories per kilogram
per hour. The rate of walking was 1.08 miles an hour and the
average weight of the subjects fifty-fovir kilograms, from which
it is readily computed that at this very slow rate of walking the
energy above basal required for transporting the body was sixty-two
calories per mile. While very few of the experiments on walking
made in the Nutrition Laboratory employed so slow a rate of walk-
ing as this, many of our observations with slightly faster rates give
values from 40 to 60 calories per mile although usually with men,
with consequently somewhat higher body-weights. Too little evi-
dence is thus available to indicate whether or not walking at so
slow a rate is performed on a distinctly uneconomical basis. The
value is not far from that commonly quoted from German sources
for the energy required in walking one mile at a moderate rate.

The method used in this research seems to be well founded and
applicable to such study. An extension of these observations is
planned to include not only women and children, but also men en-
gaged in various activities, these studies forming a part of the
general study of muscular work now in progress at the Nutrition
Laboratory.

Nutrition Laboratory,

Carnegie Institution of Washington,
Boston.

MAGELLANIC PREMIUM

Founded in 1786 by John Hyacinth de Magellan, of London

I9I9

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful K^jowledgk

ANNOUNCES THAT IN

DECEMBER, 1919

IT Wri.L AWARD ITS

MAGELLANIC GOLD MEDAL

TO THE AUTHOR OF THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING TO NAVIGATION, ASTRONOMY, OR NATURAL PHILOSOPHY (mERH
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 1919, deliverfree of postage
or other charges, his discoverj', invention or improvement, addressed to the
President of the American Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance by some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the same motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author.

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publicly
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as may be after the adjudication, either in a separate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

Ml correspondence in relation hereto should be addressed

To THE Secretaries of the

AMERICAN PHILOSOPHICAL SOCIETY

No. 104 South Fifth Street

PHILADELPHIA, U. S. A,

Proceedings

OF THE

American Philosophical Society

Subscription — Three Dollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

"Pricey One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

For Promoting Useful Knowledge

New Series, Vol. XXII,

Part III, 4to, 44 pages.

{Lately Published)

Tertiary Vertebrate Faunas of the North Coalinga Region of Cali-
fornia. A Contribution to the Study of the Palseontologic
Correlation in the Great Basin and Pacific Coast
Provinces. ByJoHN C. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California.

Subscription— Five Dollars per Volume
Separate parts are not sold

(Address The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A.

ijji/fr PROCEEDINGS

OF THE

American Philosophical Society

HELD AT PHILA^t/EI^MiA

FOR PROMOTING USEFUL KNOWLEDGE
Vol. LVIII. 1919. No. 2.

CONTENTS

PAGE

Alternating-current Planevector Potentiometer Measurements at Telephonic

Frequencies. By A. E. Kennelly and Edy Velander . 97

Some Scientific Aspects of the Meteoi'ological Work of the United States

Army. By Robert A. Millikan . . . . . • 133

Detecting Ocean Currents by Observing their Hydrogen-ion Concentra-
tion. By Alfred Goldsborough Mayor .... 150

PHILADELPHIA

THE AMERICAN PHILOSOPHICAL SOCIETY

104 South Fifth Street

1919

The American Philosophical Society

announce? that ai^ award will be made in the year 192 1 of the

tiENRY M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

The Control of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent, of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after, the
award.

Attention is called to the following regulations governing the award
of the Prize:

Competitors for the prize shall affix to their essaj^s some motto or name
(not the proper name of the author, however) and when the essay is forwarded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judges, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essaj-s shall be in their author, subject to
the right of the Society to publish the crowned essay in its " Transactions "
or " Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,
David Jayne Hill,
Simeon E. Baldwin,
John Cadwalader,
W. W. Keen,
William B. Scott,
President ex-officio.

Committee.
The essays must be sent, addressed to the
President of the American Philosophical
Society, No. 104 South Fifth Street, Phila-
delphia, U. S. A.

ALTERNATING-CURRENT PLANEVECTOR POTENTI-
OMETER MEASUREMENTS AT TELEPHONIC
FREQUENCIES.

By a. E. KENNELLY and EDY VELANDER.
(Read April 26, 1919.)

History of the Apparatus.

The measurements presented in this paper have been made in the
electrical engineering laboratories of the Massachusetts Institute of
Technology, with a new form of a.— c. planevector potentiometer,
giving its potential readings in rectangular coordinates. This in-
trument lends itself readily to the measurement of vector potential
differences up to frequencies of at least 2,000^-'. The principle of
the potentiometer is the same as that described by Professor Larsen^
in 1910; but the present form of the instrument has been worked
out in the M. I. T. laboratories, mainly through the thesis studies of
Mr. A. E. Hanson.2

The essential connections of the instrument and its method of
application are outlined in Fig. i. The anti-inductive resistance ah
is connected in series with the primary winding he, of a toroidal non-
ferric induction coil included in the main potentiometer circuit P.
The secondary winding hd, of the induction coil, is connected at h
with the junction of the primary coil and the resistance. The sec-
ondary winding of the induction coil, and the resistance ah, are both
provided with suitable taps, which are brought out to dials, for ad-
justment of the potential connections at k and /. When a sinusoidal
alternating current Ip r.m.s. amperes Z , from the oscillator source
O, at a suitable impressed frequency, passes through the potentiom-
eter circuit, the alternating p.d. between the taps k and / will be

Ep=Ip(R-^jX), r.m.s. volts Z (i)

where R is the resistance included between k and h in ohms, while

1 Bibliography 9.

2 Bibliography 13.

PROC. AMER. PHIL. SOC. VOL. LVIII, G, JULY 22, I919.

97

98

KENNELLY AND VELANDER— POTENTIOMETER

jX = jMw is the mutual reactance included between / and b. M
denotes the mutual inductance between the primary winding and the
part bl of the secondary, and oj = 27r/, is the angular velocity (in

Fig. I. Simplified diagram of connections, showing the rectangular poten-
tiometer P arranged for exploration of the potential distribution over the
working circuit IV.

radians per second) of the impressed frequency /, in cycles per
second.

The oscillator also supplies current to an associated working cir-
cuit W, containing the apparatus in which the distribution of plane-
vector^ a-c. potential is to be explored. A tuned vibration gal-
vanometer, VG, enables a potential balance to be obtained between
a selected pair of terminals on the working circuit, and the adjust-
able taps on the potentiometer circuit, so as to secure the relation

Ua

■■Ip(R-\-jX), r.m.s. volts Z (2)

whereby the p.d. at the terminals, say A and B, can be evaluated in
terms of the calibrated constants R and X of the instrument, and
the measured potentiometer current Ip, taken at standard phase. If,
however, another p.d., say U ^c, for instance, across a standard re-

3 A planevector may be defined as a geometrically directed complex
quantity in a plane of reference, and subject to the laws of complex arith-
metic, as distinguished from a vector which is subject to the laws of vector
arithmetic. In this paper the term "vector" is used as an abbreviation of
" planevector."

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 99

sistance r ohms, as indicated at BC, in Fig. i, is measured imme-
diately after U ^b, we have

UBc = Ip{Ro + i^^), r.m.s. volts Z (3)

Then

Uab^R±JX_
Ubc Ro+jXo

numeric Z (4)

In this way, the relative numerical values of the p.d.'s on different
parts of the working circuit can be evaluated, preferably in terms of
the p.d. on a standard resistance, without requiring a measurement
of the potentiometer current I p.

As thus far described, the instrument can only measure vector
p.d.'s confined to a single quadrant. By means of reversing switches,
however, on X and on the entire potentiometer, all four quadrants
in the voltage complex plane can be explored.

A more detailed description of the instrument, and of its mode
of operation, technique and limitations, appears in a paper by the
writers published* elsewhere ; so that only a brief outline of this part
of the subject will here be necessary, in order to present more fully
in this paper some of the results which have been secured with the
instrument, at telephonic frequencies, in the laboratory.

Simple Series Circuit of Resistance, Inductance and Capaci-
tance, AT Constant Frequency and Varied Capacitance.

In Fig. 2 the connections are shown of a simple working circuit
containing an inductance L of 0.8106 henry, a total resistance R of
3183 ohms, and an adjustable condenser C. Such a circuit may
briefly be described as a RLC circuit. The connections pp' to the
potentiometer are also indicated. The inductance L was very loosely
coupled with the Vreeland oscillator Osc, and the oscillator fre-
quency was maintained at Soo^-'.

With the condenser C shorted ; i.e., made infinite, the poten-
tiometer reading on 100 ohms in r was the vector OA, as obtained
from rectangular coordinates in R and — jX, viz. — 22.5 — /176 mil-
livolts. This is indicated as a vector of 179.4 millivolts, at a slope

* Bibliography 14.

100

KENNELLY AND VELANDER— POTENTIOMETER

of — 97°. 2 with respect to the reference axis OR, which coincides
with the phase of the current I p in the potentiometer circuit.

^

- ^•^'* "^^^ Osoc^/^ '^'^ a/

Fig. 2. Current locus, under constant impressed e.m.f. of 500 — ', of a
simple circuit containing inductance, resistance, and capacitance, the latter
being varied from 0.03 m/. to ^.

As the capacitance in the condenser C was reduced, the vector
voltages at the terminals pp' advanced along the locus APBDO,
which is seen to be a circle passing through the origin of coordi-
nates R and /X.

It is easy to demonstrate that a circular locus must apply to the
current and therefore to the potential difference at pp' . The im-

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 101

pedance in the working circuit of Fig. 2 is

Z=R + j

(-^)

ohms Z (5)

If we keep r, L and w constant, but vary C, the impedance will ad-
vance vectorially along a straight line parallel to the / axis, or react-
ance axis. This is shown in Fig. 3, where distances along the hori-
zontal real axis represent values of resistance R; while distances
along the ; axis represent values of reactance. As the value of C

Fig. 3. Planevector impedance locus of a simple RLC circuit in which the
capacitance is varied.

was changed from infinity to 0.03 X lO"*' farad, the vector value of
impedance moves along the straight line abd e. The impedance Z

102 KENNELLY AND VELANDER— POTENTIOMETER

therefore pursues a straight-line locus. The current in the work-
ing circuit will be

/„' = 7y amperes Z (6)

where \E-^y\ is the constant size of the induced e.m.f. in that circuit,
and /?° is the angle of lag of that e.m.f. behind the current in the
potentiometer circuit, considered as of reference phase. This cir-
cuit happened, in this case, to be strongly condensive, owing to the
adjustment of the condenser C in Fig. i. The angle of lag /? is the
angle DOP in Fig. 2. In equation (6) the vector impedance Z
appears as a reciprocal, or in the denominator. It is well known
that the reciprocal of any vector straight-line locus is a circular
locus passing through the origin. The diameter of the circle through
the origin is also the reciprocal of the perpendicular distance from
the impedance origin to the impedance locus. Consequently, in
Fig. 2, with \E^y\ =y.2y2) volts, the diameter OP, expressed in am-
peres, will be 7.273X1/3181^0.002285 ampere, or 2.285 miHi-
amperes, lagging ^° behind the phase of the potentiometer current
OR. The measured p.d. across 100 ohms in the resistance R will
be 0.2285 volt, also lagging ^° .

In reducing the capacitance of C from infinity to zero, the cir-
cular locus of potential across pp' has nearly covered three quarters
of its entire circle. It is evident from a consideration of Fig. 3, that
if, with the condenser shorted, additional inductance could have been
inserted in the circuit, while retaining all other conditions constant,
the remainder AO of the circular locus might also have been
traversed.

The use of the new potentiometer thus enables this circular
potential and current locus, at telephonic frequency, to be demons
strated observationally, instead of remaining on a purely abstract
mathematical basis.

Simple RLC Circuit of Sharper Resonance.

Another example of varying the series capacitance in a simple
a.-c. circuit, but at a frequency of 201O'— ', is given in Fig. 4. Here
almost the entire circular locus of potential and current is covered

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 103

by the observations, which, shown as round dots, lie close to the
circle OABD, except for an accidental wide unobserved gap between
0.019 and 0.022 microfarad. In this case, the resistance of the cir-
cuit was 250 ohms, and this relatively small resistance involves a
relatively large change in the vector impedance near resonance, com-

FiG. 4. Current-locus for conditions similar to those of Fig. 2, but ap-
plying to a circuit of smaller resistance, and to an impressed frequency

of 2010'—'.

puted at 0.0214 microfarad, with a small change of capacitance. It
may be noted that in this case a change of 250 ohms in the condenser
reactance changes the vector current, from resonance, through 45°
difference of phase, or over a quarter of the circular locus. As the
resistance R in the circuit is reduced, a correspondingly smaller
change in reactance, from the resonant point, will produce this 45°
change in the vector current. An a.-c. potentiometer can thus
serve to measure a small capacitance with precision, by observing

104 KENNELLY AND VELANDER— POTENTIOMETER

the vector change of current thereby produced in a sharply tuned
resonant circuit of low resistance, when the condenser is inserted or
removed. Other applications of the same principle will suggest
themselves.

Simple RLC Circuit with Variations Made First in R Alone,

■ AND THEN IN C AlONE.

An example of a simple RLC circuit, with successive variations
in resistance and capacitance, is presented in Fig. 5. Here the total
resistance in the circuit could be given different assigned values be-
tween 52 and 802 ohms. The capacitance could also be varied be-
tween infinity and 0.20 microfarad. The p.d.'s were measured
across taps pp', Figs. 2 or 4, and the current strengths determined
therefrom. The constant inductance in the circuit was o.i henry.
The frequency was 1006'—' throughout. This produces resonance in
the circuit at C = o.2^ microfarad.

If we plot the impedance of a RLC circuit, like that of Fig. 4, at
any constant frequency, we obtain an impedance diagram of the
type represented in Fig. 6. If we maintain constant resistance in
the circuit, such as r4 = 402 ohms, and vary only the capacitance,
the impedance locus will lie along the straight line AB. Such a
variation of impedance, substituted in equation (6), will, as we have
already seen, give rise to a circular current locus, the diameter of
the circle coinciding with the vector E of impressed e.m.f. If,
however, we keep the reactance constant, at some value such as
A\ = j 216.2 ohms, Fig. 6, and vary the resistance, the vector im-
pedance will move over the straight-line locus DE. This vector
impedance, inserted in (6), will also give rise to a circular locus;
but the diameter of this circle will be in quadrature with the vector
of impressed e.m.f. In the particular case of resonance, where the
reactance is kept at zero, and the resistance is varied, the current
locus will be a circle of infinite radius ; i.e., a straight line, coincid-
ing with the vector of impressed e.m.f. These results are brought
out experimentally in Fig. 5. It will be observed that with a capaci-
tance of 0.25 microfarad, the locus coincides with the straight line
OR. The vector impressed e.m.f. was so adjusted, by preliminary
tuning of the potentiometer circuit, as to coincide with this line.

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 105

All of the circles of constant resistance have their centers on the
line OR, or axis of reals.

i^o

}^

^

^

<^

~/

1 /r"^

D^

C'O.zsr/

mj

//..-.^,v.

^

15 „g

0 '

yi

K^^S^-^

x(

^

\

i

V5

-^10

7^5

Fig. 5. Graphs of vector current or vector admittance for a RLC circuit, in
which either i? or C is varied alone.

In Fig. 5, the full lines are the theoretical loci and the small cir-
cles the actual potentiometer readings. Fig. 6 is obtained by inver-

106

KENNELLY AND VELANDER— POTENTIOMETER

sion from Fig. 5, and shows the vector loci of impedance in the
circuit,, the full lines being computed. The small circles indicate
the experimentally deduced results.

4 rA-rnj

;^

/ '°°

/-^os

^

.

? ^^C—

9

>rr

'

,

-■ 0

1 . •—

yC^ksof^r ,

X* i

)
1

!

"i

1

/■ ^^

r

ft:

(t

ft:

1

^C'Oia ^f

X,

.C- 0 2t -

X,

£

.<- 0x6 ■

X,

,C- OU

"'

» 0

0 ^

0 MO 4

0 j<

« X,, 6i

'°

'"1

>z^-

Xi

-^ KB

J. 22/./

x>

'"<:-(

>20//

^_oo

v

A

/^oc

''

Z

* 1

Fig. 6. Graphs of vector impedance for a 7?LC circuit, in which either i? or C

is varied alone.

Divided Circuit with Inductance and Capacitance in

Parallel.

If we take a simple RLC circuit, such as that shown in Figs. 2,
4 and 5, and shunt some part of the fixed inductance with an adjust-
able condenser, we obtain the connection diagram of Fig. ya, where
the inductance coil /, separate from the oscillator secondary, is
shunted by the variable condenser c. Commencing with c = o, or
the condenser removed, the vector current in the circuit, as deduced
from the p.d. across the 10 ohm taps-/'/'', was the vector 01 = 3-6

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 107

_|-yi3.6= 14.07 Z 75°.2 milliamperes, at 525^, Fig. yh. As the
capacitance in the condenser c, Fig. "ja, is increased from o to in-
finity, the current locus goes over the circular path mlpk, Fig. yh.

^/6 chm

Osc.7?.

3

mdr^

r

p f

Fig. Ta. Circuit of inductance, resistance, and. capacitance, where one part of
the inductance is bridged by a variable condenser c.

Fig. 7&. Locus for the main current in the circuit of Fig. ^a, the con-
denser c being varied from o to i^, and the impressed e.m.f. being maintained
constant, at 525'—.

The phase of the e.m.f. induced in the coil L, Fig. ya, is represented
by the vector Ok, Fig. yh.

108

KENNELLY AND VELANDER— POTENTIOMETER

The circular graph of Fig. yh is a current locus to OR as stand-
ard current phase. The same graph may, however, be regarded as

,

y-Ma

A "°

09

c

■N

osjj. °'"f

/

V

f/o

\jM^. i^^e-^

\\ '^ e.

^s^-^r \

o.

jfU^

(

^^X-

^4 '

\^£e

Tna.

-2-

•V -/

' ^tfS^^-^'^i^

'JV'"^-«L jp \ ' <-

0.3

^^

^^^^

A'7rf

~d^'^

\

w

/

\

^

^ /

'I.

^

-■ '^^mI

'

i^o

'}JO

Fig. jc. Vector circular loci of main and branch currents in the circuit of
Fig. 7a, as the condenser c is varied.

an admittance locus to Ok as standard admittance phase, if we
change the scale of coordinates accordingly, since

Y =

ohms Z (7)

and E is the e.m.f. in the circuit, 3. 744 Z I9°.4 volts, which was kept
constant both in size and in slope during the observations. If, re-
garding Fig. yb as an admittance graph to the proper scale, we take
the reciprocals of successive vector values, we obtain therefrom an
impedance graph which must also be circular, according to the theory
of geometrical inversion. Fig. 8 shows the vector impedance graph
so obtained. It represents the vector impedance in the main circuit
of Fig. ya when the condenser c is varied. At c^ 00, the vector

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 109

impedance is 0^ = 464.7 \- 74°. 25. At c = o, the vector impe-
dance is 0L = 266.2 ^ 55°.6 ohms. The angle mol in Fig. yb is
the same (19°) as the angle MOL in Fig. 8. As c is increased from
c-^o to c=co, the vector impedance pursues the wide circle
LPGM.

Fig. 8. Locus diagram obtained by inversion from Fig. 7b, and showing the
variation of total main-circuit impedance of the main-circuit of Fig. 7a.

The explanations for these interesting circular graphs of Figs.
yb and 8, appear step by step in Fig. 9. Fig. ga shows the vector
admittance of the Ic combination in Fig. ya, as taken between the
points qq' in the main circuit. 01, Fig. ga, is the fixed admittance
of the coil /, Fig. ya, at the constant impressed frequency of 525-^' ;
namely 0.4656 — 74.343 = 4.369 "\ 83°. 53' millimho. The admit-
tance of the condenser c will be jcm. The vector sum of 01 and ;Vw
will evidently lie on the straight line Int, as c varies from o to co.
The corresponding impedance of the Ic combination, between points
qq' in the main circuit, Fig. ya, will be the vector reciprocal of the
straight line locus Im, Fig. ga, and must therefore be a circle, as has

110

KENNELLY AND VELANDER— POTENTIOMETER

been already pointed out. The impedance of the combination will
vary over the circle LUSM, Fig. gb, from OL with c = o to OM = o
with c=cc. The diametral maximum impedance OR or 2147

Zlf^tct,^

CmO-tl

i'-

\

0

Fy

♦

"■] ^

c-*>'

tA -

— ^r^^~~~

Fig. ga. Vector admit-
tance of / and c in par-
allel as c is varied from
zero to infinity.

Fig. gb. Vector impedance of / and c in par-
allel as c is varied from zero to infinity. Vector
inversion of Fig. ga.

ohms, occurs at the value of c (1.317^1/) at which On, Fig. ga, is
horizontal, and at which value the vector admittance in ga has mini-
mum size. At this value of c, the admittance of the Ic combination
is a minimum, its impedance a maximum, and the p.d. at terminals
qq' in phase with the main current. This condition occurs when
the susceptance of the condenser jcw is equal and opposite to the
susceptance of the coil. Thus

•-77,

y=

;'.<jo

lo

-J

r+jlui r''+l-oi- r--\-l-or ' r-^-l-d"

= g—jb mhos Z (8)

MEASUREMENTS AT TELEPHONIC FREQUENCIES. Ill

so that

b = TTT = o I 7o "o = o mhos, (q)

I 2- 1 r- + /-o)- r~

/oj + —

Ice

Consequently, at minimum admittance and maximum impedance,

ceo = , mhos (lo)

J,.

/co + -

10}

or

I

c = s farads, (ii)

'"'- + 1

It may be noted that the condition of maximum-impedance in the
Ic combination, and which has been defined as " parallel resonance,"
corresponding to zero total susceptance, differs from the condition
of simple series resonance within the Ic circuit, which occurs when
the total reactance is zero, or when

CO) = J- mhos. (12)

/co

This condition is found in the diagram, Fig. ga, at the total vector
admittance os, when the capacitance in c is 1.332 /xf. This occurs
when the angle sol. Fig. 9, is 90°, or when the angle iiso is equal to
the angle nol of the coil's admittance. The angle sou is 6°. 7', the
complement of mo/.

In Fig. gb, the impedance OS of the Ic combination is that which
is presented at series resonance. It has the value 2135 'v 6°. 7'
ohms. The angle SOU is the same as the angle sou. At the ca-
pacitance c= 1-317 /if-, the p.d. at qq', on the Ic combination, will
be in phase with the main current.

If we increase the impedances of Fig. gb by the fixed impedance
in the remainder of the circuit, Fig. ya, we obtain the total circuit
impedance as shown in Fig. gc, where OM' is the vector impedance
126 — y447.3 = 464.7 \ 74°. 16'. If we add to this fixed impedance,
OM', the vector circle LUS of Fig. 9b, we obtain the resultant vec-

112

KENNELLY AND VELANDER— POTENTIOMETER

tor locus M'K'H'U'Q' of Fig. gc. As the capacitance in c varies
from o to CO, the impedance in the main circuit varies from OL' to
OM', around the circular arc H'Q'.

The vector OP' =^206.7 \ 20°.26'.49" ohms, is the point of ap-
parent resonance or minimum impedance in the main circuit; i.e.,
maximum main current, as the result of adjusting condenser c. It
is not a true resonant point, since there is neither cophase, zero re-
sistance, zero susceptance nor equality of undamped frequencies.

The vectors OK' and OH' are two cophase vectors, at each of
which the current in the main circuit is in phase with the main im-
pressed e.m.f." In general, there will be either two such cophase
points, or none. A single cophase point may, however, present
itself as an intermediate case. Such cophase points are, in general,
not to be classed as resonance points.

The vector points, U' and S' , corresponding to U and 5 in Fig.
gh, and also to u and .j in Fig. ga, are only of secondary signification
in relation to the main-circuit impedance. That is to say, although
they correspond to local-circuit resonances in Ic, they do not, in
general, represent main circuit resonances.

Fig. gc. Total impedance of circuit includ-
ing / and c in parallel, as obtained from the
vector addition of OM to the locus of Fig. gb.

Fig. gd. Total admittance of circuit
as obtained bj- inversion from Fig. gc.

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 113

Finally, the vector OQ' represents the maximum impedance of
2357 'x 20°. 26'. 49" ohms. It is in cophase with the vector OP of
apparent resonance.

If we invert the impedance graph L'P'H'Q' with respect to the
origin O, by plotting reciprocals of the corresponding vectors, we
obtain the total admittance graph of Fig. 90?. This pursues the cir-
cular locus IpkJun, as c varies from o to x. This may also be re-
garded as a main current graph to a suitable scale of amperes, with
the voltage at reference phase OG. The vector admittance and cur-
rent will have a maximum at Op, the point of apparent resonance,
cophase points at k and h, minimum at q and local internal reso-
nances of the Ic loop at ^i and s.

Fig. yb shows the circular locus observations in the main current
corresponding to Fig. gd; while Fig. 8, deduced from yh, corre-
sponds to Fig. 9c. It is therefore brought out experimentally, with
the new potentiometer, what has perhaps hitherto been known only
in abstract theory, that the variation of pure reactance in a branch
circuit gives rise to circular-locus current variations in the main
circuit.

The experimental case, presented in Fig. 7, corresponds to that
of a radio receiving system, in which L and C correspond to the
antenna path to ground. The loop circuit Ic then corresponds to
an oscillation circuit conductively connected with the antenna, c
being tuned to bring about maximum antenna current.

Composition and Resolution of Vector Circular Loci.

The vector loci of current in the fixed inductance / and the con-
denser c, as the latter is adjusted from zero to infinity, are presented
in Fig. yc. These take the form of circles marked respectively /?
and Ic- The vector sum, /„,, or main current in the circuit LC of
Fig. ya, is reproduced from Fig. yb to e.m.f. standard phase. It
may be noted that in these three circles the maximum or diametral
values of current strength occur at dififerent condenser settings. Im
has its maximum near f ^0.5/xf., Ii near 0.7 /xf., and Ic near 0.9 /^f.

It has already been shown that the circular variation of impe-
dance in the condenser, by reason of its adjustment, gives rise to a

PROC. AMER. PHIL. SOC, VOL. LVUI, H, JULY 22, I9I9.

114 KENNELLY AND VELANDER— POTENTIOMETER

current of circular locus in the main circuit. The p.d. at the termi-
nals qq' must also have a circular locus, as it consists of the vector
difference between the constant impressed e.m.f. E, and a vector
drop of circular locus due to a circular-locus current in the constant
impedance of L and C. Consequently, the fixed branch / must re-
ceive a current of circular locus. The current in the variable con-
denser branch must also have a circular locus, as will be seen from
the following relations. The main current / is

E , ,

I = amperes Z (13)

= 7 — , 7 N I -7 amperes Z (14)

where g is the fixed admittance of the coil / or i/{r -\- jIm) in mhos
Z , and Z is the fixed impedance, in ohms Z , of the main EC circuit,
outside of the Ic combination.

This main current / has already been shown to have a circular
locus, in reference to Fig. 9, and the conclusion is confirmed alge-
braically from (14), by Mobius' theorem, as will be further dis-
cussed in the appendix.

The branch current in the condenser Ic is

jcoi Ejcoj E

Ie = I — 7-^ =7 — , ^ X 1 -y = n^ amperesZ (15)

g+JCo: (l+Zg)+jZcco .I+Zg

Z-J--

Ceo

In its last form, the expression for h indicates the inversion of a
straight line, which shows that h has a circular locus, passing
through the origin. The branch current in the fixed inductance I is

1 1 = I — — ^ = 7 — , „ X , -y amperes Z (16)

g-\-JCoo (l +Zg) -fjZcco

also evidently a circular locus when c varies.

Consequently, by analysis based on purely electrical relations,
the vector circular locus Im is resolved into a pair of component
vector circular loci h and Ic, such that at any assigned value of the

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 115

capacitance c, the vector sum of Ii and Ic, corresponding thereto, is
equal to the value of Im then existing. IMoreover, the vector sum
of 00c and 00 1, which are respectively the vectors from the origin
to the centers of the branch current circles, is equal to 00m, the
vector to the center of the main current circle.

This process of splitting a main current vector circular locus into
branch current circular loci can be indefinitely extended, by sub-
dividing the admittance of any one branch into any desired number
of equivalent branch admittances. Thus, by following the known
electrical rules of splitting admittances, a vector circular locus, fol-
lowed by the main current, can be resolved into any desired number
of component vector circular loci.

Conversely, when a number of branch circuits, forming divisions
of a main circuit containing a steady e.m.f ., connect a pair of termi-
nals, and one of them has its impedance varied circularly, the cur-
rents in all will undergo circular variation in locus. The vector
sum of all the currents in the unvaried branches can be obtained
by replacing their several fixed admittances by a single joint admit-
tance, and determining, from electrical considerations, the vector
circular locus established in this joint admittance by the circular
variation of impedance in the outstanding branch.

Voltage Distribution Along a Series of High-Resistance
Coils at 2010^.

A series of high-resistance coils, such as a megohm box, divided
into a number of sections, develops an interesting electrical condi-
tion, when used with alternating currents, which does not obtrude
itself upon the attention of the observer when such resistances are
used in continuous-current tests. This condition is a superposed
alternating-current distribution in the coils, due to their distributed
capacitance and the relatively high a.-c. potentials which are im-
pressed upon their terminals. In the continuous-current case, the
capacitances and applied potentials are present, but the charges are
fixed, and these do not interfere with the testing current. The a.-c.
phenomenon has been noticed and studied, but does not appear to
have been presented in the light of potentiometer measurements.

116

KENNELLY AND VELAXDER— POTENTIOMETER

The connections for the test are indicated in Fig. ii. The high-
resistance box AB of 100,000 ohms, is divided into 10 coils of 10,000
ohms each. The end B is grounded, and also connected to the re-
sistance tap of the potentiometer. The exploring tap, passing

jaoo

-J/co

S r .l3n-:o'>f A/

A

B

===

^=p

^^

7

'

1==-^ "

t

a.

' +

f

TruJUA^a^

600

700

^00

Fig. 10. Vector voltage drop observed along a sectioned high resistance
box of 100,000 ohms in 10 coils of 10,000 ohms each, at the frequency of 2,010
cycles per second. Crosses indicate observed, and circles computed values.

through the vibration galvanometer, makes contact in succession
with the junctions i, 2, 3 . . . 10 of the resistance coils, Fig. 11.

The results of an exploration, at 2010-^, are presented in Fig.
10. With the potential of the grounded point B at the origin, the
potential at A was 855 + ;98 millivolts. If the fall of potential
along the coils were regular, the intermediate potentials should have
fallen on the straight line AB, whereas they actually fell upon the
curve 5, I, 2, 3 ... ^. The vector deviations are indicated at
each junction, and they afifect both size and slope. The reason for
this behavior evidently is that, as indicated in Fig. loa, the resist-
ance box resembles an artificial line, in which the line sections are
resistances shunted by condensers ; while each section has a con-
denser leak to ground. The observations indicate that each 10,000-
ohm section subtends, at this frequency, an average hyperbolic angle
of 0.13/15°, which would be accounted for by a shunting con-
denser of 13.7 millimicrofarads per section, and a leak condenser of
0.268 millimicrofarad. The results indicate that the drop of poten-
tial in the section nearest to the ground point is very distinctly less
than one tenth of the total drop BA, and that in the sections near A

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 117

very distinctly greater. Consequently, when using this subdivided
resistance as a volt box, the B coils give too small, and the A coils
too large a result.

It may be observed that it is impossible completely to avoid the
introduction of these virtual shunting and leak condensers, in the
construction of hi8:h-resistance boxes for a.-c. tests. If these vir-

FiG. II. Exploration, by means of the potentiometer P, of the voltage drop
along a high-resistance box AB, of 100,000 ohms.

tual condensers are dissymmetrical, and vary from coil to coil, it
becomes very difficult to make a proper correction for the error due
to charging currents. If, however, care is taken to preserve sym-
metry of structure throughout the resistance box, so that the distri-
butions of virtual capacitance remain uniform from coil to coil, the
correction may be readily computed, at any impressed frequency, by
hyperbolic functions, according to the well-known theory of arti-
ficial lines, from a single set of observations for determining the line
constants. A knowledge of the error will thus virtually eliminate
the error. Suitably supported and connected metallic ground shields^
surrounding each resistance coil, would serve to distribute the leak
capacitance uniformly. High-resistance section boxes should, there-
fore, be designed so as to subtend small hyperbolic angles per sec-
tion ; but whatever the section angle is, it should be regular
throughout.

118

KENNELLY AND VELAXDER— POTENTIOMETER

Receiving-End Impedance of a Cable with a Variable-
Inductance Load.

The a.-c. potentiometer lends itself advantageously to the experi-
mental study of a.-c. artificial lines in the laboratory. An example
of such an application appears in Fig. 12. The plan of connections
shows an artificial telephone cable ten miles (16.1 km.) in length, in
/-sections of two miles each, or less. The lumpiness correction for
this cable, at 1200-—', was found to be negligible. The constants of
the cable are r" = 88 ohms per loop mile (54.7 ohms per loop km.)
and c" = o.o6 microfarad per loop mile (0.0373 fii. per loop km.).
The total conductor resistance was thus 880 ohms, and the total
distributed loop capacitance 0.6 microfarad.

Fig. 12. Vector current circular locus at receving end of an artificial
telephone cable with variable inductance load, under constant impressed
e.m.f. at sending end and at a steady frequency of 1,208 cycles per second.

At the sending end A of the artificial cable, a pliotron oscillator
impressed a steady e.m.f., reduced to i.o volt at reference phase
and of frequency 1208-^. At the receiving end B', an adjustable

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 119

inductance L, of negligible resistance, was used as a receiving-end
load. The received current was measured by observing the p.d.
across the last resistance section pp' of the artificial line.

With the inductance L reduced to zero, the receiving-end cur-
rent was observed to be the vector 01? =1.03 X lO"^ ^ 2)7° ^^'
liampere. The formula for I ^ is

/« =

i.o Z 0°

£0 sinh d -\-jLcx} cosh d

amperes Z (17)

where ^o is the surge impedance of 439 \ 43°-6 ohms,

0 is the angle subtended by the line 2.0 Z 44° -3 hyps Z ,
w is the impressed angular velocity ^ 27r X 1208 = 7590 ra-
dians/sec.
With L^o, I^ becomes 1.035 X lO"^ 'x 37°. 4 ampere.

By giving to /Loj successively increasing values up to J2640 ohms,
the observed received current strength was found to pursue the cir-
cular locus DPO. This circle has its diameter at OP, when the
value of the reactance is 352 ohms. At this condition, the received
current strength was a maximum at 1.4 X lO"^ Y 80° ampere.

The sector of the circle in Fig. 12 between 0 and D might have

Fig. 13. Rectilinear graph of receiving-end impedance for case pre-
sented in Fig. 12, as the load of receiving-end inductance is varied, without
changing the resistance from zero vahie.

120

KEXNELLY AND VELANDER— POTENTIOMETER

been covered, if a condenser had been substituted for the inductance
load.

Fig 13 shows why a circular current graph might be expected in
this case at the receiving end. The denominator in formula (17)
contains the vector sum of a fixed impedence ^r^ sinh 6 and the vari-
able impedance jLio cosh 8. In this term oj and cosh 6 are constant,
so that L is the only variable. The vector OA, Fig. 13, represents
s^, sinh 0. AC is the vector jLoi cosh 0, for L corresponding to Lw/R
= 0.4. The vector sum s^ sinh 9 -\- jLm cosh B is then OC. To suc-
cessive values of L, from L = o to L = 0.175 henry, correspond suc-
cessive distances from A, along the straight line AB. The vector
locus of the sum appearing in the denominator of (17), is therefore
a straight line. The reciprocal of this vector sum must conse-
quently follow a circular locus, and corresponds to the circular locus
of Fig. 12. OP is the perpendicular from the origin to the line AB.
The point P thus marks a minimum receiving-end impedance, and
a maximum received current. This corresponds to a certain type
of resonance.

^u.C.1.

J.o //i

Fig. 14. Resonance curve of received current strength squared against the
ratio of receiving end reactance to total resistance.

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 121

If we plot the square of the received current strength of Fig. 12,
as scalar ordinates, to L(d/R as abscissas, we obtain the resonance
curve of Fig. 14. It follows from the linear relations of Lw/R
along the straight line AB, in Fig. 13, that the ordinates of the reso-
nance curve in Fig. 14 will lie symmetrically at equally spaced ab-
scissas on each side of the maximum. If a condenser had been used
as a receiving-end load beyond the point L = o, the part of the reso-
nance curve, Fig. 14, on the left-hand of the origin, might have been
covered.

Prevalence of Circular Vector Loci.

It may be noted that in the cases of Figs. 2, 4, 5, 7, 9 and 12, cir-
cular graphs of vector potential loci have presented themselves, and
many measurements in the laboratory, at constant frequency, lead
to such circular graphs.

If we consider any fixed network of conductors, we may assume
that all joints are electrically good, all leaks are steady, and all iron
cores removed from the inductances ; or, in other words, assume
that each and all of the elements of the network obey Ohm's law.

Let us select arbitrarily any pair of terminals A B in the net-
work, at which a constant simple alternating e.m.f. is applied, in the
steady state, at a fixed frequency, and also select any other pair of
terminals CD in the network, at which the effects of the applied
e.m.f. are to be looked for. It will be shown, in the Appendix, that
if we introduce, between C and D, an impedance which varies along
a circular locus in the impedance plane, a straight line variation being
included as a particular case, then both the current and voltage will
be caused to vary along a circular locus at CD. Moreover, the cur-
rent and sending-end impedance at the terminals A B will be caused
to vary along a circular locus. Again, if a circularly varied impe-
dance is inserted between the impressed e.m.f. and the AB termi-
nals, the voltage and current at those terminals will be caused to
vary circularly, and also the voltage and current at CD. A further
development of this remarkable phenomenon will be found in the
Appendix. It means that a conducting network has a wonderful
propensity for reproducing impressed circular variations. A net-
work of n elements, in one of which a circular variation of inipe-

122 KENNELLY AND VELANDER— POTENTIOMETER

dance or admittance is impressed, will give rise to 7i-i new circular
variations of current ; i.e., one in each of the other elements, assum-
ing all impressed e.m.f.'s simultaneously acting on the system re-
main constant. The circular variations in terminal p.d.'s through-
out the system will be of the order n-, neglecting intermediate po-
tentials between terminals.

APPENDIX.
Circular Variation in Networks of Conductors.

It is proposed to establish the following main proposition :

In any network of conductors, all the elements of which obey
generalized Ohm's law, subjected in the steady state to any fre-
quency of constant e.m.f. (including zero frequency as the
limiting continuous-current case), circular variation of the im-
pedance of any element will produce a corresponding circular
variation in the current in every element of the network.
Moreover, the p.d. between any two points of the network will
be caused to vary circularly. The sending-end impedance or
admittance between any two points on the network will be like-
wise caused to vary circularly. By " circular variation " is
meant variation over a planevector circular locus, including the
straight-line locus as a limiting case.

The prevalence of circular loci in relation to alternating-current
circuits has been recognized in the literature of the subject,^ which
contains various scattered theorems bearing upon special cases. A
few of these theorems may advantageously be collated here in pre-
senting the demonstration of the main proposition.

It was shown by Clerk INIaxwell'^ that if, in any continuous-
current network, two terminals, say ^ and 5^ are selected as sending-
end terminals, and two other terminals, say C and D, are selected as
receiving-end terminals, a current / applied to the network at A B
would produce the same p.d., at CD, as would be produced at A B,
if the same current / were applied at CD. This theorem may be

5 Bibliography, 5, 6, 7, ja, 10.
^ Bibliography, 3 ; also 4, 8.

MEASUREMEx\TS AT TELEPHONIC FREQUENCIES.

123

given an interesting interpretation in terms of the equivalent T of
the network, with regard to terminals AB, CD. By generalization
from several well-known examples of equivalent T-s, LaCour^ dem-
onstrated that any a.-c. or d.-c. network can be replaced by a certain
equivalent T, in general dissymmetrical, so that this T, insofar as
regards conditions at A, B, C, D will be the exact equivalent of the
actual network.

too .

oo

10U .

o

o
o

o
o

o
o

100 B

100

C -.00

^

^---^

i

n

o
o
ri

8

■^

^ 1P9 „

oc

(CO iJ

8

o

I

s

100

+00

100

1

o
o

%

o

ioa<-'.„

CO

100

i^r

"L^a' too

9 00 I 100 ^^

J >-

■70000097 J.

097 jAm^ 9 233-53J8 o^m:.

A o.

0 gA/>VL /)

JL

Z293 • 753 a/urr^. /°

0- 4'3'[ 9i% Z ^ fO'' iWu«

0 efvm/ . oc ■mA.M^

id

Fig. 15. Continuous-current network with its equivalent T and n.

As a simple example, we may consider the continuous-current
plane network of Fig. 15. The resistances of the various elements
in ohms are marked thereon. Some of them are made infinite, in

'' Bibliography, 7 and 7a.

124 KENNELLY AND VELANDER— POTENTIOMETER

order to simplify the computation. With respect to the pairs of
terminals A B and CD, this network reduces to the A' B' C D' con-
nection of six elements. By any of the known processes, this may
be reduced to the equivalent T of Fig. i^h, in which the staff resist-
ance is 40.00025 ohms. It is evident that i ampere applied to this
T at the A B terminals, will produce 40.00025 volts at C D, and that
reciprocally i ampere applied to this T at the CD terminals will
also produce 40.00025 volts at B A. The "mutual impedance"® of
the network between the pairs of terminals B A and CD is thus
40.00025 ohms.

Maxwell also showed^ that if continuous e.m.f. w^ere applied at
A B, and a current thereby produced through the terminals C D when
short-circuited, the same e.m.f. applied at CD would produce the
same current through the terminals A B short-circuited. A similar
theorem had been enunciated already by Kirchhoff.^°

This proposition may, in a similar way, be given an interpreta-
tion in terms of the equivalent n of the network. In Fig. i^c,b c ad
is the equivalent n with respect to the terminals A B, CD. It is
evident that an e.m.f. of i volt applied at &o would produce a cur-
rent of 0.4379562 milliampere at c d, and that reciprocally, i volt
at c rf would produce the same current at a & shorted. In other
words, the mutual admittance of the network with respect to the
two selected pairs of terminals is the architrave admittance of the
equivalent IT. Moreover, the equivalent IT of mutual admittance
is the conjugate system of the equivalent T of " mutual impedance,"
and may be directly computed therefrom. Consequently, if the
" mutual impedance " T of a network, with respect to two pairs of
terminals, has been ascertained, the mutual admittance 11 of the
same network and terminals can be deduced by Ti-T substitution
and without further experimental investigation."

In Fig. 16, a relatively simple alternating-current plane network
example is worked out. The impedance of each element of the
nine-element network is marked thereon. The selected terminals
are AB and CD. The equivalent T is shown in Fig. i6h, and the

8 Bibliography 12.
» Bibliography 3, 4, 8.

10 Bibliography I, p. 32.

11 Bibliography 6a.

MEASUREMENTS AT TELEPHONIC FREQUENCIES, 125

equivalent n in Fig. i6c. The " mutual impedance " of the network
for these pairs of terminals is 10.48605 Z 95°. oo'.47" ohms, and the
mutual admittance 0.936265 X io"^Z 34°.59'.3i" mho.

'13-i'oi>'iL77'o'-32\j,miOllTli!>}^ZZ'-Ul^'l^'c/,m

'^■iOi-iM\-J1^0'nn.lw

M

7-ii']l.jL^°u.i.'ii-",^c

0'

-o C

csth

2i5-X/o''Z'*''-

iy-3/'

rrxktj.

5

10(^9 ■

oA/m.^.

^

ON

JR.

>-

^

X

g-«.

'^S

G

k'

S

43

r

0 A/-V .OC r.yljn

r

i]

g-.

Fig. 16. Alternating-current network with its equivalent T and n.

In any actual network, it is theoretically easy to determine the
elements of the equivalent T experimentally. It suffices to measure
EcD and I^^ in order to determine R^,. A measurement of the
sending-end impedances at each pair of terminals in turn, will then
serve to evakiate r^ and n- Similar treatment applies to the experi-
mental determination of the IT, if this is preferred.

126 KENNELLY AND VELANDER— POTENTIOMETER

In the equivalent T's of Figs. 15 and 16, all of the series im-
pedance appears in the upper arms, and none in the lower arms.
All of the impedance might, however, be transferred to the lower
arms, without altering the equivalence of the circuit. ^Moreover,
any desired share of impedance might be transferred from the
upper to the lower arms, according to the rules of T and / equiva-
lent circuits.^- Similarly, the architrave impedance of the n's of
Figs. 15 and 16 may be either wholly or partly transferred to the
lower line.

In any dissymmetrical T and its conjugate 11, we have the re-
lation

Pi R\ gl • / / ON

— = — = — numeric Z (18)

P2 -^2 gi

In the case of Fig. 15, this ratio is 1.2857. In the case of Fig. 16,
it is 0.53348 Z 99°. 46'.47". This also means that in any a.-c. case,
the difference in the slopes of p-^ and p., will be the same as the dif-
ference in the slopes of R^ and J?o, or

Pi - po = ;^i - ^2 = i2 - gi degrees (19)

this difference being 99°.46'.47".

Again in any dissymmetrical T and its conjugate or equivalent
n, the geometrical mean of the two T-arm impedances, times the
architrave admittance equals the geometric mean of the two IT-leak
admittance times the T-staff impedance, or

V Vpipo = Rr'^gi-g2 numeric Z (20)

In the case of Fig. 15, this product is o. 11 587. In that of Fig. 16,
it is 0.094224 Z 62°. 6'.4o". From this relation it also follows that

^T + fi = ^ + P2 degrees (21)

In the case of Fig. 16, this qviantity is 12°. 13'. 16".

Returning now to the main proposition, if an alternating-current
network is connected, at the receiving-end terminals D C, to a. circu-
larly varied impedance load, the impedance of the circuit ODCO',

^2 Bibliography 11 and iia.

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 127

Fig. i6b will also have circular variation, and the admittance of this
branch, including this load, must also vary circularly by the geome-
try of inversion. ^^ If, then, we add the constant admittance of the
staff leak g^, Fig. i6b, it follows that the total admittance on the
right-hand side of 00' will have circular variation, under circular
variation of the load at B C. Taking the reciprocal of this circular
admittance, the impedance of the system on the right-hand side of
00' will also have circular variation. Adding to this the constant
impedance A O , the total sending-end impedance at terminals A B
must have circular variation, as likewise the sending-end admittance.

Consequently, a circular variation of load a.t D C must produce
a circular variation of current a.t A B under constant impressed
e.m.f., or a circular variation of voltage at A B under constant im-
pressed current.

Again, referring to Fig. i6c, if the constant e.m.f. impressed at
ab is E, and an admittance load 3', which varies circularly, is applied
at terminals dc, then the admittance at ab, excluding the constant
leak ab is the circularly varying admittance :

Y = mhos Z (22)

P + —J—

Hence the entering current in the architrave ad will be

^ = ^3'=,+pfe + y) amperes/ (23)

Of this current, the fraction y/igo-^y) will pass through the load
and the current i in this load will therefore be

^■ = i-+pJ+3,) amperes/ (24)

The ratio E/i is the receiving-end impedance of the loaded system
and is

I -\- pgo

Z = p -\- ohms Z (2=5)

y ^ >^/

^3 Bibliography 2.

128 KENNELLY AND VELANDER— POTENTIOMETER

In this expression, if 3' varies circularly, so does its reciprocal, and
therefore so does Z. Consequently, if the sending-end impressed
voltage is held constant, and y is varied circularly, the current deliv-
ered to the load at the receiving-end will undergo a circular varia-
tion. From similar considerations, it may be seen that if the sending-
end impressed current is held constant, the voltage at receiving
terminals will also undergo circular variation, when y is varied
circularly.

We have hitherto considered the efifect of a circularly varied
load at the receiving terminals CD, in producing circular variations
of impedance, voltage, and current, both at those terminals and at
the sending terminals A B. We may now consider, in like manner,
the effect of circularly varied impedance, voltage and current at the
sending end.

Referring to Fig. 16b, let the receiving terminals DC be con-
nected through a fixed impedance load a ohms Z , such that

P2 + 0- = So = — ohms Z .

Then the total admittance at 0 will be g,- -\- y., mhos Z • The total
mipedance at the terminals A B is then p^ -f- [^/{gj,-}- yo)] ohms Z.
Let an e.m.f. of fixed frequency and constant vector value E be
applied at the terminals A B, through an impedance r^ ohms Z ,
which impedance is varied circularly. Then the total impedance Z ^
of the circuit at the sending-end is

Za= ZI + pi-\- ohms Z (26)

gr + yi

Since z.^ is supposed to be the only variable, the circular variation
of ^1 causes circular variation in Z^; so that the current entering
the network a.t A B is circularly varied under constant e.m.f. E ;
or, if the entering current should be maintained constant, the im-
pressed voltage E must be varied circularly to correspond.
The entering current cct A B being

E E(g, -\- y.j)

za~ (si+pi)(gr+>'2)+i

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 129

The current h- passing through the constant load a at C Z) is

3'2 E.-y2
Ic = Ia ' — , — = 7 , 77 , ^— , — amperes Z (28)

The denominator in this expression is a vector which undergoes cir-
cular variation, when s^ is varied circularly, and since Ey^. is con-
stant, Ic must also vary circularly. This means that the receiving-
end impedance of the network at CD, with respect to constant e.m.f.
at AB, undergoes circular variation as 2^ is varied circularly.

If constant current and not constant e.m.f. is impressed at AB,
through a circularly varied impedance s^, a similar expression for
receiving-end current and voltage at CD is obtained.

Finally, if with s-^ and a constant, the impressed e.m.f. and cur-
rent are varied circularly at AB, it is manifest that the p.d. and the
current in each and every element of the network, including the
load at CD, will be varied circularly and in the same simple propor-
tion. In this case, the circular variations reproduced are all of the
same type throughout the network, whereas in the preceding cases
discussed, the circular variations produced are, in general, different
in the dift'erent elements.

We have hitherto considered only the pairs of terminals A B
and CD. Since, however, these are chosen arbitrarily in the net-
work, the above propositions must apply between any pairs of termi-
nals ; so that if the impedance load at any pair of receiving termi-
nals is varied circularly, the impedance of the system, including the
load, will be caused to vary circularly at any other pair of terminals.
Moreover, if either the current or the voltage impressed at any pair
of terminals is varied circularly, the potential and current at any
other pair of terminals will be caused to vary circularly. The case
of constancy, or zero variation, must be included as belonging to a
circular locus of zero radius. Zero variations will occur if the termi-
nals AB are "conjugate" to the terminals CD}^

Finally, the current in any element of the network must vary
circularly when a constant e.m.f. is impressed at any pair of termi-
nals and any other element has its admittance, or impedance, circu-
larly varied. This is seen from an examination of Kirchhoff's or

1* Bibliography i and 3.

PROC. AMER. PHIL. SOC, VOL. LVIII, I, JULY I I, I9I9.

130 KENNELLY AND VELANDER— POTENTIOAIETER

Maxwell's demonstration/' when the reasoning is extended to com-
plex quantities. It may be seen that in the symbols here employed,

A-\-By
irs = Epq • r -i. J) amperes Z (29)

where irs is the current in the element rs, when the e.m.f. Epq is
inserted in the element pq, and y is the admittance of any third ele-
ment ttv, which is subject to circular variation ; while A, B, C, and D
are system constants, independent of y. It is well known, from the
theory of functions of a complex variable, that the linear transfor-
mation represented by (29) gives rise to a circular locus, if y varies
along a circle.

It may be noted that by reason of the known principle of vector
superposition of currents in a network, the multiplication of circular
loci extends to cases where a plurahty of e.m.f. 's coexist. The
proposition may also be extended to include mutual inductances.
Strictly speaking, iron-cored inductances are excluded, however,
owing to the imperfect application of Ohm's law to them, under
varying permeability and magnetic skin effect.

If the impressed alternating e.m.f. or e.m.f. 's are impure, so that
harmonics are present, any rectilinear variation in the impedance or
admittance of any element in the network will also be rectilinear for
any harmonic frequency, and so will give rise to circular variations
in the harmonic voltages and currents throughout the network as
well as in the fundamental. In other words, the multiplication of
circular variations in a network, due to rectilinear variation in the
impedance or admittance of one element, applies not only to the
fundamental frequency, but also to any superposed frequencies.

Summary.

1. Experimental results obtained with a new type of a.-c. poten-
tiometer are discussed.

2. The circular graphs of current and admittance are analyzed,
for the case of a simple RLC circuit, at the successive constant fre-
quencies of 500, 1,000 and 2,010 cycles per second, when either R
or C is varied alone.

15 Bibliography i, p. 25 and 3, Vol. i, Section 282a.

MEASUREMENTS AT TELEPHONIC FREQUENCIES. 131

3. The circular graphs of current, admittance, voltage and im-
pedance are analyzed for the case of a divided circuit, operated at
constant e.m.f. and frequency. The problem of composition and
resolution of component and main circular graphs is discussed in
the light of the experimental results.

4. The distribution of potential over a sectional high-resistance
box is explored, at an impressed frequency of 2,010 cycles per
second.

5. The circular graph of receiving-end current is given for the
case of an artificial telephone cable operated at constant voltage,
and at the frequency of 1,208 cycles per second, when the receiver
has variable purely reactive impedance.

6. Certain general propositions are offered concerning the devel-
opment of vector circular loci in the voltage, current, impedance and
admittance of an alternating current network in any steady state.
These propositions are shown to be connected with the equivalent T
or n of the network, with respect to any two pairs of selected
terminals.

Bibliography.

Without Pretensions as to Completeness.

1. Kirchhoff. Pogg. Ann., Vol. 72, 1847. Ges. Werke.

2. Mobius. Ges. Werke, Vol. 2, 1855. p. 245.

3. J. Clerk Maxwell. Electricity and Magnetism, 2d ed., 1881. Vol. i, p.

373.

4. A. Gray. Absolute Measurements in Electricity and Magnetism, 1888,

Vol. I, p. 160.

5. Bedell & Crehore. Alternating Currents, 1893.

6. Houston & Kennelly. Electrical World, Oct. 30, 1897.

6a. A. E. Kennelly. The Equivalence of Triangles and Three-Pointed Stars
in Conducting Networks. Elec. World and Engineer, Vol. 34, Sept.
16, 1899, pp. 413-414-

7. J. L. la Cour. Leerlauf und Kurzschluss, 1904.

7a. la Cour & Bragstad. Theory and Calculation of Electric Currents,
Transl., 1913.

8. Jeans. Electricity and Magnetism, 1908.

9. A Larsen. Der Komplexe Kompensator, ein Apparat zur Messung von

Wechselstromen durch Kompensation. Elek. Zeifschrift, Oct. 13,
1910, Vol. 31, pp. 103^1041 ; also The Electrical World, Vol. 56,
Nov. 3, 1910, pp. 1085-1088.
10. G. A. Campbell. Cisoidal Oscillations. Trans. A. I. E. E., Vol. 30, part 2,
p. 873, April, 1911.

132 KENNELLY AND VELANDER— POTENTIOMETER

11. A. E. Kennelly. The Application of Hyperbolic Functions to Electrical

Engineering Problems, 1912, Appendix E, p. 222.
iia. A. E. Kennelly. Artificial Electric Lines, 1917.

12. Standarization Rules, American Institute of Electric Engineers, 1916,

Section 916.

13. A. E. Hanson. The Design and Construction of an Alternating-Current

Potentiometer, Sept., 1918. A thesis towards the degree of M.S. in
electrical engineering at the Mass. Inst. Tech.

14. A. E. Kennelly and Edy Velander. A Rectangular-Component Two-

Dimensional A.-C. Potentiometer. Journal of the Franklin Insti-
tute, July, 1919.

SOME SCIENTIFIC ASPECTS OF THE METEOROLOG-
ICAL WORK OF THE UNITED STATES ARMY.

By ROBERT A. MILLIKAN, Ph.D., Sc.D.
{Read April 26, 19 19.)

There is no more interesting illustration of the application of
new scientific methods to warfare than is furnished by the develop-
ments in meteorology during the great war. Prior to 1914 a me-
teorological section was not considered a necessary part of the mili-
tary service. No corrections had ever been made by the artillery
of any army for any save surface winds. Firing by the map was
almost unknown. No Sound-ranging Service, no Air Service and
no Anti-aircraft artillery had ever existed to demand aerological
data.

At the time of the signing of the armistice on the western front
the Air Service and all the artillery were being furnished every two
hours with the temperature, density, wind-speed and direction,
taken at the surface and at various altitudes, from 100 to 500 meters
apart, up to 5,000 meters. Further, tables were prepared from
which each battery could obtain the correction suited to its trajectory
for the so-called ballistic wind. This is, the average wind for the
trajectory, weighted for the density of the air at the elevations trav-
ersed. Even machine guns when used for barrage work made use
of these ballistic-wind tables.

In addition, daily forecasts were furnished to the armies in ac-
cordance with the following outline :

A. Character of weather for each arm of the service.

B. Winds : Surface at 2,000 m., at 5,000 m.

C. Cloudiness including fog and haze.

D. Height of cloud.

E. Visibility.

F. Rain and snow.

G. Temperature.

133

134 MILLIKAN— METEOROLOGICAL WORK

H. Warning of weather conditions favorable for use of gas by

enemy.
K. Probable accuracy or odds in favor of forecast.

Most of the aerological data was obtained from theodolite ob-
servations on pilot balloons. The extent to which our knowledge
of the upper air has been, and is being, extended by this pilot bal-
loon work may be seen from the fact that before the war there
existed but one station in the United States where pilot balloon
explorations were regularly carried on. Within a year of the in-
ception of the meteorological service in the United States Army,
thirty-seven complete stations for the obtaining of both surface and
upper air data in aid of aviation and the artillery had been estab-
lished in the United States and equipped with special aircraft theo-
dolites and pilot balloons, neither of which had ever been produced
before in this country. Further, twenty such stations had been
established by our forces abroad. For the manning of this service,
about five hundred specially selected men had been trained in this
country, and three hundred and fourteen of them sent abroad, while
about two hundred were held for work in the United States.

The scientific interest in this service centers about four distinct
problems :

1. The extension of our knowledge of the law of motion of pilot
balloons.

2. The procurement of data and the development of methods
for the preparation of artillery range tables.

3. The development of long range propaganda balloons.

4. The charting of the upper air in the United States and over-
seas in aid of aviation.

I. The Extension of Our Knoivledge of the Lazv of Motion of
Pilot Balloons. — Prior to the development of the meteorological
service of the army there had been made in the United States per-
haps one hundred pilot balloon flights in which the balloons had
been followed by the two-theodolite method — the only method which
permits of real accuracy — and in several European countries there
had been a somewhat greater number, but the data was incomplete
and fragmentary.

Within the past year approximately five thousand such observa-

OF THE UNITED STATES ARMY.

135

tions have been taken by the meteorological service of the Signal
Corps. From these observations the altitude of the balloon is deter-
mined with great accuracy by triangulation, the base line being
usually a mile or more in length. The balloon is kept in sight up to
distances as great as sixty miles, and up to heights as great as 32,000

.'.500
4000

JSOC

t

1

fort Sill, Okia.

luly 30, lql8 <}:I7

3. m. 0

^

No, 1

.2000

"

c'

^ ZiW

0

5

9

0 ■

■ j

■ ^ ItSOt.

■

'

ISOC

- " ■

. J 000

000

TypkalKats of A»te

i

nt

[

f^

^ . 5 10 li no 2b isa

. 35

"~48

Uniform rate of ascent of pilot balloon up to 4,500 meters.

meters, or approximately twenty miles. For the practical uses of
the artillery and the air service, observations need not be carried
higher than 10,000 meters (six miles), which is the extreme height
to which airplanes have thus far ascended, or to which projectiles
usually go.

In view of the number of variables which enter into the rate of
ascent of pilot balloons, such as the changing density and the chang-

136

MILLIKAN— METEOROLOGICAL WORK

ing temperature of the surrounding air, the changing size of the
balloon and consequent changing tension of the rubber envelope, the
changing temperature of its interior because of the absorption of
the sun's rays, the diffusion of hydrogen through its walls, etc., it
is one of the most striking facts to be found anywhere in the annals

c", Aberdeen Pr«i/i,,n2 Ground, Mci,

tinv^orm Rate o| Ascent

70 80

Uniform rate of ascent of pilot balloon up to io,ooo meters.

of empirical science that these balloons rise to great heights without
deviating appreciably from the simplest possible law of ascent,
namely, that of constant speed. Graphs Nos. i, 2, 3, 4 and 5 show
beautiful examples of this constancy. Graph No. 6 shows a kink
at about 5,500 meters, which is presumably due to a descending
current struck at that altitude. Graph No. 7 shows a balloon fol-
lowed to a height of 20,000 meters where it apparently developed a

OF THE UNITED STATES ARMY.

137

leak and failed to ascend further. Graph No. 8 shows the fluctua-
tions which are often found at low altitudes, these fluctuations being
undoubtedly due to ascending and descending currents.

j IZOOQ

1 1 000
.. joooe

I 800C

TOOO

1 £ 6900

' r 6090

400»

3000

HazeiKurb-t Fi-elfi

MineoU.N.Y

^

October -i 4-. li^lS ' 3:00 f.m

/ . ' No.

3

i s"

.

'i

i ■-! ' ' '

. ...

■]

,i-j_a_M-

' 2000
1000
1 0

0^ Unkfortrii Ra.ir

e a

t A see lit ...

J

^

/ , : ; i " ^ , 1 , 1 i '

!'T' i ■ ;": \"-]

0 10 £0 410 -to «0 CO i

TP«E(fll«UTt5) • ,

Uniform rate of ascent of pilot balloon up to ii,ooo meters.

The extreme constancy in the rate of ascent, shown in a great
majority of flights, although surprising enough, is not as inexplicable
as it at first appears, for since the pressure within the balloon due
to the tension of the rubber itself is only from five to eight centi-
meters of water, and since this pressure is at sea level less than i
per cent, of the pressure of the atmosphere, it will be seen that the
balloon will expand practically freely, that is, as though the walls
did not constrain it at all, up to heights of say 10,000 meters where

138

MILLIKAN— METEOROLOGICAL WORK

the pressure is about a third of an atmosphere. This means that
the ascensional force must be entirely independent of temperature
and pressure.^ For the speeds with which these balloons ascend,
namely, about three meters a second, the resistance to motion must
be directly proportional to the density of the air and experiment

I r

L

Akefdeerv Prot/tnig Crfo and, W <k

^"October" 2 3, Iq lit " l0:6'0Z"i^7~'"^

Ho A

^'

L° r ■ M i-

[--i:-4--J.-^

Ith ii.|br#; aate IdI 'Ascent*

-U^

0 6 10 IS 2U 25 30 35 « 4S SO SS SO <S

Time (Minutes)

Uniform rate of ascent of pilot balloon up to 8,ooo meters.

shows it to be nearly proportional to the cross section of the balloon,
that is, to the square of the radius. This makes the resistance vary

1 For if /i, di, Vipiti represent ascensional force, density, volume, pressure
and temperature at the surface of the earth, and f-2, d-, Vi, p2, ti, the correspond-
ing quantities at any given elevation, then since dJdi=^Vi/v-.=^ pJjpit. (i)
and fJf2^=Vidi/Vid2 (2) there results from a combination of i and 2,
fjf., = vAJvid-, = p;UlpxU X p^hlp-^x — I.

OF THE UNITED STATES ARMY.

139

as the cube root of the density,- which means that at a height of
6,000 meters, where the density is about one half, the resistance is
.83, of what it would be at the surface. If, as is approximately true
for these speeds, the resistance varies as the square of the velocity,

12M»

1

"" 1

nooc

0 '

0

i

'

Moai

a
0

SOti

Maielhivra f.eU, N.Y.

■-0

i
1

£000

Ottahcf 8. m* .?: 01 f m 0°

; 1

•'.'So:

f., s

a"

': '

- alKJ

o'""

iOCil

zaaa

lOOO

o"" , ,

' !

0 ' ' ■ 1

0 , , ,

& ■'

0° ^

c

0

0 '

0° U-Ri.j.oi'rn R-Ci-te. uj- Aiccnt

.0

: 1

: 1

• ' : ■

;

0.

10 iO 3(1 ■iO 50 6u

^

Uniform rate of ascent of pilot balloon up to ii,ooo meters.

or the velocity as the square root of the resistance, this would mean
that the velocity should vary as the sixth root of the density. In

2 For if Ri is the resistance at the earth's surface and R2 that at any
given altitude,

Ri ^ vM

which is seen from (i) to equal

m

140

MILLIKAN— METEOROLOGICAL WORK

other words, since the sixth root of 2 is 1.13, at a height of 6,000
meters, the velocity should be about 13 per cent, greater than at the
surface. Such an increase in velocity would be very easily observ-
able in the experimental data. The fact that it is not found there

j

'""

1 aooc

i

fiflflO

\\ il

ki

7000

£<00

:'-

S 5000

,-,-

1-

./ ! , L . :

1*^ ■

'""i

" ■ '? i

er

ieiin Pro

Uinq rTToaiid,

,M<i.

3000

3

11 Oi ^

^' '■ ' Oct

obe

r -Z", 1 t( 1 8

IS&S^

\

1000

o"

U.inu&u.ai Ca.s.,e $Ko

OJV

1 u 1 .<; 0 i n T e >1

1

L . J C 6 n U e ct L 6 r

\ Cu-rre nt 1

'

o

. 0

_!_.

0 10 JO 30 40 40 60 70 80'

- - - . - T,^^ ;(,-,. „•,_,)_ ___ _ •__ _ _ _ J

Pilot balloon ascent showing isolated convection current.

is due to the wholly fortuitous circumstance that the slow diffusion
of hydrogen through the walls, as observation by Blair and Sherry
has shown, is just sufficient, with the balloons here used, to retard
the ascensional rate enough to make it quite exactly constant.

This makes it possible, provided one could always duplicate the
size and weight of his balloon, to obtain a very exact determination
of wind velocity and direction by a one-theodolite method, the
height being always known from the time and the known rate of
ascent.

OF THE UNITED STATES ARMY.

141

When, however, the weight and inflation of the balloons are
varied, as they must be in practice, since the balloons vary in weight
from twenty to thirty-five grams, and since it is convenient also to
vary the filling according as low altitude or high altitude wind-data

;j : I

No.

Col lea'e ■^to^tlan; !TV]i:a.P"i"

July Ifc. Iq\ a _3":_03_p.m. i

E Kceptiokally. Kig hi. As ee ns-i 0 il

0 CS so 73 100 125 ISO " !76 200

TifttCAVmuTt?) _ ■ .^^

Uniform rate of ascent of pilot balloon up to 20,000 meters where balloon

sprung a leak.

are desired, it is found that no accurate formula can be found for
computing the speed in terms of the ascensional force, the weight
to be lifted, and a single invariable constant. For approximate
work, however, the one-theodolite method, because of its conve-
nience and because of the impracticability of measuring an accurate
base line at the front, is much in use, and one of the advances made
in the meteorological w'ork of the army during the past year has

142

MILLIKAN— METEOROLOGICAL WORK

consisted in developing with the aid of the large amount of data
available, a general formula for the rate of ascent in terms of the
ascensional force and the weight to be lifted, which though far
from accurate is more reliable than that which has heretofore been
used. The formula heretofore used is that of Dine's, namely,

4oafi|

[^ 3000!

M 0- s

Hazeihiirit f .cli. L.L..N.VLi

October Zl , la I a 3 0'.

-xaimple Skou/vnq Ctvect

a

0 2 4 6 a 10 12 1+ 16 18 ^Q ^^ n k, ze so

Correction currents at low altitudes.

in which V represents the rate of ascent in meters per minute, / is
the free lift, or the weight of the displaced air less the weight of the
balloon and contained hydrogen, L is the weight of the balloon plus
the free lift and A' is a constant.

The formula as modified by the observers of the Signal Corps is

OF THE UNITED STATES ARMY. 143

V

■'W

This formula is found to fit the observational data within the ranges
used in the Signal Corps work to an accuracy of somewhat less than
10 per cent., which is sufficient for most work at the front.

2. Meteorology in the Aid of the Artillery.— In former times
when guns did not shoot to a greater distance than eight or ten
miles, it was usually possible to observe where the projectile hit and
to correct errors by " spotting." This made unnecessary the cor-
rection of the trajectory for the influence of the w'ind and the chang-
ing density of the air with increasing altitude. In the present war,
however, guns have been built to shoot much farther and in addition
camouflage has prevented the visual location of guns even at the
old ranges. Hostile batteries have been located in many instances
solely by the new art of sound-ranging which has itself demanded
for the high accuracy attained aerological data. The answering
battery has been obliged to fire wholly by the map, so that it is
obvious that it has become necessary to make careful allowances
both for the density of the air and the direction and speed of the
wind at various altitudes. Some of the modern projectiles remain
in the air as long as seventy seconds and a moderate wind blowing
across the path of such a projectile might easily cause it to drop
half a mile away from the point at which it would strike if fired in
still air. The wind-direction and speed at various altitudes have
been obtained, as already indicated by pilot balloons, while the tem-
perature has been determined at the proving grounds by sending
self-recording instruments aloft in specially constructed box-kites,
as well as by sending self-recording instrvmients and meteorological
observers aloft in airplanes. It has been with the aid of observa-
tions of this sort that the new range tables for the Ordnance De-
partment of the United States Army have been constructed. The
importance of this work may be understood when it is considered
that these range tables will be used in connection with the firing of
all guns, and errors in them would produce errors in the range of
every gun fired with their aid.

3. TJie Development of Long Range Propaganda Balloons. — In
view of the fact that above an altitude of 10,000 feet 95 per cent, of

144 MILLIKAN— METEOROLOGICAL WORK

the winds both over western Europe and over the United States
blow from west to east (i.e., have a westerly component), Captain
Sherry in 1917 suggested the development of a large program for
the extension of the use of pilot balloons for the purpose of flooding
the whole of Germany and Austria with propaganda dropped from
such balloons. The project was submitted to the meteorological
and military agencies in France and pronounced infeasible, chiefly
because the rapid diffusion of hydrogen through rubber had hereto-
fore rendered it impossible to obtain pilot balloon flights of more
than about 100 miles. Undiscouraged, however, by these reports.
Mr. W. J. Lester, Dr. S. R. Williams and Sergeant Redman attacked
the problem of extending the range of pilot balloon flights by de-
veloping an automatic ballast-control and by reducing the diffusion
by means of a special dope.

The automatic control was ingeniously simple, its essential fea-
ture being a belly band which kept the girth of the balloon constant
(at a diameter of four feet) through the discharge, in the act of
shrinking, of a few drops of kerosene, thus causing reascension and
consequent expansion.

With this device the balloon not only does not fall but rises very
gradually to higher and higher levels until its ballast of kerosene
or alcohol is exhausted.

In the week beginning October 3, 1918, sixty such balloons, ad-
justed to fly between the initial and final altitudes of 15,000 and
25,000 feet respectively were sent up from Fort Omaha, Nebraska,
carrying return cards and watches, which were arranged to stop and
be let down on small parachutes as soon as the ballast was ex-
hausted. Thirty-four out of sixty of these balloons were picked
up and returned to Washington. Instead of flying 100 miles, one
of them came down within ten miles of New York, 1,100 miles
from Fort Omaha, another was returned from Virginia, 930 miles
from its starting point, and the rest were scattered over Ohio, Ken-
tucky, Illinois, Wisconsin and Iowa. Not one went west of Omaha
though the balloons were sent up on days on which different surface
conditions prevailed.

The credit for this achievement, the significance of which will
be discussed later, is due primarily to Mr. Lester, Captain Sherry,

OF THE UNITED STATES ARMY.

145

Dr. Williams and Sergeant Redman. At the time of the signing of
the armistice the Military Intelligence Service was preparing for
the extensive use of these balloons for flooding the whole of Ger-
many, Austria and even parts of Russia with suitable leaflets, sev-
eral hundred of which could have been scattered by a single balloon,
the total cost of which would have been but two or three dollars.

TABLE I.

War Department, Signal Corps, U. S. Army, Meteorological Service.

Station EUendale, N. D. (90th Meridian Time.)

Wind Aloft Report.

Time 7:00 A.M. Date November 13, 1918.

Altitude,

Direction,

Velocity, | „

Altitude,

Direction,

Velocity,

Meters.

Compass.

M. P. H.

Kemarks.

Meters.

Compass.

M.P. H.

Remarks.

0

SW^

29

5.250

NW

57

250

s

17

5.500

NW

60

500

sw

16

5.750

NW

59

750

sw

15

6,000

WNW

63

1,000

w

15

6,250

NW

69

1,250

w

16

6,500

NW

68

1,500

WNW

18

6,750

NW

68

1.750

WNW

19

7,000

NW

74

2,000

WNW

22

7.250

NW

77

2,250

WNW

25

7.500

NW

85

2,500

WNW

27

7.750

NW

65

2,750

WNW

34

8,000

NW

73

3,000

WNW

35

8,250

NW

76

3.250

NW

35

8,500

NW

69

3.500

NW

40

8,750

NW

75

3.750

NW

41

9,000

NW

73

4,000

NW

41

9,250

WNW

74

4.250

NW

47

9,500

WNW

68

4.500

NW

53

9.750

WNW

65

4.750

NW

56

10,000

WNW

78

5,000

NW

54

10,250

NW

81

4. The Charting of the Upper Air in Aid of Aviation. — In a
recent Brisbane editorial the following sentence occurs : " Flying
machines of the future going long distances will travel at least
32,000 feet up, where no wind blows except the gentle eastern wind
causfd by the earth's motion on its axis." It is quite likely that the
future aviator will fly high, but his motive will be to find an air
current, not to escape one. The gentleness of the zephyrs existing
at high altitudes may be seen from tables i, 2, 3, 4 and 5 which record
three sets of pilot balloon observations recently taken by the Signal

PROC. AMER. PHIL. SOC. VOL. LVIII, J, JULY II, I9I9.

146

MILLIKAN— METEOROLOGICAL WORK

Corps. These tables show air currents increasing in intensity with
increasing altitude and approaching the huge speed of lOO miles
per hour. Such speeds are perhaps exceptional, but not at all un-

TABLE H.

War Department, Signal Corps, U. S. Army, Meteorological Service.

Station Groesbeck, Texas. (90th Meridian Time.)

Wind Aloft Report.

Time 7:00 A.M. Date November i, 1918.

Altitude,

Direction,

Velocity, „ ,

Altitude,

Direction,

Velocity,

Meters.

Compass.

lyj p jj Kemarks.

Meters.

Compass.

M. p. H.

Remarks.

0

E

9

6,750

WNW

25

250

ESE

16

7,000

WNW

19

500

ESE

13

7.250

W

8

750

ESE

2

7.500

W

12

1,000

WSW

5

7.750

W

9

1,250

WNW

II

8,000

WSW

4

1,500

WNW

18

8,250

WSW

16

1.750

WNW

23

8,500

WNW

20

2,000

NW

25

8,750

W

22

2,250

NW

20

9,000

WSW

20

2,500

WNW

20

9.250

WSW

20

2,750

NW

23

9.500

WSW

22

3.000

NNW

21

9.750

WSW

28

3.250

N

18

10,000

w

39

3.500

NNW

29

10,250

w

47

3.750

NW

25

10,500

w

50

4,000

NNW

20

10,750

WNW

59

4.250

NW

20

11,000

WNW

57

4.500

NW

21

11,250

W

44

4.750

NW

20

11,500

W

39

5.000

WNW

16

11.750

W

41

5.250

WNW

18

12,000

W

47

5. 500

WNW

35

12,250

W

51

5.750

WNW

35

12,500

w

56

6,000

WNW

32

12,750

w

59

6,250

WNW

25

13,000

w

60

6,500

WNW

26

I 13.250

w

64

known. The pilot balloon mentioned in 3 travelled from Omaha
to Virginia at an average speed of thirty miles per hour, the average
height being 18,000 feet. On November 6, 1918, at Chattanooga,
Tennessee, a velocity of 154 miles an hour at an altitude of 28,000
feet was observed by one of the meteorological units of the Signal
Corps.

These facts bring out the importance of a forecast of such cur-
rents for the purposes of long flights. A flier aided by such a wind

OF THE UNITED STATES ARMY.

147

TABLE III.

War Department, Signal Corps, U. S. Army, Meteorological Service.

Station EUendale, N. D. (90th Meridian Time.)

Wind Aloft Report.

Time 8:26 A.M. Date December 5, 1918.

Altitude,

Direction,

Velocity,

Altitude,

Direction,

Velocity,

Meters.

Compass.

M. P. H.

Remarks.

Meters.

Meters.

M. P. H.

Remarks.

0

NW

19

1.750

NW

64

250

NW

47

2,000

WNW

68

500

NW

49

2,250

WNW

81

750

NW

57

2,500

WNW

87

1,000

NW

48

2,750

WNW

96

1,250

WNW

49

3,000

WNW

93

1,500

WNW

50

TABLE IV.

War Department, Signal Corps, U. S. Army, Meteorological Service.

Station Mineola, L. I. (75th Meridian Time.)

Wind Aloft Report.

Time 7:06 A.M. Date September 7, 1918.

Altitude,

Direction,

Velocity,

Altitude,

Direction,

Velocity,

Meters.

Compass.

M. P. H.

Remarks.

Meters.

Compass.

M. P. H.

Remarks.

0

N

18

2,000

SW

25

250

N

51

2,250

SW

37

500

N

65

2,500

SW

63

750

N

29

2,750

SW

55

1,000

W

22

3.000

SW

54

1,250

W

20

3.250

SW

55

1,500

w

II

3.500

SW

81

1.750

wsw

13

as that last mentioned would move toward his objective 2 X 154.
or 308 miles an hour more rapidly than if he were opposed by it.
When it is recalled that the aviator above the clouds has no means
of knowing anything about the motion of the air in which he flies,
it will be seen that it is of the greatest importance to him to know
the nature of the currents at different levels. Table 4 furnishes a
very typical illustration of this importance.

From the above data it is evident that an aviator flying toward
the west at this time and place should have flown at an altitude of
i,0CHD meters, while an aviator flying toward the east should have
flown at an altitude of 4,000 meters or more.

148

MILLIKAN— METEOROLOGICAL WORK

TABLE V.

War Department, Signal Corps, U. S. Army, Meteorological Service.

Station Fort Oglethorpe, Ga. (90th Meridian Time.)

IVind Aloft Report.

Time 7 :39 A.M. Date November 29, 1918.

Altitude,

Direction,

Velocity,

Altitude,

Direction, Velocity,

Meters.

Compass.

M. P. H.

Remarks.

Meters.

Compass. M. P. H.

Remarks.

0

NW

7

1-750

w

36

250

NW

8

2,000

w

41

500

NW

II

2,250

w

46

750

WNW

19

2,500

wsw

47

1,000

W

29

2.750

wsw

56

1,250

W

34

3.000

wsw

76

1,500

W

36

3.250

wsw

96

TABLE VL

Altitude

n

Wind Direc-

Wind Velocity In

Meters.

tion.

Miles per Hour.

Surfac

e

NW

2.2

500.

E

5.8

1,000.

E

8.3

5.4

5.4

2,000.

NE

3,000.

W

4,000.

NW

24.6

12,000.

NW

• 49-2

In order to meet the obvious need of the aviator for a knowledge
of upper air currents the Signal Corps in the summer of 191 7 under-
took for the first time in history a general program of mapping the
upper air currents of the United States, the Atlantic and western
Europe in aid of aviation and particularly with reference to trans-
Atlantic flight. By the fall of 1918 twenty-six upper air stations,
carefully distributed over the United States, were in full operation
in place of the one station which has existed before the war. From
these stations reports are telegraphed twice daily to the Weather
Bureau in Washington. From the pilot balloon observations charts
are constructed showing the wind speed and direction at the vari-
ous levels : for instance, one chart shows the wind direction and
speed near the ground, another chart shows the wind direction
and speed 500 meters above the ground and additional charts
show the wind direction and speed at the following levels: 1,000,
1,500, 2,000, 3,000 and 4,000 meters above the ground. The fore-

OF THE UNITED STATES ARMY 149

caster at Washington has the variovis charts before him showing
wipd and weather conditions prevaiHng over the United States
within an hour and a half after the observations are made. From
these charts he prepares the forecast of weather conditions for the
various sections of the United States and at the same time prepares
a statement of the wind and weather conditions at various altitudes
along the various air routes for the use of aerial navigation. This
service is already being used by the aerial mail service, and it is also
used by the military flyers, as is evidenced by telegraphic requests
received at various military meteorological stations for special re-
ports on the weather and wind conditions when long distance flights
are contemplated.

The problem of exploring the upper air currents over the At-
lantic was at first thought insoluble on account of the absence of
fixed bases, but the success of the Meteorological Service in devel-
oping its long-range propaganda balloons has now made possible
the mapping of the upper-air highways across the Atlantic, for ar-
rangements are being made to send up both from coastal stations
and from trans-Atlantic steamers these long-range balloons designed
now for from two to three thousand mile flights, and adjusted to
maintain a constant altitude and to drop in western Europe their
records of average winds in these heretofore unchartable regions.
The importance of this work for the future of aviation needs no
emphasis.

The success which the Meteorological Service has attained would

have been wholly impossible had it not been for the intimate and

effective cooperation which has been extended to it in all of its

projects by Director Marvin and the whole staff of the United States

Weather Bureau. The chief credit for the work abroad should

go to Major William R. Blair, commissioned from the Weather

Bureau for the observational work with the A. E. F. For the

success of the service in this country Captain Sherry and Lieutenant

Waterman have perhaps the chief responsibility. Captain Murphy

and Professor Fassig have, however, contributed very important

elements.

University of Chicago,
April, 1919.

DETECTING OCEAN CURRENTS BY OBSERVING THEIR
HYDROGEN-ION CONCENTRATION.

By ALFRED GOLDSBOROUGH MAYOR.
{Read April 24, 19 19.)

Observations.

The surface water of the middle regions of the Tropical Pacific
commonly flows in a westerly direction due to the effect of the trade
winds, and this water is strongly alkaline, its hydrogen-ion concen-
tration being about 8.22 PH, or 0.602 X lO"*. Occasionally, how-
ever, in the Tropical Pacific one finds a region wherein the surface
water is temporarily flowing toward the east and thus counter to
the trend of the usual current and of the trade winds. I find that
this easterly moving water is commonly less alkaline than is that
of the general region in which it occurs. Thus while the water
moving toward the west is about 8.22 PH these easterly counter
currents are 8.1 to 8.18 PH, or 0.83 X lO"- to 0.66 X 10"^

The tension of the carbon dioxide of the great westerly current
appears to about the same as that of the air above the sea ranging
in our tests from 2.75 to 3.25 ten thousandths of an atmosphere,
whereas the water which is moving in an easterly direction is more
strongly changed with free COo, its tension ranging from 3.45 to
above four ten thousandths of an atmosphere. Moreover, this east-
erly moving water is slightly colder than that of the general region
in which it occurs and Professor L. R. Cary found its oxygen con-
tent was higher than normal. Thus it seems that these counter
currents are caused by deep water which has temporarily appeared
upon the surface either by welling upward or through a local dis-
placement of the westerly-moving surface layer.

This suggests that eddies may be set up due to the gusty nature
of the trade winds, as shown in Fig. i. Every " puff" pushes some
water ahead of it and leaves a hollow in its wake which must be
filled up to the general level by deeper water rising to the surface,

150

MAYOR— DETECTING OCEAN CURRENTS. 161

and this deeper water is in turn replaced by the water which has
been temporarily heaped up in front of the gust. There is nothing
to prevent an under-water counter current, whereas in order to fill
the hollow by surface water the current would have to move against
the prevailing wind as at a, Fig. i ; and the friction between air and

^

^^T)£

sM\

MO c^^\

water is much greater than that between water particle and water
particle. Despite this process of local adjustment, however, the
westerly trend of the surface current tends to raise the general ocean
level in the western regions of the Tropical Atlantic and Pacific, and
this is counterbalanced by the great oceanic surface eddies of which
the Gulf Stream and the Japan Current are well known examples.
The general up-welling of deep water in tropical regions has been
known since the cruise of the Challenger.

Thus due to local causes, water from the depths of the Tropical
Pacific sometimes comes to the surface in large quantity, and retains
some of its easterly movement, even against the prevailing wind.
Then upon being heated by the sun and mixing with the warm sur-
face waters its capacity for retaining free CO^ is reduced, and its
carbon dioxide passes out into the atmosphere.

As is well known McEwen, 1910, 1916, etc., has demonstrated
from studies of salinity that great quantities of deep water are con-
stantly welling up along the Pacific coast of America, and in con-
firmation of this fact I find that the COo tension of the surface
water along the Pacific coast of the United States is considerably
higher than we would expect from its low temperature, and much

152 MAYOR— DETECTING OCEAN CURRENTS.

above that of the water farther off the coast. Thus a PH of 7.85
at 10.5° C, and COo tension of 5.4 ten thousandths of an atmos-
phere were observed 54 miles ofif Golden Gate, San Francisco, on
May I, 1917; and somewhat similar conditions were seen off Van-
couver in September, 1918.

An up- welling of deep water due to off shore winds has also been
demonstrated by Bigelow, 1917, p. 241 ; along the coast of New Eng-
land north of Cape Cod, but this effect is neither so marked nor so
constant in the shallow water along our Atlantic seaboard as it is
off the abrupt slope of the Pacific coast. Indeed Bigelow, 191 5,
191 7, has shown that the cold water which drifts down the Atlantic
coast from Nova Scotia to Florida is chiefly derived from the Gulf
of St. Lawrence and receives accessions from surface drainage and
from rivers along its whole course.

McClendon, 191 7, 1918, showed that the COo tension of surface
water at Tortugas and in the Gulf Stream, is on the average about
in balance with the atmosphere, 30 determination indicating that
the pressure of carbon dioxide of the air ranges from 2.8 to 3.5 ten-
thousandths of an atmosphere, while that of the surface water of
the Tortugas lagoon on the east side of Loggerhead Key ranged
from 2.(i to 3.5 ten-thousandths of an atmosphere, and that of the
Gulf Stream from Key West to Cape Hatteras was from 3.2 to 3.5.
McClendon also found that photosynthesis by marine plants in sun-
light is a very important factor in controlling the hydrogen-ion con-
centration of the water of shallow lagoons or tide pools where the
bottom is covered with sea weed ; for the plants reduce the COo,
thus setting free oxygen and causing the water to become highly
alkaline. For example, while the PH of the sea around the Tor-
tugas is about 8.22 that of the lagoon rose at times to 8.35 by day
and sank to 8.18 at night, and the hot shallow lagoon of the Mar-
quesas, Florida, had a PH of 8.46 accompanied by precipitation of
calcium carbonate. McClendon was, however, unable to find any
appreciable diurnal range in hydrogen-ion concentration of the sur-
face water in the open sea nor can I detect it from my studies in the
Atlantic and Pacific.

