Vol. 36

PROCEEDINGS

OF THE

AMERICAN ACADEMY

OF

ARTS AND SCIENCES.

Vol. XXXVI.

FROM MAY, 1900, TO MAY, 1901.

BOSTON:
PUBLISHED BY THE ACADEMY.

1901.

Jlnibrrsitg JJrrss:
.Ii'iiN Wii.s«i.\ am) Son, CamhMDGK, U.S.A.

CONTENTS.

Pagb
I. A New Fossil Crab from the Miocene Greensand Bed of Gay

Head, Martha's Vineyard, with Remarks on the Phylogeny of

the Genus Cancer. By Alpheus S. Packard 1

II. On the Thermal Diffusivities of Different Kinds of Marble. By

B. O. Peirce and R. W. Willson 11

III. Geometry on Ruled Quartic Surfaces. By Frank B. Williams 17

IV. On Supposed Merostomatous and Other Paleozoic Arthropod Trails,

with Notes on those of Limulus. By Alpheus S. Packard . 61

V. Certain Derivatives of Metadibromdinitrobenzol. By C. Loring

Jackson and W. P. Cohoe 73

VI. On the Continuity of Groups generated by Infinitesimal Transfor-
mations. By Stephen Elmer Slocum 83

VII. On Hardystonite and a Zinc Schefferite from Franklin Furnace,
N. J. By John E. Wolff. With a Note on the Optical
Constants of the Schefferite. By Dr. G. Melczer . . . Ill

VIII. On the Thermal and Electrical Conductivity of Soft Iron. By

Edwin H. Hall 119

IX. A Neio Conception of Thermal Pressure and a Theory of Solutions.

By Gilbert Newton Lewis 143

X. International Atomic Weights. By Theodore William

Richards 169

XI. Peripheral Distribution of the Cranial Nerves of Spelerpes bilin-

eatus. By Mary A. Bowers 177

XII. On Certain Derivatives of Orthobenzoquinone. By C. Loring

Jackson and Waldemar Koch 195

XII T. On the Action of Sodic Sulphite on Tribromdinitrobenzol and Tri-
bromtrinitrobenzol. By C. Loring Jackson and Rtciiard
B. Earle 229

S

N CONTENTS.

Paoe
XIV. False Spectra from the 11 owl a mi Concave Grating. By

Thkodori Lvman 239

XV. Investigation* on th> Composition of Petroleum. By Charles
1". M.\i;i i:V

1. On the Composition oj California Petroleum. By Charles

1 •'. Mabbry and Edward J. Hudson

2. On the Chlorine Derivatives of the Hydrocarbons in Cali-

fornia Petroleum. By Charles F. Mabbry and
Otto J. Sieplein
8. On the Composition of Japanese Petroleum. By Charles

F. Mai-.i ky AND Shinn hi 1 'aka.no 253

XVI. The "Eclipse Cyclone and th< Diurnal Cyclones. Results of

Meteorological Observations in the Solar Eclipse of May £8,
1900. I!y II. Hi i.m Clayton 305

XVII. An Apparatus for Recording Alternating Current Waves. By

Frank A. Laws 319

XVIII. Suggestion concerning the Nomenclature of Heat Capacity.

By Theodore William Richards 325

XIX. Symmetrical Triiodbenzol. By C. Loring Jackson and G.

E. Bbhb 331

XX. A Study of Growing Crystals by Instantaneous Photomi-
crography. By Throdore William Richards and
Ebenkzbr Henry Archibald 339

XXI. Design as a Science. By Denman W. Boss 355

XXII. The Occlusion of Magnetic Oxalate by Calcic Oxalate, and the
Solubility of Calcic Oxalate. By Theodore W. Rich-
ards, Charles F. McCaffrey, ind Harold Bisbri 375

XXIII. Preliminary Diagnoses of New Species of Laboulbeniaceae —

///. By Roland Thaxter 395

.IV. A Study of Variation in tht Fiddler Crab, Gelasimus pugilator

Lair. l'-Y Roberi M. Xerkbs 415

XXV. I'fi' Development and Function of Reismer't Filire, and its

■/■ Connections. By Porter Edward Sargeni , 448

XXVI. 1. Synop ( u Melampodium.

2. Synoj oj Gt V ca.

Voted Synonymy among the Sper-
mat op) Mexico < d ( ■ itral America. By B. L.

153

CONTENTS. V

Page
XXVII. Some New Spermatophytes from Mexico and Central America.

By M. L. Fernald 489

XXVIII. The Solubility of Manganous Sulphate. By Theodore

William Riciiards and Frank Roy Fraprie . . 507

Records of Meetings 517

A Table of Atomic Weights. By Theodore William Richards 545

Report of the Council 547

Biographical Notices 547

Charles Carroll Everett 549

Nathaniel Holmes 552

Silas Whitcomb Holman 553

Sylvester R. Koehler 556

John Elbridge Hudson 558

John Harrison Blake 565

Charles Franklin Dunbar 569

Officers and Committees for 1900-1901 576

List of the Fellows and Foreign Honorary Members . . . 577

Statutes and Standing Votes 585

Rumford Premium 595

Index 597

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 1. — July, 1900.

A NEW FOSSIL CRAB FROM THE MIOCENE GREEN-
SAND BED OF GAT HEAD, MARTHA'S VINEYARD,
WITH REMARKS ON THE PHTLOGENT OF THE
GENUS CANCER.

By Alpiieus S. Packard.

WITH TWO PLATES.

PALEONTOLOGICAL NOTES, NO. V.

A NEW FOSSIL CRAB FROM THE MIOCENE GREENSAND
BED OF GAY HEAD, MARTHA'S VINEYARD, WITH
REMARKS ON THE PHYLOGENY OF THE GENUS
CANCER.

By Alpheus S. Packard.

Received May 3, 1900. Presented May 9, 1900.

In looking over a small collection of fossil crabs made by Mr. J. H.
Clarke of Providence, R. I., about thirty years ago, there occurred one
specimen which represented quite a different group from Stimpson's
Archaeoplax signifera. It came, he told me, from the same bed and
locality at Gay Head as the other crabs.

On consulting with Mr. Walter Faxon, assistant in charge of the
Crustacea of the- Museum of Comparative Zoology, who kindly showed
me the fossil crabs of the collection, we concluded that though the
body was very round and the surface of the carapace much more
convex, it could not belong to any other genus. Mr. J. B. Wood-
worth also kindly allowed me to examine a trayful of fragments of
Archaeoplax collected by him while connected with the U. S. Geologi-
cal Survey, from the greensand bed at Gay Head. Among them was
the hand of what was plainly enough that of a large Cancer, like our
common C. irroratus, and it seemed evident that it must have belonged
to a large individual of the same species as the small crab. Both still
retained more or less of the greensand, and the hand was loosely em-
bedded in a matrix of that material. For the opportunity of describing
the hand, I am indebted to the U. S. Geological Survey.

Cancer proavitus, n. sp., Plate I., Figs. 1,2, 3. One young male repre-
sented by the carapace, sternum, and basal joints of the legs. All the
teeth on the front edge of the carapace absent except the two on the left
side of the left orbit.

Body much narrower, and therefore rounder, more orbicular than in
the existing species.

4 PROCEEDINGS OF Tin: AMERICAN ACADEMY.

Carapace entire, with the exception of the greater part of the front
margin. It is much higher and much more convex than in the two
iting Bpecies of our coast (('. irroratus and G. borealis), being
almost rounded, the convexity being between one-fourth and one-third
the entire thickness of the body. The surface is much more irn g-
ular than in the two existing Bpecies, being thrown up into fourteen
well-marked, more or less flattened, hosses or tuberosities. Of these,
four are .situated along the median line; the one most posterior is the
broadest and highest, the fourth or most anterior one the smallest and
lowest ; also the granulations or crateriform papilla) covering the sur-
face of all the tuberosities are larger and more prominent than those
elsewhere on the surface of the carapace. On each side of the median
row are five other flattened low tuberosities, the largest of which is
opposite the second median one (counting from behind), and half-way
between this and the outer side of the same tuberosities is a small one ;
behind the one first named is a small one situated opposite to the first (or
most posterior ) median tuberosity. In frout there are two small tuberosi-
ties, Bituated opposite the anterior median one, but placed a little nearer
the front edge, and directly behind the orbit or base of the eyestalk; of
these two bosses the outer one is the larger.

Fortunately on the left side of the left orbit two of the marginal teeth
of the carapace arc perfectly preserved ; thej correspond to the second
and third teeth, from the left orbit, of C. irroratus. The third tooth
(corresponding to the third tooth in ('. irroratus) is straight on the
free edge, with about lo papilla; or granulations along the edge, while the
other (the second) is more conical or tooth-like, with from 16 to 18
granulations; indeed the granulations on tho edge and near it are
almost exactlj the same in number, Bize, and position as in C. irroratus.

The hinder edge of the carapace is much as in C. irroratus ; the row

granulations on each Bide of the middle of the carapace, and the shorter

oblique row extending to the insertions of the hinder pair of le_rs. are

almost exact!) as in the existing Bpecies. In the fossil Bpecies there

are about four of these granulations to the millimetre, there being

the Bame number in 0. irroratus, at a point near the median line
of the carapace; while in the middle of tin- Beries there are five, but
in 0. irroratus only three, being -lightly larger and farther apart.

The Bternum is as in the male 0. borealis, as are the basal joints of
the legs, and F can detect do good specific differences. In the Bternum
of the fossil the three segments, including the apical one,

which ends at the insertion of the maxillipeds. are well preserved.

PACKARD. — A NEW FOSSIL CRAB. 5

On the posterior edge of the penultimate sternal segment are two
minute tubercles, but with the tips broken off, which are exactly as
in G. irroratus ; these little tubercles we take to be the "genital tuber-
cles " of Stimpson. The abdomen itself is wanting. The sternum is on
the whole rather more like that of G. borealis than G. irroratus.

The large hand evidently belonging to this species is marked in
exactly the same manner on the outer aspect as in G. borealis, having the
four distinct raised granulated lines or ridges, with the same arrangement
as in the living species. The single specimen lacks the articular face at
the base, as also unfortunately the fingers, and also the spine at the base.

Of the four ridges on the inner face, on the lowest ridge there
are 32 papillae, or conical granulations; on the ridge above, where
they are much smaller, about 33; in the third ridge there are two
rows of conical granulations, one consisting of 20 large ones, with
a parallel row of minute ones below; on the fourth ridge there are
about 12-14 large granulations, with smaller ones situated irregularly on
the lower side, while some others are interpolated between the large
ones. The concavities between the ridges are well marked.

Length of carapace, 30 mm. ; width, 38 mm.

Thickness of the body in the middle, 17-18 mm. ; amount of con-
vexity of the dorsal surface of the carapace, about 5 mm.

Length of sternum, 20 mm. ; greatest breadth, 12 mm.

Length of hand without the fingers, of the other specimen, 30 mm.;
breadth, 21 mm.

On comparison with seven small G. irroratus (2 $ and 5 9 ) of nearly
the same age, it is seen to differ markedly in three points, i. e., the
very narrow, much rounded, or orbicular body, the much more convex
carapace, and the much greater number and prominence of the flattened
tuberosities.

In the small G. irroratus 34 mm. long the carapace is 50 mm. in
width, or 16 mm. wider; in G. proavitus, which is 30 mm. long, it is
only 8 mm. wider than long.

In G. irroratus we see only traces of the tuberosities on the sides of
the carapace, those of the median line being obsolete. On each side of
the median line or region in G. irroratus are tw.o obscurely marked
tuberosities, but they are very low, broad, and flat. The most striking
difference, however, is the much more rounded shape of the body.
On the other hand, the teeth and granulations on the teeth, and on the
narrow ridges of the posterior edge of the carapace, are nearly identical.

Comparing the hand of the fossil species from the U. S. Geological

b PROCEEDINGS OF TEE AMERICAN ACADEMY.

Survey with that of a ( Irroratus about four inches wide, notable

are to be Been Though the four ridges are the Bame in length,
width apart, and in general arrangement, the teeth on these ridges are
in C. i much larger, and are represented in 0. irroratus

by more numerous and crowded granulations, which are flattened,

ind« d. and polished.

In Borne important respects the tertiary species resembles O.bon
This differs from C irroratus in being finely moricate, the minute
iua tobercles being much larger, higher, and sharper, almost form-
sharp spines; they are also more numerous and crowded, and often
bear a hair. In these respects C. prottvittu approaches 0. borealis. As
in O. irrt the anterolateral margins are nine-toothed; of these

b the ninth, or that m-xt to the orbit, is sharper than in the other
living and ends in a sharp Bpine, with several accessory spin ales,

The postero-lateral margins are more sinuous than in C. irroratus, and
the granulations on the ridge are larger, fewer, and end in a point. In
both of tlu; Living species the convexity of the carapace is about the
Bame. The surface in 0. borealis is perhaps a little more uneven.

The hand of C borealis differs from that of 0. irroratus in being
inueh more imiricatc or spiny, the granulations on the four external
ridges of the latter being represented by well-marked Bharp Bpines, these
being especially large and high on the uppermost rid_

( . borealis is a decidedly hairy species, whereas C. irroratus is
naked, but a few hairs being visible; on the other hand, in C. borealis
nearly every tubercle bears a pale hair.

The abdomen of C. borealis differs from that of the more common
(<'. irroratus) in being less acute and mucronate at the tip.

The sternum i- a little more hairy.

jth of a Bmall C. borealis, 80 mm. ; breadth. 89 mm.; thickness
of l.oip . 10—11 mm.

The phylogeny of the Eastern American species of the genus
Cancer. — A comparison of the miocene tertiary species of Cancer with

the tWO Bpeciel now living in the waters of Vineyard Sound, brings OUl

the interesting fad that the extinct Bpecies appears to be the stem or

■ d form from which the recent Bpecies mentioned have descended.

' ineer proavitus presents characters in which it resembles C.bore-

i w.ll as c. irroratus. It resembles C. borealis in the higher,

e pointed granulations on the postero-lateral margin of the carapace,

in the <|uite high and Bharp Bpines on the ridges ol the hand, as well

in - and hail-, ; 00 the other hand it i.-, .dmi-

PACKARD. A NEW FOSSIL CRAB. 7

lar to G. irroratus in the shape of the nine teeth on the antero-lateral
margin of the carapace, and in the straight postero-lateral margin of the
same. It is rounder, narrower, the carapace more convex, and the body
in general more hairy than either of the existing species.

It thus seems most probable that the miocene species, being a more
generalized, composite form, is the ancestor from which either towards
the end of the pliocene or the beginning of the quaternary period the two
living species sprang. G. irroratus has inherited the exact shape of the
lateral teeth, and the shape of the postero-lateral margin of G.proavitus,
while G. borealis has retained the higher spine-like granulations or sub-
muricate feature of the carapace and hand, and the hairiness of the body.

On the whole the evidence that our two northeastern species have de-
scended from a much more rounded, convex, and hairy miocene form liv-
ing in the same geographical area seems well established.

It would be most interesting to compare this fossil species with very
young individuals of our living species, but after inquiry I find they
are not in existence in our museums. It is to be hoped that specimens
of the very young may be collected and compared with the fossil species.
It is known that in Cancer the body grows wider with age.

Note on Archaeoplax signifera Stimpson.* — While the collections
of the fossil Crustacea made at Gay Head comprise only one speci-
men of Cancer proavitus, with the hand of a much larger individual, the
fragmentary remains of the Archreoplax are much more abundant,
showing that it was the most prevalent form.

The specimens, however, in the Museum of Brown University, and
those collected by Mr. Clarke, and those in the Cambridge Museum, are,
so far as we have observed them, not sufficiently well preserved to enable
one to make a restoration which would be a very decided improvement
on the excellent diagrammatic drawings by Dr. Stimpson. None of the
specimens of the carapace — and in two large specimens they are tolerably
well preserved — show the four teeth of the antero-lateral margin; on
the other hand, in Mr. Clarke's specimens the legs of the four posterior
pairs are well enough preserved to show five of the joints, the terminal
ones wanting; the fourth joints are of unusual length.

In one young specimen the carapace has been broken away on the
back so as to still show the gills in place.

Regarding the temperature of the water of the miocene period at Gay

* On the fossil C»ab of Gay Head, Boston Journal of Natural History, VII. No.
4, April, 1863.

8 PROCEEDINGS OF THE AMERICAN ACADEMY.

Head Dr. Stimpson remarks: "The abundance of specimens of the
it i lav Head indicate a warm climate in that region at the
time the] were living. At the present day all Carcinoplacidse arc inhab-
itants "i" warn: -. The nearest allied genus. Heteroplax, lives on the
coast of China at tlie northern limit of the torrid zone."

The occurrence of a species of Cancer in the same bed does not nec-
trilj tend to prove that the water of the sea on our mioceue coast was
DOt somewhat wanner than at present, though not of subtropical temper-
ature. There are four species of Cancer on the Pacific coast of North
America, where ft magist&r ranges from Sitka to Monterey, O. gracilis
from Pllgef Sound to San Francisco Bay, ft prodtictus from Puget
Sound to San Francisco, while ft antennariits occurs at San Francisco,
.Monterey, and Toinales Bay. I am indebted to Mi— Mary J. Kathbun,
-tant in the Smithsonian Institution, for the information that there are
no tropical Bpecies of Cancer, and she has kindly sent me the following
list of Bpecies of this genus exclusive of those of North America and
Europe : —

' ovae-zelandiae Lucas. New Zealand.

ft bettianus Johnson. Madeira.

C.plebeius Poeppig. Chile.

C. polyodon " "

C. edwardsii Pell. "

( '. longipes Bell. "

ft japonicut Ortmann. Japan

< '. pygmaeus " "

('. gibbo8ulu8 (de Haan) " as well as west coast N. America.

O. amphioetut Rathbun " " " " " "

Of the two species now living on the shores of southern New England.

the most ( moti one (ft irrorafaui) ranges from southern Labrador to

South Carolina, while < '. boreallS is rarer, more local, and has thus far
only been found to extend from Nova Scotia to Vineyard Sound and \o
M . Land. Loth, then, appear to be on the whole boreal species.

The invertebrate fauna with which Gancer proavitus is associated

ha- been enumerated by Dr. \V. II. hall.* of twenty-two species of

mollusks, about eighl appear to he recent Bpecies ^till living in the waters

of that region; among them occur Buch boreal forms as Mya arenaria^

M. truncate, Toldia limatulay sapotella^ etc., and Dall Btates:

- on tin- Miocene and Pliocene of Gsj Bead, Martha's vineyard, etc.
Amer Jour. Bd Xl.vm . Octobt r. 1894 p 2

PACKARD, — A NEW FOSSIL CRAB. 9

" It will be observed that this is a distinctly northern assemblage ;
any of the species might be at home in the waters about Gay Head to-
day, as far as we can judge by analogy in the case of extinct species."

It would appear, then, from the evidence thus far obtained, and taking
into account the abundance of Archaeoplax, that the climate of the miocene
of Gay Head, or at least the temperature of the ocean, was probably
somewhat warmer, but yet not greatly different from what it now is south
of Cape Cod.

PLATE I ' incer proavitus.

Figure 1. From above, nat. size.

2. From below.

3. Hand of a larger specimen.

1 A chaeoplax signifera Stm., carapace from above, nat. size.

Packard — New Fossil Crab

Plate I

Fig. I

Fig. 2

Fig- 3

Fig 4

PLATE II. Archa

Fioohe 1. Carapace from above, nat Bize
2 Sterna el two individuals.

3. Four legs, Bhowing the very long fourth joints.

Packard — New Fossil Crab

Plate 2

Fig. i

Fig. 2

Fig. 3

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 2. — June, 1900.

ON THE THERMAL DIFFUSIVITIES OF DIFFERENT

KINDS OF MARBLE.

By B. O. Peirce and R. W. Willson.

Investigations on Light and Heat made and published wholly or in part with Appropriations

from the itumpord fund.

ON THE THERMAL DIFFUSIVITIES OF DIFFERENT

KINDS OF MARBLE.

By B. O. Peirce and R. W. Willson.

Received May 4, 1900. Presented May 9, 1900.

Last year we published in the Proceedings of the American Acad-
emy of Arts and Sciences an account of some determinations of the
thermal conductivities of different kinds of marble, made by the so-
called "Wall Method." The horizontal bases of a rectangular prism,
of height small compared with the area of a horizontal cross section,
were kept for a long time at constant temperatures, the final tem-
peratures at two or more points in the vertical axis were determined,
and the flux of heat through a definite central portion of the colder base
was measured. For the details of the apparatus we refer to that
paper.

In each experiment of one of our sets a rectangular prism 60 cm.
square and not more than 6 cm. high was built up of a slab
of the material to be tested, enclosed between two other slabs of
the same material. Between each two consecutive slabs of the prism
was placed a thin metal sheet. This consisted of two rectangular leaves
of tinfoil, about 60 cm. long and a little less than 30 cm. wide,
placed side by side, and separated by a narrow ribbon thermal element
made by butt-jointing, end to end, with the help of silver solder, a
strip of German silver, and a strip of copper of the same thickness as
the tinfoil. After the edges of the ribbon had been varnished so as to
prevent electrical contact, the ribbon and the tinfoil could be placed
close together so as to form a continuous sheet of metal 60 cm.
square and about 1-10 of a millimetre thick.

The ribbon thermal elements gave consistent results at all times,
provided that the junctions themselves were in contact with the slab9
between which they lay. If the tinfoil wings were considerably
thicker than the junction-ribbon, or if the junction itself were scraped
thin, the reading might be in error by an amount not easy to be
accounted for by the mere resistance of the thin air-film on each side of
the junction.

14

PROCEKl'INCS .if THE AMERICAN ACADEMY.

TABLE I.

Sp. Gr.

1. Carrara ....
Mexican Onyx

3. Vermont Statuary

4. American White .
Egyptian . . .

6. Sienna ....

7. Bardiglio . . .

8. Vermont Cloudy White

9. Vermont Dove Colored

10. Lisbon

1 1. American Black . . .

\'l Belgian

13. African Rose Ivory

11. Tennessee Fossiliferous

15. Knoxville Pink . . .

10. St. Baume

2.71
•J 71
2.72
2.74
2.68
2.69
2.75
2.74
2.75

2.75
2.75
2.71
2 73
2.70

Conductiv-
ity.

0.00505

0.0055G

0.00578

0.00596

0.00623

0.0067G

O.Of

0.0c

0.001

0.00685

0.00

0.00755

0.00756

0.00750

0.00757

0.00761

Heal

to 100 ).

Bp. H
per I nit

\. ..

0.21 1

0.579

0.211

0.672

0.2H)

0.6

0.214

582

0.212

0 58 1

0.215

0.570

0.218

0.6

(1.210

0.578

0.208

0.570

0.211

0.580

0.214

0.57 1

0.206

0.567

0.212

0.683

0.211

0.580

0.212

0.679

0.210

0.5(17

Diffnaiv.

itv.

0.0087

0.0094
0.0102
0.0102
0.0107
0.0117
0.0116
0.0118
0.0120
0.0118
0.0119
0.0133
0.0130
0.0130
0.0131
0.0134

After the absolute conductivity of a particular specimen has been care-
fulh determined, between various pairs of temperature limits, the conduc-
tivity of any other specimens can be easily obtained bj determining the
temperature-gradient, in the final state, on the axis of a prism built up of
the Blab already tested, and the slab to be examined, with their attendant
thermopiles and such other thin slabs a-- may be conveniently used. I;.
varying the order of the Blabs on different occasions, the temperatures at
the il the slab to be examined can be altered, it being always

undi that the thermal elements must be placed between slabs

approximately the same conductivities. The relative conductivities
lifferenl materials il these conductivities are not widely different, can
i with great accuracy by thi~ method ; and it is possible to

PEIRCE AND WILL80N. THERMAL DIFFUSIVITIES OF MARBLES. 15

TABLE II.

Higher
Temperature.

Lower
Temperature.

Average
Specific Heat.

Calculated.

1.

80.2

26.90

0.199

0.1956

2.

80.2

26.91

0.199

0.1956

3.

75.7

26.00

0.196

0.1942

4.

76.0

26.10

0.199

0.1947

5.

63.0

24.40

0.196

0.1919

6.

61.0

24.36

0.185

0.1915

7.

61.0

24.30

0.191

0.1915

8.

53.0

24.60

0.181

0.1900

9.

53.0

24.60

0.193

0.1900

10.

48.1

22.27

0.189

0.1887

11.

48.0

22.24

0.189

0.1886

12.

44 3

22.40

0.188

0.1880

13.

44.1

22.20

0.188

0.1879

14.

44.1

22.40

0.189

0.1880

15.

91.5

27.90

0.197

0.1979

16.

100.00

29.66

0.199

0.1999

17.

100.00

28.20

0.191

0.1996

18.

100.00

28.10

0.198

0.1996

19.

100.00

28.38

0200

0.1996

20.

79.1

25.54

0.196

0.1951

21.

79.1

25.80

0.193

0.1952

22.

79.2

26.40

0.198

0.1954

use much smaller slabs than are necessary for the determination of
absolute conductivities.

For many purposes to which the results of our experiments might be
applied, it is desirable that the specific heats of the various marbles should

16 PROCEEDINGS OF THE AMERICAN ACADEMY.

be known ; this information is given in the following table. For con-
venience the values of the conductivities, as printed on Page 56,
Volume xxxiv. of the Proceedings of the American Academy, are
appended. At hast two specimens of each sort of marble except No. 15
wnr examined.

'I'm determine the variation (jf the specific heat with the temperature,
the following determinations were made by the "Method of Mixtures,"
upon specimens of Carrara marble dried at temperatures somewhat above

100° c.

e observations are well represented by assumiug the following
expression lor the specific heat of dry Carrara marble: —

S = 0.1844 + 0.000379 t-

The fifth column of the above table gives the values' of the mean spe-
cific heat between the given temperatures computed by the formula
<? = 0.1848 (t — 25°) + .0001.S95 (t — 25°) 2 where Q is the quantity
of heat in calories required to raise 1 gr. from 25° to the temperature t :
this expression corresponds to the equation above given for S.

The 17th observation was marked for rejection, a piece of the sub-
stance having been dropped from the cup in which it was heated and
hastily picked up and dropped into the calorimeter. As it is difficult to
believe that this accident could cause so great a difference in the result
as appears, we have thought it best to retain the observation.

Jkfferson Physical Laboratory, Cambridge,
February, PJO0.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 3. — July, 1900.

GEOMETRY ON RULED QUART IC SURFACES.

By Frank B. Williams.

GEOMETRY ON RULED QUARTIC SURFACES.

By Frank B. Williams,
Fellow in Mathematics, Clark University, Mass.

Presented by William E. Story, May 9, 1900.

I. Introduction.

1. Ruled quartic surfaces have been studied systematically by Cayley,
Cremona, and Salmon, and to some extent by Rohn, Chasles, Schwartz,
Reye, Voss, Holgate, and others. Of the quartic scrolls, Cremona,* in
his excellent synthetic treatment, enumerates twelve species, while
Cayley,f in the most complete and masterly analytical treatment of
scrolls, divides these scrolls into ten species and gives a comparison of
his species with those of Cremona, stating that the latter's two remain-
ing species, though properly considered as distinct from the others, may
be regarded as sub-forms of his seventh and ninth species. As we shall
not have occasion here to take account of these sub-forms, we shall follow
Cayley's classification. The other ruled quartic surfaces are : the de-
velopable quartic, or torse, whose edge of regression is a twisted cubic,
and the quartic cones, which, although developable, must be considered
separately. The developable quartic has been quite thoroughly studied
by Cayley, Salmon, and others, but very little seems to have been done
with the quartic cones, which, as will appear later, present some very
interesting features.

The consideration of curves in space has been the subject of a great
many articles by the most eminent geometricians, but curves on ruled
quartic surfaces in particular have received little attention. As early as
1861, Chasles $ gave a method of describing curves of order 4 m + n, on

* Sulle superficie gobbe di quarto grado, Mem. della R. Istoria di Bologna,
Series II., T. VIII.

t Second and Third Memoirs on Skew Surfaces, Otherwise Scrolls. Coll.
Math. Papers, Vol. V. and Vol. VI., respectively, and Phil. Trans., 1863 and 1869,
respectively.

I Description des courbes a double courbure de tous les ordres sur les surfaces
reglees du troisieme et du quatrieme ordre. C. R. LIII, 884-889.

20 PROCEEDINGS OF THE AMERICAN ACADEMY.

ruled quartic surfaces, by means of a pencil of surfaces of order m and
groups lit';/ generators in involution. In 1897, Rohn * treated some ol
the properties of curves on the general quartic surface, considering also
the surfacea that can be passed through these curves and presenting some
theorems regarding the residual intersection.

In 1883, Professor Story discovered a method by means of which he
was able to classify all curves lying on a quadric surface and to give a
formula for the number of intersections of these curves, thus obtaining,
i synthetic process, the results already found by Cayley.t Professor
Story applied his method to the cubic scrolls, classifying all curves lying
on these surfaces and obtaining a formula for the number of intersec-
tions of any two of these curves, analogous to that found for curves on
quadrica.

By an extension of the analytical method of Cayley, Dr, Ferry X suc-
ceeded most admirably in treating analytically the " Cubic Scroll of the
First Kind,'' verifying the results of Professor Story for this surface, and
in a paper soon to be published, Dr. Ferry has also verified, in the same
way, Professor Story's results for the other cubic scroll.

It is the purpose of the present paper to consider the classification of
curves on all ruled quartic surfaces; to find the formula for the number
of intersections of any two curves that lie on the same ruled quartic sur-
face; and to point out some of the most notable results obtained in the
course of the investigation. The equations of many of the ruled quartic
surfaces are so complicated that very serious difficulties arise when we
attempt to treat them analytically, and it has been found most convenient
to employ the -synthetic method of Professor Story.

2. For convenience, we shall use the symbol SM to denote a surface
of order v, and 2<M> to denote a ruled surface of order fx. C<") will be
used to denote a curve of order '/ lying on the ruled surface in question.
By an arbitrary generator we shall mean any simple generator that bears
iki Bpecial relation to the curve in question, e. g. in considering a plane
curve, it is any simple generator not lying in the plane of the curve.
It must be proved, first of all, that every curve (" ' meets each gen-
erator of the Burface on which it lies in a constant number of points, say

* Die Raumcnrven auf 'leu Fl&cben IV'" Ordnung. Verhandlungen der K.
Sachs, Qesell. der Wiss. zu Leipz

t On the curves situate mi a surface of the Becond order. ('"11 Math. Papers,
Vol V . and Phil. Mag., 1861, pp

ometry on the Cubic Scroll of the First Kind Archiv for Mathematik og
Naturvidenskab, B. XXI Nr. 3, I

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 21

a points, equal to the number in which it meets an arbitrary generator.
(If the cui've goes through a point common to all the generators, e. g.
the vertex of a cone, this point is not counted as one of the a points on
any generator.) This furnishes us a method for classifying all curves on
2^, and we shall use the symbol aa to denote a curve of order a that
meets each generator of the surface on which it lies in a points, a being
a constant. Similarly, bp will denote a curve of order b that meets
each generator j3 times. Each generator is itself a plane curve of order
1, and since it is not met by any other generator except those lying in its
plane, it is represented by 10. In case 2^ has more than one system of
rulings, the lines of one system are chosen as the generators, with refer-
ence to which the classification is made for all curves on that surface,
and the lines of any other system are regarded simply as curves of
order 1.

We shall use the symbol (aa, bp) to represent the number of intersec-
tions of any two curves aa and bp on the ruled surface.

Professor Story * proved that for all curves lying on a quadric surface
(aa, bp) = a fi + b a — 2 a ft and for all curves lying on the cubic
scrolls (aa, bp) = a (3 + b a — 3 a ft and stated that it is probably true
that (aa, bp) = a (3 + b a — fx a ft for curves on a scroll of any order
p. It will be proved here that the formula is true for all ruled quartic
surfaces, i. e. that we have

(1) (a,bp) =afi + bu — 4 aft

and when we say that this formula holds for a certain curve we shall
mean that it gives the number of intersections of this curve with any
other curve on the scroll. It must be borne in mind that these formulae
for the cones of different orders give, in each case, the number of inter-
sections of the curves aside from those at the vertex, since the vertex is
not one of the a points on any edge. This will be proved for the
quartic cones.

3. We shall first consider three general theorems,! which must be
proved before the formula can be established. The first may be called
the fundamental theorem and may be stated thus :

Theorem I. — If a be the number of points of C(a) on an arbitrary gen-
erator, there are a points of C(rt) on each generator.

* On the Number of Intersections of Curves Traced on a Scroll of any Order.
Johns Hopkins University Circulars, August, 1883.

t These theorems were given by Dr. Story in his lectures, October to Decem-
ber, 1899.

PROCEEDINGS OF THE AMERICAN ACADEMY.

No general proof of this theorem has yet been found, and it must be
proved for each of the ruled surfaces, separately.

From Theorem I are readily deduced the other two theorems, as
follows :

Theorem II. — If aj is the complete intersection of 2<M) and SH"\ and
if bp is any curve on 2^) that has no component in common with aa, then
, bp) = u ft + b a — /x aft.

Proof. — The intersections of aa and bp are simply the intersections
of £('•> and bp and are in uuml>er equal to b v, i. e. (aa, bp) = b v. Now
since each generator meets £('> in v points, a = v, also a = p v = ti a,
and we have

a ft -\- b a — pa ft = fxaft + bv — fi a ft = b v;
therefore (aa, bp) = a ft + b a — ll a ft.

Theorem III. — If aa is irreducible and the partial intersection of
2W and £<"), a' a' being the residual intersection, and if the formula holds
for each irreducible component of a' a> with an arbitrary curve bp on 2(^,
it also holds for aa with bp.

Proof. — The residual a'0> may break up into several curves, but
bp, being arbitrary, does not in general contain any part of the inter-
section of 2<^) and £H"\ If a'a> is reducible, the order a' is the sum of
the orders of the component curves, and the number of points a in which
any generator meets a1 a< is the sum of the numbers of points in which this
generator meets the component curves. Since the complete intersection
of 2M and £(') is aa + a a> we have, by Theorem II,

(aa + a' a , bp) = (a+a')ft+b(a + a')-ti(a + a') ft.

Ly supposition

(a'„ bp) = a' ft + I a - fxu ft.

Now the number of points in which bp meets the complete intersection
lesa the number in which it meets a'a> must be the number of points in
which it meets '/„ ; therefore

( "a, bp) = a ft + b a — fL a ft.

Corollary. — If the complete intersection of 2'M and X" ' consists of
two curves and the formula holds for <>ne of these curves it holds for the

other also.

1. In Order then to prove the formula for any 2(l1 it snllices first to

prove Theorem I. and then to show thai everj curve on 2"" can be cut

out by an & '"■ ach that the residual i.-, a curve, or is composed of CUrvt .

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 23

for which the formula holds. Now the formula holds for every genera-
tor, i.e. for a 10, since 10 meets bp in (3 points, and the formula gives
(10, bp) =■ 1./8 + b.O — fjL.0,/3 = /S. Therefore, if every conic can be cut
out by an SW such that the residual is uothing but generators, if every
cubic curve can be cut out by an S^ such that the residual consists
entirely of conies or generators or both, and in general, if every aa on
2^) can be cut out by an <SW such that the residual * is of order less
than a or is composed of curves of orders less than a, the formula is
true.

5. For certain species of C^'s it may be possible to choose v smaller
than for certain other species, e. g. all the quartic curve* lying on a
quadric surface can certainly be cut out by cubic surfaces, but the
" quartics of the first kind " can also be cut out by quadric surfaces.

We shall first determine the lowest value of v for which we can be
certain that an S^ will cut out any species of 0-a\ This can be done
for a surface of any order, fx, without difficulty, but since we are here
going to treat the ruled quartics only, we shall consider the case of
fx. — 4 only.

iSW is determined by }.- (v + 1) (v + 2) (v + 3) — 1 arbitrary points. f
When v > 5 we must take care that SW does not break up into 2,4) and a
surface of order v — 4, i. e. of the points necessary to determine S^ we
must take one more than enough to determine a surface of order v — 4 as
not lying on 2i4) ; also, we must take a v + 1 points of £<■') on C1"1 in
order that SW may contain this curve ; so that for v > 5 the number of
arbitrary points of 2,4) through which we can make $» pass is

(2) . . . £(v + l)(v+2)(v+3)-l_£(v-3)(v-2(v-l)-(av+l)

= 2 v2 - a

V.

For v — 4 we must take one point of »SW not lying on 2(4), but then the
term \ (y — 3) (v — 2) (v — 1 ) = 1 ; for v = 1, 2, or 3 we do not have
to take any points of »$W off 2,4\ but then the term £ (v — 3) (v — 2)
(v — 1) = 0; therefore formula (2) gives the number of arbitrary
points for all values of v, when /x = 4. We have, therefore, 2 v2 — a v > 0,

which gives at once v >„; so that, for the lowest value of v, we have

Li

v =■ - when a is even, and v — when a is odd. In some cases it

2 2

has been found more convenient, and apparently necessary, to take v

* The letter r, when used, shall always denote the order of the total residual.
t Salmon's Geom. of Three Dimensions, Chap. XI.

21 PROCEEDINGS OF THE AMERICAN ACADEMY.

greater by one than this lowest value, in order to be able to make the
residua] consist of curves of orders less than a.

[n determining the number of points at oar disposal, given by formula (2),
we have said nothing about multiple points on C"'\ but have supposed
that the points of <SW that had to be taken on 0 were ordinary points
on this curve, and we shall always consider v chosen without regard to
multiple points on C"K If a surface be made to pass through an ordi-
nary point of a curve it meets the curve once at that point, and therefore
we have to make S") pass through a v + \ ordinary points of C'" in
order to make $") contain C(r" ; but if Cw has a double point, any surface
through this double point will meet the curve twice there, and therefore,
if we make SM pass through this double point (which counts for only
one point in the determination of S1-'')), we have to make it pass through
only a v — 1 other ordinary points of C"' in order to make it contain
C"\ since C{,,) will then intersect <Sv*) in a v — 1 + 2 = a v + 1 points.
Consequently, when C{a) has a double point, only a v of the points neces-
sary to determine *SW need be taken on C""' if we take the double point
to be one of these : this is one less than the number of points of Sfc ' taken
on C'("' in deduciug formula (2), above; and therefore, when C{'" has a
double point, we shall have at our disposal one point more than the
number given by formula (2). In like manner, if C'1"1 has an ///-tuple
point, a surface S^ through that point meets Cio) m times there, and we
need only make S ''» pass through a v— (in — 1) other ordinary points
in order that it shall contain C{" ; and, consequently, when C,a) has an
///-tuple point, we shall have at our disposal m — 1 points more than the
number given by formula (2).

In accordance with this principle, it is evident that, if v + 1 branches
of C'" meet any line L* (i. e. if v + 1 of the points of intersection of

* It is necessary to observe here a very important fact, which is often over-
looked, viz., if a curve has an m-tuple point /' and the //< tangente to the curve at
P all lie in the same plane, a Burface on which the curve lies may have this point
Pm an ordinary point, and any line /- through /'. th.it .lues net lie in the tangent
plane, meets this BUrface only once at /'. but there an //, br I the curve

that meet /. at /' and a plane through /. meets the curve m times at /' In gen-
eral, if the in tangents at P do not all lie in the same plane, /' will be a multiple
point on any surface that contains the curve, the multiplicity I of /' on any such
surface being at least equal to the onler of the cone of lowest order that cai

ed through the //, tangents, ami this lowest order i- always less than m. Then
any line /. that does not lie on the tangent cone to the surface .-it /' meets the r-ur-

timea at P, while ai fthecot ts the surfa I ;

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 25

Cia> with au arbitrary plane through the line L lie on Z), we can then
make L lie on the surface *SW that cuts out C[a) and still have left at our
disposal the number of points given by formula (2). For, if the v+1
points of C(a) on L are all ordinary points, L meets $» v+1 times and
therefore lies on $"> ; but, if C(c° has an ra-tuple point and v — m + 1
other points on L, even if L meets <$» only once at this multiple point,
m — 1 points not included in formula (2) are still at our disposal and
may be taken on L, so that v+1 points of <SM lie on L and L will
therefore lie on -SW. Therefore, in considering whether a line can be
made to lie on S^ or not, we need not take account of the multiple
points of C{a) that lie on this line, but may regard the line as meeting
£("> in points of C{,,) equal in number to the number of points on the
line in which an arbitrary plane through the line meets C"a).

6. We have seen that, when certain theorems can be proved, formula
(1), p. 21, gives the number of intersections of any two curves on the
same ruled surface. In special cases, where the curves bear a particular
relation to one another, and in most cases where the multiple curve is
involved, the result given by this formula requires a special interpreta-
tion, namely : if two curves on 2(41 pass through the same point of the
multiple curve, any branch of either curve is regarded as intersecting
only those branches of the other curve that lie on the same sheet with it,
and two branches that pass through the same point of the multiple curve
are not regarded as intersecting at that point if they lie on different
sheets of 2(4) there. In particular, two generators through the same
point of the multiple curve are not regarded as intersecting, when con-
sidered as loci on that quartic surface.

The double curve on a ruled quartic cannot be of order greater than
three, and therefore a plane section can never have more than three
double points on the double curve. If the plane section has a double
point not on the double curve, this double point is a point of tangency of
the plane, and, since a tangent plane to a ruled surface contains the gen-
erator through the point of tangency, the section must be a degenerate
quartic curve having at least one generator as a component. Therefore
the section of a ruled quartic by a plane can never consist of two proper
conies ; for the section would then have four double points, one of which
must be a point of tangejicy of the plane, and therefore the plane would
cut out a generator.

Cayley * uses the general symbol £ (m, n, p) to denote a scroll gener-

* Second Memoir on Skew Surfaces, Otherwise Scrolls. Coll. Math. Papers,
Vol. V., and Phil. Trans., 1863

26 PROCEEDINGS OF THE AMERICAN ACADEMY.

1 by a line that meets each of three curves of orders m} n, and p once,
.-.m to denote a scroll generated by a line that meets a curve of
order m twice and a curve of order n once, and S (m9) to denote :i scroll
g rated by a line that meets a curve of order /// three times. In his
Bymbols for the quartic scrolls be has also used a subscript, in most cas< -.
to denote the order of multiplicity of the curve on the scroll : but he lias
not, in all cases, adhered to his general method, and it senus best, while
preserving his classification, to change his symbols, making them con-
form to his general rule for such symbols.

II. Quartic Scroll, with a Triple Linear Director and a
Simple Linear Director, S(\3, 1, 4). (Catley's Third Sfecu s,
S(l8,l,4).)

1. This scroll has three sheets through the triple linear director,
which we shall denote by T, and T is scrolar* on each sic

Through each point of T pass three generators, one on each sheet,
and if we pass a plane through two of these generators it will also con-
tain the simple director, since each generator meets the simple director
once, and therefore the third generator at the point lies in this Bame
plane, for it meets it once at the point and once on the simple director;
i. e. any plane through the simple director meets the scroll in this
director ami in three generators that intersect in the point where the
plane meets T.

'1. Proof of Theorem I. — Pass a plane through 7'; it meet- the
.-•roll in 7' ami one generator and meets C1 » in a points. Now if we
revolve the plane about T it will cut out, in succession, each genera-
tor of the scroll, and since the plane always met I in the same num-
ber of point-, say t points, on T, it inert- c > in the Bame number of
points, say « point- on each generator, where t -| a "• Since three

_ " _ ~ "

generators lie in a plane a < - and t > — .

:;. Plane Ourves.- -A plane that dor- not pass through any line on
the scroll, i. e. an arbitrary plain-, meets the bctoII in a plane quartic
curve having a triple point on '/'. and Bince an arbitrary generator m<
the plain- once, ever} plane quartic is a I .

A plane through one and only one generator cuts oul a plain- cubic
having a double point on '/'. through which the generator passes, mak

tyley calls a line on a surface when the tangent plane to tin su

i- different si each point along the lin

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 27

a triple point ou the complete intersection ; the generator meets the
cuhic again where the plane is tangent to the scroll. Since an arbitrary
generator meets the plane once and does not meet the generator lying in
that plane, it meets the plane cubic once, and therefore every plane cubic
is a 3!. If a plane cuts out a proper conic, it must also cut out another
conic, which must be an improper conic consisting of two lines through a
point of the proper conic, since the section by the plane must have a triple
point on T; but the only lines on the scroll that pass through a point of
rl\ besides T itself, are generators, and we have seen that a plane through
two generators also cuts out a third generator and the simple director ;
therefore, there are no conies on this scroll. The triple director T is
met once by each generator, aud is, therefore, a 3^ The simple director
is met once by each generator, and is, therefore, a \x.

Each of these plane curves is either the complete intersection of the
scroll and a plane, or else the residual intersection consists entirely of
generators, and therefore by Theorems II and III, formula (1) holds
for every plane curve on the scroll.

A plane quartic has a branch on each sheet where it crosses T, and
therefore meets T three times, as the formula shows,

(4, , 30 = 4 + 3 - 4 = 3.

A plane cubic has a branch on each of two sheets where it crosses T, and
therefore meets T twice, (3t , 3j) = 3 + 3 — 4 = 2. The simple di-
rector does not meet 2] (11} 3j) = 1 -f- 3 — 4 = 0. The simple direc-
tor meets a plane once, and therefore meets a plane quartic once,
Oij 4X) = 1, but since it meets each generator once, it cannot meet a
plane cubic, (11, 3X) = 0. Two plane cubics intersect in two points,
on the line of intersection of their planes, (3X , 3X) = 3 + 3 — - 4 = 2,
the other two points, where this line meets the scroll, being the two
points where each cubic is met by the generator that lies in the plane of
the other. A plane cubic and a plane quartic intersect in three points
on the line common to their planes, (3t , 4X) = 3 -f- 4 — 4 = 3, the
fourth point where this line meets the scroll being the point where the
generator in the plane of the cubic meets the plane quartic. Two plane
quartics meet in four points on the line of intersection of their planes,
(41,41)=4+ 4-4 = 4.

4. Iwisted Cubic 3V — We saw that when a is odd we can take

a -f 1
v = — - — ; so, for the twisted cubic, v = 2, and by formula (2), p. 23,

wc have two points at our disposal in the determination of this quadric,

28 PROCEEDINGS OF THE AMERICAN ACADEMY.

S 2\ that cuts out the cubic. Since a < - we have a = 1, and every

twisted cubic is a 3j ; therefore t = 3 — 1 = 2, i. c. the twisted cubic
meets T in two points, which must be distinct, since a twisted cubic
cannot hav< a double point. Therefore the quadric that cuts out the
twisted cubic meets T in two points on tins curve, and since we can
make the quadric pass through any two points we please that are not on
the curve, we can make it pass through another point of 7) and it will
then contain T. The residual intersection, which is of order 5, then
consists of T, which counts for three lines, and two generators, since
there are no conies on the scroll, and, moreover, each generator meets
the quadric once on 7' and once on the twisted cubic, and cannot meet it
again without lying on it; and if a conic or the simple director formed
part of the residual, an infinite number of generators would lie on the
quadric, which is impossible. Since formula (1) holds for T and the
generators, by Theorem III it holds for every twisted cubic.

A plane through the simple director cuts out three generators and
meets the twisted cubic three times, once on each generator, and there-
fore the twisted cubic does not meet the simple director.

5. Twisted Quartic, 4V — We have a = 4, v = 5 = 2, and r = 4,

where r is the order of the residual ; a = 1, and every twisted quartic is

a 4r Hence t = 3, and if the quartic has no double point on T there

must be three distinct points of the curve on T, i. e. T meets the quadric

that cuts out the quartic three times, and therefore lies on it. It' the

quartic has a double point on T, it is a "quartic of the first kind." and

we can pass a quadric through it and through any arbitrary point not on

the curve; the quadric already meets T twice on the quartic curve, and

if we make it pass through another point of T, 7' will lie entirely on it.

In any case, the twisted quartic. can be cut out by a quadric such that

the residua] will consist of T and one generator, and, since formula 1 l )

holds for T and all generators, it holds for every twisted quartic; if T

lit-, (in the quadric the simple director cannot form part of the residual,

Bince each generator already meets the quadric once ou T and once on

the twisted quartic. The twisted quartic has a point on each generator,

and therefore meets the simple director once, tl,, l,) = 1.

" + 1
<',. Twisted ( 'urves, in general. — When a is odd we take v = - —

a

whence /• a f- 2 ; and when " is even we take r . whence r a.

a "_' 11

We Baw that 11 . - and r - where r is the ber of points on 7',

• > .1

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 29

in which aa is met by a plane through T. Yor a = 3, we saw that T
could be made to lie on S(2\ the quadric that cuts out the twisted cubic.

Now for a > 3, v = — - — < t when a is odd and v = - < t when a is

even. It follows at once, therefore, from what was said on p. 24, that
we can always make 77lie on £HV\ Therefore when a is odd the re-
sidual can be made to consist of T and a curve of order r — 3 = a — 1,
and when a is even the residual can be made to consist of Tand a curve
of order r — 3 = a — 3. Therefore, by Theorem III, if formula (1)
holds for every curve of order less than a, it holds for every curve of
order a ; but we have proved that it holds for all plane curves and for
all twisted curves of orders 3 and 4 ; it therefore holds for every curve
of order 5 and therefore for every curve of order 6 and so on, and it
therefore holds for every curve on the scroll.

7. The above proof is also applicable to the Quartic Scroll, with a
two-fold 3 (+1) -tuple linear director, S (13, 1,4). ( Cayley's Sixth Spe-
cies.) This scroll is, in fact, the limiting case of the scroll just consid-
ered, where the simple director has moved up into coincidence with the
triple director. Cayley denotes this symbolically by drawing a bar over
the two l's.

The triple linear director on this scroll is torsal along one of the three
sheets through it, i. e. the tangent plane to this sheet, along this director,
is the same for every point of this director; the generator, lying on this
sheet, that is cut out by this tangent plane, coincides with the simple
linear director, and is regarded as intersecting the triple linear director
at the point where the other two generators cut out by this tangent
plane, one on each sheet, intersect.

III. Quartic Scroll, with a Triple Linear Director,
#(13,2,2). (Cayley's Ninth Species, S (13).)

1. The triple director, T, is scrolar on each of the three sheets that
pass through it, and this scroll differs from the Quartic Scroll £(13, 1, 4)
in not having a simple linear director, in consequence of which we have,
on this scroll, three generators through each point of T that do not lie in
the same plane. The plane of any two of these three generators meets
the scroll otherwise in a conic that passes through their point of inter-
section on T, making up the triple point of the complete intersection.
Therefore, there are three conies through each point of T, one in each
of the three planes that contain two generators through the point.

30 PROCEEDINGS OF THE AMERICAN ACADEMY.

The proof of Theorem /is the same as for the < Quartic Scroll S (13, 1, 4),

a

p. 26. Since two generators lie in a plane, a < -.

•J. Plane Curves. — As before, each plane curve is met once by an
arbitrary generator, i.e. o= 1 for any plane curve. There is no l! on
this scroll. Each conic is a 21? the triple director T is a .'>,. and the other
plane curves, 3! and 4U are the same as for the Quartic Scroll S (1 .1
Every plane curve is either the complete intersection of the scroll h\ its
plane or else the residual is composed of generators, and therefore, by
Theorems II and III, formula (1) holds for every plane curve.

Two conies do not intersect; even if they pass through the same point
they lie on different sheets, and cannot be regarded as intersecting on the
scroll ; the formula gives (2l5 2x) = 2-f2 — 4 = 0; the line of inter-
section of the planes of the two conies meets the scroll in the four points
where the two generators in the plane of either conic meet the other
conic. In the plane of a conic each of the two generators that lie in that
plane meets the conic on T and at one other point where the plane is
tangent to the scroll; therefore the plane of every conic is a double
tangent plane to the scroll. T meets each conic once, (Slt 2X) = 3 + 2
— 4=1. A conic meets a plane cubic once, (2lt 3X) = 1, and meets a

plane quartic twice, (2X, 4^ = 2 + 4 — 4 = 2.

a

3. Twisted Curves. — Since a < -, we have for the twisted cubic

a = 1 and t = '2. where t is the number of points of intersection of the
curve and an arbitrary plane through T that lie on T. l'.y the same
oning as that employed on page 2s. we see that 7' can be made to lie
on the quadric thai cuts oul ihe twisted cubic, and thai formula 1 1 ) holds
for every twisted cubic.

When '/ is odd we take v = — - — . Then, since two generators lie in

a plane and << is an integer, « < — - — and t > — - — ; bul by formula
(•J) we have - = i points at our disposal in the determination of

and therefore We can make 7* lie On S • ' : the residual will then con-
sist of '/'.ind a curve of order r — 8 = a + 2 — 3 = a— 1.

,ii • , " " . _« ,, " a

When n \~ even We take i -; a . - and t . -. If a - -, r> -,

w & — — —

i.e. t.i, and it follow- from what was said on page -I thai '/' can be
made to lie on S" ' ; the residual will then consist of '/'ami a curve of

order a - •'!. [fa : c, every generator meel y in v points, which

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 31

are points of aa; if, then, any generator meets SM in an additional point
it must lie on £("), and therefore if any generator has on it a point of the
residual it lies on SW and is itself a part of the residual ; therefore, when

a = - the residual consists entirely of generators which are a in number,

since the residual is of order a ; and if the curve has no multiple points

on T, there are — pairs of generators that pass through the ■ ■ points

where aa meets T.

We have shown then that every twisted curve of order a can be cut
out by an S(") such that the residual will consist of curves of orders less
than a, and it therefore follows, as on page 29, that formula (1) holds
for every curve on the scroll.

IV. Quartic Scroll, with two Double Linear Directors and
with a Double Generator, »S (12, 12, 2). (Caylet's Second
Species, »S"(12, 12, 4).)

1. Let us call the double linear directors D and D' ; they do not
intersect, and a plane through either of them cuts out also two generators
that intersect in the point where the plane meets the other director,
i. e. any generator A meets a definite generator B on D and another
definite generator E on D' ', so that A and B lie in a plane through jy
and A and E lie in a plane through D, while B and E do not meet. In
a special form of this scroll four generators may form a gauch-quadrilat-
eral having two vertices on each double director, e. g. taking the gener-
ators above, if A and B meet D at the point P, A and E meet D' at the
point i?, B meets D' at the point S, and E meets D at the point Q, the
scroll may be of such a form that a generator F will pass through Q and
S, as can easily be shown analytically. It is also very probable that
there are special forms of this scroll on which any even number of gen-
erators, greater than four, form a gauch-polygon, but it is not the purpose
of this paper to discuss these special forms. The double generator, which
we will denote by G, arises from the fact that the plane quartic direct-
ing curve has three double points, one on each of the double directors
and one through which G passes ; a plane through G and either double
director does not meet the scroll again.

2. Proof of Theorem I. — On any quartic scroll where the double
curve is a twisted cubic, either proper or degenerate, we can prove
Theorem I by passing a quadric through this twisted cubic. In the

32 PROCEEDINGS OF THE AMERICAN ACADEMY.

presenfrcase tin- twisted cubic is degenerate, consisting of D, D ', and G.
Let 11- pass a quadric through eight points, three on 1), three on B', one
oa ff, and one on any generator .1. the last two points not being on

l> or D' ; then D, D\ G, and A will all lie mi the quadric and count for
7 lines in the intersection of the scroll and quadric, and therefore the
quadric cuts out one more generator; the quadric passes through eight
fixed points and we can make it pass through an arbitrary ninth point, so
it' we vary this ninth point continuously the (juadric will cut out. in suc-
cession, each generator of the scroll. .Now < "'" meets the quadric in 2 a
points, of which a fixed number lie on D, D', G, and A, and therefore
the same Dumber of points of C", say a points, lie on each generator.
It is evident that there must be a points on A, for if any other generator
be chosen, through which the quadric is always to pass, then there is the
same number of points, a, on A, as on each of the other generators.
Since we can pass a plane through J) and two generators, there are
a — 2a points of aa on D, and, similarly, there are a — 2 a points of
aa on D'. A plane through D and G meets the scroll in these two lines
only, and then- are, therefore, 2u points of aa on G, as is otherwise evi-
dent from the fact that G counts for two generators. Since a twisted

curve of order a cannot have a points on any line, 2 a < a or a < for

every twisted curve on the scroll.

3. Phtne Curves. — Each double director is met once by any gener-
ator and is therefore a 2,. No generator can meet G\ for. BUppose a
generator .1 does meet it; then A meets either Dor /)' in a point differ-
ent from that in which G meet- it. and therefore the plane through G
and A contains also I) or 27, making the order of the complete intersec-
tion of the plane and scroll as great a- 5, which is impossible. G is

therefore a 2,,. Then any plane through <•■ that doe- not contain I) or //,
meet- the scroll in a proper conic that doe- not meet either double
director, since the section has only a double point on each double director ;
-ince each <_'<'iierator meets the plane mice and doe- not meet (,\ each
conic is a 2,, and the COnicS and D and //are theonh cur\es on the

u

BCroll for which u = -. A plane through one and onlv one generator

2 i .

cuts out a plate cubic, a -5,. having a double point on G and passing once
through the two points where the generator in the plane me< ts 1 > and I > \
the generator meet-, the cubic again where the plane i- tangeul to the

scroll. A plain- that do.- not contain a line of the scroll cuts out a plane

quartic, a l,. having three double points, one on G and one on each

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 33

double director. Every plane quartic is the complete intersection of its
plane and the scroll, and therefore formula (1) holds for it (Theorem II).
A plane cubic is cut out by a plane through a single generator, and D
and D' are cut out by planes through two generators, and therefore, by
Theorem III, formula (1) holds for every plane cubic and for D and D'.
We can cut out G by a plane through D, and, since formula (1) holds for
D, by corollary to Theorem III, it holds for G. Each conic is cutout by
a plane through G, and, since formula (1) holds for G, it holds for each
conic. Therefore formula (1) holds for every plane curve on the scroll.

Either double director and G lie on both sheets of the scroll through
them, respectively, and G, therefore, intersects either double director
twice, once on each sheet, (20, 2X) = 2. A plane quartic has a branch
on each sheet, at each of the three points where it meets D, D', and G,
and it therefore intersects each of these lines twice, once on each sheet,
as the formula shows, (4X, 2X) = 2 and (4l5 20) = 2. A plane cubic
meets G twice, since it has a branch on each sheet where it crosses G,
(31? 20) = 2, but it meets each double director once, since it has a
branch on one sheet only, where it crosses either double director,
(3j, 2X) = 8 + 2 — 4 = 1. The plane of a conic passes through G and
is tangent to the scroll at two points along G, one on each sheet ; these
points of tangency are the two points of intersection of the conic and G,
(2r 20) = 2, and the conic has a branch on each sheet; one point of
tangency lies on the finite segment of G, between D and //, and the
other lies on the infinite segment, so that, as we turn the plane about G
in one direction, these two points both move toward the intersection of
G and D and coincide at this intersection, forming a pinch point, when
the conic becomes D, i. e. a line on each sheet; if we turn the plane in
the other direction, or continue to turn it in the same direction after it
cuts out D, the two points of tangency will both move toward I)' and
will coincide at the pinch point, the intersection of G and I/, when the
conic becomes D '.

It is easy to see, by the aid of formula (1), how the other plane curves
intersect.

4. Twisted Cubic, Sv — We have seen that a < - for all twisted

curves, and, consequently, every twisted cubic is a 3X. Also, if 8 be the
number of points of the curve on D or D ', 8 = a — 2 a = 1 for the

. / a + 1 \ ,

cubic. The twisted cubic is cut out by a quadnc [v = — ^ — 1, and we

can make the quadric contain D, since, by formula (2), we have two

VOL. XXXVI. — o

34 PROCEEDINGS OP THE AMERICAN ACADEMY.

points at our disposal in determining the qaadric. Every generator then
:- the quadric once on the twisted cubic and once on J), and cannot
meet it again without lying on it; therefore, the residual consists of D
and three generators. Since formula (1 ) holds for D&nd the generators,
it holds for every twisted cubic on the scroll (Theorem III).

5. Twisted Quartic 4j. — Since a< . every twisted quartic is a

1, ; 8 = a — 2 a = 2; i. e. a plane through h or 1/ meets tlie twisted
quartic twice on that line. If tlie curve is a "quartic of the lirst kind "'
it may have a double point on D, D' , or G. but in any case it has two
distinct or consecutive points on one of the double directors, Bay />. and
-hire we can pass a quadric through a M quartic of the first kind " and
any arbitrary point, we may take this arbitrary point on D, and D will
then lie on the quadric ; if the quartic has no double point on G tlie
residual will then consist of D and G, and if the quartic- has a double
point on G the residual will consist of D and two generators, since
each generator will then meet the quadric once on 1) and once on
the quartic. If the curve is a "quartic of the second kind," it is more
convenient to cut it out by a cubic surface, i. e. we take v = 3 ; the re-
sidual then is of order 8, and by formula (2) we have six points at our
disposal in the determination ot .s'11; since this quartic has do double
point, it meets both D and D' in two distinct or consecutive point-, and
if we put two more points of «S on each of these double directors they
will both lie on S3' ; then G will meet S once on D, once on //, and
twice on the quartic, and will, therefore lie on >'•"' ; each generator meets
8 once on D, once on />', and once on the quartic, and, since we -till
have two points at our disposal, we can make two generators lie on S ;
the residual will then consist of D,D\ G, and two generators. Since
formula (1) holds for />. I >' <>. and the generators, it hold- for every
twisted quartic on the Bcroll (Theorem [II).

6. Twitted Curvet in General — When a is odd we take v = — - — ,

a
and the oiil.r of the residual is r = a + 2. We have -ecu that a < -

or l .- <i . .and that there- i> the same number "1 point-. -i\ 8

points, on each double director, where 8 =a— 2a, i.e. 1 < 8< " — -•

It has also been Bhown, p. 25, that we need not consider whether ",, has

a \- 1
multiple points on 1) and l> or not. By formula (2), we have

point- at our disposal, in ih< del ( i initiation of > ■ . and therefore we can

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 35

make both D and U lie on £W if 8 + °—*— % ^— + 1, i.e. if

4 2

0 __ a + 5 = 3 a — 5 _ 3 a — 5

3 tt - - or a < 5 . .therefore for a < — — the residual

■^ 8 o 8

can be made to consist of Z), ZX, and a curve of order r — 4 = a — 2.

When a > , at least one point of aa lies on D, and since we have

8

■ — — — points at our disposal, we can make D lie on <S("), and the residual

will then consist of D and a curve of order r — 2 = a, say ap, where p
is the number of points of this curve on each generator; now formula (1)
holds for D, and if it holds for ap it will hold for aa (Theorem III),
so we need only consider the curve ap ; each generator meets SW in
a + 1

v =

2

points, once on D, a times on aa, and p times on ap, so that

a+1 , a+1 3a-5 a+ 1

p == — 1 - a < -j- - 1 r-orp< -g-,

3 a 5

i. e. p < for a > 3, and ap is therefore a curve like that consid-

8

ered above, that can be cut out by an »SW such that the residual consists
of curves of orders less than a.

When a is even, it is convenient to separate the curves into two divi-
sions, according as - is odd or even. If - is odd we take v = - ; then
2i Z u

a . . — a .

r = a, a < -z, and 8 must be even since 8 = a — 2a; if 8 > - + 1, i. e.
^2 *

a- 2

T

and the residual consists of D, D' ', and a curve of order r — 4 = a — 4 ;

if a > the residual i9 a curve of order a, say an ap, where

4

a a a — 2 " + 2,<7 + 2. . ,

pz=z a< _, or p < —j- ; but — — is an integer and p

"■ +2 _ a — 2

is an integer, and therefore p < — 1, i. e. p < -—r — ; tnererore,

ap is a curve like that just considered, that can be cut out by an £W

a
such that the residual will consist of curves of orders less than a. It -

is even we first take v — r, for which r — a; 8 is even, and if 8 > -- + 2,

i. e. if 0 = a~ 4 both i) and D' lie on 5(") (or can be made to lie on
^ 4

ft — Zt

if a < — - — , both Z) and D' lie on 5<") (or can be made to lie on SW),

a

36 PROCEEDINGS OF THE AMERICAN ACADEMY.

5 i, ami the residual then consists of JJ, JJ', ami a curve of order a — 4 ;

a — A _ a , (i ....

if a > - — , a > — , and wlii'ii a ;> - the residual is an "p, where
14 4

a a a _ a — 4 , . . . ,

p = — a < ^ — - or p p — , and aD is therefore a curve that can lie

2 2 4 4

cut out hy an 5M such that the residual will consist of curves of orders

a a

less than '/ ; finally, if a = - we take v = - + 1 ; then r — a -f 1 and

8 = ; if then we put two more points of S^ on D and two more on

I)', both of these douhle directors will lie on S(l'\ and this we can always
do, since, hy formula (2), we have a -f 2 points at our disposal in the

determination of S^ and a > 4 ; then C will meet <SW once on D, once

o

on JJ', and 2a = - times on aa, and will therefore lie on S(v); each gen-
erator will meet <SW once on Z), once on Z)', and - times on aa, and, con-

4

sequently, if we put - more points of *SW on any generator it will lie on

51") ; now we still have at our disposal a — 2 > 2 ( - J points, since

a > 1. and therefore we can make two generators lie on <Sv"); therefore
the residual can he made to consist of U, JJ', G, two generators, and a
curve of order r — 8 = a — 4.

Therefore, on this scroll, we may divide all twisted curves into two
groups, viz. group (1), those that may be cut out by an <SW such that the
residual consists of curves of orders less than a, and group (2i, those
thai may he cut out by an »$(") such that the residual is a curve of group
(1 ). with or without D. Now we have seen that formula (1 ) holds for
all plane curves and for all twisted curves of order '■'> or I ; it therefore
holds for all curves of order 5 of group (1) (Theorem III), and it then-
fore holds for all curves of order 5 of group (2); it then holds for all
curves of order 6 of group (1 I, and therefore for all curves of order 6 of
group (2), and bo on. Therefore formula (1) holds for even curve on
the scroll.