Also Wells, 1918, p. 6, shows that the water of the Gulf of Mex-
ico contains about 0.092 grams of CO2 per liter and its surface

MAYOR— DETECTING OCEAN CURRENTS. 153

tension is thus in balance with the atmosphere, and I find that when
uninfluenced by up-welHng of unusual quantities of deep water the
COo tension of the surface waters of the Tropical Atlantic and Pa-
cific is also practically in balance with the atmosphere.

Thus on the voyage of the S. S. Niagara from Fiji to Honolulu,
September 6 to 12, 1918, we encountered only the prevailing westerly
set uninterrupted by any currents moving toward the east, and the
average PH was about 8.22, the average temperature 28° C, and
the COo tension of the water three ten-thousandths of an atmos-
phere and thus practically the same as that of the air. When cold
deep water wells upward to the surface, however, a dift'erent con-
dition ensues, for due to relief of pressure and increase in tempera-
ture this water must discharge its excess CO2 into the atmosphere.

Thus on the voyage of the S. S. Sonoma between Honolulu and
Pago Pago, Samoa, June 25 to 30, 1918, we at times met with strong
currents set toward the east and the average PH was about 8.19
and the COo tension of the surface water 3.26, the average tempera-
ture being 28° C, as on the voyage of the Niagara. Similarly on
the voyage of the S. S. Ventura from Pago Pago, Samoa, to Hono-
lulu on April 19 to 25, 191 7, we met with several strong sets to the
eastward and the average PH was 8.17, the COo tension 3.35 and the
temperature 25.7° C.

Henderson and Cohn, 1916, p. 621, conclude from laboratory
experiments that upon the whole in most places and at most seasons
carbon dioxide must be escaping from the sea into the air, although
they also state that the balance is doubtless restored by COo entering
the water from the air in the polar regions. These authors did not
consider the effect of photosynthesis by marine plants which Mc-
Clendon afterward showed to be such an important factor. Were it
not for photosynthesis it is probable that large quantities of carbon
dioxide would escape from the sea in the tropics, but instead of this
McClendon, Wells and I find that the warm waters are practically
in balance with the atmosphere.

My observations along the Atlantic coast between Nova Scotia
and Florida in December-March show also that the coastal current
during these cold months has a CO2 tension somewhat below that
of the atmosphere, and this may be due to the great concentration
of plant life in these cold waters.

154 MAYOR— DETECTING OCEAN CURRENTS.

. Thus according to my observations the averages for the shore
current the saHnity of which ranges from 30 to 33 grams in 1,000
grams of water, between Nova Scotia and northern Florida in win-
ter are : Temperature 6.7° C, sahnity 31.7, PH 8.05 and CO2 tension
2.5 ten-thousandths of an atmosphere, while similar data for the
Gulf Stream of salinity 36 at the same season between the Straits
of Florida and Cape Hatteras are : Temperature 22.3° C, salinity
36.35, PH 8.21, and COo tension 2.77. Thus the cold shore water
seems to be in a condition to absorb COo from the air, while the
warm Gulf Stream waters are more nearly in balance with the at-
mosphere. In summer when the shore current is warmed to about
22° C, its CO2 tension rises to be quite in balance with the atmos-
phere, as is indicated by McClendon's Table VII., p. 226, 1918.

It is well known from the extensive work of Blackman and
Smith that photosynthesis about doubles in effect for 10° C. rise in
temperature, but due to the action of denitrifying organisms such
as Drew's Pseudomonas calcis the tropical waters are deprived of
nitrogen and can thus support only a meager plant life in compari-
son with that of colder regions. Thus McClendon found less than
o.oi mg., of nitrogen per liter as nitrates and nitrites at Tortugas,
Florida, while Raben, 1910, found more than ten times these
amounts in the North Sea ; and as shown by McClendon the tropical
ocean despite its high temperature can only eliminate a small part
of its free COo by photosynthesis due to the scarcity of plant life.

Krog, 1904, calculated that if the average COo tension of the
ocean is the same as that of the air (about 0.0003 atmosphere), it
must contain twenty-seven times as much COo as the air. Thus if
the ocean gave off one tenth of its COo to the air the carbon dioxide
tension of the sea would sink to 0.0002 atmosphere. He found that
the COo in the air of Disko Island, Greenland, ranged from 0.00025
to 0.007, being high with winds from the north and west, and low
when the wind blew from the south and east. The turbid sea water
at Disko Island had a COo tension of o.oooi to 0.00035, while the
clear water in the same region had a tension of 0.00035 to 0.0006
atmospheres, thus apparently being lower than that of the surround-
ing air.

Also the COo tension of the surface water between Cape Fare-

MAYOR— DETECTING OCEAN CURRENTS. 155

well, southern Greenland and the Shetland Islands was distinctly
lower than that of the air.

The average COo tension of the air over the ocean seems to be
about 0.000295, this being the mean of 51 observations made by
Thorpe between Brazil and England. In 1917, however, using ap-
paratus given to me by Professor McClendon, I tested the CO.
tension of the air over the Pacific between Samoa and San Fran-
cisco, but there was apparently no relation between the local CO.
tension of the air and that of the water under the air, this lack of
coordination being due in all probability to the great mobility and
rapid flvictuation in COo tension in the air as compared with that of
the water. It would apparently be necessary to obtain several thou-
sand determinations of the CO._, tension of air over the ocean taken
at all seasons and in all weathers to determine its mean CO. tension
with accuracy, but the determinations that have been made indicate
that it is not far different from that of the air over the land. My
average for three voyages over the Tropical Pacific is : Temperature
27.5°, PH 8.22, and COo tension 3.15 ten thousandths of an atmos-
phere. Thus the tropical waters appear to have a CO. tension
slightly above that of the atmosphere.

We lack sufificient data for a definite statement as to whether
COo is on the whole passing from air to sea or vice versa, but the
surmise may seem reasonable that a balance is maintained ; the ab-
sorption of COo from the air by the Polar Seas being offset by its
passing out of the ocean over the wide area of the tropics, while the
temperate regions stand in an intermediate relation, the water ab-
sorbing COo during the winter and giving it out to the air during
the summer.

It may be of interest to compare our observations with those of
Palitzsch, 191 1, who was the first to apply Sorensen's colorimetric
methods to the study of the hydrogen-ion concentration of sea
water.

Palitzsch used naphtholphthalein and phenolphthalein as indica-
tors and tested the PH of the Black Sea, Sea of Marmora, Medi-
terranean, Atlantic and North Sea in summer with the following
results :

156 MAYOR— DETECTING OCEAN CURRENTS.

Locality. PH of Sea Wa-er.

Black Sea l^"'-^"^^ ^"^l

^deep water 745

Sea of Marmora and Bosphorus 8.35

Eastern Mediterranean 8.27

Western Mediterranean 8.22

Coast of Portugal 8.25

Off Scotland and Faroe Is 8.08-8.22

S.E. part of North Sea and Skagerak 8. -8.05

Methods.

The hydrogen-ion concentration of sea water was tested by the
simple process of placing 0.4 c.c. of Yio of i per cent, thymolsulpho-
nephthalein in a non-soluble glass test tube of 24 mm. caliber and
adding 30 c.c. of the sea water to be tested. Highly alkaline sea
water gives a blue-green solution, while relatively acid water gives
a greenish-yellow color. Then by comparing this test tube with a
series of sealed tubes of similar caliber containing mixtures of
borax, boracic acid and sodium chloride of various known hydrogen-
ion concentrations we can at once determine the concentration of
our sample sea water with an error of not more than 0.025 PH.

McClendon, Gault and Mulholland, 1917, were the first to stand-
ardize sodiurii borate and boracic acid mixtures for use in measuring
the hydrogen-ion concentration of sea water with thymolsulpho-
nephthalein as an indicator. Their solution consists of 0.075 """•
sodium borate (28.67 grams of Na.B^Oj-ioH.^O), and 19 grams of
sodium chloride, dissolved in water so as to make up 1,000 c.c. of
solution. The other solution consists of 0.3 m. boric acid (18.6
grams of H3BO3) and 22.5 grams of sodium chloride dissolved in
water so as to make up a liter of solution. Definite mixtvires of
these two solutions give correspondingly definite hydrogen-ion con-
centration, as is shown by McClendon, Gault and Mulholland, p. 44,
1917.

Professor McClendon kindly gave me a set of these tubes which
I have used on voyages to test the hydrogen-ion concentration of
the surface water of the ocean. The readings of these tubes were
compared with those of a Leeds and Northrup potentiometer stand-
ardized by the U. S. Bureau of Standards. IMcClendon thus tested
their accuracy in 1917 and I repeated the process in April, 1919,

MAYOR— DETECTING OCEAN CURRENTS. 157

and found that the colors of the tubes, which had been carefully
guarded from light when not in use, had not change in the interval.

As is well known the unit of hydrogen-ion concentration is i
normal hydrogen-ion per liter of water, or about i gram of hydrogen-
ion per liter. The purest distilled water contains only about i gram
of hydrogen-ion in 10,000,000 liters of water at about 22° C, and
thus its hydrogen-ion concentration is about lO"". Sea water, how-
ever, is alkaline and contains only about one tenth this concentra-
tion of hydrogen-ions, or as we say its hydrogen-ion concentration
is about lO"^.

In order to avoid writing negative exponents Sorensen, 1909, p.
28, devised the symbol " PH " to indicate the negative logarithm of
the hydrogen-ion concentration. Thus according to Sorensen's sys-
tem a normal hydrogen-ion concentration, or iO"° of H-ions per
liter, is written PH o; a decinormal H-ion concentration, or lO"^,
is PH I ; while sea water with somewhat less than 0.000,000,01
normal hydrogen-ion concentration is about lO"^ normal and is des-
ignated 8 PH.

H for example a specimen of sea water had a hydrogen-ion con-
centration of

— - — — — - = o.6I6XIO-^
1.622 X 10''

we would call its PH 8.21 because 0.21 is the logarithm of 1.622,
and 8 is the logarithm of 10^ and 0.21 -|- 8. =8.21.

McClendon's borax-boracic colorimetric tubes indicate the true
PH of sea water containing 32 to 33 grams of total salts per 1,000
grams of sea water (salinity .32 per cent, to .33 per cent.), and the
following table gives the correction which McClendon found must
be applied in order to find the true PH of sea water of any other
salinity :

Salinity of Sea Water in Correction to be Applied to the

Grams of Total Salts Reading of the Colorimetric

per 1,000 Grams of Tube to Find the True PH of

Water. the Sea Water.

30 -}- 0.02

31 4-0.01

32-33 +0.

34 — o.oi

35 — 0.02

36 — 0.03

37 — 0.04

158 MAYOR— DETECTING OCEAN CURRENTS.

The carbon dioxide tension of the sea water was calculated from
the true PH and the temperature according to the ratio determined
experimentally by McClendon, 1917, p. 36. McClendon found that
the PH of sea water normally declines .01 for 1° C. decline in tem-
perature. Thus if the PH is 8.22 at 28° it may be expected to be
8.21 at 2"]° C. I find this to be true under normal conditions, but if
the sea water is diluted with river water rich in CO2 or mingled
with large amounts of up-welling water from the depths, this rela-
tion may even be reversed. Thus in the Tropical Pacific I have
'observed a rise of 0.13 in the PH while the temperature sank 0.45°
C. In general, however, under normal conditions, McClendon's
rule holds good both for the Atlantic and the Pacific.

The salinity of the sea water is expressed in grams of total salts
per 1,000 grams of water, and was determined by titration with
AgNOs, using KoCrO^ as indicator, and standard sea water obtained
from Professor Martin Knudsen for comparison.

The thermometers read to %o° C. and were compared with a
thermometer standardized by the U. S. Bureau of Standards.

The currents were determined by the drift of the ship from her
expected position. Naturally only a decided current could be de-
tected by this crude method, but it was the only one available. The
sea water was dipped up in a glass vessel from the stern of the vessel
and tested at once for hydrogen-ion concentration and temperature,
and a sample was preserved for determination of salinity. Experi-
ments showed that no contamination from the ship could be detected
if the water were dipped up from the stern rather than from the
bow, waste from the sides of the ship being forced away from the
stern by the back wave of the wake.

Summary.

Through the simple process of placing a few drops of the red
dye thymolsulphonephthalein in the bottom of a test tube and filling
the tube with sea water we can determine its hydrogen-ion concen-
tration. The more alkaline the water the more blue-green the color
while relatively acid water gives a yellowish-green color.

In the Tropical Pacific the surface water drifting toward the

MAYOR— DETECTING OCEAN CURRENTS. 159

west is decidedly alkaline, about 8.23 PH, and its carbon dioxide
tension is about 0.0003 atmospheres, and thus about in balance with
the air. Back set currents moving in an easterly direction are, how-
ever, often encountered in the equatorial mid-Pacific especially in
about 5° north latitude. This water is somewhat cooler than that
of the prevailing westerly current, its carbon dioxide tension is de-
cidedly acid, ranging from about 8.1 to 8.18 PH. Thus it comes
from depths wherein the w'ater is at least from 5° to 13° C. colder
than the temperature of the surface of the ocean, and therefore from
at least 100 to 200 fathoms beneath the surface (see chart No. 19,
" Narrative of Challenger Expedition," Vol. i, part 2, p. 758).

The gusty character of the trade winds combined with the gen-
eral up-welling of deep water in the equatorial region may be a
primary cause of this easterly current which consists of deep water
that has come to the surface.

My conclusions support those of McClendon that the PH of sea
water is dependent chiefly upon the temperature, and not upon the
salinity of the water. A fall of 1° C. in temperature normally cor-
responds with a decline of about 0.0 1 in the PH as found by Mc-
Clendon, but this may be altered by local conditions, such as dilution
by relatively acid fresh water, or by the coming to the surface of
cool deep water heavily charged with COo which it discharges into
the atmosphere upon being heated.

The cold shore current along the Atlantic coast of America be-
tween Nova Scotia and Florida is relatively acid in comparison with
the Gulf Stream, being 7.9 to about 8.1 in winter, and its carbon
dioxide tension is lower than that of the air, due possibly to photo-
synthesis maintained by its abundant plant life. It is thus absorb=
ing carbon dioxide from the air. Thus the colder surface waters
are absorbing carbon dioxide, while the tropical regions are probably
setting it free into the atmosphere.

The detection of the sudden and marked change from blue-green
to yellow-green when one encounters an easterly set in the Tropical
Pacific, or passes from a warm into a cold current, can be so easily
made that this method may prove of value in navigation.

160 AIAYOR— DETECTING OCEAN CURRENTS.

Literature Cited.
Bigelow, H. B.

1917. Bulletin Museum of Comparative Zoology, Harvard University,
Vol. 61, pp. 163-357, 2 plates.
Cary. L. R.

1919. Year book of the Carnegie Institution, No. 17, p. 168.
Henderson, L. J., and Cohn, E. J.

1916. Proc. Nat. Acad. Sci., Vol. 2, pp. 618-622, i Fig.
Krogh, A.

1904. Meddelelser om Gronland, Heft 2(i, pp. ZZZ-lZ'i; 409-434-
Mayer, A. G.

1917. Proc. Nat. Acad. Sci., Vol. 3, pp. 548-552.
McClendon, J. F.

1916. J. of Biological Chem., Vol. 24, pp. 519-526. Figs. 1-5. Ih'xd.,
1917, Vol. 30, pp. 265-288.

1916. Medical Review of Reviews, Vol. 22, pp. HZ-l^l^ 14 Figs.

1918. Publication No. 252, Carnegie Institution of Washington, pp. 213-

258, 25 tables, 8 Figs.
McClendon, J. F., Gault, C. C, and MulhoUand, S.

1917. Papers from Dept. of Marine Biology, Carnegie Institution, Vol.

II, pp. 21-69, 24 Figs.
McEwen, G. F.

1910. University of California Publications, Zool, Vol. 6, pp. 189-204,

2 text Figs. Ih'id., Vol. 15, pp. 255-356, 38 plates, Figs. A-C.

1918. Bulletin Scripps Institution for Biological Research, Nov. 8, 1918,

No. 7, 20 pp.
Palitzsch, S.

1911. International Council for Study of Sea, Publications de Circon-

stance. No. 60, 27 pp.
Raben, E.

1910. Wissen. Meeresuntersuchungen Kieler Xomm., Bd. 11. pp. in,
303, 305.
Sorensen, S. P. L.

1909. C. R. Lab. Carlsberg, Kopenhagen, Bd. 8, pp. 1-168,. also Biochem.
Zeitschrift, Bd. 21, pp. 130-304.
Wells, R. C.

1918. U. S. Geological Survey, Professional Paper No. \20-A, pp. 1-16.

MAGELLANIC PREMIUM

Founded in 1786 by John Hyacinth de Magellan, of London

I919

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful Knowledge

ANNOUNCES THAT IN

DECEMBER, 19I9

IT WriL AWARD ITS

MAGELLANIC GOLD MEDAL

TO THE author of THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING no NAVIGATION, ASTRONOMY, OR NATURAL PHILOSOPHY (mERE
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 1919, deliver free of postage
or Other charges, his discovery, invention or improvement, addressed to the
President of the American Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance by some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the same motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author.

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publicly
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as maybe after the adjudication, either in a separate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

^11 correspondence in relation hereto should be addressed

To THE Secretaries of the

AMERICAN PHILOSOPHICAL SOCIETY

No. 104 South Fifth Street

PHILADELPHIA, U. S. A<

Proceedings

OF THE

American Philosophical Society

Subscription — Three Dollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

Price, One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

For Promoting Useful Knowledge

JVew Series, Vol. XXII,
Fart III, 4to, 44 pages.
(Lately Published)
Tertiary Vertebrate Faunas of the North Coalinga Region of Cali-
fornia. A Contribution to the Study of the Palaeontologic
Correlation in the Great Basin and Pacific Coast
Provinces. ByJoHN C. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California.

Subscription— Five Dollars per Volume
Separate parts are not sold

(Address The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY

No. 104 South Fifth Street

PHILADELPHIA, U. S. A.

PROCEEDINGS

OF THE

American Philosopl!ii^al Society

HELD AT PHILADELPHIA

FOR PROMOTING USEFUL KNOWLEDGE
Vol. LVIII. 1919. No. 3.

CONTENTS

PAGE

Recent Discoveries of Fossil Vertebrates in the West Indies and their

Bearing on the Origin of the Antillean Fauna. By W. D.Matthew i6i

The Relative Contribution of the Staple* Commodities to the National

Food Consumption. By Raymond Pearl . . . . .182

Self-Luminous Night Haze. By E. E. Barnard ..... 223

Graphical Representation of Functions of the nth Degree. By Francis

E. NiPHER ........ ... 236

PHILADELPHIA

THE AMERICAN PHILOSOPHICAL SOCIETY

104 South Fifth Street

1919

The American Philosophical Society

announces that an award will be made in the year 1921 of the

HENky M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

The Control of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent, of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after the
award.

Attention is called to the following regulations governing the award
of the Prize :

Competitors for the prize shall affix to their essays some motto or name
(not the proper name of the author, however) and when the essay is forwarded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judges, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essays shall be in their author, subject to
the right of the Society to publish the crowned essay in its "Transactions "
or " Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,
David Jayne Hill,
Simeon E. Baldwin,
John Cadwalader,
W. W. Keen,
William B. Scott,
President ex-officio.

Committee.
The essays must be sent, addressed to the
President of the American Philosophical
Society, No. 104 South Fifth Street, Phila-
delphia, U. S. A.

RECENT DISCOVERIES OF FOSSIL VERTEBRATES IN

THE WEST INDIES AND THEIR BEARING ON THE

ORIGIN OF THE ANTILLEAN FAUNA.

By W. D. MATTHEW.

(Read April 25, 1919.)

Introduction. Interest of West Indian Fossils — Review of
Earlier Knowledge.

Ten or twelve years ago almost nothing was known of the ex-
tinct vertebrates of the West Indies. There was good reason to
suppose that a considerable fauna had existed, and that if found it
would be a very interesting one. One would expect to find peculiar
insular types, different from the mainland faunas, and the affinities
of the various types would provide evidence as to former land con-
nections and other interesting problems. The modern fauna of the
islands has undoubtedly been greatly modified by man, both before
and since its settlement by white men. Such indigenous mammals
as existed have mostly disappeared in consequence, and with them
many of the lower vertebrates and invertebrates. And many ani-
mals have been introduced either by intention or by accident. Ca-
promys, Plagiodontia and Solenodon, two peculiar rodents and a
very peculiar insectivore, are the only surviving mammals that are
certainly indigenous. The lower vertebrates are more numerous,
but many of them also have been exterminated, and others intro-
duced since man arrived on the islands.

The climate and geology of the islands limit the prospects of dis-
covery of fossil land faunas to cave and stream or spring deposits
of Pleistocene age. The underlying sedimentary formations are
practically all marine, chiefly Tertiary limestones and an older
metamorphosed series which seem to be chiefly altered volcanic ma-

PROC. AMER. PHIL. SOC, VOL. LVIII, K, July I9, 1919.

161

162 MATTHEW— RECENT DISCOVERIES OF

terial. The freshwater and Httoral sediments of continental areas
are unknown in the West Indies, even in the larger islands (with
one or two unimportant and doubtful exceptions) and it is from
these that our fossil records of land animals are almost wholly
derived. There was, and is, therefore, very little prospect of dis-
covering remains of the Tertiary or pre-Tertiary land faunas of the
West Indies. The possibilities are practically limited to the Pleis-
tocene.

Up to a few years ago two important discoveries had been made.
One was the giant rodent Amblyrhiza found in cave-breccia on the
island of Anguilla, and described by Cope in 1869. The affinities of
this animal, about the size of a capybara, have been disputed. Cope
regarding it as allied to the South American chinchillas, while J. A.
Allen considered it as a relative of the extinct Castoroides or " giant
beaver" of North America. Dr. Allen's view, although mistaken,
seems to have been more generally accepted. The other discovery
was a ground-sloth jaw and some other remains found near Cien-
fuegos in Cuba, and described by Dr. Leidy under the name of
Megalocniis, as a near relative of the North American ground-sloth
Megalonyx. For reasons that I shall explain later the Cuban dis-
covery was little known and more or less discredited in scientific
discussion. The Megalocnus was supposed to be identical with
Megalonyx, but many doubted whether the specimen really came
from Cuba. Such as it was, the evidence from fossils seemed to
point strongly to a land connection with North America as late as
the Pleistocene, extending at least as far as Anguilla. This no
doubt played its part in the very prevalent belief in an Antillean
continent, united with the greater continents to north and south of
it, destroyed only in times geologically very recent.

The characters of the modern Antillean faunas were very diffi-
cult to reconcile with this hypothesis. Their insular character is
very marked, whatever theories might be entertained as to the origin
of the various animals. And what was known of the geology did
not at all accord with the view that they were fragmented remnants
of a great continent recently broken up. A diligent search for
fossil remains and a more thorough and systematic study of the
geology was very much needed.

FOSSIL VERTEBRATES IN THE WEST INDIES. 163

The Porto Rico Discoveries.

With these and other considerations in mind, the New York
Academy of Sciences decided in 191 3 to undertake a systematic
survey of the geology and natural history of Porto Rico, in coopera-
tion with the local government and with the American Museum of
Natural History. This was very ably conducted under direction of
Dr. Britton. The geology was quite thoroughly investigated by
Drs. Berkey, Reeds and others and the reports afford a better
knowledge and understanding of the geologic history of this island
than we have of any of the others. The marine limestone yielded
one interesting fossil mammal, a sirenian or sea-cow allied not to
the manatees as one would expect, but to the old-world dugongs.

The most important find in the way of land animals was made
in the course of investigations in a cave near Utuado, where human
remains probably prehistoric were unearthed, and beneath them re-
mains of several kinds of extinct mammals. This stimulated further
search of the numerous caves in the island, undertaken by Mr.
Anthony for the American Museum of Natural History, and the
results of his very successful work have been published in a well-
illustrated memoir issued last December.

The fossil mammals include one insectivore, one edentate and a
number of rodents, all new and very peculiar types, widely different
from any known animals ; also a number of bats closely allied to the
living bats still on the island. jMost of this fauna is represented by
well-preserved skulls and various bones of the skeleton.

The insectivore Ncsophontcs is so distinct that it has to be placed
in a family by itself. It has some distant affinities with the Sori-
coidea (moles and shrews), but is a very primitive type, and its
nearest allies are perhaps to be found among the Eocene insectivora
of North America. There is no suggestion of relationship to Sole-
nodon, the only other known insectivore of the West Indies, nor to
the extinct Necrolcstes of South America, the only insectivore
belonging to that continent. Like the Eocene insectivores it retains a
great deal of the primitive tri-tubercular type of dentition which
nearly all — if not all — the primitive mammals once possessed.

The edentate, Acratocnus, is not less interesting. The skull is
superficially much like the two-toed tree sloth Choloepus, with the

164 MATTHEW— RECENT DISCOVERIES OF

same type of triangular tusks and is of quite moderate size. Its
nearer affinities, however, are with the great extinct ground sloths.
It belongs to the same family as Megalonyx of the North American
Pleistocene, but is not very closely related to it, and is very much
smaller and in some respects more primitive. Although thus related
to a North American genus of ground sloths Acratocnus points to
a South American origin, for Megalonyx is known to be an immi-
grant type in the north, derived from South American ancestors in
the Miocene.

The rodents all belong to the Hystricomorph or porcupine sec-
tion of this order, which has its headquarters in South America.

The largest of them, Elasmodontomys, is closely related to Ajh-
hlyrhisa, the gigantic extinct rodent of the island of Anguilla.
Another smaller genus, Heptaxodon, also belongs in this neighbor-
hood, but is less closely related. These are put into the family
Chinchillidse, but their nearest affinities are with the extinct Mega-
mys and Tetrastylus of the Pliocene of Argentina.

A second group of rodents, Heferopsomys and Homopsomys, is
related, but rather distantly, either to the living agouti, Dasyprocta,
or the spiny rats of South America ; it has also, as will appear, some
extinct relatives in Cuba and Hayti.

A third extinct rodent from Porto Rico, Isolohodon, is nearly
related to Plagiodontia, 3.n eximct or almost extinct rodent of Hayti,
and more distantly to the hutias (Capromys) still living in Cuba
and Jamaica. Like the preceding group it is rather distantly related
to any continental rodents ; it is referred to the Octodontidae.

That is the full extent of the mammal fauna of Porto Rico. Its
affinities are chiefly South American, but distant, pointing to long
isolation. The near relationship to the extinct Anguillan rodent is
significant. The bank of shallow water which extends eastward
from Porto Rico includes the Virgin Islands; and Anguilla with the
adjoining islands also stands on a shallow bank of considerable
extent. Vaughan has shown the probability of part or all of these
banks being above water in the Pleistocene. But between the Virgin
Islands and the Anguilla group there is a narrow trench of very
deep water. Possibly this trench is due to faulting of recent origin
and the islands were connected in the Pliocene or Pleistocene. The

FOSSIL VERTEBRATES IN THE WEST INDIES. 165

alternate is to suppose that Amhlyrhiza is a gigantic descendant of
some small rodent that got drifted over the intervening sea barrier.
Anguilla, it must be remembered, is but a small remnant of an island
of very considerable size, as indicated by the extent of the sur-
rounding shallow bank.

The Cuban Discoveries — Ciego Montero — Casimba — Cave

Fossils. The Fauna — Insectivora — Edentata — Ro-

dentia — Reptilia.

Ciego Montero. — The first discovery of fossil mammals in Cuba
was made in i860, in the early days of American palaeontology. A
lower jaw was found in a warm spring at Ciego Montero, a few miles
from Cienfuegos, with other fragmentary fossils. This jaw was
exhibited at the Paris Exposition in 1867, and is now, I beheve, at
Madrid. A drawing of it was sent to Joseph Leidy, who imme-
diately recognized it as a relative of Megalonyx and named it Mega-
locnus rodens on account of the peculiar position of its tusks which
simulated those of the rodents. In 1865, 1871 and 1875 De Castro
published articles on extinct animals in Cuba, in which he figured
this jaw and also described supposed fossil remains of Equiis and
Hippopotamus, and contended that these proved that Cuba had for-
merly been a part of the North American continent. These asso-
ciates unfortunately tended to discredit the whole paper, especially
the hippopotamus tusks. They did not appear to be very old, and
were regarded as probably post-Columbian. The Megalocmis jaw
was confused with the true Megalonyx and the suggestion was made
that it was probably brought over from Central America.

Carlos de la Torre, professor of zoology at the University of
Havana, was well aware that the Megalocmis had really been found
in the Ciego Montero spring, and from time to time, as his duties
permitted, made further investigations to see if something more
could not be found at this locality or elsewhere. He succeeded in
finding a few additional specimens here and better material at some
other localities, and in 1910 published a notice of these discoveries.
The specimens were exhibited by him at the International Congress
at Stockholm and Gratz, and were deposited at the American Mu-
seum for further study and full publication in a memoir by de La

166 MATTHEW— RECENT DISCOVERIES OF

Torre and myself. The following year i\Ir. Barnum Brown, on
Professor de La Torre's invitation and with his aid, made a more
thorough exploration of the Ciego ]\Iontero spring, and secured a
large collection. In 191 8 Mr. Brown completed the exploration of the
deposits around the spring, securing much additional material. Pre-
liminary notices of these collections have been published, the full
descriptions being postponed until all the available material had been
secured. The exploration of the deposit was a rather difficult
matter as the spring is a powerful one and the water had to be
pumped out and drained away by means of a gasoline pump, and
the spring openings cemented up, before the deposit around it could
be thoroughly explored.

The Ciego Montero collections consist chiefly of bones of Mega-
loams and Crocodilus, plates of a giant tortoise and a terrapin, and
a few remains of other ground sloths, of rodents, lizards and birds,
and a good many bones of small amphibians.

The Casimha. — Another important locality discovered by Dr. de
La Torre was in the Sierra de Jatibonico, in the central part of the
island, a fissure-spring at the bottom of a ravine where there was a
considerable deposit with fossil bones. These correspond with the
remains found at Ciego IMontero, except that the smaller ground
sloths were relatively more common. The material is equally well
preserved, but not nearly so much of it. This locality has also been
worked out.

A number of similar fissure springs were examined by Dr. de La
Torre and a few fragmentary fossils secured, but nothing of impor-
tance. Probably deposits of this character may be discovered in
other parts of the island, and some further discoveries may be made
in this way.

Cave Deposits and Kitchen Middens. — There are also many caves
in the limestone formations of Cuba, in all parts of the island. A
considerable proportion of these contain guano or similar phos-
phatic deposits which have for many years been dug out and used
for fertilizer. A few of them have recently been explored for fos-
sils, and some very interesting material secured. The old Indian
kitchen-middens, some in the caves, some elsewhere, have been
more or less exploited for prehistoric human remains by the anthro-

FOSSIL VERTEBRATES IN THE WEST INDIES. 167

pologists, and incidental to this work a few remains of extinct ani-
mals have been secured and described by Dr. Gerrit S. Miller, of
Washington. Dr. Glover M. Allen has also published a note on
some interesting cave fossils secured by Dr. Thomas Barbour. The
largest cave collection yet secured was made by Mr. Barnum Brown
last winter in a deposit discovered by Dr. Barbour and explored by
us on his invitation.

These cave faunas consist chiefly of the remains of the small
animals, rodents, insectivores, bats, etc., which are rare in the spring
faunas. Remains of the ground sloths are rare, but they do occur.
Quite recently Mr. Harrington, in exploring certain caves at the
far eastern end of the island for the Museum of the American In-
dian, secured some remains of extinct animals that warrant further
search. They were chiefly Mcgalocmis, showing that this animal
ranged through the whole of Cuba in the Pleistocene.

The fauna of these cave and spring deposits is a very interesting
one. Fortunately the two supplement each other, so as to give a
pretty full representation both of the large and of the small animals.
Each consists of many hundreds of jaws, with a proportionate num-
ber of other bones. It is fair to conclude that it gives us a pretty
good line on the characteristic mammals of the Cuban Pleistocene.
Further discoveries may bring other animals to light, but as the
same types recur in each deposit, common in one, rare in another,
it does not seem likely that they will make any great additions to
the fauna save through genus — and species — splitting, analogous to
what has been going on among modern mammals for the last few
decades. For obvious reasons it is better for palaeontologists to be
conservative in this respect — even at the risk of being stigmatized
as " lumpers."

Insectivora. — The existing Solenodon has not been found fossil.
This is not surprising, as it is very rare now, and may have been
rare during the Pleistocene. The extinct Nesophontes of Porto
Rico is represented in Cuba by a much smaller species distinct in
various particulars. It has been described by Dr. G. M. Allen as
Nesophontes micrits. Mr. Brown secured a great series of jaws,
etc., of this species in his cave collection, but only a single jaw and
a tooth at Ciego Montero.

168 MATTHEW— RECENT DISCOVERIES OF

Edentates. — The edentates are all ground sloths related to Mega-
lonyx, but belong to four distinct genera. The largest and most
abundant is Megalocnus, about the size of a black bear. The tusks
are flattened into a meniscus cross-section, and set near together,
especially in the lower jaw, so that it has a curiously rodent-like
effect. The cheek teeth are exceptionally long, and in consequence
the palate is depressed below the level of the cranium, and the jaw
extremely deep. This is like the gigantic Megatherium, but the
form of the cheek teeth is almost exactly as in Megalonyx. De-
tailed comparisons show that the Cuban genus is most nearly related
throughout to Megalonyx.

Mesocnus is a smaller and more primitive genus, which is about
the size of the Miocene ground sloths of Patagonia, and like them
has a long tongue of bone projecting forward between the tusks.
The cheek teeth are not so long in this genus, but they have almost
the same form as in Megalocnus or Megalonyx, different from the
more primitive form of the cheek teeth in all the Miocene Mega-
lonychidse. The tusks are rather small and not so broad or menis-
coid in section as in Megalocnus. On the whole this is a more
primitive genus, except the humerus, which has lost the entepicon-
dylar foramen, as in Megalonyx.

Miocnus is of about the same size as M^.jocn«.y but quite distinct.
It is very closely related to Acratocnus of Porto Rico, perhaps the
same genus. It has a short wide jaw and stout triangular tusks
with a short, bony tongue between.

Microcnus is the fourth genus of ground sloths, an animal about
the size of a cat, smaller than the living two-toed sloth. It resem-
bles Megalocnus in the form and position of the tusks, but the
molars are not so long-crowned and differ considerably in form,
with a marked suggestion of the square cross-section of the INIega-
theriidcC. I have not seen the skull or upper teeth ; a couple of foot
bones doubtfully referred to this genus come nearer to the living
tree-sloths in proportions than they do to any of the large ground-
sloths. It would be very interesting to know more about this little
animal.

There is so much individual and age variation in ground sloths
that it is difficult to say how many species of these genera are prcs-

FOSSIL VERTEBRATES IN THE WEST INDIES. 169

ent. Megalocnus occurs abundantly both at Ciego Montero and the
Casimba, but apparently only one species at Ciego Montero, while
at the Casimba there may be two or three. The species found at
the eastern end of the island agrees better with the Casimba forms
than with the Ciego Montero species. Of the smaller forms there
are clearly two species of Mesocmts, but I see no proof of more
than one of Miocnus or Microcnus.

All these Cuban ground sloths, along with the Porto Rican genus,
are most nearly related to Megalonyx among the continental genera,
and may be regarded as co-descendants with it of Eucholccops and
Megalonychotherium of the Patagonia Miocene. Microcnus is ap-
parently the least closely related ; and the relations between Miocnus
and Acratocnus are very close.

Rodentia. — The fossil rodents are all hystricomorphs related to
the South American rodents, but not closely related. There are
two groups of them (in a broad sense) each with two or three spe-
cies. One is still common on the island, the Hutias with two very
closely related genera, Capromys and Geocapromys. Of these Ca-
promys is common in the latest cave deposits, but in the older levels
of the cave material only Geocapromys. The species in the older
levels are all small — some are smaller than any surviving species.
The large species appear only at the top. The other group is en-
tirely extinct ; it consists of three or more species apparently of a
single genus which Mr. Miller has called Boromys. It is related to
the spiny rats of South America, but not closely. Capromys on
the other hand has a near relative in Venezuela Procapromys.

None of the Chinchillidse of the eastern Antilles have been found
in Cuba.

Remains of Geocapromys occur rarely in the Ciego Montero and
Casimba collections ; they are very abundant in the cave collections.
Boromys, although common in the .caves, has not been found at the
Casimba or Ciego Montero. Dr. Barbour tells me that there is
some reason to believe that this genus still survives in Cuba, although
it has not found its way into any scientific collections.

Reptiles. — In the Ciego Montero collection is a series of fine
skulls of crocodiles, which Dr. Barbour regards as all growth stages
of the living Cuban crocodile C. rhomhifer. This species has a

170 MATTHEW— RECENT DISCOVERIES OF

broad head, and in some other characters approaches the alHgator.
It is closely allied to the alleged Central American species C. moreletli.

The most interesting of the reptiles is a giant tortoise resem-
bling in several respects the living species of the Galapagos islands.
It has a very thin shell, like most of the species of oceanic islands,
but aside from this, which is probably due to parallelism, it shows
some other points that are suggestive of a true relationship,
although not a close one. It has one or two points suggestive of
relationship to the South American tortoise T. tabulata, but mostly
differs widely from it. Comparison with the fossil tortoises of the
North American Tertiary does not indicate any special relationship,
although it may be regarded as a descendant of some species of this
group. It has one very curious character, unique in so far as I
have made comparisons, in that the margins of the horny shields
are marked on the plates not by furrows but by sharp, raised crests
over the greater part of the carapace. This alone would forbid any
close relationship with any of the species I have compared.

Leidy named this species Testudo cnhensis in 1868, from a
broken plate of the carapace that had been sent to him.

The terrapin is probably identical with the living Cuban terra-
pin, which is regarded by some as a distinct species, by others as a
subspecies of the yellow-bellied terrapin C//ryi'(»;;;3'.y (or Trcc/j^^wzy^)
scripta (=scabra) of the southeastern states. It is at all events
closely related, and belongs to a group which has several fossil rep-
resentatives in the Pleistocene of Florida and other states of the
southeast.

Some fragmentary remains of lizards, snakes, birds and am-
phibians occur in the collections, but have not yet been studied. I
doubt whether very much of interest can be got from them, as they
seem to be very doubtfully identifiable.

Hayti, Jamaica and the smaller islands are as yet unexplored
territory. Except for the Amhlyrhiza of Anguilla and a few frag-
ments described by Dr. Miller from an Indian kitchen midden at
S. Pedro de Macoris, in Hayti, nothing is known of their extinct
fauna. Limestone caves are numerous, and will undoubtedly fur-
nish important finds in the near future. A primitive sirenian, Pro-
rastomiis, was described many years ago by Richard Owen from the

FOSSIL VERTEBRATES IN THE WEST INDIES. 171

older Tertiary marine limestones of Jamaica. It is distantly related
to the manatees, but it throws no light upon the problems here con-
sidered, and may be passed over without discussion.

Review of the Fauna an now known. Its incomplete and

INSULAR character. UNIFORMITY THROUGHOUT THE LARGER

ISLANDS. Probable derivation of the several groups.
Methods of colonization.

So much for the new discoveries. We may now review the
fauna as a whole, living and extinct, and consider its general char-
acter and what bearing it has upon the past history of the islands.
I shall deal principally with the mammals, partly because I am best
acquainted with them, partly because the new discoveries are chiefly
in this group, and partly because I think they afiford the best evi-
dence upon certain critical points.

The most obvious features to me in the mammal fauna are its
incompleteness and insular character. There are no perissodactyls,
no artiodactyls, no proboscidians, no carnivora, no primates, no mar-
supials, no rodents except three groups of the Hystricomorphs
which are an especially South American group, no insectivores save
for two types which have very remote affinities to any living groups,
none of the various types of edentata, except a single group of
ground sloths. Certainly this poverty of mammalian fauna is not
to be explained by scantiness of material. We have very large col-
lections, many hundreds of jaws and a due proportion of other
bones, from two types of deposit, a cave and a spring fauna. Cor-
respondingly large collections from cave or spring faunas in North
America or South America yield very different results. Compared
with the Port Kennedy, Conard Fissure or the California cave
faunas, or those described by Lund and Winge from Brazil, the
contrast is very striking. Spring faunas may be less varied, but
whatever else is absent the larger ungulates are sure to be found in
such situations.

The whole Antillean mammal fauna consists of only six types
or groups of sub-family rank at most, as follows :

1. Ground sloths allied to Megalonyx — 5 genera.

2. Rodents of the Capromys group — 4 genera.

172 MATTHEW— RECENT DISCOVERIES OF

3. Rodents of the Boromys group — 2-4 genera.

4. Rodents of the Amhlyrhiza group — 3 genera.

5. Nesophontes ] ......

^ ^ , , kwo very aberrant and primitive msectivores.

o. 6olenoaon J

These six groups constitute the surely native fauna. Most of
them are extinct; all of them except Solenodon are found as fossils.
Sixteen genera and about thirty-five species.

There are various other mammals recorded as from one or an-
other West Indian island. Some of them are known to have been
introduced by man, others may have been. They are all species
identical or closely related with continental species. If any of these
were not introduced by man, they must have reached the islands in
comparatively recent times, geologically speaking, not earlier than
the Pleistocene, and some quite surely post-Columbian. Some are
found fossil, indeed, but never much altered by petrifaction, in the
uppermost levels of the cave and spring deposits. Such animals
as the European rats, domestic cat, domestic pig, etc., belong ob-
viously to post-Columbian time; the agoutis and armadillos of the
Windward Islands and other continental American species are
probably introduced by man, but they may not all be post-Columbian.
Such species as the Bahaman and Martinique raccoon, the Jamaica
rice rat, etc., described as distinct species, but certainly very closely
related to continental species, may also have been brought by man,
but long enough ago to have developed distinct races which have
been accorded the rank of species. Some may indeed have come
by natural means, and if so they can only have come through over-
seas drift, but for the present it is best to limit the discussion to the
purely native groups, all of which are genera, subfamilies or fami-
lies peculiar to the islands.

The first group, the ground sloths, consists of four very distinct
genera allied to the Pliocene and Pleistocene North American Mcga-
lonyx and descended from the South American Miocene Eucho-
Iceops. No Megalonychidse are found in the South American Plio-
cene or Pleistocene ; they are replaced by more progressive and
specialized families. From th eevidence one may infer that they
reached the Antilles at the end of the Miocene or beginning of the
Pliocene, and that the four genera specialized there in adaptation

FOSSIL VERTEBRATES IN THE WEST INDIES. 173

to the insular conditions. Their peculiar specializations and the
absence of any of the later and more progressive types of ground
sloths that spread all over the North American continent involves
isolation since that time, from North, South or Central America.

The second group, the rodents of the Capromys type, affords
some contrast in its scope and relations. Instead of having four or
five very divergent genera quite wide apart structurally, we have
in the western Antilles, two closely allied genera, much closer in the
skeleton than they are superficially. They are the only native
mammals on the islands that are not on the verge of extinction ;
under natural conditions they would apparently be a very pros-
perous group. The cave records indicate they were rapidly progress-
ing in size and increasing in relative numbers in the Pleistocene.
And finally Procapromys of Venezuela, their nearest continental
relative, is quite closely related. They have not had time to diverge
very far from it in structure, although the record shows that they
were diverging. The surface inference is that this group was a
comparatively late arrival, perhaps early Pleistocene, perhaps late
Pliocene, clearly much later than the ground sloths. Yet they had
reached several of the western islands — Jamaica, Cuba, one of the
Bahamas, even Little Swan Island, a small islet off the coast
of Honduras. In Cuba- several species of Capromys now exist.
Geocapromys was formerly the principal or only type in Cuba, but
is now found only in Jamaica, Little Swan Island and Plana Key
on the Bahamas. In the eastern islands we find two genera less
closely related. Isolohodon of Porto Rico is extinct and Plagio-
dontia of Hayti is practically extinct ; these two are less closely re-
lated to the two western genera, but are regarded by Mr. Anthony
as the eastern representatives of the group. Systematic exploration
in the caves of Hayti would probably clear up the true relationship
of these eastern genera ; casual sketchy exploration or work under-
taken simply to get material and describe new species, is very likely
to destroy such evidence as exists on their palasontologic history.

The third group, the Boromys group of rodents, is extinct (pos-
sibly a single survivor) and includes a number of species belonging
to two very closely related genera, doubtfully separable, one found
at several points in Cuba, the other from Hayti. It is rather dis-

174 MATTHEW— RECENT DISCOVERIES OF

tantly related to the spiny rats of South America (Echinomys).
Heteropsomys of Porto Rico is rather doubtfully related to this
group; certainly the affinities are distant. Jamaica is unknown
palseontologically.

The fourth group, the rodents of the Amblyrhisa group, is lim-
ited to Porto Rico and Anguilla, that is to say, to the eastern end
of the east-west chain, and is w^holly extinct. Amhlyrhiza, the
largest, is quite gigantic for a rodent, about the size of the Capybara
or of the extinct Castoroides. Elasmodontomys of Porto Rico is
nearly the size of a beaver. Heptaxodon of Porto Rico is a smaller
form. These West Indian chinchillids are only distantly related
to the living Chinchillidse of South America, but they are quite
nearly related to the Megamys group of the South American
Pliocene.

The fifth and sixth groups of mammals, the two insectivora, are
the most isolated of all the West Indian mammals, each having a
family to itself. They are not at all related to each other, and their
nearest relatives appear to be certain imperfectly known Eocene and
Lower Oligocene genera of North America. There are no autoch-
thonous insectivores at all in South America, save for the curious
little Necrolcstcs of the Miocene, which may have had something to
do with the Chrysochloridse, but certainly not with these West Indian
genera. On the other hand, insectivora occur in all the Tertiary for-
mations of North America, and were more abundant and varied in
the older Tertiaries. Micropternodus, of the Lower Oligocene, may
have been a distant ancestor of Solenodon, and several of the Eocene
soricoid genera from the Bridger show much resemblance to Neso-
phontes. But any exact relationship is to be regarded as provi-
sional, and by no means proven.

Turning now to the reptiles, we find that the entire order of
chelonia is represented to-day by a single species of terrapin widely
distributed through the islands, found fossil in Cuba, possibly also
in the island Sombrero, and closely related to the yellow-bellied
terrapin of the United States. It belongs to a North American
group, its nearest fossil relatives are in the southeastern states, and
the few terrapins of Central and South America, which are prob-
ably rather late immigrants from North America, are less closely

FOSSIL VERTEBRATES LN THE WEST INDIES. 175

related to the West Indian species. All this points strongly to
Florida as the source, and the Pleistocene as the time of arrival.

A giant tortoise, new extinct, is found in Cuba. Unlike the
terrapin, it is only distantly related to any other tortoises, somewhat
nearer to the species of the Galapagos islands than any other. It
does not appear to be especially related to any North American tor-
toises, except that it may be descended from some primitive species
of the Miocene. True tortoises are found in the Pliocene and Pleis-
tocene of South America, but are believed to be immigrants from
the north. The Galapagos islands affinities suggest that this species
may be derived from the unknown Tertiary fauna of Central Amer-
ica. But the Cuban species must have been isolated a long time, at
least since the beginning of the Pliocene, one would judge from its
peculiar specialization.

The crocodiles of the West Indies are two, one, C. rhomhifer,
peculiar to Cuba. The fossil species in Cuba is C. rhombifer, a
broad-headed alligator-like form. It is said to be nearly related to
the alleged C. morclctii of Central America. The other species, C.
americanus, is found in Florida, Mexico, Central America and the
west coast of South America as far as Ecuador, and appears to be
rather widely distributed in the West Indies. It is probably signifi-
cant that this widely distributed species inhabits the salt marshes,
whereas the more local C. rhombifer occurs in fresh water.

The smaller reptiles and amphibians, fresh water fish, and the
various groups of invertebrates are much more widely distributed
throughout the West Indies. Their distribution and its causes have
been extensively discussed by various authors, but unfortunately
without giving due weight to two features in which it differs from
the mammals and the two groups of large reptiles.

First that the dispersal of these lower groups on oceanic islands
is probably chiefly brought about by storms. The eggs, attached to
small debris, or even the adult animals, if small enough, will often
be picked up by violent storms, and carried for considerable dis-
tances. A cyclonic storm or tornado will sometimes partly empty
a shallow pond, carrying the materials a mile or more into the air,
the lighter and smaller debris being carried along for a great many
miles before it comes to the ground again. On account again of

176 MATTHEW— RECENT DISCOVERIES OF

their small size they would not be crushed by the fall. The distri-
bution of the infra-mammalian groups in the West Indies, as on all
other oceanic islands, appears to me to be much more in accord with
this method of transportation than with any continental bridge
theories or current borne drift. Mammals, however, could hardly
be successfully colonized in this way, nor would it serve for chelo-
nians or alligators. But it does probably apply to most birds and
nearly all bats.

The second feature in regard to the infra-mammalian groups is
that we know little or nothing of their past distribution. From such
slight direct evidence as we have, and from the analogies of past
distribution of the mammals, we can be reasonably sure that it was
not the same in the later Tertiary as it is now, and on the same evi-
dence I am confident that the methods of inferring past from pres-
ent distribution generally used by zoogeographers are erroneous.
If applied to mammals such methods would lead to conclusions ab-
surdly in conflict with the known facts ; they are in conflict with the
few available data among the lower land animals ; and they appear
to me to be based upon erroneous theoretical reasoning.

Summarizing the data briefly we have among the mammals:

1. The ground sloths, probably of South American origin (but
possibly via Central America) which must have arrived about the
late Miocene or early Pliocene, and demand isolation since that
date to account for their diversity, archaic type, and absence of later
developed and more progressive relatives.

2. The rodents, all Hystricomorphs broadly of South American
affinities, but including three groups.

(c) Chinchillids, limited to the eastern islands, and most diverse
and peculiar. They also have come from South America, but from
the eastward, in the late Miocene or Pliocene.

(b) Hutias, including an eastern and a western group (but the
division line is not the same as with the chinchillids), the western
group with one near relative in Venezuela, the eastern genera (Hayti
and Porto Rico) more isolated. I have formulated several difter-
ent hypotheses to account for the distribution, but none can be
sufficiently tested by facts to be of much account. The clue lies, I
think, in a knowledge of the Pliocene and Pleistocene fauna of Cen-

FOSSIL VERTEBRATES IN THE WEST INDIES. 177

tral America, but the eastern representatives may have come from
South America independently at an earlier date. Too few data to
go upon. (See addendum, p. i8i.)

(c) The Boromys group, known as yet very imperfectly, appar-
ently limited to Cuba and Hayti. These also are quite distant from
any continental forms, suggesting. their having been isolated since
the Pliocene, but so far as we know this is not supported by diver-
sity of genera or insular specializations as among the ground sloths.
The relationship to this group of the Porto Rican Hetcropsomys is
in dispute, and it is well to defer conclusions.

3. The insectivora are much more isolated than any of the pre-
ceding, but their relations are, although remotely, North American.
It is with the early Tertiary fauna of North America that their
apparent affinities lie.

4. The giant tortoise has distant relations with the species of the
Galapagos islands, but has evidently been isolated and insular for a
long time, since the Miocene or early Pliocene, but not earlier than
that. The terrapin, on the other hand, is quite closely related to a
living species of the southeastern states and to a group of Pleisto-
cene species in the same region. It can hardly be supposed to be
older on the islands than the Pleistocene, especially since the sev-
eral islands in which it is found have not developed even distinct
subspecies, although there is a good deal of individual variation.

5. The Cuban crocodile is common in the Pleistocene and prob-
ably reached the island somewhat earlier ; it is nearly related to a
Central American species, but it is quite as likely as not that that
species was more widespread in the Pliocene, perhaps through the
Southern States. (See addendum, p. 181.)

6. Birds, bats, lizards, snakes, amphibians and fresh water
fishes, molluscs, insects and various other groups of invertebrates
are more widespread on the islands, as they are generally on oceanic
islands. I do not think the reason is so much their greater geologic
age as it is their greater facilities for dispersal, especially through
storms, as I have indicated. Generally speaking there are in each
group forms allied closely to modern mainland forms and others
which are more isolated, primitive and archaic in type, usually more

PROG. AMER. PHIL. SOC, VOL. LVIH, L, JULY 25, I9I9.

178 MATTHEW— RECENT. DISCOVERIES OF

sedentary in habit or otherwise unfavorably situated for dispersal.
Time is lacking to discuss these further.

Two questions are involved in the interpretation of this fauna.
First as to former connections between the islands themselves,
second as to former connections with the mainland.

There is considerable to be said in favor of Pleistocene or late
Pliocene connections between the greater Antilles, extending as far
east as the Anguilla group. In favor of this are the presence of
representative species or genera of mammals as follows :

Nesophontes on Cuba and Porto Rico.

Solenodon on Cuba and Hayti.

Miocmis on Cuba and the closely allied Acrafocnus on Porto
Rico.

Boromys and Brotomys on Cuba and Hayti.

Capromys and Geocapromys on Cuba, Jamaica and one island
of the Bahamas, also on Little Swan Island.

Isolohodon and Plagiodontia on Porto Rico and Hayti.

Amhlyrhisa on Anguilla and Elasmodontomys on Porto Rico.

Against such connections are to be cited

(i) The absence of three out of the four ground sloth types from
Porto Rico.

(2) Limitation of the Capromys group proper to the western
islands, of the chinchillids to the eastern islands.

(3) Representative or replacing types are in some cases closely,
in others much more distantly related as between the various islands
of the chain. No specific sequence of geographic separation can be
formulated that will fit the known facts. In some cases it would
call for an earlier separation between Cuba and Hayti, in others
between Hayti and Porto Rico.

On the whole I feel that this question is better left open till we
get more data, especially as to the fossil faunas of Hayti and
Jamaica.

As to the second question, of continental connections, it seems
to me that the weight of evidence is very heavily against it. There
are two broad considerations affecting the matter. First, the geo-
logic evidence and the submarine topography, while they do not
forbid certain former connections, are on the whole unfavorable to

FOSSIL VERTEBRATES IN THE WEST INDIES. 179

any. The islands are not the remnants of a former Antillean con-
tinent ; on the contrary they have been built up, partly by uplifting-
and uptilting of blocks along fault-lines, partly through volcanic
eruptions, largely submarine, along the lines of weakness and fault-
ing. They are not very old geologically, the oldest known rock
being the marine Jurassic of western Cuba ; nearly all if not all of
the metamorphic rocks mapped by the U. S. Geological Survey as
Palaeozoic are either certainly or probably Cretaceous ; and the lesser
Antilles are the youngest geologically. A land connection with
Florida is practically forbidden by the geology ; a land connection
via the lesser Antilles is unlikely in the later Tertiary, and very un-
likely earlier ; a land connection via Hayti, Jamaica and Honduras
is unobjectionable, but there is no particular evidence in the geology
or submarine topography to show that there zvas such a connection
in the Tertiary. There is, however, considerable to indicate that
the relatively shallow banks that stretch eastward from Honduras
and Nicaragua almost to Jamaica may have been partly or wholly
out of water about the Pliocene or Pleistocene. In short, if a con-
tinental connection is required by the faunal evidence, this is by all
odds the most likely place for it.

The second general consideration is the incomplete character of
the fauna. It is not a continental fauna, either North or South
American, but an insular fauna arisen from development in isola-
tion of a few individual elements of each. Such a fauna is not the
result of invasion from North or from South America over a con-
tinuous land area. Its incompleteness can no longer be ascribed to
extermination by man, for we see that the Pleistocene fauna, while
much more extensive, was obviously and significantly incomplete
and insular in type. It cannot be ascribed to scanty fossil material,
for we have now very large collections of both cave and spring fos-
sils. It cannot be ascribed to drowning out of the fauna, for we
have surviving on the islands two types of Insectivora that date
back to the middle or early Tertiary.

A more particular examination of the different groups of mam=
mals, etc., confirms this general conclusion. The several groups
indicate derivation from different sources geographically, and at
different times geologically. If we insist upon the need of conti-

180 MATTHEW— RECENT DISCOVERIES OF

nental bridges to account for the \\'est Indian mammals, it is not
one but several bridges that they would indicate, each lasting only
long enough to permit the passage of one or two of the least migra-
tory elements of the continental faunas, and the later ones so
arranged as not to disturb the isolation of the types which had
already reached the islands. If continental connections be de-
manded, the chinchillids call for a Pliocene bridge via the lesser
Antilles, not extending west of Porto Rico. The Capromys group
calls for a Central American bridge in the late Pliocene or Pleisto-
cene, while the ground sloths would demand a Central American
bridge in the late Miocene or early Pliocene and subsequent isola-
ticui. The giant tortoises similarly would call for connection with
the Galapagos islands and isolation since IMiocene or early Pliocene,
while the terrapin demands a Pleistocene bridge with Florida im-
passable for everything but one terrapin. The two insectivora simi-
larly would call for an Oligocene connection with North America
and subsequent isolation. Such conclusions seem to me inconsistent
and improbable.

I do not think that one should trust blindly in these indications
of former geographic relations from faunal affinities. But I do
think that so far as they are of value, they should be fully and fairly
presented. It will not do to pick certain points that may fit in with
a particular theory, and gloss over or ignore discrepancies ; nor do
I see much profit in inventing elaborate hypotheses depending upon
a series of unknown factors or unprovable assumptions to account
for discrepancies. One may of course say that we do not know
what the Tertiary fauna of Central America was like, and that it
may have contained just the necessary elements to account for the
West Indian fauna through a Pliocene connection. But that merely
transfers the difficulties that confront one in attempting to work
out any solution that conforms with all the faunal indications. It
shifts the problem from the West Indies to Central America. It does
not solve it.

The only explanation that seems to me conformant with all the
data, physiographic, geologic and faunal, is that the islands have
been populated by colonization through storms and ocean drift with-
out land connection with the continents, but aided by extension of

FOSSIL VERTEBRATES IN THE WEST INDIES. 181

the land areas to or near the borders of the continental shelf in the
Pliocene and Pleistocene, and perhaps by some further connections
between the greater Antilles. The mammals and chelonians would
seem to be practically limited to ocean drift as a method of trans-
portation. I have elsewhere discussed the probabilities of this
method in its relation to the length of geologic time ; and may add
here only that if the estimates made by Barrell of the length of
geologic periods are accepted, the chances become from twenty to
one hundred times more favorable. For birds and bats, the smaller
egg-laying vertebrates, invertebrates and plants, the agency of
storms appears to afford the simplest explanation, although wave-
borne drift has also doubtless played a part.

Addendum.

Dr. Barbour's admirable memoir on the Herpetology of Cuba
was published too late to incorporate in the foregoing discussion
the very satisfactory solution which he supplies of the alleged Cen-
tral American species Crocodihts. moreletii. He shows that this is
an error of record, the species being based upon a specimen of C.
rhoinbifer from Cuba. The Cuban species is therefore an isolated
one in the genus, with no near continental relatives. A study of the
series of fossil skulls from Ciego Montero has been undertaken by
Dr. C. C. ]\Iook, and will be .published later.

]Mr. Anthony has suggested in conversation with the writer that
the Venezuelan " Procapromys" may like Crocodihis moreletii be
an error of record, based upon a specimen erroneously credited to
Venezuela, but actually coming from one of the West Indian islands
and identical with Geocapromys or Capromys. This, if verified,
would clear up the affinities of the Capromys group conformably
with the distributional relations of the other rodents, solving a very
perplexing problem.

THE RELATIVE CONTRIBUTION OF THE STAPLE

COMMODITIES TO THE NATIONAL FOOD

CONSUMPTION.^

By RAYMOND PEARL.

(Read April 24, 1919.)

The purpose of this paper is to present a part of the more im-
portant results of a detailed and comprehensive investigation of the
food resources and utilization of the United States. The complete
study will appear presently in another form, and then the supporting
data on which the results here presented are based will be available.
Only the final results as to consumption of Jiuinan food will be
given in the present communication.

It is quite impossible to present here any account of the critical
precautions used to ensure accuracy, nor can the details of how the
final results on consumption were arrived at be given in this brief
paper. It is hoped that the indulgence of the reader will be granted
until the appearance of the complete publication,- where all the
details are set forth.

I. The Plan.

The basis of any adequate survey of food resources must be
essentially physiological, rather than one of commodities or trade.
Broadly speaking the ultimate sources of food are the soil and the
sun. The energy derived from the sun through the mechanism of
the green plant builds up the inorganic chemical elements of the soil,
air, and water into compounds which can be utilized as food by man,
either directly or secondarily in the form of the products of animals
which have been nourished on the primary foods of the plant world.

For the purpose of statistical analysis all nutritive materials pro-

1 Papers from the Department of Biometry and Vital Statistics, School
of Hygiene and Public Health, Johns Hopkins University. No. 4.

~ Pearl, R., " The Nation's Food." In process of publication by the W.
B. Saunders Co., Philadelphia.

182

PEARL— NATIONAL FOOD CONSUMPTION. 183

duced and consumed fall into one or another of the following cate-
gories, which are obviously based on the considerations set forth in
the preceding paragraph.

I. Primary Foods. Including all plant materials used as human

food or fractions of such materials, and all animals or
animal products in which the animal gets its nourish-
ment from some source other than the primary feeds and
fodders as defined below, either

(a) Directly as harvested, with only such sophistication as

comes from cooking, such as, for example, potatoes,
fish, oysters.

(b) In derivative form, where by process of manufacture a

food product is prepared from a raw plant product ;
such as, for example, wheat flour or cottonseed oil.

II. Primary Feeds or Fodders. Including all plant materials or

fractions of such materials used for the nourishment of
domestic animals, either

(a) Directly as harvested, such as the coarse grains, or

(b) In derivative or manufactured form, such as manufac-

tured feeds.

III. Secondary Foods. Including all edible products of animals

used for human food nourished with primary feeds and
fodders, including both produced,

(a) Directly, without involving the death of the producing

animal, such as, for example, honey, eggs, or milk, and

(b) Derivatively, involving the death of the animal, such as,

for example, the meats.
The basic idea in this classification is, of course, to allocate the
nutrient resources of the nation according to the usage made of
them. We have certain products of the soil, and of the seas and
fresh water lakes and streams, which are directly produced and
directly consumed as human food. To produce a crop of potatoes
or of cod fish or oysters it is not necessary to feed out to the grow-
ing crop some other crop such as hay or grain. Therefore these are
direct, primary food products. On the other hand there are many
foods such as the meats, eggs, etc., where to obtain a pound of pro-
tein, or fat, or carbohydrate for human consumption it is necessary

184 PEARL— STAPLE COMMODITIES AND

to vise a certain amount of other protein, fat, and carbohydrate pri-
marily produced as fodder or feed. Human food produced in this
manner is obviously secondarily produced and cannot be allowed to
count in the net nutritive balance sheet on the same basis as the
primarily produced food. It is a relatively more expensive form
of nourishment.

It is evident that under this classification many raw food mate-
rials will of necessity fall in part into two or more categories. For
example, to take the case of wheat, the major part of the raw grain
is ground into flour and as such used as human food, but in the
process of making the flour there is produced a certain amount of
feeding stufifs, bran, middlings, etc., which only indirectly contribute
to human nutrition through the products of animals which eat these
wheat feeds. Finally a certain small proportion of the wheat grain
is fed directly as such to livestock. Similar considerations apply to
very many other food materials. That all this adds a considerable
complexity to the problem is evident. But it is equally clear that if
anything approaching reliability in the final result is to be attained,
due regard must be paid to these complicated subdivisions in usage
of the raw food materials. Otherwise the same nutritive material
will be duplicated in the accounting and a misleading result reached.

The general plan of this study has been first to determine as
accurately as possible from existing official statistics for each year
from 1911-12 to 1917-18, inclusive, the amount of the basic nu-
trients, protein, fat and carbohydrate: (a) produced, (b) imported,
(c) exported, classifying the results under the main headings given
above. From this tabulation as a base one may then proceed to
calculations of consumption.

In all cases where investigation showed it to be necessary deduc-
tions were made for the following kinds of reasons :

(a) Loss of commodity in storage.

(b) Spoilage of commodity in storage.

(c) Loss of commodity in transit.

(d) Spoilage of commodity in transit.

(e) Loss by vermin.

(/) Amount fed to livestock.

(g) Amount used for technical, non-food purposes, including the
manufacture of alcoholic beverages.

NATIONAL FOOD CONSUMPTION. 186

(h) Inedible refuse.

The effort was made, in the most careful and critical manner
possible, to have the final figures for human food consumption rep-
resent net values. It is believed that this desideratum has been sub-
stantially attained. In the complete report of the study detailed
statements regarding the steps taken to get net values vi^ill be given.

In making up the basic tables each commodity or derivative of a
commodity was listed separately and converted as such into nutrient
values. In the matter of units of measure the following general
plan was followed : In all basic tables the quantities of production,
export and import are first given in the American units (bushels,
pounds, gallons, etc.) of the original statistics. These quantities
were then all converted into metric tons.^ All nutrient values, pro-
tein, fat, and carbohydrate are given in metric tons. Energy values
are expressed in millions of small calories.*

Regarding the sources of the basic statistics the following gen-
eral statement may be made here.^ For production figures the fun-
damental sources, in the case of primary products, are the successive
Year Books of the U. S. Department of Agriculture. Each volume
of this publication carries as an appendix statistical tables giving
the Department's official figures of crop production. A secondary
source for crop production figures is found in the successive volumes
of the Monthly Crop Report of the U. S. Department of Agricul-
ture. Its figures are again official and form the basis of the tabula-
tion of the Year Book, but frequently give more detailed informa-
tion. Reliable statistics of the derivative products such as flour,
meals, etc., are much more difficult to obtain than crop production
figures, for the reason that they are not officially collected and pub-
lished. In this field resort has been had to a variety of sources,
such as trade papers, census returns, special ad hoc inquiries of
manufacturers, etc.

Export and import figures were taken from the official reports
(annual and monthly) of the foreign commerce of the United

3 The metric ton = 2204.6 lbs.

* A small calory is the amount of heat necessary to raise i gram of water
1° Centigrade.

° This statement is supplemented by more detailed source references in
the complete report of the work.

186 PEARL— STAPLE COMMODITIES AND

States compiled by the Department of Commerce. In a few cases
where it has been clear from information available to the Food Ad-
ministration that the official figures of the Department of Commerce
were in error we have not hesitated to use other and, we believe,
more correct statistics, but in each such case specific notation of the
fact is made in the detailed report.

In the computation of nutrient values use has been made chiefly
of the factors given by Atwater and Bryant.*' It has been neces-
sary, in some cases, to supplement their tables from data given by
Leach'' and Henry and Morrison.^

All calculations in this work have been repeatedly checked and
every possible precaution taken to guard against error. It is too
much to hope that so extensive a piece of statistical work should be
without errors, but I hope that their number is small and their net
significance in the final results negligible.

II. The Consumption of Human Food in the United States.

Hitherto there have been available only the roughest guesses as
to the total domestic consumption of all but a few items of food,
such as wheat and sugar. If anyone were confronted with the
naive and simple question, " How much corn, or oats, or molasses,
or fish, or milk, or nuts," or any one of a long series of other foods,
" is consumed annually in the United States as human food," no
accurate answer could be given. Yet the question is obviously a
fair one, and one which somebody in the nation ought to be able to
answer with a considerable degree of accuracy. For some twenty-
odd great staple commodities or groups of like commodities we are
now in a position to present figures of a relatively high degree of
accuracy as to consumption. On the basis of these figures it is
possible to discuss effectively many interesting and important prob-
lems ; such as, for example, that of the relative importance of great

6 Atwater, W. O., and Bryant, A. P., " The Chemical Composition of
American Food Materials" (Corrected April 14, 1906), U. S. Dept. Agr.
Office of Expt. Sta. Bulletin 28 (revised edition), 1906.

^ Leach, A. E., " Food Inspection and Analysis." Third edition, revised
and enlarged by A. L. Winton, New York, 1913.

8 Henry, W. A., and Morrison, F. B., " Feeds and Feeding." Sixteenth
edition. Madison, 1916.

NATIONAL FOOD CONSUMPTION.

187

5 H

u

c .

rt vO O 00

3s

00

in

N in

IN

vO

PO vO vO CS

f^

r^ oo f^ (N oj

0\

■+VO oo

0\

t^

ro O in 0\

^

Ul c

r^ ro "^ 0\ ■+

C\

M_ 0\ 00_

0\

vO OO w ro

q

•- .2

6\ (> lO m" oo"

4

d ti lo

IN

in

M ro r^ M

rf

o —

lOOO \C M o

<N

oo 0\ CO

t^

ro

ro ro ro lO

m

i^ M U-) r^ f^

o_

9. '1. "^

vq

00

in M in 00

q

H Ol

ro

H lO m"

i>

vd

^

ro

1

ro

^

.«■

M ro ro O 00

lO

IT) vo r^

r^

00

00 oo 0\ oo

ro

!^C §

Tf lO Tt (N vO

(N

M P) r^

M

in

vo in o i-i

in

J= — £_,

M M 00 Tt q

o

O) O t^

q

0\

o in in Tt

lO

ro '^ ro vo" to

N

"^ Tf in

'f

Tt

M o" in Tt

^

ro

k4 C3 ._•

lo M M in -^

0\

oo M 00

00

o

Tt ro M 00

t^

I

UO 00 M M M

i^

M IH (N

in

ro MM

vO

S

vd H

oo"

M

JH

'f

"

O lO O t^ ro

r~

O 00 <N

O

o

vO t^ 0\ 00

o

.S •" ")'

<M Tf M 0\ ov

r^

n t^ 00

0\

O 0\ O 00

o

r! i^ c

r^ M "* n 0)

o_

oo m o

q

^ ro vO 0)

t^

O 00 m" o"

t^

•^ t^ 0\

(N

d\ ro !>■

o"

00 0\ M

0\

0)

Ol

_c

0\ 00 i-l O Tf

M

0\ (N r^

OO

in

0\ 0\ 0\ M

00

■i'Hs

^ ^ O ■* IH

lO

r^ n 1^5

IN

in

M 00 M t^

0\

OJ £-^ t^ 00 M

o

(M Tt LO

IN

■*

in in 0) M

^

U D o

to ro o to o

w

o" md 6\

vd

d\ M i> 00

vd

On O »H HI CNl

lO

t^ ro CN

ro

M

1

CM

0\ <N

(N

M

M

.5 ■

•* -* O ro ^

^

CM N PO

t~-

T^

0\ 00 Tt 00

0\

O 00 0\ O M

o

IN f~ O

Ov

N

O OV CC 00

t^

^ "=

t^ ^ 00 m M

I^

vq <N o

00

It

cs ro O M

^

•r- .2

ro ro O, >o rf

o

^ ro r^

'^

M

oo" m" ro o"

rf

O 7^

ro oo ro O cs

00

ro Tj- O

^

P)

oo ro vO M

Ov

0\ M lO O 00

^^

O CO ro

IN

q

ro M in r^

I^

M d

ro

m" ro M

vd

vd

[_^

Ol"

ro

^

»;

0\ 0\ N vO M

t^

in CO lo

00

M

O ro "* ^

r-

>.c §

O -* '^ ro -*

t^

M 00 ro

1^

M

ro oi M 00

in

[

J3 — lI

<N On <N O CO

t^

q o IN

00

in

M n 01 o

0)

0\ H oo" ro d

m"

Tf ro in"

o"

vd

d d m" m"

m'

•

00 ro M ro u->

N

t^ (N 00

00

o

O ro 01 to

H

^/^ CO M M M

oo_

M 00 CN

IN

q\

ro MM

>o

M

s

no" h

oo"

"

ro

M t^ H 00 t^

in

0\ ^ r^

o

o

O 01 oo ro

CO

c .H ,„•

ro O lo ro lo

00

MOO

ro

O 0\ 00 t^

m

"Z - c

iH O ro ro •*

00

o_ vq vq

O0_

in ro r^ M

00

rC d M 6

oo"

•^ ■^ 6\

6\

00 ro t-^

dv

00 0\ M

0\

c

O 0\ lO N 00

Tf

I> M (N

O

T^

vo M in M

'^

'i^ " „•

ro O M t^ (vi

lO

M vO VO

^

in

Tt !>• t^ 0\

00

•S'SS

r--oo N ■^ f-

O;

t^ OO O

IN

t

vq in in ■*

Ol

S D O

o" lo o" ro O

o"

vd o" OO

o"

00 M t^ vd

Tt

g§H

O O M M M

lO

vO O 01

C\

01

tin

O IN

OJ

M

<o

^

>.

•3

o

o

(^

It

E

3

m

1^

o

'O

u

CO

U

2s .
a t1 Z

3 •
O J5

•S

(^

s

0

at and
1 produi
produc
and pr
r cereal

t3

to a-

a,
c

M
>

1m

a

I

0-

'3

to
s

s

0,

u

c
U

a,
>

^ c.

5

0,

o

s

110 ,1

Oh

O

0-

-**

3

r

G

o

o

CO

188

PEARL— STAPLE COMMODITIES AND

c .

c^

l_l

t~

o

o

«^

lO ro O

oo

o

^

1 n

r~

^

1

O O 00

^

(V

CO

O 00 Tt

r'

r-

o

vO

o

lO

V C

■* >c o

> M

T

r-

IN O M

vC

r-

CV]

N

o

r^

•- .9

oo" « 00

o

00

\r

^ OO O

M

d

1^

d

M

i-C

o -^

O O vO

o

r^

M

00 o, u-

u-

•;

o

r-

cc

o

O (N CN

<J

u-

j>

t^ ■^ 00

►H

> I/"

) H

r^

w

M ^

1

vd

o d

r-

oj

d

d

vd

r-

'^

M

"

■^

>-= o

N O r'

1 O

oo

0\ r^ N

O

o

o

Tt

o

00

ro O

\ CN

(N

O

^ O T)

o

o

vO

in

On u-

^

vO

O in ir

5 r-

o

T

M

m" u-

^ oc

t-

M (N

■^

(vJ

t^

d

?

n H

"^

1-

o

f-

f-

t^

t^

00

00

q

i

s

K-

>

vd

M

M

00

"^

00 vO vc

C

r-

■n

Os O IN

O

0

\ c

\ vO

Tl

00

0\

c .y '

IN ro r-

I- t-

00

ro t^ M

r-

5 f

3 N

o

\ t^

"^ i: n

(M M ^

c

•^

r (V

)

Tf IN vO

i>

vO

1 H

o

■*

ri u O

M oo' c/-

oc

oo

O

vdvd o

^

00

f

3 OO

3 M

o

t^SH

00 in H,

T M

M

0\ (N r-

o

\ O

3 u*

3 r-

0\

•+

00

Tt a

'^

M

r'

3 O

00

M

C)

M

^

^

c:

0\ O oc

r-

^O

vC

00 w r-

o

'n

^

N

O

VO

"" u .

O r~

T

1- M

o

o\ o r-

r-

00

r-

r'

3 vO

in

*"*

C -r in

ro C

■> o

t~- cj >o

1 \r

> r-

VC

3 t^

ro r-

^ J_

M m" 00

M

0

o

P

ro

£^H

^

^ 00

ir

5

IN 00 ^

5

VO

0.

vO

in m

c

\ IN

r-

C

\ in

<

M

fO

H

CO

lo ro f

o

ir

5

O <N 00

N

r-

(N

IN

f

3 00

« .

ro O IN

(N

M

O

ro w "•

1 r-

o

r-

VO

m

Q

[/I c

t ° ^

: cc

00

M

M in M

(N

c

5 '^

f O

o

N

a

•- 2

M M c

5 !^

^

■

Pj" ^ (N

1 oc

O

r-

M

in

H

o —

M N oc

1 M

r- r-vc

o

c

^

r r*

3 O

\ o

6^

O- ro w

^c

: Ti

■

o 00 r-

t~

1 'ij

f

3 O

\ ^

M 4

■)

^

•

ri d

r~

cs

C

\ O

\ 4

r-

M

M

■^

N

S

w

^

t-t

>'C o

M O IN

:> c

vC

in Tj- M

o

c

^

vC

N

OS ^

^ IN

(v

1

IN \0 IN

c

M

in

^■-:I

■rt- CN

r-

l:^ \0 I'

■) oc

3 -;

[• in

o
H

d ^c

ro 1-

•

c

M Ci

•^

o
00

00

3 rf

m

T g

T

•

oc

00

m
in

hIh- §

g

o\

*-

t-i

o

w

"

lO ro C

oc

vC

w

in o I'

vC

00

1/

1 o

(N

s

c y •

m o t-

r-

r-

^ OvvC

f

c

1 o

N vC

ro

a D 0

in -^ ^

^

>-

H-

in vO ':

a

^ O

C

3 t^

^^-^

t^ ^ C

■i 1-

o

\ c

vd ro 1-

1 r-

oc

f

5 N

fclSH

t^ O ^-

-c

M O vC

'^

>c

^

vC

(N

^

-*

1

oc

in 0\

f

0\

tH

cs

M

■<;

h 't

c

O O (N

c\

oc

w

■^ in IN

t-

c

N C

•^

I- rl

h vo

"^ o •

O 1-

r-

•^

h I'

>

O ^oc

oc

\ OC

c

vC

HI

t^

•S-SS

^ ^

t r-

0

^ VC

M t^ t-

c

> VC

vC

VC

0

\ vO

o

il <u o

d oc

cc

vC

f

5

o" oo" If

^ c

^ If

i

vC

r

vd

S2H

■*

1- oc

h-

'^ 00 r;

|- r-

f

■>

C

■^

1- in

^

Oh

vC

in rn

C

N CS

t-

0

> ^

o

H

ro

g

S

,

P

c«

:?

o

u

w

W

H

-3

to

!0

1/

?

o

E

e

o
U

c

c
c

■J

a

O

1

s

c

c:

c

■S

■* o

■il'

(^

c

a

0
0

^

t

O

i2 c^

C

^

la

T3 -a <^

t

1

X

"«

t) t^

■a

c C c

1

s

e

C

g

3

a

^ ^

[c

1

K
O

n! n! C

"c

0.

° (t

'^

3 OJ X

c

3

<o

'a

o u -

o 0 j:

2

^

s

^ ^

2

6

«C_

fe

c:

d

<^

(?

c

a.

c

NATIONAL FOOD CONSUMPTION.

189

u

5S H

fc

M H t^ lO O

■*

O vO O O

p

^ t- o M r-

1 w

_c .

O O 0\ 0\ M

vO

O ro CO M

ij

0 ro ^ "O '■

t- t^

!'' C

t^ ro N ■* ro

M

q oo_ o_ ■*

a

5 sO_ ro OS vC

in

•- 2

o" O f^ ro o"

"i^

ro r^ "^ M

u

0 o" vd N o

T ro

o —

Tt- 00 O O ■*

o

O ■* 00 o\

h-

< m (-< H a

t^

"55

M O 10\0 00

ro

q q Tt lo

r

~ o_ w lO t-

q

H d

cs"

M in w r^

t

■- M

ro

ro

■*

1

h

S^c o

ro M sO t^ O

ir~

ro O lo 00

•c

) t- o» "* o

1 ro

(N lo ro 0\ LO

in

Cq 00 M t-H

0

J \0 ro O u

T vO

O u~

N O O lO 00

•o

'^ t— 00 o

r

lo 00 ro r

r t

vO to <:> <N C>

ro

<> O" rf IT)

C

j\ r^ oo" d C

Jv vd

t, rt .«

"^t 0\ M ro "*

ir>

i^ oo o o

t-

< O (N M <:

r^

t?

^ r~ W M M

o_

M o ro >n

r-

D ro M 1-

>o

1

vd m"

oo"

M

1 "

^

f

2^

1

vO M so t^ >-<

HH

0\ ro O 00

C

5 M ■<t t^ ^

Ov

^ •- ui

M r- 00 ro O

M

■■o yn 0^ ^

Tl- t^ •* C

vO

'^ h C

ro to ■* ro o

ro

M ro w ro

w ro ■*vC

m

rt u o

lOO M m"

iv^

•^ r^ o" rs"

o" ro vC

d

fc-SE-'

00 0\ w

0\

M M

0)

e

O ro 00 r^ (N

VO

ro O M ro

^

1 ■■O Oi r)- X

t-

*" o •

O O CS <N t>

ro

ro ro ^ w

u

1 lO c\ M r

'^f

•St s

00 iJO c< "^ u-i

in

oo_ ro "^ vO

^

f c^_ "^ q oc

in

ii t; o

00 M M ro oo"

ro

oo" 0)" ro -^

o" m" o" ^

- vd

P2H

r^ O M M M

ro

\0 ro r-

0 ro

M

£

0\ CN

tN

oj

a

_c .

i/^ 0\ ro 0\ M

3

O tn u

0 O

C

3 f) t^ \0 r

CJ

ON ro H vo oo

0\

t~- 00 ^

J- o

<

) O O ro r

0)

yj c

lO 0) ■* to vO

^

roo u

2 \o

H

< lO vO 00 c

q

■- .2

t-^vO O t-^ lO

r^

ro t^ C

j\ d

c

>" d oo" m" OC

d

o —

M cs 0\ noo

lO

0\ ■* r

(N

■^

J- ro IN M c

t^

"as

N H lO 00 00

o

M M O

0 t-

t-

1 0\ M o r

ro

t-^ 6\

oo"

h" 4 1-

h" o"

oc

1

cs"

l-O

t

H

>>C 0

i~~ ro ^ rj H

r

Ov H -C

o

C

) O 00 t^ 1/

T O

•^ lO O <N Tf

o

^ ro 1-

Ov

C

> M m o X

t- oo

q ro ro q ■*

c> IN o\ ^ o6

ro

f^ oo" <:

^

c< ro "1 C

0 t^ d w c

M

oo

I- rt .-

I> O CI 00 lO

lO

O 00 oc

r^

0

O cq (o ti

-> M

>4-

\q 00 M M M

0\

r^ 00 ^

ro

^

(N M 1-

m

^

t^ m'

d

"

r

t

"^

vo ro 0\ lo 0\

M

<M O «

o

C

M W M «

M

^ o •

^ O r^ O i-i

r^

00 -Tj- ^

o

O 00 w C

o

■^'H c

t^ (N Tf T)- M

o

cq O H

lo

t^ ro 1-1 vC

00

rt 4; °

M r^ 1-1 M

ro

lO vO c

i\ o"

in ^ ^

- r^

fagH

O 0\ M

(N

tN

.5

ro 00 roo M

M

r~ o 1/

1 f4

u

") <N O 0) vC

O

- <-> _■

■5j- o t- ro 00

o

U-) lO C

M

ly

-> 0\ IN C^ 1/

■> 0\

•I'SS

(N ro w -O Ov

^

t^ 00 vC

01

■^

f r^ in M vc

S u o

\d (N m" oo" (5.

oo"

VC" oo" CC

-^

in w 00 vc

rq"

I^H

VO O M M CN

Ol

r- O r.

l-l

IN

1 <M

J

M

Ol

M

1/)

>^

-a

o

!<

O

ft.

»

£

^

3

to

<o

o
O

■ o

■ 3

r,; "O

^

•

S

at and p
product
product!
and pro
r cereals

a

tr

c

Si
<L
>

1=
0 ^

cr

CL

0 c

to

^^9 ^M

o

-g 3 rt 0,

2 M -^ .a

o

J^

— C ™ a.

o

rr o ^.s! 4J

^ OJ O •"

^

•^ G u 03 *J

c3

^

O

«

rt

0

^1-

Ph

o

"5

CO

^

0 s <;

o

m

o

"2

190

PEARL— STAPLE COMMODITIES AND

o

Tt W

ir

o

t--

„

t

M

M

o

r^

1 M

o

O

1

c ,

O C 00

00

r-

o

rt- t^ 00

c

00

o

^

i-t

Z S

0\ m IN

o

^

M

ro ^ (^

u-

^

'J

Ti

i-(

•-.2

d ^"^ c

t-^

oo"

w

ro r^ r^

r^

d

XT

t-^

vd

ocT

O ^

lo ro M

o

■*

c

t~- 0\ f-

O

ir

o

00

CM

M

6^

c< m (v

q

Lr

f

\0 r^ r~

HH

>q

-^

r-

M

lO

H lO

t^

oo"

O" M

d

CM

C

cm"

d

r~

IN

CM

ir

fO

>>-5§

o o -^

•^

O

r-

(N ro cs

O

o

o

CM

<N

o

O u-

M

r^

O t^ o

00

M

o

CM

g^H

^ o

c

M

\q oo_ ^

o

f

CM_

"?

•£ 2.H

oo" r-

o

)H

m" cm

^

ry

oo"

d

n M

r-

O

o

r^

1?

(N

O

o

M

1

■^s

IC

o"

OS

o M n

N

M

lO vO -^

00

o

r~

CM

VO

t--

00

C _o •

o ^ I>

r^

t^

\0 i-i i^

) o

rj

M

o

vO

rO

'Z t a
n HJ o

° ° ^

) r-

r-

o

n ov <N

T

c

q

'C

M

00

0\

lO l/j M

t-i

00

00

d r^ ^

t-^

i/

CM

d

vd

■*

t— I

fcSH

0\ O M

M

vO

O u^ vO

o

r~

u-

o

f

O

o

lO

r-

O

\ ^ w_

r-

HH

■^

c

CO

H

pT

cm'

M

^

in

w

c

0\ O I/-

5 T)

vO

I> 0\ o

O

ly

) o

VO

r-

3^

'Z " -•

M C

eg

o

1>

M O <^

•> o

o

o

c

vO

ro

■e"5S

C\ M

M

r-

o

t^ 00 CM

M

o

t-

M

q

1^

„ u o

d c

\ O

vO

CM

oo" O CM

r-

d

CM

d

w

t/T

w

2SH

lo

00

"^

O " -^

o

'^i

Tt

ir

o

o

liO rt

O

M

I>

q

vO_

<

M

CO

H

lO ■* M

o

o

t^ U-) O

s r'

o

IN

VO

o

o

C/J

C .

0\ Oi f

> M

o

o

M in 00

r-

r-

00

00

*"* j2

q\ o_ o

S O

oo_

n

lo ro r'

3 r-

f

\0

C

\ VO

00

.Ji 3

r^ oo" c

) o

r'

w

•^ ^ M

M

I>

00

p^

> >-l

CM

c3-

rt ro ^

1^

5 Tl

t>

ro O CM

CM

o

O

\ r-

o

r^

(N O f

5 O

\y

) O

vq r^ O

\ c

^

^

'J

o

V q

w lO

mO

r^

5 vd oo"

O

(N

c

V 00

IH

^

00

CM

■^

rn

M

I c

M O M

cs

oo

3 ro cs >^

3 O

o

o

o

vO

Ov

w

00 r'

M

r-

00 00 r'

1 00

o

Tl

M

S H

fe

ro O

\ f

■>

U~l Tt U-

3 ""

1

t-

q

cSS

o

oo" O

\ Of.

O

M cm"

' ^

•^

00

oo"

ro M

5

5

CO

00

M

•It-

c5-o|

3

oo

00

CM

• o

o

I— (

1

vC

M

1-1 o

-1 ^

w

*^

N t^ u-

t 'Tl

^ r'

) "■

5 M 00 t~

ir

> Tl

w

VO

vO

M

J

C u •

O r- 00

2 «

o

s ro CM O

\ VO

00

VO

o

\ o

o

"Z'£ c

rt I' o

q ^00

5 O

00

o

> vq i-H c

c

\ CM

(N

00

5 0

M

1 in

tSsf^

O '^ <N

vO

f

3 00 lo r-

I>

ir

3 r-

(S

in

lO

^O

o

\ •^oo^

^

M

f

) o

0\

t-t

<s

'd

^ Tf

s

■* o c

Tl

h oo

vC

00 .O Tl

1- 0

\ V

5 0

f

3 r-

(VO

O

•- o .

■* c

Tl

o

m O 00

r-

VO

t-

H

CM

^

•rss

O (N

•^

H

00

I> lO >C

p'

5 O

s t~

VO

r-

IT)

i o o

6 c

c

vC

r^ 'f oc

o

^

tH

r-

d

^z;

2SH

lO 1-

vC

00

o o -^

!■ CM

'^

: 0

V o

o

oc

>n CO

0

\ CM

r-

00

r-

H

H

ro

(U

S

P

CO

^

O

U

u

W

H

><

-3

to

c^

?

o

E
g
o

U

"2

c
c

c
c

s

s

1/

-

1

c

m X

c

!«

■?

■^ -a -a "

0

^

5

n: c

C

a

0

"

►2
O

O a
flj a

C

■5

^
■^

t.
It

1

^
■^

]

"e

3 n

-a c c c

s

5

c

E^

3 a

5"^

t

[c

§^^t

3

" i

"c

c

-a

<o OJ 1- -

M <" 0 j:

>

5

^ 2

s

s

^

c?

t^

c?

■^

c?

ft-

c

C

^

^

1

00

I— I

On

o ^ o

U

C ajH

fc.:sH

■r; s c
i; u o

-2~

" " c

« u o

■7: *2 c
ii l> o

•- .2
o5

O'l

feSH

0\ H O 0\ Tf

0\ to CO O 0\

►H oc On *-' 1^

01 ro O 00 c^

O 0\ O) ro OS

0) O 00 00 r~

00 M vo N 10

M ro vO '^ 00

><~. 10 tJ- ro <N

0\ to CO O 10

M M M ro ro

10 10 \0 r^ ^
ro ■^ to \0 O
00 00 CJ t^ t-*

fO to r~ >o 00

^ Os On M 00

i/^ ro to t^ O

o" N -^ o" 10

^ •^ 0) rO NO

On n

O M On 00 M

M O M O O

to ■^ M to 10

■^ ■^ r^ co" M

00 0\ ■^ ro On

t^ O NO M O

fi C^ MM

r)- t^ to O NO

M On N ■* ro
O ^ t~- M o

t^ t^ M M

NO On " •^ t^
O nC Oi ro 00

to M NO NO to

\0 On to M CO
t- O T)- CO On
On r^ w ro c^

NO n ci to (5,
M O w (N) ro

r^ On n 00 to

cc to ro r^ nO

ro CNi c: to M

r^ o" vO (> o'

O t^ O ^ 0)

O O O t> On

00 M nO M M
w 00 to On Tt-

i-i ro 00 NO On

no r- w M

MD ro On f- M
Tj- I^ M M 0\

M M to Tf 04

c

NO ro

0 t>-

M

"~ D •

00 NO

00 t-

ro

■7^.a

0) M

^ vO

M

On w

M NO

M

2§H

tr> 0

M »H

ro

Ph

>- a tj -2 o ^

s ^

■11
1^

u O

u 5-"

o >N

U Pi

So

NO 00 00

M M On
t^ ^ On

01 oj 0)

ro On to

t^ N M

to ro O)

M !N to

cc" to r^

On NO 1-1

M ro ^
(N O 00
00 t^ M

<* t-~ >*
^ ro M
On C| Tj-

rc 00" o"
M OnnO
On rOvO

r- On ro
to 0) rj-

r^ no o
ro 01" no"
to ^ M-

M o\ ro

tri o) 00

':}- M On
M rl- On

Tf On O 1-1

to O NO 04

Tt to ro O)

to ro 0> "^

O On 0) On

oi rf On

to o CO r~

0) \0 og 00
■^ ro ro 04

^ 01 00 r-
t^ 01" NO

00 On w ro

10 o t— 00

^ M t~. 04

!>■ m" to <>

O ro On O
r^ CO O] On

00 *-< '^ r^

On 04 r^ t^
04 OS "* to

M ■* On On
NO ■* 00 04
Tf 10 00 -^

Tf M t^ oi
00" 04" 10 t-~

vO ro to On
00 On O "*
00 04 O O

o" 00" C> w

to 01 to nO

to M ■:}• On

10 NO t~- to

ro 00 M 01
00 04 t^ O

r^ On to rj-

no r^ 00 t^
to ro O M
On ro On

O ro On to

H 04 NO On

O to M 04

On w NO On

•^ 3 OJ O
2 M -iJ .:::

^ o o ^e

7; c S <"
■S o t« S —

s<cmo

00
0\

^
^

■■w

w

-^^

W

c«

H

2:

g;

o

rn

u

Q

1,

n

OS

t-H

n

H

W

fc

^

J

Z

m

<

<

s

H

^

m

u

w

00

a

OC 00 M

t-

a,

t

00 O 00

CS

0)

00

r~

t

CM

00 IN Tf

ir,

M

■o

O " Ov

01

<3

o

0\

t^

r^

y? c

0\ lO vC

*->

■^

t^

'^ "'^. f

t^

0)

in

00

q

d

■-.2

c^ ci oo"

•^

ro

t-^

r^ Tt CO

cs"

oo"

d

d

M

o —

^ O t^

o

ro

O

M 0\ m

0)

^

r^

t-t

in

00

5^

N l-l >0

•-^

m

00

O in in

HH

>q

h-i

o

q

0"
00

N ^

oo"

ro

ri d

00*

cs"

hh"

,»r

04"

00

cs

0)

0)

in

>.C o

'^ O r-

w

lO

r~

r^ O in

t^

O

O

O

0

00

u-l x>-

ro

C>1

cs

t^ m M

M

O

00

h-4

-3 t: u

q w

M

q

in oo_ ro

t^

oo_

00

ro d.

cn"

ro

m" cs"

4

r^

h-t

in

od

u « ■ —

*0 CO

o

c^

Ov

00

i.

s

vO"

r^

*-<

M

^

r- M ro

„

vO

0\

O ro CO

r^

o

t

O

0

0\

c y •

ro lO r~

^

O

r-

O m in

00

M

cs

cs

vO

oo

'~'i: c

ro 00 M

-^

°°.

ro

in o 00

CS

q

1/1

o\

a " =

d 'i- r^

M

t^

cn"

CO in o"

cs"

in

in

in

t-^

d

J^2H

t^ lO •rf

CO

H-l

f^

M 'tj- ^

O

r^

cs

o

0

t~

H lO

t^

q

in o

cs"

M

M

in

M

in

c

CN O ro

(^

„

^

ro 0\ 00

r^

P)

00

0\

vO

0

"" cj .

ro 00

c^

P)

vO

o a C\

t^

r^

o

o

04

0\

C ■— (/I
*51 i3 C

0\ O

q

q

00_

r^ t^ o)

cs

t^

00

q

00

0

S u o

" o"

ci

lO

d

d oo" oo"

in

oo"

i_h"

oo"

^

Tf

PSH

00 M

O

00

0\

ro t^ cs

^

■*

00

00

00

Oh

"

r~

m rn

0\

M

t^

q

M

ro

O " I>

-*

lO

t>

^ ro ^

ro

i_i

in

so

r~

■*

c ,

^ Ov H

vo

M

in

t^ M 00

O

o

o

o

ro

0\

« c

t^ M l-H

q

r^

q

m cs 00

o

M

q

01

q

0

™ .2

c> ci \o"

oo"

r^

oT

cs" ro 00

in

6

oT

d

00"

d

o ~

M 0\ <N

-*

cs

ro

Tf- r^ t^

h-f

M

o

o

0\

00

— "^

ro q o

q

lO

cs

00 w vO

M

r^

00

00

't

t^

u^

H lO

t^

d

I> w

d

cs"

d

ro

cs"

r~

N

Ol

c^

in

ro

>. C 0

r^ O in

N

ro

^

H in\0

O

o

o

00

'^

00

M rr,

lO

N

HH

00 o\ o

-*

in

0

M

t^ M

q

ro

oo_ 00 ^

M

oo

q

ro

o" ^'■

CN

HH

m" cs

in

t-^

ro

'S^

:

r^ ro

I--

O

ro

•*

0

t^

^

C\

0\

^

1

lO

0"

M

ro ^ O

r^

<N

in

w roo

ro

O

00

M

t-(

vO

c " •

(N 0\ ro

^r

00

0\

m cs cs

00

O

m

ro

t-

•0

■"'ii q

O N lo

^

lO

<~0

" °)^.

q

C3\

o

ro

q

ro

rti; °

lO !>■ 00

■c"

t^

d

cc" 00 o"

in

d

in

oj

ro

00

fc^r^

o •* ■*

C\

M

in

ro 0\ in

00

r~

00

00

ro

0\

l-H lO

o

a

in o
cs"

cs

M

'^

00

in

c

t^ o -*

M

in

0\

00 M 0\

t^

O

00

o

-*

ro

•- o .

IT) lO

M

t^

r-

•^ 00 00

0)

0\

ro

in

M

0\

T i3 C

Cs to

lO

PI

M

r~ r^ in

in

'^

0)

oo

0

00

1! D o

r^ oo"

o"

'^

0C3

cs" OcT sO

o"

in

i_r

ro

0

■*•

ft

^ "

o

00

J>-

O 0\ 00

o

in

00

ro

M

O^

in ro

0\

M

i>

o_

r;

M

cs"

00

O O vo

,_,

o

CS

Cs "* O

in

N

o

„

00

0

c ,

o w r^

o

o

ro

ro m ro

ro

01

t-

ro

in

ON

w c

q q q

q

r^

oo

m 0) M

in

00

■*

0

r^

•- .2

M ro o"

in

lO

o

d ^ d

o"

in

oo"

o"

0"

ro

o5

r^ m o

(N

0\

-*

in M in

00

00

M

VO

0

m

M 0\ '^

vO

•^

00 -o r~

\q

m

00

r~

°°

5^

w ro

VO

M

00

o" m"
cs

d
cs

cs"

d

0)

cs"
in

ro
00

^a.°

r- O O

r^

o

cs

O 0\ Ov

o

O

o

in

M

ro

O O

o

CVI

00

■* r^ 00

r^

Ov

t^

m

m o)

i^

ro

r- q ^

q

in

0

q

iiio

vd ro

d

r^

m" cs"

in

d

rf

(N

t. s.g

ro <N

m

04

•*

VO

IH

o\

a

q

>d

'^

N 00 M

^

in

ro

in ro t~-

■*

vO

UO

o\

'I-

t^

<- o •

0\ lO 00

ro

^

OO

O 00 ro

ro

o\

I>

■o

t^

m

■" 't^ !£

^ =q 00

cs

O^

in

q '^O

'^

\q

ro

o

w

t^

S 1".°

d -^ J>

c^"

o"

^

cs" ro cs"

d

r^

in

in

d

ro

feSH

0\ M <N

VO

CM

o •*o

o

r^

in

^

r^

0

'*

lO

00

m w

t^

M

'S^

ro

04_

cs"

Of"

M

4

in

c

r^ O O

r~

00

in

O H vO

t^

^

o

t^

00

00

■" w ■

u-i 00

CO

o

M

04 O 00

t^

o

r^

0

00

•S'S S

q oo_^

oo_

q

q

M M 0)

o

ro

00

ro

in

'^

a u o

r^ M

d\

6\

■^

in in d

d

oT

4

0

M

2SH

PL,

■* H

lO

i>

0\

oq o •*

^o

in

O

00

00

t^

in Tt

Ov

04

I>

q

r^

M

tH

ro

tr>

•

■5

to

s

"2

S

C

a

^

■3
2

Oh

cr

-

u

3
0

8
0

0

B

a

^'^

a

O

1,

a

d prod
d prod
and p

to

to

1^

1^

S
o
O

-2

o <fl

T

•X3

•S

s

7

"a

3 cs

1

•^

c - c

!S

a

O

1

f°

3 a

5 0

^

3

a

cs t« O

'a

a

a

]b

1^

s

03

o

-a
S

CO

"a

u
0

0.

1

CO

"a
<5

S
o

O

a

0

a

CO

"a
s
a

;5

NATIONAL FOOD CONSUMPTION.

193

groups of staples, like the grains and the vegetables, in the nutrition
of the people of the nation. We can calculate with accuracy the
total national food bill, and so forth.

The final net results as to consumption of human foods in the
United States during the seven years are set forth in Table I. In
that table the results are given for the several nutrient values, pro-
tein, fat, carbohydrate and calories, only. This is the most scien-

CARBOHYi >f?AT£

Fig. I. Relative curves for human food consumption. The figure for the
year 1911-12 is taken as 100 in each case and the relative figure for each
year calculated to that base.

tific, and as soon as one become accustomed to it, by far the most
useful way of thinking about food consumption.

The data of Table I. are summarized by years in Table II., and
are shown graphically in relative form in Fig. i.

The first thing which impresses one about the consumption fig-
ures is their extreme uniformity from year to year, as compared
with production, exports and imports. This is exactly what would
be expected, of course. No matter how much production, exports
and imports may fluctuate, within wide limits, the people of this
country eat about the same amount each year. To have the sta-
tistical calculation come out to this result so beautifully is strong
evidence of the correctness of the long and tedious preliminary cal-

PROC. AMER. PHIL. SOC, VOL. LVIII, M, JULY 26, I919.

194

PEARL— STAPLE COMMODITIES AND

TABLE II.

Summary of Consumption of Human Foods, Primary and Secondary
(Metric Tons).

Protein.

Per
Cent,
from

Fat.

Per
Cent,
from

Carbohydrate

Per
Cent,
from

Calories
(Millions).

Per
Cent.

from

I911-I2

1912-13

1913-I4

1914-15

I915-16

1916-17

I917-18

Total for 7
years

Average, whole
period

Average, 191 1-
12, 1916-17

'3,556,616
3.563.756
3.709.543
3,601,137

3.781.483
3,714,893
3,784,690

45 55
46'54
49'5i
46J54
47|53
45!55
4852

4,942,732
4.891.478
4.955.501
5,304,837
5.203,757
5.393.366
5.479.939

I7|83
1783
19 81
18,82
i6'84
18 82
20180

15.554.552
16,079,158
17,218,919
16,179,529
17,042,053
16,404,318
17. 135. 813

124,405,258 60 40
126,107,751 6139
131,972,88063137
130,518,116 60 40
133.853.79061 39
132,730.09460140
136,819,738161139

25,712,118
3,673,160
3.654.571

53

53

36,171,610
5.167.373

46!54l 5. 115. 279

17

115,614,342

16,516,335
16,413,088

95, 5

95 S
95I5

916,407,627 61 39

130,915.37561 39
129,931,31561 39

dilations. There has been a rather steady small increase in total
gross food consumption, but this has been very closely proportional
to the increase in the population.

In the seven-year period here discussed the greatest relative
advance in consumption was in respect of fat, and the least relative
advance in respect of protein. Carbohydrate content and calories
increased in the seven years in amount consumed to a degree inter-
mediate between fat and protein. The protein relative line falls
below the population relative line each year after 1913-14. This
means that since 1913-14 somewhat less protein has been consumed
in gross in proportion to the population. The relative line for fat
was below the population line till 1914-15, and thereafter followed
it closely.

The relative figures from which Fig. i is plotted are given in
Table III.

With such gratifying assurance of the smoothness of the con-
sumption results we may proceed to an analytical discussion of the
numerous highly interesting problems which center about human
food consumption, and for which data have hitherto been lacking.

NATIONAL FOOD CONSUMPTION.

195

TABLE III.

Consumption of Human Foods, Primary and Secondary, Relative to
1911-12, Taken as 100.

Population,

Carbohy-
drate.

Calories

(Millions).

19II-12

I912-13

I913-14

1914-15

1915-16

I916-17

I917-18

Average, whole period

Average, 1911-12 to 1916-17

100. 0
101.7
103.4

105. 1
106.8
108. 5
no. 2

100. o

100.2
104.3
IOI.3
106.3
104.5
106.4

100. 0

99.0
100.3
107.3
105.3

109. 1
no. 9

100. o
103.4
no. 7
104.0
109.6

105.5
no. 2

100. 0
101.4

106. 1
104.9
107.6
106.7
no.o

105.1

103.3

104.6

106.2

105.2

104.3

102.8

103.5

105.5

104.5

The first of such problems to which attention may be turned is :
To what relative degree do primary, as distinguished from second-
ary, human foods contribute to the total nutritional intake of our
population ? From Table II. it is seen that of the protein consumed
47 per cent, comes from primary sources and 53 per cent, from sec-
ondary sources. Thus, broadly speaking, the American people get
over one half of their protein from animal sources exclusive of
fish, which are included in the primary foods. This fact indicates
at once the importance of maintaining our animal 'herds intact and
keeping the price of animal products at not too high a level, unless
we are prepared to face the alternative of a radical and fundamental
alteration in the established dietary habits of the people.

In general there has been but little change in this protein-source
dietary habit in the seven years included in this study. What
change there has been is in the direction of a smaller proportion
of protein from secondary sources and a larger from primary, but
the movement has been but slight. As would be expected, a much
larger proportion of the total fat consumed in human food comes
from secondary sources than is the case with protein. The figures
are 82 per cent, from secondary sources and 18 per cent, from pri-
mary. Again there has been little change in the seven years. In
spite of all propaganda from dietary cranks and from commercial
interests, it is clear that the American people depend to an over-
whelming degree upon animal sources for their fat intake, rather

196

PEARL— STAPLE COMMODITIES AND

than upon vegetable oils, nuts and the like. This condition is natu-
rally reversed as regards carbohydrate. Ninety-four per cent, of
this nutrient comes from primary sources and only 5 from second-

80

^

1

r^

\]

\
\

1

\
\

1

p

!>

1 ^

1

\

1 1

1

1 i

1

1 1

i

1

! 1

1

1

PFVTtIN

r-AT

CARBOHYDRATE- CALORICS

l/^/?OM PPlMAPy SOUffCfSi

rf?OM SeCOr/DARY SCOURCtS

Fig. 2. Diagram showing the percentages of the total nutritional intake of
the American people derived from primary and secondary sources.

NATIONAL FOOD CONSUMPTION.

197

ary. In the total nutritional calory intake 6i per cent, comes from
primary foods and 39 per cent, from secondary.

It is interesting to compare the percentage of American nutri-
tional intake derived from primary and secondary sources with cor-

I

0:!

BO .

m

t

PROTt/N

r-AT

CALORltS

U/i/TtD 3TA7£-S

lUniT^D H//iGDOM

Fig. 3. Diagram showing the relative proportions of the American and British
food intake derived from animal sources (exclusive of fish).

198 PEARL— STAPLE COMMODITIES AND

responding British figures. Calculating roughly from Table I. of
the official British report^ on the subject I find that 42 per cent, of
the protein intake, 92 per cent, of the fat intake, and 35 per cent, of
the energy value of the total nourishment of the population of the
United Kingdom comes from secondary sources. In other words,
the British get less of their protein and calories and more of their
fat from animal products exclusive of fish than the Americans do.
The differences, however, are not very great, indicating generally
similar dietary habits in the two populations, a fact which we know
on general grounds to be true.

The above comparisons regarding primary and secondary sources
of human food are shown graphically in Figs. 2 and 3.

The next problem concerns the relative proportion of the total
nutritional intake furnished by the several different large food com-
modity classes. The data on this point for the main groups are col-
lected in Tables IV.-VII. and IX.-XII. inclusive. The arrangement
of these tables is to give first the annual average for the six years pre-
ceding the entrance of the United States into the war, and then to give
1917-18, our first year in the war, separately. The reason for sucli
a time division is obvious. There is no reason to suppose that the
consumption of food in this country was affected by the war till
the time we entered and the United States Food Administration
began its work. Before then the population had gone on consuming
food at about the usual normal rate. There was no reason or in-
centive to do otherwise, except in so far as price had an influence.
But in 1917-18 a^ wholly new and extraordinary influence was
brought into play to alter the national food habits. This was the
Food Administration, which through its recommendations, on the
one hand, and regulations on the other hand, sought to modify the
consumption rate of certain commodities and succeeded in doing so,
as will presently appear in detail.

In Tables IV. to VII. the percentage figures are first given sepa-
rately and then accumulated to 100 in another column.

The data of Tables IV. and VII. are shown graphically in Fig. 4.

^ " The Food Supply of the United Kingdom." A report drawn up bj' a
Committee of the Royal Society at the request of the Board of Trade. Lon-
don (Cd. 8421), 1917, pp. 35.

NATIONAL FOOD CONSUMPTION.

199

TABLE IV.

Consumption of Protein in Human Foods, Primary and Secondary, in the
United States, Arranged by Groups in Order of Magnitude.

Average for the Six Years, 1911-12 to 1916-17.

Group.

Absolute
Consump-
tion of
Protein
(Metric
Tons).

Percentage
Consump-
tion.

Cumulated
Per Cent.

For 19 1 7-1 8.

Absolute
Consump-
tion of
Protein
(Metric
Tons).

Percentage
Consump-
tion.

Cumulated
Per Cent.

Grains

Meats

Dairy products. . .
Poultry and eggs.

Vegetables

Fish

Oils and nuts ....

Fruits

Oleomargarine . . . .
Sugars

1,316,140

964.117

744.784

246,178

214,404

84.852

57.839

24.965

838

455

36.01

26.38

20.38

6.74

5.87

2.32

1.58

.69

.02

.01

36.01
62.39

82.77
89.51
95-38
97.70
99.28

99-97

99.99

100.00

1.303.348

945.277

788,969

285,413

248*772

102,022

85,021

23,621

1,808

439

34-44

24.98

20.85

7-54

6.57

2.69

2.25

.62

•05
.01

34-44
59-42
80.27
87.81
94-38
97.07
99-32
99-94
99-99
100.00

Total 3.654.572

100.00

3,784,690

100.00

From these tables and diagrams it is seen that the grains stand
at the head of the hst in contribution of protein, carbohydrate and
calories. Meats come first in contribution of fat, second in protein
and calories. Thirty-six per cent, of our protein intake normally
is in the form of grain, 26 per cent, in meats and 20 per cent, in

TABLE V.

Consumption of Fat in Human Foods, Primary and Secondary, in the
United States, Arranged by Groups in Order of Magnitude.

Group.

Average for the Six Years, 1911-12 to
1916-17.

Absolute
Consump-
tion of Fat
(Metric
Tons).

Percentage
Consump-

Cumulated
Per Cent.

For 1917-18.

Absolute
Consump-
tion of Fat
(Metric
Tons).

Percentage
Consump-
tion.

Cumulated
Per Cent..

Meats

Dairy products. . .
Oils and nuts. . . .

Grains

Poultry and eggs.
Oleomargarine. . .

Vegetables

Fruits

Fish

Sugars

2,591.613

1,405,918

623,385

203,585

173.349

57.965

21,126

20,242

18,096

o

50.66

27-49

12.19

3-98

3-39

I-I3

.41

•49
•35

0

50.66
78.15
90.34
94-32
97.71
98.84
99-25
99-65
100.00
100.00

2,602,187

1,505,129

781,461

227,807

175,220

125,024

27,864

17,866

17.381

o

47-48

27.46

14.26

4.16

3-20

2.28

•51

•33

-32

0

47.48
74-94
89.20
93-36
96.56
98.84

99-35
99.68
roo.oo
100.00

Total '5,115,279

100.00

5.479.939

100.00

200

PEARL— STAPLE COMMODITIES AND

TABLE VI.

Consumption of Carbohydrates in Human Foods, Primary and Secondary,
IN THE United States, Arranged by Groups in Order of Magnitude.

For the Six Years, 1911-12 to 1916-17.

For 1917-18.

Group.

Absolute
Consump-
tion of Carbo-
hydrate ( Met-
ric Tons).

Percentage
Consump-
tion.

Cumu-
lated Per
Cent.

Absolute
Consump-
tion of Carbo-
hydrate (Met-
ric Tons).

Percentage Cumu-
Consump- lated Per
tion. Cent.

Grains

Sugars

Vegetables

Dairy products

Fruits

Oils and nuts

Meats

Fish

9,208,939

4,193,095

1,421,851

899,691

627,487

57,097

4,907

20

0

0

56.1073

25-5473

8.6629

S-4815

3-8231

•3479

.0299

.0001

0

0

56.1073

81.6546

90.3175

95-7990

99.6221

99-9700

99.9999

100.0000

100.0000

100.000 0

9,291,445

4,374,194

1,844,682

917.169

601,350

102,231

4,717

25

0

0

54.2224 54.2224

25.5266 ' 79-7490

10.7651 90.5141

5.3524 95-8665

3-5093 99-3758

.5966 99-9724

.0275 99-9999

.0001 100.0000

Poultry and eggs

Oleomargarine

0 100.0000
0 100.0000

Total

16,413,087

100.0000

—

I7.i35.8i3

100.0000

—

dairy products. These three great commodity groups together
make up nearly 83 per cent, of our total protein intake.

The total consumption of human food was higher in 191 7-18
than the average of the preceding six years. This is to be expected
from the greater prosperity of the people incident to high wages,

TABLE VII.

Consumption of Human Foods, Primary and Secondary, in the United

States, in Terms of Caloric Value, Arranged by Groups

IN Order of Magnitude.

Group.

For the Six Years, 1911-12 to 1916-17.

Absolute Con-
sumption
(Million Cal-
ories).

Percentage
Consump-
tion.

Cumu-
lated Per
Cent.

For 1917-18.

Absolute Con-
sumption
(Million Cal-
ories).

Percentage
Consump-
tion.

Cumu-
lated Per
Cent.

Grains

Meats

Dairy products. .

Sugars

Vegetables

Oil and nuts ....

Fruits

Poultry and eggs
Oleomargarine. .
Fish

Total

45,057.265
28,104,069
19,834,010

17.196,595
6,910,026
6,269,270
2,862,540
2,620,311
542,719
534.509

129,931,314

34-68

21.63

15.26

13.24

5-32

4.82

2.20

2.02

.42

.41

100.00

34-68
56.31
71-57
84.81
90.13
94-95
97-15
99.17
99-59
100.00

45.569.323

28,122,722

21,010,397

17,939,129

8,999,132

8,104,157

2,722,604

2,648,262

1-170,593

533.419

33-31
20.55 +
15-36
13-II

6.58

S-92

1.99

1-93
.86-

•39

33-31
53-86
69.22

82.33
88.91

94-83
96.82

98.75

99.61

100.00

— 136,819,738

NATIONAL FOOD CONSUMPTION.

201

PROTtin consuMPT/on

ZO 30 40 so 60

6PAI/iS
MEATS
DA//?Y PffODUC,
POULJfry s t6<;5
VtetTABLtS

nsH

OIL5 <a nUTS
r-RUITS
OLtOnARBA/fint
3U6ARS

1

\

^^^^w

■^

^

5

b

1

ne-ATs
cwRY /^oDucrs

OlL5&nUT5

6RA/ns

POOLT/FYsf-iSeS

OLtOflAPSARl/yi

V^TSfTABL£-S

rRU/TS

t/SH

3U3ARS

5
5

FAT consunPTioM
PtR CE-nr
2o 30 40 so eo

ZZI

CAPBOHYDPATt COnSUMtD
PtR CtriT

lO 20 30 40 £0 eo

1

SUGARS

VKtTABLtS

^^^vl

mRyp/?0Doc7s

FRUITS
0/L5 & nUTS

\

ne-ATs

ri5H

P0ULTFY<&tG6S

OL£OnARGARjn£

CALORits consuMPTion
ptp ctriT

/O 20 30 40 so eo 70

1

_

ntATS

\

^^^^^^w

DAJffy RRODOCTS

vab^

SUGARS

wmkt^

Vt6tTA5trs

W,

O/IS a nUT5

Wa

rPv/Ts

5

POULTPY&£i56S

5

OLtCnAROARint

nsti

^^ 191 1- 17 V7772 I3n -/&

Fig. 4. Showing the percentage contribution of the dififerent great lood
commodity groups to the nutritional intake of the United States, for (a) six
years before our entry into the war, and (&) 1917-18.

etc. But the proportion of the total contributed by the grains and
meats is smaller in 1917-18. In other words, the two great com-

202

PEARL— STAPLE COMMODITIES AND

modity groups on which the most stress was laid in the conservation
campaign show a reduction in the part which they play in national
nutrition. The effect of the conservation work will, however, be
more clearly shown when we come to the consideration of individual
commodities.

Of the fat normally consumed 51 per cent, is furnished by the
meats as a group ; 27 per cent, by the dairy products ; and 12 per
cent, by the vegetable oils and nuts. The grains normally furnish
3.98 per cent, of the fat intake and in 1917-18 this rose slightly to
4.16, due to the increased consumption of corn meal.

The sugars stand second in the list as contributors of carbohy-
drate to consumption, with 26 per cent, of the total, to which 56 per
cent, is furnished by the grains. Of the remainder of the carbo-
hydrate intake vegetables normally contribute about 9 per cent., the
dairy products 5 per cent., and the fruit 4 per cent.

The energy values of the groups are especially interesting as
furnishing a general index of food values. Of the total energy fur-
nished by the human food consumed 35 per cent, comes from the
grains, 22 per cent, from the meats, 15 per cent, from the dairy
products and 13 per cent, from the sugars. These four groups
make up about 85 per cent, of the total energy value of all the food

TABLE VIIL

Showing the Changes in Food Consumption in the United States in

1917-18 AS Compared with the Average Annual Consumption in

Six Preceding Years (Millions of Calories).

Group.

Increase of
Consumption in
1917-18 Over 6
Year Average.

Decrease of
Consumption in
1917-18 tjnder 6
Year Average.

Percentage
Increase.

Percentage
Decrease.

Grains

Meats

Dairy products. .

Sugars

Vegetables

Oils and nuts . . . .

Fruits

Poultry and eggs.
Oleomargarine. . .
Fish

512,058

18,653

1,176,387

742,534
2,089,106
1,834.887

27.951
627,874

139.936

1,090

1. 14

.07

5-93

4-32

30.23

29-37

1.07
115-69

Total .

6,888,424

5-30

Population 1 5,662,979

5-73

NATIONAL FOOD CONSUMPTION.

203

consumed. Vegetables contribute only about 5 per cent., fruit and
poultry about 2 per cent, each, and vegetable oils and nuts nearly 5
per cent.

On the basis of Table VII. it is of interest to examine somewhat
more carefully the changes in consumption rate in 1917-18 as com-
pared with the average of the six preceding years. Such a com-
parison is made in Table VIII. and shown graphically in Fig. 5.

Fig. 5. Diagram showing the increase or decrease in food consumption in
1917-18 as compared with the average of the preceding six years.

From Table VIII. and the diagram it is observed that the total
increase in human food consumption in 1917-18 was less (nearly 3^
per cent.) proportionately than in the increase in population, both
being compared with the average of the six preceding years. The

204

PEARL— STAPLE COMMODITIES AND

consumption of meats practically did not increase at all, and the
consumption of grains only about i per cent.

The great increases were first in the consumption of vegetables
and oils and nuts, amounting to 30 per cent, in the one case and 29
per cent, in the other, and second in oleomargarine where the con-
sumption increased nearly 116 per cent, in 1917-18 over the average
of the preceding six years. In the case of vegetables and oils and
nuts the increased consumption in 1917-18 is probably to be attrib-
uted largely to the activity of the Food Administration in urging

TABLE IX.

Consumption of Protein in Human Food, Primary and Secondary, in the
United States, Arranged by Commodities in Order of Magnitude.

Order

No.

Commodity.

Average for the Six Years, 1911-12
to 1916-17.

Absolute
Consump-
tion of Pro-
tein (Metric
Tons.

Percentage
Consump-
tion.

Cumu-
lated Per
Cent.

Absolute
Consump-
tion of Pro-
tein (Metric
Tons).

Percentage
Consump-
tion.

Cumu-
lated Per
Cent.

3

4
5
6

7
8
9
10
II
12
13
14
IS
16

17
18

19
20

Wheat

Dairy products . .

Beef

Pork

Poultry and eggs .

Corn

Potatoes

Fish

Legumes

Nuts

Mutton

Other cereals . . . .
Other vegetables .

Rice

Rye

Cocoa

Apples

Other fruits

Bananas

Oranges

Oleomargarine. .

Sugars

Oils

1,054.548

744.784

528,709

392,665

246,178

202,717

114.598

84,852

69,669

46,819

43.712

30,471

30,137

17,231

11,174

11,020

8.719
7,620

6,979

1,647

838

455

o

28.85
20.38

14.47
10.74
6.74
5-55
3-14
2.32
1.91
1.28
1.20

•83
.82

•47
•31
•30
.24
.21
.19
.04
.02
.01

28.85

49-23
63.70

74-44
81.18
86.73
89.87
92.19
94.10
95-38
96.58
97.41
98.23
98.70
99.01
99-31
99-55
99-76
99-95
99.99
100.01
100.02
100.02

940,543
788,969
539.703
378,799

248,772

242,395

143.167

105.578

85,021

81,939
65,088
36,668

30,725
28,298

24-597
20.083
9.283
7.458
5-771
1,808
1,109

439
0

24-85

20.85

14.26

10.01

6-57

6.40

3-78

2.79

2-25

2.16

1.72

•97
.81
•75
-65
• S3
-25
.20

-15
-05
•03
.01

24.85
45-70
59-96
69-97
76-54
82.94
86.72

89.51
91.76
93-92
95-64
96.61
97-42
98.17
98.82

99-35
99.60
99.80

99-95
100.00
100.03
100.04
100.04

Total '3,654,572

3,784,690

the consumption of these commodities to afford a relief of the
pressure on wheat and meat products. In the case of oleomarga-
rine the increased consumption is clearly due entirely to a favorable

NATIONAL FOOD CONSUMPTION.

205

price differential as compared with butter and lard, taking into
account palatability.

The only two great commodity groups showing decreases in
consumption in I9i7-i8are fruits and fish. In both cases the result
is probably to be explained by price influences, taken together with
palatability and popular ideas as to relative necessity in the diet.
For example the price of meat may rise relatively much more than
that of fruits or fish without leading to any reduction in consump-

TABLE X.

Consumption of Fat in Human Foods, Primary and Secondary, in the
United States, Arranged by Commodities in Order of Magnitude.

Order

No.

Average for the Six Years 1911-12 to 1916-17.

For 1917-18

Commodity.

Absolute
Consump-
tion of Fat
(Metric
Tons).

Percentage]
Consump-
tion.

Absolute
Consump-

Percentagel Cur

Cumulated^ tion of Fat Consump- 1 lated
Per Cent. (Metric ] tion. Per Cent.
Tons).

13
14
15
16

17
18

19
20

Pork

Dairy products . .

Oils

Beef

Poultry and eggs .

Corn

Nuts

Wheat

Mutton

Oleomargarine . . .

Cocoa

Fish

Other cereals ....
Other vegetables .

Apples

Other fruits

Potatoes

Legumes

Bananas

Rye

Rice

Oranges

Sugars

Total 5,115,279

,024,236

,405.918

505,201

505.033

173.349

97.355

92,348

91,929

65.499

57.965

25.836

18,096

12,391

9.935

8,629

7.713
6,366
4,824
3.488
1.479
431
411

39-57

39-57

27.49

67.06

9.88

76.94

9.87

86.81

3-39

90.20

1.90

92.10

1. 81

9391

1.80

95-71

1.28

96.99

1. 13

98.12

•51

98.63

•35

98.98

.24

99.22

■ 19

99.41

.17

99.58

.15

99-73

.12

99-85

.09

99-94

.07

100.01

.03

100.04

.01

100.05

.01

100.06

0

100.06

—

—

2.045,653

1,505,129

554.851

513.596

179.337

175,220
125,024
118,845
81,835
47.273
46,853
23,104
17,866
12,586

7,953
7,451
7,325
6,767
3.256
2,885

767
278

— 5.479.939

37-33

27.46

10.13

9-37

3-27

3-20

2.28

2.17

1.49

.86

.86

.42

-33

•23

37-33
64-79
74.92
84.29
87-56
90.76
93-04
95-21
96.70
97-56
98.42
98.84
99.17
99.40

• 15

99-55

.14

99-69

.13

99.82

.12

99-94

.06

100.00

-05

100.05

.01

100.06

.01

100.07

0

100.07

—

—

tion, owing to the general belief that meat is a more necessary article
of diet than the other two sorts of food mentioned.

We may next consider the gross consumption of individual com-
modities on the same plan that has just been used in handling the
groups. The data are given in Tables IX. to XII. inclusive. In

206

PEARL— STAPLE COMMODITIES AND

these tables it will be noted that the cumulated percentage columns
run to more than loo per cent, by trifling amounts. This is to take
care of the item "other meat products" which appears in the net
export table but not in production (in basic tables not here given).
In the main consumption table it is carried into the subtotal
" Meats," but does not appear as a separate item, because of the
impossibility of calculating it as such.

TABLE XL

Consumption of Carbohydrate in Human Foods, Primary and Secondary,
IN THE United States, Arranged by Commodities in Order of Magnitude.

Order
No.

Commodity.

Average for the Six Years, igii-12
to 1916-17.

Absolute Con-|
sumption of Percentage

Carbohydrate
(Metric
Tons).

Consump-
tion.

Cumulated
Per Cent.

Absolute Con-
sumption of
Carbohydrate
(Metric
Tons).

Percentage
Consump-
tion.

Cumulated
Per Cent.

3

4
5
6

7
8

9
10

13

14
15
16

17
18

19
20

Wheat

Sugars

Corn

Potatoes

Dairy products.

Apples

Other vegetables

Legumes

Other fruits ....

Rice

Other cereals . . .

Rye

Bananas

Nuts

Oranges

Cocoa

Pork

Beef

Mutton

Fish

Oils

Poultry and eggs
Oleomargarine .

6,944.794

4,193,095

1,805,564

935,881

899,691

312,589

303.868

182,103

171.574

170,151

159. "3

129.318

111,628

35.571

31.696

21,526

2,740

1,707

484

20

o

42.3125

25-5473

11.0008

5.7020

5-4815
1.9045

1-8514

I. 1095

1.0454

1.0367

.9694

.7879

.6801

.2167

• 1931
.1312
.0167
.0104
.0029
.0001

42.3125

67.8598

78.8606

84.5626

! 90.0441

I 91.9486

93.8000

94-9095

95-9549

96.9916

97.9610

98.7489

99.4290

99.6457

99-8388

99.9700

99-9867

99-9971

100.0000

100. 0001

100. 0001

100. 0001

6,195,182

4,374,194

2,155.310

1,169,204

917,169

398,275

352,857

303-428

284,668

277,203

268,425

219.237

92,328

63.054

39.177

21,360

2.859

1.577

315

25

o

o

o

36-1534
25.5266
12.5778

6.8232

5-3524
2.3242
2.0592
1.7707

1. 6612
1. 6177

1.5665
1.2794
.5388
.3680
.2286
.1247
.0167
.0092
.0018

.0001

36.1534

61.6800

74.2578
81.0810
86.4334
88.7576

90.8168

92-5875
94.2487
95.8664
97-4329
98.7123
99-2511

99.6191

99.8477
99.9724
99.9891
99.9983,

1 00.0001
100.0002
100.0002
100.0002
100.0002

Total 116,413,087

17. 135. 813

The data of Tables IX. to XII. inclusive are shown, exhibited
graphically, in Figs. 6 to 9.

Taking first the protein consumption, as given in Table IX., we
see that wheat stands at the head of the list as a source of protein
for the population of this country, contributing nearly 28 per cent.

NATIONAL FOOD CONSUMPTION.

207

TABLE XII.

Consumption of Human Foods, Primary and Secondary, in the United

States, in Terms of Caloric Value, Arranged by Commodities

in Order of Magnitude.

Order

No.

Commodity.

Average for the Six Years, 1911-12
to 1916-17.

Absolute Con-
sumption (Mil-
lion Calories).

Percentage
Consump-
tion.

Absolute Con-: Percentage
Cumulated juj^ption (^ji- Consump-
Per Cent, ijqjj Calories. tion.

Cumu-
lated Per
Cent.

3
4
5

6

7
8

9
10
II
12
13
14
15
16

17
18

19
20
21
22
23

Wheat

Pork

Dairy products.

Sugars .■

Corn

Beef

Oils

Potatoes

Poultry and eggs
Other vegetables

Apples

Nuts

Legumes

Other cereals . . .
Other fruits ....

Mutton

Rice

Rye

Oleomargarine .

Fish

Bananas

Cocoa

Oranges

33.657.299

20,453,649

19,834,010

17,196,595

9,141,678

6,892,851

4,700,590

4.366,750

2,620,311

1,465,344

1,403,750

1,196,911

1,077,932

893.383
800,831
791,032
774.430
590.475
542,719
534.509
519,109

371.769
138,850

25.90

15-74

15.26

13-24

7-03

5-30

3-62

3-36

2.02

I-I3

1.08

-92

•83

.69

.62

.61

.60

•45
.42
.41
.40
.29

25.90
41.64
56.90
70.14

77.17

82.47
86.09

89-45
91-47
92.60
93-68
94.60

95-43
96.12
96.74
97-35
97-95
98.40
98.82
99-23
99-63
99-92
100.03

30,021

21,010

20,594

17.939

10,938

7.017

5.455

5.162

2,648

2,262

1,927

1,904

1.638

1.381,

1.299,

1.205,

1,170,

994,
678,
553
533.
429
93

979
397
616
129
521
398
418
528
262
988
964
998
716

039
820

454
593
221
641
498
419
360

569

21.94

15-36

15.05

13-11

7-99

5-13

3-99

3-77

1.94

1.65

1. 41

1-39

1.20

1. 01

•95
.88
.86
•73
•50
.40
•39
•31
•07

21.94
37^30
52.35
65-46
73-45
78.58
82.57
86.34
88.28

8993
91-34
92.73
93-93
94-94
95-89
96.77
97^63
98.36
98.86
99.26
99.65
99.96
100.03

Total 129,931,314

136,819,738

normally to the total. Dairy products are second with 20 per cent,
of the total. Beef with 14 per cent, and pork with 11 per cent,
stand next. The other commodities contributing more than 2 per
cent, to the total protein intake of the population are, in the order
named : Poultry and eggs, corn, potatoes, and fish. Taken together,
these eight commodities furnish 92 per cent, of the total protein
intake. We see that a very few commodities furnish a very large
percentage of the nutritional intake. This fact, in and of itself,
helps enormously towards the possibility of making a study such as
this substantially accurate in its results. It is clear that the minor
items omitted from our calculations have no significance in the final
general result. If four food commodities furnish nearly 75 per
cent, of the total protein ingested, it is obvious that a large error.

208

PEARL— STAPLE COMMODITIES AND

or even the entire omission, of single ones of the other minor items
can have but httle effect.

Comparing the order of the commodities in 191 7-18 with the
average of the six preceding years, it is seen that the only change
of position among the eight commodities normally furnishing over

f^i^CtrfTAOt CaiT/?/BUT/On TO TOTAL PROTtffi COTiSUMtD
Pt-R CttYT

20 ^5

30

35

4C

__^

^///////////////////////////////A

(II

>,

///////////////////////.^y

Bttf-

I^^^H

■

FORK

V'//^^/////'

1

PDULJI?Ya£GeS

CORn

rnrnk^

POTATOtS

^

nsH

B

LtOUMtS

m^

nuT6

i

MUTTOn

!

OTHtRCtPtALS

h

omt/f vtetTABLts

5

R/Ct

\

RYt

\

COCOA

1

1

APPLtS

^MAmc//^^ Al/t/?A6t 6 YtAf?S

5 0THtR TWOS

cona/ntD

!

t: ' ' \/9/7-/3

Fig. 6. Diagram showing tlie percentage of the total protein consumed in
the United States contributed by each of 22, commodities. The solid ba'rs
denote the average consumption in the six years preceding our entry into the
war. The cross-hatched bars denote the consumption in 1917 and 1918.

NATIONAL FOOD CONSUMPTION.

209

90 per cent, of the protein is in respect of the last one on the list,
namely, fish. In 1917-18 the legumes (beans and peas) moved up
to the eighth place and fish moved to the ninth place.

Turning to the fat consumption, it is seen that approximately 40
per cent, of the total fat in the nutritional intake of this country

PtRCt/yTA6t CO/YTT^fBirr/O/y TO TOTAL /W COiSUt^tD
P^CtrfT

■5 10 ■ /S go £5 30 :3^ 4^ ^S

PORK

DA/RY PRODUCTS

OILS

Bttr-

POULTRY & fTSeS

ZCRM

NUTS

Wt/tAT

TIUTTOf^

OLtOMARGARlNt

COCOA

t/S/i

UOTHtR TOODS

cor/B/ntD

/////////////////////////////////////A

p

y/////////////////////////////:yz7///////////////77A

1

i

ImAnnUM A\/£-RAG£- 6 YtAPS

Fig. 7. Diagram showing the percentage of the total fat consumed in the
United States contributed by each of 23 commodities. The solid bars denote
the average consumption in the six years preceding our entry into the war.
The cross-hatched bars denote consumption in 1917 and 1918.

comes from pork and its products. The hog is in a class by itself
as a source of fat for human nutrition, with the population of this
country. Dairy products stand second in the list, with approxi-
mately 27^ per cent, of the total. After the dairy products there
is a considerable drop in percentage contribution as we pass to the
next item on the list, namely, the vegetable oils, which normally

PROC. AMER. PHIL. SOC, VOL. LVIII, N, JULY 26, IQiQ.

210

PEARL— STAPLE COMMODITIES AND

furnish only about lo per cent, of the fat intake. Beef contributes
almost exactly the same percentage. The four commodities named
together furnish nearly 87 per cent, of the total fat intake. Only

ft/?CmTAGt CO/iT/?/BUTI(yi TO TOTAL CARBOHYDRATt COTiSUfim
PtR ctrrr

_5_ lO_ I5_ £C 25 30 35 40 4-5

SU6AR

.

y//////////////////////////^^^A

co/?/y

POTATOtS

/////////A

DA/ffY PRODUCTS

^^^

APPLtS

i

OTHtR l/terABLtS

%

LtGUMtS

\

OmtR F-RUITS

\

RICt

^

CTHtP CtPfAlS

%

RYt

1

BAMmS

UtMA/y.^a/ll AVtRAGt 6 YtA/fS

IC OTHtR moDS
coMB/ntD

5

V// A 19/7- 19/8

Fig. 8. Diagram showing the percentage of the total carbohydrate con-
sumed in the United States contributed by each of 23 commodities. The
soHd bars denote the average consumption in the six years preceding our
entry into the war. The cross-hatched bars denote the consumption in 1917
and 1918.

one other commodity group — namely, poultry and eggs — furnishes
more than 2 per cent, normally.

In 1917-18 there are some changes of significance in the relative
position of the commodities as fat contributors. The first four
items, pork, dairy products, oils and beef, stand in the same order
in 1917-18 as in the six years preceding. Nuts moved up in 1917-
18 to the fifth place from the seventh, which they had occupied

NATIONAL FOOD CONSUMPTION.

211

Pt/?CtnTA6t C0/yT/?/W7/0/i TO TOTAL CALO/F/tS COnSUfltD
Pt-R ctrrr

26 30 -as ■40 45

WHtAT

■

:'/.V///////////////////y A

1

pom

'■'///////////////,

DA/RY PRODUCTS

y////////A

SUGAR

■ -X//. //////////A

CCRN

^^^^^

Bttf-

^^

1

0/L5

^^

POTATOES

^^

POULTRY &^S6S

OTHtR VtOtTABLtS

5

AFPLtS

5

nuTs

■:

LteuMts

5

OTHtR CtR&AlS

b

OTHtP PRUITS

5

MUTTOn

f

RICt

6 OTHtP J^OODS

conantD

WM^yt^lMl AI/t/MGt e YtARS
\/Y'Af9n-/8

Fig. g. Diagram showing the percentage of the total energy vakie of the
food consumed in the United States contributed by each of 23 commodities.
The solid bars denote the average consumption in the six years preceding our
entry into the war. The cross-hatched bars denote the consumption in 1917
and 1918.

before. Oleomargarine moved from the tenth place to the seventh.
Corn, in spite of the increased consumption in 191 7, dropped from
the sixth place to the eighth in percentage contribution. Twelve of

212 PEARL— STAPLE COMMODITIES AND

the great commQdity groups before our entry into the war, and thir-
teen in 191 7-1 8, contribute less than i per cent, to the total fat
intake.

In carbohydrate consumption wheat stands at the head of the
list with over 24 per cent, normally. The sugars stand second with
about 26 per cent., and corn with 11 stands next. These three com-
modities, together with potatoes and the dairy products, contribute
altogether 90 per cent, of the carbohydrate intake. There is no
change in the relative position of the commodities falling in the 90
per cent, group in 1917-18 as compared with the average of the six
preceding years.

A noteworthy feature of this Table XL, dealing with carbohy-
drates is the relative position of the sugars. Many persons regard
sugar as a pleasant but not essential part of the dietary. It is ob-
vious enough that this is a mistaken point of view. Any commoa-
ity which furnishes nearly 26 per cent, of the carbohydrate intake
of the population must be regarded as an important essential. To
get an idea of the importance of the sugar relatively it is only neces-
sary to compare it with some of the items farther down in the table.
For example, we see that the sugars contribute more than twenty
times as much to the carbohydrate intake of the nation as does rice.

In Table XII. we get a summarized view of the general nutri-
tional importance of the several food commodities, because here we
are dealing with the energy content as measured in calories. The
order of the products in this table may be taken as the general order
of nutritional significance of the great staple foods in this country.
Wheat stands at the head of the list, contributing nearly 26 per cent.
to the total. Pork comes next with normally 16 per cent., and
dairy products third with 15 per cent., and the sugars fourth with 7
per cent. Then follow corn, beef, the vegetable oils, potatoes, poul-
try and eggs. These nine commodity groups together make up
over 91 per cent, of the total nutritional intake of the population.
The smallest contribution to the total nutrition is made by oranges,
furnishing about %o of i per cent, to the total. Bananas and fish
furnish only about %o of i per cent, of the total, and rye and rice
only a little more.

The changes in 191 7-1 8, as compared with the average in the

NATIONAL FOOD CONSUMPTION.

213

six preceding years, as shown in Table XII., are extremely inter-
esting. The figures show in much more detail than any that we
have had hitherto the precise eft'ects of the conservation and sub-
stitution campaign of the United States Food Administration during
191 7-18. While wheat normally contributes 25.9 per cent, of the
total nutritional intake (as measured by energy value), in 1917-18
it contributed but 21.9 per cent. To go farther down the table, rice,
which normally contributed but 0.6 of i per cent, to the total nutri-
tional intake, contributed i per cent, in 1917-18.

The changes in consumption, as indicated in Table XII., are of
such great interest that it is worth while to examine them more in
detail. To this end a table on the same plan as Table VII. is shown.

TABLE XIII.

Showing the Changes in Food Consumption in the United States in

1917-18 AS Compared with the Average Annual Consumption of

Six Preceding Years for 23 Staple Human Foods

(Millions of Calories).

Commodity.

Increase of Con-
sumption in
1017-18 Over
6 Year Average.

Decrease of Con-
sumption in
1917-18 Under
6 Year Average.

Percentage
Increase.

Percentage
Decrease.

Wheat

Pork

Dairy products . .

Sugar

Corn

Beef

Oils

Potatoes

Poultry and eggs .
Other vegetables .

Apples

Nuts

Legumes

Other cereals. . . .

Other fruits

Mutton

Rice

Rye

Oleomargarine. . .

Fish

Bananas

Cocoa

Oranges

Total net increase

Population

140,967
1,176,387

742,534
1,796,843

224,547

461,938

1,088,668

27.951

439,654

1,066,077
560,784

1,034,581
193,390

606,609

709.345
627,874

306,872

3,635,320

198,296

237.534

1,090
89,749

45,281

.69
5-93
4-32

19.66
3-26
9-83

2493
1.07

89.07

52.02

115.80

24-15

78.33

120.13

115-69

82.54

10.80

14-13

30.03

.20
17.29

32.61

6,888,424

5-30

5,.662,979

5-73

214

PEARL— STAPLE COMMODITIES AND

The data of Table XIII. are exhibited graphically in Fig. lo.
In this diagram the total length of the bars from the 0 line shows
the total percentage increase or decrease in consumption in 1917-18
as compared with the preceding six years. The cross-hatched por-
tion of each bar shows the percentage increase in population, and
therefore the part of the increased consumption to be expected as

ptRctnZ'^tGe incPt^st or DfARtASt /// cohsumptjon

Bl-YJfiD PAPlTY IV.'Tfi POPULAT'On /nCPtASE-

PORT/on Of /rrcptASfv ccnsuriFT/on
dusTiHto Br popuu)riori /rtcfftASf-

Fig. 10. Showing the percentage increase or decrease in consumption in
1917-18 as compared with the annual average of the six years preceding. For
explanation see text.

a result of population increase. Where the black bar is below the
top of the cross-hatched population bar it means a conservation.
Thus the true conservation on wheat amounted to 10.80 -|- 5.73
= 16.53 per cent, of the normal average consumption.

NATIONAL FOOD CONSUMPTION. 215

The table and diagram bring out very clearly the effectiveness of
the Food Administration's campaign for conservation and substi-
tution in foods. It will be noted at once that the commodities show-
ing great increases in consumption in 191 7-18 over the preceding
years are, for the most part, those which the Food Administration
urged to be substituted for articles of which the supply was less
abundant, and for which the needs of the Allies were greater.
Thus, rye, which constituted the most popular of the substitutes
for wheat in the public mind, shows the greatest increased consump-
tion in 1917-18. Next to it stands the ''Other cereals" of our
classification, including barley and buckwheat. Nuts, rice and the
vegetables generally show increases beyond the population increase,
showing that the people very generally followed the suggestions of •
the Food Administration to consume more of these products and
save Vv'heat. The articles on which the Food Administration most
strongly urged conservation — namely, wheat, beef, mutton, pork
and the sugars — all show either a consumption actually below the
normal average, or else a very slight absolute increase well below
the population percentage increase. In either case a real and sub-
stantial conservation is, of course, shown. The decrease in con-
sumption of the most popular fruits, oranges, apples and bananas
is largely if not entirely explained by high prices for these products.

We get now to the most interesting stage of any discussion of
food, namely, the per capita per day consumption. Calculating the
results on this basis puts them in a form where we may form a
better judgment of their meaning and compare them with accepted
dietary standards. In this connection it is to be remembered that
hitherto we have had no careful studies on a per capita basis of the
actual nutritional intake of the population as a whole. All dietary
standards are based not on the actual practice of the whole popula-
tion, but rather upon dietary studies made on restricted groups of
selected individuals. While a very large number of such studies
have been made by the United States Department of Agriculture
particularly from ten to twenty years ago, it must be obvious that
since such studies are made on selected small groups they can only
inferentially give any picture of what is taking place in the popula-
tion as a whole. The theory of random sampling makes it clear

216 PEARL— STAPLE COMMODITIES AND

that any inference from dietary studies, as they have been carried
on, to the whole population rests on an exceedingly dubious founda-
tion. It will therefore be of great interest to compare the results
of the present careful investigation of the population as a whole
with the results of previous dietary studies.

In reducing consumption data to a per capita basis it would ob-
viously be foolish to take the actual total population as a base, for
the reason that the amount of food consumed changes with the age
of the individual, particularly in early life. On account of this fact
the usual practice in computations of this kind is to reduce, not to a
per capita basis, but to an adult man basis. In doing this a frac=
tional factor is used to multiply the number of individuals of certain
lower ages, the magnitude of the factor being proportional to the
relation which the nutritional intake of the individual at the younger
age bears to that of an average adult man.

In the present study the following age-intake factors have been
used :

Age in Years. Man Value Factor.

0-5 0.50

6-13 0.77

14-18, male i .00

14-18, female 0.83

19 on, male i.oo

19 on, female 0.83

The man factor values here used have been adopted after care-
ful study of the subject. They differ in detail somewhat from those
adopted by English physiologists in similar calculations, but in the
net end result come to much the same thing.

Applying these factors to the total population of the United
States, and assuming that the age distribution of the population is _
the same in each of the years studied, we get the population in terms
of adult men as set forth in Table XIV. for the midyear point of
each of the years included in this study. The population equiva-
lents in Table XIV. are used for the base for the per capita per diem
calculations which follow.

Before entering on the detailed discussion of per capita con-
sumption figures it is well to recall a fact which is liable to escape
attention, unless special attention is called to it. This is the fact

NATIONAL FOOD CONSUMPTION. 217

TABLE XIV.
Population of Continental United States in Terms of Adult Men.

Population Equivalent in
Year. Adult Men, January i.

1912 79,571,000

1913 80,930,000

1914 82,289,000

1915 83,648,000

1916 85,007,000

I917 86,366,000

I918 87,724,000

that the final figures in this paper, which are called " consumption
fig-ures," really include something mare than consumption in a nutri-
tional sense. They include the food actually eaten plus that which
is wasted by loss in cooking, in garbage, etc. It is necessary to be
entirely clear on this point. In calculating the nutrients in the in-
termediate calculations use has been made of factors which allowed
for inedible refuse, so that all of the inedible portion of the foods
as produced or imported have already been deducted in our calcu-
lations up to this point. Furthermore, gross losses from storage,
spoilage, transportation, etc., have been deducted. Even after all
these deductions have been made, however, it is obvious that there
is still a considerable amount of loss and wastage of strictly edible
material, which might be saved and consumed under a theoretically
ideal system of preparing food for the table plus a conscientious
ingestion of every bit of edible material. Of course, as a matter of
fact, neither of these theoretically ideal conditions at all prevail.
There is a considerable loss of nutrient values in the process of
cooking as ordinarily practiced. This loss is undoubtedly greater
for fats than for any other of the nutrients. It is a troublesome
and time-consuming process for the housewife to conserve and
utilize all of the fat which gets melted and floats about in the water
in which foods are cooked, or adheres to the utensils in which they
are cooked. Nor, in the minds of most people, is there any neces-
sity or desirability of saving this fat. In fact, a great many people
in this country object very strongly to what they designate as
"greasy cooking." Consequently, floating fat of soup stock is
skimmed off and thrown away in the vast majority of instances.

218 PEARL— STAPLE COMMODITIES AND

The result is that in calculations made in the way those of this study
have been made, which include the total nutrient value in the edible
portion of food materials, after deducting inedible waste and
the losses which accrue up to the time the food reaches the con-
sumer, there is bound to be an apparently high consumption of
fats. The figures here presented are really statements of .consump-
tion plus edible waste and should be so regarded.

Another important factor is that of edible waste in garbage :
that is to say, the uneaten portion of the prepared food which is
edible and might be consumed, but is not for reasons of taste, over-
estimation of ingestive capacity, etc.

It is quite impossible to arrive at any accurate estimate of what
the amount of losses of nutrients in cooking and in avoidable wast-
age of edible material is. On the first point it would be extremely
difficult ever to gather accurate data because the practice of house-
wives and cooks varies so enormously in this regard. That a great
deal can be accomplished in reducing the amount of edible material
going into the garbage can has been demonstrated with both the
civilian and the Army population of the United States during the
past year.^°

The recent study of Murlin (loc. cit.) gives the data regarding
edible waste obtained from the nutritional surveys of the training
camps. The average figure for 213 messes show that 7 per cent,
of the protein supplied was wasted, 9 per cent, of the fat and 6 per
cent, of the carbohydrate. Because of special conditions surround-
ing the investigation, however, and because of the differences of
camp life, these figures are not at all applicable to civilian conditions.
Looking at the matter from the national point of view, it seems
probable that of the protein in human foods left in the country for
consumption in the statistical sense, it is safe to say that 5 per cent.
is lost in edible wastage ; of the fat left in the country for consump-
tion as human food, it is believed that at least 25 per cent, is lost
through wastage. This figure seems large, but it probably under-
go Pearl, R., " Statistics of Garbage Collection and Garbage Grease Re-
covery in American Cities," Jour. Ind. Eng. Chcni., Volume 10, page 927, 1918,
and Murlin, J. H., " Diet of the U. S. Army Soldier in the Training Camp,"
Jour, .liner. Med. Assoc, Vol. 71, pp. 950-951, 1918.

NATIONAL FOOD CONSUMPTION.

219

estimates rather than overestimates the fact. Of the carbohydrates,
probably there is 20 per cent, of edible wastage.

The total statistical consumption (ingestion plus edible wastage)
of human food in the United States, by years from 191 1 to 1918,
is shown on an " adult man " per capita basis in Table XV.

TABLE XV. ■
Summary of Consumption per Adult Man.

Protein.

Fat.

Carbohydrate.

Calories.

Year.

Per
Annum

(Kilos).

Per Day

(Grams).

Per „ T^
Annum P^r Day

(Kilos). 1 (Grams).

Per

Annum
(Kilos).

Per Day

(Grams).

Per Annum
(Kilos).

Per Day

(Grams).

1911-12

1912-13

1913-14

1914-15

I915-16

1916-17

1917-18

44.70
44.04
45-08
43-05
44-48
43.01
43-14

122
121
124
118
122
118
118

62.12
60.44
60.22
63.42
61.22
62.45
62.47

170
166
165

174
168
171
171

195.48
198.68
209.25
193-42
200.48
189.94
195-34

536
544
573
530
549
520

535

1,563,450
1,558,232
1,591,621
1,560,326
1. 574.621

1.536,833
1,559,661

4.283

4.269

4.361

4.275

4.314 ■

4,211

4.273

Average, whole
period

Average, 1911-
12 to 1916-17

43-91
44-05

120
121

61.78 169
61.65 169

197-45 541
197.82 542

1.565.075
1,566,032

4,288
4.290

Applying the estimated percentage deductions for edible wastage
stated above to the per capita average for the whole period we have
the following results for ingested human food :

114 grams protein per man per day,
127 grams fat per man per day,
433 grams carbohydrate per man per day,
3424 calories per man per day.
These figures are probably very close to the fact as regards pro-
tein and carbohydrate. They are undoubtedly somewhat too high
still as regards fat, because the edible wastage of this component
is higher than the 25 per cent. used. The intention, however, has
been to use the most conservative figures in estimating waste.

For purposes of comparison Table XVI. is inserted. This table
is based upon certain American dietary studies analyzed m the
writer's statistical laboratory.

The general agreement of these results with those set forth in
the present study, which were reached by totally different procedure,

220

PEARL— STAPLE COMMODITIES AND

is evident. The statistical estimate of per capita protein consump-
tion over the whole population is distinctly higher than in this small
group. The fat consumption is higher but not by so large an
amount as protein. The farmers and professional men show a
higher net energy intake than the general average for the whole
country, which would, of course, be expected. Mechanics are a
little lower than the average for the country in energy intake.

In any case there is one fact which must not be lost sight of,
namely, that while the figures of Table XV. do in fact represent

TABLE XVI.
Summary of Some Dietary Studies in ii Groups and ii6 Families.

No. of
Fami-
lies.

Average
Yearly
Income.

Days

per

Man.

Per Man per Day.

Pro ein,
Grams.

Fat,
Grams.

Carbohy-
drate,
Grams.

65

472

8i

495

102

479

121

420

120

419

113

460

125

430

148

438

128

395

III

405

131

506

Energy,
Calories.

Mother wage earners . . .

Garment makers

Laborers

Retired

Clerks (office)

Mechanics

Teachers

Professional men

Engineers (professional)

Salesmen

Farmers

32

17

5

5

$ 640

724

1.497

1,647

1.934
2,133
2,150
2,208
2,253
2,527

212
168
305
130
225

259
620
438
97
121

384

105
109
94
81
92
97
88
99
85
90
102

2,89s
3.145
3.210
3,095
3,125
3,245
3.195
3.480
3.070
2,980
3,610

Average 1 1.6

$i,77i"| 260

95

113 i 447

3-185

,11 Average of 104 families (farmers excluded).

ingestion and waste, it still is true, and the constancy of the figures
in successive years proves its truth, that to maintain naturally a
contented feeling in respect of nutrition the population actually uses
up the amounts of nutrients indicated in Table XV. To make these
gross consumption figures materially less would require a profound
readjustment of the dietary and culinary habits of the people, fixed
by centuries of usage. Discussion of the minimum protein, fat and
carbohydrate requirements of a nation are in considerable degree
academic if they base themselves upon net consumption rather than
gross consumption. A considerable excess over any agreed upon
minimum physiological requirements must always be allowed, be-

NATIONAL FOOD CONSUMPTION.

221

cause there will inevitably be, in fact, a margin between actual gross
consumption and net physiological ingestion or utilization. The
present study, through the figures summarized in Table XV. gives a
clearer and probably more nearly exact picture of what this margin
between net and gross consumption must be, in a population of the
habits of the American people, than has hitherto been available. It
may well be theoretically true that a man needs only 75 grams or
50 grams of protein per day to sustain life and health, but in actual

^_

^ ^

•

/9II-

/2 /QIZ

-/d /ati.

-J4 /9/4

-/£ /9IS

-/6 /Q/e

-/7 /9I7-

je

Fig. II. Diagram showing the energy value in calories of the gross consump-
tion of human food, per adult man per day.

222 PEARL— NATIONAL FOOD CONSUMPTION.

fact the American man uses up, in one way or another, about 120
grams a day. Furthermore, if the last seven years' experience is
any criterion, he will continue to use up about 120 grams per diem
until such time as his general habits of life are in some manner
rather profoundly changed. Doubtless they can be changed. But
until they are, one must count on supplying about 120 grams of
protein per day to each man equivalent component of the population.

The data of Table XV. for calories are shown graphically in
Fig. II.

From this diagram it is apparent that there has been only a very
slight decrease in per capita gross food consumption since 191 1.
Even this probably does not mean that the population is eating any
less, but that because of the gradually rising prices through all this
period there has been a narrowing of the margin between gross and
net consumption ; or, put in another way, there has been a slight
reduction in the wastage of edible foods.

Summary.

In this paper are presented a portion of the final results of a com-
prehensive statistical study of the consumption of human food in the
United States in the period from July i, 1911, to July i, 1918. The
results are expressed in terms of nutrients and allowances have been
made for losses of edible material in storage by spoilage, in transpor-
tation, etc., as well as for inedible refuse, and losses in cooking, gar-
bage, etc. The final net figures are believed to represent with sub-
stantial accuracy the course of human food consumption in this
country during the period covered. The net result is to show that
on a per man per day basis the consumption of nutrients in this
population was extremely even and steady during the seven years
covered by the investigation. The amount of total nutrients con-
sumed decreased slightly in the period, but the decrease was insig-
nificantly small. The conservation campaign of the Food Admin-
istration produced notable changes in 1917-18 in the proportionate
contribution of different classes of food materials to the total nutri-
tional intake.

SELF-LUMINOUS NIGHT HAZE.

By E. E. BARNARD.
(Read April 25, 1919.)

In the Proceedings of the American PhilosopJiical Societv, Vol.
50 (1911), p. 246, a paper was contributed by the present writer on
the subject of "Self-Luminous Night Haze," where observations
were given of a kind of luminous, streaky haze seen at night in 1910
and 191 1 which seemed to have no connection with the ordinary
aurora. Apparently this is a little-known phenomenon. It is well
worthy, however, of record and study. The only reference to it
that I have found, and it was not known to me at the time of my
previous paper, is an account of what was probably the same phe-
nomenon by T. W. Backhouse, who gives occasional records of it
in a long paper on the aurora, under the name of " irregular au-
roras," as far back as December 13, 1862. I will quote his entire
reference to the phenomena as given in No. II, p. 109, Publications
of West Hendon House Observatory, Sunderland, England. " * irr.'
means a night when the aurora was only an irregular one, not per-
ceptibly of a magnetic character; I do not know whether these are
true auroras or not. They consist of bands and patches of light,
the positions and direction of which have no particular relation to
either the magnetic decHnation or dip; they appear to lie in a hori-
zontal stratum, but differ from clouds in being self-luminous and
transparent, and also not exactly like any kind of cloud in appear-
ance, and I should think they are at a greater height. They do not
appear specially in the north. As auroras of this class are always
faint, their spectrum proves little ; but I have never made out that
it was more than a continuous one. An examination of this table,
however, shows that these irregular auroras are more frequent about
the time of the maximum of sun-spots than at the minimum, which
goes towards proving that they are of the same magnetic nature as
the ordinary auroras." Air. Backhouse did not seem to notice that

223

224 BARNARD— SELF-LUMINOUS NIGHT HAZE.

this luminous matter is perhaps only ordinary thin haze which, from
an unknown cause, is sometimes luminous at night. I believe this
same haze is also present in the daytime.

In my previous paper to the Society attention was called to the
fact that the objects observed by me did not seem to have any con-
nection with the luminous night clouds of O. Jesse of twenty-five
or thirty years previous. His objects were at very great altitudes,
some fifty-two miles above the earth's surface, and were visible by
direct sunlight which shone on them long after it had ceased to
illuminate the ordinary clouds. They were therefore not self-
luminous. They were only seen at the times of the equinoxes.
The present observations refer to objects entirely self-luminous
which are seen in all parts of the sky where sunlight could not reach
them, and appear to be at no greater elevation than the higher clouds.

It is only during short periods that the luminous condition of
this haze seems to be of frequent occurrence. Apparently it will
be absent for a year or so and then for a short time there will be a
great amount of it. In 191 1 there were frequent displays of it
during the spring, summer and autumn, but from 191 1, September
22, until 191 5, July 2, I have no record of it, though a lookout was
always kept for it. It was frequent in the summer and fall of 1910
and also in the spring of 191 1. Its prevalence seems to be inde-
pendent of any auroral conditions. There are a few records of
aurora when the haze was present, but only one case in which a
large auroral disturbance seems to have been nearly coincident with
it. One is impressed with the idea that it is not necessarily an au-
roral phenomenon and that an appearance of it during an aurora is
perhaps purely accidental. Whether this luminous appearance of
what is undoubtedly haze, is due to electric conditions or to phos-
phorescence of some kind does not seem clear.

Sometimes it appears in rather narrow strips several degrees wide
and many degrees long; sometimes the form is that of broad sheets
covering a large part of the sky. Often both these forms are pres-
ent at the same time. On one or two occasions I have taken it for
the zodiacal light in the east until its motion revealed its true nature.
When seen at the eastern horizon it has at times produced the eft'ect
of dawn. It is frequently brighter than the Milky Way. From

BARNARD— SELF-LUMINOUS NIGHT HAZE. 225

this it appears that the phenomenon is really conspicuous when seen
under good conditions. It seems to have no preference for any
special part of the sky. The motion is always easterly. The den-
sity of this haze is not great enough to hide the brighter stars over
which it passes. In fact, it does not differ from ordinary haze
except in being self-luminous. There does not seem to be any fluc-
tuation or pulsation in its light. Under proper conditions it is
visible at all hours of the night and in all parts of the sky. Some-
times it is very faint and has to be looked for ; at other times it is
conspicuous. Often the sky is very pure and dark between sheets
or strips of it. It continues luminous as long as it is under observa-
tion, which may be for a considerable part of an hour. It is pos-
sible that the luminous nights mentioned, where no streaky haze
was seen, were due to a thin uniform sheeting of the luminous haze
all over the sky.

To make this paper more complete I have written to Dr. W. J.
Humphreys for information as to the name and nature of this haze
shown on one of my day photographs. He has kindly supplied me
with the following:

The streaky haze to which you refer is the cirro-stratus cloud. Near
the edge, where it is thin, it might better be called cirrus.

There is no sharp division between cirrus and cirro-stratus. The thin
fibrous cirrus often gradually thickens into a more or less continuous cloud
veil, or sheet, in which form it is called cirro-stratus.

Its average altitude in middle latitudes is 8 to lo kilometers. In higher
latitudes its altitude is less, say 6 to 8 kilometers; and in equatorial regions
roughly ID to I2 kilometers.

It nearly if not quite always consists of snow crystals, as might be in-
ferred from its altitude and consequent low temperatures, and as is known
from the halos often seen in it.

I am also indebted to Professor Eric R. Miller, of the Weather
Bureau at Madison, Wisconsin, for valuable data on the frequency
of cirrus cloud in Wisconsin. Professor ]\Iiller says :

You will learn from the data sent you that there is no part of the year
when cirro-stratus is absent; that it is fairly evenly distributed throughout
the year.

This information is important as it shows that if the cirro-
stratus were naturally self-luminous we should have luminous haze

PROC. AMER. PHIL. SOC. VOL. LVIII, O, JULY 25, I919.

226 BARNARD— SELF-LUMINOUS NIGHT HAZE.

more frequently at night. If there is no error in my identification
of this form of cloud with the luminous haze, it would seem that the
material must be frequently present, but that the cause of its lumi-
nosity is much less frequent.

Fig. I. Photograph (x^ugust 19, 1917) believed to show the day appearance
of the luminous night haze.

The accompanying photograph shows the streaky haze (form-
ing the background for the cumulus clouds) to which Dr. Hum-
phreys refers in his letter and which the present writer believes to
be of a similar nature to the haze that sometimes is luminous at
night. That shown in the photograph would doubtless be so dense
as to blot out the stars at night. Unless, therefore, the haze were
much thinner at night it could not be the same. I have seen in the
day time a thin gauzy streaky haze, sometimes in large sheets, at the
time that the regular denser cirro-stratus clouds were present.
These would more readily represent what I refer to as luminous
night haze. This I infer is still some form of the cirrus cloud. It

BARNARD— SELF-LUMINOUS NIGHT HAZE. 227

is very important that no mistake be made in the identification and
I therefore would call attention to the objection to the denser form
of cirro-stratus. This photograph, therefore, represents a denser
condition of that cloud and would probably obliterate the stars at
night.

In the previous paper (mentioned before) there is a suggestion
that the probable passage of the earth through the tail of Halley's
comet had something to do with the luminous condition of this
haze, because it made its appearance very soon after that event.
With the later information there is now no need to call upon the
comet for an explanation of the phenomenon.

Following is given a continuation of my records of luminous
haze. The times are Central Standard Time, which is 6'^ o" slow
of Greenwich mean time.

191 1, June 17. Traces of it were suspected before moonrise,
but not certain.

Aug. 28, 10'^ 15™. Masses of it were extending over the bowl
of the Great Dipper. They were irregular and very extended to
the right and left of the Dipper. The motion was slow, to the right
horizontally. At 10^ 35"^ it extended the full length of the Dipper
and 15° to the right and 10° to the left of it; visible for some time
and feebly luminous ; steady in its light. At 14'' 35™ there was a
long, rather bright luminous strip 4° wide extending along the bowl
and handle of the Dipper ; horizontal and quite bright. This was
perhaps auroral. A few minutes later it had gone. No other signs
of aurora.

Sept. 15, 9'^ 15™. Bright auroral streamers from below the
horizon, but no arch. In various parts of the sky there were broad
streaks and areas of luminous haze. Several very long, broad
strips through the Dipper extending east and west. At 9^" 25" there
were thin sheets of it all over the sky, one passing over Brooks'
comet which I was photographing, and another one parallel to the
first 5° north of it. They extended east and west and were drifting
northerly. At 9'^ 55™ there were great streams and areas of it with
dark spaces between. There was a little auroral light low in the
north. At lo'^ 10"^ the two long masses of haze had drifted north.
There were masses of it all over the sky, specially to the south.

228 BARNARD— SELF-LUMINOUS NIGHT HAZE.

Possibly there was also a slight auroral light, but it was mostly
luminous haze. This was the most striking display of luminous
haze I have seen. In the intervening spaces the sky was pure and
dark. The aurora seemed dead at moonrise, when I stopped the
exposure on the comet.

Sept. i6, S^ 30™. The sky in the southeast was luminous nearly
as high as Gamma Pegasi, like approaching moon rise. There was
no streakiness in the sky, but I think there was uniform luminosity.
The sky was not like that of the night before. No aurora at any
time.

Sept. 22, 10'^ o"". There were great broad and extended masses
of luminous haze over the northwest with clear, dark spaces be-
tween. Great areas of it were also present south of the zenith. No
aurora.

191 5, July 2. No aurora, but the night was uniformly luminous.
July 12, 11'^ 30"^. No aurora. Luminous haze for 10° below

Polaris to Gamma Ursse Majoris. A feeble sheeting of it extended
to the north horizon ; its upper edge was quite definite. Above it
the sky was dark. No aurora. This haze seemed to be confined
to the region described. The other parts of the sky were free from
it. At 11^ 50™ the luminous haze had drifted to the right and
toward the horizon. It was then horizontal and about 20° below
Polaris. Above it the sky was dark, while below it was as bright
as the Milky Way to the right of it. At 12^ 25"^ there seemed to be
more of this, especially to the left of and below Jupiter. There
was no aurora. The sky was not very pure. 13^ o™: In the east,
all about the stars in Aries was a large luminous region almost as
bright as the ]\lilky Way ; very large, like a great diffused cloud.
There were no fluctuations in its light. At 13*^ 30" this haze was
about, above and to the right of the Pleiades. It was as bright as
the Milky Way. It was large and almost like dimly luminous
clouds. There seemed to be a considerable amount of it near the
Great Dipper and to the right of the dipper under the pole.

1916, Feb. 2, 14'' 20'". No aurora, but the sky very luminous — a"
luminous night. The brightness increased near the horizon.

Aug. 6, 13*^ 30™. There was a region of luminous haze under
the pole which seemed to be densest 5° above and 5° to the right

BARNARD— SELF-LUMINOUS NIGHT HAZE. 229-

of Alpha Ursae Majoris. There did not seem to be any evidence
of ordinary aurora, but this must have been auroral. It looked as
if a feeble moonlight were shining on large hazy clouds.

Dec. 4, 15'' 50"". Horizontal strips of luminous haze parallel
with the horizon under Cassiopeia in the northwest. Sky whitish
everywhere, but no definite aurora. 16'' 10™: Like dim dawn all
under Cassiopeia. It consisted of sheets and strips of thin haze,
like hazy clouds illuminated by a quarter moon — very noticeable.
This was drifting toward the northeast. At iG'' 57™ the upper defi-
nite edge of this haze passed just below the Pleiades. It was 5°
below and extended 20° to the right of Cassiopeia to under the
Hyades. At 17'' 10™ it was nearing the northwest horizon. This
was the best specimen of luminous haze in many years.

Dec. 21, 8*^ 40™. There seemed to be a long horizontal strip of
luminous haze 3°-4° wide passing through Zeta Ursse Majoris. No-
regular aurora. 9^ 25"": A considerable amount of broad luminous
haze west of the previous strip.

1917, Aug. 10, 9*^ i8"\ Large faint auroral streamers from the
north horizon. 9'' 34™ : Several slender auroral streamers. lo'^
50™ : Large areas of luminous haze in the east and near Beta Ursse
Minoris. 10'' 57™ : From the east horizon up to Alpha Arietis was-
a great sheet of faintly luminous haze slanting upward at the south
end. A large area of luminous haze through Beta Ursse Minoris
to the north horizon, several degrees wide. Feeble luminous haze
in high east sky. There were dark areas of pure sky free from
luminous haze. Another feeble luminosity in the low north. There
were no auroral streamers. By 11^ 20™ the luminous haze had
drifted to the east of Polaris. Large areas of it all over the sky.
It could be seen though the moon was up. (Moon last quarter
August 9.) This luminous condition was not due to moonlight. It
was the regular luminous haze seen in previous years.

Aug. 15, 8^^ 50". Some masses of luminous haze. A great
V-shaped portion with one broad leg through the Square of Pe-
gasus, and the other 10° south of Altair. They met near the hori-
zon south of the square of Pegasus. These were perhaps " left
overs " from the previous night's aurora, but there was no trace of
aurora to-night. At 9^^ 55"^ the great stream of luminous haze lay

230 BARNARD— SELF-LUMINOUS NIGHT HAZE.

between Cassiopeia and Perseus, and was io° under Polaris. It
was io° broad. Another ran through Alpha Cassiopeia and met
the first near the southwest horizon. There were several streams
of it from the intersection running upwards. These were as bright
as the Milky Way near Alpha Cygni. The sky between was very
clear. At ii'' lo™ the luminous band in the northeast was between
Alpha Persei and Capella. It was bright — brighter than any part
of the Milky Way. There were many less bright strips to the east
of the zenith. No aurora. (On August 14 there was a large au-
rora.) 12^^ 45™ "• All the sky except overhead was unequally covered
with luminovis haze. In the east and northeast there were great
long streaks of it, 10° or more wide. Also in the south. This was
as bright as the Milky Way. It was very strikingly dark where the
sky showed between.

Aug. 16, lo*^ o™. A very feeble glow low in the north, like a
very faint aurora. No luminous haze to-night so far. At 10'' 20""
there was a broad strip of luminous haze running from Altair east
by north through a point in a 22'^ 20"\ 80°. It was 10° broad and
quite noticeable when looked at for a few minutes. 11'' 40™:
Strong aurora with streamers and very low arch. 12'' 20™: Aurora
not very active, but strong glow. At 12'' 50™ there was only a
strong auroral glow.

Aug. 18, 10'' 55™. No aurora, but sky luminous ; perhaps some
luminous haze in the south. 15'^ 52'": East edge of a broad sheet
of luminous haze passing over. Alpha Arietis. 15'^ 58™: The south
edge of this was passing the Pleiades. 16'' 2"^: East edge over Ju-
piter. 16'' 23"^: It was now massing up into little patchy clouds.
There was a great amount of it all over the sky. It was pale white.
There were some dark clouds lower in the east against the dawn.

Aug. 19, 10'' 15"' to 10'' 45™. Looked carefully for luminous
haze, but none was visible. Sky very thickish ; INIilky Way very
dull. 15'' 15'": Sky very dull. No luminous haze or clouds. 15'^
33"' : No luminous haze. Sky thickish ; specially thick over Jupiter
and the Pleiades. Some misty haze — not luminous. Some zodiacal
light at 15" 42™. Some feeble luminous haze above Jupiter and
the Pleiades. A great deal of dark haze ; a long mass of it was
probably feebly luminous in Cassiopeia and evidences of it at a

BARNARD— SELF-LUMINOUS NIGHT HAZE. 231

number of other places. 15'* 50™: There were streaks and masses
of haze feebly luminous near the Pleiades ; some strong strips and
great sheets of it all over the sovith. These were not self-luminous,
but were all brought into view, evidently, by reflecting the first
feeble dawn. They were, however, beyond question the same stuff
that had recently appeared luminous at night.

Aug. 24, 12^ 50"^. No aurora. There were horizontal strips
of luminous haze through the upper part of the Great Dipper with
a clear space below them, with more luminous haze lower down.
Great long masses through Hercules with clear space below ; then
more of it to the northwest horizon. 14'' 30"^: A rather strong
strip of luminous haze just above the handle of the Great Dipper
in the north. There was also a horizontal strip across Vega with a
clear space below it. No aurora. The sky where there was no
luminous haze was very clear, is*" 50"^: No aurora.

Dec. 21, 14'' 10™. No aurora. 14^ 44": No aurora. I noticed
a little later what appeared to be a strong zodiacal light passing
through Spica in the southeast at 14'^ 57"^ ; this proved to be an
extended luminous haze. At 15'' 30"" the upper part of it passed
through Alpha and Epsilon and Beta Crateris and Gamma Hydrae.
The upper edge was rather definite. The whole mass extended
from the east horizon to beyond Crater. The motion of all this
was to the southeast. The sky was dark below Spica to the edge
of the haze. There was much, very diffused light extending above
Spica for 20° or more. 16^^ 25™: A diffused luminosity like dawn
all along the southeast horizon and upwards where the luminous
haze had drifted, no aurora. 17'' 5"": There was a dawn-like light
all about Alpha and Beta Librje. 17^' 30™: No aurora. The zo-
diacal light was very strong. No luminous haze anywhere. This
exhibition of luminous haze was very noticeable. At first it was
taken for the zodiacal light and thought to be unusually bright, but
it drifted to the southeast and disappeared before the zodiacal light
manifested itself. The light of the luminous haze was soft, like
that of the zodiacal light, but stronger, and there was a rather defi-
nite edge to it. There was no evidence of aurora at any time,
though looked for carefully. There was no other luminous haze.
This was extended from the east horizon to above Spica at first, to
the southeast, almost horizontally, beyond Alpha Crateris.

232

BARNARD— SELF-LUMINOUS NIGHT HAZE.

1 91 8, Sept. 2, 9^^ I/"". A strip of luminous haze 4° wide ran
diagonally across the Square of Pegasus from Beta and beyond
Gamma Pegasi. 9'' 24^ : A similar mass ran from the Dolphin to
Alpha Pegasi. No aurora. 10'' 35™: No aurora. A region of
luminous haze over the handle of the Dipper in the northwest.
There seems to have been a great deal of it on this night. 11^ 25":
Luminous haze all over the north ; a great sheet of it with clear
spaces here and there. No aurora on this night, but a great deal
of luminous haze.

Sept. 8, 9'' 30™. At this time there was a great deal of luminous
haze all over the northwest, as high as 4° or 5° above the Dipper.
It was quite strong. A long strip of it 5° wide extended from
above Gamma Pegasi to Aries. Though there was no aurora on
this night, there was a great deal of luminous haze.

Oct. 5, 12'' 30™. The sky was luminous. There w^as no aurora.
It was about as bright as if a quarter moon were, say, in the west.
It was uniform and w^as apparently not the regular light of lumi-
nous haze. Could readily read my watch by this light. The ]\Iilky
Way was very dim with it.