7. The above proof is also applicable to the Quartic Scrollj with a
two-fold 2 | • 'l)-tnjilr linear director, and with a double generator

S I . I . •_' i, ( ( 'at/ley's Fifth Species, S \ . I . I > >. for this Bcroll is Bimplj
the limiting case of the Bcroll just considered, where one of the double
director- baa moved up into coincidence with the other, A plane quartic
has a tac-node, where it meet- the two-fold director and has a double
point on the double g< aerator. A plane cubic, iv^ml.-d as lying in the

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 37

plane, is tangent to the generator in its plane where it meets the two-
fold director, but regarded as lying on the scroll, it does not meet
the generator there, for they lie on different sheets, and the generator
meets the cubic in one point only, where the plane is tangent to the
scroll — the formula giving (10, 3X) = 1. The system of conies is the
same as on the scroll just considered. The two-fold director may be re-
garded as a 42, since it has two lines on each of the two sheets through
it and is met twice by each generator. A plane quartic has a branch
on each sheet, and therefore meets the two-fold director four times,
(-!•!, 42) = 8 + 4 — 8 = 4, while a plane cubic has a branch on one sheet
only, and therefore meets this director twice, (3X, 42) = 6 + 4 — 8 = 2.

V. Quartic Scroll, with two Docble Linear Directors and
without a Double Generator, S (1», 12, 4). (Caylet's First
Species, S (12, 12, 4).)

1. We shall call the double directors D and 1/ . They do not inter-
sect, and if we pass a plane through either it cuts out two generators.
The scroll is similar in many respects to the Quartic Scroll S (12, 12, 2)
already considered, and what was said there in regard to gauch-polygons
applies equally well here. But the scroll now under consideration has
no double generator, the plane quartic directing curve having only two
double points, one on each double director, and this is the only quartic
scroll on which the multiple curve is of order less than three.

2. Proof of Theorem I. — A plane through D outs out two genera-
tors that meet in a point where the plane meets D', and if we revolve
the plane about D, it will cut out, in succession, all the generators of the
scroll, two at a time. The plane meets C[a) in a points, of which a defi-
nite number, say 8 points, lie on D, and therefore there are a — 8 points
of C(a) on the two generators in the plane, taken together ; for any
given curve C'"\ the number a — 8 is a constant non-negative integer,
say k. Let x and y be the number of points of Oa\ respectively, on the
two generators lying in a plane through D ; then as the plane revolves
about D, we always have x + y — &> and since x, y, and k are all non-
negative integers and k is constant, there is only a finite number of
values of x and of y that will satisfy this relation. Let us, for the
moment, designate any generator by the number of points of C"5 on it,
i. e. the generator x has x points of Oin) on it, etc. To any value of x,
say (j, there corresponds a certain value ofy, say g', such that g + g1 =k;
if then a plane through D cuts out a generator g, it also cuts out a gen-

38 PROCEEDINGS OP THE AMERICAN ACADEMY.

orator </'. As there is a finite number of pairs of values of ./• and y, we
may arrange them in order of magnitude, calling// the greatest, i. e. we
shall say that g is the greatest cumber of points of C""1 on any generator;
correspondingly, g is the smallest number of poiuts of C"1 on any gen-
erator, since g + g' = k. Now a plane through // cuts out two genera-
tors, and as we revolve the plane about D' the number of points of C[,,) on
the two generators in the plane is constant and equal to a — 8', where
8' is the number of points of G'(a) ou JJ . If then we pass a plane through
J)' and a generator g' , having the least number of points of C1"1 on it,
the other generator in this plane must have the greatest number g of
points of < '" on it ; therefore g + g — a — 5', and since g + g' = a — 8,
we have 8 = 8', i. e. C'"' meets each of the double director.-. /) and 1)',
in the same Dumber of points, 8. If now a plane be passed through a
generator g and the director D it will cut out a generator g\ and if
through this generator //' and L/ we pass a plaue it will cutout another
generator </, so that there are at least t.vo generators g. Through these
two generators g, the directors D and />' (which four lines form a
gauch-quadrilateral), and an arbitrary point we can pass a quadric ;
I), D' , and the two generators counting for 2 + 2 + 1 + 1 = ('. lines in
the intersection of the quadric and scroll. Each generator meets the
quadric in two points, one on D and one ou D\ and it we take the arbi-
trary point on any generator, this generator will lie on the quadric, and
the remaining intersection of the quadric and scroll will be another gen-
erator. Thus by varying this arbitrary point all the generators of the
scroll, two at a time, will be successively cut out by a variable quadric
that always contains D, /)', and the two chosen generators^. This
quadric always meets C"" in 2a points, of which 28 lie on D and I >'
and 2g lie on the two chosen generators </, so that the remaining
2 " — 2 8 — 2 g points lie on the other two generators cut out b\ the
quadric , bul a — 8 = g + </' and therefore 2 a — 2 8 — 2 g = 2 </' .
Now there is no generator thai has fewer than g points of ('• mi it and
the number of points of <'" on the two generators together Is 2 g/ j
therefore each must have gf points of C?"' ou it. Therefore everj gen-
erator of the scroll, excepl the two chosen generators g, has g* points of
( on it. In like manner, if wechoose two generators g*, through which
(he variable quadric is always to pass, we have the Bum oi the number
of points of '"""on the other two generators cul onl \>\ the quadric
always equal to 2 a - 2 8 — 2 ;/ 2 g, and, since do generator hat

; •<• than g points of CW on it, everj generator on the scroll, excepl

the tw( has g points of 0 on it. fherefore <i </ and

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 39

every generator meets (?("> in the same number of points, say a points.
Two generators lie in a plane through D ; therefore, 8 — a — 2 a, and

_ a
a<2'

2. Plane Curves. — D and D' are both 2/s. There is no system of
conies on the scroll, for there is no double generator, and a plane through
two generators cuts out D or D'.* A plane through one and only one
generator cuts out a plane cubic, a 31? that meets each double director
once and has no double point, since the section has only three double
points which are the points where the generator in the plane meets the
cubic, one on each of the double directors, and one where the plane is
tangent to the scroll. An arbitrary plane cuts out a plane quartic, a 41?
having two and only two double points, one on each double director.
Every plane curve is, therefore, either the complete intersection of the
plane and scroll, or else the residual consists entirely of generators, and
consequently, by Theorems II and III, formula (1) holds for every plane
curve on the scroll.

3. Twisted Curves. — It may be shown in exactly the same way as
for the Quartic Scroll, S (12, 12, 2), pp. 33-3G, that formula (1) holds
for every twisted curve on the scroll. It will be observed that, in the
proof referred to, the double generator is shown to be a part of the
residual ; now, there is no double generator on this scroll, but disregard-
ing the double generator the residual is still composed of curves of orders
less than a, and the conclusion follows as before, without change.

4. The proof just employed is also applicable to the Quartic Scroll
ivit/i a two-fold 2 (+ 2) -tuple linear director and without a double gener-
ator, S (I.,, 12, 4), (Cayleys Fourth Species), which is the limiting case
of the scroll S (12 12 4), just considered, where one of the double linear
directors has moved up into coincidence with the other.

VI. Quartic Scroll, with a Double Conic and a Doup.ee
Linear Director meeting it, S (12, 22, 2). (Cayley's Seventh
Species, S (1, 2, 2).)

1. For convenience, let D represent the double linear director and let
A' represent the double conic. Any plane through D meets K in one
more point, besides the intersection of K and D, and this is a double
point on the section by the plane. The plane therefore cuts out D and a

* The section cannot consist of two proper conies. (See p. 25.)

10 PROCEEDINGS <>F THE AMERICAN ACADEMY.

conic having a double point on K, i. e. two lines that meet in this point,
which arc the two generators in the plane.

2. Proof of Theorem 1. — The double curve on this scroll is a degen-
erate t \vi>t«'l cubic, consisting of the double conic K and the double
director D. We can pass a quadric through nine arbitrary points, and if
we take five of these on K and two on A>, distinct from the point of
intersection of A' and A*. A" and D will both lie on the quadric and we
shall still have two points at our disposal; now every generator meet- K
and D. and if we take one more point of the quadric on any generator A,
it will lie on the quadric ; the quadric will then intersect the scroll in
A' counted twice, D counted twice, and the generator A, and will, there-
fore, cut out one more generator. Making the quadric always contain A',
I), and A, we have one point at our disposal, and by varying this point
continuously we make the quadric cut out, in succession, all the gener-
ators of the scroll. C(a) meets A". D, and .1 in a definite number of points,
and as it meets every quadric in 2 a points it meets each generator in the
same number of points, say a points. It is evident that A also meets
C"" in a points, for any other generator may be chosen as the one through
which the quadric is always to pass, and then A will meet C" in the same
number of points as the other generators, i. e. in a points. A plane
through D cuts out two generators, and there are therefore a — 2 a points
of aa on D. A' is the complete intersection of its plane and the scroll,
and there are therefore a points of aa on K. The number of points of

aa on D cannot be less than zero, and therefore a < =.

3. Plane Curves. — The double director D meets every generator
once and is therefore a 2,. The double conic K is met once h\ each
generator and is therefore a \x. The section by a plane not through h
or K has three double points on the double curve, one on D and two
on A'. We have seen (p. 25), that the section cannot consist of two
proper conies, and we know that a plane through two generators thai
meet On A", cuts out A>, for each generator meets D\ therefore, if a
plane cuts out a simple conic, it cuts out also two generators that meet
in a point on l>. for there arc no lines on the scroll but the generators,
and D and two generators meet only on A' or /■>; conversely, through
every point of l> pass two generators and their plain- cuts out a proper
conic; consequently, there is a system of conies that do not meet /), but

meet A' twice, for clearly the plane cannot meet /; again, and the .section

cannot have a triple point on £> unless the plane contain- />-. each of
the two generators in the plane of an) come meets the conic twice, once

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 41

in a point where the plane meets A and once where the plane is tangent
to the scroll ; therefore the section lias five double points and the plane
is a double tangent plane. The conic and either generator lie on differ-
ent sheets at the point where they meet K, and, regarding them as lying
on the scroll, they do not intersect there. An arbitrary generator meets
the plane once, and, since it does not meet either generator in the plane, it
meets the conic once, and therefore every conic is a 2X. Any plane
through one and only one generator cuts out a plane cubic, a 3U having
a double point on A" and passing once through each of the points where
the generator meets A' and D. Any plane that does not contain D, K,
or any generator, cuts out a plane quartic, a 41} having three double
points, two on A' and one on D.

The double conic A is the complete intersection of its plane with the
scroll, aud every plane quartic is the complete intersection of its plane
with the scroll, aud therefore formula (1) holds for A" and for every plane
quartic (Theorem II). A plane cubic is cut out by a plane through a
single generator, D is cut out by a plane through two generators, and
every conic is cut out by a plane through two generators, and therefore
formula (1) holds for every plane cubic, for D, and for every conic
(Theorem III). Formula (1) holds, therefore, for all plane curves on
the scroll.

4. Twisted Cubic, 3X. — Since a < - , every twisted cubic is a 3^ We

have seen that there are a points of aa on A, i. e. there are three points
of the twisted cubic on K, and, since we have two points at our disposal
in the determination of the quadric that cuts out the twisted cubic, we
can take these two points on A" and thus make the quadric contain K.
The residual, which is of order 5, will then consist of K, which counts
for 4, and one generator, and therefore formula (1) holds for every
twisted cubic on the scroll (Theorem III).

5. Twisted Quartics, 42 and 4X. — Since a < --, every twisted quartic

JU

is either a 42 or a 41# A 42 may be cut out by a quadric ; every generator
will then meet the quadric twice on the quartic curve 42 and cannot meet
it again without lying on it ; consequently, every generator that has on
it a point of the residual must lie on the quadric and form part of the
residual, and the residual therefore consists of four generators ; therefore
formula (1) holds for every 42 (Theorem III). For the 4n we take
v = 3 ; then r = 8, and we have 19 — 13 = 6 points at our disposal in
the determination of the cubic surface £l3) that cuts out the 4X. There

42 PROCEEDINGS OF THE AMERICAN ACADEMY.

are four points of the twisted quartic 4! on K, and if we take three more
point- of S on A". N :i will contain A"; D meets S{S twice on the curve
Lx, and if we take two more points of S on 1). Ȥ(8) will contain D ;
every generator will then meet SlS) once on D, once on A\ and once on
the curve 4^ and as we still have one point at our disposal we can make
5 contain a generator ; this generator, D, and K count for 7 in the
order of the residual, and therefore Sr" cuts out one more generator.
The residual then consists of K, D, and two generators, and therefore
formula (1) holds for the twisted quartic 4, (Theorem III). Formula
(1) holds, therefore, for every twisted quartic.

G. Twisted Curves in General. — When a is odd we take v = - - ;

then r = a + 2, and, by formula (2), we have — - — points at our dis-
posal in the determination of £HV\ If K does not lie on S<?\ it meets
aS*") in 2 f — - — J = a + 1 points ; but we have seen that there are a

points of aa on A', and consequently, if we take two more points of <SW
on K, »SM will contain A"; this we can always do, since the number of

. a + 1

points at our disposal is — - — > 2 for a > o". The residual will then

consist of K and a curve of order r — 4 = a — 2.

When a is even, we distinguish two kinds of curves according as is

° ° 2

odd or even. If is odd we take v = ; then r = a, and if 8 be the
number of points of aa on D, we have seen that 8 = a — 2u: now, it
8 > " + 1, i-e. if a < • -^— , D meets #<") in at least - + 1 = v + 1

points," and therefore lies on S('\ so that the residual consists of 1> and

a — 2

a curve of order r — 2 = a — 2; if a > - . the residual either

4

breaks up into curves of orders less than a, or else it is a curve of order
i\ an ap, where p = - — a, since each generator meets $'') in v = -

£ —

a - i'

points, of which a lie on ",, and p on ",, : then, since a > ,

a h 2 , . . . a + 2 .

p < : but p is an integer and — — IS an integer, so that

= a + 2 , . ="--'

p < — 1, I.e. p. . and (it, is therefore one ol the curves

• mi -• can be made to pass through » H I points of D (Se< p

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 43

just considered, that can be cut out by such an $» that the residual will

a . a

consist of D and a curve of order a — 2. It - is even and v = -, the

residual is of order a ; then 8 must be even, being equal to a — 2 a, and,

if S > — \- 2, i.e. if a < — — , D lies on S("\ and the residual consists

A

of D and a curve of order a — 2. If a > — — , then a > . When

a > - the residual either breaks up into curves of orders less than a, or

4 « •

else it is a curve of order a, say an ap, where p == v — a < - , i.e.

_ a — 4
p ^ , and consequently ap is a curve like that just considered, that

4

can be cut out by an £(") such that the residual will consist of D and a

a a

curve of order a — 2. Finally, when a = -,we take v = - + 1 5 then

r — a _j_ 4} and^ by formula (2), we have « + 2 points at our disposal
in the determination of S(v\ If K does not lie on S(v\ it meets S^) in
a + 2 points ; now K has a points of aa on it, and if we make S^ pass
through three more points of K, not on aa, S^ will contain K; this we
can always do and still have at least three more points at our disposal,

since a + 2 > 6 for a ^ 4. Since a = - , 8 = - , and if we make S(v)

pass through two more points of D, not on aa , S^ will cut out D. The
residual will then consist of K, D, and a curve of order r — 4 — 2 = a — 2.
The twisted curves on this scroll can therefore be divided into two
groups, viz., group (1), those curves of order a, each of which can be cut
out by such an £("> that the residual will consist of curves of orders less
than o, and group (2), those curves of order a, each of which can be cut
out by such an $>") that the residual will be a curve of order a and of
group (1). Therefore, by the same reasoning as that employed for
Quartic Scroll S (12, 12, 2), p. 36, Formula (1) holds for all curves on
the scroll.

VII. Quartic Scroll, with a Double Twisted Cubic met

TWICE BY EACH GENERATOR AND WITH A SlMPLE LlNEAR DIREC-
TOR, S(3,2, 1). (Cayley's Eighth Species, S(l, 3'2).)

1. Let Q represent the twisted cubic, which is the double curve on
the scroll. Through each point of Q pass two generators ; the plane of
these two venerators contains the linear director, since each generator

44 PROCEEDINGS OP THE AMERICAN ACADEMY.

meets the linear director once, and therefore this plane cuts out also a
third generator, i.e. any plane through the linear director meets Q in
three points, say L, M, and N, and cuts out the three generators L M,
J/.V. kdANL.

2. Proof of Theorem I. — We can make a quadric pass through Q
by making it pass through seven points of Q ; and since Q is a double
cubic it counts for six iu the order of the complete intersection of the
quadric and scroll ; if the eighth point for the determination of the quad-
ric be taken on any generator A, the quadric will contain A, since each
generator meets Q twice ; the remaining intersection will be any gener-
ator on which we choose to take the ninth point for the determination
of the quadric, and if we keep the first eight points fixed and vary the
ninth point continuously, the quadric will cut out in succession the differ-
ent generators of the scroll. A fixed number of points of CM lie on Q
and the chosen generator A, and therefore every generator contains the
same number of points of CW, say a points. Any generator, other than
A, can be chosen, through which the quadric is always to pass, and
therefore there are a points of C(") on A. Since the quadric meets aa in
2 a points, there are '2 a — 2a = 2 (a — a) points of aa on Q. Three gen-
erators lie in a plane through the linear director; therefore a < — , and

o

there are a — 3 a points of aa on the linear director.

3. Plane Curves. — Since the section by a plane cannot consist of two
proper conies (p. 25), a plane through a proper conic would either cut
out two generators or the simple director and one generator; but we
have seen that a plane through two generators or through the simple
linear director cuts out three generators and the linear director; there-
fore there can be no proper conic on the scroll.

A plane through one, and only one, generator cuts out a plane cubic,
a 3i, having a double point on Q and passing once through each of the
two points where the generator meets Q; the generator meets the plane
cubic in one other point where the plane is tangent to the scroll. An
arbitrary plane cuts out a plane quartic, a 4j, having three double points
on Qf and Bince any plane quartic is the complete intersection of its
plane with the scroll, formula (1) holds for it (Theorem II). The
simple lin< ar director, a 1 ,, is cut out by a plane through three generators,
and every plane cubic is cut out by a plane through one generator, and
therefore formula (1) holds for the simple linear director and for every
plane cubic (Theorem III). Therefore formula (1) holds for all plane
curves on the scroll

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 45

The double cubic Q, although not a plane curve, will be considered
here. We have seen that it is cut out by a quadric through two gener-
ators, and therefore formula (1) holds for it (Theorem III). It is met
twice by every generator, and is, therefore, a 62. A plane quartic has a
branch on each sheet at each of the three points where it meets Q, and
the number of its intersections with Q is 6, (62, 4X) = 6 + 8 — 8 = 6.
A plane cubic meets Q four times, twice at the double point of the plane
cubic and once at each of the other two points where the plane meets
Q, (62, 3j) = 6 + 6 — 8 = 4. The linear director does not meet Q,
(02,10=6 + 2-8 = 0.

__ a

4. Twisted Cubic, 3V — Since a < -, every twisted cubic is a 3^ We

o

take v = 3 ; then r = 9, and we have 19 — 10 = 9 points at our disposal
in the determination of S{:i) that cuts out the twisted cubic. The number
of points of the twisted cubic on Q is 2 (a — a) = 4, and if we take 6
more points of S{?j) on Q, S{3) will contain Q ; this leaves 9 — 6 = 3 points
at our disposal, and, since each generator now meets S{S) twice on Q and
once on the twisted cubic, we can take one more point of *S(3) on each of
three generators and S{:i) will then contain those three generators ; the
residual will then consist of Q and three generators, and therefore
formula (1) holds for every twisted cubic (Theorem III). Since three
generators lie in a plane through the linear director, a twisted cubic does
not meet the linear director, (3l5 lx) = 3 + 1 — 4 = 0.

5. Twisted Quartic 4X. — Since a < -, every twisted quartic is a 4X.

o

We take v — 3 ; then r = 8, and the number of points at our disposal in

the determination of Sl3) is 19 — 13 = 6. There are 2 (a — a) = 6 points

of the twisted quartic on Q, and if we take four more points of Sl3) on Q,

S{3) will contain Q. Each generator will then meet Si3) twice on Q, and

once on the twisted quartic, and, since we still have two points at our

disposal, we can make S[3) cut out two generators. The residual will

then consist of Q and two generators, and therefore formula (1) holds

for every twisted quartic.

a + 3

6. Twisted Curves in General. — When a is odd we take v = — ^ — »

3a + 9 ,.

then r = a + 6, and by formula (2) we have <r points at our dis-
posal in the determination of £("). If 8 be the number of points of aa on
Q,8 = 2 (a — a) ^-^, since a ^ |. That £<"> may contain Q, it must

16 PROCEEDINGS OF THE AMERICAN ACADEMY.

Q 111

pass through 3 v + 1 = points of Q, and therefore, if we take

- 1 1 . _ a + 33 , , . „ , _ _, . ...

— 8 < — of tlie points at our disposal on Q, oW will con-

_ u

(J; if A be the number of points left at our disposal, after making

n * = 3 a + 9 « + 33 = 1 a

aS<*'' contain Q, A > — — , i. e. A > — — 1. JSow each

2 0 o

venerator meets *S(') twice on Q and a times on '/.,. i. e. at Least three

- « « + 3
times since a > 1 ; if, then, we make o(") pass through — h 1 — 3

= — - — other points on any generator, that generator will lie on .s .

2 1 a

We can, therefore, make at least two generators lie on S(l'\ since — — 1

> 2 ( — — ), and the residual will then consist of Q, two generators,

and a curve of order r — G — 2 = a — 2.

n
When a is even we take v = - 4- 1 ; then r = a -f 4, and, by form-

ula (2), we have a + 2 points at our disposal in the determination of

I a

$''). The number of points of ott on Q is 8 = 2 (a — a) > — . In

o

order for &<'') to contain (?, it must pass through 8v-J- 1 =

22

] k »ints of <?, and therefore in addition to the 8 points we must make 5

. 3a + 8 . _a + 24 . „ _ ..

pass tlirougb — o < other points ol \' : this we can

2 0 a I 2 1

always do, since the number of points at our disposal is a + 2 >-

for a > 4. The residual will then consist of (? and a curve of order
r - 6 = a — 2.

We have shown, then, that every twisted curve of order a can be cut
out 1>\ an SM such that the residual is composed of curves of orders less
than "; we have also shown that formula (1) holds for every pi
curve and for every twisted curve of order S or 1; it holds, therefore,
for every twisted curve of order •">, and therefore ti>r every twisted curve
of order <'.. and so on (Theorem III); therefore formula (l ) holds for
v curve on the scroll.

YIIJ. Qi kRTic Scroll, with a Doubli Twisted Cubic kei

Tumi BT BACB GENERATOR, S (8 .-'). (CaYLEY'8 TENTB

Sri- n>. 5(8 |.)

I. l.i t Q be the donble twisted cubic Through every point of Q
two generators. The Bcroll diffei from the Quartic Scroll S

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 47

in not having a simple linear director, in consequence of which a plane
through two generators cuts out a proper conic.

2. Proof of. Theorem I. — By passing a quadric surface through Q
and any chosen generator, Theorem I is proved in exactly the same way
as it is for the Quartic Scroll S (3.22, 1), p. 44., and if a be the number
of points of C(tt> on each generator, we see, as before, that the number of

points of aa on Q is 2 (a — a) . Since two generators lie in a plane,

_ a

a<2\

3. Plane Curves. — Since there are no lines on the scroll except the

generators, a plane through two generators cuts out a proper conic, and
since the section has three double points on Q, the conic passes through
each of the two points in which the generators meet Q, different from
their point of intersection. Each conic is a 2l9 and regarding the conic
and generators as lying on the scroll, the conic is met once only by
each generator in its plane, since at the point where the conic and gen-
erator cross Q they lie on different sheets. The two points where
the two generators meet the conic, not on Q, are points of tangency
of the plane, and therefore through every point of Q there is a double
tangent plane to the scroll. By Theorem III, formula (1) holds for
every conic. There is no lx on the scroll, and the other plane curves,
the cubic, 3l , and the quartic, 41} are the same as those of the same order
on the Quartic Scroll S (322, 1) and therefore formula (1) holds for all
curves on the scroll. Since Q is a double twisted cubic met twice by
each generator, it is a 6», and since it is cut out by a quadric through
two generators, formula (1) holds for it (Theorem III). We have seen
that Q meets each conic twice, and the formula gives

(G2, 2X) = 6 + 4-8 = 2.

4. Twisted Curves. — It is proved in exactly the same way as for the
Quartic Scroll S (322, 1) that formula (1) holds for every twisted curve
on the scroll ; therefore, it holds for every curve on the scroll.

IX. Quartic Developable.

1. This surface is formed by the tangents to a twisted cubic, E\ E is
then a double curve on the surface, and from one point of view, this sur-
face is the limiting case of the Quartic Scroll S(322, 1) where the two
points in which any generator meets the twisted cubic have become con-
secutive ; the twisted cubic then becomes the " edge of regression " and
the three double points of the section by a plane become cusps.

■Is PROCEEDINGS OP THE AMERICAN ACADEMY.

'_'. Proof of Theorem I. — By passing a qnadric through E. which is
the cuspidal edge of the developable, and any chosen generator. Theo-
rem I i> proved in exactly the same way as it is for the Quartic Scroll
I), p. 44, and if a he the number of points of C ' on each gen-
erator, E has J ('/ — a) points of a,x on it. Since E is cut out by a
quadric through two generators, formula (1) holds for it (Theorem III).
1 icing a double twisted cubic met twice by each generator, E is a >'_., and
formula (1) gives, for the number of intersections of E with </(i,

(6, , n0) = 6o + 2o — 8 a = 2 (a — a).

3. Plane Ourves. — A "plane of the system," i. e. an osculating plane
of E, cuts out two consecutive generators and a proper conic* This
conic, which is tangent to E,f can never break up, for there are no lines
on the surface except the generators, and a plane cannot contain more
than two generators (which must, moreover, be consecutive), since it
cannot meet E in more than three points. Each conic is a 2j , and there
is a conic through every point of E.

A plane through one and only one generator cuts out a plane cubic,
a 31} that has a cusp at the point not on the generator, where the plane
meets E, and has the generator for its real inflectional tangent at the
point where the generator meets E ; for, since the two consecutive points
in which the generator meets A7 must be double points of the section of
a plane through the generator, the plaue cubic must pass through each
oi these points, i. e. it meets the generator in at least two points there;
but the generator at this point crosses from one sheet to the other, say
from the upper to the lower, while the plane cubic crosses from the lower
to the upper; therefore they cross each other and intersect in an odd
number of points, and consequently they intersect in three points, i. e.
the generator is the real inflectional tangent.

Every plane quartic has three cusps, one at each of the three points
where it- plane meets /-'. and Bince it is the complete intersection of its
plane with the developable, formula <1) holds for it (Theorem II).

Every conic is cut ont by a plane through two generators and every

plane cubic is cut oul by a plane through one generator, and therefore

formula (1) holds for every conic and for every plane cubic (Theorem

III). Formula (1) holds, therefore, for every plane curve on the

lopable

i Imon'i i leom. of Three I limeniioi
« V .lanut. C R., 1886 Vol 100 pp I

WILLIASIS. — GEOMETRY ON RULED QUARTIC SURFACES. 49

4. Twisted Curves. — It may be shown in exactly the same way as
for the Quartic Scroll S (322, 1), that formula (1) holds for every twisted
curve on the developable.

Formula (1) holds, therefore, for every curve on the developable.

X. Quartic Cones.

1. Using the plane quartic curves as a base, we may say that there
are ten different species of quartic cones, corresponding to the ten species
of plane quartic curves.* But for our present purpose we need not
distinguish between double edges and cuspidal edges, and it will be
convenient to divide the cones into five groups, viz., (I) cones hav-
ing a triple edge, (II) cones having three edges double or cuspidal,
(III) cones having two edges double or cuspidal, (IV) cones with one
double or cuspidal edge, and (V) non-singular cones, i. e. cones with no
multiple edges. f

A curve C("\ lying on a cone, has a £-tuple point at the vertex, where
0 < k, and any plane through the vertex meets C(") in k points there.

2. Theorem I. — Every edge of a quartic cone meets C"(") in the same
number of points, say a points (in addition to the k points at the vertex) ;
the number of points of C(") on a double edge is 2 a, and the number of
points of C(") on a triple edge is 3 a.

Proof. — Consideriug group (I), if we pass a plane through the triple
edge and revolve it about this triple edge, it will cut out, in succession,
all the edges of the cone ; the plane meets C(") in k points at the ver-
tex and in a fixed number of points, say t points, on the triple edge, and,
as it meets C(") altogether in a points, the same number of points, say a
points, lie on each edge of the cone. An arbitrary plane through the
vertex meets the cone in four edges, and therefore

4 a + k = a, i. e. 4 a = a — k.

Taking the plane through the triple line, we have

a + t = a — k = 4 a, i. e. r = 3 a.

Consider now group (II), and pass a quadric cone through the three
double or cuspidal edges and any chosen edge A of the quartic cone ;
these count for seven edges in the intersection of the two cones, and

* Salmon's Higher Plane Curves, § 243.

t When not qualified by the words double, cuspidal, triple, multiple, etc., the
word edge will always mean a simple ed<je of the cone.

VOL. XXXVI. — 4

.10 PROCEEDINGS OF TITE AMERICAN ACADEMY.

they must have one more edge in common; if, then, we vary continu-
ously the fifth arbitrary edge that determines the qnadric cone, this cone
will always pass through the three doable or cuspidal edges and the edge
A of the quartic cone and will cut out all the edges of the quartic cone
at a tin..'. Every quadric cone meets < '' in 2a points, of which a
ilrtinite number lie on the three double or cuspidal edges and the edge A,
and therefore the same number of points of ( .'l ", say u points, lie on
each edge oi the quartic cone; it is evident that, if any other edge lie
chosen through which the quadric cone is always to pass, the edge A will
have a points of Cy(") on it. If we pass an arbitrary plane through any
of thi' double or cuspidal edges, it will also cut out two edges of the
quartic cone. and. if 8 be the number of points of C'O on this double or
cuspidal edge, we have

8 + 2a=a-i- = 4o, i. e. 8 = 2 a,

since an arbitrary plane through the vertex cuts out four edges, giving
4 a = a — k.

To prove the theorem for group (III), we may employ a method
analogous to that used for the Quartic Scroll S(l„, lg, 4), p. 37; corre-
sponding to the planes there, through the double directors, we Bhould
use here the planes through the double or cuspidal edges, and instead of
the quadric used there we should now employ a quadric cone.

We shall not consider this proof in detail, but shall now give a proof
of the theorem for groups (IV) and (V), — a proof which holds for the
other three groups as well, and therefore for all quartic cones.

If p is the number of points of C<") on any edge of the c , it is evi-
dent thai 0 < p < a — 1, and p can have, at, most, a different values,
For convenience, we shall say that all the edges for which p has the
same value belong to the same set ; and if there is an infinite number of
edges in :i set. we shall call it an infinite Bet, or, if there is a finite num-
ber, a finite set. Then- are, at most, '/ different B6ts, and, BlUCO " is

finite, at least one of these must be an infinite Bet Let us Buppose, first,
that there i.-, only one infinite Bet, and let a be the value of p for this
infinite set ; we can then pass a plane through one of these a-edges
i i. e. an edge having <i points of < ' i on it i and turn it about so that it

shall not contain any edge of any finite set ; this plane will then cut out

four dredges, and as it meets C'|,n in a points, of which /. lie at the ver-
tex, we have

La + k = ". i- 8. " — k ss 1 a.

i plant through any edg< of any ol the finite sets and

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 51

turn it about so that it will not contain any other edge of any finite set,
and this plane will then cut out three u-edges in addition to the chosen
edge ; if /3 is the number of points of C{a) on the chosen edge, we have
^3 4- 3 a = a — k = 4 a, i. e. (3 = a.

Therefore every edge of the cone has a points of C{a) on it. If the cone
has a double or cuspidal edge, an arbitrary plane through it will cut out
two a-edges, and if 8 is the number of points of C{a) on the double or
cuspidal edge

8 + 2 a = a — k = 4 a, i. e. 8 = 2 a.

If the cone has a triple edge, having r points of Cia) on it, an arbitrary
plane through this triple edge will cut out one a-edge, and we have
t + a = a — k = 4 a, i. e. t = 3 a.