I have gone to some length in the descriptions and locations of
the masses, etc., of luminous haze because so little is known of the
origin of its light that it is believed the observations may sometime
be more important than they appear to be at present.

The following tabular scheme will roughly show the frequency
of luminous haze as indicated by my records. The numbers indi-
cate the number of nights on which it was seen during the month.

1910. 1911. 1912. 1913- I9I4- 1915- 1916

1917. 1918

January. . .
February. .
March . . . .

April

May

June

July

August . . .
September .
October. . .
November .
December .

As will be seen from the above table, January, April and May

BARNARD— SELF-LUMINOUS NIGHT HAZE. 233

seem to be the only months in which the haze was not observed. It
was not seen during the years 1912, 1913 and 1914. The greatest
amount of it was recorded in October, 1910, and August, 191 7.
The first half of 1910 must not be considered because the haze was
not noticed until June and as it was not looked for it may have
been present.

A strict comparison cannot be safely made between the fre-
quency of luminous haze and that of the aurora by noting the rela-
tive number of nights on which they were visible, because often
auroras appear in the presence of bright moonlight in which the
luminous haze would not be visible. Nevertheless, it may be im-
portant to have even a rough comparison. For this purpose, there-
fore, I have collected the following dates on which auroras were
seen, which covers the entire period over which the luminous haze
has been under observation by me. They are from my own records
of the aurora which I have kept here for over twenty years. The
dates alone are given. A [ ?] means uncertainty as to whether the
observation was of a real aurora. Some of the auroras are of very
short duration and hence the apparition of an aurora on the date of
an observation of luminous haze may not have been coincident with
the latter by many hours. The details of the above auroral records
will be published later as a continuation of the two lists already
printed in the Astrophysical Journal, Vol. 16, p. 135, 1902 and Vol.
31, p. 208, 1910.

May

3

Oct.

2[?]

^lar.

20[?]

<<

18

(<

7

u

24

((

24

<(

27

((

30

((

27

((

30

Apr.

8

(t

30

Nov.

I

<<

16

June

19

IC

29

May

31

July

4

191 1, Feb.

20

June

4

Aug.

9

<(

21

July

17

((

10

((

22

Aug.

2

Sept.

6

((

23[?]

<i

22

((

10

(<

24

ii

23

((

27

((

28

i(

24

((

30

- Mar.

19

(<

28

234 BARNARD— SELF-LUMINOUS NIGHT HAZE.

Sept.

15

July

I

Aug.

22

(<

19

«

5

((

23

((

21

«

8

((

26

Oct.

10

((

II

((

27[?]

1912, Jan.

12

Aug.

6

<<

28

May

13

<(

17

Sept.

2

June

7

((

19

a

4

a

8

Sept.

9[?]

<(

7

Nov.

14

«

16

<(

16

((

15

((

21

«

22

1913, Feb.

13

((

28

«

30

Mar.

13

<(

29

Oct.

I

Aug.

29[?]

Oct.

i[?]

((

5

Sept.

22

<(

14

((

6

1914, Apr.

18

Nov.

I

((

7

July

24

(<

5

Nov.

20

((

29

((

8

Dec.

27

Aug.

28

((

II

<(

28

Sept.

15

((

16

191 7, Jan.

7

a

17

Dec.

7

(C

25[?]

i(

22

1916, Jan.

10

Feb.

17

<(

2S[?]

<(

22

<(

19

((

28

Mar.

8

((

27

Oct.

20

((

10

Mar.

i9[?]

(f

27

«

16

"

20

Nov.

15

<(

29

]\Iay

I

((

16

Apr.

I

<<

2

191 5, Apr.

2[?]

"

5

June ■

13

(<

7

((

27

July

I

(<

16

IVIay

22

a

12

<<

17

<(

30

<i

28

((

25

June

19

Aug.

8

May

16

<(

24

(I

9

June

16

July

5

<<

10

<<

17

<(

8

(<

14

((

21

((

23

"

16

((

24

Aug.

19

<<

25

BARNARD— SELF-LUMINOUS NIGHT HAZE. 235

Oct.

13

((

24

<(

27

Nov.

18

Dec.

8

((

14

1918, Jan.

13

Feb.

12

Mar.

3

((

7

(<

14

(>

15

Apr.

3

<<

4

<(

5

<(

6

10

Yerkes Observatory,

Williams Bay, Wisconsin,
April 4, 1919.

Apr.

17

Oct.

3

i(

29

"

6

July

30

<(

8

Aug.

2

((

9

11

5

((

15

(I

24

((

16

it

25

Nov.

10

11

26

<c

II

((

28[?]

((

12

<c

31

((

13

Sept.

7

((

24

((

12

a

29

<<

15

a

30

((

26

Dec.

3

((

28

a

7

((

29

a

8

a

30

a

19

Oct.

2[?]

1919, Jan.

3

GRAPHICAL REPRESENTATION OF FUNCTIONS OF
THE A^TH DEGREE.

By FRANCIS E. NIPHER.
(Read April 26, 1919.)

It has been generally assumed that any algebraic symbol y, the
exponent of which is vmity, can only be properly represented graph-
ically by a line. When the exponent is 2, the quantity 3'- has been
assumed to represent a rectangular area, the four sides of which
each have a length 3'. If the exponent is 3, the proper geometrical
representation was said to be a rectangular volume, having six faces
whose areas were 3'-. If the exponent were 4, the philosophic mind
began to speculate on the nature of space having four dimensions.
No definite solution of this problem having been reached, it has
apparently been considered impracticable to consider the properties
of space having six or eight or ten dimensions.

During this period of uncertainty it has been known to students
of elementary algebra that

I 3;l/3 3,2/3 y y/3 y/3

^ — etc.

yi.10 y-l^ y ytlU ^U / U ,,? '

,1/3 ^,2/3 ., .,4/3 .,5/3 .,2

It has also been known that similar right triangles have sides whose
lengths are directly proportional. This is illustrated in Fig. i where
an arbitrarily chosen length is adopted as a unit of length. Its
length in terms of the centimeter is unknown. Another length is
similarly chosen to represent ys, where y is to be given a numerical
value in terms of the chosen unit, but this value need not be deter-
mined.

These lengths having been laid oft' upon rectangular axes, the
lengths y"'^, y, y*''^, 3^^ 3'', etc., can be laid oft' upon the two axes by a
well-known geometrical method. By means of three parallel rulers,
two of which are linked to the third in the usual manner, the lengths
unity and y may be first adopted and laid oft' on both axes, and the

236

FUNCTIONS OF THE A^TH DEGREE.

237

lengths 3'i and 3'A can then be laid off. By means of two rulers
in parallel, the length 3;^ can similarly be laid off upon both axes.
The lengths f^-, f'", etc., can then be similarly laid off upon the two
axes. Since

y xy x^-y x^y
I X x^ x^

it is evident that any value a""3'"' can be laid oft" by purely geometrical
methods. The exponents n and m may be in decimal form and may
range in numerical value from o.i to 100 if the experimenter so
desires, and all of these values may be represented by lines.

1 y'^^i y \ y^ y~
Fig. I.

y- y^ y^ y y'
Fig. 2.

In Fig. 2 we have a system of squares constructed on lines deter-
mined by the method described above. The areas of the squares
vary from y- to 3;^-. The difference between the areas of the outer
and the inner squares is 3/^" — v".

This value is divisible by y~ — 3'. The value so determined is

3' — r

= y'' + y' + y' + y+ y' + y' + y' ^y' + f- + y-

y-~y

Multiplying both members of this equation by 3'- — 3', it may be put
into the form

4- 3,5 ( 3,5 — y) _|_ 3,4(3,3 _ 3,4)

238

NIPHER— GRAPHICAL REPRESENTATION OF

-\-yHy^ — y-) -\-y^(y^ — y^)
+ 3''(y— 3' ) +3' (y^'—y )•

(I)

The first two terms of the second member of this equation rep-
resent the area of the two outer rectangles between the two squares
having areas y'^^ and y'^°. These rectangles each have a width y^

— 3'^, and their combined length is
^6 __!_ ^,5_ -pj^g remaining terms

represent the remaining strip areas,
between the squares 3;^" and y^,

„18.

/

y y'~ y'^ y* y^ y^

y

Fig. 3 represents a cube, the
volume within the outer surface of
which is y^^. In the lower left-
hand corner is a cube having a vol-
ume y^. Between these cubes are
fifteen blocks filling the space be-
tween the two cubes. If this cube

were to be placed upon the area forming Fig. 2, each block which

stands on edge would cover a rectangular area shown in Fig. 2.

Dividing the difference between the volumes by 3;- — y as in the

former case we have

-2/
Fig. 3.

r

r

, yie -^ ys -^ 3,2

f

y —y

Multiplying by y^ — y as before the resulting terms may be written
.3;3 = 3,63,6(^,6 _y) ^ 3,63,5 Q,6_y 5) _^yY-(y^ — yr')
_j_ 3,53,5 (^5 _y) _|_ 3,53,4(3,5 _y) _|_ 3,43,4(3,5 _y)
_j_ 3,43,4 (3,4 _y) _^3,4y(y_3,3) j^yY(y* — y^)
+ yy(y — 3") -{-y^y-iy^' — y-) -\-y-y-(y^ — y-)

-{-y-y~(y'—y ) -\-y-y (y-—y ) +y y iy-—y )• (2)

FUNCTIONS OF THE A^TH DEGREE. 239

The first term in the second member of the equation represents
the volume of the block at the rear of the cube 3;^^ Its dimensions
are 3;'', y'^ and 3;^ — y^. The second term represents the volume of
the block on the right side of the cube 3;^^. The dimensions are
y^, y^ and 3/*^ — •3;^ The third term gives the volume of the block
covering the upper face of the cube 3/^^. The volume of the three
blocks is 3;^^ — 3;^^. This will be the result obtained by adding these
three terms together. Each of these three terms may be given a
very different interpretation. They are each a difference between
two quantities. They are therefore each the difference between the
cubes of two other quantities,

y^y^{y^ — ■3'^) =3'^® — 3;^".

This result shows that the volume of the block forming the back of
the cube whose volume is 3;^^ is also equal to the volume of three
blocks between the cubes 3'^^ and y'^\ We may therefore write the
value of this term as follows :

+ / 3/^/3 (3;0_3;-/a)

_|_yy33;-/3(y_yy3).

In this case the inner cube would have edges whose length is y^'''^.
This length can be laid off on the three axes by purely geometrical
means, as has been pointed out in the present paper.

The second term of Eq. (2), which represents the volume of the
block on the right side of Fig. 2, may also represent the volume of
three blocks between cubes y^"^ and 3;^". The third term represent-
ing the block at the top of the large cube may also represent the
volume v^" — 3'^^ These nine blocks would fill the volume occupied
by the three outer blocks of Fig. 3.

The thickness of the three shells filling the volume 3;^^ — •3;^^ is
(3;6_3,"/3)^(yy3_y%)_|_(y%_,^5)_ Qf course, all of the terms
of Eq. (2) can be similarly treated, and the operation may be re-
peated on the terms resulting.

Equation (2) may also represent the difference between the
areas of two squares, y^^ and y^, having sides whose lengths are y^
and y^^^. As the second member has an odd number of terms, the

240 NIPHER— FUNCTIONS OF THE A^TH DEGREE.

final term is resolved into two terms by considering it to be the dif-
ference between two squares having areas 3;* and 3;^. The equation
may therefore be written

+ r(r — /) +/(/— 3'')

+ 3'°(3'' — 3^') +3'' (3'' — 3'')

+ 3'' (3'" — 3'") +3'*(3'' — 3'")

+ 3'" (3^* — 3'") +3''(3'* — 3'"")

+ 3''(3'' — 3'") +3''(3'' — 3'')
+ 3;2(y_y/.)+y/.(y_y/.).

Equation (i) may also represent the difference between the vol-
umes 3'^- and 3'" of two cubes having edges whose lengths are 3'* and
3;^^^ That equation may be written

yl2_y2^yY(y^ — y^) -^yY(y^ — y^)-^yY(y^ — y')

^ ySyZ (^y2 _ y2~^ _^ y3y2 (^yZ _ y2 ^ J^ y2y2 (^y3 _ y2 ^

+ y-y-{y-—y ) +30' (y- — y ) +y y iy-—y )

+ 3'3' (3' — 3''''0+3' 3''^' ( 3' — 3''^' ) + 3'''''3''''' ( 3' — 3' ''' ) •

Here the final term (3'^ — 3;") of Eq. (i) has been considered as the
difference between the volumes of two cubes having edges the
lengths of which are y and y"''^.

The last equation may be written

3'"-3'— [(y^)'-(3'=')'] + [(3'^)'-(3'')^] + [(3'^)^-r1 + [3'^-3'^]-

Of course, this discussion does not in any way modify the phys-
ical or geometrical meaning of the units linear foot, square foot or
cubic foot. A contractor who should order of dealers in such ma-
terial 64 feet of garden-hose, 64 feet of sheet tin and 64 feet of
sand or gravel, would not learn that any of his orders had been mis-
understood. Possibly they might not be understood if he were to
order 2*^ feet of each of these materials.

MAGELLANIC PREMIUM

Founded in 1786 by John Hyacinth de Magellan, of London

I9I9

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful K'Nowledge

ANNOUNCES THAT IN

DECEMBER, 1919

IT wr::,L award its

MAGELLANIC GOLD MEDAL

TO THE author of THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING TO NAVIGATION, ASTRONOMY, OR NATURAL PHILOs6PHY (mERH
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 19I9, deliver free of postage
or other charges, his discovery, invention or improvement, addressed to the
President of the American Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance by some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the same motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author,

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publicly
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as may be after the adjudication, either in a separate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

A.11 correspondence in relation hereto should be addressed
To THE Secretaries of the

AMERICAN PHILOSOPHrCAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. At

Proceedings

OF THE

American Philosopliical Society

Subscription — TbreeDollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

Price, One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

For Promoting Useful Knowledge

JVew Series, Vol. XXII,
Fart 111, 4to, 44 pages.
{Lately Published)

Tertiary Vertebrate Faunas of the North Coalinga Region of Cali-
fornia. A Contribution to the Study of the Palaeontologic
Correlation in the Great Basin and Pacific Coast
Provinces. By John C. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California,

Subscription— Five Dollars per Volume
Separate parts are not sold

(iddress The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A.

/j;i/f-T-

PROCEEDINGS

OF THE

American Philosophical Society

HELD AT PHILADELPHiIa ' '

FOR PROMOTING USEFUL KNOWLEDGE

Vol. LVIII.

1919.

No. 4.

CONTENTS

PAGE

The Crocker Eclipse Expedition from the Lick Observatory, June 8,

191S. By W. W. Campbell ....... 241

The Expedition of the Mount Wilson Observatory to the Solar Eclipse

of June 8, 1918. By J. A. Anderson ...... 255

The Lowell Observ-atory Eclipse Observations, June 8, 1919. Prominences

and Coronal Arches. By CO. Lampland ..... 259
The Flash Spectrum. By S. A. Mitchell ..... 265

Photo-electric Photometry of the 19 18 Eclipse. By Jakob Kunz and

Joel Stebbixs .......... 269

The Sproul Observatory Eclipse Expedition, June 8, 1918. Coronal

Arches and Streamers. By John A. Miller .... 272

Results of Observations of the Eclipse by the Expedition from the Yerkes

Observatory. By Edwin B. Frost . . . . . .282

PHILADELPHIA

THE AMERICAN PHILOSOPHICAL SOCIETY

104 South Fifth Street

1919

The American Philosophical Society

announces that an award will be made in the year 192 1 of the

HENRY M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

TLq Ccntrol of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent., of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after the
award.

Attention is called to the following regulations governing the award
of the Prize:

Competitors for the prize shall affix to their essays some motto or name
(not the proper name of the author, however) and when the essay is for\varded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judges, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essays shall be in their author, subject to
the right of the Society to publish the crowned essay in its " Transactions "
or " Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,

David Jayne Hill,

Simeon E. Baldwin,

John Cadwalader,

W. W. Keen,

William B. Scott,
President ex-officio.

Committee.

The essays musl be sent, addressed to the
President of the American Philosophical
Society, No. 104 South Fifth Street, Phila-
delphia, U. S. A.

THE CROCKER ECLIPSE EXPEDITION FROM THE LICK
OBSERVATORY, JUNE 8, 1918.

Some Eclipse Problems.

By W. W. CAMPBELL.
(Read April 25, 1919.)

Eclipse observers have always before them two principal ques-
tions :

What is the solar corona?

Why does the sun have a corona ?

It must be confessed that the best answers at present available
are far from satisfactory and complete. We have established many
important facts, especially as to the nature of the corona, but these
are more or less isolated facts, with suggestions here and there as
to the relations of the facts to each other. That the progress made
has not been greater is due to the tantalizingly short time available
for observation. In the past twenty-one years I have had a total
of twelve minutes in which to secure observations of the corona,
and during four of those minutes — in Spain — the observations were
made through thin clouds.

It is not a narrow interest which directs our efiforts. The corona
is a part of the sun, and we can never claim to know what the sun
is until w^e understand all parts of it, including the corona, and the
reasons for the corona's existence. The sunspots, the faculse, the
flocculi, and the prominences are undoubted evidence of great activ-
ity of movement in the sun's outer strata. It seems not too much to
hope that a thorough understanding of the corona will contribute
greatly to an. understanding of the sun's circulatory system.

Again, it is not alone the more thorough understanding of our
own sun which supplies the motive for eclipse study. Is there a
corona around every sun? There may be; solar coronas may be
plentiful throughout the universe, but we do not know. A complete
understanding of our own star is certainly a pressing duty to the

PROC. AMER. PHIL. SOC, VOL. LVIII, P, JULY 30, IQip. 241

242 CAMPBELL— CROCKER ECLIPSE EXPEDITION.

investigator who looks toward the understanding of the stars in
general.

The Form of the Corona.

Near the end of last century eclipse observers recognized that
the outline of the corona is a function of the sunspot phase. Our
first slide (No. i) illustrates well this general relationship. The
coronas here shown, proceeding from right to left, are, first, that
of 1889, at sunspot minimum ; second, the corona of 1893, at sun-
spot maximum; third, the minimum corona of 1900; and fourth, the
maximum corona of 1905. In all of these the sun's equator is
nearly vertical with reference to the slide, and the sun's rotation
axis is nearly horizontal. The phenomena in question are illus-
trated in a more interesting manner by the second and third slides
of the corona on a larger scale. The circular corona (No. 2) is
that of the 1893 eclipse, and the greatly elongated one (No. 3) is
of the year 1900.

It is perhaps natural to expect that the polar streamers of the
corona should be more extensive at times of great solar activity, as
indicated by the maximum phase of spottedness, but that the equa-
torial streamers should be the longer at sunspot minimum would
scarcely have been predicted.

It cannot be said, however, that the outline form of the corona
is accurately predictable from the known phase of the sunspot-cycle.
We photographed the corona (slide No. 4) with cameras of focal
lengths varying from four feet to forty feet on June 8, 1918, at
Goldendale, Washington. The spot phase was but a few months
past the maximum, and we expected the corona to be essentially
circular in outline. Our predictions were not fulfilled so satisfac-
torily as we had expected. Slide No. 5, an exposure of four sec-
onds with a camera whose focal length was four feet, shows stream-
ers extending out to approximately three solar diameters to the right
and to the left, that is, to the west and east of the sun, whereas the
streamers to the north and south (above and below) are very much
shorter. The sixth slide is one obtained with the forty-foot camera,
exposure four seconds ; the departure from circularity of outline is
very marked even for the inner corona. Does the circular type of

CAMPBELL— CROCKER ECLIPSE EXPEDITION. 243

corona prevail only during a short period of time at a well-defined
phase of maximum spottedness, or is the relation of outline form
to spot phase not very definite? Future eclipses should try to de-
termine the relationship more precisely than we can be said to know
it at present. The equipment needed for this work is inexpensive.
Cameras of moderate length, say from four to seven feet, are abun-
dantly powerful.

There has been no satisfactory hypothesis proposed to account
for the undoubted relationship of coronal form and sunspot phase.

Relationship of Coronal Streamers to Other Solar
Phenomena.

The relations of coronal structure to other phenomena of the
sun, such as the prominences, sunspots, and faculae, are subjects of
inquiry with every corona recorded. Our slide No. 7, the eclipse
of 1898 in India, shows clearly, and I think for the first time, the
hooded forms of coronal structures which encircle some of the
prominences. The following eclipse, that of 1900 in Georgia, did
not show this feature, unless perhaps very mildly in the case of one
prominence. The larger prominences in 1900 were certainly unat-
tended by these hooded coronal forms. Later eclipses showed the
occasional presence of this phenomenon, but not strongly, until we
come to the eclipse of last year, when it was very marked indeed.
Slide No 8 is an exposure of one second on Seed thirty plates ob-
tained with the camera of forty feet focus. Nearly all of the larger
prominences and some of the smaller ones are enclosed by strong
and well-defined hoods. There can be no doubt, I think, that the
prominences and the curved streamers encircling them possess some
kind of intimate relationship. The greater part of the inner corona
recorded by this photograph consists of these curving streamers.

Slide No. 9 is a four-second exposure made with the same in-
strument. The most of the strong coronal structure here shown
appears likewise to be under the control of, or to have intimate rela-
tionship of some kind with, the prominences which seem to lie at
centers of controlling influence. Nearly all of the structure to the
right of the sun is divided into three main compartments, each with
a prominence at the central point of contact with the sun's limb.

244 CAMPBELI^CROCKER ECLIPSE EXPEDITION.

Slide No. lo is an enlargement showing one of these sections on
a still greater scale. The division of this part of the corona into
three sections by the two intervening vacancies is clearly marked.

The corona of 1893, at sunspot maximum, slide No. 11, does not
show this interesting phenomenon to have been present. Why it
should have been observed so much more conspicuously in 1918
than at other time remains unanswered.

My colleague, Dr. Moore, had attempted to correlate the main
features of last year's corona with the prominences then visible and
with the positions of sunspots and faculse as they existed on the day
of the eclipse. The Mount Wilson Observatory has generously lent
its solar photographs secured on that day, and on several days pre-
ceding and following, for this purpose, and we gladly make due
acknowledgment of our indebtedness. Lantern slide No. 12 illus-
trates the position of coronal structure, prominences, sunspots, and
faculas, at the instant of the 1918 eclipse, as measured roughly, or
estimated, from the Mount Wilson observations obtained on the
several consecutive days. The features on the front side and near
the limb of the solar sphere are illustrated on the central circle of
the slide, and the features thought to be on the rear side of the sun
are set down on the right and left wings of the slide. The well-
defined symbols represent the spots and the poorly-defined markings
the faculse. It cannot be said from this sketch that the prominences
were situated above visible sunspots or f aculae. The hooded coronal
forms therefore appear not to have been controlled from centers of
disturbance coinciding with spots or faculas. The vacancy in the
coronal structure directly to the right of the sun is approximately
in a plane passing through the observer, the center of the sun, and
the group of spots and faculse visible in the sketch to the right of
the center, but, in view of the positions of the two principal promi-
nences on the right edge of the sun, occupying the central points of
their hooded structures, we seem not justified in looking further at
present for a relationship between the vacancy in the coronal struc-
ture and the group of spots and faculae. This group is really more
than 40° from the sun's east limb.

Professor Perrine's coronal photographs of 1901, in Sumatra,
recorded what he called a " disturbed " volume of the corona. It is

CAMPBELL— CROCKER ECLIPSE EXPEDITION. 245

shown in slide No. 13. Examining the photographs of the sun ob-
tained on and near the day of the echpse, he found that a prominent
sunspot, in fact the only conspicuous sunspot known to be on the
sun at that time, was situated very close to the sun's limb, and im-
mediately under the center of the disturbance in the corona. The
same slide shows a small prominence to the left of the sun — near
the lower left edge of the slide — with a faint hooded enclosure.

This subject of relations between coronal structure and other
solar phenomena is a very interesting and important one for eclipse
observers of the future to hold in mind.

The Forms of Coronal Streamers.

The force which carries materials into the coronal region may be
volcanic, as Schaeberle and others have suggested, or it may be
chiefly radiation pressure, or a combination of these, or indeed a
force of quite unknown nature. The arrangement of the coronal
materials in the well-defined streamers may be a result of the sun's
magnetic properties, as Bigelow and others have thought, but if so
the principal features of the 1918 corona, especially to the east and
west of the sun, would require that merely local magnetic fields be
in control. The polar streamers certainly are very suggestive of
control by the sun's general magnetic forces, but this influence does
not seem to be the prevailing one in the remaining coronal structure.

Motion within the Corona.

Inasmuch as the corona of one eclipse is different in every detail
from that of the succeeding eclipse, we cannot doubt the existence
of motion and change within the corona. An interesting question
is, how rapidly do these changes occur ? Can they be detected from
a comparison of coronal photographs of the same eclipse, obtained
at two or more stations distributed along the path of totality ? This
is a problem which eclipse observers have held in mind during sev-
eral decades. The eclipse of 1905 seemed especially promising for
an attack upon this problem. Crocker expeditions from the Lick
Observatory were located in Labrador, in Spain, and in Upper
Egypt, with the principal motive the obtaining of coronal photo-

246 CAMPBELL— CROCKER ECLIPSE EXPEDITION.

graphs separated by the considerable intervals of time, with a view
to the detection of motion in coronal structure during the intervals.
Clouds prevented observation in Labrador; excellent photographs
were obtained through thin clouds in Spain, but the Egyptian photo-
graphs were not well-defined, owing to disturbances in the atmos-
phere, probably arising from the highly heated desert conditions of
the middle afternoon. Slide No. 14 compares an interesting region
of the corona as photographed in Spain and Egypt, with cameras of
identical dimensions, focal lengths forty feet. The great promi-
nence, it will be noticed, was sharply defined in Spain (on the left),
but the definition was poor in Egypt (on the right). IMany details
of prominence structure changed during the interval of approxi-
mately seventy minutes, but we are not concerned especially with
changes in prominences, as this is one of their w-ell-known charac-
teristics which may be and has been studied successfully on many
days of the year at home. A comparison of the coronal structure
in the region of the prominence is made difficult by the good defini-
tion in one case and the mediocre definition in the other, but there
is no doubt that many changes occurred in the coronal details while
the moon's shadow was passing from Spain to Egypt. Dr. Perrine
and I were unable to say, however, that the motion was outward or
inward, or that it was due to decrease of brightness for certain de-
tails and increase for others. Expressed dift'erently, there were
evident changes of detailed structure, but these changes were appar-
ently haphazard rather than due to motion of illuminated materials
systematically outward or inward. Nevertheless, that changes did
occur in the interval is certain.

Dr. Moore has compared our 1918 coronal photographs ob-
tained in the State of Washington with copies of those obtained by
the Lowell Observatory Expedition, in Kansas, thanks to Director
Slipher's courtesy. The scales are different, and the atmospheric
conditions were poorer in Kansas, so that comparisons of our origi-
nals with copies of the Lowell photographs are difficult and unsatis-
factory. Nevertheless, many changes in details are noted to have
occurred in the interval of one half hour. Some of these seem to
be the result of motion outward and others of motion inward, but
we are not prepared to say that one or the other motion is the pre-
vailing tendency.

CAMPBELL— CROCKER ECLIPSE EXPEDITION. 247

The search for changes in coronal structure evidently remains
one of the most interesting eclipse problems of the future. The
photographs secured at points as widely separated in time as possible
should be made with instruments of long focus, forty feet or more,
and under conditions as nearly identical as practicable. Slow plates,
such as Seed 23, preferably of the non-halation type, or suitably
backed to avoid internal reflections, should be used, and the program
of exposures at the different stations should be of identical effi-
ciency. The utmost pains should be taken to have the plates in
perfect focus and the diurnal motion accurately allowed for. The
camera should be so designed that heated air within the camera
itself would not detract from the definition. There remains the
factor beyond control, the different states of the atmosphere at the
various stations, and this is a serious factor indeed, as will be recog-
nized by every observer who has endeavored to compare measures
of sharply defined photographs with others of the same object but
less definite.

Copies of the photographs obtained in 1918 by the Lick Observa-
tory have been distributed to other institutions known to be inter-
ested in the subject, and if a comparison of these with photographs
obtained at other stations should reveal definite evidence of coronal
changes in the intervals we shall be very pleased indeed.

Polarized Light in the Corona.

A study of the polarized light in the corona has long been recog-
nized as of great importance. Much remains to be done in this
field of inquiry. The photography of the corona through double
image prisms (slide No. 15) has both advantages and disadvan-
tages. The latter arise in part from the factor of chromatic aber-
ration when utilizing coronal rays having a great range of wave-
length values. The definition under these conditions is unavoidably
not all that should be desired, and some uncertainty in the quanti-
tative results necessarily follows. The method of inclined plane
glass reflectors in front of the coronal cameras, as used successfully
by Dr. Perrine (sHde No. 16), has its advantages. In this slide the
richness of polarized light in the corona is indicated by the greater
vertical dimensions of the right hand image, and the greater hori-
zontal dimensions of the left-hand image.

248 CAMPBELL— CROCKER ECLIPSE EXPEDITION.

A point of first importance in any method of studying the polar-
ized light consists in making sure that the apparatus itself does not
introduce polarization, and thus vitiate and render uncertain the
observational data. One should be especially on his guard when
using quartz and possibly other optical pieces in any way. Before
definitely adopting any form or design of analyzing apparatus, labo-
ratory tests should be made to ensure that this apparatus may he
trusted not to produce its own polarization effects.

The Spectrum of the Corona.

The spectrographic study of the corona at eclipses of the past
two or three decades has led to tolerably definite ideas as to the
quality of the coronal light. That there are shallow strata of incan-
descent gases overlying the photosphere and chromosphere of the
sun is certain. However, the use of the term strata in this connec-
tion may be misleading. Some of the strata seem to be fairly uni-
form in thickness over large arcs of the sun's limb, or change their
thickness very gradually in passing from one region to another, but
in other cases the thickness is very irregular. Slide No. 17 is an
exposure, effectively short, secured with a grating spectrograph in
1918, to record the so-called green coronal ring consisting of radia-
tions with wave-length 5303A. The ring is seen to be " lumpy."
Similar photographs of the ring at wave-length 4231A have shown
it to be more nearly uniform in thickness. The principal conden-
sations in the green ring are adjacent to prominences, but not in
coincidence with them. In the illustration the positions of the prin-
cipal prominences are indicated by the dotted line enclosures lying
ooitside of the green ring. The positions of the 5303 maxima are
indicated by the dotted line extending inward from the ring and
bearing numbers indicating the north and south solar latitudes. The
green condensations are slightly farther from the solar equator than
are the adjacent prominences. If the condensations are related to
the prominences, the relationship is not immediate and intimate.
There is poverty of green material near the north and south poles
of the sun.

The faint dark bands passing through the green condensations
have been noted at earlier eclipses. The condensations are undoubt-

CAMPBELL— CROCKER ECLIPSE EXPEDITION. 249

edly strengthened a little by the superposition of these bands of
continuous light with the green ring. The origin of the continuous
bands has been under discussion. It has seemed to me that the ma-
terials responsible for the green ring yield continuous radiations
more abundantly than special radiations, for it must be remembered
that the latter are condensed in the slender ring, whereas the faint
continuous bands stretch from end to end of the spectrum. The
energy represented by the green ring, and in fact by all the bright
coronal rings, is exceedingly slight in comparison with the sum total
of coronal radiations, and apparently represents but a small fraction
of the radiations proceeding from the bright ring materials them-
selves. It has been suggested that the faint bands passing through
the condensations in the green ring may proceed principally from
prominences ; and, again, that they may have had their origin in
photospheric light passing through depressions in the moon's edge,
owing to the exposures having begun too soon after contact II., or
continued too close to contact III. I do not think the point has
been well taken, for the relationship of the continuous bands to the
green condensations is plainly evident, and there is no apparent rea-
son why these condensations should be related to depressions in the
moon's edge. Again, there is a considerable number of extant coro-
nal spectrograms bearing upon the subject. SHde No. i8 is from
Lockyer's photograph of 1905. The green coronal ring to the right
contains a prominent condensation through which passes a strong
band of continuous radiation. This band is shown on the left in
the H and K calcium region of the same spectrogram. The band
misses the calcium prominences very skilfully. In the same way
the hydrogen images of the prominences on the same spectrogram
do not show the continuous band as passing through them.

The coronal spectrograms of recent decades are essentially
agreed that the continuous radiations of the bright inner corona are
but feebly and inappreciably afit'ected by Fraunhof er absorption, thus
establishing, in my opinion, that the inner coronal materials are
chiefly incandescent and supply us with radiations of their own, to
which reflected or diffused photospheric light makes but a small
addition. The spectrum of the outer corona, on the contrary, say
regions lying ten or fifteen minutes, or farther, from the sun's edge,

250 CAMPBELL— CROCKER ECLIPSE EXPEDITION.

undoubtedly contains the Fraunhofer lines. The interpretation is
that the outer corona is shining, possibly in part by its own light,
but chiefly by virtue of the photospheric rays falling upon the outer
coronal materials. SHde No. 19 was obtained in 1918. The broad
spectrum is that of the sun itself. The spectrum of the corona to
the west of the sun is in the upper narrow band. The green bright
line is shown on the extreme right, and the ultra-violet bright line
at 3601A is very close to the extreme left end. The bright points
on the lower edge of the coronal spectrum are from the promi-
nences. The radiations in tw^o long bright lines at H and K un-
doubtedly originate chiefly in the prominences, and the great length
of these lines is due to the dift'usion of the calcium light in its own
atmosphere. The exposure was in effect too short to record the
spectrum of the outer corona. No dark lines are visible. Spec-
trograms secured by us at earlier eclipses have recorded dark lines
in the outer corona.

Trustworthy observations of this kind require an absolutely clear
sky. It is dangerous to draw conclusions from spectrograms ob-
tained through light clouds. This is illustrated by slide No. 20,
copied from the coronal spectrogram secured in Sumatra in 1901.
The Fraunhofer lines are shown strongly in the outer corona, a>id
likeivise upon the dark moon! Much of the Fraunhofer effect is
undoubtedly due to the diffusion of the sun's photospheric rays by
the thin clouds which covered the sky at that tmie.

The observed polarization effects are in harmony with the spec-
trographic as to the nature of the continuous coronal radiations.

Rotation of the Corona.

A few observers have sought to determine the rotational speed
of the corona as a Doppler-Fizeau effect, by measuring the accurate
wave-lengths of the green bright line to the east and to the west of
the sun. If the corona is rotating with a speed approximating that
of the sun's underlying surface, then the corona adjoining the east
limb of the sun should be approaching us at the rate of two km. per
second, and the corresponding coronal structure west of the sun
should be receding with an equal speed. It cannot be said that the
observations for rotation have been successful, though the dift'er-

CAMPBELL— CROCKER ECLIPSE EXPEDITION. 251

ence of observed wave-lengths is in the right direction. The chief
difficulty lies in the apparent fact that the green line is not strictly
monochromatic. Recalling the lumpy appearance of the green co-
ronal ring, we should perhaps be prepared for the probability that
there is motion within the green ring, such that the line as observed
by us is widened by radial velocity differences, and not reliably
measurable.

The Wave-lengths of the Coronal Bright Lines.

Much remains to be done in determining the accurate wave-
lengths of the coronal bright lines, in preparation for the chemical
identification of these lines. There are at least half a dozen coronal
lines whose position in the spectrum are determinable, with suitable
apparatus and care, much more accurately than they are now known.
Probably the best procedure is the unambitious one of attempting
to determine the position of only one line, or at most two neighbor
ing lines, with one instrument, exposing with as high dispersion as
good judgment dictates; covering with a narrow diaphragm the
region of the plate occupied by the coronal line while impressing
the appropriate arc or spark spectrum of an element, both shortly
before and shortly after the total phase of the eclipse.

The Brightness of the Corona.

The photometry of the corona is a problem worthy of further
attention, especially with reference to coordinating the laws of coro-
nal brightness with the sunspot phase. Studies in this field should
take into account the distribution of the coronal radiations through-
out the spectrum. The contrasting of the spectral photometry of
the outer, middle, and inner coronal structures is a most promising
problem, and the preparing of suitable apparatus, on the basis of
optical parts already existing in abundance, should be a simple
matter.

The Flash Spectrum.

The so-called flash spectrum of the sun's edge at the second and
third contacts should, in my opinion, be observed with instruments
specially and carefully designed, and adjusted with great care, to

252 CAMPBELL— CROCKER ECLIPSE EXPEDITION.

obtain improved data as to the limiting depths of the strata respon-
sible for the various solar absorption lines.

Contact Times.

The times of contact of sun and moon have been more exten-
sively observed than any other eclipse phenomena. It has long
seemed to me that the first and fourth contacts, that is, the instants
of time when the moon's image first touches the sim's, and when
the moon's image finally leaves the sun's, are not worthy of much
attention, and our expeditions have taken no interest in them. The
time of the first contact is bound to be uncertain. The observer
who is watching for it suddenly realizes that the moon has covered
a bit of the sun. How many seconds earlier the contact really oc-
curred he does not know. He is simply aware that it has occurred,
and several seconds of uncertainty are unavoidable. The fourth
contact can be observed with considerably more accuracy, but a first
contact of equal accuracy does not exist to balance it. The second
and third contacts, on the contrary, can be observed quite accurately.
Nevertheless, the brilliant points continuing to exist at contact 11.
after the general outline of the moon's edge has passed beyond the
sun's corresponding edge, owing to depressions in the lunar sur-
face— and similarly at contact HI. — introduce some uncertainty. It
seems to me that very valuable comparisons of the relative positions
of the sun and moon could be obtained by a series of large-scale
photographs, using cameras forty feet or more in length, made
near contacts II. and III., under conditions carefully devised for
reducing the brilliancy of the solar crescents and for the accurate
recording of the observation times.

The Accurate Position of the jNIoon.

It is very advantageous to eclipse observers that meridian obser-
vations of the moon's position be made in months immediately pre-
ceding the eclipse date, as a basis for predicting the time when to-
tality will begin at any station on the shadow path. The eclipse of
1918 came seventeen seconds of time earlier than the Nautical Al-
manacs had predicted three years in advance, three seconds later

CAMPBELL— CROCKER ECLIPSE EXPEDITION. 253

than Professor Tucker's recent observations of the moon's position
had indicated, and two seconds later than the chronometer time set
down on the program prepared an hour before contact II. for the
guidance of the observers.

The Einstein Effect.

The search for the so-called Einstein effect has become an im-
portant eclipse problem. It is well known that recent hypotheses
of the nature of light require that rays from distant stars on their
way to the eclipse observer, and passing close to the sun's edge,
should be drawn toward the sun in appreciable amount while pass-
ing through the sun's gravitational field. This would cause a mi-
nute displacement of the stellar images upon the photographic plate.
Photographs of the region immediately surrounding the eclipsed sun
were secured by the Crocker Expedition on June 8, 1919, and pho-
tographs of the same region of th'e sky were obtained with the same
instrument set up at Mount Hamilton in January of the present
year. If the Einstein effect is a reality, a comparison of the two
sets of plates, one obtained with the sun in the field of observation,
and the other with the sun absent, should show slight and sys-
tematic differences in the angular separation of pairs of stars on
opposite sides of the eclipsed sun's position. Owing to war service
on the part of our astronomer who secured the photographs for this
problem at the eclipse, the plates have not yet been measured. It is
hoped that they will receive attention in the month of May.

In securing both sets of Einstein photographs, the driving clock
should be reliable, and the observer should " guide " in right ascen-
sion on a bright star in the immediate neighborhood of the sun. A
guiding telescope of three, four, or five inches aperture and of focal
length equal to that of the Einstein cameras and making an appro-
priate angle with the axes of the cameras, should be able to pick up
the image of the selected bright guiding star a few seconds before
contact II.

The Vulcan Problem.

In view of Dr. Perrine's results of searches for unknown bodies
revolving around the sun, made under the auspices of the Crocker

254 CAMPBELL— CROCKER ECLIPSE EXPEDITION.

Eclipse Expeditions to Sumatra, Spain, and Flint Island, it seems
sufficient to let the progress of this problem in the immediate future
depend upon the Einstein photographs of the sun's surroundings.

Other Eclipse Problems.

Many other eclipse problems, perhaps of lesser importance from
the astronomical point of view, cannot be considered here for lack
of time. It is hoped that enough ha's been said to show that total
solar eclipses have raised difficult and interesting questions.

Lick Observatory,

University of California.

THE EXPEDITION OF THE MOUNT WILSON OBSERVA-
TORY TO THE SOLAR ECLIPSE OF JUNE 8, 1918.

By J. A. ANDERSON, Ph.D.
(Read April 25, 1919.)

The apparatus of the Mount Wilson Observatory expedition at
Green River, Wyoming, consisted of :

1. Littrow Plane Grating Spectrograph. Observers: Adams, St,

John and Brackett.

2. Corona Rotation Prism Spectrograph. Observers : Adams and

Ellerman.

3. Objective Grating Spectrograph. Observers: Anderson and

Babcock.

4. Objective Prism Polarizing Spectrographic Camera: Anderson

and Babcock.

5. Eight inch thirty-foot Photoheliograph Camera : ]\Ir. Ellerman.

6. Small Silvered Quartz Lens Camera : Miss Margaret Hale.

Light was supplied to all of these except (3) and (6) from the
thirty-inch ccelostat mirror of the Snow telescope. (3) was fed
from a small speculum mirror mounted on the coelostat axis, while
(6) was simply pointed directly at the sun, and the image allowed
to drift.

The general arrangement is shown in the plan sketch (Slide i).
The beam of light from the coelostat mirror is directed horizontally
westward ; the lower central portion falls on the eight-inch photo-
heliograph lens, the portion immediately north of this serving to fill
the two-inch lens of the polarizing spectrograph camera. The
upper central portion falls on the six-inch image-forming lens of
the corona rotation spectrograph, while the considerable portion left
of the south half of the beam sufficiently fills the concave mirror
which forms the image used by the Littrow Plane Grating Spectro-
graph. This spectrograph was located east of the coelostat, and

255

256 ANDERSON— ECLIPSE EXPEDITION OF THE

the axis of the concave mirror was incHned to the south and up-
wards sufficiently to allow the beam to clear the coelostat mounting,
the angle being a matter of about 3°.

This arrangement which is unusually compact even for eclipse
work was found to work very satisfactorily indeed, and on account
of its compactness it was a simple matter to protect all the instru-
ments from the heat of the sun, as well as from vibration due to the
wind.

The Littrow spectrograph consisted a 6-inch aperture, 18-foot
focus lens, and a plane grating having a ruled surface of 4 X 6
inches, I5,cx)0 lines per inch, one first prder being extremely bright.
The dispersion was about 3.4 A.U. per millimeter, which was
judged sufficient for accurate wave-length determinations. The
diameter of the moon's image formed by the projection mirror was
about three inches, and this image was placed just tangent to the
slit, the point of tangency being nearly coincident with the position
of second contact. During mid-totality, the whole spectrograph
was moved parallel to itself a distance of three inches so that during
the last half of totality the slit was tangent to the moon's limb at or
near the point of third contact. An auxiliary 90° prism and observ-
ing eyepiece enabled one of the observers to watch directly the
region of the spectrum including D3 for purposes of accurate guid-
ing. Two small right-angled prisms placed over the slit allowed on
iron arc comparison spectrum to be impressed on each plate simul-
taneously with the exposure on the sun. The exposures were six
in number, three made during the total phase and three before and
after, in order to get accurate wave-lengths extremely close to the
sun's limb.

The corona rotation spectrograph was a three-prism Littrow
type, using a lens of forty-inch focus, the dispersion at A 5300 being
about six A.U. per millimeter. Only that portion of the spectrum
in the immediate vicinity of the green coronal line was photo-
graphed, the slit being parallel tO' the sun's equator. An iron arc
comparison spectrum was impressed immediately before totality so
as to permit an accurate determination of the wave-length of the
green line.

The objective grating spectrograph used a six-inch Rowland

MOUNT WILSON OBSERVATORY. 257

concave grating of twenty-one-foot radius of curvature, and was of
exactly the same type as has been used at various ecHpse observa-
tions before. One improvement in the method of observation needs,
hov^ever, to be noted: Ordinarily the spectrograph is adjusted with
the axis of the incident beam pointing to the center of the sun. For
a station near the center line of the eclipse track the direction of the
incident light at second and third contacts, make angles of ±15'
with this direction, which alters the focus of a grating such as was
employed here by a quantity of the order of one millimeter. In the
present case the speculum mirror being adjusted to the center of
the sun before totality was rotated eastward through an angle of
seven and one half minutes for the exposures at second contact ;
then back to the original position for the exposure at mid-totality ;
then westward seven and one half minutes for the remaining expo-
sures at third contact.

The objective prism polarizing spectrograph consisted of an
ordinary 4X5 view camera having a two-inch aperture, eight-inch
focus Planar lens, in front of which was placed in order a direct
vision prism and a Rochon double image prism. Two objective
prism spectra are thus obtained, the planes of polarization being at
right angles to each other.

The silvered quartz lens camera was used simply to obtain data
in regard to exposure time and relative aperture required for use at
future eclipses. As is well known a silver film opaque to visible
radiation is quite transparent from about A 3200 to A 3100. Used
on a quartz lens it is hence possible to obtain fairly monochromati*.,
images in this part of the ultra-violet spectrum. In the present case
the aperture was about F/12, the exposure time through clouds with
an estimated transparency of 10 per cent., about ninety seconds,
and yet a great deal of coronal structure is well shown on the plate
obtained. Hence it should be easy to obtain good photographs in
this way when conditions are right.

On the day of the eclipse a large cloud obscured the sun from a
few minutes after first contact until a few minutes after third con-
tact. During totality the sun was visible through an irregular thin-
ner portion of the cloud, but it is doubtful if the intensity of the

PROG. AMER. PHIL. SOC, VOL. LVIII, Q, JULY 30, I919.

258 ANDERSON— ECLIPSE EXPEDITION

light was more than lo per cent, of normal. However, the program
was carried through as planned.

The photographs obtained with the photoheliograph were fairly
good as may be seen from the slides (2 and 3). The plates taken
with the Littrow plane grating spectrograph were too weak to be
of value, as was also the case with those obtained with the objective
grating spectrograph. The latter shown about a dozen of the ultra-
violet hydrogen lines, and the prominences in the stronger hydrogen
lines and in H and K show well. The irregular prominence shown
on slide 2 shows some distortion due probably to motion in the line
of sight.

The plate taken with the corona rotation spectrograph shows the
coronal line only at the east limb, hence the rotation cannot be deter-
mined. The wave-length at E limb neglecting the rotation comes
out equal to 5303.022 I. A.

Mount Wilson Observatory,
Pasadena, Calif.

THE LOWELL OBSERVATORY ECLIPSE OBSERVA-
TIONS, JUNE 8, 1918.

Prominences and Coronal Arches.

By C. O. LAMPLAND.
(Read April 2S, W^Q-)

The solar prominences or protuberances have considerable his-
torical interest in the part they have taken in the advances made in
our knowledge of the constitution of the sun. These beautiful
formations, varying in color from deep ruby to pale pink, project-
ing outside of the dark disk of the moon upon the background
of the pearly luster of the corona at times of totality caught the
attention of earlier observers and we may have references to
them extending back two hundred years. They came to occupy a
conspicuous place in the eclipse literature of 1842, and Dr. Lockyer
remarks of this eclipse in connection with the prominences, " then
the golden age begins." But whether they belonged to the moon
or the sun was a question that was not definitely decided until De la
Rue's photographic observations of the eclipse of i860. His series
of exposures during totality showed beyond doubt that the promi-
nences on the opposite edges of the sun were progressively covered
and uncovered by the disk of the advancing moon. At the present
day it may help us to sympathize with the difficulties of these earlier
investigators to recall our own helplessness, and how we are appar-
ently marking time, on some of the outstanding problems of the
corona.

Our knowledge of the more complex structure of the inner
corona by direct observations has been largely obtained by photog-
raphy, and noteworthy progress has been achieved during the last
twenty-five years. Large-scale photographs, suitable sensitive
plates, and graduated series of exposures may be mentioned among
some of the things that have contributed to the advances made.

259

260 LAMPLAND— THE LOWELL OBSERVATORY

In the present paper a brief account will be given of some pre-
liminary observations of the prominences, of coronal formations
over them, and of some of the complex structure of the inner
corona, shown in the series of photographs obtained by the Lowell
Observatory Eclipse expedition stationed at Syracuse, Kansas.

The photographs were made with two of the historic "Transit
of Venus " objectives kindly loaned by the U. S. Naval Observa-
tory. These objectives have an aperture of five inches and a focal
length of 38.7 feet. They were mounted according to Schaeberle's
method. The tower camera and its moving plate carrier have been
so frequently described, and the method is so well and favorably
known among astronomers that no time will be taken here for the
discription of the apparatus.

In connection with the description of some of the detail in the
Lowell Observatory photographs quotations and illustrations from
published observations of earlier eclipses will also be given, but such
references must necessarily be very brief and incomplete at this
time. It is thought that these comparisons should be both inter-
esting and instructive.

A brief description will now be given of the more interesting
features of the Lowell Observatory photographs.

P.A. Description of Detail.

25° A coronal arch. No prominence visible.

30°. 5 Brilliant short prominence.

51°. 5 Large brilliant triangular-shaped prominence, base resting
on chromosphere. Fine series of coronal arches, four or
more, over this great prominence. Branches of the higher
arches not symmetrical with respect to prominence but as if
partially superposed from perspective displacement.

99°. 5 Small bright prominence inclined (N. and W.) to solar
surface about 45°. It is centrally in great mass of brilliant
coronal matter. Indications of arches but detail cloud-like
or flocculent in parts.

ECLIPSE OBSERVATIONS. 261

162°. A group of three brilliant prominences near south pole;
to positions of centers of these prominences about 165°,

182° 172°. 5, and 179°. 5 (position of south pole of sun i6y°.4).
Very bright coronal matter over the group. Strong hood
or arch over the first two prominences ; part of an arch over
the third (named in order of increasing P. A.).

186° A large bright " petal " of coronal matter over the brilliant
to flame-like prominence in position angle 205°. A closely

226° packed series of arches, all of the pointed or ogival type,
except possibly the inner one. At least six of these arches
stand out distinctly. The highest ones very sharply
pointed, and the apexes, with increasing height, are to one
side of the center of the prominence, as though the dis-
placement were an effect of perspective.

231° Another large brilliant "petal" of coronal matter, with the
to great eruptive prominence centrally in its base (P. A. of

270° center of prominence about 251°). At least five arches are
visible. In this instance the arches are oval curves, and not
pointed at the vertices. The arches are somewhat skewed,
the inclination being in the same direction as the trunk of
the great skeleton prominence they enclose. A comparison
of the Lick and Lowell plates shows marked changes in
the eruptive prominence.

233° -5 Base of the " rocket " prominence.

The higher part of the trajectory of this eruptive forma-
tion or jet is not certainly shown on the Lick plates taken
about 27™ earlier. On the Lowell plates the upper part
of the curve is clearly shown and the brilliant nucleous has
apparently passed its greatest height and is descending.

270° Very brilliant " petal " of coronal matter, with large bril-

to liant prominence centrally (P. A. of base of prominence

320° extends from 288° to 301°). Complex detail above the

262 LAMPLAND— THE LOWELL OBSERVATORY

prominence — cloudy and flocculent — with indications of
confused and overlapping arches.

351°. 5 Prominence near the north pole.

225° Pronounced rifts in the corona occur near these points,

and the latter inclined, near limb, 20° or more southward with

270° the radius.

35° Four narrow and well-defined dark rays in brilliant coronal

to matter, inclined a few degrees, northward, with the radius

40° where rays start from the limb, but with gentle bending

southward towards the equator with increasing distance

from the limb. A number of broader and more diffuse

dark lanes and streaks occur in different parts of the corona.

Any detailed description of the appearance, distribution and
directions of the coronal rays for different regions will not be given
here. The variability and irregularity of their inclinations with
respect to the limb of the sun suggests the maximum type of corona,
though in some other respects characteristic features of the inter-
mediate type are present, as for example the general shape of the
corona as a whole is roughly triangular.

In the Lowell Observatory photographs all of the prominences
except the one near the north pole are surrounded by arches or
envelopes or of disturbed coronal matter. Above some of the
prominences there are series of hoods or arches the forms of the
inner or lower ones being generally some oval curve, while the outer
or higher ones have forms resembling the pointed or ogival arch.

One striking feature of these photographs is the disturbance of
the coronal streamers at and in the vicinity of the sun's poles. In
that respect the corona of 1918 resembles earlier ones observed near
sun-spot maxima, though as previously remarked it exhibits other
features characteristic of intermediate types. In the present photo-
graphs it would be difficult to estimate the position of the poles of
the sun from the appearance of the streamers.

That there is an intimate relation between the prominences and
the surrounding coronal structure has at different times been ex-

ECLIPSE OBSERVATIONS. 263

pressed by observers of the more recent eclipses. But these rela-
tions are not always evident as will be shown by extracts from
published observations of eminent observers and students of eclipse
phenomena. It is probably true that hoods and envelopes over the
prominences are generally present but are greatly reduced in in-
tensity or may even be absent for a short time near sun-spot minima
and it is also probably true that such detail and complex structure in
the inner corona are generally conspicuous at times of greater solar
activity.

From an extensive examination and careful study of many
photographs of the corona beginning with the eclipse of 1851 and
continuing with the later eclipses of i860, 1870, 1871, 1875, 1878,
1882, 1883 and 1885, an eminent authority on eclipse photographs
writes : " There is no sign of any connection between the coronal
rays and the solar prominences." It would seem, however, that the
present trend of the interpretation of observational results is that
the prominences generally do affect to a marked degree coronal
structure in their vicinity.

The complex coronal structure, arches, etc., of the eclipses of
1871, 1883, 1893, 1905 and 1918 occurring at or near times of sun-
spot maxima, and the almost complete absence of such intricate and
complicated detail of the lower regions of the conona of the eclipses
occurring at or near the minima of sun-spot activity in 1878, 1889
and 1900 may be mentioned in this connection. The eclipses at the
minimum epochs were so fully and so successfully observed that it
is difficult to believe that such structure could have been overlooked.
The reports for 1900 by Hale, Langley, Newall, Wesley and others
are especially complete in this respect as specific mention is made
with reference to search for such detail. It also appears from the
observational results that the complex detail in question is present
in varying degrees for eclipses observed preceding or following
minima of solar activity. The eclipses of 1896 and 1898 are good
examples of eclipses that fall between a maximum and minimum.

The question of changes taking place in detail in the corona
in the interval between observations made at different stations is
one of great interest and we are not aware that a definite answer

264 LAMPLAND— ECLIPSE OBSERVATIONS

has yet been given in the previous attempts to decide the matter. It
was hoped that it might have been possible to include some satis-
factory observations in this communication. The preliminary work
done thus far does not warrant a stronger statement than to say
that further comparisons should be made, at least in the case of
one of the series of prominent arches. Unfortunately the material
at present available is not wholly satisfactory for such difficult
comparisons. As prominence detail is frequently subject to rapid
change it does not seem worth while to add anything further than
already given relative to such comparisons. The marked changes
that occurred in the great eruptive prominence are obvious upon
casual inspection.

Our appreciative thanks are due to Dr. Campbell for generously
granting permission to make use of three very fine positives from
plates of short exposure taken at Goldendale, sent us recently.

Dr. Slipher gives herewith a summary of his observations of the
spectrum :

Spectrum Observations of the Corona.

The spectrograms recorded a strictly continuous spectrum for
the inner corona crossed by the three coronal emissions of wave-
lengths 4086, 4231 and 5303.0; and a faint solar dark-line spectrum
for the outer corona which seems to be of true coronal origin. The
objective prism plates registered the continuous spectrum of certain
prominences in addition to their usual emissions, and the green
coronium ring between the fragmentary ones of hydrogen F and
helium D3. The distribution of the intensity of the green ring
implies that the substance " coronium " is generally abundant along
those parts of the sun's limb occupied by prominences and from
which flow the great extensions of the corona, and that it is sparse
along those sections of the limb occupied by the bristling streamers
typical of the polar regions.

The observational results given in the present paper are the joint

work of Dr. V. M. Slipher, Mr. E. C. Slipher and the writer. A

general account of the organization and work of the expedition has

been published in one of the astronomical journals.

Lowell Observatory,
Flagstaff, Arizona.

THE FLASH SPECTRUM.

By S. a. MITCHELL.
(Read April 25, iQiQ-)

The total eclipse of 1918 saw the completion of fifty years since
the spectroscope was first used at the eclipse of 1868, visible in
far-off India. If all the minutes of totality during this half century
of eclipses were added together, they would amount to less than
one brief hour. It is safe to say, however, that no other branch of
astronomy has shown the remarkable value of the new instrument
of research as has the work of eclipse spectroscopy.

In 1868, Janssen discovered that the prominences gave a bright
line spectrum thus proving that they are gaseous in nature. As is
well known, he and also Lockyer independently, showed how to view
the prominences without an eclipse.

The year 1869 brought with it the discovery of helium in the
sun. Since the urgency of the war has resulted in having helium
supplied in such large quantities and at such a cheap price that it
may be used in balloons and airships, it is almost a surprise to think
that helium was not discovered by Ramsay tmtil the year 1895.

The eclipse of 1870 visible in Spain is noted for the discovery of
the flash spectrum. The dark lines in the solar spectrum, the Fraun-
hofer lines are caused by the absorption of the light of the photo-
sphere by a thin layer of cooler gases, the so-called reversing layer.
This layer is cool only in contrast with the very hot photosphere.
As the moon gradually covers the sun at the time of an eclipse the
dark line spectrum persists so long as there is even the slightest
trace of the photosphere visible. At the instant when the photo-
sphere is entirely covered up, the bright photospheric background of
the Fraunhofer spectrum is removed, and the lines of the spectrum
now appear as bright lines on a black background, where before
there had been dark lines on a bright background. This change
coming with totality was foretold by Young of Princeton before the

265

266 MITCHELL— THE FLASH SPECTRUM.

eclipse of 1870, and his was the first eye to perceive it. The sudden-
ness of the change caused him to name the bright line spectrum the
"flash spectrum." At the beginning of totality, the flash spectrum
lasts for only a few seconds while the moon is advancing in front
of the rather shallow layer. A second flash spectrum is seen at
the end of totality.

The first attempt to photograph the flash was in 1893, but with
very imperfect results. Shackleton in 1896 was more successful,
though the first photographs with good definition were those of the
eclipse in India in 1898 by Evershed.

In the year 1900, the American astronomers had an opportunity
in their own country at the eclipse of May 28. For the first time
in eclipse work, Rowland gratings were used. Gratings are ruled
on both plane and concave surfaces, and it is possible to use them
either with or without a slit. The great advantages of gratings
over prisms are the increased dispersion, but particularly their
normal spectrum. When used in the ordinary Rowland mounting,
the concave grating shows marked astigmatism. Although work on
stars had shown that the astigmatism of the concave grating when
used directly without slit was exceedingly small, most observers in
1900 were apparently afraid to use a concave grating without a slit.
Those who did use slits found their photographs entirely without
lines, the light that went through the slit being entirely too feeble to
leave any impression on the photographic plate. Gratings used
without slits gave more satisfactory results.

Since 1900, the eclipses most easily observed were the Sumatran
eclipse of 1901, the Spanish eclipse of 1905, the Flint Island eclipse
of 1908, the Russian eclipse of 1914, and the American eclipse of
1918.

When great dispersion is desired, the concave grating without
slit is perhaps the most desirable form of spectrograph. Its mount-
ing is very simple. Light from the coelostat mirror falls direct on
the concave grating where it is diffracted and brought to a focus on
the photographic plate. It might not be out of place here to call
attention to the great importance, and also the great difificulty, of
obtaining sharp focus with the slitless instrument. The best results

MITCHELL— THE FLASH SPECTRUM. 267

have been obtained by the writer through the use of a colhmator
designed by Jewell.

7n 1905 with a Rowland concave grating of 15,000 lines per inch,
Mitchell secured a photogrph of the flash spectrum in which he
measured 2,841 lines between A 3300 in the violet, and the D3 line
in the yellow. On account of the good definition, it was possible to
determine wave-lengths to 0.02 Angstroms correspondiog to an error
of measurement of 0.002 mm.

The general conclusions from these measures were :

1. The flash spectrum is a reversal of the Fraunhofer spectrum.

2. The flash is not an instantaneous appearance, but the chromo-
spheric lines appear gradually. At the beginning of totality, those
of greatest elevation appear first, and at the end of totality remain
the last. The "reversing layer" which contains the majority of
the low-level lines is about 600 km. in height.

3. Wave-lengths in the chromospheric and solar spectrum are
practically identical.

4. The lines in the chromospheric spectrum differ greatly in in-
tensities from the lines in the solar spectrum. The chromospheric
spectrum shows the hydrogen series of lines and the helium lines
and corresponds to a spectral type earlier than the solar spectrum.

5. The dififerences in intensities find a ready explanation in the
heights to which the vapors ascend.

6. Especially prominent in the chromosphere are the enhanced
lines.

In view of the very excellent work done recently at Mt. Wilson,
it has become of the utmost importance to have as accurate a knowl-
edge as possible of the heights of the various layers of the sun's
chromosphere above the photosphere. There is no other method of
determining these heights directly except from the photographs of
the flash spectrum at the time of a total eclipse taken without slit.
By knowing the angular diameters of sun and moon, it is easy to
calculate the height of the solar layers of atmosphere by measuring
the lengths of the cusps. On the photograph a spectrum line of
considerable length corresponds to a greater height of atmosphere,
while a shorter line belongs to a low-lying vapor.

268 MITCHELL— THE FLASH SPECTRUM.

The program of work on the flash specrum at the eclipse of 1918
was for the purpose of finding the heights of the various gases, and
also of extending our knowledge of the spectrum as far into the red
as possible. The Naval Observatory party located in Baker,
Oregon, had three concave gratings, the largest being of 2i>^ feet
radius and 15,000 lines in order to procure as large a dispersion as
possible; the second grating was of 10 feet radius and 15,000 lines,
while the third was of small dispersion to photograph into the ex-
treme red by the use of plates stained by dicyanin. Unfortunately,
thin clouds covered the sun during the whole of totality. The
clouds at the end of totality were exceedingly thin, but they were
sufificient to cut out most of the weaker lines from the spectra. At
no place in the country where photography of the flash spectrum
with great dispersion was attempted did clear skies prevail ; and con-
sequently the carefully prepared plans for 191 8 will have to be tried
again at the next available eclipse.

McCoRMiCK Observatory,
University of Virginia.

PHOTO-ELECTRIC PHOTOMETRY OF THE 1918
ECLIPSE.

By JAKOB KUNZ and JOEL STEBBINS.
(^Read April 25, 1919.)

The expedition from the University of Ilhnois had for the sole
item on its program the measurement of the total light of the corona
by means of photo-electric cells. The arrangement consists of a
light-sensitive cell connected in series with a galvanometer and a
battery giving about 150 volts. When exposed to a light about
equal to that of the crescent moon, a measurable current is pro-
duced, and it was anticipated that the corona would be bright enough
to be measured with accuracy. The advantage of this electrical
photometer is that no matter what the distribution of light is in the
source, whether is be a point or an irregular surface, the effect is
integrated and is combined into a single galvanometer deflection.

The color-sensitivity of the potassium cell is nearly like that of
the ordinary unstained photographic plate, the maximum effect
being at wave-length 4500 A. In a general way, the photo-electric
measures are between photographic and visual, but nearer the
former.

Slide I shows one of these cells, the diameter of the bulb being
about one inch.

Slide 2. — The station at Rock Springs, Wyoming, was selected
so as to be near the Yerkes camp, where Mr. Parkhurst was to un-
dertake visual photometric measures of the corona, for it was felt
that there would be advantages in having observations by different
methods, but with practically the same atmospheric conditions.
However, we did not want a local cloud to spoil all of the pho-
tometry, so we located about fifteen miles east of the Green River
people, and about two miles south of the town of Rock Springs.

Slide 5. — General view of the station.

Slide 4. — It was proposed to take measures in duplicate with two
cells, each mounted in a box at the lower end of a simple tube. The

269

270 KUNZ-STEBBINS— PHOTO-ELECTRIC PHOTOMETRY

twin tubes were fastened on an equatorial mounting, but a driving
clock was unnecessary. (Note: the simple tubes were made of
four-inch down-spouts.)

Slide 5. — The electrical connection from each cell was carried in
conduit to the galvanometer and battery inside the hut. Each cell
box could be detached from the mounting outside, and brought in
to the photometer bench. Here it was arranged that the cell could
be exposed to a standard amyl acetate candle at varying distances,
and also to standard low-voltage electric lamps. The same appa-
ratus was also used to compare the lamps with the full moon, and
thus to get an indirect measure of the corona in terms of moon light.

Slide 6. — Two assistants operated the apparatus outside, while
the observers read the galvanometers.

Slide 7. — Our expedition had the same unfortunate experience
with the weather that other parties had on the days preceding the
eclipse. The nights and early mornings were usually clear, but
clouds, increasing during the day, were usually worst about the hour
of the eclipse, 5 P.M. The picture shows the conditions thirty min-
utes before totality.

Slide 8. — At six minutes before the critical time the sun was still
behind the cloud at the upper left-hand corner of the picture, but at
tw^o minutes before time the cloud had moved away, and during
totality the corona stood out in a perfectly clear sky.

Results.

We secured four complete measures of the light of the corona,
and of the sky background. When proper allowance is made for
the absorption of the earth's atmosphere, it is found that the total
light of the corona was 1.07 candle meters, just half the light of the
full moon. There has been a curious disagreement between observers
at previous eclipses, as values have been found ranging from one
fortieth to ten times the moon's light, a differejice of 400-fold.
These were results from photographs, but the more reliable determi-
nations seemed to indicate a value of one tenth full moon, five times
smaller than our value. The visual results are in much better agree-
ment, ranging from half up to one full moon, or from the same up
to twice our value.

OF THE 1918 ECLIPSE. 271

Our comparison of the corona with the sky before and during
the ecHpse shows that, in terms of a circle of sky of the apparent
size of the sun^, the corona gave one tenth as much hght in full sun-
shine, but more than 600 times as much during totality. The de-
crease in the light of the sky due to the moon's shadow was there-
fore 6,000-fold. As the decrease from sunlight to corona light is
fully 100 times 6,000, this means that not more than i per cent, of
the general sky illumination during an eclipse can come from the
corona, the remainder being from sunlight reflected from the earth's
surface and atmosphere which is outside the moon's shadow.

The result that the corona gives one tenth of the light from a
circle of daylight sky of the same area as the sun, and 8° away, has
a direct bearing upon the problem of detecting the corona without
an eclipse. Experiments in using photo-electric cells for this pur-
pose have already been begun by Dr. Hale at the Mt. Wilson
Observatory.

University of Illinois,
Observatory.

THE SPROUL OBSERVATORY ECLIPSE EXPEDITION,

JUNE 8, 1918.
Coronal Arches and Streamers.

By JOHN A. MILLER.

(Read April 25, 1919.')

The Eclipse expedition from Sproul Observatory was located at
Brandon, Colo., a station on the Missouri Pacific, approximately
fifty miles from the Colorado-Kansas line.

This region is very dry, and the weather conditions are favor-
able. During the month preceding the eclipse the sun was either
clear or only thinly clouded, at 5 : 26 P.M. on all but six days ; and
in general the afternoons were better than the forenoons. The
wind was usually high, often very high, and the air many times was
full of dust. It was so cloudy all day on June 5, 6 and 7, and the
forenoon of June 8, that at no time could successful observations
have been made. This was the longest period of cloudy weather
that we had while there. At noon of June 8, however, it cleared,
and at the time of totality only two small clouds were visible west
of the meridian, one about twenty degree above the sun and the
other near the horizon. The seeing was fairly steady.

The following instruments were mounted :

I. One lens, nine inch aperture, focal length 62^/2 feet. This
was mounted as the Lick Observers since 1893 have mounted their
long focus lenses. Exposures of two seconds, twenty seconds,
forty-five seconds, and three seconds were made with this lens.

II. On a polar axis, two lenses of four inches and three and one
half inches aperture, respectively, and of focal length eleven feet
two inches were mounted. The exposure with each of these lenses
was for eighty-two seconds.

III. In another polar axis were mounted six lenses varying in
focal length from ten and one half to fifty inches ; also a transmis-
sion grating.

IV. A three-prism slit spectrograph with which we hoped to

272

MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION. 273

determine the rotation of the corona, and a sHtless spectrograph
were mounted on another polar axis.

The photographs secured show a moss of coronal detail. The
shorter exposures show the coronal arches that surround the great
number of prominences that are found on the rim of the sun.
These have been so well shown in the slides accompanying the pre-
ceding papers that it seems unnecessary to show them again and
for economy of time I shall show one only, a slide from the forty-
five second exposure with the sixty-two and one half feet focal
length.

The problem that I had more specifically in mind was to secure
if possible some data that would contribute to a knowledge of the
origin of the corona, to see if we could from a study of the details
of its structure get a hint of the nature of the forces that produce it
and give it its shape. I shall limit myself to a brief statement of
our study of two things.

1. To see if there was any indication of change of form in the
corona itself during the eclipse.

2. To find if the general form of the streamers in the corona
gave any indication of the way in which it is formed.

I. Change in the Corona.

Director Campbell, of the Lick Observatory, and Professor J.
C. Hammond, Director of the Naval Observatory Station, most gen-
erously put at my disposal glass positives of short exposure photo-
graphs, made respectively at Goldendale, Washington, and Baker
City, Oregon. The Lick plate was made with a camera whose focal
length is forty feet. Totality at Goldendale lasted for one minute
and fifty-seven seconds. The plate loaned me by Director Camp-
bell was exposed from one minute fifty seconds to one minute fifty-
two seconds after the beginning of totality. This plate was com-
pared with two plates made by the Sproul Observatory expedition,
one exposed for two seconds just at the beginning of totality and
the other for three seconds just at the end of totality. The focal
length of the camera with which these plates are made is sixty-two
and one half feet, so that the pictures to be compared were of very

PROC. AMER. PHIL. SOC, VOL. LVIH, R, AUG. I, IQlQ.

274 MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION.

different scales. The plates were reduced to the same scale by
measuring the diameter of the moon's shadow on each plate and
finding the ratio of these two measured diameters. The center of
the sun for each of the three plates was found and plotted on each
plate, the necessary corrections being made for different times of
exposure, parallax, etc. The conspicuous prominences were as fol-
lows : Two small quiescent northern prominences, a large north-
western quiescent prominence, a large eruptive prominence in the
southwest called from its shape the Skeleton prominence, and a
large northeastern prominence called the Pyramid prominence.

Around the eruptive prominences there are series of arches de-
scribed so well in the preceding paper by Professor Lampland that
further comments are unnecessary.

Following my suggestion Aliss Margaret E. Powell, a graduate
student at Swarthmore College, made a study of these plates. All
the measures which I shall give presently were made by Aliss
Powell. We selected three arches which seemed definite enough to
measure.

The measures were made in this way. A point is selected which
can be certainly identified on each of the three plates. Choosing a
line through this point and the center of the sun as an initial line
and the center of the sun as origin, we may locate any point in the
streamer by its polar coordinates 6 and p. The arches were about
equally well defined in the Lick plates and each of the two Sproul
plates. The polar coordinates of a series of points on an arch were
measured on the Lick plate, then setting off the same series of vec-
torial angles on a Sproul plate and measuring the radii vectors, we
obtained the polar coordinates of the corresponding points on the
Sproul plate. If these radii vectors are the same the arch is un-
changed, but if they are different (assuming the measures are ex-
act) the shape or the position of the arch has changed.

Miss Powell measured three arches in this way. The accom-
panying tables give the details of the measures. The vectorial an-
gle is given by 6. The cjuantities p and p are the radii vectors in
inches measured on the Sproul and Lick plates respectively and
reduced to the scale of the Sproul plate.

The eclipse at Brandon occurred twenty-five minutes after it

MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION. 275

occurred at Goldendale ; p-p is therefore the amount the arch has
changed. If p-p' is positive for a given point, that point is farther
from the center of the sun, when the sun was ecHpsed at Brandon
than when it was ecHpsed at Goldendale twenty-five minutes before.

After the first set of measures was made, the zero line was rede-
termined and all angles and distances remeasured, so that each set of
measures was independent of the others except the same size of each
vectorial angle was retained. The different series of measures were
made on different days, the last series being made about six weeks
after the first were made, so that I believe little, if any, prejudice
originating in the earlier measures was carried into the last set.

Measures of the Arch Around the Southeastern Prominence. —
This prominence is a few degrees east of the South Pole. There
are five well defined arches on the equator side of this prominence,
the inner one of which is measured. The table shows the results.

I. An Arch Around the Southeastern Prominence.

First Measures.

Second Measures.

Third Measures.

P

P'

p-p'

9

P P'

p-p'

P

P'

p-p'

e

I.

3-81

3.71

-l-o.io

+ i°48'

3.84 3-76

-1-0.08

3.84

3.76

-ho. 08

-hi°48'

2.

3-88

3-71

■ 17

I 18

3.91 : 3.76

.15

3.91

3.80

.11

I 18

3-

3-91

3-76

.15

0 48

3-94 3.84

.10

3.94

3.84

.10

0 48

4-

3-94

3-84

.10

0 24

3.97 1 3.88

.09

3.97

3.88

.09

0 24

5-

3-97

3.88

.09

— 0 18

3-97 3-93

.04

4.00

3.93

•07

-0 18

6.

3-97

3.S8

.09

-0 36

4.00 1 3.93

.07

4.00

3-93

.07

-0 36

The quantities p and p in the " first measures " of this arch are
from center of the sun to the inner boundary of the arch ; the other
measures from this arch and all the others are to the middle of the
stream forming the arch. The first measure in each series is always
for a point near the limb of the sun, and is in consequence very
indefinite.

2. An Arch Around the Pyramid Prominence.

Toward the Pole side of this prominence there are four well-
defined arches ; one of these was measured. The arches on the
Brandon plates seem to curve more than those on the Goldendale
plates. The results are shown of four sets of measures on three
different days.

276 MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION.

First Measures.

Second Measures.

Third Measures.

P

p'

P-P

d

P

p'

p-p'

P

P'

p-p' 1

d

I.

4.06

4.14

-0.08

+1° 36'

4.09

4.01

+0.07

4.09

4.05 1 +0.04

+1° 36'

2.

4.28

4.27

+ .01

2 12

4.28

4-23

•05

4.28

4.18

.10 :

2 12

3-

4-38

4-31

.07

2 36

4-38

4-23

•15

4.38

4-23

• 15

2 36

4-

4.41

4-35

.06

3 30

4-47

4.27

.20

4-47

4-31

.16

3- 30

5-

4-47

4.40

.07

4 18

4-47

4-31

.16

4-47

4-35

.12

4 18

6.

4-47

4.44

•03

4 54

4-50

4-35

•15

4-50

4-35

• 15

4 54

7-

4-50

4.48

.02

6 06

4-50

4.40

.10

4-50

4.40

.10

6 06

Fourth Measures.

Fifth Measures.

P

P'

p-p'

P

'''

p-p'

e

I

4.00

4.10

+0.10

4.06

4-05

+0.01

+ i°36'

2

4-25

4.18

.07

4.22

4.18

.04

2 12

3

4-31

4-23

.08

4.28

4-23

• 05

2 36

4

4-34

4.27

.07

4-34

4.27

.07

3 30

S

4.38

4-35

■03

4.38

4-35

■ 03

4 18

6

4.41

4.40

.01

4-47

4.40

.07

4 54

7

4.41

4.40

.01

4-47

.4.40

.07

6 06

3. An Arch of the Skeleton Prominence.

This prominence has five distinct arches on the equatorial side.
On the Pole side they are much less distinct, due perhaps to the
presence of another prominence beyond the edge of the sun. The
measures of the fourth arch from the prominence on the equatorial
side are shown in Table 3.

First Measures.

Second Measures.

Third Measures.

P

p'

P

-p'

8

P

P'

P-p'

P

p'

p-p'

d

I

4.09

4.01

+0

.08

+ 9°

30'

4.09

4.01

+0.08

4.00

4.01

— O.OI

9°

30'

2

4-15

4.10

+

•05

10

24

4.16

4.10

.01

4.09

4.10

+ .01

10

24

3

4.19

4.18

+

.00

II

30

4.19

4.18

.01

4.19

4.14

•05

II

30

4

4.22

4-23

—

.01

13

06

4-25

4-23

.02

4-25

4.18

.07

13

06

5

4.22

4.27

—

• OS

13

42

4.28

4.27

.01

4.28

4.27

.01

13

42

6

4.28

431

—

•03

14

42

4-31

4-31

.00

4-31

4.27

.04

14

42

7

4-34

4-35

—

.01

16

00

4-34

4-31

■03

4-37

4-31

.07

16

00

8

4-38

4-35

+

.02

17

06

4-38

4-35

.02

4.38

4-35

•03

17

06

9

4.41

4.40

+

.01

17

48

4.41

4.40

.01

4.40

4.40

.01

17

48

10

4.44

4.44

—

.00

18

48

4.44

4.40

.04

4.44

4.40

.04

18

48

Miss Powell found that on the average the recession of the arch
around the Southeastern Prominence is (assuming these measures
are correct) ninety miles a second ; that the recession of the arch
around the Pyramid is sixty miles per second and that the Skeleton
is receding fifteen miles per second.

MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION. 277

The Skeleton Prominence and its arches cover 37.5° along the
margin of the sun, the highest arch is 108,000 miles ; this means that
if these arches are a mechanical effect that a volume of gas of more
than 2,540.470,000 millions of cubic miles has been affected.

There are other things suggesting changes of form, but hardly
such things as one could measure. It is not unusual for streamers
to issue from small projections on the prominences. From each of
three tips of the Southeastern Prominence there issues streams each
of which assume the approximate form of the arches above them.

In interpreting these measures, one must not be unmindful of
the difficulties of these measures and the consequent uncertainty.
The streamers themselves are somewhat indefinite objects to meas-
ure. An error in plotting the center of the sun on the plates might
cause the differences of p and p'. On the other hand, the measures
have been very conscientiously made. No one could have been more
painstaking than Miss Powell has been. The center of the sun was
plotted separately on the plate for the first and for the second meas-
ures. Any two series are reasonably consistent, and in one thing
they are very consistent — p-p', though small, is almost always posi-
tive. Moreover, the three prominences are so distributed about the
sun's limb that an error in plotting the center should have affected
p-p' in different ways for different arches. In my judgment these
measures make it probable that these arches of the corona changed
in the twenty-five minutes between the eclipse at Goldendale and
that at Brandon, and that they are going outward from the sun.

The other study is of quite a different nature. There are in
every corona a number of streamers, those around the poles of the
sun being most conspicuous, but streamers are by no means confined
to these regions. In 191 1 I published a paper^ in which I assumed
that what we see or photograph as the^streamers of the corona are
projections on a plane perpendicular to the line of sight, of streams
of particles the motion of which is produced by ejection, by the rota-
tion of the sun, by the attraction of the sun and by radiant pressure
from the sun. At that time Director Campbell generously placed
at my disposal the most excellent series of long focus photographs
that had been made by eclipse expeditions from Lick Observatory.

1 Astrophysical Journal, XXXIII., 4, page 303.

278 MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION.

I was able to show that if these streamers were thus mechanically
produced that there would result in each corona a few streamers of
such a shape that one could draw a radius vector from the center of
the sun which would be tangent to the streamer and that one could
find for streamers of this shape the following things :^ the point on
the sun from which the streamer issued, the velocity with which the
particles in it were ejected, the velocity of the particles at any time,
the paths described by the particles and if we assume the law of the
repulsive force, we can find its magnitude. Theoretically, there
should be very few streamers of this shape. In an examination of
the Lick plates made at six eclipses between 1893 and 1908, I found
sixteen streamers of this type ; and on the plate made by Father
Cortie at a time of minimum sun-spots in 1914, Aliss Caroline Smed-
ley, an assistant at the Sproul Observatory, found and measured two
streamers of this type, and in the corona of 1918 there is unmis-
takably one and probably two streamers of this form. These also
Miss Smedley measured and reduced. There always exists the pos-
sibility that the form of the streamer has been affected by local
causes and the arches which I have just discussed makes it ap-
parent that at least near the surface of the sun at this eclipse local
forces were very effective, but the streamers that we found in this
corona and which we measured stand out separated from these dis-
turbed regions and away from the prominences.

This theory works admirably with one exception which I shall
now discuss. Since the solution gives the form of the paths in
which the particles in the streamer are traveling, one should be able
to compute an ephemeris for each particle and using the constants
determined in the solution compute a streamer that exactly repro-
duces the streamer from which the constants were found. It turns
out that with these constants the streamer on the sunward side of
the tangential point P and for some distance on the other side of
the P the streamer can be perfectly represented, but that when we
compute the position of the particles at a distance of two or tiiree
radii from the sun, that the streamer thus plotted turns back nnich
more abruptly than those do that are shown in the ]:)hotograph. \\'e
have ])lotted with the constants obtained by the solution several

- See Fig. 4, loc. cit.

MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION. 279

streamers and they all possess this property. This may be inter-
preted to mean that this mechanical theory of the corona is wrong
or that there are other forces acting on these particles than the at-
tractive forces of the sun, and of a repulsive pressure that varies
inversely as the square of the distance from the sun's center. It is
easy to conceive of many forces such as resistance to motion by
gases or that the particles at greatest extension of the streamer were
lighter than those forming the part of the streamer measured.

We attempted also to apply Bigelow's earlier theory to the for-
mation of these streamers. Miss Smedley was able to show that if
this (Bigelow) theory is true that these magnetic forces alone could
produce no streamer of the form described in Figure 4 ; that is, that
the vectorial angle of the radius vector from the center of the sun
to any point P on the streamer either constantly decreases or con-
stantly increases as P is moved along the streamer away from the
sun.

Another phenomena which seems very exceptional in the case
of this eclipse occurs in connection with the polar rays. In almost
no case are the polar rays arranged exactly symmetrical with regard
to the axis of rotation of the sun. But in most coronas the vectorial
angle of a point p decreases as it moves out along a streamer on the
west side of the axis of rotation ; while for those on the east side the
vectorial angles increase ; that is, these rays are gently curved away
from the axis, but. as I have said, not exactly in a symmetrical way.
All these things can be produced under the assumption of the me-
chanical theory ; but in this eclipse the vectorial angle decreases as P
moves out along the streamer whether it is on the east or on the
west side of the axis of rotation, and this is true not only of the
rays around the north, but also around the south poles. I have not
had time to investigate whether or not this is possible under the
mechanical theory.

It seems probable to me that the long graceful streamers of the
corona, at least the part of them that is at some distance from the
sun are largely the product of mechanical influences. The sugges-
tions made by JNIr. Lampland in the preceding paper that of certain
prominences the inner arches seem to circle around the prominence
while the outer arches were pointed, is in my opinion very signifi-

280 MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION.

cant and seems to me to exhibit the general tendency of coronal
streamers to straighten themselves out at some distance from the sun.

We examined the long streamers on the forty-five second expo-
sure with the 62^-foot focal-length camera to see if one would
guess from the shapes of the streamers that there was a violently
disturbed region immediately below them. Our conclusions were
in the negative, that is, that no one would guess from the shape of
the outlying streamer that there was any violent agitation at the
solar surface. There are many phenomena that lead one to believe
that the corona is a magnetic or electric product and it is possible
that it results from a combination of these things. At any rate
there seem to be abundant reasons to believe that the problem is not
beyond solution.

Dr. L. A. Bauer had consented to give a summary of the mag-
netic work done during the eclipse of 1918 at stations under his
direction. He was prevented from doing this because of his de-
parture to establish stations in Africa and Brazil to make magnetic
and electric observations during the eclipse May 29. Before leav-
ing he sent a summary of the chief conclusions that he reached from
the observations of the eclipse of June 8, 1918. I shall read this
summary.

The following conclusions are drawn covering the chief results of the
magnetic observations made in connection with the solar eclipse of June
8, 1918:

(a) Appreciable magnetic effects were observed during the solar eclipse
of June 8, 1918, at stations distributed over the entire zone of visibility and
immediately outside. (How much further some of the effects may have
extended must be left for future study.) The chief characteristics of the
effects took place generally in accordance with the local eclipse circumstances
and in general accord with effects observed during previous eclipses. The
evidences of a direct relation between the magnetic effects and the solar eclipse
are so numerous as to warrant drawing the definite conclusion that an appre-
ciable variation in the Earth's magnetic field occurs during a solar eclipse.
This particular variation is termed here the " solar-eclipse magnetic vari-
ation."

(b) The range of the solar-eclipse magnetic variation, according to the
particular magnetic element, is about o.i to 0.2 that caused by the solar-
diurnal variation on undisturbed days. The effects are of a more or less
complicated character, according to location of observation-station in the
zone of visibility. The effects caused during the local eclipse-interval are
superposed upon those caused by the continued disturbance of the Earth's

MILLER— SPROUL OBSERVATORY ECLIPSE EXPEDITION. 281

magnetic field in the region over which the shadow-cone has already passed.
It is thus possible to discern effects having a period approaching that of the
local eclipse-interval and others having a period approximately that of the
encire or terrestrial eclipse-interval.

(c) The general character of the system causing the solar-eclipse mag-
netic variation is the reverse of that causing the daylight portion of the
solar-diurnal magnetic variation. The range of the eclipse variation is
comparable with that of the lunar-diurnal variation, and, like the latter, the
variation usually consists of a double oscillation during its period of de-
velopment.

(d) The range of the apparent effect on the intensity of magnetization
of the Earth during the solar-eclipse magnetic variation, is about equal to
that found associated with a lo per cent, change in the solar radiation as
shown by changes in the solar-constant values.

(e) The results at the high mountain-station. Corona, Colorado, indicate
that the magnetic effects during a solar eclipse may be modified and even
intensified by altitude of station, topography and meteorological conditions.
In view of the bearing of these results upon the theory of the solar eclipse
magnetic variation and possibly upon the theory of other variations of the
Earth's magnetic field as well, it will be highly desirable in the planning of
future eclipse work to include as many mountain-summit stations as con-
veniently possible. — Terrestrial Magnetism, March, 1919, Vol. XXIV., No. i.

RESULTS OF OBSERVATIONS OF THE ECLIPSE BY

THE EXPEDITION FROM THE YERKES

OBSERVATORY.

By EDWIN B. FROST.
(Read April ^5, J919.)

The expedition from the Yerkes Observatory for the observa-
tion of the solar echpse of June 8, 1918, occupied three stations:
the principal one at Green River, Wyoming, the second at the
Chamberlin Observatory at the University of Denver, and (3) a
site near Matheson, Colorado, which had been selected by Professor
Barnard and the writer on a reconnaissance trip in 191 7.

At the last-named station, Professor Edison Pettit, of Washburn
College, Topeka, Kansas, who was at the time an assistant at the
Yerkes Observatory, aided by ]\Iiss Hannah B. Steele, then fellow
in astronomy at the University of Chicago, and by others, obtained
some good photographs of the corona with the use of the twelve-
inch objective of Washburn College, stopped down to an aperture
of four inches, combined with a moving plate according to Schae-
berle's method. The first slide shows one of Mr. Pettit's best pic-
tures, with an exposure of one second. The weather was unfavor-
able in the weeks preceding the eclipse for a considerable extent of
the track of totality, but it fortunately cleared at IMatheson at the
time of the eclipse. Other parties which made successful observa-
tions at this station were that of Professor Loud and Mr. Hartley,
that from Drake University under Professor Morehouse, and that
from the University of Toronto under Professor Chant.

At Denver we received the great courtesy from Director Howe
of the use of the twenty-inch Clark equatorial. A special auto-
collimating spectroscope with a Michelson plane grating belonging
to the Yerkes Observatory was adapted in our shops to fit the Den-
ver equatorial. Professor Schlesinger kindly loaned us an attach-
ment from the Porter spectrograph of the Allegheny Observatory

282

FROST— OBSERVATIONS OF THE ECLIPSE. 283

whereby we could bring into juxtaposition images from the east and
west Hmbs of the sun. The scale was about 22 A per mm., and the
definition good. The intention was to photograph the green coro-
nal line out to about two minutes of arc from the eastern and west-
ern limbs of the sun in order to determine, if possible, the rotation
of the corona. With clear weather, we should have had as good a
chance of making this test as has been had thus far by other expe-
ditions. The writer spent ten days at Denver getting everything
ready, at the end of May, and the observations were to be made by
Professor R. S. Nyswander. Unfortunately the day was com-
pletely cloudy in the vicinity of Denver.

On the observatory grounds at Denver, we had also installed a
small coelostat for the direct photography of the corona, utilizing a
five-inch objective of twenty-two feet focus, belonging to the Den-
ver equatorial. The installation of this apparatus was attended to
by Professor Paul Biefeld, of Denison University, who kindly vol-
unteered his services.

The main station at Green River had also been selected in the
previous year, when it was visited by Professor Barnard and the
writer on a day of extraordinary atmospheric clearness. The
weather records for many years seemed to promise extremely well
for this station, although there was much bad weather during the
six weeks after our camp was established. The elevation of the
station was about 6,200 feet. The day of the eclipse was fine until
after noon, when white cumulus clouds began to drift across the
sky. A large triangular cloud covered the sun at the time of first
contact and moved away with aggravating slowness, so that there
was a fair question for twenty minutes previous to totality whether
or not the sun would be covered by the cloud. Unfortunately the
cloud did not drift away until some three minutes after totality.

Although we had a fine view of the corona and of the brilliant
prominences through the edge of the cloud, the spectroscopic obser-
vations were very greatly impaired. To me, the visual phenomena
of the eclipse were much more impressive than they were in a per-
fectly clear sky at Wadesboro in 1900.

The direct photographs of the corona and prominences were
made under Professor Barnard's direction: (i) with the coelostat

284 FROST— OBSERVATIONS OF THE ECLIPSE.

and a six-inch objective of sixty-two-feet focus, with the exposures
by Miss Mary R. Calvert, who had her station in the dark room ;
(2) with the twelve-inch Kenw^ood equatorial of this observatory,
operated by Professor Barnard, with several smaller cameras at-
tached to the tube. Professor Barnard makes the following com-
ment on the apparent connection of some of the coronal streamers
with prominences :

" Apparent Connection of Some of the Coronal Streamers
WITH Prominences.

" Perhaps no photographs of a previous eclipse have shown with
such beauty and distinctness the succession of expanding arches
that must have extended above some of the prominences like great
spreading envelopes. This is strikingly shown by a large coronal
form, made up of numerous arches that seem to center about the
remarkable " skeleton " prominence, the position angle of whose base
is 253°. A similar but smaller form and arches are centered about
a small prominence in position angle 206°. There are other cases,
but the arches are not so well developed.

" The intimate connection of some of the coronal streamers with
some of the prominences is best shown by a small prominence in
position angle 234°, from which a coronal streamer apparently
emanates and in which it seems to have its actual origin. Close to
this, to the west, is a long, low-lying prominence, in position angle
240°, in which similar coronal streams seem to have their origin.
Originating apparently in a projection in the southerly part of this
prominence, a broad stream bends westward over the entire promi-
nence. From the great prominence in position angle 298° there
are strips of matter apparently streaming southward toward the
equator, as if impelled by some directing force.

" On account of the large scale on which they are taken, these
features are shown to good advantage on the photographs with the
6i>2-foot coelostat."

It had been my hope that we could apply to the problem of the
rotation of the corona the method of the interferometer so success-
fully used by Messrs. Fabry, Buisson and Bourget in photographing
interference fringes in the Orion nebula (Astrophysical Journal,

FROST— OBSERVATIONS OF THE ECLIPSE. 285

Vol. 40, pp. 241-258, 1914). While this method would be quite dif-
ficult of application and there might be a reasonable doubt as to
whether sufficient exposure could be given to secure a proper im-
pression of the rings from the green coronal line, nevertheless a suc-
cessful photograph would be of the greatest value, both for studying
internal motions of the corona and its rotation as a whole.

I had the great advantage of discussing this with M. Fabry
during his visit to the United States as chairman of the French Sci-
entific Mission. He very kindly offered the loan of the apparatus
used at Marseilles, and, if it had arrived in time, we should have
tried the method, employing a rather small image of the sun in the
hope of getting sufficient intensity. Unfortunately the apparatus
did not reach us until long after the eclipse.

I hope that this method will be properly tested at some future
eclipse.

The flash spectrum as hitherto photographed has represented a
composite of the successive images of the dift'erent reversing lines
during the critical second or two at second and third contacts. Ex-
ceptions to this have been the instances where Professor Campbell
has employed a falling plate. I am not sufficiently familiar with
the results thus obtained, of which I have not seen reproductions,
to know how definitely the different stages of the brief phenomenon
are recorded. It seemed to me that the movie camera was at pres-
ent in a sufficient state of development to be successfully applied to
this problem. A " Universal " type of camera was employed, with
its short-focus lens removed, and this was attached, without altera-
tion or injury to the machine, to an objective-prism spectroscope
having three large Mantois prisms and a special camera lens of 5
cm. aperture and 40 cm. focus. This gave a scale of 13 A per mm.
in the vicinity of Hy. Only a small portion of the spectrum can
be photographed with the commercial machine because of the size
of the film, which allows an image i inch X ^ inch (25 mm. X 18
mm.). A region would naturally be selected in which important
lines occurred, or such as it was desired to study particularly. The
correct exposure was naturally a question of some uncertainty in
advance of the event. This is, of course, determined by the rate at
which the crank of the machine is turned. It is arranged for eight

286 FROST— OBSERVATIONS OF THE ECLIPSE.

pictures per turn, and after some preliminary experiments, I decided
that it was safe to have the crank operated at the rate of two turns
per second, giving sixteen pictures per second. Inasmuch as about
half of the time was used in moving the film between exposures, this
would represent about %o of a second for the exposure. On ac-
count of the clouds still lingering over the sun, it was not possible
to obtain satisfactory pictures at the beginning and end of totality
for bringing out the delicate details of the flash. But it is perfectly
evident from the pictures that an excellent record of the successive
stages of the development of the flash would have been obtained if
the sky had been clear. This slide shows a few of the pictures taken
several seconds after totahty : it will be possible to notice the tips
of bright gamma still reversed. The next slide shows a few of the
exposures, successively of the 15th, i8th and 23d seconds after
totality. Over 2,000 impressions of the spectrum were obtained.

Another and important advantage of this method is that it re-
moves all the uncertainty of making the exposure at the correct
instant to secure the flash. This has been a matter of difficulty
and consequent nervousness on the part of the observers, even if
the signal should be given by a person observing the flash itself
visually. By beginning to operate the machine half a minute before
the expected time of totality, and running it for a few seconds after
totality has begun, there could be no doubt about catching the phe-
nomenon at all the stages and hence precisely at the best instant.
Thus the history of the reversal of each line should be shown, and
it would be very dilTerent for those of high level and those of low
level. This could be demonstrated to an audience or to a class by
the use of the film itself, and the film of course could be measured
as well as an ordinary photographic plate.

It is very easy to connect a chronometer with the machine so as to
impress a dot on the film every second or half second, so that the
precise instant of each exposure can be known. At Green River,
Mr. Blakslee of our staff operated the crank and he was to receive
the signal from me as I watched the spectrum with a spectroscope
for this purpose. However, I decided during the partial phase that
it would be safer to begin the exposure one minute before the pre-
dicted time of totality. As a matter of fact, owing to the cloud, I

FROST— OBSERVATIONS OF THE ECLIPSE. 287

was unable to see the flash spectrum at the beginning of totality and
could see the reversal of but few lines at the end of totality.

It would also be perfectly feasible to operate the film by a direct
connection with a chronograph, controlled by a conical pendulum ;
or, if desired, a graduated change in rate of rotation could be given
so that a longer exposure would result before totality (say, an
exposure of ^Ao second) and a still shorter (say, %o second), thirty
seconds before and after the contacts. I sincerely hope that this
simple method will be included in the plans for the next eclipse
expedition, because with clear weather it practically guarantees good
results.

I may say here that it is unfortunate that better records of the
eclipse were not obtained with movie machines. So far as I have
learned no first-class picture was obtained. I took pains to notify
some of the film companies, explaining how important it was that
they should not depend upon their usual short-focus lenses (two to
three inches), but that they should use lenses of at least twenty or
thirty inches focus. Some commercial operators at Denver were
prepared to do this, and dovibtless would have obtained results both
instructive and interesting if the weather had permitted.

Our program also included the photography of the infra-red
region of the spectrum, with the use of films stained with dicyanin,
in connection with a small concave grating of sixty inches focus,
used directly. This was the instrument I had employed at Wades-
boro in 19CO. It was operated by Professor S. B. Barrett, but no
results in the infra-red could be obtained through the cloud. It
might be a question whether there would be ordinarily time enough
for sufficient exposure for the infra-red, and probably future plans
in this direction should insure an abundant light-power in the spec-
troscope.

In connection with the apparatus arranged by Professor Park-
hurst for the photometric study of the corona was a reflecting tele-
scope of six inches aperture and sixty inches focus, covered by a
15° prism of ultra-violet glass of the same aperture. One of the
exposures made with this instrument is shown herewith, as it brings
out an interesting point with respect to the large prominence in posi-
tion angle 253°, already shown from Professor Barnard's negative

288 FROST— OBSERVATIONS OF THE ECLIPSE.

and designated by some as the " heliosaurus." As may be seen
from the slide, the prominence is visible only in the H and K lines
of calcium. If present at all in the other emissions it is too weak
to impress the plate. On the contrary, the quiet prominence at
position angle 297° (= north latitude 40°) appears also in the H/?,
Hy, and Hh lines, and the prominence at position angle 51° (^ north
latitude 26°) is bright in helium D^ as well as in the hydrogen series.

The low prominence or uncovered photosphere at position angle
240° (^ south latitude 17°) shows a remarkable extension into the
ultra-violet, the bright group between wave-lengths 3350 and 3390
A being especially strong. This prominence is an origin of coronal
streamers, as Professor Barnard has shown in his paper on page 222,.

The continuous spectrum of the corona is so strong that it masks
the narrow rings due to the bright coronal lines at wave-lengths
5303, 4231 and others.

Yerkes Observatory,
April 18, 1919.

MAGELLANIC PREMIUM

Founded in 1786 by John Hyacinth de Magellan, of London

I9I9 .

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful K-nowledge

ANNOUNCES THAT IN

DECEMBER, 1919

IT Wr;.L AWARD ITS

MAGELLANIC GOLD MEDAL

TO THE AUTHOR OF THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING TO NAVIGATION, ASTRONOMY, OR NATURAL PHILOSOPHY (mERH
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 19I9, deliver free of postage
or other charges, his discover^', invention or improvement, addressed to the
President of the American Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance by some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the same motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author.

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publiclj
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as may be after the adjudication, either in a separate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

{Lll correspondence in relation hereto should be addressed
To THE Secretaries of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A,

Proceedings

OF THE

American Philosophical Society

Subscription — Three Dollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

Price, One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

For Promoting Useful Knowledge

JVeTv Series. Vol. XXII,

Part III, 4to, 44 pages. -^

{Lately Published)

Tertiary Vertebrate Faunas of the North Coalinga Region of Cali-
fornia. A Contribution to the Study of the Palseontologic
Correlation in the Great Basin and Pacific Coast
Provinces. ByJoHN C. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California.

Subscription— Five Dollars per Volume
Separate parts are not sold

(Address The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A.

PROCEEDINGS

OF THE

American Philosophical Society

HELD AT PHIL4pEfL^^IA

FOR PROMOTING USEFUL iWo^L^DGE
Vol. LVIII. 1919. No. 5.

CONTENTS

PAGE

The Basis of Sex Inheritance in Sphaerocarpos. By Charles E. Allen 389

The Application of Sanitary Science to the Great War in the Zone of the

Army. By Bailey K. x\shford . . . . . -317

PHILADELPHIA
THE AMERICAN PHILOSOPHICAL SOCIETY

104 South Fifth Street
1919

The American Philosophical Society

announces that an award will be made in the j^ear 192 1 of the

HENP.Y M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

The Control of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent, of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after the
award.

Attention is called to the following regulations governing the award
of the Prize :

Competitors for the prize shall affix to their essays some motto or name
(not the proper name of the author, however) and when the essay is forwarded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judges, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essays shall be in their author, subject to
the right of the Society to publish the crowned essay in its " Transactions '"
or "Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,
David Jayne Hill,
Simeon E. Baldwin,
John Cadwalader,
W. W. Keen,

William B. Scott,
President ex-officio.

Committee.

The essays must he sent, addressed to the
President of the American Philosophical
Society, No. 104 South Fifth Street, Phila-
delphia, U. S. A.

THE BASIS OF SEX INHERITANCE IN
SPH^ROCARPOS.

By CHARLES E. ALLEN.
(Read April 25, 1919.)

Sph.^rocarpos in Cultivation.

The studies here described were made upon cuhures of Sphcr-
rocarpos Donnellii Aust. growing under greenhouse conditions.
Through the kindness of Dr. A. B. Stout, living plants of this spe-
cies were received from the New York Botanical Garden on Jan-
uary 13, 1916. These plants, as Dr. Stout informed me, had been
obtained from Mr. Severin Rapp, Sanford, Florida. To a request
for additional material Mr. Rapp very generously responded, as he
has to later similar requests, and the greater part of my present
cultures of this species began with material which he sent. Plants
have been received from Mr. Rapp on February 4 and March 10,
1916, on January 21 and February 23, 1918, and on April 10, 1919.
On February 16, 1916, plants of Sphcerocarpos were obtained for
class use from the Plant Study Company, Cambridge, Massachu-
setts. A letter from the company informed me that they had been
collected at Miami, Florida, by Miss Clara Hart. Some of these
plants were used also as starting points for greenhouse cultures.
Plants of all the cultures here referred to have been identified as
those of 5. Donnellii, and this identification has been confirmed by
Miss Caroline C. Haynes, to whom were sent representatives (in-
cluding sporophytes) of each set of cultures except those derived
from the last lot of plants supplied by Mr. Rapp.

My experience has shown that the thalli of 5". Donnellii (and of
S. texamis Aust. as well) can be kept growing indefinitely, the pos-
terior portions dying as growth continues at the anterior end. This
applies to the gametophytes of both sexes, although the male thalli
are much more susceptible to unfavorable conditions and it is only
with some care and difftculty that cultures of the male plants can be

PROC. AMER. PHIL. SOC, VOL. LVIII, S, AUG. 7, IQIQ. 289

290

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

kept in healthy condition. In mixed cultures the female plants
crowd and choke out the males, which latter in such cultures are
sure sooner or later to disappear. As already indicated, I have
cultures (now purely female) which have been growing continu-
ously since the early months of 1916, the plants having multiplied
as a result of their branching and apical growth and of the death of
the older portions, as well as by regeneration which occurs freely
under a variety of conditions from the lateral lobes, from the body
of the thallus, and from the involucres. Sporophytes have been
formed and the spores scattered in some of the cultures, in w^hich
cultures therefore not all the individuals now living are the result
of the vegetative growth of those originally present. It is easy, how-
ever, to obviate this possibility, so that a culture (clon or pure line)
of any desired extent can be obtained which is known to have been
derived by vegetative means from a single gametophyte, or from

Sphccrocarpos Donnellii. Living female (Fig. i) and male (Fig. 2)
plants. Received from Sanford, Florida, April 10, 1919; drawn April 15,
1919. A very small portion of the anterior end of each plant (the upper ends
in the figures) represents growth since transference to the greenhouse; other-
wise the plants are typical of those in nature. Drawings by Miss Martha
Engel. X 7-

the germination of a single spore. ]My oldest male cultures date
from spores which were sown June 16, 1916, the plants derived
from which were transplanted and have been kept in culture since
March 10, 191 7.

ALLEN— SEX INHERITANCE IN SPH.EROCARPOS. 291

C. and R. Douin (1917) report having kept female plants of S.
tcrrcstris (=^. Micheiii Bell.) and ^. caUfornicns {=S. texanus
Aust.) in cultivation for nine or ten months, and conclude that the
female thalli " may live indefinitely at a suitable humidity." The
male thalli, however, they say, are much less resistant and " dis-
appear little by little in the presence of humidity." Campbell
(1896) has noted the appearance of Sphcerocarpos plants (appar-
ently of both sexes) in a culture of other liverworts grown under
glass, and Goebel (1907) describes a culture of female plants which
had grown vigorously for two years. So far as I know, these are
the only previously published accounts of Sphcerocarpos in culture.
In my own work, best results have been obtained by growing the
plants in pots containing a mixture of about equal parts of clay loam
and sand. The pots stand on earthenware plates or enameled metal
pans in a Wardian case ; a little water is kept in the plates or pans.
The soil in the pots is thus constantly moistened from below. Un-
der these conditions contamination of the cultures by other organ-
isms is kept at a minimum, though of course by no means entirely
prevented ; and, even in case plants of both sexes are growing in
the same pot, fertilization does not occur. Fertilization can be
brought about when desired by flooding from above with sterilized
water, a pot containing male and female thalli. The plants must be
transplanted occasionally, both to relieve overcrowding and to free
the cultures from contamination, especially by blue-green algae,
which are the greatest source of trouble. The only other weeds
that cause serious annoyance are mosses, and they are not difficult
of exclusion when a culture is once free from them.

Campbell (1896) and C. and R. Douin (1917) have noted that
Sphcerocarpos under greenhouse conditions takes on a form notice-
ably dififerent from that which it exhibits in nature. Campbell's
plants were either (or both) those of S. caUfornicns Aust., which
Miss Haynes (1910) and the Douins agree is identical with 6".
texanus Aust., or of 6^. cristatus, a California species later separated
and described by Howe (1899). The Douins' statements apply to
both S. Micheiii Bell, and S. texanus.

The observations of these authors are confirmed by my own on
S. DonnelUi. Figs, i and 2 represent female and male plants, re-

292

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

spectively, drawn (living) a few days after their receipt from their
native habitat (Sanford, Florida), and therefore fairly typical of
the wild form. The female and male plants shown in Figs. 3 and 4
were taken from greenhouse cultures nearly three years old. The
most conspicuous differences result from the fact that the plants of
both sexes, and particularly their vegetative parts, grow more rap-

Spharocarpos Donncllii. Living female (Fig. 3) and male (Fig. 4)
plants from greenhouse cultures. Rhizoids not shown in Fig. 4. Drawings
by Miss Martha Engel. X 7-

idly and more luxuriantly under favorable conditions indoors ; the
branching is thus more noticeable, and the lateral lobes especially,
which in the wild plants are ordinarily insignificant, become, as pre-
vious observers have noted, decidedly leaf-like. The involucres
surrounding the sex organs become slenderer and often longer, the
archegonial involucres being characteristically tubular in the green-
house form. The orifices of the archegonial involucres are fre-
quently quite wide in plants grown in the greenhouse, though this
was not true of most of the involucres of the particular plant shown
in Fig. 3_. As a result of the greater development of the vegetative
parts and the lessened diameter of the involucres, the latter (in both
male and female) appear less crowded in the cultivated than in the
wild form. Similar modifications appear in cultures of 5. tcxanns,
as is shown by a comparison of Figs. 5 and 6 (wild form) with
Figs. 7 and 8 (greenhouse form). The plants shown in Figs. 5

ALLEN— SEX INHERITANCE IN SPH.EROCARPOS. 293

and 6 were sent from Austin, Texas, by Professor F. McAllister;
those shown in Figs. 7 and 8 were from cultures which began with
plants received from Professor R. S. Cocks, New Orleans, January
15, 1917, the drawings having been made in February, 1919.

It must be noted, in considering the difit'erences in question, that

Sphccrocarpos texanus. Living female (Fig. 5) and male (Fig. 6) plants
as found in nature. Received from Austin, Texas, February 13, 1919; dravirn
February 15, 1919. Drawings by Miss Martha Engel. X 7-

in such characters as the size and form of the lateral lobes, size,
form, and closeness or distance apart of the involucres, and size of
orifice, there is great variation between individual plants in the cul-
tures, as well as evidently also in nature. Compare, for example.
Figs. 3 and 9, both representing female plants oi S. Donnellii from
greenhouse cultures, that shown in Fig. 9 bearing sporophytes.
Much of this variation is obviously due to environmental conditions,
but some of it apparently to the existence of differing strains of
sub-specific rank.

The changes that appear under cultivation seem not to be corre-
lated with any loss of function on the part of the sex organs or of
the gametes. At any rate, it is easy, as already noted, to secure fer-
tilization, as a result of which sporophytes are formed in abundance,
especially in late winter and spring. At other seasons either fertili-
zation or sporophytic development seems to meet with more diffi-
culty, and the proportion of sporophytes obtained is ordinarily
smaller.

294

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

The sporophytes appearing in the cultures have given rise to
spores which were apparently normal in every respect and which
germinated, although thus far there has always been a considerable
proportion of ungerminated spores. Whether this failure of some
of the spores to germinate is due to a lack of viability or to a failure
to provide the most favorable conditions for germination is a ques-
tion yet to be determined. The spores of S. Donnellii in my cul-

Sphccrocarpos texanus. Living female (Fig. 7) and male (Fig. 8) plants
from greenhouse cultures. Rhizoids not shown. Drawings by Miss Martha
Engel. X 7-

tures invariably remain united in tetrads (Fig. 11), even to a time
when the capsule wall, calyptra, and surrounding involucre have
completely broken down and the spores consequently are being
scattered.^ Such a persistent union of the spores is characteristic,
as is well known, of other species of Sphcurocarpos, but not, accord-
ing to Miss Haynes (1910), of 5*. Donnellii as found in nature.

The Sex Ratio in Sph^rocarpos.

C. Douin (1909) investigated the sex of the respective members
of groups of the thalli of Sphccrocarpos, the members of each group
having resulted presumably from the germination of the spores of
a tetrad. His observations showed that, at least as a general rule,

1 Since this was written, it has been observed in one culture, containing
numerous sporoph^^tes, that the tetrad walls had largel.v broken down, allow-
ing the spores to become separated from one another.

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

295

two of the four spores of a tetrad develop into female plants and
the other two into male plants. In 9 cases of 81 studied, more
than four plants were present in a single group ; but on the assump-
tion that two tetrads had germinated in close proximity, these cases
were brought into harmony with the rule. In only four instances
of the 81 the distribution of sexes was "abnormal" (5 females and
3 males, i female and 3 males, and in two cases 3 females and i
male). The plants thus studied by Douin apparently included rep-
resentatives of both 6". Michelii and S. texanus. The evidence thus
furnished was cited by Strasburger (1909) as proof that in Sphccro-

FiG. 9. A female plant of Sphccrocarpos Domicilii from a greenhouse
culture; sporophytes are present within the swollen bases of several of the
involucres, as at s, s, X 7- Fig. 10. A portion of a similar plant in vertical
section showing the relation of the sporophytes (s, s) to the structures of
the gametophyte. X 7- Fig. ii. A group of four spores, resulting from the
division of a single spore mother cell and remaining permanently attached.
X 312. Drawings by Miss Martha Engel.

carpos the sex tendencies, brought together in the union of &gg and
antherozoid and remaining united, thoi:gh of course not finding ex-
pression, during the life of the sporophyte, are separated by the

296 ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

reduction divisions which resuh in the formation of four spores
from each spore mother cell — a notion strongly suggested by pre-
vious investigations of dioecious liverworts and mosses.

Spores being available in considerable numbers in my cultures,
it seemed worth while to attempt to determine whether the conclu-
sions thus arrived at for the European species hold likewise for
6^. DonnelUi. Spores liberated by breaking the capsule wall in a
drop of water were sown on soil June i6, 1916. Spores sown at
this time of year have been found to be relatively slow in germina-
tion ; the sporelings appeared during the following autumn and win-
ter. On March 10, 191 7, most of the plants that resulted having
begun to produce involucres, those that had grown from the spores
from two capsules were carefully examined, with the results shown
in Table I. As a rule, plants of a group of four could be reason-

TABLE I.

Sex of Plants Grown from Spores from Two Capsules of
Sphccrocarpos Donncllii (Culture PSC2B).

Spore tetrads sown June 16, 1916; plants examined March 10, 1917.

2 females, 2 males 3 groups

2 females, i male 2 groups

I female, 2 males i group

1 female, i male i group

2 females, 1 group

2 males 3 groups

I female, i group

I male i group

1 female, 2 males, i ? 3 groups

2? I group

4? I group

3 females, i male, i group

4 females, 2 males, i group

2 females, 3 males, 2? i group

3 females, 2 males, 3? i group

Totals: 30 females, 32 males, 14 (?).

ably assumed, because of their contiguity, to have come from the
spores of a single tetrad. Sometimes, however, the plants in a
group were sufficiently separated to make their relationship, though
probable, open to question. Isolated plants, or groups of two or
three, evidently indicated that one or more spores had failed to

ALLEN— SEX INHERITANCE IN SPHyEROCARPOS. 297

germinate, or that the sporeHngs growing from them had died. A
question mark ( ?) in the table indicates that no involucres had yet
appeared upon the plant in question, so that its sex could not be
determined.

It is plain that all the cases shown in Table I., with the excep-
tion of the last four, harmonize with the expectation of two females
and two males. The last three cases can be made to harmonize by
the assumption made by Douin, namely, that two spore tetrads
germinating in close proximity may give rise to a group of eight
plants (or fewer), which will be expected to include four females
and four males. There remains one case (of three females and one
male) which does not agree with the expectation, unless it be
assumed to represent the four survivors of a group of eight.

However, it appeared from this study, as well as from a more
extended observation of the behavior of plants in culture, that the
question of the sex potentialities borne by the spores can be finally
settled only by the use of more exact methods. The sources of
possible error are at least three : the impossibility of determining
with certainty just which plants have come from the spores of a
single tetrad ; the multiplication of plants as a result of their branch-
ing and the separation of the branches as independent plants, either
by accident or by the death of the older portions ; and the produc-
tion of new plants by regeneration. All these sources of error are
likewise present in the observation of plants growing in nature,
although the second is perhaps less important under these conditions
because of the slower growth out of doors. It is thus quite possible
that Douin's " abnormal " cases may not be real exceptions to the
general rule ; although, on the other hand, it must be admitted as at
least conceivable that in an occasional instance an error from one
of the causes mentioned might make a really aberrant case seem to
agree with the rule. The possibility of a modification of the appar-
ent sex ratio as a result of branching or of regeneration has recently
been recognized by Douin (C. and R. Douin, 1917).

Attempts to develop better methods of studying this question
led to the sowing of isolated tetrads on various substrata. Diffi-
culties have been met with in securing favorable conditions for
germination and in preventing contamination, and the results thus

298

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

far are small. In one culture, however, a sufficient number of
germinations were obtained to make the results of some value (Ta-
ble II.). In this case 24 tetrads, all from a single capsule, were
sown on filter paper which was moistened with a nutrient solution
made according to the formula given by £. and £. Marchal (1907),

TABLE II.

Sex of Plants Derived from Spores from One Capsule of
Sphccrocarpos Donnellii (Culture R19).

Spore tetrads sown separately on moistened filter paper March 17, 1917;
sporelings separated and removed to pots of soil May 25 to June i, 1917.

2 females, 2 males i group

2 females, i male 2 groups

I female, 2 males i group

1 female, i male i group

2 females, i group

I female, i group

1 male i group

I female, i male, i ? i group

I female, 2? 2 groups

I female, i? i group

2 males, i ? i group

I male, i ? 3 groups

2?- I group

Totals: 15 females, 14 males, I2(?).

placed in a petri dish, and sterilized before the spores were sown.
The sporelings were removed from the filter paper while they were
still very small and while their attachment to the spore wall could
still be distinguished ; there could thus be no question as to the origin
of the plants of a group, and their youth precluded the possibility
of a previous multiplication by vegetative means. The sex of the
plants was determined, so far as this could be done, at the time of
their removal from the filter paper, and each was transplanted to a
separate pot of soil. In the cases in which the plants continued to
live after transplanting, observations as to their sex were made
after further development. Some of the plants died after trans-
planting and before producing involucres, and these are the ones
whose sex is indicated as doubtful in the table.

Although more extended observations are desirable, it seems

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 299

quite clear from these results that the general conclusion arrived at
by Douin and Strasburger for the species of Sphccrocarpos consid-
ered by them applies also to S. DonnelUi — namely, that the differ-
ences which distinguish female from male plants result from differ-
ences in the spores that are to give rise to them, and that of the
spores formed by the division of a single mother cell two bear fe-
male potentialities only and two male potentialities only. It seems
probable, too, that exceptions to this rule, if any occur, are less
frequent than might be thought from Douin's observations — his
apparently aberrant cases probably having resulted from some of
the causes of error already noted. It is not intended to be sug-
gested, however, that positive exceptions to the rule will not be
found. The possibility of such exceptions will be discussed on a
later page. The occurrence or non-occurrence of an exceptional
distribution of sex characteristics, and the proportion of such occur-
rences, if any, are evidently to be determined by the observation of
thousands of plants rather than of a few score.

It will be noticed from Tables I. and II. that the number of
plants certainly referable to the respective sexes is substantially
equal: in the one case 30 females and 32 males, in the other 15 fe-
males and 14 males. Thus it appears that so far as power of germi-
nation is concerned there is no marked dift'erence between female-
producing and male-producing spores. This fact is striking in view
of the great difference in favor of the female plants in rate of
growth and in power of resistance to unfavorable conditions. Ap-
parently this difference, which for want of a better word may be
spoken of as one of "vigor," does not exist, or at least come to ex-
pression, as between the spores, but becomes apparent at some time
after germination and during the development of the gametophytes.

While there seems to be no difference between the female- and
male-producing spores in their capacity for germination, it remains
possible that there is a difference in time or rate of germination.
On this question the foregoing observations throw no light, but a
suggestion is furnished by a recent set of experiments in which an
attempt has been made to study this particular point. On December
27 and 28, 1918, 67 tetrads from two capsules were sown in as many
pots of soil, one tetrad to a pot. Germination has been slow, and

300 ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

additional sporelings are still occasionally appearing. However,
the slowness of germination has given opportunity for the entrance
of blue-green algae, whose presence in quantity either checks further
germination or results in the death of the young sporelings while
very minute. The germinations observed in this series have re-
sulted as follows :

In 12 pots, I female sporeling;

In 4 pots, I male sporeling ;

In I pot, I of undetermined sex;

In I pot, I female followed by i female ;

In I pot, I male followed by one male, and this by one of undeter-
mined sex (the latter possibly, though apparently not, a regenerated
shoot) ;

In I pot, 2 males appearing nearly or quite simultaneously ;

In I pot, I female followed by 2 of undetermined sex.

Thus in 14 cases a female-producfng spore seems to have been
first to germinate ; in 6 cases, a male-producing spore. While too
fragmentary to furnish a basis for a conclusion, these figures sug-
gest that there may be a difiference in rate of germination in favor
of the female-producing spores.

The Chromosomes of Sph^rocarpos.

SphcErocarpos suggested itself as a favorable plant in which to
look for a possible chromosome difference between the sexes, both
because of its marked sexual dimorphism and because of the strong
evidence of a relation between chromosome reduction and the sepa-
ration of sex potentialities. So far as published records show, the
only previous attempt to study the chromosomes of SpJiccrocarpos
was by Strasburger (1909), who reports that neither in the nuclear
divisions in the spore mother cell nor in the structure of the nuclei
formed by these divisions did he find any evidence of the separation
of structures that could be interpreted as the bases of sex differ-
entiation.

In April, 1914, through the courtesy of Professor Mangin and
of M. Capus of the Museum d'histoire naturelle of Paris, I was
enabled to locate living plants of Sphcrrocarpos (probably of 5".
Michelii) in the neighborhood of Bois-le-Roi. Some of this mate-

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 301

rial was fixed and studied in Professor Mangin's laboratory. The
amount available was too small and the stage of development for
the most part too advanced to furnish definite results ; but it gave
an opportunity for acquiring a familiarity with the plants and for
testing methods of fixation. When greenhouse cultures of 5". Don-

<s^

14

Figs. 12 and 13. Chromosome groups from female gametophytes of
Spharocarpos Donncllii: Fig. 12, from a basal cell of an archegone ; Fig. 13,
from the basal part of a young archegonial involucre ; x, the large chromo-
some. Figs. 14 and 15. Chromosome groups from male gametophytes of
S. Donnellii: Fig. 14, from the wall of a young antherid; Fig. 15, from a
projecting superficial cell of the apical meristem, possibly an antheridial
initial; y, the small chromosome. X 3800.

nellii were later obtained in vigorous condition, the cytological study
was renewed. Some of the results of this study have already been
briefly reported (Allen, 1917&).

Figs. 12 and 13 show typical chromosome groups from cells of
the female gametophyte ; Figs. 14 and 15, corresponding groups
from cells of the male gametophyte. In each case, as the figures
show, eight chromosomes are present. In the female, one of the
eight chromosomes {x, Figs. 12, 13) is much longer and thicker than
any of the others. The other seven differ in length among them-
selves. It is probable that all the individual chromosomes of differ-
ent cells can be identified by their length;. but since in sections the
chromosomes lie at various angles and thus some of them are always
foreshortened in the camera lucida drawings, it will require the
assembling and study of a considerable number of figures showing

302

ALLEN— SEX INHERITANCE IN SPHyEROCARPOS.

the chromosomes in favorable positions to make such identification
reasonably certain. However, in all of the numerous division fig-
ures that have been seen in various parts of female plants the single
large chromosome is present and conspicuous.

The chromosome groups of the male contain no element at all
like the large chromosome of the female. Seven of the chromo-
somes of the male, varying in length among themselves, seem to cor-
respond to the seven smaller ones of the female. The eighth chro-

FiG. i6. A group of cells from a developing antheridium ; nuclei in late
prophase ; y, y, the small chromosome. X 3800.

mosome (3', Figs. 14, 15, 16) is very small. This small chromosome
is not always easily distinguishable. In the group shown in Fig. 14,
for instance, the body y is very lightly stained ; the same is true of
the body similarly identified in cell A, Fig. 16. However, in the
majority of cases, as in cells B, C, D, and E, Fig. 16, and in the cell
shown in Fig. 15, in which the chromosomes are unusually widely
scattered, this small element is stained like the other chromosomes
and is plainly one of them. It appears certain, therefore, that a
very small chromosome in the male in some way corresponds to, or
replaces, the very large one of the female.

Figs. 17 to 21 show stages in the division and separation of the
chromosomes in dividing cells of the female. In the cell represented
in Fig. 17 the chromosomes or most of them, including the large
one, are longitudinally split. Figs. 18 to 20 illustrate a peculiarity
in the behavior of the large chromosome — namely, that its daughter

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

303

halves are the last to separate in the metaphases (Fig. i8), and that
they lag behind the other daughter chromosomes in their passage
to the poles of the spindle (Figs. 19, 20). Even at the late stage
shov/n in Fig. 21, the large daughter chromosome moving toward
the lower pole of the figure is still a little in the rear of its fellows,
which have reached the spindle pole. In this respect the behavior
of the large chromosome recalls that of the X-chromosome in the
heterotypic divisions of certain insects, and in some cases at least

... -x

17

Figs. 17-21. Stages in nuclear division in the female gametophyte :
Fig. 17, a late prophase, the chromosomes showing longitudinal splitting, in a
cell of the calyptra; Fig. 18, a metaphase in an involucral cell; Fig. 19, an
anaphase in an involucral cell ; Fig. 20, a later anaphase in a young thallus
lobe; Fig. 21, a diaster in a cell at the base of a young lobe; x, the large
chromosome. X 3800.

in the homceotypic divisions of these animals as well. It is quite
possible, to be sure, that this lagging of the large chromosome of
Sphcvrocarpos is merely a result of its proportionately great size.
Relatively few division figures have been observed as yet in the

304 ALLEN— SEX INHERITANCE IN SPH.^ROCARPOS.

developing sporophyte. Those which have been seen show that
about sixteen chromosomes are present and that one and only one
of' the group is much larger than the others, being, therefore, ob-
viously derived from the female parent. Fig. 22 represents the
two anaphase groups in a dividing cell of a young sporophyte ; each
group shows fifteen chromosomes, one of which is the large one.
The sixteenth chromosome which would be expected to appear in

£

0^%

'V 'i%

V

I

Fig. 22. Two anaphase chromosome groups in a dividing cell of a young
sporophyte ; A, the group seen at an upper, B, that seen at a lower, focus ;
X, the large chromosome. X 3800.

each of these groups is, it seems likely, the small element derived
from the male parent, which here is not seen either because of its
too light stain or because it is hidden by one of the others.

The first division of the spore mother cell nucleus (the hetero-
typic division) has not yet been seen; but a considerable number of
spore mother cells have been found in which the second (homoeo-
typic) division was in progress. In every one of these latter cases
in which the complete series of sections of a spore mother cell was
present, one of the two spindles in the cell bore a large chromosome,
and the other did not. Figs. 23 and 24 show polar views of the two
homoeotypic metaphase groups in a single spore mother cell ; in the
group of Fig. 24 a large chromosome (apparently longitudinally
split) is present; the group of Fig. 23 includes no large chromo-
some. An exact count of the smaller chromosomes has not been
found possible at this stage because of the irregularities in their
shape and distribution. The fact that a single group is commonly
represented in several sections (those chosen for Figs. 23 and 24
are unusually favorable in this respect) involves additional diffi-
culties.

The conclusion seems warranted, from the presence of a large

ALLEN— SEX INHERITANCE IN SPH.^ROCARPOS. 305

chromosome on only one of the two homoeotypic spindles in a spore
mother cell, that in the heterotypic division the large and small chro-
mosomes pass respectively to different daughter nuclei. This means
that the separation of these two very different bodies is effected in
the same division as that which, on the basis of the best cytological
evidence, seems to result in the qualitative separation of other ("or-
dinary ") chromosomes. It remains to determine from a study of
the heterotypic division just how this separation is brought about.

24

■@ B

Figs. 23 and 24. The two chromosome groups, seen in polar view, in a
metaphase stage of the homoeotypic division in a spore mother cell. The
great majority of the chromosomes on the two spindles (shown in Figs. 235
and 24 A) were in the same section; Fig. 23 A shows a single chromosome
belonging to the group of Fig. 23 B, appearing in the previous section ; and
Fig. 24 B, two chromosomes (or a dividing chromosome) belonging to the
group of Fig. 24 A, and appearing in the following section; x, the large
chromosome. X 3800.

Of the four nuclei formed in the spore mother cell, two sister nuclei,
and therefore the two spores into which these nuclei pass, receive a
large chromosome each ; the other two nuclei (and spores) receive a
small chromosome each. Since the large chromosome is always
present in the cells of the female and never in those of the male, it
follows that a spore receiving a large chromosome must develop, if
it develops at all, into a female gametophyte, and that a spore re-
ceiving a small chromosome must develop, if it develops at all, into
a male gametophyte.

The Chromosomes and Sex Characters.

What I have hitherto referred to respectively as the "large"
and the " small " chromosome of Sphccrocarpos are so closely simi-
lar in appearance and behavior to the sex chromosomes of various

PROC. AMER. PHIL. SOC . VOL, LVIII, T, AUG. H. I919.

306 ALLEN— SEX INHERITANCE IN SPH.EROCARPOS.

animals — especially of such insects as Lygcciis and Euschistus (Wil-
son, 1906, 1912; Montgomery, 191 1) — that in the remainder of the
present discussion they will be designated, as commonly are the
apparently corresponding bodies in animals, the " X- " (large) and
"Y-" (small) chromosomes. This terminology is borrowed for
the sake of simplicity, in spite of the fact that it may require revision
in the light of future investigations.

Whatever may later appear in this respect, at present the prob-
lem of the relation of the X- and Y-chromosomes to the complexes
of characters that distinguish male from female seems to be similar
in its broadest terms in Sphcerocarpos and in those animals whose
possession of sex chromosomes has been demonstrated. The sug-
gestion of an underlying similarity, however, does not involve main-
taining that the precise relation of these particular chromosomes to
certain hypothetical sex-determining factors is identical in all organ-
isms in which X- and Y-chromosomes are found. Indeed, there
appear at present to be, as regards the relation between special
chromosomes and characters distinctive of sex, three quite different
types of cases :

(a) Those in which, in terms of the chromosomes, the female
is homozygous because it possesses what Wilson (1909) has called
two " X-elements " — the X-element in different instances being a
chromosome or a group of chromosomes ; and the male heterozygous
because it possesses either one X-element only, or one X- and one
Y-element ; the Y-element, if present, sometimes consisting of one
body, sometimes of a group. To this class are now referable a
large number of animals representing several phyla.

{h) Those in which the male is homozygous, possessing two
X-elements, and the female heterozygous, possessing one X-element
only, or one X- and one Y-element. This class, apparently smaller
than the first, is established upon the basis of suggestive cytological
observations which in no case as yet cover the whole life history,
and of a larger amount of evidence derived from experimental
breeding.

(c) Those in which the female possesses one X-element (or
chromosome), the male one Y-element, and in which the diploid
generation (the sporophyte) is heterozygous in terms of the chro-

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 307

mosomes although phenotypically asexual. To this class can be
assigned at present with certainty only two species of Sphccrocarpos:
S. Donnellil, discussed in the present paper, and S. tcxanus, in
which Miss Schacke (1919) has recently found a condition as to
chromosomes similar to that in S. Donnellii. It may be hazarded,
however, that here belong also the dioecious mosses upon which the
epoch-making studies of the Marchals (1907, 1909, 191 1) were
made ; the heterozygous sporophytes of these mosses were induced
to give rise by regeneration to diploid and hermaphroditic game-
tophytes.

The differences in the apparent relations of the chromosomes in
question in the organisms representing these three classes are
brought out strikingly by the facts that in class a two X-elements
are necessary to the appearance of femaleness, but only one X-ele-
ment (with or without a Y-element) to that of maleness ; that in
class b, two X-elements are requisite to the appearance of maleness,
and only one X-element (with or without a Y-element) to that of
femaleness ; whereas in class c the presence of one X-element means
femaleness, that of one Y-element maleness, and the presence of
both, under ordinary conditions, is correlated with non-sexuality.

Another difference, at least as between the plants of class c on
the one hand and the animals of class a on the other, lies in the
apparent lack of function in h-eredity of the Y-chromosome or
-element in animals, as shown by its utter absence in many forms
and by the genetic evidence of its failure when present to influence
the transmission of sex-linked characters; although Bridges' (1916)
evidence in this connection should be cited, namely, that the absence
of the Y-chromosome in Drosophila, while not affecting the appar-
ently normal development of the male animal, does result in steril-
ity. In Sphccrocarpos, however, there is no reason at present for
considering the Y-chromosome to be in any sense a functionless
body. Its presence seems, as will be indicated more fully in a later
paragraph, to be related to the appearance of definite characters in
just as positive a way as is the presence of the X-chromosome.
And in the normally dioecious mosses studied by the Marchals,
heterozygosis in the diploid gametophyte manifested itself by the
appearance of both male and female characters.

308 ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

The visible differences between male and female plants of Sphcs-
rocarpos fall naturally into two categories :

(o) Differences in rate of growth and in the size at maturity of
homologous parts (compare, for example, Figs, i and 2, or 3 and
4), and, probably intimately connected with these, a dift'erence in
power of resistance to unfavorable conditions ; in each of these re-
spects the advantage is with the female.

{h) Differences in form and structure of the sex organs (an-
theridia and archegonia), and of course of the gametes themselves,
as well as in the form and structure of the involucres surrounding
the sex organs. These differences in structure are associated with
size differences — for example, the archegonial involucre is much
larger than the antheridial involucre, as well as different in form —
but it does not follow that characters of size and those of form,
though constantly associated and even causally related, have been
brought to expression by the same chain of causative factors.

Of none of these distinctive characters, it may be noted, is there
any apparent reason for suspecting that it has anything in common
with the " secondary sexual characters " of the higher animals.

That the possession by each sex of its own complex of distinctive
characters — hereinafter referred to for the sake of brevity as " sex
characters " — constituting a constant phenotypic difference between
the sexes, is the expression of a genotypic difference, can hardly, I
think, be doubted ; for all the available evidence indicates that
SphcBvocarpos is strictly and under all circumstances dioecious, and
that an individual gametophyte possesses one and only one set of
potentialities so far as sex characters are concerned. This being
the assumption upon which discussion must for the present rest, it
follows that the respective groups of sex characters (or their phys-
ical bases) are separately inherited by the sexual from the asexual
generation through the spores, which latter are the only possible
vehicle of transmission from sporophyte to gametophyte. Since
two spores of each tetrad develop into plants of either sex, the con-
clusion already drawn by previous writers seems inevitable, namely,
that the physical bases for the sex potentialities were united in the
sporophyte down to the time of the formation of the spore mother
cells, and that the separation of these physical bases occurred in the

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 309

course of the divisions which formed the spores. The invariable
connection of the X-chromosome with femaleness and of the
Y-chromosome with maleness seems to prove a causal relation of
some sort between the presence of either chromosome and the ap-
pearance of the characters of the corresponding sex. The further
logical step seems likewise inevitable, although admittedly it raises
questions of considerable difficulty — the conclusion, namely, that it
is the presence of an X- or of a Y-chromosome in the cells of a
particular plant which determines the appearance in that plant of
the characters of the corresponding sex. This essentially corre-
sponds with the hypothesis which the students of the sex chromo-
somes of animals seem in general to have adopted — although natu-
rally the form of the statement, as of the precise fact, is different
for an organism whose sexual generation is diploid rather than
haploid.

The facts in the life of Sphcvrocarpos that seem to force this con-
clusion upon us are, first, that, as already noted, the physical bases
for the sex characters must have been separated in the course of
the divisions in the spore mother cell and distributed among the
four resultant spores ; and second, that the only structures that are
shown to have been so separated and distributed are the X- and
Y-chromosomes — the four spores being to all appearances alike in all
other points of structure, form, and size. The difficulty of imagin-
ing how the sex chromosomes can function in impressing sex char-
acters upon the plant — a difficulty which, however, is but of the
same order as that which confronts the very widely accepted theory
of the function of the chromosomes in inheritance in general — will
continue to breed a healthy skepticism and will stimulate the attempt
to find another basis for sex inheritance ; but at present this hypothe-
sis of the significance of the sex chromosomes seems the only intel-
ligible one and the one therefore upon which further investigations
must rest.

The question then arises as to how the sex chromosomes of
Sphcvrocarpos can be conceived as exercising their controlling influ-
ence. Apparently this question may be answered in different ways
as regards the two categories of sex characters already mentioned.

The notion suggests itself at once that the differences in size and

310

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

in rate of growth of the respective gametophytes may well result
from the difference in bulk of chromosome substance present in the
cells of female and male respectively. The size of the X-chromo-
some is such that the chromosome group of the female exceeds in
bulk that of the male, at a rough estimate, by perhaps fifty per cent.
In view of this difference, and of the differences in cell size which
in many cases have been shown to result from differences in chro-
mosome number — as, for example, between haploid and diploid
moss gametophytes (fi. and £. Marchal, 1909), between haploid

Fig. 25. A typical fully developed lateral lobe of a female gametophyte.
Fig. 26. A similar lobe of a male gametophyte, drawn to the same scale.
Both from greenhouse cultures of Sphccrocarpos Domicilii. X 36.

and diploid Spirogyra cells (Gerassimow, 1901), and between the
ordinary and gigas forms of tomato and nightshade (Winkler,
1916), it would not have been surprising to find a marked difference
in cell size between homologous members of the male and female
gametophytes of Sphccrocarpos. However, some measurements
made to test this possibility showed that in corresponding parts,
such as the lateral lobes, the range of cell sizes is substantially the
same in the two sexes. This fact appears clearly in the camera
lucida drawings of typical mature lobes from a female and a male

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 311

plant respectively (Figs. 25, 26), and of mature archegonial and
antheridial involucres (Figs. 27, 28). These figures, while not
negativing the possibility of a small difiference between the sexes in
average cell size — a possibility to be tested only by a great number
of measurements — demonstrate that such differences, if they exist,
are negligible as compared with the difference either in total chro-
mosome volume or in the total surface area of the chromosomes ;
and, what is more significant from the present point of view, that

Fig. 27. A typical fully developed archegonial involucre. Fig. 28. A
typical fully developed antheridial involucre, drawn to the same scale. Both
from greenhouse cultures of Sphccrocarpos Donncllii. X 36.

such possible differences in cell size play no important part in bring-
ing about the marked differences in size of homologous organs. On
the other hand, as the figures referred to demonstrate, the charac-
teristic differences in size between male and female plants result
from the presence of a much greater number of cells in correspond-
ing parts of the female. Since these size differences appear be-
tween plants of the same age, it follows that cell growth and divi-
sion go on more rapidly in the female than in the male. If, there-
fore, the size and related differences between plants of opposite sex
are determined by the diff"erence in chromosome bulk, this quanti-
tative difference produces its ultimate visible effects by means of its

312 ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

influence on the rate of cell growth and of cell division. That this
is the case seems conceivable and indeed probable, because the over-
w^helming evidence of an influence of the chromosomes on inheri-
tance points to the exertion of this influence largely, at least, through
a determination of the rate and nature of constructive metabolism ;
and one effect of a modification of the rate of metabolism would of
course be an increase or decrease in the rate of cell growth and thus
also in the rate of division.

But if such a simple quantitative explanation can be adduced to
account for the influence of the sex chromosomes on the one class
of sex characters, it cannot with equal ease be made to account for
the characters of the second category — those which concern the
form and structure of gametes, of sex organs, and of involucres.

It is true that, in connection with the observed facts in animals,
several writers have suggested the possibility of a " quantitative
theory " of sex determination, which would make all primary sex
characters the expression of the amount or degree of activity of the
chromosome material present. Without entering into the discus-
sion of the validity of such a theory in the case of the metazoa, it
seems quite impossible of application to the class of characters under
consideration in Sphccrocarpos. For one thing, it is to be remem-
bered that in this plant, difl:'erently from the condition so common
in the higher animals as well as in the dioecious seed plants, there
are not two sets of reproductive structures, one functional, the other
rudimentary but conceivably capable (certainly capable in many
seed plants) of a normal development under particular conditions.
In such a case the stimulus leading to the functional development
of one set of structures or the other might conceivably result from
the presence of a greater or less quantity of particular nuclear sub-
stances. On the contrary, the male plant of Sphccrocarpos shows
no trace of archegonia or of archegonial involucres ; the female plant
shows no trace of antheridia or of antheridial involucres. Nor can
the differences between male and female structures be explained by
modifications of cell size operating differently in different cell axes.
Factors of some sort must be supposed to be at work which deter-
mine, very differently in the two sexes, the planes of successive cell
divisions; which modify, also in very different ways, the nucleo-

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 313

cytoplasmic ratio ; and which determine in the one case the consid-
erable but relatively slightly specialized development of the egg, and
in the other the remarkable metamorphosis of the androcyte into the
antherozoid. This is but a partial analysis of the nature of the
processes whose causes are to be explained, but it is sufficient to
illustrate the nature and complexity of the problem.

Perhaps it is not safe to say more at present than that the ap-
pearance of the sex characters falling within the form-structure
category are to be ascribed to factors which in some way are de-
pendent upon, carried by, or inherent in, the X- and Y-chromosomes
— the word " factor " here being used in its ordinary, not in a tech-
nical Mendelian sense. Nothing in our present knowledge of the
mechanism of inheritance in Sphccrocarpos would justify us in hold-
ing that the factors here in question are of the nature of those which
are postulated by any particular theory of heredity. The conclu-
sion, if it can be so called, to which we are for the present led, is
therefore that one category of sex characters is reasonably explain-
able by the difference in mass between the sex chromosomes ; and
that those of a second category seem to result from some other, but
unknown, specific peculiarities of the same chromosomes.

On the analogy of the irregularities which, it seems well estab-
lished, occur, though rarely, in the distribution of the sex chromo-
somes during the reduction divisions in certain animals, it is perhaps
to be expected that irregularities more or less like these will be
found in Sphccrocarpos. The occurrence of " non-disjunction," for
example, m.ight lead to the formation of a tetrad two of whose
spores have both an X- and a Y-chromosome each, the other two
spores possessing neither. It would be idle to speculate as to the
effect of such a distribution upon the viability of the spores or upon
the sexuality or sterility of the resulting plants ; but it is plain that
the result might well be a disturbance of the normal 2 : 2 sex ratio.

Other irregularities than non-disjunction might conceivably like-
wise lead to modifications of this ratio ; and so it will not be surpris-
ing should a study of the results of the germination of large numbers
of spores bring to light occasional exceptions to the general rule of
the distribution of sex characters. It is possible that some of the

314 ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

apparent exceptions found by Douin were the result of an aberrant
behavior of the sex chromosomes ; but reasons have already been
given for doubting whether these exceptions were other than appar-
ent. In the examination of spore tetrads I have come across a few
instances of spores joined in twos rather than in fours ; and other
instances in which a group consisted of two spores of normal size
and one or two very small. Such conditions may result from an
irregular distribution of the chromosomes ; but they may equally
well or perhaps better be explainable by the death or accidental
injury or destruction of one or two spores after the completion of
the division of the mother cell.

The existence of definite sex chromosomes having been estab-
lished for two species of Sphccrocarpos, it is to be expected that
similar bodies will be found in other plants. The most promising
organisms for such investigations are probably the other dioecious
bryophytes, especially those with a marked sexual dimorphism.
Next would come perhaps some of the dioecious algae. That pre-
vious searches for sex chromosomes in plants have been fruitless
has been largely because they dealt chiefly with dioecious seed plants.
Enough of these have now been examined to make it quite plain
that no visible chromosome difference is to be expected as between
the staminate and the pistillate individuals of any species. This
negative result is quite in harmony with what we now know regard-
ing Sphccrocarpos, because the distinction between staminate and
pistillate sporophytes is of a quite different sort from that between
male and female gametophytes ; and if anywhere in the seed plants
a chromosome difference is to be looked for exactly comparable
with that described in the present paper, it must be between the
micro- and the macrogametophytes. Evidence as to the possible
existence of such a difference is still meager; but such evidence as
does exist is, it must be admitted, of a negative character.

In a previous paper (Allen, 1917a) I have referred to Hirase's
(1898) description of a cytological difference between the two an-
therozoids produced by the same microgametophyte of Ginkgo,
which might be imagined to be the basis of the dioecism of the sporo-
phytes of this species. This difference consists in the presence of a

ALLEN— SEX INHERITANCE IN SPH^ROCARPOS. 315

conspicuous cytoplasmic body in one antherozoid and its absence in
the other. The difference is brought about, not during chromosome
reduction, but in the final division in the microgametophyte, and
has no apparent parallel, so far as I know^ in the cytology of any
other dioecious seed plant. However, little or nothing is known
regarding the corresponding division in other dioecious plants.

Correns' (1907) well-known experimental studies of sex inheri-
tance in Bryonia seemed to show the existence in B. dioica of male
gametes of two kinds with reference to sex potentialities. Correns,
as well as most of those who have attempted an explanation of his
results, has assumed that this involves the existence of two kinds of
pollen grains, so strong has seemed the probability that the segrega-
tion of any particular hereditary factors is brought about by the
reduction divisions. It is, however, quite possible to assume that,
so far as sex potentialities are concerned, the pollen grains of
Bryonia are alike, and that the segregation of sex potentialities
occurs in the division of the generative cell within the pollen grain
or pollen tube. This suggestion, though based only upon the ob-
scure phenomenon recorded for Ginkgo, is perhaps worth testing,
in view of the failure of all attempts to find other visible bases for
the separate inheritance of staminate and pistillate characters.

Department of Botany,

University of Wisconsin.

Literature Cited.

Allen, C. E. (1917a.) The Spermatogenesis of Polytrichnni jiinipcriniim.

Annals of Botany 31 : 269-291. pis. 15, 16.
. (1917&.) A Chromosome Difference Correlated with Sex Differences

in Sphccrocarpos. Science n. ser. 46: 466, 467.
Bridges, C. B. (1916.) Non-disjunction as Proof of the Chromosome

Theory of Heredity. Genetics i : 1-52, 107-163. pi. i. figs. 1-9.
Campbell, D. H. (1896.) Notes on Sphcerocarpus. Erythea 4: 73-78. pi. 2.
Correns, C. (1907.) Die Bestimmung und Vererbung des Geschlechtes nach

neuen Versuchen mit hoheren Pflanzen. pp. i-iv, 1-81. figs. 1-9. Berlin.
Douin, C. (1909.) Nouvelles observations sur Sphcerocarpus. Revue

Bryol. 36: 37-41. figs. i-io.
Douin, C, and Douin, R. (1917.) Notes sur les Sphcerocarpus. Rev. Gen.

de Bot. 29: 129-136. pi. 16.
Gerassimow, J. J. (1901.) Ueber den Einfluss des Kerns auf das Wachsthum

der Zelle. Bull. Soc. Imp. des Naturalistes de Moscou n. ser. 15: 185-

220. pis. A, B,

316 ALLEN— SEX INHERITANCE IN SPH^ROCARPOS.

Goebel, K. (1907.) Archegoniatenstudien XL Weitere Untersuchungen

iiber Keimung und Regeneration bei Riella und Sphccrocarpus. Flora

97: 192-215. figs. 1-23.
Haynes, C. C. (1910.) Spharocarpos hians sp. nov., with a revision of the

genus and illustrations of the species. Bull. Torrey Bot. Club 37: 215-

230. pis. 25-32.
Hirase, S. (1898.) fitudes sur la fecondation et I'embryogenie du Ginkgo

biloba (Second memoire). Journ. Coll. of Sci. Tokyo 12: 103-149.

pis. 7-9.
Howe, M. A. (1899.) The Hepaticse and Anthocerotes of California. Mem.

Torrey Bot. Club 7 : 1-208. pis. 88-122.
Marchal, £., and Marchal, £. (1907.) Aposporie et sexualite chez les

Mousses. Bull. Acad. Roy. de Belg., Classe des Sci. 1907, pp. 765-789.
, . (1909.) Aposporie et sexualite chez les Mousses II. Bull Acad.

Roy. de Belg., Classe des Sci. 1909, pp. 1249-1288.
, . (191 1.) Aposporie et sexualite chez les Mousses III. Bull. Acad.

Roy. de Belg., Classe des Sci. 191 1, pp. 750-778. i pi.
Montgomery, T. H., Jr. (1911.) The Spermatogenesis of an Hemipteron,

Euschistus. Journ. Morph. 22: 731-815. pis. 1-5.
Schacke, M. A. (1919.) A Chromosome Difference between the Sexes of

Spharocarpos texanus. Science n. ser. 49: 218, 219.
Strasburger, E. (1909.) Histologische Beitrjige VII. Zeitpunkt der Be-

stimmung des Geschlechts, Apogamie, Parthenogenesis und Reduk-

tionsteilung. pp. i-xvi, 1-124. pis. 1-3. Jena.
Wilson, E. B. (1906.) Studies on Chromosomes — III. The Sexual Dififer-

ences of the Chromosome-groups in Hemiptera, with Some Considera-
tions on the Determination and Inheritance of Sex, Journ. Exp. Zool."

3 : 1-40. figs. 1-6.
. (1909.) Recent Researches on the Determination and Heredity of

Sex. Science n. ser. 29: 53-70. 2 figs.
. (1912.) Studies on Chromosomes — VIII. Observations on the

Maturation-phenomena in Certain Hemiptera and other Forms, with

Considerations on Synapsis and Reduction. Journ. Exp. Zool. 13 : 345-

448. pis. 1-9.
Winkler, H. (1916.) tJber die experimentelle Erzeugung von Pflanzen mit

abweichenden Chromosomenzahlen. Zeitschr. Bot. 8: 417-531. figs. 1-17.

pis. 4-6.

THE APPLICATION OF SANITARY SCIENCE TO THE
GREAT WAR IN THE ZONE OF THE ARMY.^

By bailey K. ASHFORD.
{Read April 24, 19 19.)

It may be somewhat of a disappointment that I should present
so little that is new to this Society, but war has requisitioned of
science its application and not its theories and each man has been
called upon to produce a result. The interest in this subject lies in
seeing how far sanitary science could be applied on the battlefield
and what part it took in winning this war. Let us answer the last
question first :

In the Civil War there were 65 deaths from disease annually per
1,000 of strength; in the Spanish-American War there were 30; in
the Greatest of Wars just concluded there were 14.8 and of these
about 12 were due to epidemic pneumonia. That is to say, the med-
ical sciences have kept 100,000 American men in fighting trim w^ho
would have in 1861 died of disease, and America needed those men.
There was no time to get trained men from home, nor vessels in
which to send them ; the expense per man, the least consideration,
was enormous, and, above all, if the liberty of the world was to be
won at all it had to be done by the brawn of these very men and like
men of our Allies. It had to be done then. The hour had struck.
All the treasure and patriotic devotion of our country was helpless
to do more in this crisis. When one reflects that never have armies
been forced to live under more menacing conditions to health than
those under which our soldiers lived and fought — at times, for mili-
tary reasons, with reduced rations and much of the time under-
ground, one cannot but wonder how it is that our generous people
do not yeld more than a passing thought to the scientist and to the
science that made his invisible weapons. And the prevention of dis-

1 Authority to publish has been granted by the Board of Publications,
Office of the Surgeon General.

317

318 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

ease was only one of the activities of our medical sciences against
the German. The part played by surgery, by medicine itself and by
the skilful evacuation of wounded has yet to be written. We are
apt to confuse our priestly mission of ministering to the suffering
of friend and foe alike, with that militant one of applying our mind
and body, our art and science, to win the war. A man who carries
arms is not less a soldier for observing the laws of humanity in his
treatment of a fallen enemy. A doctor who keeps healthy men at
the front and salvages the sick and wounded to fight again is no
less a soldier because he carries no rifle.

We went into this war with one officer per 27,000 men, one man
in the division, charged exclusively with the responsibilities of the
sanitation of our great fighting unit. He was only an advisory officer
by law. He was a high ranking officer because rank is the current
coin in the Army, and rank has its privileges. But he had strong
support because he represented a reasonable and powerful science.
Most commanders, jealous of the health of their troops, gave him
every possible chance to work out his salvation. Some even gave
him authority to command in their name. But the weak point lay
in his lack of a force of trained men of his own to carry out his own
suggestions. It was too much to require of fighting troops that they
should lay aside their arms, often after exhaustive physical strain
in the line, to build privies, delousers, etc. It was their duty to do
their part but not all the creative work.

Again, this officer was expected to transform a new area into
which the division would be suddenly thrust, into a sanitarily safe
living-place for these soldiers and to do it right away. It was only
when he failed to do it that he realized the full prominence of his
position, for he was charged with the responsibility to his military
chief for failure. This sudden assumption of responsibility for the
sanitation of an unprepared area, or an area soiled by preceding
troops, was another weak point. It is true that all medical officers
of whatever condition are charged by law with sustaining the sani-
tary excellence of the command to which they are assigned, but " in
addition to their other duties," and these other duties become in
active periods paramount. In practice every one tried to keep things

THE GREAT WAR IN THE ZONE OF THE ARMY. 319

sanitary, but only one, the division sanitary inspector, had nothing
else to do and all he could do was request and advise.

But the Great War promptly threw the spotlight on these weak
points and to its everlasting credit the Army began to evolve a plan
to remedy them most skilfully. First, labor battalions began to be
employed to prepare areas for troop cantonment, as well as to " clean
up " after a battle. Two sanitary squads of one officer and 26 men
were assigned exclusively to certain divisions for sanitary duty.
These men were plumbers, pipe-fitters, carpenters, masons, tinsmiths
and artisans of all sorts needed for carrying into effect sanitary
plans. Engineer troops, quartermaster troops and even, when nec-
essary, soldiers of the line, were to be called upon to help in work
clearly beyond the scope and endurance of these two squads. But
of all the promising steps taken in these reforms the plan of provid-
ing a permanent sanitary organization for Army areas was the best.
Instead of a confused and hurried attempt every time a division was
moved, to sanitate its new area by its own efforts at enormous ex-
pense, such efforts to be repeated more or less by the next division
when in a week or so the first moved on, the Army, through its chief
sanitary inspector, was to divide up its area into three sections, each
of which was to be divided into from 8 to 12 subsections. These
sections were plotted out for topographical reasons, and with an eye
to highways, and had no reference to the divisions which might oc-
cupy them. The personnel was permanent and did not leave when
divisions occupying their section left. They were Army units, under
the military zone commander. They prepared the sanitary appli-
ances for the incoming troops and helped keep them in repair as
well as keeping a watchful eye over them. They exercised no au-
thority over the troops using them, — they were there to help the
troops keep decent and comfortable, as well as to prevent disease.
The reason why they were Army units is that an Army is the least
mobile unit in regard to its area. The divisions and corps were con-
stantly changing. Each section had its chief in a town where he was
to run a small school of sanitation and was to have a sanitary squad
to build sanitary appliances for demonstration as well as to aid in
repairing them in the sections which were being actually used. The
sanitary squad at Langres was early stationed at the dump and from

320 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

its culls recovered enough lumber, metal, etc., to build loi flyproof
latrines, 67 urinals, 23 grease traps and 16 incinerators in a month.
This was due to their commanding officer, Captain Starr A. Moul-
ton, and shows how i officer and 24 men can provide a whole area
occupied by the equivalent of a division with sanitary appliances and
sanitate an area out of nothing, remaining on good terms with ever}^-
body.

The sanitary section of an Army area should be traversible in
all directions by automobile and is thus constantly under the eye of
the young sanitary officer.

The subsection is manned by a non-commissioned officer and
two or three men, responsible to the section chief. They are his
eyes — they give no orders, but are distinguished by an armlet and
know everything that goes on in every nook and cranny of their ter-
ritory, traversible on foot or horseback. They make themselves
popular with shifting battalions who are only too glad of a helping
hand and they keep the sanitary section chief thoroughly in touch
with his problems.

The man who will make area sanitation a success must be a man
who lives permanently in that area and who has this and nothing
else to think of ; who has planned and built its appliances ; who will
fight for them ; who will enthuse the men personally in their proper
use ; stimulate competition ; prevent waste ; prevent heady unit com-
manders from blocking proper sanitary functions in his section ; who
will make himself a helpful and not a carping critic ; who has a
remedy and can do things as well as talk ; who is not unreasonable
but just a normal, sane, well-bred, well-educated gentleman with a
keen sense of humor and an indomitable will ; with his mental proc-
esses backed by scientific facts, his actions by common sense. In
short, a man with enough of the prose of material results and the
poetry of an invincible spirit to make a plan succeed, even if that
plan is not a perfect one.

That is the sort of a man we want.

This leaves the division sanitary inspector with his two squads
free to run the machine and not to make and run it at one and the
same time. He administers the area and if he is at the front he has
some time to devote to his legitimate duty, that of providing trench

THE GREAT WAR IN THE ZONE OF THE ARMY. 321

sanitation, or the sanitation of an open battlefield, as the case might
be. He has a divisional water-supply officer on his staff. He keeps
a close watch on diseases which might become epidemic in his com-
mand and in epidemics works in cooperation with the chief sanitary
officer of the Army. He is in close touch with Army laboratories
and calls on them for special work.

Epidemiology,

The classification of communicable diseases and the administpa-
tion of the epidemiologic service is based on an exposition of this
subject made before the Arm,y Sanitary School in Langres by
Colonel Hans Zinnser, chief sanitary officer of the Second Army.
This officer was W'Orking out these plans when the war came to a
close and to him w'e are indebted for a prominent part in a great
reform.

The Army chief sanitary officer should be an epidemiologist and
laboratory man, as well as an administrator. He should keep spot
maps showing the prevalence and source of infectious disease. In
this he is largely dependent upon his section chiefs and divisional
sanitary officers. He should have a stationary laboratory capable
of becoming mobile at headquarters of the field Army, with two
traveling motorized laboratories to dispatch to special work. I am
not in accord with Colonel Zinnser in depriving a division of its
mobile laboratory, but these two motor laboratories can be kept busy
reinforcing division laboratories and checking their work. In-
fectious diseases are roughly divided into, (i) those disseminated
by the respiratory tract : Pneumonia, influenza, measles, scarlet
fever, meningitis, mumps, smallpox and chickenpox. (2) Those dis-
seminated by the intestinal tract : Typhoid and its congeners, the
dysenteries, simple diarrhea, etc. (3) Insect-borne diseases: trench
fever, typhus, malaria, scabies, etc.

Communicable diseases depend for their extension either on the
susceptibility of the individual or upon means and methods of trans-
mission. For the first, such as meningitis and pneumonia, general
hygienic measures were enforced with a degree of insistence and
elaborateness of detail dependent upon the circumstances and serious-

PROC. AMER. PHIL. SOC, VOL. LVIII, U, AUG. 12, I9I9,

322 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

ness of the local problem. The general measures employed to limit
the spread of diseases disseminated by the respiratory system are:

1. Hygiene of Living Quarters: (a-) A minimum of 40 square
feet per man. If less, the suspension of shelter halves between
bunks, (b) Head and foot sleeping in bunks side by side, (c)
The spread of a command in tents in warm weather, (d) Ventila-
tion by perflation once a day and once at night between taps and
reveille, (e) Airing of blankets and bedding several hours on pro-
pitious days. (/) The invariable punishment of promiscuous spit-
ting, (g) The sweeping of floors once a day after sprinkling.

2. Hygiene of Mess Kits: (a) Immersion after use in boiling
water. Men were lined up after meals and made to dip their eating
utensils in a ten-gallon, or larger, container of boiling soapsuds kept
at ebullition by a hot fire, (b) Rinsing of same in flowing water.
(c) Air-drying of same. No common towels permitted.

3. Inspections of Command: (a) Men lined up and inspected
twice a day for head colds, cough and inflamed eyes. (&) Tempera-
ture taking of men not feeling well, (c) Segregations of all coughs,
colds and red eyes in a separate barracks or curtained-of¥ portion of
a common barrack, (d) Evacuation of all with fever to a hospital.

4. Inspection of Men's Personal Equipment: The verification of
possession by each man of a raincoat, an overcoat, two pairs of
shoes, three blankets and three pairs of socks.

5. Provision for a Drying Room for Clothing — a vital necessity
in France.

6. Avoidance of Excessive Labor. Drills and fatigue duty should
be cut to a minimum to avoid exhaustion, fatigue being a well-known
condition of inviting infection by lowering individual resistance.

The majority of these measures could only be enforced in rest
areas. In combat, naturally, no sensible man could consider the em-
ployment of more than a tithe of these protective measures.

The general measures employed to prevent dissemination of dis-
eases by the intestinal tract are :
I. Care of Latrines:

{a) Fly-proofing (area sanitation).

(&) Disinfection of interior of receptacles and latrine seats
during an epidemic (divisional sanitary squads).

THE GREAT WAR IN THE ZONE OF THE ARMY. 323

(c) Ablutioji benches or other provision for handwashing after

use of latrines (area sanitation).

(d) Prohibition of towels.

2. Sinitation of Kitchens:

{a) Neatness and cleanliness of kitchen itself and utensils.

{h) Prohibition against use of uncooked food.

{c) Enforced hand cleanliness of kitchen crew.

((/) Medical inspection of kitchen personnel and all handlers

of food and removal of all sick or sickening, as well as

carriers.

3. Surveillance of Drinking Water:

(a) Proper chlorinization of water and a bacteriological ex-
amination of disinfected water from time to time as a
check thereon. (Division sanitary squads and Army
mobile laboratory.)

(&) Careful supervision of water sources and guard therefor
(unit commander, division sanitary inspector, sanitary
section chief, engineer water supply details. Army mobile
laboratory).

(c) Supervision over water carts (division sanitary personnel).

4. Disposal of Garbage and Offal:

(a) Soakage pits for liquid waste.

(b) Incineration, or sale, or other, disposal to civilians of solid

garbage.

(c) Prompt burial of carcasses.

(d) Removal and proper disposition of manure (burning was

impossible).

5. Prompt Notification of Military Chief zvhen in the Presence of

Epidemic Conditions.

It will not be amiss to consider briefly the Special Procedures for
Special Diseases.

Typhus and trench fever require careful destruction of lice, the
carriers of these diseases.

Typhoid and congeners were practically eliminated from all
Armies by vaccination. The few persons who acquired typhoid
fever were especially susceptible and did not receive a large enough
preventive dose of the vaccine ; were recipients of an overwhelming

324 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

dose of virulent organisms, or, through carelessness, were not
properly inoculated.

Influenza demanded prompt evacuation in separate ambulances,
segregation, and masking.

Smallpox called for strict quarantine and vaccination.

Trench foot: Issuance of dry socks at dinnertime, when in the
line, or in very muddy terrain elsewhere ; powdering of feet with
a combination of camphor, salicylic acid and talcum ; mutual massage
of feet, after removing shoes, in groups of two when in the line ;
precluding of foot and leg covering which constricts the circulation ;
drying of shoes at every favorable opportunity.

Venereal disease: Propaganda, prophylactic stations and sur--
veillance of civilians. Applicable in cantonments.

Tetanus: Universal prevention dose of antitoxin in all wounds
and in trench foot.

Boils and other infective inflammations of the deeper layers of
the skin: Inspection of men at baths for scabies and other sources of
irritation of the skin, and segregation under treatment. These skin
affections, plus lousiness which often caused trench fever, were
responsible, it is said, for a full 90 per cent, of all incapacitation
from disease in a British Army.

It is extremely interesting to review the experience of the French
and British during the long months of war preceding our entrance
into these scenes, a time when experience, that bitter teacher, was
elaborating for us the saving knowledge that Briton and Gaul turned
over to us for our protection, without a thought inspired by selfish
pride, or the slightest attempt to cloak their own shortcomings, that
by these truths our men should live to fight.

When the war broke upon Europe, France, we were told, had no
compulsory vaccination against typhoid. The result was that for
the first six months, due, in part, to a life-and-death struggle in
which only military resistance could be considered, typhoid ran
rampant. No one knows how many cases developed in the French
Army ; one French officer named a figure that is so high as to be
almost incredible. Compulsory vaccination wiped out the disease
from this Army in the succeeding months. In the British Army 90
per cent, of her men were inoculated against this disease with the

THE GREAT WAR IN THE ZONE OF THE ARMY. 325

result that there were only 7,700 cases of a strength of 2,000,000,
throughout a period of 40 months of campaigns. Compare this with
their record in the South African War when, of a force of 100,000
men, and very few vaccinations, there were 56,000 cases and 8,000
deaths, and with the record of the first American Army of one mil-
lion men, which according to its Chief Surgeon, Colonel Alex.
Stark, had only 17 cases of typhoid and 10 of para-typhoid to Janu-
ary I, 1919. With tetanus in 1914 there were 32 cases per 1,000
wounded; in 1917, a small fraction of i per cent., the favorable
change being due to strict orders that all wounded should be given a
preventive dose of the antitoxin.

In the first British Army in 191 5-16, 3,311 cases of trench foot
were evacuated; in 1916-17, only 395. In the fourth French Army
in 1916-17, 2,861 were treated for trench foot. As a result of the
clear enunciation, in time, of the preventive measures to be taken,
practically no cases developed in the American Army.

Availing ourselves of their knowledge, plus the proof by Strong,
of our Army, on suggestive evidence presented by the British that
trench fever was carried by lice, plus the experience of both French
and British that the destruction of insect life in clothing would
enormously reduce morbidity, the American Army, with its own
enlightened medical corps, reinforced by the frank testimony of the
Allied troops, was able to produce a record for preventive medicine
in war the like of which the world has never before seen.

All of these precautions above noted being in force and all of the
special sanitary appliances for this war in function, there remains a
general statement to be made : The chief cause of sick wastage,
when sanitary science has already yielded its utmost, is reduced
vitality from overwork, little sleep, poor food and constant cold.
Many of these undesirable conditions are inseparable from certain
military necessities, but it is an ill-conceived idea of a commanding
officer that troops to be efficient should do arduous service and at the
same time be deprived of possible comforts and prime necessities
" in order to harden them," such as sleep. In such a war as the one
through which we have just passed, this is not the time nor the
place to weed out the men with slight imperfections capable of
gradual remedy. The wise commanding officer will endeavor, where

326 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

the whole male fighting population of his nation is requisitioned for
war, to keep all of his men and not only those who can rise above
his own peculiar selective processes. It is vicious to get men up
before sunrise, give them their breakfast in the dark and start them
out in the cold morning mist, chilled, unfed and with clothing still
damp from the rain of the day before to do a hard day's work,
again terminating in wet and cold. It is still worse to gather such
men after supper with clothes still wet for lectures running far into
the night. This " intensive training " may break down a man before
he has the opportunity to deliver a blow, not only the palpably weak,
but, to our surprise, the apparently strong. Our Field Regulations
preach against this, our best officers of the old Army condemn it,
and it is only mentioned to deter a well-meaning but inexperienced
newcomer from a dangerous point of view. The best officer knows
how to care for and feed his men and the best troops are faithful
witnesses of the effect of the care bestowed upon them.

The Sanitation of Trench Warfare.
In the Trenches.

Sanitation is reduced to its simplest terms in the trenches. It
has to be. Men are there to fight. In the brief week or ten days
they are there before their relief, lights, smoke, evacuation, building,
are taboo. Moreover every nerve is taut to get the best of the enemy
and there is little time for things of immeasurably less importance,
however simple and enticing the procedure may seem. Sanitary
science is reduced to its barest necessities and then greatly reduced
from that point.

Latrines: These are either buckets under a flyproof seat or fly-
proof boxes over a pit in a blind offshoot from a communicating
trench. There is a camouflaged head cover and there should be a
can or trough leading to a soakage pit for urine. Pit latrines should
be deep and buckets emptied in adjacent shell holes at night, as is
garbage from food brought up to the men.

Water: This, in well consolidated positions, is pumped from a
chlorinating source. It may be stored in cement water tanks or
carried up from water carts in the rear. A trick of our soldiers was

THE GREAT WAR IN THE ZONE OF THE ARMY. 327

to fill a large number of canteens in the rear, string them on a pole
and thus carry them up. Water may be obtained from wells in dug-
outs and, after filling the 40 gallon canvas Lyster bag, and disin-
fecting with hypochlorite powder, be doled out to the men. Best of
all it may be piped up, but this could only be depended upon in very
quiet sectors, as bombardment quickly broke down such a system.

Food: This is usually sent up once a day from kitchens in a com-
municating trench somewhere in the rear and is kept hot by storing
food for 16 men in a marmite, or thermos can. There is a special
danger in these cans, when the lid is left ofif until the temperature
falls to blood heat and the lid then replaced, from fermentation,
which may cause serious food poisoning.

Clothing: Dry socks should be supplied with the evening ration
when men are standing in wet trenches. The necessity of standing
in muddy water is the only excuse for the rubber boot, otherwise
an abomination.

Drainage of trenches: Well-constructed trenches are floored with
a duckboard path and under-drained by a gutter with occasional
sumpholes which are emptied by suction pumps in favored positions.

Drying rooms: Every trench position of any permanence should
provide these rooms, necessarily in a dug-out, and a brazier coke
stove or two will be sufficent to accomplish satisfactory results.

Heating of dug-outs is by braziers. Ventilation by stove pipes
with closing apparatus in case of a gas attack. Light by candles,
kerosene lamps, acetylene or electricity, but oil lamps require oil to
be brought up, a thing possible only in times of inactivity, acetylene
is practically confined to medical dug-outs and electricity is only
available in well-consolidated positions, long occupied.

The Sanitary Duties of the Battalion Medical Officer: The sacred
relations established between this lone representative of our great
profession and the soldier in his hour of trial is readily seen in the
influence this officer wields for the prevention of disease. It is a
personal relation. He talks to no audience and he indulges in no
flights of fancy. A good battalion surgeon knows all the men ; he is
over in the trenches ; he knows when the breaking strain has been
reached in Jones and he laughs Smith out of his " shell-shock."
He invents a thousand ways to make the men more comfortable,

328 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

supervises the simple sanitary arrangements of his sector and catches
the first signs of a communicable disease in his command. The
trench is no place to keep a sick man, but it takes a real doctor to
determine whether he is sick, worn out or ailing. All communicable
diseases are immediately evacuated if possible, bvit many an incipient
" shell shock " or case of homesickness has been exorcised by a keen
young medical officer in the trenches. One of his great duties is to
" keep men on the line " and the man who aids in doing this has a
friend in the battalion commander. There is no one for whom the
men have greater respect, in whom they have greater faith, for
whom they entertain a more loyal affection than " their doctor "
unless it be their own personal commander, and this is as it should
be. This picture has been attempted to show how strong a represen-
tation sanitary science has at the elbow of the fighting men and that
it is this human touch, humanely administered, that makes the
practice of hygiene possible among our troops. The most that we
ask, that we require, of our men under fire is to be as clean as pos-
sible and decent always.

The Sanitation of Open Warfare.

Here sanitation is reduced to a mere shred, but the little that can
be done often makes the difference between ultimate victory or
defeat.

The essentials are :

1. Hot food,

2. Disposal of feces and urine,

3. Safe water,

4. Burial of bodies and carcasses,

5. A reasonable amount of rest.

Food is supplied by wheeled kitchens, a great boon to soldiers.
Hot liquids are a tremendous stimulus to exhausted soldiers and the
measure of good done by our Red Cross, Salvation Army, Knights
of Columbus, Y. M. C. A. and other patriotic societies to soldiers
going up to and returning from the battlefield has left a grateful
memory in the soldier.

Pioneer battalions are absolutely necessary in these modem wars.
They should follow the troops, bury animals, bodies, etc., burn refuse

THE GREAT WAR IN THE ZONE OF THE ARMY. . 329

or bury it, dig latrines and are at times indispensable in many ways.
Unburied bodies and carcasses rapidly breed myriads of flies,
streams polluted by dead are contaminated with their intestinal bac-
teria, the thirsty and tired men drink the water on the line, sicken
and their germs of disease raised by passage to high virulence, and
teeming in uncovered feces, are carried back by these flies to spread
like a wave of cloud gas over the food of advancing reinforcements.

Latrines are mere straddle trenches a foot wide and a foot deep.
They must be scrupulously covered by some one.

The question of safe water is settled by chlorinization. If the
Lyster bag cannot be used, an ingenious American device is the
dosing of the contents of some sergeant's canteen with a full tube
of the powder, a plan suggested by Zinnser. The men can be ad-
vised that all they have to do to get safe water from the battlefield is
to fill their canteens from any possible, unpoisoned source and put
"a little," about a teaspoonful, from the sergeant's canteen into
their own. In about 20 minutes they can drink reasonably safe
water. No one will ever know the thousands of American lives
saved by the bag devised by Colonel Lyster, of our medical corps,
in the Great War. Vaccination against typhoid is not everything.

The Back Area in Trench Warfare (Position Warfare).

This is itself a subject for another paper. Only essential sani-
tary arrangements necessary to complement trench life will be men-
tioned.

Bathing and Delousing.

This should be at least under the supervision of the medical de-
partment for it is a powerful weapon against disease, as has been
seen.

All men returning from a tour of duty in the trenches are put
through the bath and delouser. The underlying principle of im-
portance is not the bath ; it is the killing of vermin and itch mites on
clothing by heat. There should be two per division with an output
of one thousand a day. There should be a medical department floor-
walker to detect all skin diseases.

330 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

Types of D do users.

The Fodden-Thresh single-cylinder and double-cylinder steam
sterilizer for disinfection under pressure.
The French Autoclave of like type.
The Jacobs hot-air chamber and Orr Canadian hot-air chamber.

Types of Bathing and Delousing Establishments. French Army.

This is known as the personal hygiene section and it is run by the
French medical department. Upon its general principles our Amer-
ican baths were run.

Personnel : One non-commissioned officer, a sergeant, two cor-
porals, three mechanics, one barber and lo men.

Bathers were admitted in groups of 40. It is a test unit. It con-
sists of one store tent, Bessonnneau, for clean clothing, one tortoise
tent for storing soiled clothing, one tortoise tent for barber and
chiropodist, one tortoise tent for valuables, and the bathhouse. The
men on first entering the bath place their valuable in an individual
bag bearing their bathing number. This bag goes into a basket
which is locked and placed in the tortoise tent until they return
from the bath. They now receive their bath number on a wooden
tag, go to Room A of the bath house and look for their bath number
painted on latticed seats in the center of the room. Here they un-
dress, throw their underclothes on the floor, and stuff their outer
clothing and shirt into a mesh bag with a tag bearing their number.
This goes into the sterilizer. They keep their hats, their gas masks
and their shoes to carry into the bath room. In the bath room these
are placed in a cubby hole numbered with their bathing number.

Upon entering the undressing room, an attendant notes the size
of the man for replacement of underclothing. There are only three
sizes ; one for short, one for medium and one for tall men.

A reserve of 10,000 suits of clothing is kept in the store tent and
includes the following: Shirts, drawers, undershirts, stockings,
towels, and handkerchiefs.