Therefore, the theorem holds when there is only one infinite set. Sup-
pose now that any number of the a sets are infinite, and let a be the
least value of p belonging to any of these infinite sets. Pass a cubic
cone through nine of these a-edges ; then, since the cubic cone meets
C{a) in 3 a points, of which 3 k lie at the vertex and 9 a on the nine
chosen edges, the remaining 3 a — 3 k — 9 a points lie on the three
other edges in which the cubic and quartic cones intersect, and at least
one of these three edges must have as many as a — k — 3 a points of
C'-a) on it. Keeping seven of the a-edges fixed, we can vary the other
two in such a way that the cubic cone, determined each time by the nine
a-edges, will have for its remaining intersection with the quartic cone
three edges different from those of any other such cone previously deter-
mined ; since there is an infinite number of a-edges, we get, in this way,
an infinite number of such cubic cones, and therefore an infinite number
of edges each having as many as a — k — 3 a points of C{a) on it ; hav-
ing an infinite number of such edges, we can so choose two of tbem that
their plane will not pass through any edge of any finite set, because there
is a finite number of edges in all the finite sets taken together ; besides
the chosen edges this plane will then cut out two edges belonging to the
infinite sets. Now the number of points of C{a) on the two chosen edges
taken together is equal to or greater than 2 a — 2 k — G a, and if /5 and
y be the number of points of C(<t), respectively, on the other two edges
in the plane we must have

2a — 2k— 6a + j3 + y<«- k,
and, since neither (3 nor y can be less than a,

2a — 2k— 6a + 2a< a — k,

52 PBOCEEDINGS OF THE AMERICAN ACADEMY.

i.e. a — k < 4 a.

If, now. we pass «i plane through any edge of any infinite set anil turn
it about BO thai it will not pass through any L'dge of any finite set. it
will cut out three other edges of the Infinite sets. Let pu pa, />,,, and p4
he the uuinher of points of 6'1"1, respectively, on the four edges in this
j ilane ; we have then

Px + Pi + p3 + pt = a — k < 4 a,
and, since neither pu p.,, pg, nor pt can he less than a, each must be equal
to a and a — k = 4 a. Therefore, every edge of every infinite set meets
C1"' in a points, i. e. there is only one infinitive set, and we have proved
that the theorem holds in this case.

3. On the cones, a curve C{a) is, as before, designated by the symbol
aa, but a now means the number of points, other than those at the
vertex, in which an arbitrary edge of the cone meets the curve C"°. We
shall now show that formula (1), (aa, bp) = a ft + b a — 4 a ft, gives,
for the quartic cones, the number of intersections of the two curves,
<ia and bp, exclusive of the number of their intersections at the vertex of
the cone.

We have seen that a — k = 4 a, i. e. k = a — 4 a, where k is the
number of branches of aa through the vertex, i.e. aa has an (a — 1 a )-
tuple point ;it the vertex. Let aa be the complete intersection of A"'
and the quartic cone, and let b$ be an arbitrary curve on the cone; the*
total number of intersections of aa and bp is the number of intersections

of Su) and bp, which is by, and, since a = 4v, b v = — ; now, at the

vertex, £(•> has a point of multiplicity (since aa is the complete

intersection of >'" ' and the cone and has an (a — 4 u)-point at the vertex ).
and bti has a (/> — 4 ft) -tuple point, so that b$ meets SW in

^-^yb_4ft) = "''-(l,e+ba-.iaft)

point- at the vertex . sine' the total number of intersections of bp and

> is - , the number of their intersections exclusive of those at the

l

vertex, i. e. the number of intersections of bp and </„ exclusive of those at

the vertex, is f (afi + &a — l <i ft) — "ft+ ba — A a ft,

which is the number given l>y formula (1).
Since the formula holds when <>., is the complete intersection of the

cone and S'\ by Theorem III, it holds when "., i the partial intersec-

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 53

tion of the cone and S^\ provided we can always cut out aa by an 5M
such that the residual is of order less than a or breaks up into compo-
nents, each of which is of order less than a ; for we know that it holds
for the number of intersections of any edge, a 10, with an arbitrary curve
ba, giving (10, b£) = (3. Each edge is a 10, each double or cuspidal
edo-e is a 20, and each triple edge is a 30, since the edges and multiple
ed"-es meet only at the vertex. Every multiple edge can be cut out by
a plane such that the residual will consist entirely of generators, and
therefore formula (1) holds for every multiple edge (Theorem III).
Since the plane quartic cannot break up without the cone breaking up
(unless it consists entirely of edges or multiple edges), the only plane
curve, besides the edges and multiple edges, that can lie on the cone is a
plane quartic which is the complete intersection of its plane with the
cone, and therefore formula (1) holds for every plane curve.

4. Twisted Curves. — There is no cubic curve on a quartic cone, since
1 < a and 4a< a, and, therefore, 4 < a.

Every twisted quartic is a 4lf since a < -, and, since k = a — 4a = 0,

it does not go through the vertex. Any twisted quartic can be cut out
by a cubic monoid,* for we can pass the monoid through 19 — 4 = 15
arbitrary points, and if we take thirteen of these on the quartic curve,
the monoid will contain it. The node of the monoid is taken at the
vertex of the cone ; each edge of the cone then meets the cubic monoid
twice at the vertex and once on the quartic curve, and cannot meet it
again without lying on it; therefore the monoid cannot meet the cone
in any other curve, and the residual consists entirely of edges or multiple
edijes of the cone.

Every twisted quintic is a5u and has one branch through the vertex,
& = 5 — 4 = 1; it can be cut out by a cubic monoid which can be
passed through 15 arbitrary points; for the quintic meets the monoid
twice at the vertex, and if we pass the monoid through 14 other points of
the quintic, it will contain the quintic. Every edge of the cone meets
the monoid twice at the vertex and once on the quintic curve, and there-
fore the residual consists entirely of edges or multiple edges of the cone.

Every twisted sextic is a 6X, and has two branches through the vertex,
since h = 6 — 4 = 2 ; it can be cut out by a cubic monoid whose node
is at the vertex of the cone ; for the sextic meets the monoid four times

* A monoid of order m is a surface of order m having an (m — l)-tuple poiut.
(Cay ley.)

54 PROCEEDINGS OF THE AMERICAN ACADEMY.

at the vertex, and if we pass tbe monoid through 15 other points of the
Bextic it will contain the sextic. The residual will then consist entirely
of edges or multiple edges of the cone.

In the same way it may be Bhown that all curves of order 7. and all
curves of orders 8, '.', and 10 for which a = 1, can he cut out by a quar-
tic monoid such that the residual will consist of edges or multiple edges
of the cone.

When a > 2 (for which a > 8), or when a > 11. i. e. for all curves
not yet considered, we must take v > 5, where v is the order of the Bur-
of lowest order that cuts out <in such that the residual consists of
••8 or multiple edges of the cone. We shall now show that such a
surface can he found for any oa, and the value of v given in terms of "
and a. Since the residual is to consist entirely of edges or multiple
edges of the cone, an arbitrary edge must meet the required surface in
v — a points at the vertex, i.e. the surface must have a (v — a)-tuple
point at the vertex of the cone. Let 3P _a be the required surface.
"When v > 4 we must take care that w ' does not break up into the

^ v — a *

• juartic cone and a surface which must be an -fl/„_a_4, Le. a surface

which has a {v — a — l)-tuple point at the vertex ; Mj~a_^ can pass
through only

i [(V _ 3) 0 - 2) (c - 1 ) - (v - a - 4) (v - a - 3) (v - a - 2) j - 1

arbitrary points, different from the vertex, and consequently if we make

M _ ( pass through

I £ [(„ _ 3) (v_ 2) (v- 1)- (v- o- 4) (v- a- 3) (v- a- 2)]

arbitrary points, not on the quartic cone, it cannol have the qnartic cone
as a component ; the number oi arbitrary points remaining, to determine
.1/ ' . must be greal enough to make it contain a» . which has a — 4a
branches through the vertex • consequently we have

(3) . . . a v + 1 - (a - 4a)(v - a) < \ [(»< + 1) (v + 2) (v + 8 1

_(v-a)(v-a • l)(v-a + 2) - (. - -1)

v — a-\ h n - a - 3) (v - a - 2)] - 1.
from which we obtain the relation

1 + u a -t 1 a i I - 1 " i -2 -i- + 4 i> — \ a — 3,

i. e.

a (a - 2.. -* I • 1

1) . . i .

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 55

or the next greater integer. When v = 3 or 4 we saw that a = 1, so
that the expression (A) that enters into eq. (3) vanishes for v = 3, as it
should, and is equal to unity for v — 4 ; therefore eq. (4) gives the
correct value of v for all twisted curves on the cone. If then we take
the value of v given by eq. (4), any twisted curve aa can be cut out by an
Ml"la sucn tDafc fche residual will consist entirely of edges or multiple
edges of the cone, and therefore formula (1) holds for all twisted curves
on the quartic cones. In determining the above value of v, no account
was taken of the actual multiple points, not at the vertex, that aa may
have. We have seen (p. 24) that an /rc-tuple point reduces by m — 1 the
number of points of aa through which it is necessary to make M*v2a pass
in order for it to contain aa, if this m-tuple point be taken as one of
them ; therefore, if a (a — 2 a + 4) + 4 = 1 (mod. 4) and aa has a
double point not at the vertex, the value of v may be taken one less
than that given by eq. (4) ; if a (a — 2 a + 4) + 4 = 2 (mod. 4)
and aa has two double points or one triple point, not at the vertex, the
value of v from eq. (4) is reduced by unity, and so on. If aa has four
double points, or two triple points, or two double points and a triple
point, or a double point and a 4-tuple point, or one 5-tuple, not at the
vertex, then the value of v is always one less than that given by eq. (4).
We have shown, then, that formula (1) gives the number of inter-
sections, aside from those at the vertex, of any two curves on any quartic
cone. It is to be observed that any branch of aa through the vertex has
an edge of the cone as its tangent at that point and that one of the two
consecutive points, in which the edge meets the branch there, is one of
the a points of aa that lie on this edge ; if a double or cuspidal edge is
tangent to a branch of aa at the vertex, then two of the 2 a points of
aa that lie on it are consecutive to the vertex ; and if a triple edge is
tangent to a branch of aa at the vertex, then three of the 3 a points of
aa that lie on it are consecutive to the vertex. If, therefore, aa and b$
each have a branch through the vertex tangent to the same edge, i. e.
a branch of aa tangent to a branch of bp at the vertex, then one of these
two intersections of the curves at this point is included in the number of
intersections given by formula (1). In like manner, if aa and bp each
have a branch through the vertex, and these branches have there a com-
mon inflectional or cuspidal tangent or have any peculiar relation to one
another, the excess of the number of intersections that these branches
have there over the number of intersections that two arbitrary branches
would have there is included in the number of intersections given by
formula (1).

56 PROCEEDINGS OF THE AMERICAN ACADEMY.

5. We BhalJ qow consider the twisted quartics more in detail. Every
twisted quartic is a li. and, since it has at least two apparent double
points, the coi i which it liea must have at least two double or cus-
pidal edges, i. e. no twisted quartic can lie on the cones of groups (IV)
and (V); moreover, the cubic monoid, which cuta <>nt the twisted
(juartic, has six * lines on it through the vertex, and these six lines must
couut for eight, the order of the residual intersection of the cone and
monoid, and therefore the cone must have, at least, two double or cus-
pidal edges with which two of the six lines coincide. Through a "quar-
tic of the lirst kind" can be passed an infinity of quadrics, i.e. we can
pass a quadric through a "quartic of the first kind" and any arbitrary
point; let this quadric be passed through the vertex. The "quartic of
the first kind" lias two apparent double points, and the two double
edges, of the cone, ou which they lie, meet the quadric twice on the curve
and once at the vertex, and therefore lie entirely on it; these double
edges are, therefore, the two generators of opposite systems, of the
quadric, through the vertex. The cubic monoid, in this case, breaks tip
into the quadric and a plane through the vertex. The " quartic of the
first kind" may have an actual double point or cusp, in which case the
cone ha- an additional double or cuspidal edge that meets the quadric
only once at this double point or cusp and once at the vertex, and then-
fore does not lie on it. The cone may have a cuspidal edge due to an
apparent cusp on the quartic curve, i. e. if a tangent to the curve pa
through the vertex, the curve when viewed from the vertex appears to
have a cusp on this tangent, which is therefore a cuspidal edge of the
COne (the apparent cusp replaces one of the apparent double points and

the cuspidal edge ia one of the generators of the quadric). If two tan-
gents to tin' quartic curve pass through this vertex, the cone has two
cuspidal edges and the curve ha- two apparent cusps. When the quadric
that cuts out the quartic '_r"es through the vertei the residual consists
entirely of the two double or cuspidal edges on which the apparent

double point- or apparent CUSpS lie, but for every other quadric that

passes through the "quartic of the first kind," the residual i- another
"quartic of the first kind" similar to tin- original quartic; this may be
sliown as follows: sine,, the quadric does not go through the vertex no
edge or multiple can lie on it. and therefore the residual cannot break
up, i.e. it must be a twisted quanic; moreover, everj edge or multiple

\. Dumber "f common i ■' it- superior and inferior cones

Caylcy, < Collected Papi rs, V

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES. 57

edge meets the quadric twice and no more ; then every double edge on
which an apparent double point of the original quartic lies is met by the
original quartic in two distinct points, and therefore the residual quartic
must cross this double edge at the same two points, since there are two
and only two branches of the complete intersection at these two points
where the double edge meets the quadric ; therefore, for every apparent
double point of the original quartic there is an apparent double point of
the residual quartic. The double or cuspidal edge, on which an actual
double point or cusp of the original quartic lies, meets the quadric only
once there; and must meet it at one more point, which is, therefore, a
double point or cusp on the residual quartic. A cuspidal edge due to an
apparent cusp on the original quartic meets the quadric in two consecu-
tive points, and the residual quartic must pass through these two con-
secutive points, i. e. it has this cuspidal edge as a tangent and therefore
the residual quartic also has an apparent cusp at this point. The re-
sidual quartic is therefore a " quartic of the first kind " similar to the
original quartic. Each of these two "quartics of the first kind" consid-
ered as lying on the quadric, meets the generators of each system of the
quadric in two points, and is therefore a 4., on the quadric ; the formula
for the number of intersections of two curves aa and b$ on a quadric
being (rta, bp) = a /3 + b a — 2 a ft, the two quartics in question, con-
sidered as lying on the quadric, intersect in (42, 42) = 8 + 8 — 8 = 8
points ; but on the cone these quartics are 4/s, and, since they do not
go through the vertex, formula (1) gives the total number of their
intersections, so that regarding these quartics as lying on the cone they
intersect in (41? 4t) = 4 -f 4 — 4 = 4 points. This illustrates what
was said (p. 25) in reference to the point of crossing of two curves on a
multiple line not counting as a point of intersection of those curves, con-
sidered as lying on the surface having the multiple line, when those
curves lie on different sheets at this point of crossing. In the present
case we have four such points, two on each of the two double or cus-
pidal edges on which the apparent double points or cusps lie ; these four
points count as points of intersection of the quartics considered as lying
on the quadric, because the quartics lie on the same sheet of the quadric,
but they do not count as points of intersection of the quartics, considered
as lying on the cone, because these curves lie on different sheets of the
cone at these points ; these four points make up the difference in the
number of intersections that the quartics have on the two surfaces.

The "quartic of the first kind" may lie on a cone with a tacnodal
edge formed by the union of two double or cuspidal edges ; the quar-

58 PROCEEDINGS OF TIIE AMERICAN ACADEMY.

tic then has an apparent tac-node equivalent to two apparent double
points.

The quartic th.it lit* s on the cone having a triple edge, 18 a "quartic of
the second kind ; " for, if it were a " quartic of the first kind," we could
pass a quadric through it and through the vertex j the triple line would
then lif on the quadric and be a generator of one system ; the generator
of the other system that passes through the vertex would meet the cone
four dm is at t he vertex and at least once on the curve, and would there-
fore coincide with an edge of the cone; but an edge of the cone meets
the quartic once only, and therefore the quartic would be met by the
generators of one system of the quadric once only, and would not be a
" quartic of the first kind " as supposed.

Every "quartic of the second kind" has three apparent double points
(or cusps), and cannot lie on a quartic cone with fewer than three double
or cuspidal edges; two of these inav unite, forming a tac-nodal edge, or
all three double edges may unite, forming a triple edge; on the tac-nodal
edge the quartic has an apparent tac-node equivalent to two apparent
double points, and on the triple edge the quartic lias an apparent triple
point equivalent to three apparent double points. Through every ''quartic
of the second kind" can be passed one ami only one quadric, and. it' the
quartic lies on a cone with a triple edge, the quadric alwaj a passes through
tin- vertex; for, the triple line meets the quadric three times on the quar-
tic curve, and therefore lies on it, being a generator of one system of the
quadric ; the generator of the other system that passes through the ver-
tex coincides with an edge of the cone, as we have seen, ami therefore
the residua] consists of this edge and the triple ed^e. When the " quar-
tic of the second kind" lies on a com: with three double (or cuspidal)
edges, the quadric cannot go through the vertex (for if it did all tl
double or cuspidal edges would lie on it), and the residual is therefore
another " quartic of the second kind" because it lias an apparent double
point (cusp) on each of the three double (cuspidal) edges; the points of
crossing are the same as those of the original quartic, and the curves lie
on different sheet-, at these points. Now, the generators of the quadric
meet the cone four times, and. considering one system of generators, they

most meet the cone three times on one quartic and ono the other

quartic (since every "quartic of the Beoond kind" on a quadric meets
the generators of one Bystem three times and those of the other system
once i, i. e. mi the quadric, one of the quartics i- a 1 . and the other is a
■1,; therefore, considered as \\\wi on the quadric, these quartics inter

■ in l" points, < 1 .. l,i \ -f 12 — 6 10. Bui on the cone each

WILLIAMS. — GEOMETRY ON RULED QUARTIC SURFACES.

59

quartic is a 4X, and these two quartics intersect in only 4 points,
(4ij 4X) = 4. The six points, two on each double or cuspidal edge,
where the branches of the two curves cross, do not count as points of
intersection of the two quartics considered as lying on the cone, for the
branches lie on different sheets of the cone at these six points, and this
accounts for the difference in the number of intersections that the curves
have on the two surfaces, — a further illustration of the principle already
stated.

The different species of twisted quartics that lie on the different cones
can be tabulated as follows, where 8 is the number of actual double
points, k the number of actual cusps, h the number of apparent double
points, k the number of apparent cusps, and T the number of apparent
triple points.

Quartic Cone.

Quartic Curves.

Double
Edges.

Cus-
pidal
Edges.

Triple
Edges.

First Kind

Second Kind.

5

«

h

k

h

k

T

2

0

0

0

0

2

0

1

1

0

0

0

1

1

0

2

0

0

0

0

2

2

1

0

0

1

2

0

2

1

0

1

0

1

1

1

2

0

1

0

0

2

1

2

0

0

1

1

1

3

0

0

1

0

2

0

3

0

0

0

3

0

0

1

0

2

0

Q
O

0

0

0

1

0

0

1

5. Salmon* has divided twisted quintics into four groups, viz. group
(1), having four apparent double points, group (II), having five appar-

* Geom. of Three Dimensions, § 352.

60 PROCEEDINGS OF HIE AMERICAN ACADEMY.

ent double points, and groups (III) and (IV), having six apparent
double points. Now, from an ordinary point on auy curve of order >//.
the number of apparent double points of the curve is // — m + 2,* where
h is the number of apparent douhle points of the curve from an arbitrary
point; therefore, since every twisted quintic on a quartic cone has one
branch through the vertex, the number of apparent double points from
the vertex, i. e. the number of double edges of the quartic cone, is
// — 3 > 1; moreover, since any twisted quintic can be cut out by a
cubic monoid, the six lines of the monoid that pass through the vertex
must count for seven, the order of the residual, and therefore the cone
has at least one double (or cuspidal) edge with which one of these six
lines coincides; therefore there is no quintic curve on the non-singular
cone. There is a twisted quintic, which is not a special case of any of
the groups given by Salmon, that has only three apparent double points,
but it has an actual triple point, where the three tangents do not lie in
the same plane, and the quartic cone on which it lies has a triple edge
due to this actual triple point. All the quartic cones, except the non-
singular cones, have species of twisted quiutics on them, and these may
be tabulated in the same way as the twisted quartics.

We have seen that any twisted sextic can be cut out by a cubic monoid,
and, since the residual is of order six, the six lines of the monoid that
pass through the vertex may be six edges of the cone, and therefore the
non-singular cone, as well as each of the other cones, may have twist d

sextics on it.

Any twisted curve of order 7 can be cut out by a quartic monoid which
has 12 lines that pass through the vertex, 9 of which may be edges of
the cone, forming the total residual, and therefore a quartic cone of any
group has on it some species of twisted curve of order 7. For a > 8
we have a > 2 for some species of aM and therefore when a > 8, a quar-
tic cone of any group has on it some Bpecies of'/,,.

* Balmon's Geom. of Three Dimensions, § 880, example 2.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 4.— July, 1900.

ON SUPPOSED MEROSTOMATOUS AND OTHER
PALEOZOIC ARTHROPOD TRAILS, WITH NOTES ON

THOSE OF LIMULUS.

By Alpheus S. Packard.

PALEONTOLOGICAL NOTES, NO. VI.

ON SUPPOSED MEROSTOMATOUS AND OTHER PALEO-
ZOIC ARTHROPOD TRAILS, WITH NOTES ON THOSE
OF LIMULUS.

By Alpheus S. Packard.

Presented May 9, 1000. Received May 22, 1900.

Trails or tracks evidently made by paleozoic arthropods have occurred
most abundantly in the Potsdam sandstone (Cambrian) of Canada and
New York, also in the Hudson River or Cincinnati stage of the Ordovi-
cian Period. We have now to add descriptions of tracks of a similar
nature from the Chemung stage (Upper Devonian) and from the Upper
Carboniferous. We will give the name trail to the entire series of foot-
prints, and restrict the word track to the individual footprints. The trails
discovered by Logan, and carefully described by Professor R. Owen,*
were very large, being six inches wide and several feet long, and were
evidently made by some large trilobite (as first suggested by Dana)
witb a caudal spine, as there is a well-marked median furrow. Whether
the trilobite was a Paradoxides or not is uncertain, because the species
of this genus are without a definite caudal spine, such as is to be seen in
Dalmanites and certain other trilobites of a later period than the Cam-
brian. But aside from this the tracks, in sets of seven and eight, seem
most probably to have been made by an arthropod with numerous pairs of
jointed cylindrical legs, such as we now know, through the researches of
Walcott and of Beecher, trilobites possessed. Professor Owen described
six species of the tracks, for which he proposed the generic name of
Protichnites, but we venture to suggest that it is not improbable that
they were all made by a single species of trilobite, as observation has
taught us that Limulus may make tracks of very dissimilar shape. We
would suggest that to trails consisting of sets of several, or as many as

* Journal Geolog. Soc. London, VIII. pp. 199, 214. 1852.

These trails, Protichnites septem-notatus Owen, are figured on a reduced scale in
Dana's Manual of Geology, Fig. 25G ; and P. octonotatus Owen in the new edition
of Bronn's Lethaja geognostica.

64 PROCEEDINGS OF TIIE AMERICAN ACADEMY.

7 or 8, individual tracks, the term Protichnitea be restricted. Logan's
Climactichnitei wilsoni,&\x. and u half inches wide and thirteen feet long,
also from the Potsdam sandstone, which Dana suggested ma\ have been
made by a large trilobite,* seems to be such. There is an interrupted
median furrow : the oblique furrows were probably made by the legs,
and the lateral furrow bounding the track is much like that made by the
cheeks or sides of the head of Limulus.

We now come to similar but less complex tracks described by O. C.
Marsh, t also from the Potsdam sandstone near Port Kent, N. Y. This
trail was about six feet long, and the tracks were separated from each
other "by a space of about one and three-fourths inches, and having an
extreme width between their outer edges of two and a half inches." In
this trail there is no median furrow, no lateral ridge, and here and there
are double tracks, as if they were footprints made by a second anterior
pair of feet. The track has a decided merostomatous appearance, but was
probably made by a trilobite, as there are no Limuloid merostonies known
to have existed in the Cambrian, and the track could scarcely have been
made by an Eurypterid.

The next set of paleozoic trails are those described by Miller and
Dyer J in 1878, in a fine hard shale of the Cincinnati's stage of the Ordo-
viciau Period at Cincinnati. The originals arc in the Museum of Com-
parative Zoology at Cambridge, Mass., and I am indebted to Dr. B. T.
Jackson for the privilege of examining them. One of them is similar to
that described below from Providence, but the trail is twice as wide, and
the tracks not so wide. They were perhaps made by a trilobite; there
is no median furrow, or lateral ridges. §

The trails of Limulus polyphemus. — Figs. 1, 2 (one-third natural
size). The trail made by this merostome has been described and figured
by the late Sir .1. \V. Dawson in the Canadian Naturalist, VII. p. 271.
He li'ives three interesting figures of the trails left by a Bingle small
Limulus four inches wide on a sandy shore. In each of his figures the
median furrow is distinct, but the lateral marks are furrows " with Blight
ridges exterior to these," while my example left only a ridge. The

* Manual of Geology, p 189 i first i edition), Big

t Amor. Journ. Be. and Arts, M.viil. July, 1869 Plate.

I .luiirn. Cincinnati Natural Bistory Bociety, 18'

§ The ir.iil- figured l>y Emmoni (Agriculture of New York, I.) as Iterates jack-
in : appear to be Annelid trails, and nunc of those figured by

James Ball (Paleontology of New York, II. Pis. 18 16) seem t<> li:i\ <■ been
made either by trilobitei <>r merottom

PACKARD. — PALEOZOIC ARTHROPOD TRAILS.

65

footprints were slightly oblique series of four punctures, or pits, "deepest
behind, in which the four marks left by the nails of the posterior feet
were most prominent, and sometimes the only marks seen."

Figure 1.

" When the Limulus creeps on quicksand, or on sand just covered
with water, so that its body is partly water-borne, it appears principally
to use its ordinary walking feet, and the footprints then resolve them-
selves into a series of longitudinal scratches after the manner of Protich-
nites lineatas." . . .

" When placed in shallow water, just covering the body, the creature
u?ed its flat abdominal swimming feet, and though the impression made
was very faint, and not readily observed under water, it was obviously
very different from those before mentioned, agreeing with them only in
the lateral and median grooves, while between these were series of fur-
rows extending ohliquely from each side of the middle groove, and re-
vol. xxxvi. — 5

0(1 PROCEEDINGS OF THE AMERICAN ACADEMY.

aembling ripple marks (Fig. 3). These were produced by the Band
swept up by the Bwimming feet." Dr. Dawson then compares the trail
represented by his Fig. 3 with Logan's Climactichnites, and they are
remarkably similar, except that the oblique furrows made by the l<
between the median' and lateral ridges are directed in the reverse
direction.

Of the tracks afterwards described and figured by Dr. Dawson* from
the upper Carboniferous of Nova Scotia, none seem to me to be referrible
to trilobitic or merostomatous trails. Proticluiites carbonarius, as already
remarked in these Proceedings, f appear to have been made by a crusta-
cean, and we have referred them to a distinct genus, Ostrakirhnitcs.

Some years ago I made some experiments with small Limuli by placing
one in a shallow tin pan, in which the sand was about half an inch deep,
ami the water not deep enough to entirely cover the body. The animal.
so far as I can now remember, used its ambulatory feet in walking, while
the Bwimming or abdominal legs were partially used. The result may
be seen in Fig. 1. The king-cr:il> was about four inches (10 cm.) in
width. The trail it made consisted, besides the tracks themselves, of an
outer ridge made by the outer edge of the head or carapace ; this ridge (d i
was about 15 mm. in width, and was due to the heaping up of the tine
sand; in section it would be low conical; one would suppose that the
action of the ed're of the head would make a furrow rather than an ele-
rated ridge.

The tracks (/) were opposite, and quite regularly concavo-triangular,
the apex of the triangle rounded, and directed backwards, the sand being
pushed slightly up on the posterior edge of the track.

The tracks of each side were directly opposite each other, and those of
each pair directly iu line with those of the pair in front It was notia d
that the distance apart of the tracks varied with tin- rapidity of the half-
walkuur, half-swimming movements of the animal. It was Been that the

tracks were made by the hindermost, or sixth pair, of limbs only, no im-

pn -ion being left in the sand by the feet in front. Tin- triangular

shape of the traek was due to tin- spreading out of the two spatulale

spine-, of the lasl Begmenl of the leg. It should be observed that the
distance apart, outside measurement, of the traek. is about two-thirds
that of the entire trail.

ipressiom 1 footprints of aquatic animals and imitative markinj

carboniferous rocks. Amer. Journ. 8c. and Arts, 3d Series, V. Jan 1873, pp. 16 24,
i: - i . Also Acadian Qeologj 2d edit. Supplement 1*78.

\\V. April, 1900, p W8

PACKARD. — PALEOZOIC ARTHROPOD TRAILS. 67

The four narrow furrows (/) intersecting the tracks were made by the
lateral abdominal spines, which are bent down and trail in the mud or
sand when the animal is walking or moving over the bottom. The large,
deep, median furrow (e) was made by the caudal spine ; it was continuous,
uninterrupted, during the continuous forward movements of the animal.

Another trail is represented by Fig. 2. It will be seen that it pre-
sents no resemblance to the other. Unfortunately I did not make any
notes as to the relations of the animal to the bottom. So far as I can
remember it was a smaller individual, and probably it moved rapidly.
The indentations on the margin of the trail were evidently made by the
feet, while the series of median furrows were made by strokes of the
caudal spine. The trail was 3.50 cm. wide.

The tracks of living terrestrial Isopod Crustaceans. — Fig. 3. In this
connection it was interesting to ascertain the nature of the tracks made
by Crustacea so much like trilobites in general shape as our
terrestrial Isopods. One of our common Armadillo was cap- • .

tured, its feet inked, and it was then set free and allowed to
" make tracks " on a sheet of paper. It will be remembered , \
that the Isopods have seven pairs of ambulatory legs, the . .
extremities of which end in a single sharp point. The width
of the body is from 4 to 5 mm. It was noticed that the
crustacean in running put down the feet of each pair at the , ,

same time, and that the legs, and especially the pointed ex- • \

tremities, were perpendicular to the surface over which it
ran. •

The trail thus made was a very simple one, being a double • •

row of slightly elongated dots, the individual tracks of each
pair being exactly opposite to each other, as seen in the figure. • •

When the feet dragged the dots became lines. The trail is yIGUKE 3.
also nearly of the same width as the body itself, though a
little narrower rather than wider.

A specimen of our common Porcellio scaber was also compelled to
undergo similar treatment and like evolutions with the same result. The
trail differed in no essential respects from that of the other Isopod.

It will be of much interest to experiment with macrurous and brachyu-
rous Crustacea, in order that the results may throw light on the numerous
tracks in the Triassic beds of the Connecticut Valley described by Hitch-
cock in his " Ichnology."

Merostomichnites beecheri n. sp. Fig 4. This trail is in a fine shaly
sandstone of the Chemung beds at Warren, Pa. It covers an area

'i^ PROCEEDINGS OF THE AMERICAN ACADEMY.

3 cm. Ion" and about 9 mm. wide. It is straight and short. The
individual tracks are opposite each other, as in those of Limnlus. The
animal must have leaned more to the right side, as the tracks on this
Bide are larger, deeper, aud much more perfect than on the left side.
The most perfect ones are in shape almost exactly like those made by
Limulus when half walking and half swimming in very shoal water.
A typical track may be described as forming a low triangle, the apex
pointing backward; it is hollow, the interior forming a hollow triangle
surrounded by a raised ridge.

There are six pairs of tracks, with traces of a seventh. The width of
the trail is 9 mm., but probably if the entire trail were perfect on the
left side it would have measured about 10 mm. in width. The tracks are
opposite to each other. The largest and best marked individual track is
5-6 mm. wide and 3 mm. deep (or long) from in front to the apex or
hinder part. The tracks of each pair are very near to each other, and
those in the front part of the trail tend to be united into a simple trans-
verse ridge.

There are no secondary tracks in between the others, and in this respect
the track differs from that of M. narragansettensis, and resembles that of
Limulus.

The original of this track is in the paleontological collection of the
Peabody Museum of Yale University. I am indebted to the kindness of
Professor C. E. Beecher, the curator, who did me the favor to scud me
an excellent cast, from which the above description has been made.
Professor Beecher informs me that in the beds of the Chemung group at
Warren, Pa., there are no remains of trilobites ; and he expressed his
belief that the tracks were those of some merostome. It is to be observed
that there is no median furrow or trail made by a tail or caudal spine,
and no furrow or ridge made by the edge of the body or carapace. The
merostome which mail'- it mosl probably had no caudal spine. And yet
iln' Hacks are Otherwise very similar to those of Limulus. It may be
observed that all tin- Eurypterida are provided with a telson, either broad
or narrow and spine-like, and their trails would evidently include a furrow
mad'- by such a spine. The I.iniulidae were represented in the Devonian
by Protolimuliu erientis Williams ("associated with typical Chemung
fossils"). Was this track made by a young individual of this genus,

with the Caudal spine too short to make a trail ? If not. was it made by

a Bunodes, which lived in the Upper Silurian of Europe, but has not yel
been discovered in America, and had no caudal Bpine? But Bunodes

would perhaps have left a lateral furrow made by the broad thin edge of

PACKARD. — PALEOZOIC ARTHROPOD TRAILS.