Each disinfector holds 40 kits, each composed as follows:
Breeches, blouse, overcoat and pair puttees. The autoclave is a very
handsome one and it has a steam ejector for creating a vacuum and
drying coils for driving oft' steam after sterilization.

THE GREAT WAR IN THE ZONE OF THE ARMY. 331

There is a hot-water tank on top of the autoclave which feeds
the showers and gives a pressure up to one kilo by manometer.
There are twenty shower heads. Water is heated to 38° C. and
kept at that temperature by a cold-water mixer. A bell rings when
45° is reached.

The autoclave is drawn by two horses and the weight is 2,000
kilos (4,000 pounds). To set up this portable bather and delouser
it is necessary to be near a watercourse giving at least one liter per
second, and to have space up to at least 5,000 square meters. It
takes two hours to set it up and eight hours to make ready for the
first bath.

These baths require about 12 liters of water per man. Men are
allowed, normally, ten minutes for actual bathing and it requires a
minimum of twenty minutes to hand them their deloused outer
clothing. Soap and towels are furnished by the bathing estabhsh-
ment and the dirty, but deloused, underclothing is stored in a tor-
toise tent to be trucked to the laundry after bathing hours if distant,
an equal amount being returned clean from their store in exchange.

The English variation from this bathing establishment lay, as a
rule, in the delouser itself. Toward the end of trench warfare dis-
insectization was accomplished by dry hot-air chambers. The most
elaborate type of these is the Jacobs and the easiest to construct and
run is the Canadian Orr which will be described. It can be built in
a few days by four pioneers. It is essentially a hot-air chamber
with double iron walls and the space between insulated by earth. It
is heated by under-burning coke braziers and ventilated by holes into
which fit wooden plugs, removable, to regulate the temperature.
The loaded chamber attains 65° C. in ten minutes.

We were experimenting at Langres on a combination of bathing
house, incinerator and laundry all in one, when the war came to a
close.

Portable Hot-Air Disinfestor.

The Pleasants hot-air disinfestor and portable shower for com-
panies or battalions. Mounted on Army general service wagon ; has
a limited application for service nearer the front.

The Serbian barrel will not be detailed as it is far from satis-
factory and when used is a makeshift. The same principle on

332 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

wheels, devised by the French with a portable bath was however a
most ingenious application.

Men should be bathed at least once every ten days and deloused
at least once a month. Hence battalion organizations were each fur-
nished, in our Army, with a Charles Le Blanc heater and 8 head
showers, in addition to divisional baths.

Every rest area should have its cobbler shops, its tailor shops,
and its barbers. They are, as are baths and delousers, run by Class
B men, generally. Dirty, ragged men lose respect for themselves.
" Smartness " has a determining influence on the soldier at the front.

Billets. — Neatness and cleanliness are the keys to their success-
ful employment. Our men did wonders in tidying up stables, barns
and outhouses in France. Scabies in a billet required " delousing "
of all blankets and a working quarantine, if feasible, of occupants.
The English greatly feared scabies and an English chief surgeon
insisted that much of it was spread on latrine seats.

Kitchens.

Time does not permit their discussion, which offers much of in-
terest. The main desiderata are :

1. Cleanliness and simplicity of equipment. A clean kitchen in-
spires a good appetite. Everything that shows should be immacu-
late and always in its place. Men should be allowed to help them-
selves from a common receptacle and be told to never take more
than they want, but always come back for more if they were still
hungry. The English said that they saved 10,000 pounds in that
way in a year at one large camp.

2. All grease should be salvaged and sullage water deprived of
its grease by traps. The water then percolated into the earth with-
out clogging its pores after being led into a filtering cess-pool or
" soakage pit."

3. All scraps of food, especially bread and meat, were saved and
made into tasty dishes at subsequent meals. This means not only
the teaching of cooking to cooks, but of •" catering" to mess ser-
geants. A full garbage pail became a poor recommendation for any
unit in the A. E. F.

I

THE GREAT WAR IN THE ZONE OF THE ARMY. 333

4. A brick or cement floor was most desirable in a kitchen and
whitewash and tar hberally appHed at least helped as an appetizer.

5. Cooks should be cotnpelled to wear white caps and aprons in
the rest area. Dirty cooks mean dirty food — and generally a " full
garbage pail."

6. Menus should be varied and stews should not figure on more
than 3 days in the week. Meat should be stored in meat safes and
" quarters " should be hung, haunch down, and covered with burlap
or muslin bags.

We labored with the generally acknowledged defect that the
culinary art in the United States had declined. " Canned stuff " was
offered as the solution, but with all its virtues it can never replace
art. Besides it is at least debatable whether a preponderance of
canned foods is more desirable than a fair representation of the
dishes "that mother used to turn out." But the great, the crying,
sin was the capacity for enormous waste of food material " not
equalled by any country on earth," an Americanism we carried
with us to Europe. Our Army manfully struggled with this gen-
erous failing and certainly improved the situation, but it was only
by persistence that real saving was accomplished.

Latrines.
The ideal latrine for troops under ordinary conditions was the
deep pit and flyproof box seat. Even the English acknowledged
this, but it should be from 10 to 12 feet deep. The seats should be
cleaned inside and outside daily and blackening of the pit with lamp
black and crude oil is desirable. The trouble was we could not get
crude oil. Burning out of pits was also impossible ; there was no
mineral oil " to waste." Hence the American Army devised their
own latrine, a cross between the American and a French commodity.
It was simply a deep pit rimmed at the surface by two-by-fours
with a platform covering it in. Every two boards were nailed down.
The intermediate one was loose and could be removed by a handle
or a pole fixed perpendicularly in the center. Thus there was no
seat, the simple concession of the French to economy. It was fly-
proof when the loose boards were replaced. The British, on the
other hand, provided what was for most cantonments, in France, the

334 ASHFORD— APPLICATION OF SANITARY SCIENCE TO

most feasible and certainly the most sanitary arrangement ; the
bucket latrine and incinerator. Pits could not be dug in many places
ten feet deep without striking water or rock. Besides, the French
bitterly opposed, and with reason, the digging of pits for feces. The
incinerators built by the British usually served 800 men and were
models of practical common sense and good sanitation. They usu-
ally presented a hopper leading down onto a hot plate for drying of
excreta before dumping the contents upon the fire and this served
for solid garbage as well. Often a few coils of pipe were led around
the fire box and heated water for an ablution bench nearby. Every-
thing about the incinerator could be built with stones, old bricks, a
little cement, old rails, old bits of iron sheeting, etc., and in an in-
credibly short time. The fuel was ostensibly at the rate of 100
pounds a day. In reality 2 pounds of coke was used to start the fire
and the thing was kept at full blast eternally, like the altar fires of
the Vestal Virgins, by camp trash. This kept the camp clean be-
cause the incinerator was served, as were these altar fires, day and
night, by a husky soldier looking for a fuel economy record ; the
camp was carefully gleaned daily and gleanings deposited by that
incinerator. The latrine was nearby, a long shack with usually a
cement floor, flyproof seats and, under each seat, a bucket. In front
of the seats was a urinal, a metal water way leading down into a
soakage pit. A little cresol stood in each bucket and before incin-
erating the contents a good 20 per cent, of saw dust was added to
insure incineration. The buckets were emptied twice a day and
smeared with crude oil, which they secured in some manner, before
replacing. I never detected the odor of feces in any British camp I
visited. Critical eyes may challenge this statement, but no one will
testify that the bucket latrine was a failure if they saw enough of
them. The English became so bold that they would place their
bucket latrines near their kitchens and mess halls. This, with the
proximity to ablution benches, where hot water was obtainable for
the morning shave, very much simplified daily routine and stimu-
lated cleanliness and an orderly observation of sanitary principles.
The answer of the British to an objecting theorist would invari-
ably be :

THE GREAT WAR IN THE ZONE OF THE ARMY. 335

"If you have clean latrines and burn your feces you need not
fear flies."

In concluding this long paper we are compelled to the reflection
that, after all, the great point is that it was not only mechanical
sanitary appliances that we were after but the sensitizing of our
soldier's conscience by appealing to his sense of decency. The im-
mediate commander of troops must inculcate in his men, in addition
to a strict regard for military discipline, a decent regard for the
ornate, and as a result thereof, for the clean and neat. It should
be deliberately taught men that it is not enough to have merely the
" rough stufif," the bare necessities ; that civilized men are expected
to improve their natural condition ; that they should keep them-
selves and their surroundings clean if they expect to keep well ; that
they should keep their initiative, their ambition, their imagination
active ; that to make their barracks, their reading rooms, their camps
clean and attractive to the eye is to preserve those illusions of life
which are necessary to make men contented and happy.

Nothing illustrates this spirit better than the following incident :
When the First Division was in a training area under the tutelage
of the famous and gallant 47th Division of Chasseurs, the disposal
of manure piles in the center of these tiny hamlets was a sore trial
to us American officers. The French had some towns for billeting
and we had others. One day in traveling over the French area with
Colonel Cultin, their division surgeon, I noticed that none of his
towns possessed the protruding manure pile beside the neat little
house in the main street. On inquiring how it was that no amount
of coaxing or threatening by French civil authorities had prevented
their reappearance in our towns, he replied : " We removed the piles
to the site the owner selected himself ovit of town, and planted
flowers where they used to be — and no good French peasant will
smother growing flowers with manure piles."

Nor will an American soldier spit on a clean floor, nor befoul a
neat home, which he has been led to beautify by his own efforts.

It may seem trite to mention that neither laws in a civil com-
munity nor regulations in the Army are alone potent to sanitate
one or the other. There is a difference between securing the pre-
cision of a machine in a military organization and considering a

336 ASHFORD— SANITARY SCIENCE AND THE GREAT WAR.

man simply as a machine. Germany had the most perfect machine
in the world and she failed. She persistently bid us to treat our
men as machines and she herself failed. It is the spirit of a glorious
France that bespeaks our own American heart and the words are
from the favorite work of a brave and humane officer, idolized by
his men, Major General Goubeau, given me personally as his divi-
sion watched the road to Metz :

" Le soldat n'est point une machine. Avant d'etre une unite de
groupe, il est une personne."

Surgeon-General's Office,

War Department, Washington, D. C.
April, 1919.

MAGELLANIC PREMIUM

Founded in 1786 by John Hyacinth de Magellan, of London

I919

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful Ki>io\vledge

ANNOUNCES THAT IN

DECEMBER, 1919

IT Wr:.L AWARD ITS

MAGELLANIC GOLD MEDAL

TO THE AUTHOR OF THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING TO NAVIGATION, ASTRONOMY, OR NATURAL PHILOSOPHY (mERE
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 1919, deliver free of postage
or other charges, his discovery, invention or improvement, addressed to the
President of the American Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance by some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the sam'^ motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author.

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publicly
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as may be after the adjudication, either in a separate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

kW correspondence in relation hereto should be addressed

To THE Secretaries of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A,

Proceedings

OF THE

American PhilosopMcal Society

Subscription — Three Dollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

Price, One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

For Promoting Useful Knowledge

iV^-i.- Series, Vol. XXII,
Fart III, 4to, 44 pages.
{Lately Published)
Tertiary Vertebrate Faunas of the North Coalinga Region of Cali-
fornia. A Contribution to the Study of the Palgeontologic
Correlation in the Great Basin and Pacific Coast
Provinces. By John C. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California.

Subscription— Five Dollars per Volume
Separate parts are not sold

i^ddress The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A.

i^^-t PROCEEDINGS

OF THE

American PhilosophicaiSoGiety

HELD AT PHILADELPHIA

FOR PROMOTING USEFUL KNOWLEDGE
Vol. LVIII. 1919. No. 6.

CONTENTS

PAGE

Star Clusters and Their Contribution to Knowledge of the Universe. By

Harlow Shapley . . . , . . . , -337

Hydration and Growth. By D. T. MacDougal . . . . -346

Some Considerations on the Ballistics of a Gun of Seventy-five-mile Range.

By Arthur Gordon Webster . . . . . . '373

On a New Method in Exterior Ballistics. Bv Arthur Gordon Web-
ster and Mildred Allen ........ 382

Characters and Restoration of the Sauropod Genus Camarasaurus Cope.

By Henry Fairfield Osborn and Charles Craig Mook . 386

PHILADELPHIA

THE AMERICAN PHILOSOPHICAL SOCIETY

104 South Fifth Street

1919

The American Philosophical Society

announces that an award will be made in the year 1921 of the

HENRY M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

The Control of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent, of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after the
award.

Attention is called to the following regulations governing the award
of the Prize:

Competitors for the prize shall affix to their essays some motto or name
(not the proper name of the author, however) and when the essay is forwarded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in .pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judged, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essays shall be in their author, subject to
the right of the Society to publish the crowned essay in its " Transactions "
or " Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,

David Jayne Hill,

Simeon E. Baldwin,

John Cadwalader,

W. W. Keen,

William B. Scott,
President ex-officio.

Committee.

The essays must be sent, addressed to the
President of the American Philosophical
Society, No. 104 South Fifth Street, Phila-
delphia. U. S. A.

STAR CLUSTERS AND THElR CONTRIBUTION TO
KNOWLEDGE OF THE UNIVERSE.

By HARLOW SHAPLEY.
(Read April 25, 1919.)

Social relationships among stars are nearly as common as among
men and the lower animals. Sidereal bodies completely independent
of all star societies are difficult of conception; for the heritage of
early and ancestral associations, as well as the immediate environ-
ment, influences the present behavior and the destiny of stars.
Planetary systems, binaries, groups of three or four nearly equal
bodies are thought to be very common — almost universal, it may
be ; and the assemblage of stars of all kinds by the tens, hundreds,
and thousands into physically organized clusters now appears to be
a property of fundamental significance in stellar investigations.

In considering the aspect of clustering among the fixed stars we
see a gradual progression from the largest and most scattered con-
stellations to such rich and highly concentrated stellar groups as the
globular clusters. Although the constellations were outlined for
the most part in prehistoric times and have been used in myths and
astrology persistently and universally throughout thousands of years,
in general they do not represent definite physical organizations that
exclude the stars of neighboring groups; and frequently even the
legendary relationships of the stars in the most anciently known con-
stellations are traced with difficulty.

There is, however, among the varied groupings, an easy transi-
tion from widely-scattered Ophiuchus and Camelopardalis to Orion,
Scorpio, and the Great Bear; and in recent years we have found
that the most conspicuous stars of these three last named constella-
tions actually form physical stellar systems. The stars of each have
motion, colors, and distances much in common, and in each case
they have evolved, no doubt, from an origin common in space and
in time. From Orion we readily trace the progression in clustering

PROC. AMER. PHIL. SOC, VOL. LVIII, V, OCT. 21, I919. 337

338 SHAPLEY— STAR CLUSTERS.

to the Hyades — a more compact and more definitely circumscribed
dynamical system — and then to the Pleiades, to Prsesepe, to the
double cluster in Perseus and similar faint loose clusters of the
Milky Way; thence, by way of Messier ii and Messier 22, we pro-
ceed by nearly equal steps to the typical globular systems exemplified
in the great Hercules cluster, Messier 13.

Although we may justly restrict the term " star cluster " to physi-
cal systems — that is, to groups which have the characteristics of
distinct dynamical organization — it is clear that the svibdivisions of
the long sequence of groups from Orion to the Hercules cluster
must necessarily be vaguely defined. For convenience we here dis-
tinguish only open and globular clusters, and designate all as open
except those seventy or eighty highly condensed groups whose stars
appear innumerable even with the aid of our biggest telescopes and
most sensitive photographic plates.

Open and globular clusters differ in matters other than richness
and apparent circularity. In average distance from the earth the
globular clusters much excel, in stellar constituency they are more
varied, and we recognize in their wide spatial distribution that from
a dynamical point of view the globular clusters are quite distinct
from the open groups which closely congregate along the middle line
of the Milky Way.

A few of the nearest globular clusters are visible to the unaided
eye as faint hazy objects, and some of them have been in the astro-
nomical records for two or three hundred years. To Messier and
earlier observers they were known only as starless nebulosities, but
Sir William Herschel and his son, with their greater telescopes, par-
tially or completely resolved the brighter clusters into myriads of
distinct stars.

The great telescopes of the present time and the powerful modern
methods of astrophysical investigation have greatly extended our
knowledge of globular clusters, but they have not appreciably added
to the total number. The numbers of known stars and nebulas have
increased enormously with the increase of optical power, but during
the last eighty years less than five new globular clusters have been
added to the original lists compiled by the Herschels. In fact, we
seem to have passed the era of discovery of such systems; the pres-

SHAPLEY— STAR CLUSTERS. 339

ent lists may be considered essentially complete — a condition that
does not prevail for any other important type of celestial object.

It has been shown through the studies made at Mount Wilson
that most of the globular clusters are remotely isolated systems,
neither intermingled with nor closely surrounded by other stars.
They may be treated, indeed, as distinct cosmic units ; and in treat-
ing them as such, we may fulfil the purpose of the present commu-
nication by discussing briefly a single system. In certain details,
there are, to be sure, conspicuous differences from system to system,
but in such matters as size, number of stars, and stellar make-up,
no greater diversity appears.

The Great Cluster in Hercules, No. 13 in Messier's well-known
compilation of 103 bright nebulae and clusters, is the system chosen
for the present illustration. To the unaided eye it is faintly visible
as a hazy star of magnitude 5.8, about two degrees south of Eta
Herculis. The photograph used for this illustration was made by
Professor Ritchey on a plate of medium rapidity with an exposure
of eleven hours ; it records something like 30,000 stellar images
brighter than the twenty-first magnitude. Nearly all of these are
actual members of the cluster and not merely objects of the fore-
ground, projected among the cluster stars.

The distribution, brightness, and colors of many hundreds of
the stars in Messier 13 have been specially studied at Mount Wilson.
Some attention has also been given to spectral types and radial ve-
locities— difficult problems even for large telescopes and powerful
spectrographs because of the extreme faintness of the individual
stars. Space cannot be given here to describe the methods recently
developed for the determination of the distance of Messier 13 and
other globular clusters ; we shall only remark that photometry, as-
trometry, and spectroscopy are all involved, and that Cepheid vari-
able stars play a fundamental role. The adopted value of the par-
allax for the Hercules cluster is o". 00009, with an estimated uncer-
tainty of less than twenty per cent.

Even to those who are accustomed to think of the great depths
of sidereal space, it is difficult to comprehend clearly the remoteness
of the Hercules cluster. Its distance, 3.5 X 10^^ kilometers, is more
than eight thousand times that of the nearest star now known.

340

SHAPLEY— STAR CLUSTERS.

Fig. I. The Globular Cluster, Messier 13.

Light, traveling with its hardly conceivable velocity of 300,000 kilo-
meters in a second of time, requires eight minutes for the passage
from the sun to the earth, but it must travel more than two thou-
sand million times as long to reach us from this globular cluster.
Hence our knowledge of the position and physical characteristics
of Messier 13 does not refer to the system as it now exists. Our

SHAPLEY— STAR CLUSTERS. 341

most recent information dates from the time the hght we now re-
ceive left its remote origin in the ckister, and what has occurred
there during the last 360 centuries is beyond our power of finding
out. On the basis of our knowledge of the probable causes of these
past conditions, we may believe with good reason that the cluster
is now much as it was 36,000 years ago. Such an interval of time
is of small consequence in the life history of a gigantic stellar sys-
tem ; but while these pulses of light have been coming across the
intervening fraction of unending space a thousand human genera-
tions have come and gone ; man has emerged from a vague, unre-
corded past and in fleeting succession all his historically known na-
tional civilizations have slowly evolved, flourished in vaunted per-
manence and supremacy, and quickly relapsed into oblivion or poor
mediocrity. We recall, too, that of all the globular clusters whose
radiation is continually streaming toward the earth, Messier 13 is
one of the very nearest.

With the distance of the cluster knowm, we readily translate an-
gular dimensions, as measured on the surface of the sky, into linear
dimensions. Thvis, by the definition of parallax, the angular value
o". 00009 corresponds in the Hercules cluster to one astronomical
unit — the distance separating sun and earth. By transferring a
linear scale to the cluster, therefore, as in the accompanying illus-
tration (which does not include the outlying stars), we m^ay deter-
mine the separation of the individual stars, the number in a given
volume, and numerous other facts concerning the physical structure
of the system.

It is worthy of emphasis that in determining the distance of the
Hercules cluster we have at the same time derived the distances of
its tens of thousands of stars. The absolute values of these dis-
tances appear to be accurately known ; and certainly the distances
of these stars relative to each other are known with an accuracy
that is unequalled in ordinary parallax work, for in any given
globular cluster the distance of all the stars from the earth may be
accepted as the same, with an error of less than one per cent. We
are thus in a position to make some general investigations of the
inter-relationships of the brightness, colors, numbers, and positions
of stars ; and, because of so little uncertainty in the relative distances

342 SHAPLEY— STAR CLUSTERS.

and because of the great numbers of stars available, we can make
these investigations with more ease and accuracy in a globular clus-
ter than with the bright stars around the sun.

A conception of the great size of a globular cluster may be gained
by indicating on the picture of Messier 13 some of the familiar dis-
tances of the solar neighborhood. The distance separating the near-
est known star (Alpha Centauri) from the sun is indicated by the
short black line near the center. In the cluster, within the corre-
sponding distance from the center, there appear to be thousands of
stars, illustrating how greatly different from the conditions of our
own immediate part of the universe is the concentration of mass
and luminosity in a globular system. The distance to the Hyades,
the well-known group of bright stars in Taurus, is represented on
the photograph by the distance from the center of the cluster to the
star marked H. The diameter of the large circle is about fifteen
minutes of arc on the surface of the sky, corresponding at the dis-
tance of the cluster to ten million astronomical units ; a sphere of
that diameter with the sun at the center would include all the stars
within eighty light-years.

The total angular diameter of the cluster is about thirty-five
minutes of arc, corresponding to twenty-three million astronomical
units, or more than three hundred and fifty light-years.

All the cluster stars shown on the photograph are giants in
actual luminosity. At the cluster's distance of 36,000 light-years,
a star as bright as our sun would be considerably fainter than the
twentieth magnitude and would not appear on this reproduction.
The three stars whose images are enclosed in small circles are pho-
tographically almost exactly one hundred times as bright as the
sun, and the most luminous giants in the cluster exceed a thousand
suns in light emission.

It is a remarkable fact that the brightest of these cluster stars are
red and the fainter ones are blue. The red stars, because of low
surface temperature, emit much less light for a given amount of sur-
face area. To excel in brightness, therefore, their volumes must
be enormously large — in extreme cases more than a hundred thou-
sand times that of the sun, and, since stellar masses in general are
probably not very unequal, the mean densities of these red giants
are correspondingly small.

SHAPLEY— STAR CLUSTERS. 343

Although these cluster stars are gigantic when compared with
the sun, and are highly concentrated into a compact and symmetrical
organization, they do not differ in physical properties, so far as we
now can tell, from hundreds of isolated stars of the Galaxy. Many
of the naked eye stars equal them in light power, in color, in volume.
The Cepheid variables in the cluster have light-curves, color varia-
tions, and spectra like those of the variables near the sun. In other
words, the members of clusters are normal stars.

Until recently the globular clusters have been accepted as spheri-
cal in shape. When projected on the sky we should therefore ex-
pect them to be circular in outline, and, except for accidental errors
of distribution, the number of stars should be the same whatever the
direction from the center. A systematic study of the photographs
at Mount Wilson has shown, however, that a majority of globular
clusters, as seen in the sky, are slightly but symmetrically elongated.
This condition has been interpreted as a flattening of the cluster
system — an indication that the clusters are not spheres but rather
are oblate spheroids or ellipsoids. Messier 13 is one of the most
flattened ; and though the elongation, in the direction indicated in
the picture by inclined white lines, can be micertainly seen, either
visually or on photographs of the cluster, it is very readily shown
by counts of the individual stars. There are about thirty per cent
more stars in the direction of elongation than at right angles thereto.

The flattening suggests that this great stellar system may be in
rotation about its shorter axis. Observations have not as yet deter-
mined whether or not such a motion exists. It is known, however,
from spectroscopic work at the Lowell Observatory and at Mount
Wilson, that the cluster as a whole is moxing with the high velocity
of two or three hundred kilometers a second. Noting that the mass
of the whole cluster is probably in excess of 100,000 suns, we appre-
ciate that the momentum of this moving cosmic unit must be exceed-
ingly great.

The only component of the motion of Messier 13 now known is
directed toward the earth and toward the greatly extended strata of
stars that constitute the galactic system. If the cluster moves con-
stantly with this velocity, it will have reached the galactic plane
within fifty million years, coming from its present isolation in space

344 SHAPLEY— STAR CLUSTERS.

to the regions where scattered galactic stars are numerous and where
all the open clusters are found.

* The foregoing analysis of a single typical globular system indi-
cates the relevancy of the investigation of clusters to general stellar
problems. The factor that contributes most to knowledge of the
sidereal universe naturally is that of distance, for distance is the
key to the actual dimensions of a cluster and to the real luminosity
of stars. We directly approach problems of even greater interest,
however, when the distance is determined not only for Messier 13
but also for the 85 other globular systems that are now recorded in
our catalogues. Those problems are the size of the known stellar
universe, and the arrangement and relationships of its various parts.

Probably the most interesting of many results deduced from the
study of the aggregate of open and globular clusters is the evidence
of the insignificance of the earth, sun, and brighter stars in the gen-
eral galactic system. We are impressed with the shortness of the
sidereal distances commonly measured as compared with those of
clusters, Cepheid variables, and star clouds. The sun, it appears, is
only a yellow, dwarfish, very old star, eccentrically situated in a
large moving star cluster which is itself situated still more eccen-
trically in an immensely larger stellar organization — that is, in the
general galactic system, which appears to include practically all
known objects of the stellar universe.

The details and general results of the investigation of the dis-
tances and distribution of all globular clusters cannot be mentioned
here ; but we may conclude and briefly summarize by stating that
the study indicates that in volume the galactic system is more than
a hundred thousand times as large as we formerly believed it to be.
The center of the great ellipsoidal system appears to lie in the di-
rection of the rich star clouds of Sagittarius, at a distance of at least
sixty thousand light-years. Its most striking feature, besides its
dimensions and probable mass, is the extensive, much flattened, mid-
galactic segment, which contains open clusters, isolated stars, and
nebulae in abundance, but appears to be empty of globular clusters.
Apparently the globular clusters are approaching this segment from
without — their radial motions, their distribution in space, and their

SHAPLEY— STAR CLUSTERS. 345

probable genetic relationships to open clusters support that view.

It may be that the thousands of sub-organizations in the galactic

system, such as open clusters, star streams, moving groups like those

in Ursa Major and Taurus, wide pairs of stars of common motion,

long-period visual binaries, all oringinate in these in-falling globular

clusters ; it may be, too, that the galactic system itself, which possibly

is little else than a great mixture of disintegrating minor systems,

owes its beginning and subsequent growth to globular clusters. But

whatever its origin and destiny, it is clear that the sidereal system,

as it now stands out, is a giant in mass and volume compared with

the region around the sun to which we have usually confined our

stellar investigations.

Mount Wilson Observatory,
Pasadena, Calif.
March, 1919.

HYDRATION AND GROWTH.*

By D. T. MacDOUGAL.

(Read April 2$, IQIQ-)

The studies described in the present paper are based upon the
results of three methods of observation and experimentation, as
follows :

(a) A comprehensive series of measurements of the variations
in volume of stems, leaves and fruits have been made in which the
course of growth and variations in rate have been correlated with
variations in such environmental factors as temperature, humidity,
water supply, etc. The records include the entire developmental
period of many stems extending in some cases over a period of two
years. A paper descriptive of some of this work was presented
before this Society three years ago and was printed in the Proceed-
ings for 191 7.

(b) Attention has been directed to the determination of the
composition of living matter and the manner in which its com-
ponents are united or mixed in the cell. Series of analyses ar-
ranged to show not only the general character of the cell contents,
but also the seasonal and developmental changes in plants have been
made, in connection with the comprehensive work of Dr. H. A.
Spoehr upon carbohydrate metabolism (now in press).

(c) Measurements of the hydration reactions of tracts of living
cell-masses have been compared with the reactions of sections of
plates of colloids made up in simulation of the composition of plants
show that the water relations of the living material of the higher
plants are those of a colloidal mixture consisting predominantly of
pentosans, of a lesser proportion of albumin, albumin derivatives
and amino-compounds and of a minor proportion of lipins, with an
inevitable small amount of salts.

The components of these three groups are not mutually inter-
dififusible to any extent and hence in my colloidal preparations, as

* Abstract prepared by author from a lengthy manuscript.

346

MacDOUGAL— HYDRATION AND GROWTH. 347

well as in the protoplast, must be considered as simply mixed in an
intimate manner in interlocking meshworks, foams or whatever
interpretations may be given to the structures revealed by cytological
technique. Living matter made up in this manner may be miscible,
or immiscible, according to the nature of the pentosan or protein en-
tering into its composition. It is obvious that the protoplasm in
which the carbohydrate element is like agar would not be " soluble,"
while a mixture composed largely of gum arable would go readily
into suspension. Hydration as used in this paper denotes the union
of water with the molecules or aggregate of molecules of substances
in a colloidal condition inclusive of the action by which first definite
proportions are taken up in so-called chemical combination, and
also of the indefinite absorption combinations.

Growth may be defined as consisting chiefly in the hydration of
colloidal material in a living condition generally accompanied by
accretions to the main components as described above, although not
actually necessary for the conception as the ground or fundamental
structure of protoplasm might be increased by rearrangements of
material already included. The later stages of increase in volume
of a protoplast are coincident with the formation of denser per-
manent structures having the eft'ect of increasing the relative dry
weight with age or approaching maturity, a procedure which may
be taken to be universal in the tissues of the higher animals, ac-
cording to the researches of Donaldson, Hatai and others.

This too is currently assumed to be the case in the tissues and
organs of the higher plants, but my own results include a type of
growth in which this is not the case. Succulents of all kinds produce
stems or leaves in which the percentage of dry weight of small young
organs such as the joints of cacti, leaves of Mesembryanthemum,
and fruits of the melon or berry type, is greater than that of mature
organs or members exemplifying a category of growth especially
interesting in the present connection. Succulence is due to the
hypertrophy or exaggerated growth of cells in which the hexoses
have been converted into pentosans, initially as a result of low water
contents of the cells, the pentosans having a high capacity for water
which is exerted during the remainder of the life of the cell, and it
is in such organs that the relative dry weight does not increase with
age.

348 MacDOUGAL— HYDRATION AND GROWTH.

The living matter of the plant cell includes pentosans which are
to be taken as weak acids and which show only slight dissociation,
and of albumins or amino-compounds, w'hich undergo a greater
hydration with increasing concentration of hydrogen ions. Hydra-
tion of the carbohydrate or pentosan constituent or protoplasm is
affected in the reverse manner. The actiial rate of growth of any
protoplast would therefore be a resultant of the opposing action of
pentosans and of albumins.

Swelling, hydration or growth of bio-colloidal masses is not
however simply a matter of hydrogen ion concentration, as we have
found that the hydration and swelling of pentosan-protein colloids
simulating protoplasm, of tissues of living plants and of dried cell-
masses, is facilitated by the amino-compounds which dissociate as
bases. The greatest increases occur in concentrations of o.oi M to
o.ooi M of such substances as alanin, asparagine, glycocoll, phenyl-
alanin, and ethylamine. Concurrent accelerations of grow'th result-
ing in increased dry weight and greater volume are reported by
many experimenters using such substances in culture solutions. It
is highly probable that bases (cations) may have some effect on
hydration or growth, a matter yet to be tested.^

The essential feature of an idealized growth is the accretion or
addition of water and material to the mass of colloid constituting the
cell. The actual mechanism of incorporation is not easily delineated.
If protoplasm consisted of a system of colloidal structures such as
those of the pentosans and the proteins interwoven, but not diffusing
into each other, the more solid material which lowers the surface
tension to the greatest extent, having the least attraction for water-
molecules, would tend to usurp the position of the surface layer.
Furthermore, the solid phase, whether it be in the form of globules
or in the continuous element, would tend to increase and crowd to-
gether with a lessening of the more liquid phase. This would imply
that when gelatine in small proportion is mixed with agar or starch
paste in the larger proportion that the carbohydrate would form
both the colloidal framework or mesh, as well as the external layer

1 See MacDougal and Spoehr, " The Effect of Organic Acids and tlieir
Amino-Compounds on the Hydration of Agar and on a Biocolloid," Proc.
Soc. Expcr. Biol, and Med., i6: 2)2)' ipiS-

MacDOUGAL— HYDRATION AND GROWTH. 349

of the mass. In actual practice the mixtures are soHdified too
quickly for this to occur.

The separate colloidal masses where they do exist have, of
course, definite boundary layers, as are formed wherever two col-
loidal phases meet. Protoplasm may not be regarded however as
altogether a mechanical admixture of minute strands of material of
different composition. Much of it, including the more fluid portions,
must consist of molecules of carbohydrates, proteins, salts and even
lipins aggregated to form submicrons in the disperse phase or in
the denser more solid fibers, mesh or honeycomb of the structure.
The external layer formed might well be in a sense a mosaic, but
it is to be noted that no actual proof of such a condition is at hand.
Both absorption or imbibition and osmosis including differentiated
diff'usions would be aft"ected by the composition and relations of the
two phases of the colloids in this outer layer, and it seems highly
probable that an adequate interpretation of permeability will be ob-
tained by a study of these features. Meanwhile no general agree-
ment as to the nature of the " membrane " or its action is to be
expected until many widely current assumptions are discarded. The
external layer of a protoplasmic unit is in every case a product of
the surface energy of the mass or systems of living material internal
to it and of the medium and it has no other permanent or morpho-
logical value. Its constitution must necessarily vary widely as does
that of the living protoplasm.

The experimental studies described in the present paper have been
devoted chiefly to the action of protoplasmic masses in which the
pentosans in colloidal combination with proteins and salts determine
the volume or mass of the living material of the cell. The other
soluble carbohydrates, including the hexoses, sucrose, dextrose, do
not occur in the cell in such concentrations as to affect the enlarge-
ment of the protoplasmic mass directly, but in the vacuoles they may
exert an osmotic effect additive to that of the amino-acids which
may accumulate in these cavities. It is to the osmotic activity of
these substances in the vacuoles that turgidity is due and a by no
means unimportant part in the maintenance of the rigidity of organs
and other features is to be ascribed to these turgor stresses and ten-
sions. The inadequacy of osmotic phenomena and of the concep-

350 MacDOUGAL— HYDRATION AND GROWTH.

tion of the semi-permeable membrane to provide a mechanism for
the translocation of complex material from cell to cell, and the in-
corporation of new material in a growing mass has long been recog-
nized. That osmotic pressure however may play an important part
in the enlargement of the plant cell may well be concluded from the
fact that in the stage following the initial swelling of the embryonic
cell, a large share of the increase in volume is due to the increase of
the vacuoles. It would be a mistake to conclude that the vacuole is
simply a sac charged with electrolytes, as these cavities invariably
hold proteins and carbohydrates in a colloidal condition in which the
degree of dispersion may vary widely, but still absorb water. A cor-
rect estimation of the manner in which osmosis and imbibition inter-
lock in growth is one of the tasks demanding the immediate attention
of the physiologist.

The Effect of Organic Acids and their Amino-Compounds on

Growth.

The accelerating effects of acids or of the hydrogen ion concen-
tration on hydration of proteins is well known, and something of the
retarding efifect on the swelling of mucilages or pentosans have been
described in recent papers.

The effect of bases (cations) on these processes has not yet been
measvired, nor has the method by which the amino-compounds act
in this matter been determined.

The hydrogen ion concentration of the liquids in the plant cell
remains constant within limits and in succulents examined by Jenney
Hempel may be expressed by Ph 3-9-7, as determined by electro-
meter and titration methods. The presence of the acids of which
malic is the more abundant, and its combinations with such bases as
potassium, sodium, calcium, magnesium, iron and aluminium makes
a " buffer " by which the degree of dissociation is controlled.

In addition to this comparative stability or narrow range of the
concentration of hydrogen ions, amino-compounds are invariably
present, and their relative amount probably varies but little.

A series of tests were planned in which a comparison would be
made possible between the action of some of the commoner organic
acids and of their amino-compounds.

MacDOUGAL— HYDRATION AND GROWTH. 351

Two groups were chosen for the tests. Succinic acid and amino-
succinic or aspartic, which are dibasic, and its amide as noted above
which is monobasic, and acetic acid and amino-acetic or glycocoll,
which are monobasic. Sections of plates of agar, gelatine, agar-
gelatine, agar-protein and other mixtures were used. Swellings
were carried out in the equable temperature chambers of the Coastal
Laboratory at 15-16° C. A tabulation of the principal results is
given on the following page.

The two organic acids, succinic and acetic, are seen to exert the
classical effect on gelatine, the greatest hydration taking place in
the higher concentrations, the effect decreasing with dilution until
at 0.0004 ^ the swelling in acetic acid was scarcely greater than in
distilled water. At 0.0004 M however the dibasic succinic acid
showed a swelling less than that in distilled water, a fact which sug-
gests a rapid solution or dispersion from the surfaces of the sections
and alterations of viscosity in the mass.

Mixtures of agar (8) and gelatine (2 parts) were now tested,
and the hydration in succinic at 0.00008 M was but 1030 per cent,
as compared with 1684 per cent, in water, while acetic acid was
slightly higher, 1167 per cent. A similar statement would hold for
the action of these acids on agar, and for agar-protein, the hydration
of water along being reached more nearly than in the agar-gelatine
sections.

When we now turn to amino-succinic or aspartic acid and amino-
acetic acid or glycocoll, some new relations are uncovered.

The aspartic acid (amino-succinic) appeared to exercise a notable
influence on the hydration of agar within the range of its solubility.
When more than this was added to the water used for solution a
swelling in excess of the expectancy resulted. It was also seen that
the surface of the liquid became covered with thin crystals. In all
probability the solution or dispersion of some agar into the water
resulted in the displacement of some of the acid with the result that
the sections were actually hydrated from a solution less concen-
trated, giving a swelling in excess of the expectancy.

Asparagin was now applied in a series of concentrations to sec-
tions of agar of the above swelling capacity in water and it was
found that hydration was actually increased or accelerated by the

352

MacDOUGAL— HYDRATION AND GROWTH.

Hydration of Agar, Gelatine, Agar-Gelatine and Agar-Oat-Protein in

Organic Acids and their Amino-Compounds at 16-17° C. Expansion

IN Percentages of Dried Thickness.

Concentration.

Succinic Acid Aspartic Acid
Mol. Mol.

Asparagin
Mol.

Acetic Acid Glycocol! Mol.
Mol. ■'

Agar.

0.3

1950%

0.5

1060%

2804

O.I

1000%

2260%

1333

0.05

1091%

827

2308

1433

O.OI

1273

1270

2365

1560

2964

0.002

1600

1400

2440

1790

3166

0.0004

1750

1788

2720

1955

2605

0.00008

2528

2080

3250

2640

Water

Average :

2600%

Gelatine.

0.1

COS

1200%

1500%

320%

952%

370%

O.OI

700

1033

480

714

0.002

500

380

500

690

360

0.0004

433

340

467

643

360

Water

Average :

600%

Agar 8 — Gelatine 2 Parts.

0.5

0.1

850%

0.05

716%

910%

1485%

850

1233%

O.OI

860

1017

1574

900

i960

0.002

917

1295

1608

922

1767

0.0004

1000

1667

1383

1117 •

1420

0.00008

1030

1786

1383

1 1 67

1484

Water

Average :

1684%

Agar 8 — Oat-proxin 2 Parts.

0.5

500%

0.1

....

809

0.05

700%

855%

1867%

1090

' 1983%

O.OI

864

900

2455

125s

2340

0.002

909

1670

2523

1738

3050

0.0004

1136

2500

2675

2238

3000

0.00008

2330

3050

2600

2480

Water

Average :

2365%

presence of this substance. That this result did not simply appear

by faulty comparisons was shown by the following replacement test.

A trio of sections which had been swelled in distilled water to a

MacDOUGAL— HYDRATION AND GROWTH. 353

total of 2630 per cent, and which had stood in the solution without
any perceptible change for a few hours after the close of the test
was now treated to a o.oi M asparagin solution. The mechanical
disturbance which might result from changing the liquid in the dishes
was minimized by fractionization. About one third of the water
was removed then the level was raised by the addition of asparagin
solution and this was repeated about a half dozen times, the final
result being a solution which was diluted slightly below hundredth
molecular. A slow expansion began at once which continued for
about 20 hours, which raised the total hydration of these sections to
2890 per cent., an increase of 230 per cent., due to the action of the
asparagin on sections which had undoubtedly been reduced in mass
somewhat by solution from the surfaces.

The presence of asparagin in the water in which swelling of
gelatine was carried out, produced an uncertain effect by reason of
supposed solution or dispersion of the gel. Neither can much be said
concerning its action on agar-gelatine mixtures, except that the re-
sults show a maximum at o.oi M.

When the asparagin is applied to mixtures in which the gelatine
is replaced by an albumin, the results included some special reac-
tions. Plates of agar and oat-protein were made up to contain 8
parts of the first and 2 of the last, coming down to a thickness of
0.22-0.23 nmi. These swelled at 17° C. to the proportions shown in
the table, which in some cases exceeded that in water. The swell-
ing in concentrations as high as o.oi M were but little below that in
water.

Glycocoll has been used in many cultural tests with plants and
various interpretations have been placed on its accelerative influence
on growth, and its influence on swelling therefore has an unusual
interest.

Thin sections of agar swelled in all glycocoll solutions less con-
centrated than 0.3 M to the amplitude attained in water and exceeded
it in some cases, a fact which for the first time gives a sound basis
for results in which growth was accelerated and the total increased
by this compound.

Another pentosan, gum tragacanth, was dried from solutions to

PROC. AMER. PHIL. SOC, VOL. LVUI, W, OCT. 21, I919.

354 MacDOUGAL— HYDRATION AND GROWTH.

form sections 0.13 mm. thick on filter paper. Swellings at 15° C.
were obtained as follows :

Distilled Water. Glycocoll.

.03 Hi. 0.05 M. o.oi M.

1380% 1382% 1077% 1462%

This gum liquefies irregularly and hence the figures show the
extent of swelling before active dispersion of the mass begins.

Next a mixture of 9 parts of gelatine and i part tragacanth was
made up at 2.5 per cent, to correspond to a similar mixture of gela-
tine and Opuntia mucilage. Swellings as follows at 15° C. were
obtained :

Distilled Water. Glycocoll

0.3 HI. 0.05 M. 0.01 M.

1320% 1520% 1040% 1320%

Nothing may be concluded on the basis of these figures except
that the hydration of this material reaches a stage where it goes into
dispersion unevenly and in a manner which makes auxographic
readings, as well as all mass or weight determinations, of doubtful
value.

The above tests were repeated with Opuntia mucilage with the
following results :

Distilled Water. Glycocoll (15° C).

0.3 M. 0.05 M. o.oi M.

923% 800% 654% 600%

It is highly probable that the high relative swelling in the more
concentrated solution is due to the hydrogen ion effects on the gela-
tine as glycocoll is known to show some dissociation in this way.

Sections consisting of 4 parts agar and i of gelatine which had
an average thickness of .3 mm. swelled as follows at 15° C. in gly-
cocoll.

0.3 M. .05 i1/. .01 J/. .002 M.

1550% 1233% 1960% 1767%

The average swelling of such sections in water was about 1700
per cent., and the irregularity characteristic of auxographic measure-
ments of the action of this amino-acid is seen in the above results.

MacDOUGAL— HYDRATION AND GROWTH. 355

A preparation was now made in which two parts of the water
sohible protein from oats was added to 8 parts of agar in a 2.5 per
cent, solution of the latter. The plates dried to a thickness of .25
mm. When sections of such biocoUoids were swelled in the gly-
cocoll series, the results were as shown in the table, the hydrations
in concentrations less than o.oi M approaching and surpassing those
in distilled water.

The action of acids being supposedly due to the hydrogen ion
concentration, a test was made of the action of a solution in which
glycocoll (amino-acetic) was added to acetic acid. Trios of surface
slices of Opuntia which had dried to a thickness of 0.8 mm. swelled
163 per cent, in .05 A'' acetic acid and 156 per cent, in a .05 A^ solu-
tion of acetic acid and glycocoll. No especial significance can be
attached to the lesser swelling in the double solution, except that no
evidence as to acceleration of swelling by the addition of the amino-
acid was obtained.

Next trios of sections of 8 parts agar and 2 parts gelatine 0.3
mm. in thickness were swelled in the acetic and amino-acetic solu-
tions 0.01 A^ at 18° C. The swelling in the acetic acid alone was
1450 per cent., while that in the combined solutions was but 1300
per cent., which agreed with the previous efifects in being less than
in the acid alone.

Trios of sections of agar swelled 1875 per cent, in a o.oi A^ solu-
tion of acetic acid at 18° C, while a combined solution of equiva-
lent molecular concentration showed a swelling of 1750 per cent.

There now remained the test with Hving tissues. Some joints of
Opuntia blakeana of 1918 which had been brought from Tucson two
months earlier and had laid on the table with the result that they had
lost much water but were still alive, were used for this test. A trio
of sections with an average thickness of 6 mm. swelled 60 per cent,
in the hundredth normal acetic acid, while a similar trio which meas-
ured 5.5 mm. on the average swelled but 45.5 per cent, in the com-
bined acetic-glycocoU solution. A second feature distinguished the
two reactions, the swelling in the acetic being continuous and ap-
proaching zero during the 20 hours of measurement, while in the
combined solution full expansion was reached in 4 hours after which
a shrinkage resulted in a loss of nearly 5 per cent.

356 MacDOUGAL— HYDRATION AND GROWTH.

A return was made to the biocolloidal mixtures and trios of sec-
tions of agar 8 parts and oat protein 2 parts with a thickness of .22
mm. were swelled at 18° C. The hydration swelling in the hun-
dredth normal acetic acid gave an increase of 13*18 per cent., while
an equinormal solution of the acetic acid and glycocoll gave a swell-
ing of 1605 per cent. This test is the only one of the series in which
the addition of glycocoll to the acetic acid enhances imbibition.

An additional test was made in which equal amounts of gly-
cocoll and acetic acid were brought together at a concentration of
0.00 1 M on agar-oat protein sections as above. The swelling in the
acetic acid was 2681 per cent, or about the same as that possible in
distilled water (2630 per cent.) while the swelling in the combined
solution was slightly less, being 2570 per cent., a difference which
is without special significance in this case, as it is near the limit of
instrumental error or might have been caused by brief temperature
variations.

The positive action of amino-compounds in affecting hydration
is well demonstrated by the above facts. These effects are not coin-
cident with the action of solutions offering the greatest hydrogen
ion concentration, or at any point in which an optimum of such
action might be assumed. The actual maximal hydrations in amino-
compounds take place in reality in attenuated solutions (o.i M to
0.00008^), but the nature of the action has not yet been determined.

The Temperature Factor in Growth.

If growth were dependent in the main upon any reaction or upon
a chain of consequent transformations the influence of temperature
upon the rate and course might be readily and definitely established.
Many authors have assumed such a state of affairs.

Growth however is an interlocking mesh work of reactions and a
rise of temperature through the ordinary range from 5° C. to 30° C.
may and generally does pass the point at which one or more of the
reactions passes the range of its accelerating effect, and in any
further rise becomes an inhibition. Such results are to be found,
for example, in the action of hydrogen ions or in respiration res-
idues. The following experiments may serve to illustrate this
matter.

MacDOUGAL— HYDRATION AND GROWTH. 357

The petioles of some young plants of a Solanmn hybrid in the
glass house at Tucson were available on April 21, 1918. Trios of
sections were placed in distilled water and acids at 18° and 38° C.
witii results as follows :

18° C.
38° C.

Distilled
Water.

Hundredth Normal
Citric Acid.

4.2%

4.2%

II.8

2.6

The swelling in distilled water was nearly three times as great
at the higher temperature, while in the acid solution a retardation
took place which limited the total at the higher temperature to
something over a half that possible at the lower point. The total
swelling in acid at the lower temperature occupied an hour and at
the higher temperature it was a matter of ten or fifteen minutes. A
similar speeding up of imbibition in water was observed. The total
capacity at the lower temperature was not reached for 8 or 10
hours, while at the higher it was something under 2 hours.

Plants of Phaseolns which formed the experimental material for
measuring the growth of pods and seeds bore some pods in which
the beans were nearly mature. Pods of the same stage of develop-
ment as one which was under the auxograph for recording daily
changes were opened and the unripe beans removed. The ends were
cut away, the outer coat removed ; the remainder of each coty-
ledon made on section of which three were taken from separate
pods for swelling. The average thickness was 3.2 to 3.4 mm. and
the swellings were as follows :

Distilled Water. Hundredth Normal Citric Acid.

( a....l4 % a.... II %

^° ^ '{ b.... 8 b.... 9.7

\ a.... 9.6 a.... 6.6

^ ^ 1 b....ii.7 b.... 4.4

The higher temperature to which a was subjected appears to be
above the point at which maximum absorption or imbibition takes
place in distilled water as the swelling was 30 per cent, less than at
the point below. The retarding efifect is much more marked in the
acid solution, however, as the reduction of the total capacity below
that shown at 18° C. amounted to 40 per cent.

358 MacDOUGAL— HYDRATION AND GROWTH.

The material in series b taken at a later date and with seeds
which seemed to be more nearly mature, showed an increase in
swelling in distilled water of about 45 per cent, over the total at the
lower temperature, while the swelling in acid was less than half
that at 18° C. The average of the two series is such that the swell-
ing in distilled water is nearly the same at both temperatures, while
in acid the average at 18° C. is 10.4 per cent., which is nearly double
that at 38° C. at which point the hydration capacity seems to be in-
variably lower than at the lower temperature. These averages rep-
resent a total of six cotyledons each.

A final test of variations in temperature upon material in an
acidified condition was made with dried sections of Opuntia. These
sections were made by slicing away the chlorophyllous layer from
one side of the flat joint and drying the remainder in the desiccator
and in sheets of blotting paper in such manner that buckling and
crumpling was prevented. After all of these precautions were taken,
however, the measurement of the sections was subject to some
error due to the fact that the fibrovascular strands remaining would
increase the thickness under the calipers without reacting in due
proportion to the action of the swelling agent. A wide range of
figures was obtained, but it was apparent that a rise in temperature
did not have an effect on material in acid equivalent to that in dis-
tilled water, as will be apparent from the following measurements
obtained from sections which were .43 to .46 mm. in thickness.

Swelling in
Distilled Hundredth Normal

Water. Citric Acid.

18° c 315% 360%

385 430

486 460

Average 395% 417%

28° C 453 460

38° C 500 477

413 400

Average 457 439

The increase in swelling in distilled water is seen to be about
twice that in the acid in the rise from 18° C. to 38° C. The influ-
ence which the condition in question may exert on the rate of
growth is obvious. Thus the course of enlargement of an organ or

MacDOUGAL— HYDRATION AND GROWTH. 359

of a cell-mass in so far as this consists in hydration may vary widely
in the first instance because of the residual acids in the colloids and
the balance or accumulation of this will in turn depend upon the
effect of the enzymatic or respiratory processes in metabolism.
Thus a rise of lO degrees from the customary morning temperature
of 15° C. which has been encountered in so many of these experi-
ments, would result in an acceleration of growth largely determined
by the state of acidosis of the plant. A rise from the same tem-
perature later in the day or under other conditions of illumination
would necessarily have a different result. An extension of the at-
tempts to bring rates of growth into a figure or formula, therefore,
would be a forced application of knowledge of one process to a com-
plex of activities in which any change in temperature might set up
opposed alterations. In consequence of this fact, the coefficient of
variation for 10° C. is seen to vary from one to seven in various
organisms.

The capacity for hydration and growth is a resultant of the com-
position and proportions of the principal components of the living
matter and the relations of the phases in which they occur, modified
by the " nutrient " salts absorbed in its structure, and by the products
of unceasing metabolic changes, especially the transformations
which are comprehended in respiration and which carry compounds
through a stage in which acids are formed. These features as in-
fluenced by temperature determine the rate, daily course and total
expansion in growth. In addition, a certain amount of material is
lost from the plant in the form of carbon dioxid, and as has been
emphasized on the preceding pages, the surface loss of water may
overbalance absorption. The rate, course, and amount of growth is
therefore affected by many agencies and includes multiple interlock-
ing reactions.

Water-content, Dry Weight and Other General
Considerations.

Two different types of organs or shoots with respect to the
variations in the water-content and dry weight are recognizable in
the material which has served for studies in growth as described in
this paper and in the work of other writers. The commoner types

360 MacDOUGAL— HYDRATION AND GROWTH.

of woody stems, thin leaves and the organs of the greater number
of the higher plants undergo a development which terminates in a
mature stage in which the proportion of solid material is very much
higher than that found in younger material. A parallel procedure
is the prevalent one in the tissues of the higher animals. Thus by
way of illustration, Donaldson found that the proportion of water
in the bodies of mammals disminishes with age, and Hatai has
shown that the percentage of water is an indicator of phases of
chemical alteration in the composition of the body.^

Growth and differentiation of cell-masses into specialized tissues
is not inseparably connected with increases in dry weight, however,
as has been demonstrated by studies of the growth of frog larvae in
the earlier stages,^ and it is highly probable that similar phenomena
are prevalent in the fleshy fungi and other lower forms of plants.

The distinction between the two kinds of growth has not been
made previously in studies of plants and the matter was finally
taken into consideration in the experiments late in 191 8.

Etiolated plants furnish examples of growth with a small increase
in proportionate dry weight, but chief interest attaches to plants
which normally show such action, and the most striking illustrations
are to be furnished by the organs of succulent plants and by fruits.
The relative amounts of solid material in the flattened joints of
Opuntia does not increase with the course of development toward
maturity, and joints which have reached full size may contain over
91 per cent, of water. Secondary thickening, especially that which
results from branching and the development of additional fibrovas-
cular tissue, may cause an added amount to be formed. The propor-
tion of dried material and water in the leaves of Mesembryantlicmiim
does not vary greatly with age. These and probably all succulent
forms are characterized by an exaggerated production of mucilages
or pentosans, and have certain implied cycles of metabolism includ-

1 Donaldson, H., " The Relation of Myelin to the Loss of Water in the
Mammalian Nervous System with Advancing Age," Proc. Nat. Acad. Sc, 2:
350. 1916.

Hatai, S., " Changes-in the Composition of tlie Entire Body of the Albino
Rat during the Life Span," Amcr. Jour, of Anat., i : 23. 1917.

- Ostwald, W., " Ueber die zeitlichen Eigenschaften der Entwickelungs-
vorgange," p. 49. 1908.

MacDOUGAL— HYDRATION AND GROWTH. 361

ing an incomplete type of respiration which leaves large acid resi-
dues. The total acidity of the cell-masses may vary greatly during
development and during the course of a day, and the actual acidity
or hydrogen ion concentration of the sap resulting from the buffer
situation may also show a marked variation, but within much nar-
rower limits.

Although the development and maturation of fruits such as
•berries obviously includes a growth in which the total effect is one
of practical maintenance or increase in the water-content, studies of
their growth seems to be lacking. It was therefore planned to ar-
range a final series of experiments in which the enlargement of
fruits with increasing dry weight and others with low relative dry
weights should be measured. The walnut was taken to represent
a structure with accumulating solid matter, and the tomato for the
other type.

The walnut consists of a thick fleshy exocarp and a heavy endo-
carp which finally becomes hard and bony with the deposition of
anhydrous wall material. The enclosed embryo also accumulates a
large amount of condensed food-material. The tomato is a large
globose berry in which deposition and thickening is confined to the
small hard seeds. The greater part of the fruit is a fleshy pulp,
which becomes more highly watery as progress is made toward
maturity.

Nuts of Juglans caUfornica var. quercina Babcock, of various
sizes from 3 mm. in diameter to that approaching maturity were on
two trees in the garden at Carmel in June, 1918. Suitable supports
being provided, the bearing lever of an auxograph was rested as
lightly on the young nuts as was consistent with a clear record, and
temperatures were taken by thin thermometers thrust into similar
nuts or into young stems near the preparation. Fifteen nuts were
measured for periods of two or three days or for as long as two
months in the case of No. 10.

Coincidentally with the measurements an eff'ort was made to de-
termine the degree of saturation or hydration of the stems on which
the nuts were borne. A well defined " negative " pressure was de-
tected in the basal branches of Juglans major which was growing
near the experimental tree. A basal branch 1.2 meters from the
trunk gave a dry looking surface when it was cut off.

362 MacDOUGAL— HYDRATION AND GROWTH.

A section of a similar branch about 8 mm. in thickness and 42
cm. long was cut away from another basal branch of the tree, the
end of the detached portion quickly sealed with vaseline and when
all was in readiness the tip was excised and the cut thrust into
water to ascertain the actual deficiency in this portion. 14 hours
later a total of 6 c.c. of water had been absorbed and 24 hours later
8.5 c.c, which was a practical saturation, at a temperature of 18°
to 20° C. The volume of the branch proved to be 35 c.c. so that the
amount of water absorbed was 24 per cent, of the total.

Sections of young internodes of Jnglans calif ornica qucrcina
which had an average diameter of about 2.5 mm. were swelled in
solutions as below, then dried and swelled again with results as
below at 16° C.

Water. Citric Acid o.oi Normal. Potass. Hydrate o.oi M. Potass. Nitrate o.oi M.

Fresh — living.
10% 14% 13.2% 12%

After Drying.
34 34 34 32

(On basis of original thickness.)

The unsatisfied capacity of these sections taken from young ter-
minal internodes was comparatively great, doubtless due in part to
the constant drain of the active leaves they bore. The older wood
including that formed the previous year showed an absorptive
capacity of 22 per cent, in water. Now it is from these older inter-
nodes that the nuts arise.

The nuts were highly turgid, exuded sap when cut into, and
hence must have had a colloidal composition which acted to with-
draw water from the stems which contained a much lower percentage
of water. The soil w^as low in moisture content at this time, as it
had been four to five months since the rains.

Tests of nuts 8 to 10 mm. in thickness from which tangential
slices had been removed to give a uniform thickness of 7.5 mm.
were made in July, and these swelled at temperature of 17 to 30° C.
in solutions as follows :

Distilled Water. Citric Acid, .01 Af. Potass. Hydrate, .01 Af. Potass. Nitrate, .oiAf.

1.5% 1.8% 1.4% 2%

MacDOUGAL— HYDRATION AND GROWTH. 363

A useful conception of the hydration conditions in the stems and
fruits may be formed, if due weight is given to the measurements
cited above. The woody branches of the previous year, on which
botn the leafy green twigs and those bearing the nuts are borne, had
a relatively large deficiency in water so that sections a few centi-
meters long absorbed about 20-25 per cent, of their volume of dis-
tilled water in 24 hours at 20° C. No swelling test was made, but
it is obvious that an enlargement of only a small fraction ought to
be shown by this or any branch with a mature woody cylinder. The
active green twigs still in a state of elongation arising from these
branches had a swelling capacity of 10 per cent. The growing nuts
arising from the drier stems exuded water from cut surfaces, the
cotyledons being sacs of watery fluid, in contrast to the dry ap-
pearance of sections of the youngest internodes, and showed a
swelling of less than 2 per cent, and soon shrunk when placed in a
cylinder of distilled water after being cut in halves. In a system
of this kind any alteration of the conditions which would facilitate
transpiration would have a differential effect on the older stems, the
green leafy twigs and the fruits. The loss from the stems would
be affected least since the bark would effectually prevent any notable
increase in evaporation from the relatively dry woody tissues. The
loss from the leafy twigs would of course tend to become greater
and the deficit in both leaves and twigs would be increased and their
absorbing power correspondingly increased. The outer integument
of the nuts being still in an embryonic condition and being highly
hydrated the loss would reach a maximum rate with the daily effect
of causing a cancellation of enlargement beginning mid-forenoon at
20-22° C. and continuing until mid-afternoon when a fall in tem-
perature brought transpiration to a rate below that of accession
from the stem.

A large percentage of the nuts which were placed under the
auxograph lever were cut off at various stages of development by
abscission of the stalk. The inciting causes of the actual anatomical
change which cause the abscission lie outside the scope of this
article. It was noted however that it was preceded by a period in
which the nut showed a shrinkage by day in the higher temperatures
and lessened humidity, alternating with equalizing enlargements at

364 MacDOUGAL— HYDRATION AND GROWTH.

nights. Finally an abrupt rapid and continuous shrinkage resulted
in the separation of the stalk.

The general features of growth of these nuts may be illustrated
by a resume of history of No. lO, which was under continuous ob-
servation from July 15 to September 9, 19 18, during which period
of 56 days its diameter increased from 16 mm. to 26.5 mm. Of
this 2.25 mm. was gained in the first five days of cool foggy weather.
This effect was confirmed by the fact that a cessation or retarda-
tion occurred at midday and was most pronounced on hot sunny
days, suggesting a direct water-loss. In the week ending July 29,
the total growth was an increase of 1.7 mm. This period was char-
acterized by heavy fogs and mists in the forenoon, both the amount
of shrinkage and rate of increase being lessened — an equalization to
be ascribed in part to approaching maturity. The temperature taken
from a thermometer thrust in a young branch of the thickness of
the nut ranged from 13 to 22° C. The completion of the record of
No. 10 was followed by cutting of the branch bearing it at a dis-
tance of 30 cm. placing excised end in water and arranging the entire
preparation in the dark room at 17° C. with the nut under the bear-
ing lever of the auxograph. Swelling continued for about 20 hours
after which shrinkage began which rapidly accelerated.

The general features of growth were also well illustrated by the
following notes on No. 15, which was brought under observation
when it was about 15 mm. in diameter and put under an auxograph
amplifying 45 on August 3. Great daily variations in size with a
net total increase were displayed every day. Actual enlargement
could be detected between noon and 2 o'clock which continued until
8 or 10 the following morning, depending upon the sunshine. If
the sun rose clear, shrinkage began immediately. If the morning
was foggy, it would be delayed, but the increase in thickness was
rapid, being mostly accomplished in 2 hours. jNIinor variations in
this general procedure might be brought about from the shade of
clouds, especially noticeable at noonday August 6 and to be seen at
other times (see Fig. i).

It is to be seen from the above that the fruit of the walnut in an
environment favorable to its development exhibits daily variations
in growth clearly attributable to the balance between transpiration

Macdougai^-hydration and growth.

365

and absorption. The nut in a growing condition has a high water
content, and a small unsatisfied capacity, but its supply from the
relatively dry stems must come slowly, so slowly that any marked
increase in transpiration would overbalance absorption in the nut
and result in cessation of enlargement or even shrinkage.^

Fig. I. Tracings of auxographic record of walnut (No. 15) during
three weeks. Variations are amplified forty-five times and the downward
movement of pen signifies enlargement. The scale is numbered in millime-
ters. Shrinkage during the midday period lessened by fogs and acceleration
in growth by the humidity and increased water supply from rain are among
the more striking features.

Such retardation or apparent cancellation of growth by rapid or
excessive water-loss has already been discussed in connection with
the presentation of my original results dealing with the growth of
Opuntia. Recent exemplification of this action in Ccstrum noc-
tnrnum has been described by Brown and Trelease.*

The fruit of the tomato (Lycopersicum) presents similar fea-
tures of behavior. The proportion of solid matter in young fruits

3 MacDougal, D. T., " Reversible Changes of Form in Succulents." Re-
port Dept. Bot. Res. Carnegie Inst, of Wash. No. 14, p. 71. 1915.

4 Brown and Trelease, "Alternate Shrinkage and Elongation of Growing
Stems of Cestrum nocturnuiii," Philip. Jour, of Science, 13 : No. 6. Bot. p.
353. November, 1918.

3G6 MacDOUGAI^HYDRATION AND GROWTH.

was greater than in those approaching maturity, representing the
extreme of this type. Four fruits less than a week old, with radial
diameters of 14, 16, 17 and 18 mm. were found to weigh 14.650
grams. These were fragmented and placed in a beaker on a water
bath at about 100° C. for 48 hours, at which time the dry material
remaining was 1.90 grams. From this it is to be seen that the young
fruit contained 87 per cent, of water and 13 per cent, of dry ma-
terial. A mature fruit of the same kind as those measured taken
September 10 from a ranch near the Coastal Laboratory was 46
mm. in axial diameter and 58 mm. in radial diameter and weighed
93.050 grams. This was dried over a water bath for 2 days, at
which time 8.400 grams remained. From this it is to be seen that
the ripe fruit contained 91 per cent, of water and 9 per cent, of dry
material. (See Anderson, S. P., "The Grand Period of Growth in
a Fruit of Cue iirb ita pepo DeterminQd by Weight,"' Minn. Bot. Stud.,
i: 238, 1894-98; and MacDougal "Practical Text-book of Plant
Physiology, pp. 293, 294, 1918, New York.)

A number of plants of the tomato were grown in suitable boxes
of soil at a ranch in the Carmel valley, and were in such a stage of
development that young fruits were available at the Coastal Lab-
oratory in August, 1918.

Six plants in all were used and continuous records from fruits
of an axial diameter of 3 to 4 mm. to maturity at 50 to 55 mm. were
obtained. The fruits were oblate-spheroid in form and the auxo-
graph was arranged to register increase in diameter nearly parallel
to the axis in some cases and radially or at right angles to it in
others. In addition to the other advantageous features of this ma-
terial their regular form and mode of growth made it possible to use
the variations in diameter as a basis for calculating the changes in
volume of the fruits taken as spheres.

Temperatures were taken by thrusting the thin bulbs of the small
thermometers into fruits near the one under measurement. The
development of such fruits was but little affected by this wounding
and the thermometer remained firmly in place as in the fleshy joints
of Opuntia, in the measurement of which this method was first
practised. The preparations stood in a well-ventilated glass house
and the soil around the roots was kept moist in accordance with the

MacDOUGAL— HYDRATION AND GROWTH. 367

cultural requirements of these plants. The results may be best set
forth by the description of the action of the several fruits measured.

No. I was placed in the greenhouse and a fruit 29 mm. in diam-
eter was fixed on a block of hard cork in such position that it gave a
radial bearing to the auxograph which was set to amplify changes in
volume by 5, on August 9. The record was kept continuously until
September 18, at which time the radial diameter of the fruit was
51.5 mm. The fruit was turning yellow on September 16 and was
showing fluctuations in volume comparable to those in No. 2, with
which it was run in close comparison and under almost exactly the
same conditions of moisture and temperature as recorded.

No. 2 was adjusted to the auxograph in the greenhouse on
August 9 in such manner as to give modifications of the axial diam-
eter, which at this time was about 27 mm. The record was con-
tinuous until September 18, at which time the diameter was 50.5
mm. This fruit like No. i was beginning to turn yellow on Sep-
tember 16.

No. 3, 10 mm. in diameter, was adjusted to the auxograph to
record variations in radial diameter on August 21, and a record was
kept continuously with frequent notations of temperature and sun-
shine, etc. It is to be noted that i, 2 and 3 were under equable tem-
peratures, 19 to 20° C, and high relative humidity during the rain-
fall of September 11 and 12.

The fact that the greatest increase in growth occurs in fruits at
diameters between 16 to 25 mm. in diameter, before half the final
size is reached is a point to which we shall recur in the discussion
of growth in terms of volume. Thus in No. 3 the increases in thick-
ness weekly were as follows : 6 mm., 6.3 mm., 2.5 mm., 3.5 mm.

If this method be followed it would at once be obvious that while
the rate of increase in diameter would be a direct measurement yet
as the fruit increases as a globe the actual material added could be
regarded as a shell on this globe. The rate in terms of volume
would therefore be the amount of this shell to be calculated by find-
ing the difference between the initial volume, and the volume at the
end of each period by the formula Pi r — Pi R in which r^the
new radius and R the initial radius. The rate by direct measure-
ment of diameter and by volume increases may be compared as
below.

368 MacDOUGAL— HYDRATION AND GROWTH.

Average daily rate of growth measured as increase in diameter
was as follows for periods beginning as follows :

Aug. gth. Aug. i6th. Aug. 21st. Aug. 28th. Sept. 4th. Sept. nth.

1.7 mm. I.I mm. 0.7 mm. 0.4 mm. 0.28 mm. 0.17 mm.

Average daily rate of increase in volume :
1.953 cu. mm. 1,885 cu. mm. 1,548 cu. mm. 1,036 cu. mm. 732 cu. mm. 521 cu. mm.

The rate on September 11 by direct measurement would ap-
pear to be one tenth that of a month earlier, yet actually water and
new material was being added at a rate equivalent to one fourth of
the earlier rate. The radial proportions would make the rate on
August 21 not much more than 40 per cent, of the rate on August 9,
while the increase in volume was over 96 per cent. The rate in the
week beginning August 28 would appear to be less than a fourth
that by direct measurement on August 9. yet actually the increment
of water and material is more .than half that in the younger stage
and smaller size.

A second plant with the auxograph bearing arranged to take
axial variations in the fruits which measured 33 mm. was arranged
to run concurrently with No. i and under identical temperature and
conditions of moisture. The daily rates of increase in diameter
were as follows for the periods beginning :

Aug. 9th. Aug. i6ih. Au_g-. 2ist. Aug. 2Sth. Sept. 4th. Sept. nth.

0.95 mm. 0.7 mm. 0.56 mm. 0.3 mm. 0.2 mm. 0.2 mm.

Daily rates in terms of volume :
1,563 cu. mm. 1,388 cu. mm. 1,274 cu. mm. 949 cu. mm. 380 cu. mm. 420 cu. mm.

Here again the actual course of growth as calculated in terms of
volume shows that simple measurements of the thickness do not
express the real values in growth in such organs.

The third test was made on a fruit taken at a much earlier stage
at a diameter of 16 mm. with a transverse or radial bearing, the tem-
perature and moisture conditions being similar to those of i and 2.
The daily rate of increase was as follows for the weeks beginning
on the following dates :

Aug, 2ist. Aug. 28th. Sept. 4th. Sept. II. Sept. i8th. Sept. 2sth.

.85 mm. .85 mm. .64 mm. 8. mm. .3 mm. .^l mm.

Daily rate in volume :

537CU. mm. 851 cu. mm. 885 cu. mm. 1,6431 cu. mm. 594cu. mm. 662 cu. mm.

macDougai^hydration and growth.

369

The actual volume of this fruit at the close of the experiment
was approximately 2,200 cu. mm. and its growth had been followed
for a period of 40 days. It is notable that in the earlier stage in the
advance of the fruit from 20 to 26 mm. in diameter that while the
increase of the diameter seems constant yet the actual accession of
material is very much greater. Then in further development the
average increment to the diameter was smaller, yet the actual ac-
cession of material was greater. Following this the rate falling
from 0.8 to 0.3 mm. daily the accession decreases less than half.

The record of growth of No. 3 shows beyond question the effect
of transpiration and water loss on growth. As the daily tempera-
tures of the fruits rose fromi 12° C. and 14° C. to 26° C. and 28°
C, acceleration ensued up to a point where the rise caused a water
loss overbalancing the gain by hydration. Higher temperatures
therefore did not facilitate or accelerate growth unless accompanied
by high relative humidity. Thus the highest are those of midday
and afternoon, with fog or showers. This is especially marked on

Nn \fi Hn WIV Nn U\ Hn H\ Mn Mt

Fig. 2. Auxograph tracing of variations in volume of fruit of hybrid
solatium. Downward course of line denotes enlargement amplified 45 times.
The scale is numbered in millimeters. Midday temperatures are given.

the record of September 10, ii, 12 and 13, in which a 50-hour rainy
period was anticipated and followed by high humidity. It was not
possible to increase the water supply by watering the soil around
the roots in such manner as to cancel the midday shrinkage or slack-
ening in growth. One especially striking effect is that in which the
rise in temperature consequent upon the cessation of the rain from

PROC. AMER. PHIL. SOC VOL. LVIII. X, NOV. 28, I919.

370

MacDOUGAL— HYDRATION AND GROWTH.

20 to 25° C. at 3 P.M. was followed by a lessened rate of growth,
and on the cloudy days was uniformly high. Similar effects were
exhibited by a small fruit of a potato in a greenhouse at Tucson in
May, 1918.

The water deficit of the stems as measured by swelling includes
that of the entire structure. The fruits however receive their supply
through special conduits which sustain only a mechanical relation to
the other parts of the stem which may be active in its swelling.
Such non-conducting tissues of course draw their supply from this
system of conduits also, but it is highly probable that the dispropor-
tion between the water content of the fruit and of the tracts in the
stem from which it receives its supply is not so great as might be
indicated by the measurements given. The hydration capacity of
the fruits would be the resultant of many factors including the
pentosan-protein ratio, the hydrogen ion concentration, the action of
salts and the effect of the amino-compounds.

N(

bn ui,^

N |yf

Ko

im Hit Ua
1

u

i

No

h

i

.

h

1

Neon

1

\

^ ' 1

1 \

\ ' --^

1
1

1 /•

'.

1

^K^K^

L Us /

\ March '^,1916 \ \

\ \ \^

-^ \

\^

1

\^

\^

-^;

NojEn wff RSn Mjf N$n t^i nJoo fS NSoii fSf NoEn m! N3on W

Fig. 3. Tracings of auxographic records of Opuntia discata during two
weeks of secondary growth. Downward course of the pen denotes enlarge-
ment amplified 45 times. The scale is in centimeters. The shrinkage at night
is greater in some instances than the enlargement of the preceding day.

The balance between water-loss and the gain by absorption is so
delicate in the fruits the action of which was measured, and in many
stems that increased humidity may be followed by accelerated
growth, while a rise of ten to fifteen degrees in air temperature may
check enlargement by increasing water-loss and it is also taken into
account that the consequent rise in temperature of the growing stem
may actually lessen the water-holding capacity of the biocolloids
which make up living matter.

MacDOUGAL— HYDRATION AND GROWTH. 371

It would be unwise to assume, however, that the general pro-
cedure followed by the walnut, tomato and by some green stems, is
universal.

The growth of the flat joints of Opitntia which were described
before this Society in April, 1917, presents many features different
to the above, and some tracings of the growth record of Opuntia
discata for three weeks are shown in figure 3.

Enlargement (denoted by a downward course of the recording
pen) begins in the morning and continues with the rising tempera-
ture until mid-afternoon, then slows down and shrinkage sets in
which continues through the night. Such shrinkage must inevitably
accompany and result from excessive water-loss, and in confirma-
tion it is found that the cacti show the greatest transpiration at
night, at which time the acidity rises until it is ten times as great as
in the daytime. The disintegration of this by light and higher tem-
peratures increases the imbibition capacity of the cells known to be
high in pentosans and a swelling or growth in the daytime results,
producing a growth record almost exactly in reverse of that of the
walnut and tomato.

The rate, course and amount of growth are at all times a re-
sultant of agencies which affect water-loss, hydration capacity,
respiration and its residues and other features of metabolism.

The procedure in transpiration from the surfaces of a growing
organ may be such that the maximum loss partly masking growth,
may for example occur in the cacti at night, while it comes in the
midday period in the commoner types. The swelling or hydration
capacity of any plasmatic mass which determines its capacity for
growth or enlargement depends in the main on the mucilages or
pentosans present and their amount relative to the proteins, the
character of which also affects growth.

The biocolloids of the plant show a degree of swelling in water
greater than that in solutions containing free hydrogen ions, so that
growth generally is most rapid in cell solutions near the neutral
point. In modification of this last statement, it is to be pointed out
that maximum swelling effects may result from the action of some
of the amino-acids.

Rises in temperature within the range ordinarily associated with

372 MacDOUGAL— HYDRATION AND GROWTH.

growth may result in lessened swellings or hydrations, a result to
be connected directly with certain fluctuations in the growth rate.

Succulent stems, leaves and fruits may show growth in which
development or age is not accompanied by an increase of relative
dry weight. In some of these structures, such as the tomato, it is
possible to analyze the auxographic record and determine the actual
total accretion of material in any stage of development. The graph
plotted from such data does not follow the contours of a figure
plotted from the variations in thickness or diameter, the chief dif-
ference being that the maximum comes at a later period.

SOME CONSIDERATIONS ON THE BALLISTICS OF A
GUN OF SEVENTY-FIVE-MILE RANGE.^

By ARTHUR GORDON WEBSTER. Ph.D., Sc.D., LL.D.
{Read April 20, igi8. Received May 2, 1919.)

On the afternoon of March 23, 1918, the civilized world was
astounded by the news that the Germans were bombarding Paris.
Inasmuch as it was known that the nearest point at which the Ger-
man lines approached Paris was over seventy miles distant, curiosity
was universal as to how this result was accomplished, and the most
unlikely and absurd hypotheses were suggested.

The writer, among others, was asked whether or not the result
was likely to be due to an aerial torpedo. The next day revealed the
truth, which was, simply, that the Germans had really built a gun
which carried a projectile this hitherto unheard-of distance.

It has since been determined that the gun is situated in the Forest
of St. Gobain in the neighborhood of Laon, at a distance of about
120 kilometers or seventy miles.

At this distance the curvature of the earth causes one end of such
a line to be about a half mile below the horizon at the other end, so
that it is impossible to see the target from the gun or vice versa ;
there being no mountains of any such height in the whole region,
visual aiming would be quite out of the question.

I wish, first, to call the attention not only to the remarkable bal-
listic achievement of the Germans in so far surpassing previous
ranges but also to the unique opportunity possessed. It is obvious
that precision of aim at such a distance is well-nigh impossible and
that the only hope of effecting any damage lies in the possession of
a very large and valuable target. Little has been allowed to come
through the cable as to the damage done by these long-range shots,
but enough has been learned in order to see their terrible potential-

1 Contribution from the Ballistic Institute, Clark University, No. 2.

373

374 WEBSTER— BALLISTICS OF A GUN

ity. I shall remind you by a few lantern slides of the concentration
of monuments of civilization to such a degree as probably exists
nowhere else in the world.

The gun was obviously aimed to strike the Cathedral of Notre
Dame, cathedrals being the German specialty in this war. The
church in which eighty people were killed on Good Friday is easily
identified as the church of St. Gervais, which is across the street
from the Hotel de Ville, the slides of which shown are taken from
the top of that church. Other objects nearly in the line of the
" axe " are the Louvre, the Sorbonne, the Pantheon, the Bon
Marche (a department store) ; I will not fatigue you with others.

Even if the so-called ellipse of dispersion should be very large
it is easily seen that, if the shot should fall anywhere within a
circle of perhaps twenty-five miles in diameter, the moral effect
would be very great.

The longest previous range used during the war was about
twenty-two miles with the gun with which the Germans bombarded
Dunkirk. In an article in Nature, March 28, 1918, which I have
just seen this morning, my friend Sir George Greenhill, the author
of the article on Ballistics in the Encyclopedia Brittanica, says :

"In the language of sport the German gunner has 'wiped the eye' of our
artillery experts and defied all the timid preconceived notions of our old-
fashioned traditions. The Jubilee long-range artillery experiments of 30
years ago were considered the ne plus ultra of our authorities, and we were
stopped at that as they were declared of no military value. Today we have
the arrears to make up of those years of delay, but the Germans watched our
experiments with great interest, resumed them where we had left off and
carried the idea forward until it has culminated today in his latest achieve-
ment of artillery of a gun to fire 75 miles and bombard Paris from the
frontier."

The Jubilee gun, referred to, fired a shell weighing 380 pounds
at an elevation of 40° with a muzzle velocity of 2,400 feet per
second, giving a range of 22,000 yards or 121/2 miles. -

2 It is a singular coincidence that as I left my laboratory to attend the
meeting of the American Philosophical Society I remarked to my assistant,
" Nothing is likely to ' queer ' this paper except the fact that Sir George
Greenhill may have calculated the trajectory and published it in Nature."
As I entered the room, the Secretary, Dr. Hays, called my attention to Sir
George's paper and I was greatly relieved to find that it contained no figures
of the trajectory.

OF SEVENTY-FIVE-MILE RANGE. 375

Inasmuch as the only thing that has made possible this great
ballistic achievement is the taking advantage of the decrease in
resistance due to the decreasing density of the air in high altitudes,
it is first necessary to make some assumptions as to the law of
decrease with the height. I have here assumed that the temperature
is constant, giving the isothermal law of Laplace. I accordingly put

(i) S = S„e-

iii

where 80= i.2932kg./meter^, and k is equal to .1251, the altitude y
being expressed in kilometers. As the shot rises to a height of
about forty kilometers, or twenty-five miles, this results in a diminu-
tion of density of about sixteen times. While it is true that for the
upper part of the trajectory the form is practically that of the
parabola as investigated by Gallileo for the vacuum, it is by no
means true, as certain discussions that I have seen would seem to
indicate, that there is a region about two miles high at which the
effect of the atmosphere suddenly stops.

The second thing that must be known is the so-called ballistic
coefficient of the projectile, involving its mass and diameter, the
resistance of the air being supposed to be proportional to the square
of the diameter while the acceleration is inversely proportional to
the mass.

For lack of sufficient information at the time that this paper was
prepared, it was assumed that the mass of the projectile was 300
kilograms and its diameter twenty-two cenitmeters, or eight and
one half inches. It is also necessary to know the form factor, which
depends upon the sharpness of the projectile. In the calculations
made here the number .9, which is that of old-fashioned, rather
blunt projectiles, is used. As, however, reports on the shell have
shown that it is furnished with a long pointed cap of sheet metal, the
form factor should be considerably reduced. If, however, we take
a mass of 180 kilograms, or 396.8 pounds, the results given here will
be exact if we assume a form factor of .54, which is undoubtedly
much nearer the correct value. Finally, if we assume the mass
to be 120 kilograms, this will give the same trajectory with a form
factor of .36, which is smaller than that of any shot with which I
am acquainted.

376

WEBSTER— BALLISTICS OF A GUN

The method of calculation is as follows : The first approximation
by successive arcs indicated by Gen. Siacci is used. The distance
of 120 kilometers is divided into twelve equal parts, and the chord
is drawn, making use of the velocity obtained, at the end of the
preceding chord. We make use of the familiar equations ot
ballistics,
(2) ds = pd9,

(3)
(4)
(5)

V =

ds
dt'

an = V-/P = — g COS d,

dvx

-di^-'-^^'^'

in the last of which, which is the only dynamical equation involved,
we put

vdd d{v cos 6) c

(6)

dt^ -

vf{v).

gcosd' d9

Instead of differentials we use finite differences. The table shows
the computation required for Curve 2, Fig. 4.

c

OMPUTATION

FOR CURVE

2

r

Vr

Vr'

DEC.

v'cole

^y

DEG.

^e„

y

KQ./MET

AVV

M

Dx

l'x-Al>

CosSrt

VrH

1

1300

169.10''

52'

640.10'

705

3.3'

.058

6.10^

190

138.3

12.5

862

661.x

.600

1002.0

2

1002

1009

48.7

438

1120

55

.097

16

40

37.5

15 1

282

624.2

.729

856.5

3

865.5

744

43.2

389

1260

7.7

.133

25.5

12

13.2

IG.O

105 5

611.0

.814

749.0

4

749

561

35.5

373

1310

10.0

.175

32.5

5.5

6.88

16.35

56 3

604.1

.903

$69.0

5

G69

447

25.5

-365

1340

12.55

.219

37.0

2.0

2.83

16.55

23 3

601.3

.974

617.0

6

617

380

13.0

360

1360

14.75

.258

39.0

1.0

1.54

1665

12.8

599.8

.909

600.0

7

600

360

-1.75

360

1360

15.6

,272

38.25

1.0

1.58

16.66

13.15

598.2

.954

627.0

S

C27

392

-17.35

359

1365

14.35

.250

36.0

2.0

3.03

16.60

25 1

595 2

.851

698.0

9

698

487

-31.7

355

1380

11.5

201

26.5

6.0

8.01

16 80

67.2

587.2

729

805.0

10

805

649

-43.2

345

1420

8.65

.151

180

22.0

25 8

17 0

219.0

561.4

.618

908.0

11

908

825

-51.8 .

315

1550

6.8

.118

•8.0

105.0

108.5

17 7

952

453.0

.522

867.0

•MASS = 300 KG. JNITIAL VELOCITY- 1300 -

Oo="' ,2.i<=l'8 SEC. APPROX

The meaning of the symbols is as follows :

OF SEVENTY-FIVE-MILE RANGE. 377

The nvimber of the chord is denoted by r; the velocity of the
projectile is denoted by v; and its A'-component, by Vx. The in-
clination of the tangent at the beginning of the arc is dr. The drop
away from the tangent, due to gravity, which is of the second order
of smallness with respect to the length of the arc, is denoted by Ay.
The mean ordinate of a given element is indicated by y. The re-
sistance R is found from the graphical table (Fig. 2). The change
in the horizontal component of the velocity is Avx, the time taken
for the projectile to traverse one arc is A^. Dx represents the " set-
back " computed from the resistance, representing the amount by
which the projectile falls short of the assumed horizontal distance
owing to the resistance of the air. The height reached at the end
of the arc is

y={A' — Dx) t3.nO — Ay,

The most important thing in a ballistic calculation is the knowl-
edge of the ballistic, or resistance, function f{v). We have here
made use of the famous result of the greatest of recent ballisticians,

1 Air Resistance Curve of StdCCL

1 /c' K(v) - lo'-M

3 Air Resisiance Curve of F. Hrupp

Ve/ociiy in Meters pe.r Second
Fig. I.

General Siacci, in his papers in the Rivista di Artiglieria e Genio
(1896), from which it appears from the results of thousands of
shots made by Bashforth in England, Mayevski in Russia (shots

378

WEBSTER— BALLISTICS OF A GUN

made at Krupp's), Hojel in Holland, and Krtipp in Germany, that
for velocities of over 300 meters per second the law is represented
with great exactness by a straight line. Although experimental re-
sults do not go above velocities of 1,200 meters/sec, I have felt no
hesitation in extrapolating for such values as are here used. In
Fig. I are shown, in Curve i, the values given by Siacci. In Curve
2, the values of the function K(v)^^f(v)/v', and in Curve 3 the
values of K{v) as given by Krupp. It is only fair to say that the
results of the French Commission de Gavre more nearly resemble
Curve 3 than Curve i.

The use of the linear law (originally suggested by Chapel) has
been recommended by the Comte de Sparre in a paper published in
1901.

Fig. 2.

In order to expedite matters, a graphical chart was prepared
(Fig. 2) with abscissas denoting the velocity in meters per second,
containing straight lines and inclinations proportional to the densi-
ties of the air at different heights. On the right is a scale giving
the height above the earth and by following the line corresponding,
the resistance R is read off on the scale at the left.