69

Figure 4.

Figure 5.

70 PROCEEDINGS OF THE AMERICAN ACADEMY.

its elliptical body. The trail seems to indicate the occurrence of a mero-
stome of a group not represented by any known fossil genus, unless it
Bhould prove to be a young Protolimulus.

Merostomichnitei narragansettensis (Pack.). Fig. 5. Proc. Araer.
A i. Arts and Sci. XXXV. No. 20, April. 1900, p. 402. Three
trails of this species were discovered by a member of my geological class,
— Mr. II. II. Mason, of the class of 1900, — and kindly given to me, while
doing field work under my direction in Providence, R. I., just north of
the city and of the North Burying Ground. The trails occurred in a
rounded pebble of dark arenaceous shale picked up from the mass of
water-worn gravel and boulders constituting the body of a large rounded
kame. It was split in two, so as to show the impressions and the relief
of the tracks. The subglacial deposits at this point are derived from a
region a few miles a little east of north, probably in the vicinity of South
Attleboro, though I have not seen beds of this peculiar blackish sandy
shale in place.

There are three trails, — one separate, and the two others crossing
each other. The long separate series is nearly straight, 6.50 cm. in
length ; the other trail is sinuous, and is crossed by a third, shorter trail.
The width of each trail is the same, being about 12 mm. outside meas-
urement. The distance between the individual tracks of the same pair
is about 9 mm., that between the footprints on the same side varies from
about 2 mm. to about 4 mm.

The tracks are opposite and not alternate. Along most of the length
of the trail there is but a single series of tracks on a side, but in portions
of the entire trail the footprints are double, there being an inner and an
outer set on each side. If we select four of the tracks in a place where
they are double, it will be seen thai they are arranged in a very low trap-
ezoid ; while the Bpace between the two outside tracks is 9 nun., that
between the two inner, a little in advance, is o nun.

The individual tracks forming the trails are of quite uniform shape, the
best marked ones being transversely oval or crescentie in outline, the
mud having been pushed back by the animal's feet so as to leave a cres-
centie ridge, the concavity pointing forward. The m/.c .it' the impression
in tran diameter is about 3 mm., the longitudinal diameter L.5,

i. e. the tracks are about twice as broad as long. Tims they are not
linear, and more or less parallel to the main series of tracks, as in those

referred to the decapod Crustacea. In two or three cases the tracks are

Connected by B slight ridge curved forward in the direction the animal

mo\ ed.

PACKARD. — PALEOZOIC ARTHROPOD TRAILS. 71

While these tracks are certainly not Isopod tracks, they with still more
certainty cannot be referred to impressions made by the feet of insects,
which are always alternate, as in hexapod insects the legs of each pair
arc raised and put down alternately. The Providence carboniferous
tracks were evidently made by an arthropod of the same group as
Limulus, as the tracks are opposite, and in general shape like those of
Limulus, as may be seen by Fig. 1. The present tracks differ from
those of Limulus in the absence of a caudal spine trail-mark, and in the
fact that an additional anterior pair of feet made impressions. Trilo-
bites are not known to exist in our Upper Carboniferous associated with
Limulus. Lirnuloids of the genus Prestwichia with a short caudal spine
exist in considerable numbers in the Upper Carboniferous of Mason
Creek, Illinois, and in the Upper Carboniferous beds of Pennsylvania,
though from a horizon higher than that of the Mason Creek beds. As
the genus Belinurus has a very long caudal spine, and there are uo traces
of a median furrow in our trails, that genus should be ruled out as the
author of these tracks. Now the adult Prestwichia is about two inches
in diameter and its caudal spine nearly half an inch in length. It is,
however, well known that in the freshly hatched Limulus, and even after
the first moult, and when the creatures are half an inch in diameter, the
caudal spine is very short, too much so, probably, to leave a trail or
median furrow.

The Providence trails are considerably less than half an inch in diam-
eter, and reasoning by analogy, and also in part by exclusion, it seems
not impossible that these trails are the footprints of a young Prestwichia
a little less than half an inch in diameter, with too short a caudal spine
to leave a furrow.

This conclusion is interesting as suggesting the occurrence of these
Lirnuloids in the Narragansett basin during the later part of the Carbon-
iferous Period. These tracks are so similar to those of the Chemung beds
above described that they were probably made by Merostomes of the
same family or genus, and may be referred to Merostomichnites rather
than to Protichnites.

The possibility that these trails could have been made by an Eurypterid
seems excluded by the absence of a median furrow, or of any prints made
by the large paddles of the hind feet, or by the paddles and chelae of the
first pair of feet of such a form as Pterygotus.

Brown University.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 5. —July, 1900.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

CERTAIN
DERIVATIVES OF METADIBROMD1NITROBENZOL.

By C. Loring Jackson and W. P. Cohoe.

, CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF

HARVARD COLLEGE.

CERTAIN DERIVATIVES OF METADIBROMDINITRO-

BENZOL.

By C. Loring Jackson and W P. Cohoe.

Presented December 13, 1899. Received June 4, 1900.

The dibromdinitrobenzol melting at 117°. 4 was made by Korner in
his classical research on Isomerism of the Aromatic Compounds with Six
Atoms of Carbon,* but' its constitution has not been determined with
certainty. Nietzki and Schedler f in a recent paper have proved that the
dichlordinitrobenzol melting at 103° has the structure Cl2 1.3.(N02)24. 6,
and as this substance is made from the action of fuming nitric acid on
metadichlorbenzol, just as the dibrom compound is made from metadi-
brombenzol, there is good reason to believe that they have the same
constitution ; but we thought it necessary to prove that this was the case
before we studied this dibromdinitrobenzol further. For this purpose we
heated the substance with aniline, and obtained the dianilidodinitrobenzol
melting at 186°, which Nietzki and Schedler had obtained from their
dichlordinitrobenzol ; the dibrom substance therefore has the structure
Br2 1. 3. (N02)24. 6. If aniline acted on the dibromdinitrobenzol in the
cold, a bromauilidodinitrobenzol, CGH2BrC6H5NH(NOo)o, melting at
157° was obtained, and this occurs in a yellow and in a red modification
which seem to differ in crystalline form as well as in color. This seems
to be a case of dimorphism, as the two forms pass into each other with
great ease. The yellow form is converted into the red by crystallization
from benzol, the red into the yellow by heating to 135°. Similar phe-
nomena have been observed in the cases of anilidotrinitrophenyltartronic
ester t and triauilidodinitrobenzol. §

* Gazz. Chim. 1874, 305.

t Ber. (1. chem. Ges., XXX. 1666 (1897).

t Jackson and Bentley, These Proceedings, XXVI. 83.

§ Jackson and Herman, Ibid. XXVII. 253.

7G PROCEEDINGS OF THE AMERICAN ACADEMY.

After the constitution of the 1.3.4.6 dibromdinitrobenzol had been
determined, we studied the action of sodic ethylate upon it, in the hope
of encountering the replacement of bromine by hydrogen under the action
of tins reagent, which has been studied for some years in this Labora-
tory. But no such behavior was observed; the action ran in the normal
way. resulting in the formation of the dinitroresorcine diethylether melt-
ing at 133°, discovered by Warren and one of u>* and proved to have
the symmetrical structure by Koch and one of usf; so thai this observa-
tion confirms the determination of the constitution of this body by the
action with aniline.

Sodic phenylate converted the dibromdinitrobenzol into diphenoxydi-
nitrobenzol, which melts at 129°. The action of sodic malonic ester was
tried on this compound to determine whether the phenoxy groups could
be replaced by the malonic ester radical CII(COOC.II5)o ; as it has
been found | that the best way to make dichlordimalonicesterquinone
CcCl..[CII(COOCoII- i..].jOj is by treating dichlordiphenoxy<piinone with
sodic malonic ester, and we wished to see whether the nitro groups
would produce the same effect as the oxygen atoms of the quinone in
promoting the replacement of the phenoxy groups by malonic e
radicals. We found that phenol was eliminated when the diphenoxy-
dinitrobenzol was treated with sodic malonic ester, and from the very
unmanageable product a sodium salt was obtained which contained an
amount of sodium corresponding to CcH.,OC',.l 1 5CNa( ( < >< K 'I I- |s| N( )..)., ;
so that we feel justified in assuming that the reaction has consisted in
the replacement of the phenoxy by the malonic ester group.

The dibromdinitrobenzol, when reduced with zinc dust and acetic acid,
gave a dibrommetaphenylene diamine, which was identical with that
melting at 135° obtained by the action of bromine on metaphenylene
diacetamid. § It follows from our preparation of this base that it has
the constitution Br2 1. 3. (NH2)8 4. 6.

In all tliis work we have been careful not to approach too near to the
field reserved by Nietzki and Schedler in their paper on dichlordinitro-
benzol.

* These Proceedings, XXV. 170.

t [bid. XXX IV. |:.l

I Jackson and Grindley, [bid. XXX l"Y

§ Jackson and Calvert, [bid. XXXI. L60.

JACKSON AND COHOE. — METADIBROMDINITROBENZOL DERIVATIVES. 77

Preparation of Metadibromdinitrobenzol.

Monobromacetanilid was first made by passing a stream of air laden
with the vapor of bromine through five litres of water in which fifty grams
of acetanilid were suspended. The end of the reaction was determined
by the appearance of a lasting yellow color in the liquid and a distinct
change in the appearance of the solid. In all our attempts to convert
this product into the dibromacetanilid by the action of liquid bromine on
it when suspended in water, we observed the formation of a considerable
amount of symmetrical tribromaniline (NH2 1. Br3 2. 4. 6.), produced un-
doubtedly by the action of bromine on the free base proceeding from the
saponification of some of the bromacetanilid by the hydrobromic acid
formed in the reaction. The monobromacetanilid was accordingly filtered
out, and after being dried, suspended in glacial acetic acid or chloroform,
to which a little more than the calculated amount of bromine was then
added drop by drop. This method gave an excellent yield of the
dibromacetanilid with little trouble.

The dibromacetanilid was saponified by boiling it with sulphuric acid
of specific gravity 1.44 in a flask with a return condenser, until the solu-
tion of the solid, which was usually accompanied by a darkening in color,
showed that the reaction was complete. The liquid was then allowed to
cool, when most of the dibromaniline crystallized out, and what remained
in solution was precipitated by the addition of a large quantity of water.

To remove the amido group, 100 grams of the dibromaniline were
dissolved in a mixture of 300 c.c. of alcohol and 120 c.c. of benzol, to
which 20 c.c. of sulphuric acid had been added ; the mixture was heated
in a flask on the steam bath, and 60 grams (a large excess) of solid sodic
nitrite added as fast as the reaction would permit. The contents of the
flask, which had taken on a reddish color, were boiled for an hour, and
then allowed to stand over night, after which a large quantity of water
was added, and the precipitated oily liquid distilled over with steam. A
small amount of tribrombenzol is usually present, which appears as a
solid in the condenser toward the end of the distillation with steam ; it is
well to stop the distillation as soon as this solid appears. The distillate
with steam was dried with calcic chloride and distilled, collecting for use
the fraction boiling from 210° to 225°. A small additional amount of
dibrombenzol was obtained by extracting with ether the aqueous portion
of the steam distillate, but the increase of the yield in this way was so
small that this extraction was hardly worth while.

The dibrombenzol was converted into dinitrodibrombenzol by boiling

78 PROCEEDINGS OF THE AMERICAN ACADEMY.

it with fuming nitric acid of specific gravity 1.52, or better a mixture of
this acid and sulphuric acid. As soon as the oil had dissolved in the
arid, the reaction was complete; the boiling was stopped, and the prod-
uct precipitated by pouring the acid liquid into a large quantity of cold
water. The yellow crystals thus obtained were purified by crystallization
from alcohol, until they showed the constant melting point 117°.

Determination op the Constitution of Metadibromdinitro-
benzol bl* the action of aniline.

Two grams of the dibromdinitrobenzol melting at 117° were heated
with an excess of aniline on the steam bath for half an hour. At the
end of this time the excess of aniline was removed by treatment with
dilute hydrochloric acid, and the yellow residue, after being washed thor-
oughly with water, was recrystallized from a mixture of alcohol and
benzol until it showed the constant melting point 185°, which is essen-
tially identical with 186°, that of the dianilidodinitrobeuzol prepared
from dichlordinitrobenzol by Nietzki and Schedler.* As they established
the constitution (NHC6H6)a 1. 3. (N02)2 4. 6 for this body, the dibrom-
dinitrobenzol belting at 117° must have the corresponding structure
Br21.3. (N02)24. 6.

Action of Aniline on Symmetrical Dibromdinitrobenzo'. in
the Cold : Bromanilidodinitrobenzol.

Two grams of the dibromdinitrobenzol cooled by a freezing mixture
were moistened with aniline, and the mixture was allowed to Btand
packed in ice for twelve to eighteen hours. The excess of aniline was
then removed by dilute hydrochloric acid, and the dense yellow residue,
after thorough washing with water, was purified by crystallization from
alcohol and benzol until it showed the constant melting point of \57°,
when it was dried at 100 , and analyzed, with the following result: —

I. 0.0994 gram of the substance gave by the method of Carius 0.0558
•_rram of argentic bromide.
II. 0.2138 gram of the substance gave 0.1229 gram of argentic bromide.

r.romine

Calculated f"r

Cll.Jlr' II Ml NO

T.

Found.

11.

23.67

23.s:.

24.46

lk-r. (1. chem. (its., XXX. L668 (1897).

JACKSON AND COHOE. — METADIBROMDINITROBENZOL DERIVATIVES. 79

Properties of Bromanilidodinitrobenzol,
CGH2BrC6H5NH(N02)2(Br 1. C6H5NH 3. (N02)2 4. 6.).

This substance crystallizes from a mixture of benzol and alcohol,
sometimes in much modified short and thick flat prisms apparently of the
monocliuic system, resembling certain crystals of felspar ; at other times
in rhombic plates occasionally with some of the angles slightly bevelled,
which pass into forms like rhombohedra ; all of these crystals have a full
golden yellow color. Other crystals were distinguished from these by
their brilliant red color and by appearing in square prisms modified on
one of their angles or on two opposite angles. From these two colors,
red and yellow, and the difference at least in crystalline habit, if not
really in crystalline form, we infer that the substance appears in two
closely related modifications. The yellow form, in which the substance
is obtained from its preparation, becomes gradually converted into the
red by crystallization from a mixture of alcohol and benzol, but several
crystallizations are necessary to make this conversion complete. If,
however, the yellow form is dissolved in benzol alone, the solution
deposits, as it evaporates, a red oil, which on stirring solidifies to the red
crystals. All the other common organic solvents except the alcohols
also convert the yellow into the red form by a single crystallization. On
the other hand, if the red crystals are heated in an air bath, they begin
to assume a yellow color at 110°, and are completely converted into the
yellow form at 185° ; this conversion is attended with no change in
weight. The substance melts at 157°, forming a red liquid, which, when
cooled and stirred, solidifies to the yellow modification. It is obvious
that the melting point given above is that of the yellow form, since the
red is converted into the yellow at 135°. The bromanilidodinitrobenzol
is freely soluble in benzol; soluble in toluol or chloroform ; slightly sol-
uble in ether, acetone, amyl alcohol, or hot ethyl alcohol. The best
solvent for it is a mixture of benzul and alcohol.

Determination of the Constitution of Metadibromdinitro-
benzol by treatment with sodic ethylate.

One gram of the dibromdiuitrobenzol melting at 117° was mixed with
an excess of sodic ethylate prepared by the action of sodium on a consid-
erable excess of absolute alcohol, and after the mixture had stood for
some time at ordinary temperatures, the excess of alcohol was allowed
to evaporate spontaneously. The residue, after repeated washings with

80 PROCEEDINGS OF THE AMERICAN ACADEMY.

hot water, was purified by crystallization from a mixture of ligroin and
benzol, when it was found to melt at 133°, the melting point of the
dinitroresorcine diethyletber discovered by Warren and one of us,*
which it also resembled in crystalline form. As Koch and one of usf
have proved that this dinitroresorcine diethyletber bus the constitution
(OCII3)2 1.3. (N02)o 1. 63 it follows that the dibroindinitrobenzol melt-
ing at 117° has the constitution Br., 1. 3. (NO.,)., 4. G, — a result which
confirms that already obtained from the action of aniline.

Action of Sodic Piienylate with: Symmetrical DIBROMDINITRO-
BENZOL.

The sodic phenylate for this experiment was prepared by adding five
grams of powdered sodic hydrate to enough melted phenol to make a
pasty mass. While this was still in the viscous state live grams of the
dibromdinitrobenzol were added little by little with constant stirring, the
beaker containing the mixture being kept cool with ice water. The
product, a dark brown solid mass, was treated with water, filtered, and
the insoluble portion recrystallized either from slightly dilute alcohol, or
from a mixture of benzol with a large excess of ligroin, until it showed
the constant melting point 129°, when it was dried in vacuo, and ana-
lyzed, with the following results: —

I. 0.2494 gram of the substance gave on combustion 0.5563 gram of

carbonic dioxide and 0.0903 gram of water.
II. 0.37 46 gram gave on combustion 0.8457 gram of carbonic dioxide.
The water was unfortunately lost.

-

i'.i], ui:it. ■! fur

-ii 01 ii NO,),.

Fou
I.

nd.

11.

Carbon

61.88

60.82

61.59

Hydrogen

3.41

L02

The method of making sodic phenylate Ldven above was adopted
because in alcoholic solution the sodic phenylate gave principally the
dinitroresorcine diethylether, and in aqueous solution there was uo action
until hieh temperatures were reached, and then the yield was unsatis-

O 1

factory.

I k . Proceedings, XXV. 170.
t Ibid., XXXIV. L84

JACKSON AND COHOE. — METADIBROMDINITROBENZOL DERIVATIVES. 81

Properties of the Dinitroresorcine Diphenylether,
C6H2(OC6Hs)2(N02)2 . ((OC6H5)2 1. 3. (N02)2 4. 6.).

This substance, when crystallized from a mixture of benzol and alco-
hol, forms slender white prisms terminated by one slanting plane, or,
when better developed, by two planes meeting at a very obtuse angle
and modified by several smaller ones, so that a blunt end is formed. It
melts at 129°. It is freely soluble in benzol; soluble in cold toluol or
chloroform, or in alcohol, ligroin, or acetone when these solvents are hot;
slightly soluble in ether or amyl alcohol. The best solvent for it was
either dilute alcohol, or ligroin containing a little benzol.

Some experiments were tried to determine whether the phenoxy
groups in the diphenoxydiuitrobenzol could be replaced by malonic ester
radicals (CH(COOC2H5)2), as a replacement of this sort had been ob-
served in the case of dichlordiphenoxyquinone.* The solid diphenoxy-
dinitrohenzol was dissolved in an alcoholic solution of sodic malonic ester,
and the mixture allowed to stand at ordinary temperatures over night ;
afterward it was treated with dilute sulphuric acid, which precipitated
an oil, and showed that phenol had been set free by the strong smell of
this substance. The oily precipitate, after washing with water, was
extracted with ether, which removed from it a pale yellow oily substance ;
but as this did not solidify even after standing for several months, we
tried to determine its nature by the analysis of a salt. A specimen of
the oil was dissolved in alcohol and treated with an aqueous solution of
sodic hydrate, which threw down a bright red precipitate ; this was
washed with benzol, dried in vacuo, and analyzed, with the following
results : —

I. 0.1383 gram of the substance gave 0.0191 gram of sodic sulphate.
II. 0.1476 gram of the substance gave 0.0228 gram of sodic sulphate.

Calculated for Found.

C6Hj(N02)2OC6H5CNa(COOC2II-),. I. II.

Sodium 5.23 4.47 5.01

The results of these analyses, in connection with the elimination of
phenol, prove that the reaction has run as was expected. The phenoxy-
dinitrophenylmalonic ester separated from the salt as an oil, and there-
fore we did not think it worth while to attempt a more careful study
of it.

* Jackson and Grindley, These Proceedings, XXX. 425.

VOL. XXXVI. — 0

82 PROCEEDINGS OF THE AMERICAN ACADEMY.

Reduction of Symmetrical Dibromdinitrobenzol avitii Zinc

Dust and Acetic Acid.

Twenty grams of zinc dust were placed in a flask iitted with a Bunsen
valve and connected with a carbonic dioxide generator, acetic acid of
eighty -five per cent was added, and then five grains of the dibrouidinitro-
beuzol in small portions at a time, as under these conditions the reaction
ran quietly and smoothly, although accompanied by blackening in every
case. The reduction was carried on at first in the cold, but toward the
end of the operation the mixture was heated gently on the steam bath.
After two hours the reaction was complete, when the insoluble portion
was filtered out, and extracted with dilute alcohol, which upon cooling
deposited crystals melting even in the crude state at 131°. The filtrate
from the insoluble reduction products was treated with an excess of sodic
hydrate sufficient to dissolve the zinc salts, the precipitate formed in this
way filtered out, and extracted with hot dilute alcohol, which gave an-
other portion of the crude product melting at 131°. After purification
by crystallization from alcohol the melting point became constant at
134° ; it was dried in vacuo, and analyzed with the following result: —

0.1407 gram of the substance gave by the method of Carius 0.1976 gram
of argentic bromide.

b*

/.•!,/T.1:l,'vn,r Found.

Bromine G0.15 59.78

This dibromphenylene diamine melts essentially at the same point
(135 ) as thai prepared by S. Calvert and one of us* from metapheny-
lene diacetamide and bromine. It also crystallizes like this in white
needles, which turn brown on exposure to the air. As the constitution
of our dibrombinitrobenzol is Brs 1. 8. I NO,), 1. G, it follows that this
dibrommetaphenylene diamine must have the corresponding constitution
I, i. .;. -Ml l ,4.6.

* These Proceedings, XXXI. 150.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. G. — August, 1900.

ON THE CONTINUITY OF GROUPS GENERATED BY
INFINITESIMAL TRANSFORM A TIONS.

By Stepiikn Ei,mek Slocum,

ON THE CONTINUITY OF GROUPS GENERATED BY
INFINITESIMAL TRANSFORMATIONS.

By Stephen Elmer Slocum,
Fellow in Mathematics, Clark University, Worcester, Mass.

Presented by Henry Taber, May 9, 1900. Received June 20, 1900.

§ 1.

The publication of the results of Professor Sophus Lie's investigations
in the theory of finite continuous groups, embodied in papers appearing
from 1870 to 1898, chiefly in the " Archiv for Mathematik og Naturvi-
denskab" and the " Forhandlingar i Videnskabs Selskabet i Christiania,"
and the systematic presentation of his theory in six large volumes, pub-
lished during the years 1888-1893, opened to mathematicians a new and
exceedingly rich field of investigation. As a creator and pioneer in this
field, Professor Lie's aim was to outline his theory as broadly as possible,
not stopping to obtain entirely rigorous demonstrations of his theorems ;
and it is not surprising to find that certain of these theorems, and of the
fundamental conceptions of his theory, require modification. Thus it
appears from a discovery of Study's, mentioned below, that the chief
theorem of Lie's theory holds, in general, only in the neighborhood of the
identical transformation, and, as a consequence of this fact, that the con-
ception of isomorphism, as developed by Lie, requires modification.

The chief theorem of Lie's theory is that r independent infinitesimal
transformations,* whose symbols are

n g

At- — ift cik \X\ . . . xn) -=

i dxk

[i = 1, 2 . . . r)

(where the £'s are analytic functions of n independent variables xl . . . xn)
generate an r-parameter (r-gliedrige) group, in which each transformation

* Lie terms the symbols of infinitesimal transformation Xx . . . Xr independent
if the |'s satisfy no linear homogeneous relations of the form

«i fi< (*)+...+ er fir (x) = 0
simultaneously for i — 1, 2 . . . n, with coefficients e independent of the x's.

86 PROCEEDINGS OF THE AMERICAN ACADEMY.

is generated by an infinitesimal transformation of the group, if and only
if the A'i . . . A'r fulfil relations of the form

r
(Xj, -Xj) = -sCjks X,

(./,*• = 1,2 . . . r),

where (Xj, Xk) denotes the alternant A", Xk — Xk Xj, and the coefficients
cib are quantities independent of the ar's.*

In Volume XXXV. of the " Proceedings of the American Academy of
Arts and Sciences," pp. 239 et seq., I pointed out an error in the demon-
stration of what Lie calls the first fundamental theorem,! upon which
he bases the demonstration of his chief theorem. This error consists in
neglecting conditions imposed at the outset upon certain auxiliary quan-
tities fxi, ju2 • • • {*ri introduced iu the course of the demonstration. Thus
in the " Continuierliche Gruppen," pp. 372-37G (and substantially in
" Transformationsgruppen," III. pp. 558-564), Lie proceeds as follows :
Being ffiven at the outset a family with an oor of transformations Ta, de-
fined by the equations

% i = J, (xli • • • xmali ' ' • ar)
(i = 1, 2 . . . n),

containing the identical transformation, rind such, moreover, that the a/'s
satisfy a certain system of differential equations, he defines by the intro-
duction of new parameters /x a family of transformations EM

x'i = Ft(p?lt . • ■ Xn, fAl} . . . fr)

(, = 1,2. . . n),

each of which is generated by an infinitesimal transformation. Lie then
establishes the symbolic equation

To E^ = T .X

* Transformationsgruppen, III. 690; Continuierliche Gruppen, 211, 305, 390.
t Transformationsgruppen, III •">»;;; Continuierliche Gruppen, 376.
i If the equations defining the families >'t" transformations /'.. and A'M are.
i, spectively,

(.■ = 1.2 . . »),

and

.r't = /•>-', . . . rn, a-, . . . ^,)

(, - 1,2, . »),

the symbolic equation /"„ / „ T is equivalent to the simultaneous system of
equations

SLOCUM. — FINITE CONTINUOUS GROUPS. 87

where the a's and /a's are arbitrary, and

ak = ®k (Pi • • • Pn «1 • • • «r)

(k = 1, 2 . . . r),
the <£'s being independent functions of the /a's. For

(0)

ak = ak

(k - 1, 2 . . . /),

the transformation Ta becomes the identical transformation ; and there-
fore we have

where

<** = <p* (/ai . . . /Ar, «i . . . or ;

(A: = 1,2 . . . r).

Thus every transformation of the family E^ is a transformation of the
family Ta. If, conversely, we could show that, for arbitrary values of
the a's, every transformation Ta belonged to the family E^, it would
follow that

that is to say, we should then have shown that the family of transforma-
tions Ta forms a group.

But, although the <£'s are independent functions of the /a's, nevertheless
the u's in certain cases may be iufinite for certain systems of values of
the a's ; and infinite values of the /a's, by their definition, are excluded

*, -ft (*i • • • x„ . «i • • • %),

x'. = F. (x\ . . . r'ri , Mi • • • M,.), (« = 1, 2 . . . »),
X'i ~f\ (*1 ' ' ' Xn ' al • * ' ar)>

or, to the functional equations

Fi (A (*■ a) • • • /„ (x, a), /*! - . ■ /*.) =/,(xi • • ■ xn, ax . . . a.)

(i = 1, 2 . . . B).

* That is,
F.(x\ . . . x'n, Ml . . . Mr) = V\ (/i (x, a(n) ) . . .fn (», n'° ) , „x . . . m,.) =/, (*, . • • *„, «i ■ • • «,l

(i = 1, 2 . . . «),

(i = 1, 2 . . . »).

t That is

/,(/!<*, 5) • ■ • /.(x, a), O! . ... a.) =/((.r1 . . . xn, aa . . . a,.)

(i = l,2. . . »).

SS PROCEEDINGS OP THE AMERICAN ACADEMY.

at the outset.* "We cannot then assume that every transformation T,
belongs to the family E .

We may. however, proceed as follows: For all values of the a's for
which the functions fij = Mi . («x . . . ari a1" . . . a, ." ) (J = 1, 2 . . . r)
arc finite, we have

that is,

ftififa a) . . ./.(ar, a), ai . . . a,.) = -FJC/xCx, a) . . ./„(>, a), ^ . . . Mr)

=ftfa . . . arB, «i . . . aj
(i = 1, 2 . . . b).

Let /3i, ft* . . . be a system of values of the a's for which one, or more,
of the corresponding /x's is infinite. Also let blt b2 . . . be the system
of values assumed by the a's for ak = (3k (k = 1, 2 . . . >•). Since the
functions / are continuous functions of the variables and parameters, and
since we assume that the system of parameters (3 give a definite transfor-
mation T& of the family, we have

fi (/l (*. «) • • • fn fa "),/?!••• fir)

= hin. /' (/, (r, a) . . ./„ (x, a), ax . . . a,.)

= lim. /, fa . . . xa} at ... . «,.) =/, (xx . . . x,„ ^ . . . &r)

(' = 1,2 . . . »),
which is equivalent to the symbolic equation

Ta Tp = Ta lim. Ta = lim. 7;: Ta = lim. JTa = r&.

or|3 a =: 3 « = A

Consequently, the composition of two arbitrary transformations Ta and
7^ of tin- Family is equivalent to a transformation '/', of this family : that
is to say, the family of transformations Ta forms a group. The transfor-
mation '/). however, may m>i be a transformation of the group that can
be generated by an infinitesimal transformation of the group. Tims,
every transformation of a group with continuous parameters and con-
taining the identical transformation is nol necessarily generated by an
infinitesimal transformation of the group.f

Professors Study and Engel were the firel to point this out, and thus
establish a distinction between a group with continuous parameters and
a eonti >us group. J They found that not every transformation of the

■ These Proceedings, XXXV. 247.

t These Proceedings, XXXV

( Engel • Leipzig) r Berichte, 1892

SLOCUM. — FINITE CONTINUOUS GROUPS. 89

special linear homogeneous group can be generated by an infinitesimal
transformation of the group, and consequently this group is not properly
continuous in the sense in which Lie uses the term. Since this impor-
tant discovery, the subject of continuity has been investigated for the
case of the general linear homogeneous group and its sub-groups, as well
as for various other groups, by Professor Taber and his pupils, and from
a geometrical standpoint by Professors Newson and Emch. *

This paper contains an investigation of the relation of the continuity
of a group generated by infinitesimal transformations to its structure
(Zusammensetzung), and the classification of all possible types of structure
of complex groups with two, three, and four parameters with reference
to the continuity of groups of these types. All possible types of groups
with two, three, and four parameters can be divided into three classes.
Every group is continuous whose structure is of a type belonging to the
first class; every group is discontinuous whose structure is of a type
drawn from the second class ; and of the groups whose structure is of a
type belonging to the third class, some are continuous and some are
discontinuous. The parameter group of a given group Gr has the same
structure as Gr ; and every group of a given structure has the same
parameter group. In every case which I have examined, the parameter
group is discontinuous unless its type of structure is of the first class. I
have considered not only complex groups, but also real groups generated
by infinitesimal transformations.

§2.

The criterion for the continuity of an r-parameter group Gr is obtained
as follows. Let Xx . . . Xr be any system of independent infinitesimal
transformations of G>. The equations of Gr in their canonical form f
are then

(1) x\ =fi(x1 . . . xn, ax . . . ar)

(i = l,2 . '. . n),

where ft (x, a), for i = 1, 2 . . . n, is defined in the neighborhood of
the identical transformation by the series

* Taber; Am. Jour. Maths., XVI.; Bull. Am. Math. Soc, July, 1894, April,
1896, Jan. 1897, Feb. 1900; Math. Ann., XL VI.; These Proceedings, XXXV. 577.
Rettger: These Proceedings, XXXIII. 493-409; Am. Jour. Maths., XXII. Wil-
liams: These Proceedings, XXXV. 97-107. Newson: Kansas Univ. Quart., IV., V.
1896. Emch: Kansas Univ. Quart., IV., V. 1896.

t Transformationsgruppen, I. 171, III. 607; Continuierliche Gruppcn, 454.

90 PROCEEDINGS OP TIIK AMERICAN ACADEMY.

1
xx + >tfij Xjxt + - . 1 1 a A A r, + . . .
\ —-11

The transformation defined by equations (1) (the general transformation
of this group) may be denoted by J... For finite values of the parame-
•'].": . ■ . " , the transformation 7', is generated by the infinites-
imal transformation

O] A', + a.. X, + ...+" A :

but for infinite values of a,, a., ..../, '/', is not generated by an infin-
mal transformation of the group unless Ta = 7'a, the parameters

olf <ij . . . ur bring all finite.41 The transformation 7',, is defined by

X"i =/, '-'"'i • • • *'«• &1 • • • *r)
1,2 .. . n);

and tli" transformation 7',. 7!,, obtained by the composition of the trans-
formations Ta and 7',,,t is equivalent to a transformation 7', defined by

(8) /■". - / (a?! . . . xH, cx . . . cr)

(t = 1,2 .. . n),
where

' • ' Ck= r/* (", . . . a,.. /;, . . . 6,.)

(A = 1, 2 . . . r).

If the e*s can be taken finite for every finite Bystem of values of the
and //s, the group is continuous. If, however, it is possible to assign
finite values to the or*s and i>\ such that in each Bystem of values of the
ior ni'.i.) of the '"s becomes infinite, the transformation /' /',
cannot be generated by an infinitesimal transformation of the group,
and consequently the group i> discontinuous.} A transformation which
ioI !><• generated by an infinitesimal transformation of the group may
be termed tuentidBy ringtdar.% It' the parameters '/ and b are taken suf-
ficiently small, the transformation Tu '/',, can always be generated 1>\ an
infinitesimal transformation, and. consequently, Lie's chief theorem hold-,
in the neighborhood of the identical transformation.

l.il.er These Pi \ XV 67 I

/ / the transformation obtained by applying to the manifold

(m first the transformation /', and then the transformation '/'. . Lie

denotes ilii* resultant transformation bj / /
Am. Jour. Matlis., \.\II. "
Bull Am Math Boc, VI i ii;.-, Proceedings, XXXV. 680.

met

ers a± . .

. a,..

That is

to say,

from

«'*

= <f>* («i

1,2

ar, a.! . .
. . . r)

. a,.)

a\

1,2

. . . r),

. -A)

a"k

= 0* («1

• • •

1,2

ar, yx . .
• • ■ r),

.*)

SLOCUM. — FINITE CONTINUOUS GROUPS. 91

If the system of equations (4) be written in the form

(5) «'*= 0t(«i . • . ar, ai . . . a,.)

(* = 1,2 . . . r),

it can be shown that they define an r-parameter group in the variables
a and a', with p

(5)

and

(6)

we have

(7)

where

(8) y;. = fy fa . . . Or, fa . . . fir)

(J = 1,2 . . . r).

The group thus defined is termed the parameter group of the group Gr*
Since the equations defining the transformations of the parameter group
involve the functions <j>, this group is especially important in the study
of groups generated by infinitesimal transformations.

In general there is more than one system of functions </> such that

-*c = J-b 'a)

provided

Cj = <£;-(«i . . . «,., &! . . . 5r)

0 = 1,2. . . r).

But it may happen that the equations defining one group of a given
structure restrict the functions c to fewer systems of values than in the
case of another group of the same structure. Thus it is possible that of
two groups of a given structure one shall be continuous and the other
discontinuous, f

These statements are exemplified by a consideration of two groups
G2 and G2 , whose infinitesimal transformations are, respectively, pu
x1p1, and p2, x%Pi-\-pi-% Both of these groups have the structure

* Transformationsgruppen, I. 401 et seq.
t Cf. Bull. Am. Math. Soc, VI. 202.

t Throughout this paper Lie's notation will be followed, in accordance with
which

. d — d —±_

Pl = !xl' P*-dx2' • • • Pr — dZr'

92 PROCEl DINGS OP THE AMERICAN ACADEMY.

(A",. A V. The canonical form of the finite equations of the group

E •

These equations define the transformation T„ of G.z . Similarly, the
equations defining tin- transformation Th of 6\ arc

n i

X .. = X .. ,

The transformation 7' '/' . obtained by the composition of the transfor-
mations '/', and 7'. is defined by

(11) ' ' ' «k + -(<*-l)* + 5(*-l).

and, if this is equivalent to a transformation /' of the group, we have
also

(12)

/', =x1ef« + -(c'»- 1),

Therefore

X ., — x%

(18) = </>, (//,. a,, /.,. 6,),

".. 4- &j i 2 kit y— 1 . = </>2 ("i. "■_ . Jj . 6j),

where i is an arbitrary integer. Consequently for (?x there is more
than one system of functions <j>. Provided os -I 6j is nol an even multi-
ple of 7r y/— 1. every system <»f values of '-, and cs is finite. For
n, + l>.. - 2 k 77 V — I, - is finite, but the denominator of <\ becomi -
zero. In this case, however, that Bystem of values of r, corresponding
to k = — k is finite. < lonsequently the parameters Cj and c9 can alwaj -
be chosen finite, and therefore '»'. is continuous.

For the group Crt , whose infinitesimal transformations arc pfJ
/ y i /'i. /' i- defined by

rg< • "'(-• -1).

SLOCUM. — FINITE CONTINUOUS GROUPS. 93

and Tb by

05) ,

a!"1 = *'1«** + £(«8»-l).

Consequently the transformation 7^ T7,, is defined by

x'\ = xx + a2 + b2)
(16) ,

x"2 = x, e"> + b° + e**-1 (e°* - 1) + £ (e*s - 1),

whence, if 7; Ta = Tc,

a2 + &2 r.6.al/^„ in , *i ,

c> = ,, + t , te V (C"2 " 1} + A- (e ~ 1} ]- ^ <a» °2' 6l' ^'
c2 = «2 + ^2- := </>2 (%; a2, 6X, b2).

In this case there is but one system of functions <£. If, now, a2 + b2 is
an even multiple of it V — 1, c2 is finite, but cx is infinite ; that is, there
is no finite parameter c1 corresponding to this choice of the parameters
a and b. Consequently, if a2 + b2 = 2 k w \/ — 1 ^0, ThTa cannot be
generated by an infinitesimal transformation of the group, and therefore
G2 is discontinuous.*

Lie states that two groups having the same structure are (holohedri-
cally) isomorphic; but the groups G2 and G2 are not properly isomor-
phic, except in the neighborhood of the identical transformation, since one
is continuous and the other discontinuous. Whence it appears that the
conception of isomorphism, as developed by Lie, requires modification.

The parameter group of G2 is defined by the equations

a'

= a2 + a2 + 2^V-l 5 _ «j _

e"2 + «2 _ i L a2 a2 yj

(13 a) _

a'2 = a2 + a2 + 2 k tt -)/— 1,

* For the group -^-^, \{x\Pi — X2P2)> which also lias the structure
(Xj , X2) EE ^ , we have

ci = aa + ^+^>v=n [fi6,o, (e„2 _ 1} + £ (e», _ 1}] = ^ K> 02) Vi Wi

c2 = a2 + b2 + 4 kit -y/— 1 ; = 02 («i , a2 , 62 , 6.2),

and if rt„ + 62 = 2 (2 k + 1) tt <\J—\, where k is an integer, cx is always infinite.
Consequently this group is also discontinuous.

'.'I PROCEEDINGS OF THE AMERICAN ACADEMY.

where k Is any integer, and of Gj by the equations

= „ + u2 «j _ «j _

e"tT«» _ j L «. u2 /J

(17 a)

Oj s= a., -\- a2.

Nevertheless, for any value of k, the parameter group of G£ is identical
with the parameter proup of G% . For let Sa denote the transformation

(21

defined by equations (13 a), and S~ the transformation defined by
equations ( 17 a). Then for any system of values of al5 a2, aud for any
value of k, we have, by properly choosing [Sx, (32,

C.O v' -

Sa =Sfi '>

which symbolic equation persists for /3i, /?2, and k arbitrary, if a%, a2 are
properly chosen.*

Let Ta lie an arbitrary transformation of Gr, defined by the equations

1 <• r

*i = *t + 2, a, A', Xt + — 2, 24 a, at A X4 X,. + . . .

i -i : l i

(/= 1,2 . . . n).

The transformation Ta , inverse to Ta, is then defined by the equa-
tions which we obtain on replacing alf a2 • . . by their negatives.!
Compound the transformation Ta and its inverse with each transforma-
tion Ta of Gr so as to obtain the transformation TaTaT'a , which is also
a transformation of Gr; aud let

?* = Tar„ra-\t

( l .' ) /,,.. = /fl /„. / g .

Then, if

7; =y.7',.

we have

(2i) /• / /:,7'y"\

1 18 a) may l»- regarded u defining a group with three parameters
a| , a., and /■, nf which two, a, and a., vary continuously, and one, namely /, takes
only integer values. Bui this group is nol a mixed group, since we have shown
that / is unessential, thai is, it i- immaterial what value is assigned to k.
t TransformAtionsgruppen, I. 52

1 The transformation '/'.,• is said by Lie to be obtained by the application ( I
of '/', to the transformation! / of Gr. Q Lie Continuierliche Grup-
pen ■ I y.

SLOCUM. — FINITE CONTINUOUS GROUPS. 95

The symbolic equation (18) may be regarded as defining a transfor-
mation between the parameters a and a of Gr, and is equivalent to r
equations of the form

(22) ari=Fj(a1 . . .' ar, ai . . . ar)

(j = l,2. . . r).

Similarly, (19) is equivalent to

a", = Fj (a\ . . . a'r, fii . . . (3r)

(j = l,2 . . . r),
and (21) to

a"i = Fj(°i •••«>•> Yi • • • Jr)

(./ = 1, 2 . . . r),
and, in virtue of (20),

y> = 4>j {al • • • ar J Pi ■ • • /?r)

(j = l,2 . . . r).

Thus equations (22) define a group T, which is termed the adjoined of
Gr.* The number of variables of the group T is r, and it contains r
parameters, but these are not necessarily all essential. The number of
essential parameters in V is less than r by one for each independent infin-
itesimal transformation of G,. commutative with each of the infinitesimal
transformations Xx . . . Xr.-\ Thus, if Gr contains just s such independ-
ent infinitesimal transformations, T is an (r — s)-parameter group.

The canonical form of the equations defining the transformation Ta of
G?

2

IS

rq , x'1 = x1e*> + ^(e**-l),

(y a) a2

X 2 — ^2i

and consequently, if Ta, = Ta T„ Ta , we have

a^ = _LZ 1 a, e-a* — (a2 + 2 kir V— 1) - (e 2 -1)

(23) a* „ , a2 ,

a' 2 — Go + 2 k TT \/— 1 = F2 (f'l, 0T2> al> "2)'

where k is an arbitrary integer. The family of transformations between
the variables a and a' which we obtain for any assigned integer value of

* Transformationsgruppen, I. 272, 275 ; III. G67-670. Continuierliclic Gruppen,
454-455.

t Transformationsgruppen, I. 277.

96 PROCEEDINGS OF THE AMERICAN ACADEMY.

k interchanges the transformations of G« ' (so that T„ becomes 7],,), but
this family of transformations does not form a group, except for 4 = 0; in
which case it is the adjoined of G2\ This adjoined group, V , is gener-
ated by the iuliuitesimal transformations

d 9

— «•>

We may regard ab a2, and k as parameters, ai} a2 varying continu-
ously, and k taking only integer values, and then we have a family of
transformations (interchanging the transformations of G2 ) that forms a
mixed group, of which P1' is a sub-group. Only those transformations
of this mixed group which belong to r(1) are generated by an infinitesimal
transformation of this mixed group. This mixed group might be called
the adjoined of G2\ in which case the adjoined of a given group G,. would
appear as a mixed group containing more than r parameters, some of
which, however, do not vary continuously.

In the case of the group G2{'} the transformation Ta is defined by the
equations

X j ^^ Xj "J- do ■

(14 a)

x\ = x2 e^ + - (ea= - 1),

a2

and if T„, = TaTaT~\ we have

a\ = a^-0' — a* — (<f~ "2 — 1) == Fx (a?i, a„, ax, a2),
(24) f a' W l \

rot

Consequently the adjoined of the group G,~ cannot be regarded as a
mixed group. Thus the equations of the adjoined, obtained from the
Bymbolic equation Ta, = TaT„T~l, are not necessarily all linear and
homogeneous. However, they will always include one system of linear
homogeneous equations that define a family of transformations generated
by infinitesimal transformations, and forming a group.

Lie has shown that if X\ ... X generate an /--parameter group

Or in the n variables xx . . . rn, and subjeel to the conditions

r

1 A . A. i _ 1 ' , A.
0',* = 1,2 . . . r),

tie' c\ being the structural constants, the adjoined group is generated by
the infinitesimal tran jformatl

SLOCUM. FINITE CONTINUOUS GROUPS. 97

(".= 1, 2 . . . r),

and Ex . . . Er satisfy the conditions

(Ej, Ek) = l>scJksEt
i

(j, k = 1, 2 . . . r).*

The infinitesimal transformations Ex . . . Er, however, are not neces-
sarily all independent. The number of independent infinitesimal trans-
formations of the adjoined of Gr will be one less for each infinitesimal
transformation of Gr that is commutative with every infinitesimal trans-
formation of Gr (ausgezeichnete Transformation), as mentioned above,
page 95. Such a transformation will be called an extraordinary trans-
formation of G,.. It follows, from what has been said, that every group
of the same structure has the same adjoined. If Gr contains no extraor-
dinary transformation, G,. and V have the same structure. If Y con-
tains an essentially singular transformation, Gr must also contain at least
one essentially singular transformation. Therefore, if Y is discontinuous,
every group of which Y is the adjoined is discontinuous.! But Y is not
necessarily discontinuous if Gr contains an essentially singular trans-
formation.

By Lie's theorem, % the infinitesimal transformations of the adjoined of
G2 and also of G2" (since both have the same structure) are

9 9

■ — (to » «i — J

9 «i 9 ax
and thus the finite equations of the adjoined are

«'i = ol e«* - a2 ^1 (c». - 1),

a 2 — fls2,
which result agrees with the equations deduced page 95.

§3.

In what follows I shall denote by a, /?, y, respectively, the following
differential operators :

* Transformationsgruppen, I. 275; Continuierliche Gruppen, pp. 466-407.
t Taber: Bull. Am. Math. Sou., VI. 203; These Proceedings, XXXV. 590.
X Cf. Continuierliche Gruppen, p. 467.
vol. xxxvi. — 7

08

PROCEEDINGS OP THE AMERICAN ACADEMY.

a = Oj A\ + a„ X« + . . . + " A .

ft = 6, A\ + />, A, + . . . + 6, Ar.

y = Ci X\ + c- A._> +•••+'' A .

where the a's, Ps, and c's denote arbitrary parameters, aud by ea the
operator

(1 + a + ^ + §-' + . . .)/ = /+ «/+ £/ + g^/ + • • • »

wherea" + 1/=o(a"'/)-

Let

and let

i c1 r,-

= e" a:,

0 = 1,2 . . . n).

Since the a:"s are function? of /, any function of the xns,f(x\ . . . x',t),
is also a function of t. And we have

/(A • • • *'„)

, o+ 3

rf^ J,.

+ . ..

Hut

(1 = 1,2 . . . «).

df(x') » 9f(xf)dx'{ _ » , (/) ?/(*') _ X> f(/)
^/ / i d x | a t i c> x <

where A"' denotes the result of substituting the accented for the unac-
cented variables in A'. Therefore

= [/0O], o+<[ay,/,], «+Ux'V GO]. o+g-jC^ ' 03. Q+.- •

= /(*) + ' -V. /■<■.-. : £ A-v '(«) + j 3", X»/<*) + ■ ■ •
[uently, if

.-. • bave

1. 2 n),

1, 2 n).

SLOCUM. — FINITE CONTINUOUS GROUPS. 99

Let now

X'i =fi(Xl ■ • • Xni «1 • • • «,) = eaxi

(i = 1, 2 . . . n),

and

x"i^fi{x\ . . . *'„, bx . . . 6P)=ePV,
(i = 1, 2 . . . »),

where /3' denotes the result of substituting the accented for the unaccented
variables in the X's which appear in the operator (3. Then, by what
precedes, we have

X"i =fi(x\ . . . X'n, by . . . br)

= e«/i(*i ■ • • *»> *i ' • ■ b>) = ea C^*')

(i = 1, 2 . . . n).

Let the operator ea e& be defined as follows :
(e«e0)/
= (1 + (a + /8) + i(a2 + 2aj3 + /Qa)+ i(a»+ 3a2/3+ 3a/32 + /33) + . . .)/

2!v ' ' ' r/ ' 3!

^/+(a + iS)/+^(a24-2a^4-^)/+^(a3+3a^ + 3a/32+^)/+...

Then

(6ae^) a^ = ea (e^)

(i = 1, 2 . . . n),
and therefore

x"; = (ea e&) x{

(i=l, 2 . . . n);

thus eae^ denotes the result of the composition in the order named of the
transformations denoted by ea and e*3.*

By § 2, page 94, the transformation inverse to ea is e~ a. Let S*
denote an infinitesimal constant. Since the transformation ea + s'y is
infinitely near the transformation ea, the transformation e~ °-ea + s/y
is an infinitesimal transformation. If we denote its parameters by 8tbx,
?>tb2 . . . 8 t br, we have

e~ a ea + &'y

11 1

= 1 + 8t{7--(a,y)+-(a,(a,y)) - -, (a, (a, (a, y))) + . . .} +...

= e5tp — i + g j £ + . . .

in which (a, y) denotes the alternant ay — ya ; and neglecting infinites-
imals of the second and higher orders, we have

(26) 0 = y - ^ (a, y) + ^ («, (a, y)) - ^ (a, (a, (a, y))) + . . -

* Cf. Campbell: Proc. London Math. Soc, XXVIII. 381-390. Also Point-are;
Comptes Rendus, Mai ler, 1899.

100

PROCEEDINGS OF THE AMERICAN ACADEMY.

By supposition y = ^ r, Xj ; and since the r infinitesiuiul transformations
i

A\ . . . A'r satisfy the Lieschen criterion,

r r r

(a,y) = (Sj-ajXj, ^ifik^t) — -jl^i^ ~ «2^i)ci2j + («ic3 — a3c1)c13j+ . . .] Xr

Whence it follows that (a, (a, y)), etc, are linear in Xx . . . Xr. It is
to he observed that each term in the right member of (26) is linear in
cx . . . cr. Since Xt . . . Xr are independent, the coefficients of cor«
responding X}8 in the two members of (26) are equal. Therefore

bx — Gnc1 + Gl2c2 + . . -f Glrcr1

b, = G.ncx + G.2,c2 + . . . + Gircr,
(27)

br = Gricl + 6>2c2 -f ... +Grrcr,

in which the G's are integral functions of ax ... ar*

Let the determinant of the G^'s be denoted by A, that is, let

A =

Gn->

6rj2 • •

• G1

G>\,

Lr no • •

■ Go,

Grl,

Gr2 . .

Gn

The

symbolic equation

e~a

ea + S/Y _

eh't

may

be written

If A 4= 0 we may take bx . . . />,. arbitrarily, and by means of equa-
tions (27) determine the e's to satisfy this symbolic equation ; in which
case

A A A

A.,.,

■ I ]
A

f
A

inn- and Rngel. tnsformationsgruppen, III. 7~» 1 <t .« </. ami 7sw ,is,,/

I Proceedings WW. 584,

SLOCUM. — FINITE CONTINUOUS GROUPC. 101

where A^v is the first minor of A relative to G , and thus the A's are
integral functions of ax, a» . . . a,.. Consequently, if A £ 0, the com-
position of the transformation ea and an arbitrary infinitesimal transforma-
tion e lfi gives a transformation ea+ ty, infinitely near the transformation
ea, and generated by an infinitesimal transformation. Let ea~^~siy be
denoted by e"1, that is, let e"1 = ea ' Y, where

ax = a + 8 t y = a: X1 + a., X2 + . . . + a,. Xr.

Applying the infinitesimal transformation e repeatedly, we thus obtain
the equations

ea>

—

a

e

«**

ea=

=

e '

e5/^

=

a

em^

e"3

=

a

•«8V

—

a

e

esm

e°» = ea"-V^ = e°V5//3.

For n infinite, «8< is finite, and may be taken equal to unity ; thus

a„ a (3

e " = e e .

Consequently, if A does not vanish for any system of values of ax . . . ar,
in which case A is a constant,* then the composition of an arbitrary
transformation ea with finite parameters with an arbitrary transforma-
tion e" = e with finite parameters, gives a transformation of the group
with finite parameters which is generated by an infinitesimal trans-
formation.

The form of A depends only on the structural constants, and thus A is
the same for all groups of the same structure. Therefore, if the A cor-
responding to a given structure is a constant, the composition of two
arbitrary transformations of any group of this structure gives a transfor-
mation of the group with finite parameters, that is, a non-singular
transformation of the group, and consequently every group of this
structure is continuous. |

If the A corresponding to a given structure vanishes for certain systems
of values of at ... . ar, some groups of this structure may be continuous
and others discontinuous. For example, the two groups G2 and G2 , con-
sidered above, page 92, both have the structure (X1} X2) = Xl. The

* For complex groups A is either unify, or else vanishes for certain systems of
values of «! . . . ar. See the expression for A as a product on page 104.
t This criterion of continuity is due to Professor Taber.

102 i'i:oci:i:i>ings op the American academy

e"- — 1
determinant A corresponding to this structure is A = -, and this

do

vanishes for o.2 an even multiple, not zero, of ~\'— 1. Nevertheless
the group (•■, is continuous, whereas the group ' '< '.J is discontinuous.

Tlie symbolic equation eax = ea+ 'y is equivalent to the system of
equations

nk = al: -j- 6t Ck

[k= 1,2 .. . r),

which (IctiiiL- the infinitesimal transformation of the parameter group.

But the infinitesimal transformation of the parameter group is defined by

the equations r

ak = ak + S,£^(a)ft,S<

(A- = 1, 2 . . . »).»
Thereforo

r

c* = -; £y (a) fy
[k =1,2... r).
If A 41 0, equations (28; give

(fc = 1, 2 . . . r).

Therefore, if 3L . . . 2lr denote the symbols of infinitesimal transforma-
tion of the parameter group, we have

3

Aik 9

&j — 1k tJk (") — — 24 — r—

i c "* i A c a4

(> = 1, 2 . . . r).t

I-, illustrate what precedes, consider the two-parameter structure

I V, . X: ) = X1.
Equation (26 I givi

6j A', ■; A .V c, A', + r, X. — — (a, '•., - a, e, | .V, - - * i a, es - nts rV Y

a*

- - ; (a, c, - atj c, ) A", i a, es - Oj r,) A', — . . .

v\ hence folli i

/,, = c, e* — a., — 1 ) r.,,

msformntionsgruppen, !
Engi 1 and Bchnr, Transformationsgruppen, III. 764 etseg. and T.Ks,/ gg^.

SLOCUM. FINITE CONTINUOUS GROUPS.

103

Consequently

A

a

«2

_-'-(e"--fl8_l)

For a2 an even multiple, uot zero, of it \/— 1, A vanishes. Thus it is
possible that some group of the above structure shall be discontinuous.
Equations (29) give

(30)

- [*i + ~i («"2~ 0. - 1)*3] = 2, fc (a) ft, ,
e 2 — 1 a 2 i

ft,

% £2j. (a) bj

Therefore the infinitesimal transformations of the parameter group are

a,

e«2 _ J 9 «1

a> _ «i (e"2 - a-2 -I) 9

fl2(e"2— 1) 3«i ' 5a2'
As a second example, take the three-parameter structure

(X1,X) = 0, (Xu X3) = 0, (X2,X8) = X_.
Equation (26) gives

ftx Xx + b, X2 + ft3 X3 = q I, + e2 X + c3 X3 — % (a2 c3 — a3 c2) Xt
Therefore

a -

(31)
and

*i = cx + y c2

ft, Z= Co ,

"3 — cn 5

«o

C3!

1,

a.

r/2 = 1.

2

2'

0, 1, 0

0, 0, 1

Whence it follows that all groups of the above structure are continuous.
Equations (31) give

104

PROCEEDINGS OF Till-: AMERICAN ACADEMY

ci = h — 2*2 + y h = 2 j iu (a) bj,

r ,

Ja

= *,**(«)**.

c3 = b8 — 2, £8j (a) hj .

Therefore, the infinitesimal transformations of the parameter group are

".:

<3 = _ —

- 9 "i ' 9 a%

2 5 Oi <T "..

By means of the methods explained above, I have examined the deter-
minant A, and the adjoined group, corresponding to all possible types of
structure of two-, three-, and four-parameter complex groups,* and the
results are summarized in the table on pages 591-5D7, Vol. XXXV
of These Proceedings. For all types considered, the several elements
(a, y), (a, (a, y)), etc., of (26) were calculated, and the A determined
by the actual summation of the resulting series. Since making these
calculations, Professor Taber has discovered a method of obtaining A
immediately from (a, y) ; f namely, we have

(a, y) = S,

1 it, now, <£ denote the matrix

C — IttjCjn,

au cl

rv:j +

«!, Ci

a2, c2

f'oi ('3

+

X,.

2 Oj c}2l

- ~m ff. (' ■ J -J . — OjCd

— i a .

— %OjCJlr, — 2ajCjir

■ - *i«iCjrr J

,,,. are
be

- - 1
I hen A is the determinant <»f the matrix ; that is. if pj .

* fe°> - 1 \

the roota "f the characteristic equation of 0, A = IF, ( - ). Tl

» \ Pi J

' These structures are enumerated by Lie on pp. €66, 671 9, Continuier

liche Gruppen ; and also on pp. 713, 7!''., 728 780, Transformationsgruppen, III.
\ This method had previous!} been given by Professor EngeL See Transfor-
ruppen, III.

SLOCUM. — FINITE CONTINUOUS GROUPS. 105

constituents of A are integral functions of the constituents of <£, and
therefore integral functions of at . . . ar. *

In every case for which the determinant A vanishes for certain systems
of values of ax . . «,., I have found at least one group of the corre-
sponding structure which is discontinuous.

I have also determined the infinitesimal transformations which generate
the parameter group corresponding to each structure enumerated in the
above mentioned table ; but since the symbols are in many cases very
complicated, and are of no especial interest in themselves, I have not
given them.

§4.

In this section let the variables and parameters be restricted to real
values. We will then consider the continuity of real groups, that is,
groups all of whose transformations are real.

Let x'i =fi(xl . . . xn, ffli . . . Oj)

(i =1,2... n),

in which the /'s are analytic functions of their arguments, define an
/•-parameter group of real transformations. Lie's chief theorem then
states that r real, linearly independent, infinitesimal transformations

X,EE|ft^(*)A

(J = 1,2 • • . r)

in the n real variables xx . . . xn, generate an /'-parameter real group
Gr if and only if Xx . . . X satisfy the conditions

r

(Xj, Xk) E5 ^scjks Xs
(;, k = 1, 2 . . . r),

where the cjks are real quantities independent of the X's.f

Since the structural constants cjks must be real, there are more types
of structure possible for real groups than for complex groups. For ex-
ample, for the three-parameter structures

(Xi, X2) = Xi, (Xi, X3) EE 2 X2,-\Xo, X3) EE -X3'
and

(-Xi, X) = - 2 Xt, (Xu Xs) = X, (X, X3) =-2Xs,

* Taber: These Proceedings, XXXV. 581.
t Transformationsgruppen, III. 360 et seq.

LOG PROCEEDINGS OP THE AMERICAN ACADEMY.

tit.- Btractnral constants <■ . are real, and one of these --tinctures can be

transformed into the other, but only by means of an imaginary transfor-

mation ; const quentlj these structures air distinct fur rial groups.

Tin' only possible types oi structure of real or complex two-parameter

groups an- (A,. A'.) = A\ , and (A',, A..) =0. For the structure

■ — 1
. \ A , , A = - , which does not vanish tor any real system

"z

of values of ",,".,; consequently all real groups of this structure are
continuous. For the structure t A', . A'.,) = 0, A — 1 ; consequently all
real ami complex groups of this Btructure arc continuous. Therefore all

two-parameter real groups are continuous.

However, there exist three-parameter real groups which are discontin-
uous. Thus, let

A . A =0, (A",. A:;, = X. (X . A )=- A',.

For thi^ structure we have

- l_T 1 e~a*V=l—\

A = . ,

a. \/~ 1 — aa \/— 1

and A vanishes for real values of OgJ namely, when e*a is an even multiple,
not zero, of -. This indicates the possibility that discontinuous real
groups of this structure may exist. The theorem in relation to the
adjoined group, given in ^ 2, holds true also for real groups; namely, if
the adjoined of a given real group Gr is discontinuous, Gr itself, and all
groups having the same structure as Gr , are discontinuous. The adjoined
group corresponding to the above structure is, however, continuous, and
consequently not every group of this structure is necessarily discontinuous.
Nevertheless, the group /<, , /'., . r, p< — xi]h + Pz> °f the above struc-
ture, U discontinuous.4' Its finite equations in the canonical form are

A _ ft

- "

- "
i • a .

'I hi- i- one <>f the real groupa of Euclidean movement in three dimensional
• f. Trantformationagruppen, III.

SLOCUM. — FINITE CONTINUOUS GROUPS. 107

If this transformation is denoted by Ta, then from the symbolic equation
Tb Ta = Tc we obtain the five relations

(27) ef* t/=I1 + e~ c° * '~l = £ (e''3 v~x + e~ h V=1) (e"3 v -=1 + e ~ "3 V=i)

(28) «r» 4/=I1 - e~ Cs V31 = i (e'3 V~l + e~ ''* ' ^ (e 3 ^ - e~ "3 V'~1)

(29) c3 = a3 + 63,

(30) ^- (e^-1 + e-^'"1 - 2) - ^^i (e^"1 - ^V~1)

'% ■" «-3

X

+ VzLi(eV-i_e-M-i}

(31) -^ (^v-i + e-^-i - 2) + ^=^ C^1^- e~r3V'-1)

ii_ (e's v - 1 + e- *8 <=i „ 2) + ^vpi (e,3 *- 1 _ e- * V_ I,

Z 03 ii 03

12<73V 2a3 J

+ :V/~J(^V-1-^M-1)

1 2 «3 2 o3 7

Denoting the right-hand members of equations (30) and (31) by x tiutl $
respectively, and solving for c1, c2> c3, we have

1"^ PROCEEDINGS OF THE AMERICAN ACADEMY.

- ( f-J _c-i('-a+*»)»-l A '

C3 = «:( + l>. ■

It the fit's and 6'a are so chosen that \ and \p are different from zero, and

:- bA — 1 h -, where k is an arbitrary integer, both t\ and c2
infinite. Con>c<]uently this group is discontinuous.

Ou pages 106—107,384, Vol. III., Transformationsgruppen, Lie enu-
rii- rates all possible types of real projective groups ol the plane. I have
examined all the two-, three-, and four-parameter groups in this list, and
find that the groups

•'l ,"■_•• -ri l'\ -- 'u /'-'• '-j/'l'
and

/'i + >'-\}h + •')■'■,/'_• /'■■ + 'V'j/'i + -i'-j/'j. -'-j/'i — J~lp2,

and these only, are discontinuous.

The first of* these groups is the special linear homogeneous real group,
and has the structure

(.V. V ,=_2A'1.(.V;, X9) = X2} (X,, Xi)=-2XS.
The determinant A corresponding to this structure is

a2 » «*, + «,», _ j — 2 1 u\ + -,», _ j

2 \ a .. f ", ": — "-' \ a\ + ",

This vanishes if the a's are bo chosen a-- to satisfy the condition

■ ././— — /

where k is an arbitrary integer.*

-

The second of the abo^ i groups has the structure

% • \ • V) "=E — A . . ■ A . A:;) E A,.

The A corresponding t>» this structure is

il linear hon jroup hat turn shown to be discon

ftnuoui by I r Study, !• Uerichte, 1802; and the real primp by

r l it- r. Bull. Am. Math. 8oc., April, L890. The linear homogene-

.; or complex) group ia continuous, Thus a group maj be continuoua and

up,

A =

SLOCUM. — FINITE CONTINUOUS GROUPS. 109

• •

V— («2i + ""2 + «"3) — V— (a\ + «22 + «2s)

and this vanishes if the as satisfy the condition a\ + a"22 + a23 = 4 &"2 7r2,
where £ is an arbitrary integer. The adjoined of this group is

9 3 3 9 9 9

#1 ~ a3 « ' °3 « °2 ^ > °2 « °1 « j

5 «3 cy «! c «2 » a3 cy ax c> ct2

and is discontinuous. Consequently, every group of this type is discon-
tinuous.

In the table on pages 391-397, Vol. XXXV., of These Proceedings,
as mentioned above, I have marked by an asterisk those types of
structure for which all real groups are continuous ; and by a dagger
those types of structure for which I have found at least one real group
that is discontinuous.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 7. — August, 1900.

CONTRIBUTIONS FROM THE HARVARD MINERALOGICAL

MUSEUM.

VIII. — ON HARDYSTONITE AND A ZINC SCHEFFERITE
FROM FRANKLIN FURNACE, N. J.

By John E. Wolff.

WITH A NOTE ON THE OPTICAL CONSTANTS OF THE

SCHEFFERITE.

By Dr. G. Melczer.

CONTRIBUTIONS FROM THE HARVARD MINERALOGICAL

MUSEUM.

VIII. — ON HARDYSTONITE AND A ZINC SCHEFFERITE

FROM FRANKLIN FURNACE, NEW JERSEY.

By John E. Wolff.

WITH A NOTE ON THE OPTICAL CONSTANTS
OF THE SCHEFFERITE.

By Dr G Melczer.

Received June 21, 1900.

a. Hardystonite.