OF SEVENTY-FIVE-MILE RANGE.

■ 379

The equations used are as follows

(7)

(9)

(II)

Ax

(8) ^v, = -jvf{v)^^,

gAx

(lo) ^^ = -

— AVr

COS Or+l

Vr+U

(12) D, =

The calculation is made as follows : An arbitrary value of x is
selected, the drop Ay is calculated from (7), the change in angle
from (10), the "set-back" from (12), and, finally, the corrected
value of y corresponding to the given x-Dj- is obtained. One row
of values is obtained for each element of the trajectory.

Fig. 3 shows the details of the graphical method. After the
chords are drawn by means of a flexible ruler, the trajectory is
drawn through the vertices of the polygon constructed.

In Fig. 4 are shown some of the most striking results. The
heavy black curve shows the surface of the earth and the change of

380

WEBSTER— BALLISTICS OF A GUN

range, amounting to about a mile, that it would cause (neglecting
the conversions of the verticals).

The method of choosing was frankly one of guessing. At first,
an angle of departure of 45° was assumed with an initial velocity
of 1,300 meters/sec. This gave Curve i. The angle of departure
was then increased to 52°, which gave Curve 2, with a range of 120
kilometers. In order to show the enormous effect of the resistance

1 e.= 4S° ff= cSi^ifM
z e.--5z' ft= CJty)fW

f e.-45' R-o

3 Surface of Earih
X

Fig. 4.

100 Kilometers

of the air. No. 4 is drawn showing the parabolic or vacuum trajec-
tory, with a range of 174.4, and, finally, No. 3, on the assumption
that the air has the constant density found at the surface of the
earth. Further, the trajectory that we should have in case the den-
sity had the constant value, taken at a height two thirds of the
maximum, as suggested by Colonel Ingalls for shots nearer the sur-
face of the earth, was calculated. The range obtained was ap-
proximately 56 kilometers, quite different from the correct value of
Curve I.

Note, May 2, 1919. — The foregoing paper, which was read over
a year ago, has, of course, lost the timeliness that it had at that mo-
ment. In fact, I have been advised by a high ballistic authority not
to publish it, as such calculations are now " a matter of routine."

OF SEVENTY-FIVE-MILE RANGE. 381

To this I have repHed, that, although they may be nozv, they were
not when I read the paper, and even now are not so in the Navy.
In fact, when I inquired of a high naval authority how long it would
take to calculate such a trajectory, he repHed, "About two days."
I said that we did it with two men in an hour.

Since then far better methods have been developed, but as I have
seen only one publication of a trajectory, viz., by Major J. Maitland-
Addison, R. A. (Journal of the Royal Artillery, Vol. XLV., No. 4),
which confirms my results, I publish the paper as read, in the inter-
est of the history of the matter, regretting that more pressing mat-
ters have so delayed the publication.

ON A NEW (?) METHOD IN EXTERIOR BALLISTICS.^

By a. G. WEBSTER and MILDRED ALLEN.
(Read April 20, 1918. Received June 4, 19 19.)

Whether the method to be described deserves to be called new
or not is a matter of opinion, since the instrument here used was
designed and built by one of us twenty-eight years ago, and has
been described in an article in the Physical Review, Volume 6, May-
June, 1898. The method of measuring time by means of the charg-
ing or discharging of a condenser has been used by Pouillet, and has
been applied to ballistics by Cranz in Germany and Sabine in Eng-
land, but so far as we know the electrometer has not been applied
to ballistics.

The essential part of the instrument for measuring the time is
shown in Fig. i. A projectile drops from an electromagnet upon
the two levers shown in the middle, the one on the right being car-
ried on a carriage adjustable by means of a micrometer screw. By
knowing the height through which the projectile falls, the velocity
of the projectile on striking the lever is known.

The arrangement of the apparatus is shown in Fig. 2. The
battery charges the condenser through a resistance, the condenser
being short-circuited through conductor {a, b), which in the calibra-
tion is the left hand lever of the drop interrupter. When the circuit
is broken the condenser begins to charge in accordance with the

equation

_ t
q = qo(i - e -^'^).

A curve of calibration is shown in Fig. 3, by which it appears
that times may be measured of the order of one-millionth of a
second.

In the ballistic application (a, b) and (c, d) are two strips of tin
foil stretched tightly between the brass supports. These are shot

1 Contribution from the Ballistic Institute, Clark University, No. 3.

382

WEBSTER AND ALLEN-EXTERIOR BALLISTICS.

383

Fig. I.

384 WEBSTER AND ALLEN— EXTERIOR BALLISTICS.

Path D« (mllet ej

nrrcngemenr of gpngrdtus

Fig. 2.

IS ..C.3

Fig. 3-

WEBSTER AND ALLEN— EXTERIOR BALLISTICS.

385

away by a bullet. It is possible to measure the velocity of a bullet
when the strips of tin foil are an inch apart, but in general a dis-
tance of two or three feet is conveniently taken. This may be con-
trasted with the distance of one hundred or one hundred and fifty
feet generally used in ballistic laboratories of arms companies.

The table of results is given both for a small saloon rifle and
for a 44 calibre gun. It is not claimed that this method would be
of great use in the open, but in a laboratory it is certainly a very
great convenience.

Saloon Rifle.

.44 Calibre Rifle.

Deflection in Cms.

Time X 248 in Sec.

Velocity in Cms. /Sec.

Velocity in Ft. /Sec.

5.2

•037

30,162

5-4

•039

28.615

1.303.0

5-3

.038

29.368

1.330.2

S-3

.038

29,368 .

1.329-2

5-3

.038

29,368

1,326.4

5-3

.038

29.368

1,284.0

5-3

.038

29.368

S-4

•039

28,615

1.375-7

5-4

•039

28,615

1.327-3

S-i

.036

31.000

1,326.4

S-S

.041

25.363

1.339-8

5-4

•039

28,615

1,320.7

Average velocity (

v)

28,985

1.337-98 av.

PROC. AMER, PHIL. SOC , VOL. LVIII, Y, DEC. 4, I919.

CHARACTERS AND RESTORATION OF THE SAUROPOD
GENUS CAMARASAURUS COPE.^

From Type Material in the Cope Collection in the American
Museum of Natural History.

(Plate I.)
By henry FAIRFIELD OSBORN and CHARLES CRx\IG MOOK.

In 1902 the Cope Collection of Fossil Reptiles was presented to
The American Museum of Natural History by President Morris K.
Jesup. It included all of Cope's types and other dinosaur material of
Morrison age from the vicinity of Canyon City, Colorado. Several
of these types antedated in definition Marsh's types from beds of
similar age. Cope's references were full but accompanied by few
figures ; Marsh's came later and were adequately illustrated. Marsh
also issued in the publications of the United States Geological Sur-
vey two more or less complete summaries of the characters of these
animals, which were fully illustrated and widely distributed. Conse-
quently they became well established in the literature, while Cope's
are still unrecognized and imperfectly known. Our object has been
to describe and determine as fully as possible Cope's types, especially
of the Sauropoda, the most important of which is that of Camara-
saurus. This generic name antedates Morosauriis Marsh, with
which it is considered congeneric, by about one year.

The fugitive descriptions and determinations by Cope, Osborn,
Riggs and Mook may now be replaced by thorough descriptions and
illustrations, in which the characters of the genus Caiiiarasaitrus are
determined in great detail, so far as the nature of the material will
permit. All the type material, including the types of six genera and
eleven species, has been figured and these animals, practically un-
known since their original mention forty years ago, have now been
brought to light.

^ Partly printed in the Proceedings of the Palsontological Society, Bul-
letin of the Geological Society of America.

38G

SAUROPOD GENUS CAMARASAURUS. 387

Occurrence and Collecting of Material.

Original Discovery and Collecting. — In the spring of 1877 Mr.
O. W. Lucas, superintendent of public schools in Canyon City, Colo-
rado, discovered some large fossil bones, which he sent to Professor
Edward D. Cope. The date of this discovery is not definitely known,
but it appears to have been some time in March. From the first
specimens which reached the Cope Museum in Philadelphia, Cope
made his original description of Camarasaurus and founded the
genus ; this description was published August 23, 1877. The name
Camarasaurus, or " chambered saurian," was given in reference to
the cavernous nature of the centra of the cervical and dorsal ver-
tebrae, in connection with the lateral cavities now known as pleuro-
coelia. After receiving the original bones. Cope employed collectors
who gathered more material, all of which is now in the American
Museum.

Subsequent Collecting. — The amount of material collected by
Cope's parties was very large. It was not all prepared at once, but a
considerable amount of it was cleaned up by Jacob Geismar under
Professor Cope's direction. In 1877 a reconstruction of the skeleton
of Camarasaurus was made by Dr. John A. Ryder, under the direc-
tion of Professor Cope. This reconstruction, the first ever made of
a sauropod dinosaur, was natural size and embodied representations
of the remains of a number of individuals; it was over fifty feet in
length. It was exhibited at a meeting of The American Philosophical
Society on December 21, 1877, and since has been exhibited a num-
ber of times at the American Museum (where it is now preserved)
and elsewhere. A greatly reduced copy of it was published by Mook
in 1914.^

After the collecting of the material which formed the basis of the
above-mentioned reconstruction, Cope's collectors sent in more ma-
terial. This collecting was continued until 1880.

The Quarries. — Unfortunately the quarry records of the Cope
Canyon City material have been lost ; no quarry diagrams are men-
tioned in any of Cope's descriptions, and it is unlikely that any were

2 " Notes on Camarasaurus Cope," by Charles C. Mook, Ann. New York
Acad. Sci., Vol. XXIV., pp. 19-22, May 21, 1914, i fig.

388 OSBORN AND MOOK— CHARACTERS AND RESTORATION

made. Two large quarries are known to have existed and their loca-
tion is known at the present time.

One large quarry is situated about 500 yards west to southwest
of a small conical hill, locally known as the " Nipple," a considerable
distance from the edge of the escarpment. This quarry is called
Cope Quarry No. i. Here the Morrison is capped by the Purgatoire
sandstone and the quarry site is very definitely marked by a great
excavation. The matrix is chiefly reddish to brownish, and probably
most of the bones of a reddish color, collectively known as the red
series, came from this quarry.

Another quarry is situated almost at the crest of the escarpment
which forms the west boundary of Garden Park, and near the base
of the " Nipple." It is not very definitely marked, but traces of the
work of excavation by Cope's collectors and others mark its site. This
quarry is called Cope Quarry No. 2. The matrix is largely grayish,
and it is likely that it furnished most of the bones which are known
collectively as the yellow series, although this is not certain. Some
of the matrix is neither gray nor yellow, and it is possible that cer-
tain of the yellow bones may have come from the other quarry. The
value, therefore, of the color of the matrix, in determining the field
association of the bones, is limited. Variation in color depends upon
the condition of the iron oxide of the matrix, and probably also upon
the original conditions of decay of the animal tissue. The quarry
was reworked by Mr. J. B. Hatcher for the Carnegie Museum in 1901.

There may have been one more quarry in this vicinity which per-
haps furnished some of the sauropod material, but the nature and
the location of it are not known ; indeed, the types of Aniphicaelias
alius and A. latits may have come from this quarry, about which no
reliable information is available. All three quarries are located a
short distance north of the quarry worked by Mr. M. P. Felch, later
known as the Marsh-Hatcher quarry, which yielded the genotypes
Diplodociis longiis Marsh and Haplocanthosaurus priscus Hatcher,
also H. iitterbacki Hatcher. The INIarsh-Hatcher quarry was ex-
cavated at a lower geological level than the Cope quarries.

SAUROPOD GENUS CAMARASAURUS, 389

Preparation and Research in the American Museum.

Acquisition. — The Cope Collection of Fossil Reptiles had been
examined in Philadelphia by Dr. W. D. Matthew and was trans-
ferred to the American Museum under his direction. The preparation
of the material was made by ]\Iessrs. Kaison, Charles and Otto
Falkenbach, Lang, Christman, Hoover, Bruckner, Carr and Home.

Research in 1904. — Doctor Matthew went over the material,
under the direction of Professor Osborn, and catalogued and iden-
tified it so far as was possible with the aid of the records available,
distinguishing the material obtained in the earlier collecting in 1877
by Superintendent Lucas from that obtained in the later collecting in
1880 under Mr. Ira H. Lucas. The bones of the earlier collection
were given the number 5760, with variations according to their iden-
tification as individuals, such as 5760' and 5760" ; the bones of the
later collection were given the number 5761, with a modification
into 5761 -a for a presumably different individual than the rest of
5761. Subsequently Professor Osborn and Professor W. K. Gregory
made a further study of the vertebrae and arranged them provision-
ally into series, using in addition to the previous records the color of
the bones, those of the red series apparently having come from a dif-
ferent quarry than those of the yellow series. Most, if not all, of the
red bones probably came from Cope Quarry No. i, and most, if not
all, of the yellow from Cope Quarry No. 2.

In connection with this work, which was carried on in 1904, Mr.
Rudolph Weber, then artist of the Department of Vertebrate Palaeon-
tology, made line drawings of many of the vertebrae. In 1906 some
wash drawings of the skull material were made by Mr. Erwin S.
Christman. These illustrations were originally prepared for the
United States Geological Survey Monograph on the Sauropoda, in
course of preparation by Professor Osborn. The cost of preparation
of these drawings was borne by the Survey.

Research in ipi^-ipip. — In 1912 work on the Cope Sauropoda
material was renewed as part of the preparation of the Sauropoda
Monograph, which was being prepared for the Survey by Professor
Osborn. This work was undertaken by the present junior author
under the direction of the senior author. The entire Cope Collection

390 OSBORN AND MOOK— CHARACTERS AND RESTORATION

of Sauropoda from Canyon City was studied, among other material.
with the object of separating the vertebrae and Hmb bones referable
to Camarasaurns, Amphiccclias and the other Cope genera, and ar-
ranging them in series similar in size, proportion and color, as well
as determining the characters of Camarasaurns and Amphiccclias
and the less known genera. To a considerable extent this work con-
sisted in verification of the previous work by Doctor Matthew and
Professor Gregory, in modification of their results, in a few cases,
and in adding to them to meet the present needs.

This research has terminated in the arrangement of the vertebrse
and ribs in morphological series, which may represent originally dis-
tinct individuals, or may not. The attempt was made to associate
the bones of single individuals so far as practicable, but in many
cases evidence for this was insufficient and in such cases an attempt
was made to assemble series that would be reasonably constituted in
a morp'hological sense. The arrangement of the bones in these series
is as accurate as it could be made, in view of the distorted, sometimes
incomplete and badly mixed character of the material. The pairing
of the girdle and limb bones was similarly undertaken, though no
attempt was made to pair the ribs. In a few cases it has been pos-
sible to determine the relation of some of the girdle and limb bones
with the vertebrae, but in most cases the original association is still
unknown, though their possible association is very evident.

Carnivorous Dinosaur Material and Types. — The type of Epan-
terias amplcxus consists of bones of a large theropod. There are
some ribs among those of Camarasaurns which certainly do not agree
in characters with the majority of camarasaur ribs, and do resemble
those of the Theropoda. There is also a theropodous femur. They
may be provisionally referred to this form. It is possible, if not
probable, that the types of Tichosteus lucasanus and T. ccqnifacies,
also of Symphyrophus musculosns, may be referable to the The-
ropoda. Cope's types of La-laps trihcdrodon, Brachyrophus altar-
kansanus and Hypsirophus discurus were also collected at this local-
ity. The first of these is certainly a theropod ; the position of the
secotid and of the third is uncertain.

Characters of the Genus Caniarasaurus. — The results of the in-
vestigations described above, include determinations of the generic

SAUROPOD GENUS CAMARASAURUS. , 391

characters of Camarasaurus and Amphiccelias, so far as these char-
acters are determinable from the material in the collection. The
genus Camarasaurus is characterized by massive proportions.
Throughout the skeleton, with the single exception of the ischium,
the bones are stoutly constructed.

Synonymy of Camarasaurus Cope and Morosaurus Marsh. — In
1898 the synonymy of Morosaurus Marsh with Camarasaurus Cope
was suggested by Osborn f in 1902 this view was favored by Riggs ;*
in 1914 it was definitely adopted by Mook.^ At present Morosaurus
is considered to be a synonym of Camarasaurus, Cope's term having
priority and therefore being valid.

Characters of the Genus Amphica^lias. — Amphicoelias is more
slender than Camarasaurus; its remains resemble those of Diplo-
docus, but are somewhat larger than any known Dlplodocus, and are
somewhat more strongly constructed.

Restoration and Reconstruction of Camarasaurus.

Ryder's Restorations. — It would be hardly justice to the very able
comparative anatomist. Dr. John A. Ryder, to publish, without ex-
planation, his reconstruction (Fig. i), roughly drawn, life size, and
exhibited before The American Philosophical Society December 21,

1877.

The reconstruction was obviously made after one series of bones
was exposed, but before Professor Cope had had time to give them
much study. It would not appear that Professor Cope himself
seriously studied the reconstruction, from the false arrangement of
the teeth on the malar jugal arch, and from the placing of consoli-
dated spines like those of the sacrum opposite the massive scapula.
Twelve to thirteen vertebrae are consigned to the neck, close to the

3 " Additional Characters of the Great Herbivorous Dinosaur Camara-
saurus," Henry Fairfield Osborn, Bull. Amcr. Mus. Nat. Hist., Vol. X., Art.
xii, June 4, 1898, pp. 219-233.

* " The Fore Leg and Pectoral Girdle of Morosaurus; with a Note on the
Genus Camarosaurus," Elmer S. Riggs, Field Col. Mus. Pub. 63, Gcol. Ser.,
Vol. I., No. 10, pp. 275-281, Pis. XL., XLL, XLH., October, 1901.

^ " Notes on Camarasaurus Cope," Charles C. Mook, Ann. New York
Acad. Sci., Vol. XXIV., pp. ig-22. May 21, 1914, i fig.

o O O Ck

u. „ OJ *-
a; " ^ O

•c^

<

H

u t^

C in

Q

H

m ^ o

CJ

yi;

.ti

!S

c

^

CS

C

o
u

0)

rt

in

s

_N

"*~*

3

IS

1-.
3

(J

.S

tu
tn

3

OS

o

00

U

rt

<u

^

5=

h-"

e

bO

<

Q

^^

OJ

CL>

g

c

-a

G

tu

"rt

Q

o

|5)

c
o

>

c

O

.d~

en

JJ

<u

E

a

rt

OJ

H

^

^

^

tn

dJ

(U

OJ

c

en

N

^

n

c

'in

-*-'

o

cu

o

U Ph

•^

rt

c

Q

rt

oj

5

o

_<-;

u,

03

u

W

3

<u

C

3

u

13

"i

JH

O

oi

OJ

•S

o

C

o
o
<u

u

"o

£

13

Ph

.S

rt

rC

M-l
O

>,

c

E

.—

O

o

C

'o

t/l

OJ

o

o

t-H

C/2

ffi

o

u

'U

O

u,

oi

OS

3

U^

-5

o

OJ

J

o

^

in

SAUROPOD GENUS CAMARASAURUS. 393

true number. Eighteen vertebrae are consigned to the back, eight
too many. Fifty-seven vertebrae are consigned to the tail, not far
from the correct number. A complete set of claws is consigned to
both the fore and the hind feet.

Still Unknown Parts of Camarasaurus.

1. In the accompanying restorations and reconstructions of Ca-
marasaurus it is observed that our knowledge of this animal is still
very incomplete regarding the structure of the pes, the foot bones
being based on our knowledge of the pes of Apatosaurus or Bronto-
saurus.

2. It is also unknown or uncertain whether these animals pos-
sessed a set of abdominal ribs. One rib has been found which may
possibly represent a member of the abdominal rib series.

Fig. 2. Restoration of the external appearance of Camarasaurus. Pho-
tograph of a model of Camarasaurus made under the direction of Prof. W.
K. Gregory by Mr. E. S. Christman with the cooperation of the authors.
This model is the result of very careful comparison of the Camarasaurus
skeleton with previous sauropod models and with various living reptiles ; it
embodies extensive myological studies by Professor Gregory, and a thorough
analysis of the characters of the skeleton.

3. The vertebral formula is still uncertain, although it is probable
that the number was as follows : cervicals 13 ; rib-bearing dorsals 10,
dorsosacral i ; consolidated primary sacrals 3 ; caudosacral i ;
caudals 53.

394 OSBORN AND MOOK— CHARACTERS AND RESTORATION

Present Reconstructions and Restorations.

The accompanying illustrations represent both a reconstruction
and a restoration, based upon the topotype materials of four or more
individuals of Camarasauriis. The skull characters are partly based
upon the Camarasauriis topotype skull bones and largely upon the
nearly complete skull originally referred to Morosaurus by Osborn ;
it is probable but not certain that this reference is correct.

It is very important to note that the four chief skeletons and two
of the remaining skeletons are of approximately the same linea/r
measurements, but differ slightly in proportions.

Chief Proportions of the Skeleton.

1. The enlargement and elevation of the shoulder above the rela-
tively depressed and reduced sacropelvic region is one of the most
surprising features of this reconstruction. This proportion cannot
be considered as actually demonstrated, because only two complete
ilia were found with the four skeletons, and they belong to the same
individual. It is possible that these ilia and other bones of the pelvis
represent another individual than those individuals represented by
the massive scapulocoracoid bones. The other pubes and ischia are
of the same size as in the pelvis figured in the restoration.

This makes the highest point in the backbone between the shoul-
ders, as contrasted with the previous restorations of Brontosaurus
and Diplodocus by Osborn, in which the pelvic region is made the
high point. Recent studies of Apatosaurus by Osborn and Gregory
make the shoulders higher than the sacrum.

2. The neck is rather massive and the neck and shoulders bore
the chief weight. The heaviest portion of the chest was at the fourth
rib, where the relatively immobile hyposphen-hypantrum articulation
of the dorsal vertebrae begins.

3. As shown in the articulated skeleton (Fig. 3) the total length
and height of the animal, with the spinal curvatures and in the walk-
ing pose, were as follows :

Elevation of tlie head above the ground 20 to 22 feet

Total length from the front of the head to tlie tip of the tail . . 50 to 52 feet
Height of the backbone at the shoulder 13 to 14 feet

SAUROPOD GENUS CAMARASAURUS.

395

Fig. 3. Three poses of Cainarasaiinis. A, walking pose, head elevated.
B, walking pose, head and neck horizontal. C, ground-feeding pose, head
depressed. These photographs are made from drawings in which every bone
is figured separately to a one-fifth scale and fastened upon a black back-
ground. This articulated drawing is successively placed in three poses. A,
B and C, each photographed in turn. In the restoration and reconstruction
the structure and arrangement of the phalanges of the feet are purely con-
jectural. The head of the animal attains a height of 20 feet {A^. The
maximum length in the extended position {B) is 52 feet. Length of back-
bone, without curvature, between 54 and 55 feet.

396 OSBORN AND MOOK— CAMARASAURUS.

4. The head of the Sauropoda, as pointed out by Dr. W. J. Hol-
land, has been mistakenly represented as parallel with the long axis
of the neck, whereas it should be flexed or at an angle with the neck.
In this respect our restoration gives a somewhat misleading impres-
sion, as the head should be bent down as in PL I, B^ C.

Comparison with Other Sauropoda.

As compared with the two other Sauropoda, in which the skele-
ton is now fully known, namely, Diplodocus and Apatosaurus, Ca-
fuarasaiints is relatively the most massive, the most elevated at the
shoulder, the most elongated over all, and the most ponderous in its
proportions. It was apparently not provided with the whip-like
terminal tail vertebrae so characteristic of Apatosaurus and Diplo-
docus; the vertebrae steadily lessen in longitudinal diameter and
would indicate that the tail came to a rather abrupt point.

External Appearance of Camarasaurus.

The external appearance of the head is shown in PI. I. The
head is extremely short and deep in its proportions, contrasting with
the elongate head of Diplodocus. The animal as a whole is sketched
in Fig. 2. This represents an animal terrestrial in gait but adapted
to an amphibious life, with smooth rounded limbs.

'i^ROGLOGLANIS^ PATTERSONI A NEW BLIND FISH
FROM SAN ANTONIO, TEXAS.^

By carl H. EIGENMANN.
(Read October j, 1919.)

Professor J. T. Patterson, of the University of Texas, has se-
cured a specimen of a small blind catfish from an artesian well in
San Antonio, Texas, belonging to Mr. George W. Brackenridge.
Pending the securing of other material the following facts may be of
interest.

The specimen is without pigment. There is no external evidence
of any vestige of an eye. It has a total length of 85 mm. Other and
larger specimens were emitted but not preserved.

The occurrence of blind fishes in Texas was predicable. There
are large springs, the outlets of underground rivers in the same
region and artesian wells tap the subterranean waters in various
places about San Marcos and San Antonio. The flow of the artesian
well of the Bureau of Fisheries at San Marcos shows that the under-
ground waters have an abundant cave fauna. From this well and
some neighboring caves I secured twenty (20) species of inverte-
brates and the blind salamander Typhlomolgc in less than a week's
stay. The surprise therefore is not that cave fishes have been secured
from the underground rivers, but that they have not been found
before. It is more of a surprise that the fish should be a catfish
rather than a member of the blind-fish family of Amblyopsidcr, found
in Tennessee, Arkansas and northward.

However, the occurrence of blind catfishes somewhere in the
Mississippi basin was also predicable. Some of the catfishes are
nocturnal in habit and live in crevices, under rocks, stumps and such,
and detect their food by means of touch and taste organs scattered

1 rpicyXr], 17 = cave ; y\ar]is, rj =r catfish, originally from Glanis, the name
of a river.

2 Contribution from the Zoological Laboratory of Indiana University,
No. 167.

397

398

EIGENMANN— TROGLOGLANIS PATTERSONI.

over their entire body and especially over their barbels. All of these
facts predispose toward a cave, existence, and various catfishes have
become blind in different parts of the world. I have found normal-
eyed catfishes in the caves of both Indiana and Kentucky. Cope
secured blind catfishes, dark in color, Gron'as nigrilahrls from east-
ern Pennsylvania. I have recently called attention to Typhlohagrus
kronei Ribeiro^ from the Cavernas das Areiras, Iporanga, Sao Paulo,
Brazil and more recently to Phreatobius cisternarnm Goeldi,* the
blind catfish living on the Island of Marajo. These belong to the
Pimelodinae, a subfamily of catfishes not found in North America.

/a..

Fig. I. Trogloglanis Pattersoiii Eigenmann. Type.

Gronias nigrilabris, from eastern Pennsylvania, is without ques-
tion a derivative of the universally distributed Ameiurns of the east-
ern and central United States. Typhlohagrus, from southeastern
Brazil, is derived from Pimelodella, a genus widely distributed in
South America, a member of the Pimelodinae. Phreatobius is more
remotely related to Hcptaptcrus, another but very difl:'erent member
of the Pimelodinae.

The new Texan blind fish, judging by the position of its dorsal
and ventral fins, as well as by its adipose fin, is derived from a fish-
like Schilbcodcs, a genus of catfishes with nearly a dozen species,
generally distributed from the St. Lawrence to Texas. Schilbcodcs

^Memoirs Carnegie Museum, VII., p. 255, Plate XXXIV., April, 1917.
4 L, c, p. 372, Plate LVL, 1918.

EIGENMANN— TROGLOGLANIS PATTERSONI.

399

noctiirnus has been found in the springs (mouth of the underground
river) at San Marcos and at various other places in Texas and is the
only species of the genus recorded from Texas. Schilheodes noc-
ttirhiis is, however, not closely related to the blind catfish living in
the underground rivers of the same region.

I have called attention to the fact that the species living in the
caves of the south are more intimately adapted to their subterranean
home than those of the north. The eyes of the Texan Typhlomolge
are more degenerate than those of the salamanders of Missouri.
Judging from external appearance the eyes of the Texan Troglo-
glanis are more degenerate than those of any of the blind fishes from
farther north. The technical description of Trogloglanis follows :

Fig. 2. Outlines of the head as seen from above.

Head similar to that of a tadpole, as broad as long; mouth in-
ferior; teeth?; adipose fin long and low, rounded posteriorly, con-
nected at its base with the accessory caudal rays ; no external evi-
dence of eyes; distance between origin of dorsal and tip of snout
half as great as origin of dorsal from the end of the adipose; dis-
tance between snout and origin of ventrals i% in the distance between
origin of ventrals and base of middle caudal rays; pectoral spine
strong and pointed, about two thirds as long as the longest ray, about
equal to the length of the head behind the posterior nares, smooth in

400 EIGENMANN— TROGLOGLANIS PATTERSONI.

front, its posterior margin with seven straight teeth, less than half
the width of the spine ; caudal truncate, with numerous accessory
rays ; dorsal spine equal to the pectoral spine ; base of adipose fin
equal to the predorsal area ; anal but slightly rounded, its highest ray
equal to the length of the head. Nasal barbel reaching very nearly
to end of opercle, maxillary barbel to the pectoral spine, mental bar-
bels a little beyond the edge of the gill-opening.

Head 4.4 in the length ; depth 6 ; D. I. 7 ; A. 14 ; P. I. 9 ; V. 8.

The specimen was collected by Mr. G. W. Brackenridge of San
Antonio, Texas. Mr. Brackenridge gave the specimen to Professor
Patterson who sent it to the author for determination. It is cata-
logued as No. 15240 Indiana University Museum.

MAGELLANIC PREMIU

Founded in 1786 by John Hyacinth de Magellan, of London

1920

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful Knowledge

ANNOUNCES THAT IN

DECEMBER, 1920

IT Viri.1. AWARD ITS

MAGELLANIC GOLD MEDAL

TO THE author of THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING TO NAVIGATION, ASTRONOMY, OR NATURAL PHILOSOPHY (mERH
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 1920, deliver free of postage
or other charges, his discovery, invention or improvement, addressed to the
President of the American, Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance by some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the sam« motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author.

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publicly
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as may be after the adjudication, either in aseparate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

kU correspondence in relation hereto should be addressed

To THE Secretaries of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A,

Proceedings

OF THE

American Philosophical Society

Subscription — Three Dollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

Price, One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

• HELD AT PHILADELPHIA

For Promoting Useful Knowledge

New Series, Vol. XXII,
Part III, 4to, 44 pages.
(Lately Published)
Tertiary Vertebrate Faunas of the North Coalinga Region of Cali-
fornia. A Contribution to the Study of the Palaeontologic
Correlation in the Great Basin and Pacific Coast
Provinces. By John C. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California.

Subscription — Five Dollars per Volume
Separate parts are not sold

(Address The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILAnFLPHIA, U. S. A.

^;^j PROCEEDINGS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

FOR PROMOTING USEFUL KNOWLEDGE
Vol. LVIII. 1919. No. 7.

CONTENTS

PAGB

Polarized Light in the Study of Ores and Metals. By Fred E. Wright 401

On the Luciopimelodinae, a New Subfamily of the South American

Siluridje. By Chas. S. Driver ....... 448

MiNUTBS . . . . . . . . . . . . i

Index , . . . . ' . . . . . . , . xv

PHILADELPHIA

THE AMERICAN PHILOSOPHICAL SOCIETY
/ 104 South Fifth Street

1919

The American Philosophical Society

ainnoupws.that an award will be made in the year 1921 of the

HENRY M. PHILLIPS PRIZE

The subject upon which essays are to be submitted is

The Control of the Foreign Relations of the United States :
The Relative Rights, Duties, and Responsibilities of the Presi-
dent, of the Senate and the House, and of the Judiciary, in
Theory and in Practice.

The Essay shall contain not more than one hundred thousand words,
exclusive of notes, and must be in the possession of the Society on or
before December 31, 1920.

The Prize for the crowned essay will be Two Thousand Dollars, in
gold coin of the United States, to be paid as soon as may be after the
award.

Attention is called to the following regulations governing the award
of the Prize:

Competitors for the prize shall affix to their essays some motto or name
(not the proper name of the author, however) and when the essay is forwarded
to the Society it shall be accompanied by a sealed envelope containing within
the proper name and address of the author and on the outside thereof the
motto or name adopted for the essay.

At a stated meeting of the Society in pursuance of the advertisement, all
essays received up to that time shall be referred to a Committe of Judges, to
consist of five persons, who shall be selected by the Society from nominations
made by the Committee on the Henry M. Phillips Prize.

Essays may be written in any language, but, if not in English, must be
accompanied by an English translation.

No essay which has been already published or printed, or for which the
author has received any prize or profit of any nature whatsoever, shall be
accepted in competition for the prize.

Essays must be typewritten on only one side of the paper, and six copies
must be furnished by their respective authors for the use of the Committee
of Judges.

The literary property of such essays shall be in their author, subject to
the right of the Society to publish the crowned essay in its "Transactions "
or "Proceedings."

The Society reserves the right not to award the prize if none of the com-
peting essays is deemed worthy of it.

John Bassett Moore,
David Jayne Hill,
Simeon E. Baldwin,
John Cadwalader,
W. W. Keen,
William B. Scott,
President ex-officio.

Committee.

The essays must be sent, addressed to the
President of the American Philosophical
Society. No. 104 South Fifth Street, Phila-
delphia, U. S. A.

POLARIZED LIGHT IN THE STUDY OF ORES AND

METALS.i

By FRED. E. WRIGHT.
(Read April 14, IQ17.)

The measurement of the optical properties of transparent min-
erals, even in minute, irregular grains, is a simple task with modern
petrographic microscope methods and is accomplished by petrolo-
gists as part of ordinary routine work. But the determination of
the optical constants of opaque substances is difficult and is rarely
attempted by microscopists ; all observations are necessarily made in
reflected light and are restricted commonly to the determination of
color, of the character of crystallization, and of the behavior of the
mineral or metal plate toward reagents and abrasives. It is gen-
erally recognized that if methods were available by means of which
the optical constants of opaque substances in fine particles could be
readily ascertained these methods would be of great value not only
to students of ores and opaque minerals, but especially to metal-

1 The manuscript of this paper was finished in March, 1917, and is here
presented without alteration. A brief resume of the resuUs of the investiga-
tion was given at the meeting of the American Philosophical Society in April,
1917. With our entrance into the war the writer joined the Army, and the
publication of this paper was accordingly postponed.

In the theoretical section of this paper certain standard equations are
derived and expressed in Cartesian coordinates. The expressions would have
been much simpler and shorter had the methods of vector anatysis been em-
ployed; but this was not done and the equations are developed' in the usual
notation in order that they may be easily accessible to the reader interested
in this particular subject. Many of the problems of crystal optics are, how-
ever, essentially vectorial in character and yield most readily, as do many
problems in electricity (alternating currents, wireless telegraphy), to treat-
ment by vector analysis. In vector analysis the imaginary quantity f -^ V — i
is treated simply as an operator rotating a vector through 90°. This greatly
simplifies the interpretation of equations containing complex quantities. An
interpretation of this kind of many of the equations in the present paper
would undoubtedly render them more intelligible, but this would have greatly
increased the length of the paper and was accordingly not attempted.

PROC. AMER. PHIL. SOC, VOL. LVIII, Z, JAN. 21, I92O

401

402 WRIGHT— POLARIZED LIGHT IN THE

lographers in the study of metal alloys. Unfortunately the explana-
tion of the phenomena presented by opaque bodies is somewhat in-
volved and the experimental measurement of the optical constants of
such substances is encumbered with difficulties which, in many cases,
allow only approximate results to be obtained at best. In view of
these facts it has seemed to the writer that a useful purpose may be
served by a critical discussion, based on the commonly accepted ver-
sion of the electromagnetic theory of light for absorbing bodies, not
only of the phenomena which might be of value for diagnostic pur-
poses, but also of the factors underlying the several methods, old
and new, which may be applied to the measurement of the few deter-
minable optical properties of opaque bodies. A proper appreciation
of the possibilities and also of the limitations of the application of
polarized light to the study of opaque substances can only be had by
a proper understanding of the fundamental principles involved and
of their relative significance in practical diagnosis.

In the theoretical treatment of the general problem emphasis will
be placed on the phenomena resulting on perpendicular reflection
(angle of incidence = o), because most metallographic observations
with the microscope are made under conditions of vertical illumina-
tion. The attempt will be made to present the fundamentals of the
subject and to indicate the mode of derivation of the equations re-
quired. Although several new relations are given, the treatment as a
whole is necessarily along lines which have been followed by others.
In the preparation of the section on theory the following books and
articles have been specially consulted : P. Drude in Winkelmann's
" Handbuch der Physik," Vol. VI, and in Annalen der Physik ;
" Lehrbuch der Optik," by P. Drude ; " Lehrbuch der Kristall-
physik," by F. Pockels ; " Physical Optics," by A. Schuster ; " The
Analytical Theory of Light," by J. Walker ; and " Physical Optics,"
by R. W. Wood.

The results of the investigation may be summarized by stating
that in general the optical constants, such as refractive indices and
absorption indices, cannot be satisfactorily ascertained on small pol-
ished random sections ; that the application of polarized light en-
ables the observer ordinarily to determine whether the crystal plate
:- 'Isotropic or anisotropic and also to ascertain the degree of aniso-

STUDY OF ORES AND METALS. 403

tropism ; that for this determination methods based either on the
contrast in intensity of the two reflected components or on the
amount of rotation of the plane of polarization on reflection of the
incident plane-polarized light may be employed ; that methods based
on the phase difference between the two reflected components are
in general of little value because of the small differences in phase
which ordinarily result for a relatively large change in birefringence
or biabsorption.

Several new methods are described for detecting anisotropism in
opaque substances ; of these, that requiring the use of the bi-quartz-
wedge-plate is the simplest and has been found in practice to be
superior to any heretofore suggested.

Theoretical.

Light waves on passing through homogeneous material are ab-
sorbed to a greater or less degree. In transparent media the amount
of absorption (in the visible spectrum) is relatively slight and can
be neglected for most purposes ; but in absorbing media there is an
appreciable weakening in intensity even in thin plates, which gives
rise to the phenomena of absorption. No body is, however, per-
fectly opaque (perfect absorber) or perfectly transparent, and the
terms transparent and absorbing are relative terms which express,
in a general way, the degree of absorption in the visible spectrum.
The conditions may, of course, be reversed in the infra-red or ultra-
violet. The experimental law of absorption, as expressed by Lam-
bert, states that in a homogeneous medium each layer of equal thick-
ness absorbs an equal fraction of the light transmitted ; if the layers
be considered to be one molecule deep, as in some crystals, then each
layer absorbs the same percentage of the light which passes through
it. In other words, for a layer of given thickness the intensity of
light transmitted (total light less absorbed light) is proportional to
the intensity of the incident light or

f = /. I = /..-'. (I)

where /q^ intensity of incident light, I, intensity of transmitted
light, /, thickness of layer, and m, the absorption modulus {i.e., the

404 WRIGHT— POLARIZED LIGHT IN THE

density of a layer of unit thickness^) ; since the intensity varies with
the square of the ampHtude of vibration of the transmitted Hght, the
equation is vahd

A = Aoe ^° ' = Aoe ^ "' , (2)

in which A^ is the amphtude in vacuo. A, amplitude in the medium
after passage through thickness I, X, wave-length in material, Aq,
wave-length in vacuo, k, the absorption coefificient, k, the absorption
index. But

therefore

/\ I - — ■<./ --—-kj

=— = gA =g^0 = g-ml. ^2a)

k = HK and m =

Xo X

The significance of the absorption index can best be realized by a
transformation of equation (2a)

luy = — l-n- K
I Ao

or expressed in ordinary logarithms

Jo 47r- 043429
logio^ = r l-n-K.

1 Ao

- The significance of these terms is clearly expressed in photography in
which the relative densities of the photographic plate serve to record more
or less accurately the relative intensities of the light impinging upon it. The
incident light produces changes of such nature in the silver bromide particles
in the emulsion that after development the exposed particles remain as opaque
specks of silver. The more intense the incident light, the greater the number
of affected silver particles. In the developed plate the more silver specks
there are per given area, the more light is stopped, and the more opaque and
dense is the plate. Density is measured in number of silver particles per
unit area (total weight of silver particles per unit area). If a layer is of
such density that it transmits the fraction a of the incident light, then two
such layers superimposed transmit aa or a~ of the light, three layers aaa or
o^; thus the density increases in arithmetical, but the transmission decreases
in geometrical progression. If ///o = percentage of light transmitted (=:f'''"
equation (i)) the reciprocal, h/I is the opacity 0 = ^'"' and the density
Z) = log opacity or ml or simply m for /= i. A plate is said to have unit
density Z? =: i = log opacity = log /o//=:log 10, i.e., when it transmits only
Mo of the incident light. Similarly a crystal plate of absorption-modulus i
and of unit thickness transmits only Mo of the incident monochromatic light.

STUDY OF ORES AND METALS.

405

For a wave length Ao = . 0005458 mm. (near the mercury green hne)
this expression reduces to

1 ^°

(2b)

V

1

6

1

5

\v

w

\\

w

\

3

\\

V

.\

Vo

v\

\

'ii

^v

\

\

K

<;

^

^

\

\

\

\J

X

^

^

■---

V

\ >

'c

h

^

\

^

:;::::::;

t-^

__

T

• c

^4'

vi

N
^

"^

—

^

^

^SE

=

mrr

=:=

0

K-

^~ —

2

—

3

4

5

B

7

s

9

.1.

0

Fig. I. In this figure the curves are equi-intensity curves indicating the
ratio of the intensity of the transmitted light to that of the (normally) inci-
dent light for a plate o.ooi mm. thick, for greenish yellow hght (^o = 0.0005458
mm.), and for different refractive indices (ordinates) and absorption indices
(abscissae). The pronounced effect of even a small absorption index in cut-
ting down the transmission is evident from the curves.

A series of vakies computed by means of this equation on the
basis of thickness / = 0.001 mm. (i micron about %o the thickness of
an ordinary rock thin-section) are Hsted in Table I. and represented

406

WRIGHT— POLARIZED LIGHT IN THE

graphically in Fig. i. From these the pronounced effect of k in cut-
ting out and absorbing the transmitted light energy is clearly evident.
For any given thickness /, refractive index n, and absorption index k,
the amount of light absorbed can be obtained directly. From Fig. i
it is evident that for a plate only i micron thick and for k > 0.5 the
amount of transmitted light is almost negligible and the plate is prac-
tically opaque. The effect of using a plate %o as thick is the same
as that produced by a plate of the original thickness but with an
absorption index %o as large. Thus a plate of refractive index
n = 2, thickness / = 0.001 mm. and absorption index k = o.5 trans-
mits 10'" of the incident light while a plate of the same refractive
and absorption indices but ^0 as thick ( 0.000 1 mm.) transmits lO"^ of
the incident light. If we assume an intensity of illumination of 1,000
meter-candles and a threshold limit of vision o.ooi meter-candles,
then 10"*^ is the least amount of light which can be detected; under
these conditions for an ordinary rock section of thickness 0.02 mm.,
equation (2b) reduces to n.K = o.OT,; therefore a mineral in a thin
rock section whose absorption index k exceeds 0.02, will appear per-
fectly opaque. The influence of even a very small absorption index
is therefore exceedingly great in absorbing light energy.

TABLE I.

In this table are listed the absorption indices k of absorbing crystal plates
of thickness / = o.ooi mm. and refractive index n Avhich transmit given quan-
tities ///o of the incident light. Thus a plate of this thickness, of refractive
index n = 2.0 and absorption index k = 0.030 transmits 0.25 of the incident
light.

///o

0.8

0.67

0.50

0.25

lo-i

10-2

lo'S

10-4

io~5

10-6

JO-'

lo-s

io-»

10-10

n
I.O

.010

.018

.030

.060

.100

.200

.300

.400

.500

.600

.700

.800

.900

1. 000

2.0

.005

.009

.015

.030

.050

.100

.150

.200

.250

.300

■350

.400

•450

.500

3-0

.003

.006

.010

.020

•033

.067

.100

■133

.167

.200

•233

.267

.300

.333

4.0

.002

.004

.008

•015

.025

.050

•075

.100

.125

.150

•175

.200

.225

.250

.S.o

.002

.004

.006

.012

.020

.040

.060

.080

.100

.120

.140

.160

.180

.200

7-5

.001

.002

.004

.008

.013

.027

.040

•053

.067

.080

•093

.107

.120

•133

lO.O

.001

.002

.003

.006

.010

.020

.030

.040

.050

.060

.070

.080

.090

.100

The characteristic feature of waves traversing an absorbing medium is
the decrease in their vibration amplitude with distance of penetration; the
waves are damped; the result is that for obliquely incident waves the vibra-
tion amplitudes along each wave front are not constant. In other words, the

STUDY OF ORES AND METALS. 407

surfaces of equal phase do not coincide with the surfaces of equal amplitude
as in transparent media. In the electromagnetic theory transparent media are
grouped under dielectrics (electric non-conductors) and in these Maxwell
showed the possibility of the development of an electric current under the
induence of an electric force, the current being proportional to the rate of
change of the electric force; the current results on the shift in the density
of the lines of electric force and varies in intensity with the rate of their
displacement. If E is the electric force, the " displacement " current is pro-
portional to dEjdt and is in fact equal to e-dEj^ir-dt where e is the dielectric
constant. A current is surrounded by a magnetic field such that the lines of
magnetic force are closed curves. Maxwell showed that the work done in
carrying an isolated magnetic pole around one of these curves is equal to 4""
times the electric current measured in electromagnetic units. Similarly a
m_agnetic current or flux varies when the strength of the magnetic field
changes and the lines of flow are surrounded by an electric field such that
the lines of electric force are closed curves ; the line integral of the electric
force around one of these curves is numerically equal to 47r times the mag-
netic flux; furthermore the magnetic flux is proportional to the rate of change
of the magnetic force or dM/dt and is in fact equal to n-dM/^Tr-df, in which
/^ is the magnetic permeability. For vibrations of the high frequency of light
waves M is practically equal to unity. From these relations Maxwell deduced
the following differential equations :

(3a)

(3^*)

in which X, Y, Z and u, v, w are components of the electric and magnetic
forces respectively and c is the ratio of the electrostatic to the electromag-
netic systems of units and is numerically equal to the velocity of light.

To account for the phenomena observed in absorbing media the above
equations are modified according to one of two possible hypotheses both of
which involve the movements of electrons; the movements thus set up (trans-
lation in conductors, vibration in selective absorbers) absorb or divert part
of the electromagnetic energy and convert it into heat (ohmic heating) and
reduce the amount of energy available for the light vibrations. In electric
conductors an electric force sets in motion streams of negatively charged
electrons which are in effect the conduction current. The reaction is directly
proportional to the impelling electric force, E, or to (^-E in which <t is the
absolute conductivity in electrostatic units (Ohm's law). A medium which
is not a perfect insulator exhibits properties interjnediate between those of
a dielectric and of a conductor and the current consists of two parts : a dis-
placement current and a conduction current. The expression for the cur-
rent which covers both cases is then

« ^^ I c f \

e dX _ dw dv
c'df ~ dy~ dz'

€ d Y du dw
c ~dF ~ Jz ~ dx '

e dZ _ dv du
C dt dx dy '

I du _ dY dZ
c dt ~ ~dz dy '

I dv _dZ dX
C di dx dz '

1 dw _dX dY
c dt ~ dy dx

408 WRIGHT— POLARIZED LIGHT IN THE

In light-waves the disturbance is periodic in effect and the time occurs

onl}'- in the form of a factor the most general expression for which is Ae^ t ;
in this factor T is the period of vibration and A, which may be real, imagi-
nary, or complex, is the amplitude vector. Under these conditions the ex-
pression (4) for the current may be written

— (€ - 2T(n)-—.
47r at

The only difference between this expression for the current and that for
dielectrics is the replacement of the dielectric constant e by the complex quan-
tity (e — zTcri), in which z = V — i.

In crystals the dielectric constant and the conductivity are different in
different directions. Experience has shown that the components of the elec-
tric displacement for any direction of wave propagation are homogeneous,
linear functions of the field components. Thus the X, Y , Z components of
the conduction current are

those of the displacement current

47r \

ax , ay , az\

wherein omc = o-^ft and ^lik = ^/.-ft. The first Maxwell equation becomes for
absorbing crystals

or if Fftfc be substituted for the complex expression (cftt — 2T<rhki) the Max-
well equations can be written

I / - dX , _ dV , . dZ\ dw dv
C \ 6t dt dt ) dy dz

I /- ax , . aF , . az\ du dw , ,

-{e-n-^^ .22^ + .23^ j = ^- ^ = ., ^ (5a)

I /. ax , . aF , . az\ dv du

- I ^31 -rr + 632-^7 +^33-^ I = -7- = f ,

C\dt Bl dt / dx dy

iaM_aF_az i^_^_^ i^-^_^ ( h)

c dt ~ 'dz ay ' c dt~ dx dz ' c lit ~ dy dx ' ^^ ^

wherein eitk = fhh.

If for abbreviation the right hand side of equations (50) be made equal
respectively to ^, v, f, the differential dXjdt, BY/dt, dZjdt can be expressed

STUDY OF ORES AND METALS.

409

in the form

— - = - (au^ + ai2V + aisH,
at c

dY I _

-— - = - (a21? + 022?? + a23f),

at c

r\7 T

-rr = - (flsi? + 0321? + Osaf),
at c

(6)

in which the complex constants (called polarization constants) ahk = akh-\-ibhk
are simple determinate functions of c^ and of the complex quantities ^n . . .
^22 . . . ; in these equations dhk = dkh.

To eliminate the components of the electric force X, Y, Z differentiate
equations (5&) after the time

c dt^

dz\dt ) dy\dt ) ' c dt^ dx \dl ) dz\ dt J ' ^'^'

I d^u _ d /dX\ d /dY\
c dt^ ~ dy\ dt ) dy\dt)'

and substitute in these equations the values of dX/dt, dY/dt, dZ/dt from
equation (6), and obtain the equations

t;^ = -r- (a2i5 + 0227? + a23f) — -r- (OsiS + azirj + CssD.
dr ax ay

d'^v d . . , , \ d , . , , .

TTi ~ T' ^^31^ + 0327? + ossf) — -T-{auk + Oi2»7 + aisD,
dr ax oz

^7;r = T- (^11^ + 0,1171 + cijf) - — (a2i$ + a22»7 + 023?),
di- dv ax

(8)

which are free from the components of the electric force and of the electric
current.

A particular solution of these equations is

u = mp, V — flip, w = H3p,
~ ' T A ''1^ + v^y + V3Z

p = Ae

{'-

).

(9)

in which Mi, ^2, /"s are the direction cosines and A (complex) the amplitude
of the polarization vector ; v^^ v„, v^ the direction cosines of the wave normal
and g" (complex) the velocity. If in (9) we assume vertical incidence, then

^j = j,2 = o, i-s = I ; if we put = = , and X = T-q, wherein q, k, and X are

real, then

u = niAe

-2nKli2^(-j.-l)

(10)

This equation states that the amplitude of the wave of light after passage
through the path X (one wave length) has decreased to the extent of f"^"^'";
K is therefore called the absorption index.

410 WRIGHT— POLARIZED LIGHT IN THE

For a beam of light entering a second medium the axes of reference may
be so chosen that the Z-axis is normal to the boundary surface while the
plane of incidence is the XZ plane. In this case ''i = sin i, i-, =r 0/3 = cos i,
if the positive Z-axis points into the second medium and i is the angle of
incidence. In adopting this convention and substituting the values of (9) in
(8 and 5a) we obtain, on eliminating /Wj, /x,, /x, the equation

(q2 — fl_) (g2 — Q^^ cQs2 { — fl^^ sin2 i + 2031 sin i-cos i)

,~ . . - .^, ) (11)

^(OosSmi — ai2 cos

0=.}

In this equation 'q and the constants an- • • are complex quantities. In similar
manner we find, by observing that the wave normal and the polarization vec-
tor are at right angles, the expression

t^ii'q^ — ojjo) ^ ^^(0^.., cos ! sin I — "a.nCOi-i). (12)

If the polarization plane include an angle 5 with the plane of incidence
(i"i = — cos I -cos 5, Mi = sin 5, /x, = sin t cos 5) equation (12) can be written

<.^„ ; g^ — fl22 012 cos i — 023 sin i . .

tan o = _ ; ; — . = — z 7r~- ; . : . (12a)

ai2 cos I — ao2 sm t ^ — on cos^ t — a^ sm^ 1 -{- 20,31 sm t cos i

By means of this equation the azimuth of the wave Iji- can be computed.

Boundary Conditions. — On passing from one medium to another, as from
air into a crystal plate, light waves encounter at the boundary surface of the
plate entirely new conditions. This passage from the system of forces opera-
tive in the first medium to that in the second is exceedingly rapid and is
accomplished within a very thin film ; but it is nevertheless a continuous
process since, physically speaking, there are no discontinuities in nature. The
boundary surface is an inhomogeneous film in which the dielectric constant
passes continuously, though rapidly, from that of the first to that of the
second medium. Now in order that the finite current be carried across the
boundary the components of the electric and magnetic forces parallel with
the boundary surface must be continuous through the boundary; for an
infinitely thin film the forces on either side of the film must therefore be
equal. If the boundary surface be the xy plane and the plane of incidence
the xs plane, the general boundary conditions are («)i=: (m),, (t/)i=: (v)2,
(X),= (X)., (Y)i=(Y)n. For periodic vibrations it is evident that {dX/dt)i
= {dX/dt)2 and (dY/dt)i = {dV/dt)2 may be used in place of (X),= (X)„
(F)i= {¥).; the last equation of (5b) can be written {du>!dt)i = {dw!dt)2,
or (w)i^(zi;); for periodic vibrations. Only four of these conditions are
independent.

In the case of light waves entering a crystal plate from air, there are
two components (magnetic vectors) on the air side of the boundary film,
Ue, that of the incident wave and ur that of the reflected wave; on the crystal
side there are the components M,, m, of the two refracted waves. Equations
(8) define the state of the light wave motion at any point and instant. For
the boundary film Z = o; in order that in the film the boundary conditions
hold for all instants of time, the relations must obtain

STUDY OF ORES AND METALS.

411

vi _ V2 _ I Sin t _ sin r _ I

2i h J ?n 2 /

in which i is the angle of incidence, rangle^ of refraction, go velocity of light in
air, q wave normal velocity in the crystal medium and / a constant, real
number. Since "q is complex, r is also complex. By means of these values
equations ii and 12a can be transformed to read
[022 + (022 — /-) tan^ r][dn — 2031 tan r + (033 — /-) tan^ r]

= {di2 — an tan^ r){i + tan- r)
(/^ — 022) tan^ r — dii

tan 5

(/2 — 022) tan^ r — a^i

(13)
(14)

(ai2 - an tan2 7) Vi + tan^ r (/' - ^ss) tan^ r + 2a3i tan r - an '

In these equations "ou • • •, r and 5 are complex quantities. A complex
value of "signifies that the amplitude in the wave is not constant; a complex
value of 5, that there is a phase difference between the amplitudes of the
components normal and parallel to the plane of incidence, hence the vibration
is elliptic in form.

For the case assumed, namely, a cr3'stal plate surrounded by air, the
boundary conditions (ri)i := (j«)o, {v~)i^{v)2, (w)i=:(w);, (<9X/c)0i =
{dX/dt)o can be written by virtue of equations (9), (5a), (6), and (7).

{E cos € — R cos p) cos i = Di cos 61 cos ?i + D2 cos 52 cos ?2,

E sin e + i? sin p = Di sin 5i + D2 sin 52,
(E cos e + i? cos 5) sin i = Di cos 5i sin ?i + D2 cos 62 sin rt,
(E sin e — R sin p) sin j-cos i
— sin ri

(15)

= Di

2r

[sin 5i(aii cos ri — ai3 sin n) + ctia cos 5i]

+ D2 — ^ [sin 52(011 cos ?'2 — Ci5 sin r2) + ai2 cos 52], ^

in which £, R, Di, Dn are the amplitudes of the incident, reflected, and two
refracted waves respectively; e, p, S, 5,, the polarization azimuths; q„, Qo, qi, ^2,
the normal velocities of the waves, i, — J, ~i, ^2, the angles between the wave
normals and the plate normal.

The last equation of (15) may also be written

(E sin e — R sin p) sin i cost ^ D^ sin ri(cos r, sin 5 + sin r^ tan s^)
+ £>2 sin7o(cosn, sin 5^ + sin n tan To),

(15^)

wherein 7;^ is the angle between the refracted wave normal and its ray direc-
tion. In this form the equation is more convenient for use in certain com-
putations. In these equations i, E, e, q^ of the incident wave are known ; i,
qa of the reflected wave; also by computation from (12) and (13) r^, n and
5j, S^^of the refracted waves; unknown are R, T oi the reflected wave and
Di, D„ of the refracted waves.

1 A dash above a letter is used to signify that the quantity represesented
may be complex.

412

WRIGHT— POLARIZED LIGHT IN THE

The computation of these four quantities from the equations is exceed-
ingly complicated. The introduction of " uniradial azimuths " facilitates the
solution considerably. Thus for a certain value of e such as «,, all of the
refracted light takes the path D (Dj^i, Z>o:=o), while for another azimuth
of the incident plane polarized wave, {D^^i, Z?i = o) and all of the light is
refracted along D^. The values of the uniradial azimuths are given by the
equations

sin^ r-tan 5

tan p = — cos {i + r) tan 6
tan e = cos {i — r) tan 5

sin {i — r) cos 5
sin^ ?-tan s

(i6)

sin {i + r) cos 5

in which T is the angle between the wave normal and the ray direction of the
refracted wave.

Let the values of E and R, computed under the assumption that D^^^o,

Dt^i, be £„ R^; similarly for £)i = o, Z). = i, let the values be E., Rn.
Since equations (15) and (15^) are linear in D^ and D.,, we find on sub-
stituting therein first Z), = 0, Z?i = i and then Dj =: 0, D.^i, and indicating
by subscripts, i or 2, in each case the proper uniradial azimuth, multiplying
the equations thus obtained by D^ and D^ respectively and then adding

[ (E, cos ej), + £„ cos e,A) — {R, cos Sjj^ + ^^cos S,A) ] cos i _

= Z), cos Sj cosTi -\- Dn cos S.. cosn;
therefore

E cos e = Ep= E, cos e, A + E., cos ej)._ = Ep,D, + Ep.3,
R cos Pz=Rp = R, cos pj5, + R., cos pjj. = Rp,D, + Rp.3.

Similarly

(17)

E sin e=^Es ^= Ej sin e^D^ -\- £0 sin (..D. = Es^D^ -\- Es.D.,
'R sin p = Rs — R, sin p,D^-j-R. sin p,5, = Rs,D^ + ^.y.A;,
in which

E^h = -Eft sin e,^, Ep,, — E,^ cos e,, ; Rs,^ — R,, sin e^„ Rp,^ = Rj^ cos e,,.

If the amplitude and polarization azimuth of the incident wave be given
and hence Esi, Ep^, 5^, and r,,, the corresponding values of the reflected wave
can be computed from equations (17) and put into the form

(Ep,Es,—Ep,Es,)Rs,= (Rs,Es,—Rs,Es,)Ep—(Rs,Ep,—Rs,Ep,)Es,
(Ep,Es,—Ep,Es,)Rp = (Rp,Es,—Rp,Es,)Ep— (Rp,Ep,—Rp,Ep,)Es,

(18)

which is deduced by eliminating D, and D._. from equations (17). The values
oi Rs and Rp are complex. _

If the incident wave be plane-polarized the complex values of Rs/En,
Rp/Ep, and Rs/Rp show that phase differences exist between the three waves
and that therefore the reflected wave is in general elliptically polarized. If

STUDY OF ORES AND METALS. 413

we put

-=- = tan p' = tan xpe'''^ = tan \p{cos A + i sin A)

RP _ _ _ _ (IQ)

_ {RsxEsi - RsjEsi) - {RsiEp2 - RsjEp,) tan e ^^

~ {RpiEsi - RP2ES1) - {RpiEp2 - RpiEpi) tan e '
in which tan e = Es/Ep and the azimuth angle 'A is real, then A is the rela-
tive phase difference of the components normal and parallel to the plane of
incidence. By equating the real and imaginary parts of this equation (o zero,
we obtain equations from which ^ and A can be computed. The equation
shows that for a certain angle of incidence the phase difference may be "'/a;
for this angle of incidence the reflected wave is plane-polarized. If e = 4S°,
then this angle of incidence, i, is called the principal angle of incidence and
the angle "^ the principal azimuth. From these two angles it is possible in
certain cases (especially isotropic bodies) to compute the refractive index
and the absorption index of the reflecting medium. Different methods have
been devised for ascertaining these angles ; practically all the methods avail-
able for ascertaining the refractive and absorption indices of absorbing sub-
stances are based on the above equations or certain modifications of the same,
deduced on the basis of simplifying assumptions.

Vertical Incidence.

In case the angle of incidence is zero, i = o, the above equations
reduce to the following :

(r — «o2)(g^ — «ii)=«i2', (11')

- q~ — ait ai2 , ,,

tan 5 = "^ = =^ — , {12a')

an r - «ii

£cose — Rcosp= Di cosSi -|- Z)o cos 82,
£sine+ Rsmp= D^ s'mJ^-j- D^s'm^, (15', 15'^?)
q^E cos e -|- q,,R cos p = QiD.^ cos 8^ + ^2^2 cos 82.
q^E sin e — q„R sin p^^q^D^^ sin 8^ + q..D^ sin 80.
For uniradial azimuths we find for 8^,
' Epi — Rp\ = cos 5i,

Esi + ^^1 = sin 5i,

5i
Epi + Rpi = — cos 5i,
2o

?i • ~
Esx — Rsi = — sin 5i.
go

414 WRIGHT— POLARIZED LIGHT IN THE

Therefore

2qoEp^=(qo-\-qi) cos 8"i, 2q^Es^={q^ ^q^) sin h^;

2go^^i=C^i — (?o) cos 81, 2q^Rs^={q^ — q^) sin \.
Similarly

2go£/'2=(^o + go) cos~8o, 2q^Ep._=(q,,^q.,) sin S", ;

2qQRpn={'q2 — ^0) cos S.,, 2q,,Rsn={qQ — g,) sin 82.

On substituting these values in equation (19), we obtain

2go(22 - qi) tan 5i tan 62 - [(go - Si) (go + 22) tan 5i

i?5 ~ (go ~ g2) (g + gi) tan ^2] tan e

Rp (go - g2)(go + gi) tan 5i - (go - gi)(go + g2) tan ^2

- 2go(g2 — gi) tan e

(19')

The right-hand side of this equation is complex ; therefore, plane
polarized light incident on a crystal plate of an absorbing crystal gen-
erally becomes on reflection elliptically polarized. The equations
are, however, so complicated, that progress is best made by the solu-
tion of a few simple cases in which the crystallographic symmetry
relations prescribe certain types of vibration.

For the special case that the plane of incidence is a plane of sym-
metry the equations become noticeably simpler ; this assumption is
valid for isotropic, uniaxial, and certain sections of orthorhombic
crystals because in these the positions of the principal axes of the
polarization and absorption surfaces of reference coincide as they
are fixed by the symmetry relations. In general these axes do not
coincide and the surface of reference can be represented only by the
use of complex quantities. On the assumption that the plane of in-
cidence is a plane of symmetry let 9 be the angle between the Z' axis
(normal to the plate) and the Z principal axis. The equations de-
fining the complex polarization constants then obtain

aii = a-^i -j- ^"^11 = ^' cos- 6 -\-'c~ sin- 6,

a'20 = aoo + ib^n = h-,

0,33 = ^33 -{- ibss = «' sin- 6 -\-c' cos" 6,

O03 = ao3 4- i&23 = o»

STUDY OF ORES AND METALS. 415

031 = 031 +^'&3i^(^^ — ^^) cos 6 sin 6,
a^o = ^12 + '^12 = O-

In these equations a^ • • • and a", &, c are complex quantities. On
substituting these values in equations (10'), (ii'a) and (15) we find

(q^ — <2ii)(2" — 022) =0 or §2" = fln Si" = ^22,

TT

5i = o, ^2 = -,

TT

ei = O, €2 = ~ .

2

The refracted waves 8^, So are thus plane-polarized. Equations (18)
reduce to the form

Rs _ RsiEp2 _ Rsi __qo — Qi
Es Ep^Esi Esi 2o + er

But from equation (10)

ii ^ gi ^ I .

go 2o(i — in) Wi(i — •iK) '

accordingly

Rs ni — I — ifiiKx
Es «] + I — iiiiKi '

The right-hand side of this equation is a complex quantity and of
the form

a + ib aA ^-hB + i{bA - aB)
A +iB ~ A'^-\-B'' - / + *^-

This in turn may be considered equivalent to the expression

{P-\-iQ)e'^=iP + iQ) (cos T-j-isin r)

= P cos T — Q sin T-|- i(P sin t + Q cos t).
If now

P^r cos A,

Q ^ — r sin A,

the amplitude may be written for £.?= i

Rs = r-^ cos (r — Aj-f rr^ sin (r — l^^)^f -\-ig = r^e ''~^' (21)

(20)

416 WRIGHT— POLARIZED LIGHT IN THE

Therefore

/ = ^iCOs(t — Ai),

^ = r,sm(r — Ai).
Hence

tan (r - Ai) = ^

The expression /^ -\- g^ is most readily obtained by multiplying
f -\-ig by its complex conjugate (/ — ig). Accordingly the intensity
of the reflected light which is proportional to the square of the am-
plitude r-^^ is

Irs (wi —

{22)

Is {n, + iy + n,H,^ ( I V , 2'

The phase difference is given by

hA — aB — 2WiKi

tan (r — Ai) = . , , p = — ; , . (23)

^ aA -{- bB Wi^ — I + WiKi

Similarly the amplitude of the component parallel with the plane of
incidence is

Rp Rp2 52—50 W2 — I — ^2^2

Ep Ep2 22+50 W2 + I — in^Ki

The intensity is

Irp _ (no - i)^ + no-Kj^

Ip ~ (W2 + i)^ + ni^Ki^'

The phase difference is

(24)

(25)

By division of (20) by (24) an expression for the amplitude
ratio is obtained

Rs (ni — I — iniKi)(n2 + i — in2K2) Es , .

_ — :== _ -: — ^ ~ . (27)

Rp (wi + I — iniKi)(n2 — I — in^Ko) Ep'

STUDY OF ORES AND METALS.

417

This IS a complex quantity and indicates that a phase difference
exists between the two components (parallel and normal to plane of
mcidence) of the reflected light; the reflected light is in general
elliptically polarized. In case the azimuth of the plane polarized inci-
dent light e coincides with either e^ or e^, the reflected light is plane-
polarized. The azimuths for which the reflected light is plane-polar-
ized occur at intervals of 90°.

The intensity ratio of the components of the reflected light is
found by division of (21) by (25)

Irs

Irp

[(ni - 1)2 -t- nr'Ki^][(n2 -f i)^ -f Ui'Ki'] Is

[(ni + 1)2 + Wi2ki2][(W2 - l)2 + n.^Ki"] Ip'

The phase difference for Ep = Es or tan €=i is A^ — A^ or

2W2K2(«l^— I+Wi^/Ci^) — 2niKi(n2^ — I + ^2^2^)

71- (28)

tan (Ai — A2)

(W]2— I + n-^K^){n2^ — 1 + n^yK'f) -f ^niKxriiKi

(29)

Transparent Media.

In order to realize clearly the effects of absorption phenomena it
is essential to ascertain first the behavior of non-absorbing media,
then to pass to the more complex problem of absorbing bodies.

Isotropic Substances. — For non-absorbing bodies k = o; in this
case equation (22) reduces to the ordinary Fresnel expression for
the intensity of rays reflected from an isotropic plate

Ir

Tr f n — i\

TABLE 2.

(30)

In this table are listed the relative intensities of light reflected at vertical
incidence from polished surfaces of substances of given refractive indices.
Thus a plate of refractive index 2.1 reflects 0.1259 of the incident light.

n

Ir
I

n

Ir

I

n

Ir

I

n

Ir
1

I.O

0

1.6

•0533

2.2

.1406

2.8

.2244

I.I
1.2

.0023
.0083

1-7
1.8

.0676
.0816

2.3
2.4

•ISS2

.1696

2.9
3-0

•2374
.2500

1-3

.0170

1.9

.0963

2-5

.1836

4.0

.3600

1.4

.0278

2.0

.nil

2.6

•1975

5-0

•4444

1-5

.0400

2.1

.1259

2.7

.2111

10. 0

.6667

PROC. AMER. PHIL. SOC, VOL. LVIII, AA, JaN. 21, 192O.

418

WRIGHT— POLARIZED LIGHT IN THE

This equation shows that the higher the refractive index the more
light is reflected. The relative intensities for a series of refractive
indices are listed in Table 2.

Equation (23) reduces for k = o to tan (r — A^)=o; in other
words, there is no change in phase on reflection. Vertically incident
plane-polarized light is reflected as plane-polarized light without
change of phase and with no change in azimuth of polarization plane.
(Equation (27) for n^^Ur, and Kj^ = k2 = o.)

Birefracting Media. — For a birefracting transparent medium
(ki=:/co^o) ; equation (28) reduces to

Irs / «i — I Y / ^^2 + I Y
Irp~\ni+ I ) \n2 - I J

4(W2 — Wi)

I —

+

4(«2 - wi)

(31)

(Wi+ I)(W2 - I) («i + I)2(W2 - 1)2

The third term of the last expression is negligible for weakly bi-
refracting substances. The change in intensity ratio with change in
least refractive index n^ and with birefringence (wo — n^) is shown
in Table 3 and presented graphically in Fig 2.

TABLE 3.

In this table are given the relative intensities of the two components,
normal and parallel to the plane of incidence, of light waves reflected from
a transparent birefracting crystal surface whose low refractive index is «i
and whose birefringence is ;;- — n^. Thus for a crystal plate, whose least
refractive index is 1.8 and whose birefringence is 0.040, the ratio of the inten-
sities of the two reflected components is 0.932.

«2— «1

0.005

0.010

0.020

0.030

0.040

0.050

0.075

O.IOO

0.150

0.200

Ml
1.2

.960

.919

•845

•794

• 723

•673

•568

•487

•374

.298

1.4

•979

.960

.917

.888

•854

.824

•755

•695

•598

.522

1.6

•989

•975

•952

.928

.908

.886

.837

•793

.716

.652

1.8

.991

.982

.964

■948

•932

.916

.881

.848

•787

•734

2.0

•993

.986

•973

.962

■949

•937

.908

.882

•833

.789

2-5

•995

.991

•984

•977

•970

•963

•945

.929

.898

.869

30

•998

•995

.990

.986

.980

.976

•963

•953

•931

.911

4.0

•998

.996

•995

•992

•989

.986

.982

•974

.962

•951

5-0

•999

•998

•997

•995

•993

•992

.988

•98s

•977

.968

STUDY OF ORES AND METALS.

The ratio of the ampHtudes (equation 27) becomes

Rs (ni — i)(w2 + i) Es

Rp («i + i)(w2 — i) Ep'

419

(32)

uu

^

z:^

==

^

==

:^^

::;;r:

.

^

;^3:^::ii

s

=

n

<<

■^

F^

:^

^

""'

^

^

/^

;//

/

y

^

-^

1

/ ,

%^

V

/

/
/

/

/A

' pA

/

11

//

a

1

/

|,

50

1/

//

1 1

//

I

/

/

/

30

1/

/

/'

20

//

//

10

ill

1

1

Fig. 2. The curves in this figure indicate the change in the percentage
intensity ratio (ordinates) of the two components of normally reflected light
with change in the least refractive index »i (abscissae) and in the birefrin-
gence (Ho — "1) (curves) of a crystal plate.

The fact that this is a real quantity proves that there is no phase
difference between the two components ; the phenomena observed on
such plates are due therefore entirely to a difference in the ampli-
tudes of the reflected wave components. The reflected light is plane-

420

WRIGHT— POLARIZED LIGHT IN THE

polarized; the azimuth p of the plane of polarization of the reflected
wave is for Es^^Ep or tan €= i (equation 32)

Rs 2(«2 — Wi)
tan p = ^^ = — I + -, , — TT r

Rp (Wi + l)(W2 - i)

(33)

There is therefore on reflection a rotation of the plane of polar-
ization through the angle (e — p). For the ordinary case tan e^i
(e = 45°), we find from (33)

tan (e — p) =

I — tan p Hin^ — i
I -f tan p 712 — fii

(34)

The angular amounts of rotation on reflection of vertically inci-
dent plane-polarized light for e = 45° and for different refractive
indices and birefringences are listed in Table 4 and are presented
graphically in Fig. 3.

TABLE 4.

In this table is given the angular rotation, on reflection from a crystal
plate, of the plane of polarization of vertically incident light waves whose
plane of polarization includes an angle of 45° with the vibration directions of
the crystal plate.

«1

1.2

1.4

1.6

1.8

2.0

2.5

30

4.0

5-0

«2— «1

.005

o°39'

0° 18'

0°ll'

o°o8'

o°05'

o°03'

0°02'

0°0l'

0°0l'

.010

I 16

0 35

0 22

0 15

0 II

0 06

0 04

0 02

0 01

.020

2 28

I 09

0 43

0 30

0 22

0 13

0 08

0 05

0 03

.030

3 36

I 43

I 04

0 45

0 35

0 19

0 13

0 07

0 04

.040

4 41

2 15

I 25

0 59

0 44

0 26

0 17

0 09

0 06

.050

5 42

2 47

I 45

I 14

0 55

0 32

0 21

0 II

0 07

•075

8 03

4 02

2 33

I 48

I 22

0 48

0 31

0 17

0 10

.100

10 08

5 II

3 19

2 22

I 47

I 03

0 41

0 22

0 14

.150

13 36

7 18

4 46

3 26

2 36

I 32

I 01

0 33

0 21

.200

16 23

9 10

6 04

4 24

3 19

2 00

I 20

0 43

0 28

Figures 2 and 3 indicate clearly that there are two distinct meth-
ods for detecting anisotropism in birefracting, transparent media by
means of vertically incident light ; either intensity differences be-
tween the two components parallel and normal to the plane of inci-
dence may be utilized (equation 31) or the rotation of the plane of
polarization (equation 34).

Differences of intensity are detected ordinarily by photometric

STUDY OF ORES AND METALS.

421

methods which involve direct comparison of two adjacent illumi-
nated fields (half shade principle). The accuracy of a single setting
obtainable with such methods is dependent on the sensitiveness of

\\

J

V

\

\

\

\

Q

1

\

\ ^

\

11 \

\ \

1

\

ml

\

1 \

\

\

V

\

\\

\

\ \

\

^

\\

\

\

<

Vn

^

\ \

\

\

\
\

\

\

\

\

V

\

\

\^

^

\,^

>■

^v

\

\

^

w

\

K

\

^

^

"~^

-___

^^

\'

■^

:::5

^^

-^

rr:

-^

■ -

— ■

—

-

—

—

^^

in^

^l^li^i:z:

^ z

=

==

—

^

"

Fig. 3. In this figure is indicated the angular rotation in degrees (ordi-
nates) of the plane of polarization of vertically incident, plane-polarized light
of azimuth 6 = 45° on reflection from a crystal plate of least refractive index
111 (abscissae) and of birefringence n. — Wi (curves).

the eye, or on the least perceptible difference in light intensity under
the conditions of observation. An extended series of measurements
by Koenig and Brodhun^ on the least perceptible increment (dif-

^ Sitzungsherichte d. Acad. d. Wissensch. Berlin, July 26, li
melte Abhandl. und Physiolog. Optik " (A. Konig), 116-143.

" Gesam-

422

WRIGHT— POLARIZED LIGHT IN THE

ference limen) for different intensities and different wave-lengths
shows that for an intermediate intensity of illumination differences
of intensity 1.6 to 2.0 per cent, can just be detected; for low (below
50 m.c.) and high (above 100,000 m.c.) intensities, the least per-
ceptible difference in intensity is greater. Under ordinary condi-
tions of illumination we may assume that 2 per cent, difference in
illumination intensity is about the limit perceptible to the eye. This
means that in order to be recognized by the eye there must be a dif-
ference of about 2 per cent, in intensity of light reflected by the two
components. To find the birefringence (Wa — ^1) required to pro-
duce this least perceptible difference in intensities of reflected light,
namely, 2 per cent., we may without sensible error use only the first
two terms of equation (31) and transform it to read

712

fix =

0.02 (wi + i)(w2 — i)

(35)

from which the necessary birefringence (n^ — nj for the different
refractive indices n^ can be computed. These are listed in Table 5.

TABLE 5.

In this table is given the least degree of birefringence of a crystal plate
which on the reflection of vertically incident light produces a detectable differ-
ence in intensity of the two waves polarized at planes normal one to the other.

«1

«2— «1

«i

n«—ni

«l

Wo — «i

I.I

.001

1.6

.007

2.5

.026

1.2

.002

1-7

.009

3-0

.040

1-3

.004

1.8

.Oil

4.0

•075

1.4

.005

1.9

.013

5-0

.120

1-5

.006

2.0

.015

10. 0

.440

This table shows that for very low-refracting, non-absorbing sub-
stances only a relatively slight increase in refractive index (bire-
fringence weak) is necessary to produce the required perceptible
difference in intensity ; for minerals of medium refringence and bi-
refringence, such as quartz, the difference limen increases, but the
birefringence required is still medium to weak ; for highly refracting
media the required difference in refractive indices increases rapidly.
The essential feature to note, however, is that minerals of medium

STUDY OF ORES AND METALS. 423

refractive index but weak birefringence, such as apatite, do not ex-
hibit the required difference in intensity (2 per cent.) of the reflected
components ; and that, if the observations were Hmited to this phe-
nomenon, the anisotropism of the mineral would escape detection.
In other words the sensitiveness of methods based on the relative
intensity of the two reflected components is far below that for de-
tecting anisotropism in transmitted light. Thus it is not difficult in
ordinary thin sections (0.02 mm. thick) to detect a path difference of
2/XjU,; this corresponds to a birefringence of o.oooi. For a mineral
of refractive index 1.6 the least detectible birefringence by reflected
intensities is about 0.070. In other words the methods of trans-
mitted light are 50 or more times as sensitive as the methods based
on intensity differences of reflected intensities. This follows as a
direct result of the lack of sensitiveness of the eye in detecting dif-
ferences in intensity of adjacent fields.

Methods based on equation (34) depend on the ability of the eye
to determine the position of complete darkness (intensity of illumi-
nation =0) ; the chief factor which limits the degree of precision
attainable by this method is the threshold limit of vision or the least
quantity of light which the eye can detect. Experience has shown
that it is not difficult to detect, under ordinary conditions of illu-
mination with the aid of certain devices, a rotation of the plane of
polarization through 5'. Substituting this value in equation (34)
we obtain

n„ — n^ = o.oois(>h-\- i)(n.,— i). (36)

A comparison of this equation with (35) shows that this method is
at least three times as sensitive as the first. The precision attain-
able by the second method is dependent, moreover, on the sensitive-
ness of the device employed to detect a rotation of the plane of
polarization. Without any special device an error of >4° is readily
possible. In this case the accuracy of this method is considerably
less than that of the intensity dift'erence method. In case a device
is used which shows a perceptible difference for a rotation oi }i°
the precision attainable by the two methods is identical. The ac-
curacy of the second method can, however, be increased by using a
very intense source of light. The several methods for determining

424

WRIGHT— POLARIZED LIGHT IN THE

positions of extinction have been discussed in detail by the writer*
and the conclusion reached that the bi-quartz-wedge-plate is the most
sensitive device available for the purpose. With it the sensitiveness
can be varied to meet the conditions of illumination. M. Berek^ has
shown that with the bi-quartz-wedge-plate a precision of 0.5' can be
attained under the most favorable conditions of observation.

Absorbing Media.

In the foregoing section the phenomena produced on reflection
from transparent crystal plates are treated in some detail and the
factors underlying the methods for the detection of anisotropism and
for the measurement of the birefringence and the refractive indices
are discussed. The introduction of the absorption index into the
equations carries with it a series of complications which render the
relations less easy to follow, but which are, as a result, the more
interesting.

TABLE 6.

In this table the intensity is given of the light reflected, at vertical inci-
dence, from the surface of an isotropic substance of refractive index n and
absorption index k. Thus a plate of refractive index 2.0 and absorption index
1.32 reflects 50 per cent, of the incident light.

/.,//,

0.1

0.2

°-3

0.4

0-5

0.6

0.7

0.8

0.9

O.I

. .

3-51

8.9

16.7

0.2

—

—

—

—

3^74

6.06

8.0

12.8

0.3

—

—

•52

1.86

2.00

3^82

5^07

6.92

10.7

0.4

—

■5

1-43

2.17

2.81

3-57

4.67

6.13

9^35

0.5

—

1.0

1.56

2.08

2.78

3^32

4.21

5^57

8.43

0.75

.69

I.I

1.47

1.97

2.29

2.79

3^51

4.70

6.92

I.O

.67

1.0

I-3I

1.63

2.00

2.45

3^05

4.00

6.00

2.0

—

•5

.78

1.04

1.32

1.66

2.10

2.78

4.22

3.0

—

—

.36

.67

•94

1-25

1.63

2.21

3^41

4.0

—

— •

—

•32

.66

•97

1-33

1.85

2.90

S-o

—

—

—

—

.40

•75

I. II

1.60

2.56

7-5

—

—

—

—

—

.21

.69

1. 17

2.03

10. 0

—

—

—

—

—

—

•35

.89

1.67

For K

= 0

Ml .

•52

1-93

.38
2.62

.29
3-43

•23

4.44

•17
6.04

•13
7.90

.09
"•3

.06

•03
38.2

W2.

17.9

^Aiii. Jour. Scl. (4), 26, 377-37S, 1908. Carnegie Institution of Wash-
ington, Publication 158, p. 139, 191 1.

^ Neues Jahrbuch fiir Mineralogie etc., Beilage Band jj, pp. 583-661, 1912.

STUDY OF ORES AND METALS.

425

Isotropic, Absorbing Media. — For such substances Mi = W2,
K^ = K2. The intensity of the reflected Hght is (equation 22)

Irs (n — 1)2 + n^K^

(37)

Is {n + i)^ + w^K^"
In order to form a clear picture of the relative influence of the

/

^-^

^

^

—

-^

=-

-=

A

/

/

.►"

^

^^

/

/

/-

../

.

//

/

/

,'!,•'

-^

---

70

/

/

/

/

/

-^

^

^

-^

/

' /

/

-"

^

^^

^^

6!o

\)

/

/
/

X

^

\

/

/

/

/

^

^

w

/

/

/

/

y

^

1

/

/

/

,0

^^

40

/

/

X

V.

/

'

/

/

/

/

/

/

/

I

'

/

/

/

/

20

1

/

/

^

/

/

/

\

/

/

/

\v

/

\

^-

On^»1 2 3 4 5 6 7 8 9 lO

Fig. 4. The percentage (ordinates) of normally incident light reflected
from a plate of refractive index n (abscissae) and absorption index « (curves)
is given in this figure.

refractive index and the absorption index on the reflecting powder
of an anisotropic, absorbing crystal surface, a series of values com-
puted by means of equation (37) are listed in Table 6 and are rep-

426

WRIGHT— POLARIZED LIGHT IN THE

resented graphically in Fig. 4 in which the refractive indices are
plotted along the abscissae, the reflecting power in per cent, along
the ordinates ; the absorption indices are indicated by the curves.
In Fig. 5 the same relations are expressed in different form ;

10

\

1

1

\

1

1

1

\

1

\

\

\

1

\

.\

\

\

\

\

\

\

\\

\

\

\

W

\

\

\

\

^

^

\

\

\

\

V'

V\

\

\

\

\ \

\ \

^

\

\.

\\

\

W

\

\

\

'A \

rV

\

\

^C

\

\\

\

C

.i«lr

\

\

\

N

\.\

\\

•Pd\

^^1

-Sn

\

x^

V

^\

\\

\

\

U,,^

••Al

,

:i

"^

^

^

)

)\

>\

\

X

.CO

_^

^~^

1

c

1

—

=:^

>

>

>

— -

=.

3

O K— » 1

Fig. 5. In this figure the reflecting power (curves) in per cent, for nor-
mally incident light, of a plate of refractive index n (ordinates) and of
absorption index k (abscissae) is indicated. The optical constants of a num-
ber of element's are also given.

the ordinates are the refractive indices, the abscissae, the absorption
indices and the curves, the reflecting power in per cent. In this
figure are also included the refractive and absorption indices of cer-
tain opaque elements and compounds. The figure is instructive

STUDY OF ORES AND METALS. 427

because it shows that high reflecting power (metalHc or sub-nietalHc
luster) may arise either from high absorption index as in Na, Ag;
or from high refractive index, as in Si and galena ; or from both
high refractive index and high absorption index, as in Ir, Zn. The
curves of these figures show that in general an increase in the ab-
sorption index is much more efifective in increasing the reflecting
power of the substance than the same amount of increase in the re-
fractive index.

The ratio of the amplitude components Rs, Rp is (equation 27)

Rl _ _Ep
Rp~ ~ Ep'

For Ep = Es this reduces to — i and indicates a phase difference
of TT between the incident and reflected waves. The phase difference
between the two components is (equation 29)

tan (A. — Aj^o.

Therefore plane-polarized, vertically incident light is still plane-
polarized after reflection from an isotropic absorbing body.

Birefractmg , Biabsorbing Media. — The intensity ratio of the
reflected components from a birefracting, biabsorbing crystal plate
is stated by equation (28) which can be written for £.? = £/> or
tan e ^ I

Irs _ 4[(wi — no)(i ■+■ W1W2) + Uin^in^K^^ — Wifi^)

lrp~^ [(«x + 1)2 + n,'Kmn2 - lY + n,^KiV~ ' ^^^^

This equation shows that for weakly birefracting substances {n-^
approximately equal to Wo) the intensity ratio depends on the dif-
ference of the squares of the absorption indices. For substances
with weak biabsorption the intensity ratio depends primarily on the
difiference of the refractive indices. This indicates that an increase
in the biabsorption is more effective in increasing the degree of
anisotropism than the same amount of increase in birefringence.

The phase difference between one of the components and that
of the incident beam is given by equation (23). Series of values
computed by means of this equation are listed in Table 7 and are
presented in graphical form in Fig. 6 in which the refractive indices

428

WRIGHT— POLARIZED LIGHT IN THE

are the abscissae, the ordinates the angles (t — A^), and the curves
the absorption indices. The curves show clearly that for substances
with high refractive indices and high or low absorption indices the

TABLE 7.