The new mineral hardystonite described in these Proceedings * was
found in small grains in a mass of zinc ore and isolated by handpicking
and the use of heavy solutions, while the (tetragonal) crystal system was
determined by the study of thin sections of the grains. When visiting
the mine in September, 1899, I received from the mine officials pieces
from a large mass of nearly pure Hardystonite, several inches in diame-
ter, which had been found in the same workings as the original mineral.
The material is grayish-white in color, often streaked or clouded by faint
pinkish tints, and breaks into angular fragments owing to the presence of
several cleavages ; the lustre is glassy on the more perfect cleavages,
elsewhere faintly resinous. It was easy to select material for thin sec-
tions oriented parallel to the basal and prismatic cleavages and for pol-
ished plates parallel to the base, from which the indices of refraction
were determined and the original statement confirmed ; namely, that the
mineral is tetragonal and optically negative, has a basal cleavage and
primatic cleavages parallel to the prisms of the first and second orders —
in addition, traces of a pyramidal cleavage were observed.

By means of the Abbe total reflectometer the indices of refraction
were determined on a plate parallel to the base as follows :

* These Proceedings, XXXIV. 479, 1899.

VOL. XXXVI. — 8

Ill

PROCEEDINGS OF THE AMERICAN ACADEMY.

For Na w = 1.G691

c = 1.G568

For Li &> = 1.6758 ± .0002
c= 1.6047 ± .0002

The figures for Li are inaccurate in the fourth decimal to two or more
places, owing to the indistinctness of the boundary line.

Unlike the original material, the mineral gives a strong sodium flame,
and the following analysis of the new material (I) was therefore
made : *

I.

II.

III.

IV.

V.

Si()2

37.78

37.73

<V2 1

024

38.10

A1203
Fe203
ZnO

0.91

0.43

23.38

0.91

0.43

23.35

8

2

286

1 313

I 0.57
24.30

MnO

1.20

1.25

17

)

1.50

CaO
MgO

34.22
0.20

34.19

0.20

610
6

t 618

33.85
1.62

K-P

0.78

0.78

8

1-5

Na.,0

1.10

1.10

17

S

If?

0.34

. . .

. . .

0.52

100.46

100.00

. . .

100.46

I. Analysis of new material.

II. Anulv-is of new material reduced to 100 omitting Jg.

III. and IV. Molecular proportions.
V. Analysis of original material.

It is Been that the alkalies replace in part the Ca and Mg, but that
there is still a molecular excess of the alkalies. The thin sections show,
in addition to numerous fluid inclusions, the presence of frequent small

grains of an undetermined mineral to which the content in alkalies may

h-- partly due.

• Both analyaii are! optical determinations were made by me in the Miner-
alogies! [nstitate at Munich.

WOLFF AND MELCZER. — HARDYSTONITE AND SCHEFFERITE. 115

b. Zinc Schefferite.

While visiting the mine in September, 1898, my attention was called
by Mr. Van Mater, Superintendent at North Mine Hill, to a peculiar
pyroxene which was then coming out from the workings at the Parker
Shaft, and abundant material was then secured from the ore sorting belts
in the concentrating mill. The mineral occurs in large foliated masses,
associated with franklinite, willemite, and small grains and masses of a
white zinc mineral (to be described in the future), which often lies in
thin films parallel to the basal planes of the pyroxene. The latter has a
light-brownish red color in the large masses, while another variety occur-
ring in small grains in the zinc ore has a deep brown color. The most
striking physical feature is the (apparent) basal cleavage, which is as
perfect as that of feldspar, in addition to the ordinary prismatic pyroxene
cleavage.

The angle between the two prismatic cleavages was determined by the
reflecting goniometer as 92° 59', between the basal cleavage and the prism
as 79° 02'. In a thin section parallel to (010) the angle (3 between the
basal cleavage and c' was determined as 74° 25'. It is readily seen in
this clinopinacoidal section that the apparent cleavage parallel to the
base is due to the development of gliding planes, for the basal cleavage
planes enclose thin lamellae which are evidently in the position of twins
parallel to the base with reference to the main mass of the mineral.

The thin sections show the usual optical character of monoclinic py-
roxene, — one optic axis is approximately perpendicular to the base ; on
the clinapinacoid the axis of least elasticity c makes an angle of 40° 35°
with c', lying in the obtuse angle (3.

The following analysis shows that the mineral is a zinc schefferite ;
it was found impossible to completely decompose the mineral with IIF,
hence FeO was not determined.

Si02

52.86

Fe203+Al,03

1.08

MnO

5.31

ZnO

3.38

MgO

13.24

CaO

24.48

Ig

0.45

100.80

Sp. Gr.

3.31

1 1 n PROCEEDINGS OF THE AMERICAN ACADEMY.

Optical Constants of the Sciiefferite.
By Dr. Gustave Melczer.

The indices of refraction were determined on polished plates, prepared
by Voigt & Ilochgesang, by means of the Abbe total reflectoineter and
with the application of the differential method proposed by Viola.*
There were used (1) a plate approximately perpendicular to the acute
bisectrix, (2) a thin section similarly oriented, and (3) and (4) two plates
parallel to the base (001). Since the plates were not transparent, the
boundaries of the total reflection could only be measured with the reduc-
ing telescope, bat with this, especially in the case of the last two plates
(owing to their excellent polish), the limits could be fixed within 1 to 3
minutes. By reading every 15 degrees and in the vicinity of the maxima
and minima every 5 degrees (which according to my experience com-
pletely sullices with the reducing telescope), the boundary curve was
constructed, and the following maxima and minima determined for Na
light:

(1) t 64° 2Gi' and 64° 22' 02° 56' and C2° 26'

(2) ? ? 62° 57 f 62° 27 J'

(3) 64° 25' 6 4° 1'/ 02° 55' 62° 23£'

(4) 64°20f 64° \' 62° 5 If 02° 23'

The boundaries for the comparison prism were determined, after each
pyroxene determination, by five readings 90° apart, as follows:

(1) 62° 11' (2) 62° 10y (3) G2° llf (4) 62° llf

The cause of these noticeable differences in 10 of the prism cannot
lie in the temperature, for this only differed by 3° on the different days
when the measurements were made, nor can it be due to lack of homo-
geneity in the glass hemisphere, for in using the enlarging telescope the
mean of two readings in diametrically opposite positions of the horizon-
tal circle were only 15 seconds apart. The cause can therefore only lie
in the use of the reducing telescope itself, the aocuracy of which, even
with the sharpest boundaries, is only within 1 to l \ minutes, while with
the enlarging telescope the same boundaries can be determined within 10
to 15 seconds.

• Zeil fttr Krw , XXX. 488 and XXXII 818

Prism.

62° 57£' and 62° 29'

62° 10fc'

62° 54i' 62° 24^'

62° 12'

62° 51f 62° 21i'

62° 11'

WOLFF AND MELCZER. — HARDYSTONITE AND SCHEFFERITE. 117

As regards the differences between the boundary angles of the individual
plates, these are too large to be ascribed to indefiniteness of the boundary
lines, for as already mentioned these were sharp, especially in (3) and
(4). In order to be quite sure I repeated the measurements on three
plates and found :

(1) | 64° 27f and 64° 23^'

(2) 64° 21' ...

(3) 64° 21' ...

There the above supposition was correct.

Since the refractive index of the glass of the Abbe hemisphere, accord-
ing to one of my previous determinations, is 1.8903 for Na light, and
that of the glass prism used 1.6724, there follows from the boundary
angles given above for the individual plates :

Tna £na Gna

(1) and (2) 1.70G0 ± 0.0002 1.6840 ± 0.0002 1.6766 ± 0.0004

(3) 1.7050 ± 0.0005 1.6834 ± 0.0001 1.6757 ± 0.0001

(4) 1.7045 ± 0.0001 1.6827 ± 0.0001 1.6752 ± 0.0002

and as a mean : y — /3 = 0.0218
j3 — a = 0.0075
y_a -0.0293

These variations of the indices in the individual plates could perhaps be
referred to local variations in the chemical composition of the zinc
schefferite, or to the uneven distribution of pigment. The latter is not
noticeably different, and yet it is known that for certain minerals at
least very small variations produce such differences in the indices.
If from the above values we take as the mean of the indices

y = 1.705

p= 1.683
a= 1.676

then by calculation 2 Va for sodium = 59° 29}/.

For the direct measurement of the axial angle the plate (1) could not
be used, for although one optic axis and the middle of the figure could
be seen, and therefore, H1 might have been measured, yet the geometric
orientation of the surface could not be determined ; the thin section (2)
gave both hyperbolas, but somewhat indistinctly, so that they could only

118 PROCEEDINGS OF THE AMEBICAN ACADEMY.

be centred within i° accuracy. The plate was immersed in monobrom-
oaapthalin whose temperature during the measurement was 19° and
whose index of refraction was determined by the Abbe total-refracto-
meter ;is : nna = 1.6597. With the thin section the following values
Were determined in the axial angle apparatus :

Li

Na

Tl

Hi 37° 30'

37° 20'

36° 50'

H2 24° 10'

23° 40'

23° 40'

whence, as w^ for the section was found to be G25 57' 15" and (3= 1.6842 :
2 Va for sodium = GO" 0' and p > v.
Munich, April, 1900.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 8. — August, 1900.

ON THE THERMAL AND ELECTRICAL CONDUCTIVITY

OF SOFT IRON.

By Edwin H. Hall.

Investigations on Light and IIeat made and published wholly or in part with Appropriations

from the rljhford fund.

ON THE THERMAL AND ELECTRICAL CONDUCTIVITY

OF SOFT IRON.

By Edwin H. Hall.

Received May 9, 1900. Presented July 17, 1900.

Trirc general method used in the investigation of which this paper is
to give an account is set forth with much detail in two articles already
published.* Certain more or less important changes of apparatus or
procedure will be described and discussed later; but the main results
will first be given.

The metal studied was Taylor (Yorkshire) wrought iron, recommended
to me by an engineer friend of much experience as the softest wrought
iron to be found in the Boston market. Chemical analysis showed the
following composition : —

Iron 99.93 %

Carbon 0.059

The density was about 7.785 at 0° C.

The values which I find for the thermal conductivity, k, of this* iron
are

at 0.1528 28°. 2 C.

" 0.1514 58°.3 C.

if the specific heat of water at each of these temperatures is called 1.
This would give for the temperature coefficient 0.0003, very nearly.

If the values of k are revised in accordance with the values proposed
by Winkelmann f for the specific heat of water at the given tempera-
tures, they become

at 0.1513 28°.2 C.

" 0.1511 58°.3 C.

* Thermal Conductivity of Mild Steel, by E. II. Hall, These Proceedings, XXXI.
271, 1896 ; On the Thermal Conductivity of Cast Iron, by E. II. Hall and C. H.
Ay res, XXXIV. 283, 1899.

t Part 2 of Vol. II. p. 340.

122 PROCEEDINGS OF THE AMERICAN ACADEMY.

ami the temperature coefficient deduced therefrom will be too small to be
worth writing down, 0.0000 ?, let us say.

lint the recent work of Calleudar and Barnes* gives for the specific
beat of water

0.9992 at 25° C.

0.9987 " 30° C.

0.9992 " 55° C

1.0000 " 60° C.

If these values given by Callendar and Barnes are adopted as correct,
niv first values of k will stand almost without change ; and I sliall there-
fore leave them for the present without correction for variation in the
specific heat of water.

My value for the temperature coefficient, 0.0003 or a little less, is, so
far as it goes, a corroboration of the substantial accuracy of the tempera-
ture coeilicient found by Lorenz, 0.0002282, although it may be doubted
whether the last three figures of this number are of much significance.
This agreement is eminently satisfactory to myself; for a comparison of
the work of Lorenz with that of other investigators has convinced me
that his value of the temperature coefficient is entitled to an especial
decree of confidence.

Measurements of the electrical resistance of the iron were made on nine
cylinders, each 2 cm. long and about 0.23 cm. in diameter, cut from the
same great bar as the disk on which the measurements of k were made.
Thf length of the cylinders, like the thickness of the disk, was taken
parallel to the length of the bar; and the cylinders were cut from a part
of the bar adjacent to that from which the disk was cut. The extreme
dilference in the specific resistances of these cylinders was apparently
about 5 per cent. The mean specific resistance was found to be 12240
at 18 C The mean specific conductivity, x, at the same temperature
would, therefore, he 817 < 10 7 C. G. S.

The ratio k -r- X is about 171'', at 0°C.

The "thermo-electric height" of this iron, as compared with copper,
i- about

1028 x 10-8 \nlts at 26 .0 C.
980 " " " 41 .."- C.
986 " •' •• 54 -"> C.
870 " " " 71 .1 C.

• Physical Review, April, 11

HALL. — CONDUCTIVITY OF SOFT IRON. 123

The relation of the values here given for k and x to those found by
others who have studied the thermal and electrical conductivities of soft
iron, I have set forth in the " Physical Review " for May-June, 1900.

The following details of my work are perhaps unnecessarily extended
and tedious ; but a considerable study of the literature of thermal con
ductivity has convinced me that most experimenters in this field have
omitted important matters in the printed description of their investiga-
tions. It is my hope that those who may have to deal with problems
similar to, though not exactly like mine, will find what is here written
worthy of their attention.

The iron used for the experiments on thermal conductivity was in the
form of a disk cut from the end of a five-inch cylinder and turned down
at first to a diameter of 10.5 cm. The thickness of the disk was about
1.996 cm., the greatest thickuess indicated by the calipers being 1.998 cm.,
and the least 1.995 cm.

Treatment of the Disk.

The disk was coated with copper electrolytically on both faces by a
method substantially the same as that previously described. Spots
which, because of slight flaws in the surface, appeared not to be taking
the copper well from the preliminary cyanide bath, were rubbed with the
point of a lead pencil to give them a coating of graphite, after which
they speedily became coppered like the rest. In the sulphate bath the
convex surface of the disk was protected as before by rubber bands ; but
outside these bands was now placed a band of paraffined paper about
5 cm. wide, the object of which was to impede the deposit at the edge of
the faces of the disk and so make it keep better pace with the rate of
deposit at the centre of the faces. As before, it was occasionally neces-
sary during the progress of the deposit, which lasted about a week, to
remove the disk from the bath in order to break off or file off projecting
pimples, or corals, of copper.

The final coating of copper on each face, after being turned down
nearly to a plane, was about 0.2 cm. thick. The whole curved surface
was now turned down until the diameter of the disk was 10.00 cm.

Mounting and Use of Disk and Adjacent Apparatus.

Figures 1 and 2 of an article already mentioned, " On The Thermal
Conductivity of Cast Iron," indicate with accuracy in most particulars

124 PROCEEDINGS OF THE AMERICAN ACADEMY.

the method of mounting and using the disk now under consideration.
Certain small changes, however, must be imagined in Figure 2, in order
to make it accord perfectly with the latest developments of the apparatus.
Thus, the water stream entering beneath the disk is no longer allowed to
flow without restriction straight against the centre of the lower face. A
little affair shaped somewhat like a three-legged stool is placed at the top
of the admission tube, just beneath the point marked by the letter C,
and most of the water escapes laterally between the legs of the stool,
although a small part of it runs through a hole in the top straight against
the disk. This device was adopted in the hope that it would make the
temperature more uniform over the lower face of the disk. With a
similar purpose regarding the upper face of the disk, the hard rubber
block 1111 has been replaced by one having a somewhat more gradual
curvature, so as to make a thinner and more rapid stream over the
central parts of the disk. It is doubtful whether these changes have
done much good, on the whole, although they appear to have made the
difference of temperature between top and bottom of disk at the centre
very nearly equal to the mean difference at other parts, as numbers
• ntly to be given will show. Another attempt to improve the distri-
bution of temperature, by making the How of water more nearly equal
along different radii of the upper face of the disk, affected the manner of
admission of the water to the funnel FF ; but, as it was of doubtful
utility, it need not be described. The air-vent hading up from the
funnel FF by the tube «> has been considerably enlarged; but the small
escape of water which had previously been maintained at this vent is no
longer permitted.

The small copper wires, extending from the copper coatings of the
disk, are now protected from actual contact with the hard rubber plugs
Iv,, K.. etc., and with the soft rubber packing surrounding these plugs,
by a wrapping of oiled silk, with the purpose! of preserving the wires from
the destructive action of the sulphur emitted from the rubber ; but, after
all, the wires cannot be depended upon for more than a few months.

Stops about 0.16 cm. thick, placed at the <-,]^r> of the laces of the
disk, prevent the blocks III! and I I'll' from approaching these faces so

near a- to endanger the safety of the small wires or cut off the flow of

water.

The parts marked d, and Jj iii figure 2 of the former paper now
contain spirals of platinum instead of thermo-electric junctions of copper

and German silver, change of resistant f platinum having been Bubsti

tuted for change oi thermo-electromotive force as a means of measuring

HALL. — CONDUCTIVITY OF SOFT IRON.

125

the change of temperature of the stream of water which flows over the
upper surface of the disk. More will be said of this later. The parts
surrounded by the water-jacket are carefully and thickly wadded with
cotton-wool so as to make a nearly cylindrical body, about 18 cm. in
diameter, up to the plugs Jx and J2. The plugs also are covered with
cotton-wool, as well as the slot in the top of the jacket. Around the
disk itself the wadding is so thick as barely to allow the jacket to
enclose it.

The copper wires leading out from the plugs J\ and J2 ran, after
July 17, outside the cotton-wool wrapping, but without touching the
jacket.

Determination of the Difference of Temperature
of the Two Faces of the Disk.

As before, this is effected by thermo-electric means, the iron disk and
its two copper coatings being used as a thermo-electric couple. As be-
fore, thirteen fine copper wires lead
off from as many points on the upper
coating, and similar wires from cor-
responding points on the under coat-
ing. Each pair of corresponding
wires can be used singly, or all the
thirteen pairs can be joined and
used in multiple by an arrangement
described in a preceding paper. The
distribution of wires over either
coating is shown in Figure 1, the
numerals being alongside the points
of attachment to the coating.

In the arrangement of these points
there was an attempt to make the

various areas represented respectively by the individual points as nearly
equal as practicable. Point 13 is intended to beat the centre of the
disk ; points 3, 6, 9, and 12, 2.55 cm. from the centre of the disk ; points
2, 5, 8, and 11, 3.80 cm. from the centre of the disk; points 1, 4, 7, and
10, 4.60 cm. from the centre of the disk.

No great accuracy is attained in placing these points ; and, in fact, each
attachment is rather a line, about 0.5 cm. long, than a point, each such
line, except No. 13, crossing nearly at right angles the radius upon which
it lies.

Figure 1.

126

PROCEEDINGS OP THE AMERICAN ACADEMY.

It is, of course, desirable to make the flow of water along the two faces
of the disk such that the indication received from any one pair of wires,
one above and one below, shall be about the same as that given by any
other such pair of wires, and much thought has been given to the attain-
ment of this end. The result is not quite all that could be wished for ;
but it is such that any important error from the inequalities observed is
very unlikely. The following tables show the. results of tests made
under conditions as nearly uniform as it was found practicable to keep
them. The galvanometer deflection credited to each pair of junctions is
the mean obtained from two short sets of observations, the various pairs
being used first in the order in which they are here given, and then in
the reverse order. Between these forward and back series of observa-
tions with single pairs a short set of observations with all the pairs in
multiple was made, and the mean deflection from this set, corrected for
the difference of resistance of the multiple and single arrangements, is

also given below.

Temperature about 28° C.

Junctions.

13 and 13'

1'
4'

1 and

4 and

7 and V

10 and 10'

2 and 2'

:> and -V

Mill 8'

11 and 11'

3 and 3'

(', and 6'

9 and '.l'

12 and 12'

Deflections.
11.1

11.0

11.1

12.li 1
12.2/

10.!)

10.8 r

11.7 J
11.6 >

10.6 v
10.7
10.5 |
11.1 '

11.6

11.3

1U.7

M-- in, 1 1.2

Along radii

1 and

1'

1 1 .0

2 and

2'

10.9

3 and

3'

10.6

4 and

4'

11.1

5 and

5'

10.8

6 and

6'

10.7

7 and

7'

12.0

8 and

8'

11.7

9 and

9'

10.5

10 and 10'

12.2

11 and

11'

11.6

12 and 12'

11.1

10.8

11.1

11.0

All in multiple, 1 1 .3

HALL. — CONDUCTIVITY OF SOFT IRON. 127

Temperature about 57° 0.

Junctions. Deflections Along radii.

13 and 13' 10.4

1 and 1' 11.

■■->)

4 and 9.5 '>iQ7 2 ana z ll.D ^. 11.3

7 and 7' 10.8
10 aud 10' 10.8

9.6

2 and 2' 11.5 \
5 aud 5' 9.5 (

8 and 8' 10.G f

11 and 11' 11.3 ^

8 and 8' 10.6 [

3 and 3' 11.0 ^ 9 aud <y 99 )

G and G' 9.7

9 and 9' 9.9

12 and 12' 10.6 > 1 and 11.3 ) 19.0

>1

1 and

1'

11.5

2 and

2'

11.5

3 and

3'

11.0

4 and

4'

9.5

5 aud

5'

9.5

6 and

6'

9.7

7 and

7'

10.8

8 and

8'

10.6

9 aud

9'

9.9

10 and 10'

10.8

1 1 and 1 1'

11.3

12 and 12'

10.6

I

Mean, 10.5
All in multiple, 10.4

It seems likely that the differences between the different radii are due
in part to inequalities of water-flow caused by the lodging of air-bubbles
at various points on or near the surfaces of the disk.

The agreement between the deflection obtained with all junctions in
multiple and the mean of those obtained with the pairs used singly ap-
pears satisfactory, in view of the not very rigid character of the test by
which the comparison is made.

The resistance of each individual pair of wires, out to the point where
all were connected in multiple, is shown in the following table, as it
was found July 12th, 1899 : —

13-13'

0.56 ohm

1- 1'

0.55 «

4- 4'

0.55 «

7- 7'

0.5G "

10-10'

0.56 «

2- 2'

0.55 "

5- 5'

0.56 «

8- 8'

0.56 ohm

11-11'

0.56 "

3- 3'

0.55 "

6- 6'

0.56 "

9- 9'

0.55 "

12-12'

0.56 "

The deflections noted above are those of a rather sensitive astatic gal-
vanometer, the sensitiveness of which was frequently determined by
means of a potentiometer and a standard Carhart cell, or rather, two

128 PROCEEDINGS OF THE AMERICAN ACADEMY.

such cells, which differed from each other in electromotive force about
1 part in 500.

The mean difference of temperature of the two streams of water on
entering the apparatus was about 8° at low temperatures and about 7°. 6
at high temperatures. The mean difference of temperature between the
two surfaces of contact of the iron and copper, as indicated thernio-
i lectrically, was about 1°.42 at low temperatures and about 1°.58 at high
temperatures, which shows that heat is communicated more readily from
the water to the copper coatings, and vice versa, at high temperatures
than at low temperatures, other things being equal.

In order to determine the difference of temperature, just mentioned,
between the surfaces of contact of the copper with the iron disk, it was
of course necessary to find the thermo-electromotive force corresponding
to a known difference of temperature of two similar junctions. This
datum was obtained by the method described in Appendix I of the paper
on the " Conductivity of Cast Iron," to which paper a number of refer-
ences have already been made in this writing. Some particulars of the
present case follow.

A second disk about 2.5 cm. thick was cut from the same end of the
same cylinder of Yorkshire iron that had furnished the disk already de-
scribed. This second disk was then cut into two half disks, and from one
of them was taken a slice, thickness-wise, about 11 cm. long, 2.5 cm.
wide, and 0.3 cm. thick. This slice was then sawed up, crosswise, into
thirty-three bars. All but ten of these bars were then reduced by filing
and milling to a diameter of about 0.1G cm., and the length of each was
reduced to 2.0 cm. The remaining ten were reduced in the same way
to a thickness of 0.23 cm. and a length of 2 cm. The bars were then
numbered from 1 to '■>'■> in the order of their original position in the slice
from which they bad been cut. Numbers 1, -1. 7. 10, 1."., l'0. 2:;, 26, 29,
and 82 were placed end to end, in the order just given, in the bore of the
den cylinder of the thermo-electric test apparatus shown in Figure 5
of the article on " Conductivity of Cast Iron."" No essential change has
In', n made in this apparatus or in the manner "f lining it since the de-
scription given in the article just mentioned was written, except in these
particulars, that the electrical resistance * of the end-to-end row of bars

• Thia n &8 likely to be as much a- 2.5 <>r :', ohms when the [in is-

about 8 was newly applied to the row. but a gentle rocking, from

le, of the copper blocks, kept w< 11 seated, would in the course of a few min-
utes reduce this resistance to less than 1 ohm, the end pressure remaining un-

HALL. — CONDUCTIVITY OP SOFT IRON. 129

and copper end-pieces has been dealt with more carefully and successfully
in recent experiments than iu earlier ones, and that in some cases the bars
have been placed in a glass tube instead of a wooden one. The results
of these thermo-electric tests, subject to slight corrections for pecu-
liarities of the thermometers* used, are given below. The temperature
put down for each case, in the second column, is the mean of the ther-
mometer readings in the two copper blocks at the ends of the row of iron
bars. The third column, headed A, gives for each case the mean differ-
ence of temperature of the two thermometers. The fourth column, headed
E, gives for each case the thermo-electromotive force corresponding to a
difference of 1° between the two thermometers, each of which thermome-
ters is supposed to iudicate with sufficient accuracy the temperature of
the copper-iron contact neighboring to it.

.A B

7°.0 k 1041 x 10"8^

7°. 52J 1017 "

7° fil I 1024 "

>26°.6 ' "": >7°.30 "Z ,, S1028 X 10"8

Date

T

June

3,

99

25°. 8-

«

7,

u

27°.7

a

30,

(t

25°.3

July

3,

a

26°.5

(c

4,

a

26°. 4

.'<

6,

a

27°.8

June

6,

«

40° .5

K

7,

c<

44°. 6

July

3,

a

40°. 6

a

4,

cc

38°.8

a

6,

cc

42°.0

June

6,

cc

55°. 6

a

7,

<«

.56°. 5

July

3,

CC

52°. 2

a

4,

a

53°.8

cc

6,

cc

54°. 6

June

3,

((

69°.0-

CC

7,

((

74°. 5

(I

30,

cc

69°.4

July

3,

cc

70°. 6

(C

4,

cc

71°.3

(«

6,

u

71°. 9

volt

8°. 22/ ' 1042

6°.64\ 1031

60.79/ 1012

7°.W\ 979

7°.23/ 964

7°.79>7°.19 982

6°.94\ 990

6° .57; 976

7°.03>\ 926

6°. 95/ 930

7°.60>6°.89 941 "

6°.64\ 938 "

6°.25j 933 "

6°.07>. 885 "

6°. 24 J 850 "

,71°.l ££>.* |» I > 871X10-

6°.04l 879

6°.66j 870

changed, and, except during the moments of measurement of resistance, no electric
current flowing through the row of bars.

* Baudin, Nos. 10,286 and 10,287, as in previous work.
vol. xxxvi. — 9

130 PROCEEDINGS OF THE AMERICAN ACADEMY.

Applying to the mean values of E, found above, certain small correc-
tions which take account of errors in the graduation of the thermometers,

we get, —

T K

2G°.6 1028 X 10"8 volt
41 ..; 980 "

54°.5 936 « «

71°.l 870 « u

These numbers plotted, with temperatures for abscissas and electro-
motive force per degree for ordinate*, indicate a curve which descends with
wry gradually increasing slope with rise of temperature. Indeed, this
curve is so nearly a straight line that its curvature cannot be satisfactorily
shown iu a small figure. It would be almost perfectly straight if the
numbers given under E were 1028, 975, 929. and 870.

The minute examination which I was obliged to irive to the individual
small cylinders, during measurements of their electrical resistance, led me
to notice defects and possible distortions which might, I feared, have
aifected their thermo-electric quality. Accordingly, several months after
tests just described were made, I undertook a similar test with ten of tho
somewhat larger cylinders already mentioned, which had apparently suf-
fered much less in the process of milling. In this later test I found it
convenient to enclose the bars in a tube of glass instead of a tube of
wood. This test gave 1064 X lO"8 for E at 15°. 4 C.

The earlier tests, above described, did not run so low in temperature,
but by extrapolation they give, for 15°. 4 C, E = 1065 X 10~8, or some-
thing very close to that, a satisfactory agreement.

Determination of the Difference of Temperature of the In-

GOIK D Ol rGOING WATEB at THE CHAMBER ABOVE THE

Disk.

It has been already stated that a differential platinum thermometer
•i ed for this purpose instead of the two copper I rerman-silver thermo-
electric junctions which had been employed in the preceding investiga-
. This change was the result of a conviction, the fruit of much
exp< and vexation, that no suitable permanent protection can be

found for the thermo-electric junctions against the action of hot water.
A thin layer of shellac well dried on appears to be the best coating; but
this is liable to give way at a most inconvenient time and bring to caught

HALL. CONDUCTIVITY OF SOFT IRON.

131

the labor and observations of hours. The platinum thermometer method
is not without difficulties, as the following pages will show ; but it ap-
pears preferable to the other.

Figure 2 represents one of the platinum spirals, S, in position for use.
It consists of about 22 cm. of wire, 0.012 cm. in diameter. The diameter
of the coils is about 0.45 cm. The ends of this platinum wire are soldered

Figure 2.

to copper wires, C and C, each about l.G m. long and 0.1 cm. in diameter,
which extend through the hard rubber plug R. Each of the copper
wires is soldered at the outer end to a copper rod 5 cm. long and 0.6 cm.
in diameter, which is well amalgamated at the free end and serves to
make connection with a mercury well of a Carey Foster bridge. Care was
taken to make the length of the wire very nearly equal in the two spirals,
and equal care to make the copper wires, all of which are from the same
piece, all alike at first. The parts of Figure 2 which are in solid black
represent metal ; the parts cross-hatched thus / represent soft rubber.

The resistance of each spiral was about 1.4 ohms, and that of its con-
necting copper wires about 0.1 ohm. Trial showed that one spiral with
its connecting wires had a resistance about 0.0006 ohm greater than that
of the other spiral and its connections. Slight changes in the copper
wires reduced this difference of resistance to something like 0.000025
ohm. As the original difference of resistance was probably nearly all
in the spirals, these continued to differ, at the temperature of the room,
by something like 0.0006 ohm, which would correspond to about one-
seventh part of a degree difference of temperature. No attempt at a

132 PROCEEDINGS OF THK AMERICAN ACADEMY.

closer adjustment of the spirals was made, as the method of experimen-
tation was expected to eliminate from the result any considerable error
arising from this inequality. That this expectation was finally justified
will be shown later, although more dilliculty was encountered than at
first appeared, the fact being, apparently, that the two spirals, although
made from the same piece of wire, did not have quite the same tempera-
ture coefficient of resistance, so that the difference of resistance between
them was not constant when their temperature was varied,* but increased
when the spirals were heated.

Calibration of Differential Platinum Thermometer.

In order to calibrate this differential platinum thermometer, the two
spirals were placed in streams of water, the temperatures of which were
measured by means of the same thermometers that were used in studying
the copper-iron thermo-electric junctions, and the difference between the
electric resistance of spiral No. 1, with its connecting wires, and that of
spiral No. 2, with its connecting wires, was measured by means of a
Carey Foster bridge. The conditions under which this trial was made
resembled very closely those under which the spirals were to be used in
the main experiment, the plugs bearing the spirals being inserted in the
same sockets in which they were afterward to be used, which sockets
were temporarily removed from the main apparatus and fitted to brass
tubes through which flowed the streams of water employed for the test ;
these tubes were wrapped around with cotton wadding, and sockets and
tubes together were surrounded by the same water-jacket that was used
later to surround the main apparatus. The copper connecting wires led
down between the cotton wadding and the water-jacket, and then out
beneath the latter without touching it ; this detail is mentioned because
tin temperature of the connecting wires is by no means a matter of in-
difference in some parts of the investigation. In this calibration test
each stream flowed past the thermometer bulb before reaching its spiral,
and past the spiral before reaching the ends of the copper wires which
carried the spiral. The distance from each thermometer bulb to the

* The wire from which the spirals was made was annealed by drawing it through

a flame, which treatment may have- introduced into it BOme luck of uniformity.
After the spirals were formed anil soldered to the copper they were kept in a bath
of melted parafflne, in the neighborhood of 146° <'., about two hours, with the
object of removing inequalities of condition caused by the bending to which the

w ire had been subjected.

HALL. — CONDUCTIVITY OF SOFT IRON. 133

spiral beyond it was perhaps 2 cm. The velocity of the stream between
the bulb and the spiral was some 30 or 40 cm. per second. One stream
was usually about 6°. 7 warmer than the other. The method of alterna-
tion was used ; that is, if the stream passing spiral No. 1 was during one
set of observations kept a certain number of degrees warmer than the
other stream, the streams were exchanged at the end of that set and an-
other set was then made, the difference uf temperature and the mean tem-
perature remaining nearly as before, and no part of the apparatus suffering
change of place except the cock by means of which the change of flow
was affected. The combination of the two complementary sets of obser-
vations gave a result from which errors due to disagreements of the ther-
mometers and lack of perfect equality of the spirals was practically elim-
inated. Sets of observations were made at various mean temperatures,
and, in order to make the results at these various temperatures com-
parable, slight corrections, amounting at the most to less than one-half
of one per cent, were made in certain cases because of errors in the
graduation of the thermometers, the combined diameters of the bores
being slightly greater at some temperatures than at others. Similar cor-
rections made in calibrating the copper-iron thermo-electric junctions
have already been referred to.