In this table are listed the phase differences between the vertically inci-
dent beam and one of the components reflected from a plate of refractive
index n and absorption index k. Thus for a plate of refractive index 1.5
and absorption index 0.2, the phase difference is 24° 07'.

K

0.2

0.4

0.6

0.8

1.0

1-5

n
0.2

- 4° 47'

- 9° 32'

-14° 14'

-18° 54'

-23° 52'

-34° 36'

0.4 .

-10 52

— 21 27

-31 31

— 41 00

-49 37

-68 12

0.6

— 21 00

-39 30

-54 40

-66 52

-76 52

85 15

0.8

-45 30

-68 03

-82 18

87 47

80 05

65 46

I.O

84 18

78 42

73 18

68 12

63 27

53 08

i-S

24 07

36 42

41 10

41 44

40 36

35 30

2.0

14 13

23 45

28 24

29 55

29 45

26 34

3-0

8 10

14 16

17 45

19 13

19 36

18 26

4.0

5 SI

10 19

13 01

14 13

14 28

13 14

5-0

4 35

8 08

10 18

II 19

II 32

10 36

7-5

2 59

5 20

6 49

7 29

7 40

7 03

10. 0

2 01

3 59

5 04

5 36

5 45

5 18

K

2.0

3-0

4.0

5-0

7-5

lO.O

n
0.2

-45° 00'

-63° 27'

-78° 42'

88° 51'

66° 43'

52° 45'

0.4

-82 53

75 58

61 44

51 41

36 20

27 49

0.6

71 34

54 II

43 II

35 41

24 40

18 45

0.8

55 30

41 37

32 55

27 05

18 36

14 06

1.0

45 00

33 41

26 34

21 49

14 56

II 19

1-5

30 21

22 43

17 51

14 38

9 59

7 33

2.0

22 50

17 06

13 26

II 00

7 30

5 40

3.0

15 16

II 30

8 36

7 20

5 00

3 49

4.0

II 27

8 35

6 44

5 31

3 45

2 50

5-0

9 10

6 52

5 24

4 25

3 00

2 16

7-5

6 07

4 35

3 36

2 56

2 00

I 31

10.0

4 35

3 27

2 42

I 55

I 30

I 08

phase difference is only a few degrees and therefore, so far as useful
in measurements, is practically negligible. In other words the reflected
light, although elliptically polarized, approaches plane polarized light
in character and the essential phenomena to be observed are inten-
sity differences and rotation of the plane of vibration. This state-
ment is valid for all substances of refractive index above 3.0 and
for substances of still lower refractive index (1.5) provided the
absorption index is either high or very low (less than 0.3). The

STUDY OF ORES AND METALS.

429

region for which slight changes in refractive index or absorption
index produce the greatest difference in phase are for low refractive
indices w < 1.5 especially for n < i and for high absorption indices.
Substances, however, with such indices are rare ; the conclusion is
therefore generally valid that the phase difference between the two
components of a birefracting and biabsorbing substance for ver-

.I8Cf

on— <1 2 3 4 5 6 7 8 9 'C

Fig. 6. In this figure is given the phase diflFerence (ordinates) between
one of the components of normally incident polarized light (azimuth € = 45°)
and the same component after reflection from a birefracting, biabsorbing
crystal plate of refractive index n (abscissae) and absorption index k (curves).

tically incident light is not great and that the anisotropism finds
expression chiefly in the intensity differences of the two com-
ponents. This conclusion is important, since it enables the observer
to &pply the methods outlined above for non-absorbing media to
the detection of anisotropism in opaque substances.

430 WRIGHT— POLARIZED LIGHT IN THE

The amplitudes Rs, Rp (equation 20, 24) are complex and of the
form f-\-ig. In order to compound these two vibrations (vectors)
and to ascertain the directions of the resultant elliptical vibration let
(equation 21)

Rs = rie'^'-'''\

(sq)

If i/' be the angle between the Rp axis and the major axis of the
resultant ellipse of vibration, then we find by taking the real parts of
(39), namely

Rs = r. cos (t — A,),

(40)
and Rp = r^ cos (t — Ag),

and compounding them into the form

Mo = n cos (r — Ao),

(41)
7:^0 = ^0 sm (t — Ao),

which obtains for the components after the principal axes of the
elliptical vibration, that

2rir2

tan 2^ = — -cos (Ai — A2). (42)

^1 — r<t

To effect this transformation the normal transformation equations

Uq = u cos ip -^r V sin xp,

• ,1 , ^43)

z'o= — « sm i/'-j-z' cos </',

are used and the coefficients of cos r and sin t of equations (41)
and (43) are equated.
If we put

= tan §, (44)

equation (42) may be written

tan 2 i/' = tan 2 (9 cos (A^ — Ao). (45)

The quantities on the right hand of this equation can be computed
from equations (27), (29), (40) and (41), (44); but the expres-
sion thus obtained is too complicated to be readily interpreted. The
fact that the phase difference for most opaque substances is, as

STUDY OF ORES AND METALS. 431

shown above, small and therefore the cosme of the angle (A2 — A^)
not greatly different from unity indicates that </' is only slightly
smaller than 0 whose tangent is (equations (22) and (44))

^''"' " r,' [(«2 + ly + n2^K2'][(wi - I)' + wi'/ci^ •

The right-hand side of this expression is identical with that ob-
tained for the intensity ratio. Therefore, in general, the square of
the tangent of the angle of rotation 6 is approximately equal to the
intensity ratio. As this increases so also does the angular rotation
of the plane of polarization.

The above equations and the methods based on these equations
do not sufihce for the determination of the optical constants n^, n^,
Kj, /Co of an opaque crystal examined only under vertical incidence.
They do indicate, however, that for the detection of anisotropism
the phase difference between the two components Rs and Rp is ordi-
narily not sufficiently large to be readily measured and that therefore
the phase difference is of little value as a diagnostic feature. The
other variable factor is the amplitude ratio. This gives rise to a
difference in intensity of the components of the reflected wave. For
certain cases in which the reflected waves are plane-polarized one
of the components is more intense than the second and the excess in
intensity in the one direction produces a certain amount of polarized
light after reflection of the non-polarized, incident waves ; this can
be detected by methods similar to those which have long been in use
for the detection and measurement of polarized light in the sky.
Since for most substances the phase difference between the reflected
components is not great, methods suitable for the detection of small
angular rotations of the plane of polarization of incident plane-
polarized light may also be used to detect anisotropism in opaque
bodies. A brief description of the several methods which may serve
for this purpose will now be given.

Methods Based on Intensity Contrast.
These methods are wndely used in the measurement of sky polar-
ization and have proved to be of great usefulness in that connection.
One of these methods has been applied by J. Koenigsberger to the
detection of anisotropism in plates of opaque crystals.

432

WRIGHT— POLARIZED LIGHT IN THE

Koenigsberger's Method.^ — In a series of articles Koenigsberger
suggested the use of a Savart plate^ in conjunction with an anal-
yzing nicol and a telescope for the detection of polarization in light

reflected from a crystal surface, the
incident light to be non-polarized and
to impinge vertically on the plate.
The arrangement proposed by Koe-
nigsberger is illustrated in Fig. 7.
The non-polarized light (monochro-
matic or white) enters the system
along B, passes through a contrast
plate Q consisting of a bi-plate of
smoky quartz cut parallel to the axis
and mounted so that the axis is at
right angles in the adjacent halves ; a
lens L, for imaging the contrast plate
in the image plane of the telescope E.
The light passes through L to a small
reflecting prism R, thence through the
objective O to the crystal plate C
where it is reflected back through O
past R, through a plane-parallel, ro-
tatable glass plate P of known refrac-
tive index (w ^1.515), through the
Savart plate S and the nicol A'' into
the telescope E.

The Savart plate consists of two
plates of calcite or quartz cut at an
angle with the optic axis sufficiently
large (45°) that if examined in con-
vergent polarized light the isochromatic
curves of the interference figure cross the field as practically straight
lines. The two plates are of equal thickness and are superimposed
so that the horizontal projection of the axis in the one is at right

6 Centralblatt fur Mineralogie, Geologie u. PaVdontologie, 1901, pp. 195-
197; 1908, pp. 565-569, 597-605; 1909, pp. 245-250; 1910, pp. 712-713-

^ Devised and first described by D. Brewster, Edinburgh Transactions, 9,
148, 1819; described later by Savart, Pogg. Ann. d. Phys., 49, 292, 1840.

Fig. 7. Koenigsberger's ap-
paratus for the detection and
measurement of anisotropism
in opaque substances.

STUDY OF ORES AND METALS.

433

angles 'to that of the axis in the second. The relations are shown in
Fig. 8. If the combination be observed between crossed nicols with
the axes at 45° with the principal nicol planes, a series of parallel,
vertical dark interference bands appears, if monochromatic light be
used, in the field ; if a white light source be used the bands, except
one central band, are colored. These bands are similar in appearance
to the bands in a quartz wedge or a Babinet compensator ; but their
mode of formation is different as they depend on very slightly con-
vergent polarized light for their development while in the quartz
wedge the change in thickness of the wedge introduces the required

fb
a

Fig. 8. Diagrams illustrating the principle on which the construction of
the Savart plate is based.

path difiference. From the mode of formation it is evident that if
non-polarized light be used no such bands will appear, but that with
a small percentage of polarized light there will be superimposed on
the white field a series of colored bands the intensity of which in-
creases with the amount of polarized light present. Under the best
conditions of setting at the center of the light bands the maximum
intensity is obtained, namely, J/^ of total non-polarized light and Yi
of total polarized light ; whereas at the center of the dark central
band only the non-polarized portion is transmitted. At intermediate
points the non-polarized light and a part of the polarized component
is transmitted. The field, therefore, alternates in intensity ; the least
amount of polarized light which can be detected depends obviously
on the least perceptible increment in intensity which the eye is able to
detect. As noted above Koenig and Brodhun's data place this dif-
ference limen at 1.6 to 2 percent, for favorable intensity of illumi-
nation. Pickering and others^ estimate that in sky polarization
about I per cent, of polarized light can be detected. Koenigsberger

^ Report of U. S. Naval Observatory of Total Eclipse of July 29, 1878.

PROC. AMER. PHIL. SOC, VOL. LVIII, BB, JAN. 21, I92O.

434 WRIGHT— POLARIZED LIGHT IN THE

asserts that 0.3 per cent, difference can be detected, but a series of
experiments by the writer using a rotating, optically plane-parallel
plate and dift'erent sources and intensities of illumination indicate
that the figures of Koenig and Brodhun are more nearly correct and
that Koenigsberger's statement of the precision attainable is about
5 times too great. The difference between different settings, espe-
cially for large intensity differences, is of course considerably less
than the least perceptible increment but this difference may not be
considered to be the least perceptible increment itself.

In order to render more readily visible the Savart bands and the
point of exact compensation Koenigsberger employs a contrast bi-
plate of smoky quartz which introduces a difference in intensity
between the components (result of pleochroitic absorption) and thus
produces a shift of the lines in the halves of the field ; these lines are
then shifted by the changes in intensity resulting on reflection from
an anisotropic crystal plate.

In the practical application of this method it is essential that the
plate to be examined be well polished and normal to the axis of the
microscope ; that the reflecting prism be not too large ; that the
rotating glass plate P be mounted with its axis at 45° with the upper
nicol plane ; that the Savart plate be accurately constructed and be
normal to the microscope axis ; that the telescope be accurately
focussed on infinity (in order that convergent light of only very
small angular aperture pass through the Savart plate). To measure
the degree of anisotropism the glass plate P is rotated until the effect
of the crystal plate is exactly compensated and the Savart bands dis-
appear or are separated by exactly half a band if the contrast band
be used.

The effect of the tilted plate in compensating intensities normal
and parallel with the plane of symmetry can be found from equa-
tions 15 simplified for the case of an isotropic body. Thus if D be
the amplitude of the transmitted plane-polarized wave and 8 its
azimuth, we find

sin 2r sin 2i
D cos i5 = £ cos e —V- ^TT—. — ^ = E cos e • 61,

sm''(i -f- r)

sin 2r sin 2i .

D sm 8 = E sui e-r-. — : :— , — : ,„ = E sin e-Co.

(sm t cos t -j- sin r cos r)^

STUDY OF ORES AND METALS.

435

The intensity contributed by the component of the wave D (azimuth
8) in the plane of incidence is accordingly

The total intensity for the components of all waves D (non-polar
ized light) in the plane of incidence is

Jo

IT 12

cos^ e-de.

Similarly the total intensity for the components of all waves D
(non-polarized light) normal to the plane of incidence is

7r/2

E^C2^ I swU-de.

The ratio of the two intensities is

Ip C\ (sin i-cos i + sin r-cos rY

Is Co

sin^ (i + r)

— = cos* {i — r). (46)

The same expression can be derived more directly from equation
( 16) if we consider the waves to pass through the system in reverse
direction which is permissible. A series of values computed by
means of this equation is listed in Table 8 and shown by the curve
of Fig. 9.

TABLE 8.

Let a plane-polarized light beam be incident at the angle i on a plane
parallel glass plate of refractive index 1.515; let the plane of polarization in-
clude angle of 45° with the plane of incidence. The ratios of the inten-
sities of the two components, parallel and normal to the plane of incidence,
of the beams emerging from the plate under these conditions are given in this
fable.

/

/p//, = Cos4(/-r).

i

IplI, = CosHi-r).

0°

1. 0000

25

■ (^1-Al

5

.9848

30

. 56.25

10

.9406

35

, 45-02

IS

.8705

40

34-51

20

.7798

45

25.00

In this figure the curve drawn by Koenigsberger and furnished with
his apparatus is reproduced as the dotted curve. (Refractive index

436

WRIGHT— POLARIZED LIGHT IN THE

of glass plate =1.515). The two curves do not coincide; no satis-
factory explanation for the discrepancy has yet been found.

Photometrically the gradation in intensity between the different
Savart bands is less advantageous than the juxtaposition of two or
a series of evenly illuminated fields, the one at minimum intensity
the next at maximum intensity. This principle is used in measure-
ments of sky polarization and can be readily applied to the present
problem.

"^

\.

V-

N

s '^

\

\

^ \

\

^

\ '■■

\'

\

\

\

\

\

\

\,

\

S

\

\

S

V

\.

B-

- 1

o"

2

0-

3

0"

4

0"

5

0*

6

0"

7

0'

8

0

9C

Fig. 9. Curve illustrating tlie percentage intensity ratio (ordinates) be-
tween the two components, normal and parallel to the plane of incidence, of
light transmitted through a tilted glass plate for different angles of rotation
(abscissae).

Nezv Method. — For this method use is made of a cleavage plate
of calcite of such thickness and an aperture of such width that the
two fields from the two rays just touch (Fig. 10) ; better fields are
obtained by use of a small Koenig-]\Iartens portable photometer"
(Fig. II).

3 Verhandlungen d. dcutsch. Physikal. Gesellschaft, i, 204-208, 1899 ;
Physikalische Zeitschrift, i, 299-303, 1900.

STUDY OF ORES AND METALS.

437

-^-B

Fig. 10. New
method for the de-
tection and meas-
urement of the
degree of aniso-
tropism of a crys-
tal plate. In this
method the cleav-
age plate of cal-
cite serves to pro-
duce two adjacent
fields whose planes
of polarization are
normal to each
other. The inten-
sity of the illumi-
nation of the two
fields is made
equal by rotation
of the nicol A''.

polarized light be
fields is different;

In Fig. lo the non-polar-
ized incident light is reflected
by the prism R to the plate C
whence it passes on reflection
through the calcite plate aper-
ture^ to the nicol A^, the weak
lens D and the microscope eye-
piece E. In the writer's mi-
croscope the upper nicol can
be rotated and the angle of
rotation read off to 6' on the
stage of the microscope. The
function of the low-power
lens D is to image the aper-
ture A in the focal plane of
the weak eye lens E.

In Fig. 1 1 the calcite plate
is replaced by the Koenig-
Martens arrangement of aper-
ture ^, field lens D (to render
rays from aperture A to Wol-
laston prism W parallel), and
the twin prism F. This gives
a better photometric field than
the calcite cleavage plate. In
both methods artificial light,
non-polarized, and either mon-
ochromatic or white is used
and rendered diffuse either
by the interposition of a
ground glass or opal glass
plate. The fields should be
equally illuminated and no
junction line between them
should be visible. In case
present the intensity of illumination of the. two
but it can be made equal by rotating the nicol A''

O

Fig. II. In this
method the Koe-
nig-Martens ar-
rangement of wol-
laston prism and
twin glass prism
is substituted for
the calcite cleav-
age plate of Fig.
lo. The intensity
of the illumina-
tion of the two
adjacent fields is
rendered equal by
rotation of the
nicol A^.

438

WRIGHT— POLARIZED LIGHT IN THE

through an angle a ; in this case the observed intensity from the
one field of intensity /, is / cos^ (45° — a) ; that of the second, h,
is / cos- (45° -|- a). Since the observed intensities are equal we
have

COS^ (45° + a)
COs2 (45° - a)

= tanM45°-«).

(47)

A series of values computed by means of this equation is listed
in Table 9. For practical purposes a large-scale curve plotted on
millimeter cross section paper is useful for ascertaining intermediate
values.

TABLE 9.

In this table are listed the relative intensities of the sources of illumina-
tion of the two fields of a calcite rhomb or a Koenig-Martens photometer for
different settings a of the nicol N at which equal field intensities are obtained.

a

Tan2 (45-a)

a

Tan2 (45-a)

a

Tan2 (4s-a)

0°

1. 0000

7°

.6104

18°

.2596

I

.9326

8

.5678

20

.2174

2

.8696

9

.5279

25

.1325

3

.8107

10

■4903

30

.0718

4

.7557

12

.4217

35

.0311

5

.7041

' 14

.3610

40

.0077

6

.6558

16

•3073

45

.0000

Of the two devices the Koenig-Martens photometer is the better
because it gives a better photometric field and is an attachment com-
plete in itself which may be used on any microscope in place of the
eyepiece. The only additional accessory required is the reflecting
prism or mirror with which every metallographic microscope is
equipped.

These methods are superior to the Koenigsberger method, not
only because of increased sensitiveness, but also because the par-
ticular plate under investigation can be viewed directly (at high or
low magnifications) and the test made directly on the plate in full
view. In the Koenigsberger method the Savart bands are not readily
distinguished unless the plate is brought out of sharp focus and even
then the irregularities of the image tend to render indistinct and
uncertain the faint Savart bands.

STUDY OF ORES AND METALS.

439

Methods Based on the Detection of a Ro-
tation OF the Plane of Polarization
(Position of Extinction).

These methods have been long in use by
petrologists and the principles underlying the
construction of the different devices need not
be repeated here."

Kocnigsberger's Method}'^ — In this method
Koenigsberger adopted the arrangement shown
in Fig. 12. The incident light passes first
through the polarization prism (vibration
plane horizontal or vertical) P, and the total
reflecting prism R to the crystal plate whence
it travels through the Biot quartz plate A (cut
normal to the axis and of such thickness, 3.75
mm., that it gives the sensitive tint between
crossed nicols), the eyepiece E, the cap nicol
A^ to the eye of the observer. The rotation of
the plane of polarization on reflection is de-
tected by the change in interference hue on
rotating the crystal plate. Koenigsberger
emphasizes the fact that this method is not so
sensitive as his first method and is only quali-
tative in nature.

H an emann's Method. — In 1913 Hanemann^-
improved Kocnigsberger's method by substi-
tuting a Soleil-Biot plate (two adjacent Biot
quartz plates, 3.75 mm. thick, cut normal to
the axis, the one of right-handed, the second
of left-handed quartz) for the single
Biot plate. This arrangement introduces the

n A

. m

R ' ^

t=

0

&

s

/

c

Fig. 12. Diagram
showing general ar-
rangement for the
detection of the rota-
tion of the plane of
polarization of nor-
mally incident plane-
polarized light waves
on reflection from a
birefracting and bi-
absorbing plate.

feature of color con-

10 These are treated at length in " The Methods of Petrographic Micro-
scopic Research," Carnegie Institution of Washington Publication 75^, pp.
115-148, 1911.

'^^ Centralhlatt fiir Mineralogie, etc., 1909, 245-250; 1910, 712-713; Metal-
lurgie 4, 605-608, 1909.

12 K. Endell and H. Hanemann, Stahl und Risen, 1913, p. 40; Zeitschrift
fiir anorganische Chetnie, 83, 267-274; 88, 265-268, 1914.

440 WRIGHT— POLARIZED LIGHT IN THE

trast in adjacent fields and is accordingly more sensitive than the
Koenigsberger method.

The Bertrand eyepiece which is in common use by petrologists
would serve the purpose as well as, and possibly better than, the
Soleil-Biot plate employed by Hanemann.

Nezv Metliod. — The methods of Koenigsberger and of Hane-
mann are based on changes in color (sensitive tint) on rotation of
an anisotropic plate. In the case of strongly colored substances,
especially yellow minerals, the natural color dominates the field so
effectively that the sensitive tint is no longer present as such and
the slight changes in hue on rotation of the stage are only with dif-
ficulty detected and may be overlooked. The sensitiveness of these
methods varies therefore with the color of the substance under ex-
amination. It is also difficult to obtain a uniformly colored field
under average conditions of illumination, especially near the junc-
tion lines in the field.

In measurements of this nature in which the effective conditions
vary within wide limits, whereas the sensitiveness of the eye (thres-
hold vision) is more or less fixed it is essential for the best results
that the sensitiveness of the device employed for the measurement
be variable so that the most sensitive conditions of observation can
be obtained. It is on this principle of variable sensibility that the
writer's bi-quartz-wedge-plate^^ was constructed.

On substituting the bi-quartz-wedge-plate for the Biot plate of
Koenigsberger or the Biot-Soleil plate of Hanemann the most sensi-
tive arrangement for the detection of anisotropism in opaque sub-
stances is obtained. The method is exceedingly simple and requires
no apparatus in addition to that furnished with a research model
petrographic microscope. The observations are made either in white
or in monochromatic light. The degree of anisotropism is indicated
by the amount of angular rotation of the cap nicol required to pro-
duce equal intensity of illumination in the adjacent halves of the
bi-quartz-wedge-plate.

Other devices, such as the half-shade Lippich prism, may be sub-
stituted for any one of the rotating quartz plates and wedges but

^^ Am. Jour. Sci. (4), 26, 377-378, 1908; Carnegie Institution of Wash-
ington, Publication 138, 140-142, 191 1.

STUDY OF ORES AND METALS. 441

since the apparatus is more complicated and the attainable accuracy
is not increased, such devices are not recommended for practical
work.

Effect of Reflecting Prism or Plate on the Character of Light. —
An extended series of experiments with different kinds of reflectors,
total reflecting prisms, silver surfaces, silvered glass mirrors with
central aperture or a small disk mirror in center of field, or an Abbe
reflecting cube, was carried out to ascertain if possible which type 'is
the best for such work, especially for use with the second group of
methods. The general equations state that in case the azimuth of
the plane of polarization of the incident wave is zero or 90°, the
reflected beam is still plane polarized ; practical experience with such
reflectors states, however, that the reflected beam always shows traces
of elliptical polarization even when the plane of polarization is hori-
zontal or vertical. This is no doubt due in the case of reflecting
glass surfaces to internal reflections ; with a blank metal reflecting
surface, such as the fresh silver side of a silvered mirror, it may re-
sult from strains in the outer film. Be the cause what it may, in no
experiment was the elliptic polarization entirely removed, and the
accuracy of the settings was correspondingly diminished. Expe-
rience with the different types of reflectors did not demonstrate
marked superiority of any one particular type. It is, of course, es-
sential that the plane of polarization of the incident beam be strictly
horizontal or vertical in order to reduce to a minimum the amount of
elliptic polarization present.

Measurement of Phase Differences in Reflected Waves.

The curves of Fig. 6 prove that in general the phase difference
between the components of the reflected waves normal and parallel
to the plane of incidence is of a low order of magnitude, so low in
fact that it may in general be neglected. The actual measurement
of the phase difference is best accomplished by the standard method
with a Babinet compensator. In view of the slight differences to be
observed, however, this method is not to be recommended for prac-
tical diasfnosis.

442 WRIGHT— POLARIZED LIGHT IN THE

The Ejfect of Thin Surface Films on the Character of the Re-
flected Light. — Drude^^ was the first to emphasize the importance
of thin surface films in affecting the polarization relations of re-
flected and of transmitted light waves. In the case of opaque sub-
stances (ores and metals) the surface film effects may be serious
and cause an isotropic mineral to show evident phenomena of aniso-
tropism. Thus a sheet of copper or other isotropic metal shows
distinct anisotropic effects. The rolled surface of the metal is stri-
ated and part of the polarization effects may be of the nature of
grating polarization effects or may result from reflection at oblique
surfaces. Many ground surfaces of isotropic pyrite are decidedly
anisotropic. In certain cases this is probably due, as Koenigsberger
suggested^^ to grinding striae ; but in other pyrite sections the aniso-
tropism persists in spite of care taken to eliminate grinding striae.
This may be the result of surface film effects or of strain bire-
fringence in the pyrite. Its presence in random sections, even care-
fully ground, serves as a danger-signal, warning the observer against
the drawing of too positive conclusions from a single section. As a
general rule pyrite sections are isotropic and the conclusion is war-
ranted that it is isometric, as crystallographic measurements prove it
to be. The statement of Koenigsberger that surface films affect
only the phase difference and not the amplitudes of the reflected
beams is borne out neither by theory nor by practical experience.
In general, however, it may be stated that surface film effects are
not so pronounced as to affect seriously the validity of the deter-
minations.

The Use of Refractive Liquids. — In case the medium surround-
ing the crystal be not air but a refractive liquid the refractive indices
appearing in the above equations have to be reduced in the ratio of
the reciprocal of the refractive index of the liquid ; thus if «„ be the
refractive index of the liquid, n^, Wo of the crystal plate must be re-
placed by Mi/wo, Wj/Wq- With each different refractive liquid em-
ployed both the intensity ratio of the reflected components and the
angular rotation of the plane of polarization are changed. Theo-

"^^ Ann. d. Physik, 43, p. 146, i8t>i ; Lehrbuch d. Optik, 1906, pp. 2721-280;
Winkelmann's Handbuch d. Physik, Vol. 6, pp. 1278-1316, 1906.
15 Centralblatt fiir Miner., Geol. u. PaldontoL, 1908, p. 597.

STUDY OF ORES AND METALS. 443

retically it is thus possible by measurements in different immersion
liquids to determine the refractive indices w^, n^, and also the ab-
sorption indices k^, Ko even with normally incident light; but prac-
tically this method is of little value because the results are encum-
bered with a large probable error due chiefly to the relative insen-
sitiveness of the eye to slight changes in the observed phenomena.
In a transparent, birefracting substance the maximum effect is
obtained by the use of a liquid of refractive index equal either to
Wi or to ^2- The intensity of one of the reflected components is then
zero; the intensity of the second component for tiQ^iii is

Ir \ Wo + Wi / '

which is a very small quantity for weekly birefracting substances ;
and even for strong birefringence such as that of calcite it amounts
only to one half of one per cent. With intense illumination, how-
ever, relatively weak birefringence can be detected by the use of
proper refractive liquids. Since clear refracting Hquids of index
above 2.2 are not available, it is not possible to obtain the most sen-
sitive conditions of observation with minerals of very high refractive
index, such as hematite and selenium. But with most minerals con-
ditions of observation can be improved somewhat by the use of im-
mersion liquids of proper refractive index. In practical work, how-
ever, it is doubtful if the improvement thereby gained is worth the
bother.

Summary.

In the foregoing pages the attempt has been made to present in
.connected form the electromagnetic theory of the reflection of light
by absorbing media and especially that part of the theory which
treats of the reflection phenomena resulting from vertically incident
light waves under the conditions usually encountered in the use of
the reflecting or metallographic microscope ; this outline of the
theory is given in order to indicate and also to emphasize the
fundamental principles on which all methods involving the appli-
cation of polarized light to the study of opaque substances are nec-
essarily based ; thus the possibilities and the limitations as well of

444 WRIGHT— POLARIZED LIGHT IN THE

such methods are ascertained. It is shown in the general case that
vertically incident, plane-polarized light waves become on reflection
elliptically polarized (as a result of a difference in phase between
the components parallel and normal to the plane of incidence) ; that
the amplitude of the component of the light vector in the plane of
incidence is different from that normal to the plane of incidence.

In special cases where crystallographic symmetry relations pre-
scribe the positions of the principal axes of refringence and of
absorption, as in isotropic, uniaxial, and the principal planes of ortho-
rhombic crystals, the above relations are simplified to the extent that
the refracted waves are plane-polarized and not elliptically polarized
as in the general case ; as a result, normally incident, plane-polarized
light waves whose vibration directions are either parallel or normal
to the plane of symmetry are reflected as plane-polarized waves ; but
the intensity of the two reflected waves is different because of the
difference in the refractive and absorption indices parallel and
normal to the plane of symmetry. Therefore normally incident,
non-polarized light contains after reflection a certain amount of
plane-polarized light and this amount increases with the strength of
the birefringence and of the biabsorption in the crystal plate.

The presence of plane-polarized light in essentially non-polarized
light can be detected by'any one of the well-known physical methods,
such as are commonly used in determinations of sky polarization.
Of these methods Koenigsberger adopted the Savart plate with
rotating glass compensator ; a second method is suggested above
which employs either a single calcite cleavage plate with proper ap-
erture (after the manner of the Haidinger lens or the Pickering
photometer) or a small portable Koenig-Martens photometer. This
method is superior to the first method in two respects : it is simpler
in adjustment and in manipulation; it is based on a photometrically
better principle, namely contrast of two or a series of illuminated
fields, the first at minimum intensity, the adjacent field at maximum
intensity with a sharp line of demarcation which disappears when
the intensity of illumination in the adjacent fields is the same. The
attainable accuracy depends on the least perceptible difference in in-
tensity of illumination which the eye can detect; this is about 1.5 to
2 per cent, under ordinary conditions of illumination. An opaque

STUDY OF ORES AND METALS. 445

crystal, therefore, of such weak birefringence or weak biabsorption
such that the difference in intensity between the reflected components
is less than 1.5 per cent, appears isotropic. In transmitted light dif-
ferences in birefringence of only 0.02 per cent, of the refractive
index are readily detectible ; therefore the methods based on dif-
ferences in intensity of the reflected components of vertically inci-
dent light are 50 or more times less sensitive in the detection of
anisotropism than the methods based on the phenomena of plane-
polarized, transmitted light waves.

In case the vertically incident light be plane-polarized, the dif-
ference in amplitude of the reflected components normal or parallel
to the plane of symmetry causes a rotation of the plane of polariza-
tion and this can be detected and measured by any one of a number
of devices in common use by petrologists. Of these Koenigsberger
adopted a Biot sensitive tint plate and used it in a qualitative way to
detect anistropism. Hanemann improved Koenigsberger's method
by adopting the Biot-Soleil sensitive tint bi-plate and obtained
thereby a color contrast in two adjacent fields. The ordinary Bert-
rand eyepiece which is in common use by petrologists is suggested
above as a still further and equally simple method for detecting
anisotropism.

A still more sensitive and better method is, however, to use the
writer's bi-quartz-wedge-plate in which the sensibility is variable and
can be adjusted to meet the conditions of illumination and thus to
produce the most favorable conditions for extreme sensitiveness and
consequent accuracy of results of measurement. This method is
simple and is the most sensitive at present available for detecting and
measuring anisotropism in opaque substances. The degree of aniso-
tropism is indicated by the amount of angular rotation on reflection
of the plane of polarization of the vertically incident light waves.
The accuracy of these methods depends on the threshold limit of
vision (least intensity of light which the eye can detect) ; this is dif-
ferent from the least perceptible difference in intensity of illumina-
tion of two adjacent fields on which the sensitiveness of the first
group of methods depends.

The sensitiveness of the second group of methods can be in-
creased by increasing the intensity of the light source and also by

446 WRIGHT— POLARIZED LIGHT IN THE

introducing some contrast device such as the bi-quartz-wedge-plate,
or the Lippich biprism. In general it may be stated that for practical
purposes the methods of the second group are simpler in manipula-
tion than those of the first and more sensitive and therefore better.
A disturbing factor is the introduction of a small amount of ellip-
tically polarized light by the reflecting prism or plate. The phase
differences between the two components of the reflected waves are
in general small and do not seriously affect measurements by the
methods outlined above which are based on the amplitude differences
between the two components. The phase differences may be meas-
ured by one of several devices of which the Babinet compensator is
the best and most widely used ; but for the detection and measure-
ment of anisotropism, methods, which are based on phase differences
between the components after reflection of vertically incident plane-
polarized light, are not in most cases of practical value.

The use of immersion liquids and a monochromatic light source
is suggested for obtaining the greatest differences in intensity be-
tween the reflected components and thereby increasing the sensitive-
ness of the methods for detecting anisotropism.

The above equations show that it is in general not possible to
measure both the refractive indices and the absorption indices of an
opaque body by means of the phenomena resulting on the reflection
of vertically incident light waves ; but it is possible in special cases to
determine the degree of anisotropism and also the positions of the
polarization directions in a given crystal plate. These data taken in
conjunction with crystallographic data enable the observer, just as
do birefringence and extinction angles in transparent crystals, to
draw conclusions regarding the crystal system of the substance under
investigation. But in opaque substances the precision attainable is
relatively slight and the phenomena which can be observed are rela-
tively few and restricted in scope. As a result one cannot expect
from the application of polarized light to such substances the harvest
of optical data which have been gathered from transparent crystals.
For opaque bodies the possibilities are few, the limitations are great,
and recourse must be had in practical diagnosis to other methods of
determination such as behavior of the substance toward abrasives

STUDY OF ORES AND METALS. 447

and solvents, color, density, hardness, etc., in addition to the deter-
mination of the degree of anisotropism.

In this paper the properties of a number of opaque substances
such as ores and metals, determined by the several methods de-
scribed above, are not included ; such lists would add materially to its
length and are not germane to its purpose.
Geophysical Laboratory,

Carnegie Institution of Washington,
Washington, D. C,
March i, 1917.

ON THE luciopimelodin;e, a new subfamily of

THE SOUTH AMERICAN SILURID^.^

(Plates II and III.)

By CHAS. S. driver.

(Communicated by Prof. Carl H. Eigenmann. Read October 3, 1919.)

In the Nematognathi or catfishes and their relatives the anterior
vertebrae are coalesced and form with the auditory ossicles the so-
called Weberian apparatus. This structure has been studied by
various authors notably Sagemehl," R. Ramsay Wright^ and Bridges
and Haddon.* A brief summary is also given by Boulenger in his
" Fishes of the Congo " and in other papers on systematic ichthy-
ology by Regan, Eigenmann, et al. Usually the centra of the first
three vertebrae are coalesced into one piece without evident sutures.
The centra of the fourth, fifth and sixth vertebrae are frequently
coalesced with the anterior three, but the sutures between the third
and fourth as well as between the fourth and fifth are frequently
quite evident, especially in the 3^oung.

The coalesced first to third vertebrae are without lateral processes
The fourth and fifth have various processes.

As in all other members of the Nematognathi, the centra of the
anterior vertebrae of the species of the Pimelodinae, a subfamily of
South American catfishes, are coalesced. The lateral processes of
the fourth, fifth and sometimes the sixth vertebrae are expanded,
sometimes largely separated from each other, more frequently joined
to each other to form an osseous shield or roof over the anterior part
of the abdominal cavity.

In Rhamdia of the Pimelodince there are three processes distinct

1 Contribution from the Zoological Laboratory of Indiana University,
No. 166.

- Morpholog. Jahrb., Bd. X.

3 Trans. Roy. Soc. Canada, 1885.

* Phil. Trans. Roy. Soc, B, 1893.

448

DRIVER— THE LUCIOPIMELODIN^. 449

from each other for more than their distal halves (Plate II, B, 9, 10,
5). The first one articulates distally with the clavicle and process
of the post-temporal, the second and third are free at the ends. The
first and second (9 and 10) belong to the fourth vertebra,^ the third
process belonging to the fifth vertebra. In Pseudoplatystoma fas-
ciatum there are four expanded transverse processes, all of them
separated from each other toward their distal ends. In small speci-
mens of this species the interlocking sutures of the centra of the
coalesced vertebrae are quite distinct. In this species it is quite evi-
dent that the first and second processes belong to the second distin-
guishable vertebra, in reality to the fourth. The fourth process
belongs to a supernumerary modified vertebra, the sixth, one usually
normal in other species. Its processes bear ribs. In Pinielodus
clarias the lateral processes are united to form a solid plate without
even marginal notches to indicate the boundaries between successive
processes.

In the genera Pimclodina, Pseudopimolodiis, Rhamdia, Pimelodiis,
etc., of the Pimelodinee the air-bladder*' is well developed and lies
free in the abdominal cavity. It is simple and more or less oblong
in shape and extends posteriorly to the eighth or tenth vertebra.

In Sornhim (Fig. i. A) and Hemisoriibim the air-bladder is
drawn out more than in any of the other genera, and extends back
past the tenth vertebra. It is larger proportionally than in any of the
other genera. In Pseudoplatystoma and BracJiyplatystoma it is very
much as in Sortibim and Hemisoruhim but slightly shorter, extend-
ing only past the ninth vertebra.

In Rhamdia (Fig. i, B), Pimelodella and Pimelodus (Fig. 1, C)
the air-bladder is very dift'erent from that found in Soruhim, Hemi-
soriibim, Pseudoplatystoma and BracJiyplatystoma. In these genera
it is wider than long. The width averaging from one to three mm.

5 Bridges and Haddon, PJiil. Trans., Vol. B, 1893.

6 The air-bladder of one specimen of Ageneiosus caucanus has been
examined to determine its relation to the Pimelodinge, and to the genera
Megalonema and Luciopimelodus. This speciinen was 280 mm. long and' its
air-bladder was 23 mm. long and 23 mm. wide. It tapers at the posterior
end, causing it to be heart-shaped. It lies free in the abdominal cavity and.
is not enclosed in a bony capsule. It thus resembles the condition found in
the Pimelodinse.

450 DRIVER— THE LUCIOPIMELODIN^.

in excess of the length. In Rhamdia it extends to the ninth vertebra
while in Pimclodclla and Pimelodus it extends only to the eighth.

As stated, the lateral processes of the first five vertebrae in the
Pimelodinse are coalesced and form a more or less smooth roof be-
neath which the air-bladder lies. In Rhamdia (Plate II, Fig. B)
the transverse processes are quite distinct, and are separated from
each other ; the coalescing is more complete in Soruhiiim, Heniiso-
ruhim, Pimelodella and reaches the climax in Pimelodus.

ABC

Text- Fig. i. A, Air-bladder of Sorbium lima; B, Air-bladder of Rhamdia
cinarescens ; C, Air-bladder of Pimelodus clarias. Outlines as seen from
below.

The air-bladder of Pimelodus is small. The coalescing of the
vertebrae is more complete in this genus than in others. The outer
edge of the shield formed by the coalesced process has a tendency to
become arched antero-posteriorly.

The air-bladder in all these genera comes in close contact with
the skin which serves as a tympanic membrane. This can be seen
especially well in young specimens.

Perugia (Plate 11,^, C,D,E), which had hitherto been placed in
the Pimelodinae, has the air-bladder very much reduced in size, con-
stricted longitudinally with one lobe on each side of the centra of the
coalesced vertebrae. Each lobe is surrounded by a partial capsule,
made up of the coalesced vertebrae above and behind, the scapula
and the process connecting the scapula with the basioccipital in
front. The floor of the capsule is membranous.

DRIVER— THE LUCIOPIMELODIN^. 451

The two lobes of the air-bladder are very small. In a fish 200
mm. long they are 2.5 mm. in diameter and the distance between the
outer margins of the two capsules measures 12 mm. They are pear-
snaped, tapering to a small canal connecting the two capsules.

Liiciopimelodiis and Megaloncma are certainly closely related to
Perugia. This becomes quite evident when we examine the air-,
bladder and the anterior vertebrae associated with it. An examination
of these structures also demonstrates that Luciopimclodus and
Megalonema are not related to the Pimelodinas with which they have
been associated.

The air-bladders in these two genera are not indirect contact with
the skin. They extend outward to near the skin, being separated
from it by a thin layer of fat.

As in the structure of the air-bladder and the Weberian apparatus
Perugia, Luciopimelodus and Megaloncma differ greatly from all
the genera of the Pimelodince, they may well be separated into a new
subfamily, the Luciopimclodincv.

Text-Fig. 2. Megaloncma platanum Giinther. Photograph by Sur. Valette,
Jefe de la Oficina Fomento de Pesca. Buenos Aires.

In Cuvier and Valenciennes, "Hist. Nat. Poissons," XV, 1840,
p. 176, Valenciennes described a new species which he called Pime-
lodus pati basing his description on a specimen 30 inches long col-
lected by d'Orbigny, a very small specimen which he received from
the museum at Lisbon, and on a drawing in the MS. of Father Feuil-
lee, in Hugard's library. Later in d'Orbigny's " Voyage dans I'Amer-
ique Meridionale," V, 2c Partie ; Poissons, 1847, P- 7. plate I., Fig.
7-9, Valenciennes figured the species.

452 DRIVER— THE LUCIOPIMELODIN^.

D'Orbigny reported that the species is taken in the Parana from
latitude 26 degrees South to Buenos Aires. He reported it to be a
permanent resident at Corrientes, but at Buenos Aires it is migratory,
arriving in September and departing in March.

Giinther in his Catalogue also placed it in the old genus Pime-
lodus which includes a number of different generic types. In a
short article on the Fish-fauna of Rio de la Plata/ Giinther briefly
describes Pimelodiis platanns. This species is without spots and it
is quite possible that it was included in the species patl concerning
which Valenciennes says (p. 178) :

M. d'Orbigny nous dit que les taches varient a I'infini, et disparaissent
quelquefois entierement.

Eigenmann and Eigenmann^ created the genus Liiciopimelodus
for pati and platanus placing the genus in the Pimelodinse. They
distinguished the genus by the flexible, non-spinous first dorsal and
first pectoral rays, the elongate, spatulate, depressed head, the nar-
row, short occipital process, and the fontanel confined to the frontal
portion of the head.

Steindacher^ described Pirinampus agassizii. The relationship
of this species with pati was not recognized by Steindachner or
Eigenmann and Eigenmann, the latter placing it in a hypothetical
distinct but unnamed genus numbered XXII in their monograph on
the Nematognathi. This was later placed in a new genus, Perugia,
by Eigenmann and Norris. Steindachner described this species as
having a pungent dorsal spine, but Fisher^*^ shows that in a specimen
in the Carnegie Museum (N^o. 7251) there is evidence that the pun-
gency is due to the fact that the tip of the first dorsal ray is broken ofif.

In his monograph on the Freshwater Fishes of British Guiana^^
Eigenmann describes the new genus and species Megalonema platy-
cephalnm from the lower Potaro River of British Guiana. The
genus Megalonema differs from Lticiopiiiielodus in the shape of the

''Ann. and Mag. Nat. Hist., Vol. VI., July, 1880, p. 10.
^ Proc. Calif. Acad. Sci. (2), I., 1888, p. 122, and Occasional Papers Calif.
Acad. Sci., I., 1890, p. 106.

9 6". B. Ak. IVien., LXXIV, 1876, Ichthyol. Beitr., IV, PI. 12.
^'^ Ann. Carnegie Mus., XI, 1917, p. 407.
^1 Mem. Carnegie Mtts., V, 1912, p. 150.

DRIVER— THE LUCIOPIMELODIN^. 453

band of premaxillary teeth. The band is rounded at the outer end
instead of prolonged into a sharp angle as in pati.

Still later Eigenmann^- described Megaloneina xanthum from
the upper Magdalena River of Columbia.^^

Key to tpie Genera and Species of Luciopimelodin.^.

a. Mouth terminal, the lower jaw pointed, its tip extending to the tip of the
upper jaw; premaxillary band of teeth with a long, backward project-
ing angle ; adipose fin beginning at tip of last dorsal ray, its length little
more than three in the length ; snout greatly depressed ; depth of head
three in its length ; post-mental barbels extending past base of the anal,
A. 12; eye 10 in head, 4.5 in snout, 2.33 in interorbital. (Based on a

specimen 230 mm. to base of caudal.) (Luciopimelodus.) pati.

aa. Jaws subequal, premaxillary band of teeth without a backward projecting
angle.
b. Space between adipose and base of last dorsal ray considerably greater

than base of dorsal (Megalonema.)

c. Head 4: depth 6; adipose 4 in length; dorsal I, 6; A. 12; eye 2.5

in snout, 5 in head platamim.

cc. Head 3.66 in the length ; depth 5.5 ; adipose 4 in the length ; dorsal

I, 6 ; A. II ; eye 2 in the snout, 5 in head platyccphalum.

ccc. Head 4; depth 5.33; A. 9; adipose 4.95 in the length; eye 4 in
the head, .75 in the interorbital; snout depressed, width of
head at the rictus 2 in the length of the head; maxillary barbel
reaching middle or beyond tip of pectoral ; teeth in a very
narrow band ; fontanel reaching base of occipital crest which
is minute ; the first dorsal ray slender, a little longer than the
head. (Based on a specimen 38 mm. in length.)

pauc'iradlatum.

bb. Space between adipose and base of last dorsal ray much shorter than

the base of the dorsal, post-mental barbel not far behind the

mental (Perugia.)

d. Lower jaw but a little shorter than the upper, premaxillary band
of teeth terminal, partly exposed when mouth is closed; fon-
tanel increasing in width backward to behind the eye ; a small
fontanel at the base of the occipital process ; head 5 in the

length ; depth 6 ; dorsal I, 6 ; A. 11 agassi::ii.

dd. Snout much projecting, the mouth inferior; the premaxillary band
of teeth very narrow; maxillary barbel extending past base
of anal ; adipose 2.5 in the length ; fontanel wider behind than
in front, extending to the posterior edge of eye ; small fon-
tanel at the base of the occipital process; A. 11 xanthus.

12 Ind. Univ. Studies, No. 16, Dec, 1912, p. 16.

^3 The species of Megalonema described by Meek from the Tuyra River
belong to the genus Pimclodus.

454 DRIVER— THE LUCIOPIMELODIN^.

The species of Lucio pimclodincc have so far been recorded from
the La Plata Basin, from the Rio Jauri at the headwaters of the
River Paraguay to the mouth of the La Plata at Buenos Aires, from
the Amazon between Calderon and Para, from the Essequibo Basin
in Guiana and from the Magdalena Basin in transandean Colombia.
Their distribution is very wide which may indicate that they have
been established in South America since before the days of the for-
mation of the eastern Andes of Colombia.

Synonymy and Bibliography.

LuciOPiMELODUS PATi ( Valcncienncs) .

Pimelodiis pati Valenciennes, Voy. d'Orbigny, pi. i, figs. 7-9,
1847; Cuv. & Val., Hist. Nat. Poiss., XV, 176, 1840 (Parana;
La Plata ; Corrientes ; Buenos Ayres) ; Kner, Sb. Ak. Wien,
XXVI, 1857, 416 (Forte de S. Joaquim, Rio Branco) ;
Giinther, Cat. Fish. Brit. Mus., V, 128, 1846 (copied).

Luciopimelodus pati Eigenm. & Eigenm., Proc. Calif. Acad., 2d
Ser., Vol. I, 122 (Buenos Ayres) ; Eigenmann, Repts. Prince-
ton Univ. Expcd. Patagonia, III, 1910, p. 383 (name only) ;
Fisher, y^wM. Carnegie Mus., XI, 1917, p. 407 (Buenos Aires).

Habitat: Rio Plata ; Rio Branco near British Guiana.
Megalonema platanum (Giinther).

Pimclodiis platanus Giinther, Ann. Nat. Hist. (5), VI, 10, 1880
(Parana, Rio Plata).

Luciopimelodus platanus Eigenm. & Eigenm., Occasional Papers
Calif. Acad. Sci., I, p. 108, 1890; Eigenmann, Repts. Prince-
ton Univ. Exped. Patagonia, III, 1910, p. 383 (name only) ;
Fisher, Ann. Carnegie Mus., XI, 191 7, p. 408 (Rio Jauru,
Asuncion).

Habitat: Rio Plata; Piracicaba ; Paraguay Basin.
Megalonema platycephalum (Eigenmann).

Megalonema platycephalum Eigenmann, Repts. Princeton Univ.
Exped. Patagonia, III, 1910, p. 383 (name only) ; Eigen-
mann, Mem. Carnegie Mus., V, 1912, p. 150, PI. X, jfig. 2.

Habitat: Essequibo Basin.

DRIVER— THE LUCIOPIMELODIN^. 455

Megalonema pauciradiatum Eigenmann, spec. nov.

15029, I. U. M., four, the largest the type, 28-38 mm. Villa Rica.
Anisits.

Head 4 ; depth 5.33 ; A. 9 ; adipose 4.75 in the length ; eye 4 in the
head, .75 in the interorbital ; maxillary barbel reaching middle or tip
of dorsal, postmental a little beyond tip of pectoral ; teeth in very
narrow bands ; snout depressed, width of head at the rictus 2 in the
length; depth of head nearly two-thirds its length; fontanel reaching
base of occipital crest which is minute, reaching only about one-sixth
of the way to the dorsal fin ; distance between snout and dorsal
2-2.66 in the length ; first dorsal ray slender, a little longer than the
head, extending beyond all the rest when depressed ; ventrals and
base of third dorsal ray about equidistant from snout ; first pectoral
ray slender, reaching the ventrals, the latter not to the anal ; caudal
peduncle 5 in the length.

Air-bladder transverse, about three times as wide as long.

Pale, without any conspicuous markings.

The type is a female with mature eggs nearly a millimeter in
diameter. While the specimens at hand have probably not reached
full size, the largest is sexually mature.

This species differs from Megalonema platanum and Lnciopiine-
lodus pati of the Plata basin in having but nine anal rays while they
have twelve.
Perugia agassizii (Steindachner).

Pirinampus agassizii Steindachner, Sb. Ak. Wicn, LXXIV,
1876; Ichthyol. Beitr., IV, 57, pi. 12 (Para) ; Vaillant, Bull.
Soc. Philoiii., series 7, IV, 153, 1880 (Calderon).

^ agassizii Eigenm. & Eigenm., Occasional Papers Calif. Acad.
Sci., I, p. 183, 1890 (Amazons).

Perugia gen. nov. Eigenmann & Norris, Revista Miiseu Paulista,
IV, p. 355, 1900.

Luciopimelodus agassizii Eigenmann, Repts. Princeton Univ.
Exped. Patagonia, III, 1910, p. 383 (name only). Fisher,
Ann. Carnegie Mus., XI, 1917, p. 407 (Para).

Habitat: Amazons.

456 DRIVER— THE LUCIOPIMELODIN^.

Perugia XANTHUS (Eigenmann). ^

Megalonema xanthwm Eigenmann, Indiana University Studies,

No. i6, p. i6, 1912.
Habitat: Girardot; Apulo.

MAGELLANIC PREMIUM

Founded in 1786 by John Hyacinth de Magellan, of London

1920

THE AMERICAN PHILOSOPHICAL SOCIETY

Held at Philadelphia, for Promoting Useful Ky^towledge

ANNOUNCES THAT IN

DECEMBER, 192O

IT WT'-l. AWARD ITS

MAGELLANIC GOLD MEDAL

TO THE author of THE BEST DISCOVERY, OR MOST USEFUL INVENTION, RE-
LATING TO NAVIGATION, ASTRONOMY, OR NATURAL PHILOSOPHY (mERH
NATURAL HISTORY ONLY EXCEPTED) UNDER THE FOLLOWING CONDITIONS :

1. The candidate shall, on or before November i, 192O1 deliver free of postage
or. other charges, his discovery, invention or improvement^ addressed to the
President of the American Philosophical Society, No. 104 South Fifth Street,
Philadelphia, U. S. A., and shall distinguish his performance bv some motto,
device, or other signature. With his discovery, invention, or improvement, he
shall also send a sealed letter containing the sam*; motto, device, or other sig-
nature, and subscribed with the real name and place of residence of the author.

2. Persons of any nation, sect or denomination whatever, shall be ad-
mitted as candidates for this premium.

3. No discovery, invention or improvement shall be entitled to this premium
which hath been already published, or for which the author hath been publicly
rewarded elsewhere.

4. The candidate shall communicate his discovery, invention or improvement,
either in the English, French, German, or Latin language.

5. A full account of the crowned subject shall be published by the Society,
as soon as may be" after the adjudication, either in a separate publication, or in
the next succeeding volume of their Transactions, or in both.

6. The premium shall consist of an oval plate of solid standard gold of the
value of ten guineas, suitably inscribed, with the seal of the Society annexed to
the medal by a ribbon.

t^W correspondence in relation hereto should be addressed
To THE Secretaries of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A*

Proceedings

OF THE

American Philosophical Society

Subscription — Three Dollars per Annum

General Index to the Proceedings

Volumes 1-50 (1838-1911)

Lately Published

Price, One Dollar

TRANSACTIONS

OF THE

American Philosophical Society

HELD AT PHILADELPHIA

For Promoting Useful Knowledge

JVetu Series, Vol. XXII,

Fart Illy 4to, 44 pages,

(Lately Published)

Tertiary Vertebrate Faunas of the North CoaMnga Region of Cali-
fornia. A Contribution to the Study of the Palaeontologic
Correlation in the Great Basin and Pacific Coast
Provinces. ByJoHNC. Merriam, Pro-
fessor of Palaeontology, Uni-
versity of California.

Subscription— Five Dollars per Volume
Separate parts are not sold

flddresi The Librarian of the

AMERICAN PHILOSOPHICAL SOCIETY
No. 104 South Fifth Street

PHILADELPHIA, U. S. A.

MINUTES.

MINUTES.

Stated Meeting, January j, 1919-

William B. Scott, D.Sc, LL.D., President, in the Chair.

The decease was announced of
Charles Francis Himes, Ph.D., at CarHsle, Pa., on December 7,

1918, aet. 81.
Lewis S. Ware, at Paris, on December 20, 1918, set. 67.
Mr. Leon Dominion, of New York, read a paper on " The
Peoples of Central and Eastern Europe," which was discussed by
Dr. Jastrow, Dr. Scott and Prof. Lingelbach.

The Judges of the Annual Election held this day between the
hours of two and five in the afternoon reported that the following
named members were elected according to the Laws, Regulations
and Ordinances of the Society to be the Officers for the ensuing year.

President.
William B. Scott.

Vice-Presidents.
George Ellery Hale,
Arthur A. Noyes,
Hampton L. Carson.

Secretaries.
1. Minis Hays,
Arthur W. Goodspeed,
Harry F. Keller,
Bradley Moore Davis.

Curators.
Charles L. Doolittle,
William P. Wilson,
Leslie W. Miller.

iv MINUTES.

Treasurer.
Henry La Barre Jayne.

Councillors.
(To serve for three years.)
Maurice Bloomfield,
John M. Clarke,
George H. Parker,
Arthur G. Webster.

Stated Meeting, February 2, ipip.

William B. Scott, D.Sc, LL.D., President, in the Chair.

The decease was announced of
Rossiter Worthington Raymond, Ph.D., LL.D., at New York,

on December 31, 1918, set. 78:
Theodore Roosevelt, LL.D., at Oyster Bay, on January 6, 1919,
set. 60.
~ Edward C. Pickering, LL.D., L.H.D., at Cambridge, on Feb-
ruary 3, 1919, set. /2.
The following papers were read :
" The League of Nations," by Russell Duane.
" Political Problems of Central and Southern Europe," by

Charles D. Hazen of New York. (By title.)
" The Eastern Question," by Prof. Morris Jastrow, Jr.
" Organized Labor and Peace," by Prof. E. P. Cheyney.

Stated Meeting, March 7, 1919.

William B. Scott, D.Sc, LL.D., President, in the Chair.

The decease was announced of
George Carey Foster, F.R.S.. D.Sc, at Herts, England, on Feb-
ruary 9, 1919, set. 83.
Hon. George Franklin Edmunds, at Pasadena, Calif., on Feb-
ruary 27, 1919, set. 91.
Charles L. Doolittle, ScD., LL.D.. at Philadelphia, on :^Larch
3, 1919, set. 76.
Mr. Langdon Warner, introduced by the President, read a paper

MINUTES. V

on " Siberia and the Czechs," which was discussed by Mr. Goodwin
and President Scott.

Stated Meeting, April 4, 1919.
William B. Scott, D.Sc, LL.D., President, in the Chair.

The following papers were read :

" Present Tendencies in the National Development of Water-
ways and Water Traffic, with Reminiscences of the Service
rendered by the American Philosophical Society in behalf of

• early Canal Construction in America," by Wilfred H. Schoff.
(Introduced by Dr. W. P. Wilson.)

" What Shall be Done with the Railroads ? " by Emory R. John-
son, Sc.D.

Stated General Meeting, April 24, 25, 26, 1919.

Thursday Afternoon, April 24, 1918.

Opening Session, 2.30 o'clock.

William B. Scott, D.Sc, LL.D., President, in the Chair.

Dr. Alfred G. Mayor, a lately-elected member, having subscribed

the Laws, was admitted into the Society.
The decease was announced of

Sir William Crookes, Kt., O.M., LL.D., D.Sc, at London, on
April 4, 1919, set. 87.

Henry Morse Stephens, B.A., M.A., Litt.D., at San Francisco,
on April 16, 1919, set. 61.

George Ferdinand Becker, Ph.D., at Washington, on April — ,
1919, set. y2.
The following papers were read :

" The Cosmic Force, Radio- Action," by Monroe B. Snyder, Di-
rector of the Philadelphia Observatory.

" The Conservation of Natural Alonuments," by John M.
Clarke, Ph.D., Sc.D., LL.D., Director of Department of Sci-
ence and State Museum, Albany, N. Y.

" Detection of Ocean Currents by their Alkalinity," by Alfred
G. Mayor, M.E., Sc.D., Director of Department of Marine

vi MINUTES.

Biology, Carnegie Institution of Washington, Princeton, N. J.

" Some Oceanographical Results of the Canadian Arctic Ex-
pedition, 1913-1918," by Vilhjalmur Stefansson, A.B., Com-
mander of Canadian Arctic Expedition. (Introduced by
Mr. Henry G. Bryant.) (By title.)

" Evolution and Mystery in the Discovery of America," by
Edwin Swift Balch, A.B., of Philadelphia.

" Benjamin Franklin's Art as Applied to Books for Elementary
Teaching," by Charles R. Lanman, Ph.D., LL.D., Prof, of
Sanskrit, Harvard Universtiy. (By title.)

" The Energy Loss of Young Women during Light Household
Muscular Activities," by Francis G. Benedict, Ph.D., Sc.D.,
Director of Nutrition Laboratory (Boston) of Carnegie In-
stitution of Washington, and Alice Johnson.

" The Relative Contribution of the Staple Commodities to the
National Food Consumption," by Raymond Pearl, Ph.D.,
Professor of Biometrics, School of Hygiene and Public
Health, Johns Hopkins University.

" Hygiene and Sanitation as Improvised in the Zone of Opera-
tions during the Great War," by Bailey K. Ashford, M.D.,
Sc.D., Surgeon U. S. Army. (Introduced by Dr. W. W.
Keen.) Discussed by Prof. Osborn.

" Bloodless Removal of Foreign Bodies from the Lungs through
the Mouth by Bronchoscopy," by Chevalier Jackson, M.D.,
Attending Laryngologist, Jefferson Medical College, Phila-
delphia. (Introduced by Dr. W. W. Keen.) Discussed by
Dr. Keen.

Friday, April 2j, ipig.

Executive Session, p.jo o'clock

Proceedings of the Officers and Council was submitted.

Morning Session, 10 o'clock.

William B. Scott, D.Sc, LL.D., President, in the Chair.

The following papers were presented :
" The New Discoveries of Extinct Animals in the West Indies
and their Bearing on the Geological History of the Antilles,"

MINUTES. vii

by William D. Matthew, A.M., Ph.D., Curator of American
Museum of Natural History, N. Y. Discussed by President
Scott.

" Characters and Restoration of the Sauropod Genus Camara-
sanriis Cope, from Type-Material in the Cope Collection of
the American Museum of Natural History," by Henry Fair-
field Osborn, Sc.D., LL.D., Research Professor of Zoology,
Columbia University, and Charles C. Mook. Discussed by
President Scott.

"Energy Conception of the Cause of Evolution," by Henry
Fairfield Osborn, Sc.D., LL.D., Research Professor of Zool-
ogy, Columbia University.

" The Parasitic Aculeata, a Study in Evolution," by William M.
Wheeler, Ph.D., Sc.D., Professor of Economic Entomology,
Bussey Institution, Harvard University.

" Two Recent Entomological Problems — -The Pink BoUworm
and the European Corn Borer," by L. O. Howard, Ph.D.,
LL.D., Chief of Bureau of Entomology, U. S. Department
of Agriculture, Washington.

Hampton L. Carson, M.A., LL.D., Vice-President, in the Chair.

" Hydration and Growth," by D. T. MacDougal, Ph.D., LL.D.,
Director of the Department of Botanical Research, Carnegie
Institution of Washington, Tucson, Arizona.

" Hydration of Colloids and Cell-Masses in Propionic Acid and
its Amino-Compounds," by D. T. MacDougal, Ph.D., LL.D.,
Director of the Desert Laboratory, Tucson, Arizona, and H.
A. Spoehr. (By title.)

" Sterility and Self-and-Cross-Incompatibility in Shepherd's
Purse," by George H. Shull, Ph.D., Professor of Botany and
and Genetics, Princeton LTniversity.

" The Basis of Sex Inheritance in Sphccrocarpos," by Charles
E. Allen, Ph.D., Professor of Botany, University of Wiscon-
sin. (Introduced by Prof. Bradley M. Davis.)

" Hydrogen-ion Concentration of Nutrient Solutions in Rela-
tion to the Growth of Seed Plants," by Benjamin M. Duggar,
A.M., Ph.D., Research Professor of Plant Physiology, Mis-

viii MINUTES.

souri Botanical Garden, St. Louis. (Introduced by Prof.
Bradley M. Davis.)
"The Relation of the Diet to Pellagra," by E. V. McCollum,
Ph.D., Professor of Bio-Chemistry, Johns Hopkins Univer-
sity. (Introduced by Dr. Henry H. Donaldson.)

Afternoon Session^ 2 o'clock.

George Ellery Hale, Ph.D., Sc.D., LL.D., Vice-President, in the

Chair.

The following papers were presented :

" The Eclipse Expedition from the Lick Observatory. Some
Solar Eclipse Problems," by W. W. Campbell, Sc.D., LL.D.,
Director of the Lick Observatory, Mount Hamilton, Calif.

" The Expedition of the Mount Wilson Observatory to the So-
lar Eclipse of June 8, 1918," by J. A. Anderson, Ph.D., Mount
Wilson Solar Observatory, Pasadena, Calif. (Introduced
by Prof. John A. Miller.)

"The Lowell Observatory Eclipse Observations, June 8, 1918.
Prominences and Coronal Arches," by Carl O. Lampland,
A.M., Lowell Observatory, Flagstaff, Arizona. (Introduced
by Prof. Eric Doolittle.)

"The Flash Spectrum," by Samuel Alfred Mitchell, Ph.D.. Di-
rector McCormick Observatory, University of Virginia.
(Introduced by Prof. John A. Miller.)

" Electric Photometry of the 1918 Eclipse," by Jacob Kunz,
Ph.D., and Joel Stebbins, Ph.D., University of Illinois, Ur-
bana. 111. (Introduced by Prof. John A. Miller.)

" The Sproul Observatory Eclipse Expedition. The Form of
the Coronal Streamers," by John A. Miller, A.M.. Ph.D., Di-
rector of the Sproul Observatory, Swarthmore College, Pa.

" Results of Observations of the Eclipse by the Expedition
from the Yerkes Observatory," by Edwin B. Frost, D.Sc,
Professor of Astrophysics and Director of Yerkes Observa-
tory, University of Chicago.

" Self-Luminous Night Haze," by E. E. Barnard, D.Sc, LL.D.,
Professor of Practical Astronomy, University of Chicago.

MINUTES. ix

" Photometric Measurements of Stars," by Joel Stebbins, Ph.D.,
Professor of Astronomy, University of Ilhnois. (Intro-
duced by Prof. Henry Norris Russel.) (By title.)

" Star Clusters and their Contribution to Knowledge of the
Universe," by Harlow Shapley, Ph.D., Mt. Wilson Solar
Observatory, Pasadena, Calif . (Introduced by Prof. George
E. Hale.)

" Tatar Material in Old Russian," by J. Dyneley Prince, Ph.D.,
Professor of Slavonic Language, Columbia University.

Evening Session, 8.^0 o'clock.

Arthur Gordon Webster, Sc.D., LL.D., Professor of Physics,
Clark University, Worcester, spoke on the " Recent Applications of
Physics in Warfare."

Saturday, April 26, ipip.

Executive Session, p.^o o'clock.

WiLLLiAM B. Scott, D.Sc, LL.D., President, in the Chair.

Certain proposed amendments to the Laws and Rules of Admin-
istration and Order and Resolutions in reference to amendments to
the Charter, recommended by the Officers and Council were adopted.
Pending nominations for membershp were read and the Society
proceeded to an election. The tellers reported that the following
nominees had been elected to membership :
Residents of the United States:

Robert Grant Aitken, Sc.D., Mount Hamilton, Cal.

Joseph Charles Arthur, Sc.D., Lafayette, Ind.

Edward W. Berry, Baltimore.

James Henry Breated, A.M., Ph.D., Chicago.

Ulric Dahlgren, M.S., Princton.

WilHam Curtis Farabee, A.M., Ph.D., Philadelphia.

John Huston Finley, LL.D., Albany, N. Y.

Stephen Alfred Forbes, Ph.D., LL.D., Urbana, 111.

Chevalier Jackson, M.D., Philadelphia.

Dayton C. Miller, A.M., D.Sc, Cleveland.

George D. Rosengarten, Ph.D., Philadelphia.

X MINUTES.

Albert Sauveur, S.B., Cambridge, ]\Iass.
William Albert Setchell, A.M., Ph.D., Berkeley, Cal.
Julius O. Stieglitz, Ph.D., D.Sc, Chicago.
Ambrose Swasey, Sc.D., D.E., Cleveland.

Morning Session, lO o'clock.

William B. Scott, D.Sc, LL.D., President, in the Chair.

Prof. James Henry Breasted, a newly elected member, subscribed
the Laws and was admitted into the Society.
The following papers were read :

" Artificial Formations Resembling Lunar Craters," by Captain
Herbert E. Ives, B.S., Ph.D., of Philadelphia. Discussed by
Prof. Magie, Mr. Davis and Prof. Webster.

" The Meteorological Service of the Signal Corps in the War,"
by Robert A. Millikan, Ph.D., Sc.D., Professor of Physics,
University of Chicago.

" Detection of Submarines," by Harvey Cornelius Hayes, Ph.D.,
Naval Experimental Station, New London. (Introduced by
Prof. John A. Miller.) Discussed by Prof. Webster.

" Errors Induced in Bullets by Defects in their Manufacture,"
by Ernest W. Brown, M.A., Sc.D., Professor of Mathematics,
Yale University. Discussed by Messrs. Akimoff, Webster,
Brush and Brown.

" Sound and Flash Ranging," by Augustus Trowbridge, A.AI.,
Ph.D., Professor of Physics, Princeton University.

"The Work of the Ballistic Institute of Clark University," by
A. G. Webster, Ph.D., Sc.D., LL.D., Professor of Physics,
Clark University, Worcester, ]\Iass.

" Alternating-Current Planevector Potentiometer Measure-
ments at Telephonic Frequencies," by A. E. Kennelly, A.M.,
Sc.D., Director Research Division, Electrical Engineering
Department, Massachusetts Institute of Technology, Cam-
bridge, and Edy Velander. Discussed by Prof. Webster.

" The Genesis of Petroleum as Shown by its Nitrogen Con-
stituents," by Charles F. Mabery, Sc.D.. Emeritus Professor
of Chemistry, Case School of Applied Science, Cleveland.
(By title.)

MINUTES. xi

" Graphical Representations of Functions of the nth. Degree,"
by Francis E. Nipher, Professor Emeritus of Physics, Wash-
ington University, St. Louis.

'' Ghmpses of the Near East During the War," by A. V. W.
Jackson, L.H.D., LL.D., Professor of Indo-Iranian Lan-
guages, Columbia University. (By title.)

" The Empire of Amurru," by A. T. Clay, Ph.D., LL.D., Pro-
fessor of Assyriology and Babylonian Literature, Yale Uni-
versity. (By title.)

"The Science of Stealing (Steyacastra) in Ancient India," by
Maurice Bloomfield, Ph.D., LL.D., Professor of Sanskrit,
Johns Hopkins University. (By title.)

" The Crib of Christ," by Paul Haupt, Ph.D., LL.D., Professor
of Semitic Languages, Johns Hopkins University.

" The Atonement Idea Among the Ancient Semites," by Ed-
ward Chiera, Ph.D., Instructor in Assyriology, University
of Pennsylvania. (Introduced by Prof. Morris Jastrow, Jr.)

Afternoon Session^ 2 o'clock.
Arthur A. Noyes, Sc.D., LL.D., Vice-President, in the Chair.

Symposium on Chemical Warfare.

" Historical Introduction," by Col. Marston T. Bogert, Ph.D.,
LL.D., Chemical Warfare Service, U. S. A.

" Chemical Warfare and Research," by Col. George A. Burrell,
Chemical Warfare Service, U. S. A. (Introduced by Col.
Bogert.) (By title.)

" Chemical Warfare and Manufacturing Development," by Col.
Frank M. Dorsey, Chemical Warfare Service, U. S. A. (In-
troduced by Mr. A. A. Blair.)

"Production of Chemical Warfare Munitions," by Col. Wil-
liam H. Walker, Chemical Warfare Service, U. S. A. (In-
troduced by Prof. H. F. Keller.)

" Means of Defense Against Chemical Warfare," by Col. Brad-
ley Dewey, Chemical Warfare Service, U. S. A. (Intro-
duced by Dr. PhiHp B. Hawk.) Discussed by Prof. Web-
ster, Dr. Keen and Mr. Rosengarten.

xii MINUTES.

Stated Meeting, May 2, ipip.
William B. Scott, D.Sc, LL.D., President, in the Chair.

The President appointed Prof. J. C. Arthur to prepare an obit-
uary notice of the late George Francis Atkinson.

The decease was announced of
Lucien Adam, at Rennes, France, on November 29, 1918, ast. 86.

Mr. Frederick Ives read a paper on " The SimpHfication of
Photography," which was discussed by Prof. Miller and Mr. Davis.

Stated Meeting, October S, 1919.

William B. Scott, D.Sc, LL.D., President, in the Chair.
William Curtis Farabee, a newly elected member, subscribed the
Laws and was admitted into the Society.

Letters accepting membership were received from

Robert Grant Aitken, Sc.D.,

Joseph Charles Arthur, Sc.D.,

Edward W. Berry,

James Henry Breasted, A.M., Ph.D.,

Ulric Dahlgren, M.S.,

William Curtis Farabee,

John Huston Finley, LL.D.,

Stephen Alfred Forbes, Ph.D., LL.D.,

Chevalier Jackson, M.D.,

Dayton C. Miller, A.M., Sc.D.,

George D. Rosengarten, Ph.D.,

Albert Sauveur, S.B.,

Wilham Albert Setchell, A.M., Ph.D.,

Julius O. Stieglitz, Ph.D., Sc.D.,

Ambrose Swasey, Sc.D., D.E.
A letter was read from Mr. Thomas A. Edison, resigning mem-
bership.

The decease was announced of the following members: .

Francis B. Gummere, A.B., Ph.D., LL.D., Litt.D., at Haver-
ford on May 30, 19 19, set. 64.

William Gilson Farlow, A.M., M.D., at Cambridge, on June 3,
1919, set. 74.

MINUTES. xiii

Sir Boverton Redwood, F.R.S., at London, on June 4, 1919,

set. 73.
Ira Franklin Mansfield, at Beaver, Pa., on June 7, 1919, set. y'j.
Rt. Hon. John William Strutt, Lord Rayleigh, O.M., LL.D.,

D.C.L, on July i, 1919, set. 'j'j.
Gustavus Retzius, M.D., at Stockholm, on July 21, 1919, set. 'j'].
Charles Conrad Abbott, M.D., at Bristol, Pa., on July 27, 1919,

set. "]().
Ernest Haeckel, Ph.D., at Jena, on August 8, 1919, set. 85.
Andrew Carnegie, LL.D., at'Xenox, Mass., on August 11, 1919,

ast. 84.
The following papers were read :
" Trogloglanis Pattersoni, a New Blind Fish from San An-
tonio, Texas," by Carl Eigenmann, Ph.D.
" On the Luciopimelodinse, a New Subfamily of the South

American Siluridse." by Charles S. Driver (communicated by

Dr. Eigenmann).

Stated Meeting, November 7, 1919.

William B. Scott, D.Sc, LL.D., President, in the Chair.
Mr. George D. Rosengarten, a newly elected member, subscribed
the Laws and was admitted into the Society.
The following papers were read :
" An International Treaty which Resulted in a Scientific League

of Nations," by Hon. William C. Redfield.
" The Metric System of Weights and Measures," by Dr. Samuel

W. Stratton.
" Child Maintenance at Public Expense," by Elisha Kent Kane.
" Tectonics of Laccolithic Mountains," by Charles Keyes (com-
municated by Prof. W. B. Scott).

Stated Meeting, December 5, 1919.

William B. Scott, D.Sc, LL.D., President, in the Chair.
The decease was announced of
Charles Henry Hitchcock, A.M., Ph.D., LL.D., at Honolulu,
on November 6, 1919, set. 83.

xiv MINUTES.

Julius F. Sachse, Litt.D., at Philadelphia, on November 4, 1919,
aet. 76.
The following papers were read :

" The Deep Kansan Pondings in Pennsylvania and the Deposits
Therein," by Prof. Edward H. Wilhams, Jr., which was dis-
cussed by President Scott.

" The Parallaxes of Fifty Stars, determined at Sproul Ob-
servatory," by John A. Miller (by title).

" The Experiences of a Sanitary Officer in the Army," by Dr.
Alexander C. Abbott, which was discussed by Mr. Field, Dr.
Jastrow, and President Scott.

INDEX.

Abbott, experiences of a sanitary
officer in the army, xiv

Aculeata, parasitic, i

Allen, basis of sex inheritance in
Sphccrocarpos, 289

Allen & Webster, on a new (?)
method in exterior ballistics, 382

Alternating-current planevector po-
tentiometer measurements at tele-
phonic frequencies, 97

America, evolution and mystery in
the discovery of, 55

Anderson, expedition of the Mount
Wilson Observatory to the Solar
Eclipse of June, 1918, 255

Antillean fauna, recent discoveries
of fossil vertebrates in the West
Indies and their bearing on the
origin of, 161

Ashford, application of sanitary sci-
ence to the Great War in the zone
of the army, 317

B

Balch, E. S., evolution and mystery
in the discovery of America, 55

Ballistics of a gun of seventy-five-
mile range, 273

on a new method in exterior, 382

Barnard, self-luminous night haze,
223

Benedict & Johnson, energy loss of
young women during the muscular
activity of light household work, 89

Bloomfield, science of stealing in an-
cient India, xi

Bogert, historical introduction to
symposium on chemical warfare, xi

Brown, errors induced in bullets by
defects in their manufacture, x

Burrell, chemical warfare and re-
search, xi

Camarasaurus Cope, characters and

restoration of the sauropod genus,

386
Campbell, Crocker eclipse expedition

from the Lick Observatory, June,

1918, 241
Chemical warfare, symposium on, xi

Cheyney, organized labor and peace,
iv

Chiera, atonement idea among the
ancient Semites, xi

Clarke, conservation of natural mon-
uments, V

Clay, empire of Amurru, xi

Commodities, relative contribution of
the staple, to the national food
consumption, 182

Coronal arches and streamers, 272

Crocker eclipse expedition from the
Lick Observatory, June 8, 1918, 241

Dewey, means of defense against
chemical warfare, xi

Dominion, peoples of central and
eastern Europe, iii

Dorsey, chemical warfare and manu-
facturing development, xi

Driver, the Luciopimeliodinae, xiii,
448

Duane, league of nations, iv

Duggar, hydrogen-ion concentration
of nutrient solutions in relation to
the growth of seed plants, vii

E

Eclipse expedition from the Lick Ob-
servatory, 1918, Crocker, 241

Eigenmann, Trogloglanis Pattcrsoni,
a new blind fish from San Antonio,
Texas, 397

Election of officers and councillors, iii

Energy loss of young women during
the muscular activity of light
household work, 89

Evolution, parasitic aculeata, a study
of, I

Food consumption, relative contribu-
tion of the staple commodities to
the national, 182

Fossil vertebrates in the West Indies
and their bearing on the origin of
the Antillean fauna, 161

Frost, results of observations of the
eclipse of the expedition from
Yerkes Observatory, 282

Functions of the nth degree, graph-
ical representation of, 236

XV

XVI

INDEX.

Growth, hydration and, 346
Gun of seventy-five-mile range, bal-
listics of a, 2>7?>

H

Haupt, crib of Christ, xi

Hayes, detection of submarines, x

Hazen, political problems of central

and southern Europe, iv
Howard, pink boUworm and the Eu-
ropean corn borer, vii
Hydration and growth, 346

Ives, artificial formations resembling
lunar craters, x

simplification of photography,

xii

J

Jackson, A. V. W., glimpses of the
near east during the war, xi

Jackson, C., bloodless removal of for-
eign bodies from the lungs through
the mouth by bronchoscopy, yi

Jastrow, the eastern question, iv

Johnson, E. R., what shall be done
with the railroads?, v

Johnson and Benedict, energy loss of
young women during the muscular
activity of light household work, 89

K •

Kane, child maintenance at pubHc ex-
pense, xiii

Kennelly and Velander, alternating-
current planevector potentiometer
measurements at telephonic fre-
quencies, 97

Keyes, tectonics of Laccolithic moun-
tains, xiii

Kunz and Stebbins, photo-electric
photometry of the 1918 eclipse, 269

Lampland, Lowell Observatory eclipse
observations, June, 1918, Promi-
nences and coronal arches, 259

Lanman, Benjamin Franklin's art as
applied to books for elementary
teaching, vi

Laws, amendments to, ix

Light in the study of ores and metals,
polarized, 401

Lowell Observatory eclipse observa-
tions, June, 191 8, 259

Luciopimelodinae, a new subfamily
of South American Siluridae, 448

M

Mabery, genesis of petroleum as

shown by its nitrogen constituents, x

MacDougal, hydration and growth,

346
hydration of colloids and cell-
masses in propionic acid and its
amino-compounds, vii
McCollum, relation of the diet to

pellagra, 41
Matthew, recent discoveries of fossil
vertebrates in the West Indies and
their bearing on the origin of the
Antillean fauna, 161
Mayor, detecting ocean currents by
observing their hydrogen-ion con-
centration, 150
Meteorological w'ork of the U. S.
Army, some scientific aspects of
the, 133
Members admitted:
Breasted, J. H., x
Farabee, W. C, xii
Mayor, A. G., v
Rosengarten, G. D., xii
Members deceased :
Abbott, C. G., xiii
Adam, Lucien, xii
Becker, G. F., v
Carnegie, Andrew, xiii
Crookes, Sir Wm., v
Doolittle, C. L., iv
Edmunds, G. F., iv
Farlow, W. G., xii
Foster, George C, iv
Gummere, F. B., xii
Haeckel, Ernest, xiii
Himes, C. F., iii
Hitchcock, C. H., xiii
Mansfield, I. F., xiii
Pickering, E. C, iv
Rayleigh, Lord, xiii
Raymond, R. W., iv
Redwood, Sir B., xiii
Retzius, Gustavus, xiii
Roosevelt, Theodore, iv
Sachse, J. F., xiii
Stephens, H. M., v
Ware, L. S., iii
Members elected, ix, x
Membership, acceptance of:

Robert Grant Aitken, Sc.D., xii
Joseph Charles Arthur, Sc.D., xii
Edward W. Berry, xii
James Henry Breasted, A.M.,

Ph.D., xii
Ulric Dahlgren, M.S., xii
William Curtis Farabee, A.M.,
Ph.D., xii

INDEX.

xvii

John Huston Finley, LL.D., xii
Stephen Alfred Forbes, Ph.D.,

LL.D., xii
Chevalier Jackson, M.D., xii
Dayton C. Miller, A.M., D.Sc,

xii
George D. Rosengarten, Ph.D.,

xii
Albert Sauveur, S.B., xii
William Albert Setchell, A.M.,

Ph.D., xii
Julius O. Stieglitz, A.M., Ph.D.,

xii
Ambrose Swasey, Sc.D., D.E., xii
Membership resigned :

Edison, Thomas A., xii
Miller, parallaxes of fifty stars de-
termined at Sproul Observatory,
xiv

Sproul Observatory eclipse ex-
pedition, June, 1918, coronal arches
and streamers, 272

Millikan, some scientific aspects of

the meteorological work of the

United States Army, 133
Minutes, iii

Mitchell, flash spectrum, 265
Mook and Osborn, characters and

restoration of the sauropod genus

Camarasaiiriis Cope, 386
Mt. Wilson Observatory expedition

to the Solar eclipse of 1918, 255

N

«th degree, graphical representation

of functions of the; 236
New method in exterior ballistics, 382
Night haze, self-luminous, 182
Nipher, graphical representation of

functions of the »th degree, 236

0

Ocean currents, detection of by ob-
serving their hydrogen-ion concen-
tration, 150

Ores and metals, polarized light in
the study of, 401

Osborn, energy conception of the
cause of evolution, vi

ores and metals, polarized light

in the study of, 401

Osborn and Mook, characters and
restoration of the sauropod genus
Camarasanrus Cope, 386

Pearl, relative contribution of the
staple commodities to the national
food consumption, 182

Pellagra, relation of diet to, 41

Photo-electric photometry of the 1918
Solar eclipse, 269

Planevector potentiometer measure-
ments at telephonic frequencies, al-
ternating current, 97

Polarized light in the study of ores
and metals, 401

Prince, Tatar material in Old Rus-
sian, 74

Prominences and coronal arches, 259

R

Redfield, an international treaty
which resulted in a league of na-
tions, xiii

Russian, Tatar material in Old, 74

Sanitary science, application of, to
the Great War in the zone of the
army, 317

Schoff, present tendencies in the na-
tional development of waterways, v

Scientific aspects of the meteorolog-
ical work of the U. S. Army, 133

Self-luminous night haze, 223

Sex inheritance in Spharocarpos, ba-
sis of, 289

Shapley, star cluster and their con-
tribution to the knowledge of the
universe, Z27

Siluridae, Luciopimelodinas, a new
subfamily of the South American,
448

Shull, sterility and self-and-cross-
incompatibility in shepherd's purse,
vii

Snyder, cosmic force, radioaction, v

Solar eclipse of June, 1918, expedi-
tion from Lick Observatory, 241

• — — ■ Lowell Observatory,

259

• McCorm,ick Observa-
tory, 265

' ■ Sproul Observatory,

272

■ University of Illinois

Observatory, 269

Expedition from Mt. Wilson

Observatory, 255

• Yerkes Observatory,

282

Spectrum, flash, 265

Sphccrocarpos, basis of sex inheri-
tance in, 289

Star clusters and their contribution
to the knowledge of the universe,
2Z7

XVlll

INDEX.

Stebbins and Kunz, photo-electric
photometry of the 1918 eclipse, 269

St'efansson, some oceanographical re-
sults of the Canadian Arctic expe-
dition, vi

Stratton, metric system of weights
and measures, xiii

Tatar material in Old Russian, 74
Trogloglanis Pattersoni, a new blind
fish from San Antonio, Texas, 397
Trowbridge, sound and flash rang-
ing, X

V

Velander and Kennelly, alternating-
current planevector potentiometer
measurements at telephonic fre-
quencies, 97

W

Walker, production of chemical war-
fare munitions, xi

Warner, Siberia and the Czechs, iv

Webster and Allen, on a new ( ?)
method in exterior ballistics, 382

Webster, recent applications of phys-
ics in warfare, ix

some considerations on the bal-
listics of a gun of seventy-five-mile
range, 373

works of Ballistic Institute of

Clark Univ., x

Wheeler, parasitic aculeata, a study

of evolution, i
Williams, deep Kansan pondings in

Pennsylvania and the deposits

therein, xiv
Wright, polarized light in the study

of ores and metals, 401

Proceedings Am. Philosoph. Soc. LVIII

Plate I

Three studies of head of Camarasaurus. A, head extended and bent up-
wards. B, neck flexed and head bent downwards. C , neck flexed and head bent
strongly downwards to show ground-feeding pose and to expose the teeth.
The head of Camarasaurus is extremely short and deep. The nostrils and
eyes are elevated. The tympanum leaves a space behind the quadrates.

Proceedings Am. Philos. Soc. Vol. LVIIt.

Plate II

A, ventral view of the anterior vertebrae of Perugia xanthus ; B, ventral
view of anterior vertebrae of Rhamdia cinerascens ; C and D, dorsal views of
skull of Perugia xanthus; E, jateral view of Perugia xanthus.

I, Basioccipital ; 2, inferior process of post-temperal ; 3, ventral view of air
bladder capsule; 4, transverse process of fourth vertebra; 5, transverse proc-
ess of fifth vertebra; 6, transverse process of sixth vertebra; 7, transverse
process of seventh vertebra; 8, concentric process of tripus; 9, anterior proc-
ess of fourth vertebra; 10, posterior process of fourth vertebra; 11, nasal;
12, ethmoid; 13, lateral; 14, frontal; 15, sphenotic; 16, pterotic; 17, epiotic;
18, supraclavical ; 19, dorsal view of air-bladder capsule; 20, supraoccipital ;
21, neural spine of fourth vertebra; 22, sixth vertebra; 23, hyomandibular ;
24, metapterygoid ; 25, quadrate ; 26, preopercle ; 27, interopercle ; 28, opercle ;
29, clavicle; 30, end of air-bladder; 31, lateral view of air-bladder capsule;
32, fifth vertebra.

Proceedings Am. Philos. Soc. Vol. LVIII.

Plate III

1-3. Perugia agassisii (Steindachner). After Steindachner.

4. Luciopinielodus pati (Valenciennes). After Valenciennes.

5. Perugia xanthus (Eigenmann). Photograph by Eigenmann.

6. Megalonema platycephalum Eigenmann. After Eigenmann.

3 2044 083 311 686

^^"

,- t

J-:^.

■^..

^■<^\ ■>