Each of the horizontal lines, numbered from 1 to 8 in the table below,
gives the result of two complementary sets of observations, such as have
been described above. T is the mean temperature of each pair of sets ; A is
the mean difference of temperature of the streams ; L is the mean length
of bridge wire included between the two points of equilibrium correspond-
ing to the two positions of the commutator of the Carey Foster bridge.

May

Date

T

A

L

L h- A

5, 1899

20°.77

6°. 68

20.30

cm.

3.039

0)

6, «

36°.72

6° .33

19.21

«

3.032

(2)

11, «

70°.41

6°.55

19.86

a

3.032

(3)

11, «

21°.32

6°.99

21.20

a

3.033

(4)

12, «

76M2

6°. 63

20.21

a

3.048

(5)

17, "

51°. 60

6°. 80

20.57

u

3.024

(6)

26, "

52°.46

6°.96

21.01

a

3.018

(7)

26, "

37°. 69

7°. 15

21.59

u

3.019

(8)

The mean resistance of the bridge wire used in this test was 0.001491
ohm per cm., the material being German silver and the diameter about
0.2 cm. The mean resistance of the bridge wire, German silver about
0.4 cm. in diameter, employed in later experiments, when the spirals

T

L -^ A

i

21°. 1

3.036

17.77

37°.2

3.026

17.71

52°.0

3.021

17.68

73°.3

3.040

17.79

134 PROCEEDINGS OF THE AMERICAN ACADEMY.

were used to determine the change of temperature of water in the heat-
conduction apparatus, was 0.0002548 ohm.* With these data we get
from the tahle above : —

Mean of

(1) and (4)

(2) « (8)
(6) » (7)

(3) « (5)

Here / stands for (L-fA)x (0.001491 -h 0.0002548), that is, the
length of bridge wire, used in the main experiments, corresponding to
a difference of 1° in the temperature of the spirals. Values of / for
temperatures intermediate between those here given were found when
needed by interpolation between the values of I here given. It will be
observed that there is a maximum difference of 11 parts in 1770 between
the given values of I. Rigid constancy of I would mean rigid constancy
of the mean of the temperature coefficients of the two spirals with ref-
erence to the mean scale of the mercury thermometers used. Such con-
stancy was hardly to be expected.

Use op Differential Thermometer in the Main
Experiments.

In the main experiments the plugs bearing the spirals were at first
placed in their supporting sockets according to Figure 2, which repre-
sents the spiral placed in the incoming stream. The water, therefore,
on entering the apparatus, passed the first spiral before reaching its
copper connections, but on leaving the apparatus it passed the copper
connections of the second spiral before coming to the spiral itself. Ac-
cordingly, any net conveyance of heat by the copper wires into or out
of the stream was superposed on the action of the conducting disk to
produce the difference of temperature noted by the spirals. Any such
effect of the copper wires would probably be very small except at high
temperatures ; and at such temperatures the effect would be, in large
measure, eliminated from the result by the practice, always maintained,
of combining one set of observations in which the disk carried heat to
the upper stream with another set of observations in which the disk car-
ried heat from the upper stream, the mean temperature of the disk re-

* Each bridge wire was calibrated by finding what length upon it corresponded
to the difference between the resistance of a 1 ohm coil and that of a similar coil
shunted by a known, much larger, resistance

HALL. — CONDUCTIVITY OP SOFT IRON. 135

maining nearly unchanged throughout the two sets.* The elimination,
however, would not be perfect ; for the reason that the temperature of
the stream which we are considering is not quite the same in the two
complementary sets of observations just described, being about 8° warmer
in one set than in the other. The question at issue comes, therefore,
very nearly to this, whether the amount of heat carried away from the
stream by the wires, when the stream is 8° warmer than the room, is
negligible in comparison with the amount carried away by the disk, the
difference of temperature of the two faces of the iron being, as we have
seen, usually 1°.4 or more. This question can be answered in the affirm-
ative; for the aggregate cross-section of the four copper wires is little, if
any, more than 0.03 sq. cm., and the length of each wire, from the spiral
out to the point where it is exposed to the temperature of the room, is
not far from 30 cm. The carrying power for heat of a rod of copper
30 cm. long and 0.03 sq. cm. in cross-section, the thermal conductivity
of copper being taken as eight times that of iron, would be about one
five -thousandth part of the carrying power of the iron disk for a given
difference of temperature, and not more than one eight-hundredth part,
if the difference of temperature were 8° for the rod and 1°.4 for the disk.
Nevertheless, after a considerable number of trials had been made with
the spirals placed as in Figure 2, the plugs bearing the spirals were put
in at the other ends of the supporting sockets, so that any changes of
temperature produced in the stream by the copper wires must now occur
either before or after the change of temperature noted by the spirals.
The mean of the results obtained after this change of arrangement was
slightly different from the mean of those obtained before ; but it is un-
likely that the difference was due to this change.

In the observations of July 31 and thereafter a loop of each copper wire
was kept in a pocket containing oil on the outside of the water-jacket, so
that each wire had at this point a temperature not more than one or two
degrees different from that of the spiral with which it was connected.

Another question was whether the heat carried in or out by the copper
wires affected the temperature of the spirals directly by metallic conduc-
tion, so as to keep them at a temperature different from that of the water
passing them. The reasons for thinking that any such effect was neg-
ligible are given in the Appendix to this paper.

It has been stated that there was a difference of resistance between
the two spirals at any given temperature, and that this difference in-

* The time between the exchange of the water streams and the beginning of the
next set of observations was usually rather more than twenty minutes.

136 PROCEEDINGS OF THE AMERICAN ACADEMY.

creased with rise of temperature. This difference, if constant at each
temperature, would be eliminated by combining two sets of observations
made at the same mean temperature, one set being made with spiral
No. 1 the warmer, and the other with spiral No. 2 the warmer. Sets of
observations were, it is true, combined in pairs, but in each such pair it
was the mean temperature of the disk that was kept nearly constant,
while the mean temperature of the spirals was seven or eight degrees
warmer in one set than in the complementary set. Accordingly, the
mean result of a pair of sets, taken without individual correction for the
inequality of the spirals, would have been subject to the error caused by
ignoring the variation of this inequality through a rise of, we will say,
8 degrees in mean temperature. This error would have been about 1.5
per cent of the final result at any temperature. To prevent such an
error, corrections, based on careful observations made for this specific
purpose, were applied in each individual set of observations for conduc-
tivity ; and, after these corrections had been applied, the complementary
sets were put together for a mean result at any given temperature of the
disk By this means the error in question was probably reduced to very
small dimensions.

It appeared likely that some part of the apparent difference of resist-
ance of the spirals at high temperatures was due to difference of resist-
ance of the copper connecting wires, which near the spirals and for some
distance away from the latter were considerably heated. As the experi-
ments went on, increasing care was taken to make the condition of the
wires leading to one spiral as nearly as might be the same as the condi-
tion of those leading to the other spiral, in order that the differential
changes of resistance should be confined to the spirals themselves. The
result was a progressive diminution of the correction for isothermal ine-
quality of resistance at high temperatures ; but this correction remained
lartre and somewhat uncertain to the end, so that the value of the eon-
ductivity calculated from any one set of observations, unbalanced by its
complementary Bet, was liable to a considerable error, as the details
presently to be given will show. If I were to go through Buch work
again, I Bhould try to reduce the importance of temperature changes in
the copper connecting wires by increasing the resistance of the spirals or
tie- change of temperature of the stream between them.

Heating \m> Flow of Water.

The method of heating and controlling the flow of the streams of watei
been described in previous papers. The onlj feature of importance

HALL. — CONDUCTIVITY OF SOFT IRON. 137

to add to the description is this, — that in these later experiments the
freedom of movement of the covers of the gas-holders, upon which the
accuracy of regulation of the gas pressure depends, has been greatly
increased by keeping the hammer of an electric bell in brisk action
against the side of each gas-holder. When the regulating devices were
working well the maximum variation of temperature in either stream
during one of the main sets of observations, occupying some 45 min-
utes, was usually about one fifth of a degree C.

Results in Detail for Thermal Conductivity.

Experiments on conductivity of the iron with the apparatus as de-
scribed in this paper were begun July 13, 1899. The results of that
day and of July 14 are not here given, as they were obtained before the
full importance of some precautions was recognized ; but all the results
obtained later are recorded below. The subscript 1 refers to cases in
which the warmer stream of water ran above the disk, the subscript 2 to
those in which it ran beneath the disk. T, the mean of Tx and T2, is the
mean temperature of the disk during a pair of complementary trials. K,
the mean of KL and K2, is the mean value of the conductivity given by a
pair of complementary trials. The results obtained July 31 and August 2
are distinguished from the others by the fact that they were made after a
certain change in the position of the thermo-metric spirals (see p. 135).

Date

T,

T2

K.

K,

T

K

July 17,

'99

29°. 2

28° .9

0.1521

0.1531

29°. 1

0.1526 ^

" 18,

<(

28° .7

28°. 5

0.1468

0.1553

28°.6

0.1511

« 22,

ti

28°. 2

28°. 1

0.1568

0.1500

28° .2

0.1534 ^0.1528

« u

(i

28°. 1

27°.8

0.1539

0.1513

28°. 0

0.1526 1
0.1545 J

" 27,

a

27°. 4

27°. 5

0.1580

0.1510

27°. 5

" 31,
Aug. 2,

it

u

27°.7
28° .3

27°. 4
28°.2

0.1465
0.1543

0.1529
0.1573

27°. 6
28°.3

0.1497 )

A .-.Q ^0.1528

0.1 008 )

0.1526

0.1530

28°.2

0.1528

July 18,

It

58°. 8

58°. 6

0.1531

0.1529

58° .7

0.1530 ^
0.1525 (
0.1507 (°-1521

" 22,

it it

it
it

58°. 9
58°. 6

58°.8

58° .7

0.1608
0.1593

0.1441
0.1420

58°. 9

58°.7

" 27,

11

56°. 2

57°. 6

0.1552

0.1490

56°.9

0.1521 )

" 31,

ti

57°.8

57°.7

0.1510

0.1498

57°.8

0.1504 ^ rtlCAO

Aug. 2,

ti

58° .3

58°. 6

0.1568

0.1432

58° .5

0.1500 ]

■ U. li)\J6

0.1560 0.1468 58°.3 0.1514

138 PROCEEDINGS OF THE AMERICAN ACADEMY.

The values of K given above take no account of the variation in the
Bpecific heal <>f water between 28°. 2 and 58°.3. The temperature coelli-
cient of K obtained from the trials made July 31 and August 2 after a
change in the arrangement of the spirals, which change was supposed to
make for greater accuracy of results, is decidedly greater than that ob-
tained from the trials which preceded this change; but the considerable
difference between the values of K at low temperature found July 31
and August l' makes it unsafe to give especial weight to the value of the
temperature coefficient calculated from the trials of these two day-.
Tin- beet course appears to be to take the mean of all the results at low
temperature and compare it with the mean of all those at high tempera-
ture ; and this has beeu done, with the result stated at the beginning of
tin- paper.

Measurement of the Electrical Resistance.

A -ati-factory determination of the mean electrical resistance of the
iron was a work of considerable difficulty. I attempted it at first by use
of several little cylinders, each taken singly. Buch as had been used in
testing the thermo-electric quality of the iron (see p. 128). I found,
however, that the last milling which these cylinders had been Bubjected
to had left them somewhat irregular in diameter, so that it was impossible
to measure this dimension accurately, even when calipers with jaws
meeting along a narrow line were used. Accordingly I used the some-
what larger cylinders, already mentioned, which had suffered much less
from the imperfections of the milling process. These were about 0.23 cm.
in diameter, and could be measured with satisfactory accuracy.

The straight thick wire of a Carey Foster bridge having been replaced
by two .-tout lira-- rods lying iii line, with a gap between their ends,

which were amalgamated, one of the iron cylinders was placed end to end
between these rods and firmly held there under considerable pressure. Then
the resistance of a certain measured length, 1.472 cm., of the iron bar was
rmined in the usual way by means of the bridge and the accompany-
ing low resistance coils. Four cylinders were tested in this way, and

their mean Specific resistance was found to be 12-1."><>, ('. < '•■ S.. at L'"J C.

Later the resistance of the same length of each of these bars and of
f I %•• other similar bars, held u jual described, was measured by a poten-
tiometer method, the potentiometer wire being oi copper drawn especially
for this U I The mean specific resistance of the nine bars, a> found bj
latter method, was 12240, C G. 8., a1 18 C, the value given in the

ginning of this paper.

HALL. — CONDUCTIVITY OF SOFT IRON. 139

Appendix.

Does the conductive action of the copper wires attached to the platinum
thermometer spirals introduce error by preventing these spirals from taking
the temperature of the water flowing past them?

If the warming or cooling produced by the action of the copper wires
were equally great at the two spirals, no harm would result, as it would
not affect the difference of their temperatures, which is the quantity
measured- but this perfect compensation can hardly be, for all of the
copper connections are subjected to very nearly the same temperature
conditions outside the supporting hard rubber plugs, while the spirals
themselves differ in temperature about 0°.5. The differential effect, upon
which alone the possibility of sensible error depends, is very much the
same as if one of the spirals were at the same temperature as its connect-
ing wires outside the plug and the other spiral 0°.5 warmer or colder
than its connecting wires outside the plug. It is possible to make a very
rough estimation of the maximum amount of error which could arise
from such a condition. For the purpose of this calculation it may be
assumed that the hard rubber plug is a non-conductor of heat, — an
assumption which tends to magnify the effect under discussion. The
length of the plug, that is, the length of the wire from air to water, is
2.5 cm., the diameter of the wire about 0.1 cm., the length of each copper
wire exposed to the water about 1.6 cm. The length of the platinum
wire in each spiral is about 22 cm., and its diameter about 0.012 cm.
The point of attachment of the platinum to the copper is near the middle
of the part of the copper exposed to the water. We will discuss the
action of a single copper wire, and assume that one half of the plat-
inum wire was attached to this by one end, the other end being free.
For this purpose it will be necessary to know something about the
"surface conductivity" of copper immersed in running water. Fortu-
nately the main experiment with the copper-coated disk gives us some
information in regard to this, — very inaccurate information, no doubt,
but sufficient for the present purpose.

In this main experiment, with a mean temperature t in the disk, and
with a temperature t + 4 in the stream on one face and t — 4 in the
stream on the other face, the two meeting surfaces of copper and iron
had respectively temperatures about t + 0.8 and t — 0.8, we will say.
Assuming the thermal conductivity of copper to be eight times that of
iron, and remembering that the copper coatings are each 0.2 cm. thick
while the disk is 2 cm. thick, we get for the temperatures of the two outer

] 1 PB0CEEDING8 OP THE AMERICAN ACADEMY.

copper surfaces, in contacl with the water, the temperatures / + 0.82 and
t — 0.82 respectively. This makes each copper coating to have a tem-
perature gradient ofO .1 per cm., with a difference of 3 .18 between its
outer surface and the water stream flowing across it. The ratio of the
temperature gradient to the external difference is therefore about ,'_,.

According to this we may infer that a Btream of water flowing across

one face of a copper wire, with a speed equal to that of the flow across

surface of our disk, and with a temperature t degrees above the tern-

perature of the face of the wire will maintain within that wire a gradient

of temperature equal to t -:- 32, all lateral action being excluded.

The point of attachment of the platinum wire to the copper is about
midway of the exposed part of the copper, and is as much as 3.0 cm.
from the cuter end of the plug, [f the copper wire terminated at this
point of attachment, and Buffered conductive contact with the water only
at its terminal surface, the change of temperature from the cuter end of
the plug, Buppoeed non-conductive, to the end of the wire would accord-
ingly be about ;;;J. ,V it as the difference of temperature between
the end of the wire and the water flowing past it. Ii. therefore, the wire

at the cuter end of the plug exceeds in temperature the stream o\ water
li\ ii .."». as uc w ill assume, the fall of temperature within the wire would
be about 0 .0 1. and the end of the wire would be about 0 . 16 above the

temperature of the stream. This conclusion, however, is based on a
false assumption as to the area of contact of the wire with the water; in
fact, this area of contact is about BlXtj times as great as the cross-section
of the wire, and the point of attachment of the platinum is near the
middle of this area, so that we shall not be very far from the truth in
assuming that the temperature of the copper at the cross-section next the
point of attachment of the platinum is the Bame that it would be if the

Wire had contact With the water only at this cn»s-eeiioii hut had sixtj

times ae gr< at a Burface conductivity as such an area reallj has in contact

with the Btream. This leads to the conclusion that the fall of temper-
ature within the wire, from the outer end of the plug to the point of

attachment of the platinum, is about i;V times as great as the difference of
temperature between the point of attachment and the Btream. This last
difference would therefore he rather less than 0°.l, but we will call it

that.

The problem now is to find how much the mean temperature of a plat-
inum wire 0.012 cm. in diameter and II cm. long will exceed that of the
uai tm iii which it is placed, if one end of this wire is kept 0 .1

above the temperature of the water This problem is of a familiar sort.

HALL. CONDUCTIVITY OF SOFT IRON. 141

and is easily dealt with if we know the ratio between the thermal con-
ductivity, K, of the platinum and its " surface emissivity," E, in the
stream of water. Assuming E to be the same for platinum as for copper,
and K to be \ as great for platinum as for copper, we get for E H- A the
value \. We now have (see Preston's " Heat," p. 513)

2_Ep 1 .012
* ~ KA ~ 4 X ~M& ~ b6'

or /j. = 9 in round numbers. Then, using the formula 6 = 60e~ >xX, where
60 is excess of temperature of the heated end of wire above temperature
of water, 0 the excess of temperature at any point distant x cm. from this
end, and e the Napierian base, we have, reckoning 0 in per cents of a
degree : —

X

0

0.1 cm.

0.2 " 1.7 ' 2 \ Mean

0.3 " 0.7 'k n\ (2.4 % of]

0.4 "
0.5 "

Accordingly, the mean excess of temperature of the wire along its first
0.5 cm., less than ^(j of its whole length, would be about 5 % of the
usual difference of temperature, about 0°.5, between the two spirals;
and beyond this point the excess would be very small ; so that the error
made by neglecting the difference of temperature between the spirals and
the water is not important, provided the calculation just made is tolera-
bly accurate. The most uncertain element in this calculation is probably
the value of the " surface emissivity," which is based on observations
made on the behavior of the disk and the streams across its face. But
as the velocity of the water in passing the spirals is probably ten times
as great as its mean velocity across the disk, it seems altogether likely
that a sufficiently low estimate of emissivity has been used in the calcula-
tion, and that the possible error from the source in question has been
overestimated in the discussion just given.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 9. — October, 1900.

CONTRIBUTIONS FROM THE CHEMIPAL LABORATORY OF
HARVARD COLLEGE.

A NEW CONCEPTION OF THERMAL PRESSURE AND
A THEORY OF SOLUTIONS.

By Gilbert Newton Lewis.

A NEW CONCEPTION OF THERMAL PRESSURE AND
A THEORY OF SOLUTIONS.

By Gilbert Newton Lewis.

Presented by Theodore W. Richards. Received July 19, 1900.

Introduction.

For an understanding of all kinds of physico-chemical equilibrium a
further insight is necessary into the nature of the conditions which exist
in the interior of any homogeneous phase. It will be the aim of the
present paper to study this problem in the light of a new theory, which,
although opposed to some ideas which are now accepted as correct, yet
recommends itself by its simplicity and by its ability to explain sev-
eral important phenomena which have hitherto received no satisfactory
explanation.

The theory suggested itself in the consideration of certain remarkable
general laws which treat of heterogeneous equilibrium in which the
several phases are subject to different pressures. These laws will be
discussed in the first section of this paper, and in the second section it
will be shown that they can all be explained by a single simple assump-
tion. In the third section it will be shown that the same assumption is
alone sufficient to explain all the laws of dilute solutions. In the last
section other consequences of the new theory will be discussed, especially
in their relation to the theory of van der Waals.

I.

The Effect of Pressure on the Tendency to Pass from

Phase to Phase.

It has been shown by several investigators* that, in a number of cases
of heterogeneous equilibrium, if the pressure upon one of the phases
alone is changed, a readjustment takes place that can be easily calculated.

* Poynting, Phil. Mag. (5), XII. 32 (1881) : Schiller, Wied. Ann., LIII. 396,
(1894) ; Hall. Jour, of Phys Chem., III. 452 (1899.)

VOL. XXXVI. 10

14G

PROCEEDINGS OF THE AMERICAN ACADEMY.

In order to introduce a general discussion of this problem, which leads to
very notable and instructive results, let us consider first a simple, special
case, namely, the question of the effect upon the vapor pressure of a liquid
caused by a change in the total pressure on the surface of the liquid.

Figure 1 represents a ring-shaped enclos-
ure containing a liquid X in the part BCD,
and the vapor of X throughout the remain-
ing space. The space D E contains also an
infinitesimal layer of some inert and insoluble
gas, which is prevented from diffusing into
the space B A E by a membrane at E,
which is permeable only to the vapor of X.
The foreign gas thus enclosed exerts a
pressure upon the liquid at D and maintains
a difference of level, d 11. between B and
D*. This pressure, moreover, must have
an effect upon the vapor pressure of the liquid, for, on account of its
weight, the pressure of the vapor is greater at I) than at B, but the
liquid is in equilibrium with the vapor at both points, therefore the
vapor pressure of the liquid is greater at D than at B. If d Px represent
the difference in vapor pressure between B and D, and s{ the specific
gravity of the vapor, then we may write,

dP1 = sidII.

If d P, represent the difference in the total pressure upon the liquid at
B and I), and s., the specific gravity of the liquid, then

,1 P., = s..d If.
From these two equations,

dj\

.1 /'.

«i

or if -r, and <r., represent the specific volumes of vapor and liquid respec

tively, then

dp

,//■

,T.,

<u

0)

■ ijection . raisi <\ to proofa of tlii- kiml ; thus, in this case, it would

be argued thai in reality the liquid will distil from I> to the Bpace above E. Thai
this may happen in no iray invalidates the proof, for it ia nol necessary thai il
should happen. By keeping E A free from drops "f the liquid the system ia in
perfect equilibrium, the i quilibrium of auperaaturation. That there is another
more stable equilibrium possible ia ol no concern.

LEWIS.

THERMAL PRESSURE.

147

In general, therefore, an increase in the total pressure upon a liquid
will cause an increase in the vapor pressure, and the ratio of the two
changes will be the ratio of the specific gravities of liquid and vapor.
This relation, which has been stated before in several different forms, has
an importance which has been hitherto overlooked, probably because,
under ordinary circumstances, the calculated effect upon the vapor pres-
sure has been too small to be measurable, and also because the result
has usually been obtained by assuming the applicability of the gas law to
the vapor, thus making the result seem only an approximation.

Although ordinarily the magnitude of the effect is extremely small, in
some cases it must be of considerable significance. Equation (1) shows
that the influence of external pressure upon the vapor pressure depends
upon the relative densities of vapor and liquid. Therefore, for liquids of
high molecular weight, as a rule, the effect will be considerable, and also
in the case of a liquid whose vapor is under high pressure. But espe-
cially must the effect be considered in the study of critical phenomena
and the influence of foreign substances in the determination of the critical
constants, for in the region about the critical point the densities of liquid
and vapor approach identity. The proof of equation (1) given above
shows that its validity rests upon no assumption 'as to the specific nature
of the two phases considered. The probability is immediately suggested
that equation (1) is simply a special statement of a general law applying
to all heterogeneous equilibrium. That this is true may be shown in the
following way.

In figure 2, let A B C and A' B' C represent
two similar enclosures. In the first, Xx and
X2 are two different phases of a simple sub-
stance X ; Px and P2 are the pressures exerted
by an inert gas on the two sides and are such
that equilibrium exists ; B is a membrane per-
meable only to the vapor of X. The specific
volumes in phases Xt and X2 are <xi and <r2
respectively. The enclosure A' B' C is the
same as A B C except that here there is another
state of equilibrium in which the pressures
upon the phases Xx and X2 are Px + d Px and
P2 -(- dP.2, and the specific volumes are <ri — d<n
and o-2 — da2 respectively.

Now let the following process occur reversi-
bly and isothermally : (1) One gram of Xx is

Figure 2.

148 PROCEEDINGS OF THE AMERICAN ACADEMY.

removed from A at pressure Pl; (2) it is compressed to Px + d P1;
it is introduced into the secoud enclosure ;it A' against pressure
1\ -f- dPx: ill one gram of X^ is taken from the second enclosure at
( ' at pressure Pa + d Pz\ (o) it is allowed to expand to P.2 ; (6) it is
introduced into the first enclosure at (' against pressure Pa.

The system will have returned now by internal adjustment to its
original condition. The quantities of work done hy the system in the
ral steps are,

)Vl = P1a1,

Wi = -P1d<rl,

W9 = -(P1 + dP1) ((Ti-do-0,

U\ = (P.2 + dP,) (<T,-dn

IV, = P2da2,

n; =- p,a2,

By writing the sum of these terms equal to zero, according to the
second law of thermodynamics, we obtain the equation,

dP rr

(Tod P., — ti dPx = 0; or, , " = — . (2)

The meaning of this equation is obvious. When under any conditions
two phases of a substance are in equilibrium and the pressure is increased
upon one phase, then in order to maintain equilibrium the pressure must
be increased on the second phase; and the second change in pressure
must be to the first as the specific volume of the first phase is to that of
the second. For example, if ice and water are in equilibrium at atmos-
pheric pressure and an additional pressure of one atmosphere is put upon
the ice alone, then equilibrium can be maintained by an additional pres-
sure of 1.09 atmospheres upon the water, since 1.09 is the ratio of the
volumes of ice and water.

The above law has been derived without any assumption whatever,
and is based only upon the second law of thermodynamics. It must be
considered, therefore, in the Bame degree aa the latter a universal and
exact law of nature. This law may be presented in another form.

Thi i tendency for the particles ol everj phase to escape into M>me

other phase. Liter a function \p will he so defined as to represent this
escaping tendency. Here it will be sufficient to consider ip merely a
quantity Bach that when two phases are in equilibrium, C ) in- the same
value in both ; when not in equilibrium,^ iter in the less stable

LEWIS. THERMAL PRESSURE. 149

phase. When two phases in equilibrium are subjected to infinitesimal
changes of pressure resulting in a second state of equilibrium,

d\l/1 — d tpz, or in other terms, -pr~ d P1 = ~~ d P.

9 ipi p _ 9^

9P1d^~9P2"

(3)

where 9 denotes a partial differential ; finally,

9P\ _ dj\ __ oj
9if/o d Px a.2

9P,

from equation (2). In general, therefore,

9ilf

where k is a constant.

That is, the change in the escaping tendency of any phase with a given
change in the external pressure is proportional to the specific volume of the
phase. For example, if solid and liquid benzol are in equilibrium at one
pressure and this pressure is increased, the escaping tendency of the
liquid is increased more than that of the solid in the ratio of the specific
volumes, 1.13 to 1.11. The liquid phase, therefore, totally disappears.
In the case of water, whose liquid is denser than the solid, the phenome-
non is exactly reversed. The above law, therefore, expresses quantita-
tively what the principle of Le Chatelier states qualitatively.

In the preceding discussions we have dealt with different phases of a
simple substance, not with a mixture, but the same method of proof and
therefore the same law can be shown to be applicable to all cases in
which the phases considered are all capable of being converted entirely,
under the conditions of equilibrium which exist,* into one substance,
whether this be a pure substance or a mixture.

Equations (1), (2), and (3) apply, therefore, to all cases where there is
association, dissociation, polymerization, or isomerization, provided that
all these different molecular species are in "true" equilibrium with one
another.

* This qualifying phrase is necessary. The different phases must not merely
have the same composition. Thus, a system composed of water and a gaseous
phase of hydrogen and oxygen in equivalent proportions differs essentially from a
system of solid amnionic sulphydrate and a gaseous phase containing ammonia and
hydrogen sulphide in equivalent proportions. The latter is subject to the above
treatment, the former is not.

150

PROCEEDINGS OF THE AMERICAN ACADEMY.

For the sake of completeness the corresponding phenomena in the case
of other mixtures will next be considered.

In Figure 3 let the tube CA B contain a homogene-
ous mixture of liquids and the outer vessel one of the
pure liquids, X, alone. A and l> represent two mem-
branes permeable to X alone. If pressure is applied in
the tube so that X neither enters nor leaves the tube at
A, then there will he equilibrium at B also, for other-
ui-e a continuous cyclic process would he kept up.
contrary to the second law of thermodynamics. The
escaping tendency of X is the same outside and inside
at A and also at B, since there is equilibrium at both
points, but the changes in pressure between A and I>
which determine the magnitude of that tendency are different inside and
outside the tube. If d His the height A B, and the specific gravity of
the mixture and the liquid X are SY and Sa respectively, the changes in
pressure are Sid II inside the tube and S3 d H outside. \$dPx and d /'..
respectively, represent these changes in pressure,

c

Figure

dP1 £,

d /'..

■s

or

d I\
d />.,

<T„

o-l

This is the same as equation (2), and at first sight it seems that the
influence of pressure upon the equilibrium between two phases is as
simple in the case of mixtures as in the case of pure substances. It is
actually, however, much more complicated. In the above example there
is a change in the escaping tendency of X from the mixture, not only
because there is a change of pressure, but also hecause, in general, betwi en
A anil B there is a change in the relative amounts of the components of
the mixture. In oilier words, the concentration of X differs it A and 1 '.. '
Moreover if there is chemical equilibrium in the mixture the chemical
reaction will have run nearer to completion in A or in B according to
whether the total volume is increased or diminished by the reaction. On
account of this complexity the further study of mixtures, although of
much interest, must he deferred, -inc.' it fa not essential to the present
paper.

Qouy and Chaperon, Ann Ch. Ph. (6), Ml 884 (18t

LEWIS. — THERMAL PRESSURE. 151

II.

Thermal Pressure.

When we consider again the laws which govern different phases of a
simple substance, as expressed in equations (1), (2), and (3), their great
simplicity and generality suggest that some equally simple physical expla-
nation is possible. As we have already seen, an increase in the external
pressure always produces an increase in the tendency of a substance to
escape from any phase. The pressure can influence this tendency only
by changing the conditions within the phase. Let us briefly analyze
these conditions.

If in the interior of any homogeneous phase of infinite extent, not sub-
ject to gravity, we imagine a septum of infinitesimal thickness, i. e. a
mathematical plane, there is every reason to believe that upon each side
of this septum there would be a pressure exerted which would depend
upon the temperature, and would in general differ from the pressure
observed at the surface of the phase. This internal pressure will be
termed the thermal pressure of the phase. It would be expressed, in the
hypothetical terms of the kinetic theory, as equal to the number of
molecules passing from one side to the other in unit time through an
imaginary plane of unit area, multiplied by twice the average momentum
of each molecule in the direction perpendicular to the plane.

An illustration of the meaning of thermal pressure is the quantity

R T

- in the equation of van der Waals ; this quantitv, if the theory of

v — b

van der Waals is correct, represents the thermal pressure of a compressed
gas. Thermal pressure will denote the pressure due to heat in distinction
from that due to attractive or repulsive forces.

The actual pressure observed at the surface of the phase may be con-
sidered equal to the thermal pressure plus or minus the resultant of all
other repulsive or attractive forces acting in the phase. This resultant
will be called the attractive pressure and designated by a, — positive if
an attractive force, negative if repulsive. If the thermal pressure is
represented by yS we have the equation,

P-a = P, (4)

where P is the external pressure. This equation combined with equa-
tion (3) gives,

d{p — a)

152 PROCEEDINGS OF THE AMERICAN ACADEMY.

that is, the tendency of a substance to escape from a phase is greater,
the greater the thermal pressure and the less the attractive pressure.
Now equation (o) applies to any phase of any pure substance, solid.
liquid, or gaseous. Iu sonic cases a small change in pressure will cause
:i very great change in both [3 and a ; in others, a vers small change in
both. The change in (3 relative to that in a will sometimes be very
great, sometimes very small. In general, the function \p depends, not
upon the absolute, individual magnitudes of/? and a, but only upon their
difference ; and we may conclude from the very universality of equation
(5) that it is true not merely when (3 and a are changed by a change in
the external pressure, but that in general a change in (3 or a. no matter
how caused, will affect the quantity i}/ only as it changes their difference
(3 — a, and that any change in (3 accompanied by an equal change in a
has no effect on ty since one neutralizes the other. This will be a neces-
sary postulate in the development of the following theory.

Let us consider a system composed of a liquid or solid phase and its
vapor, under such circumstances that the latter obeys the laws of a per-
fect gas. By equation (1), if we replace -- by - , the molecular volumes,

RT . °"! '"'

and write ?q = — — where R is the gas constant, T the absolute tem-

■* i
perature, we obtain

dPx vA rfP, P^ dj\ dP.,

JP2 = V °r' dP2= R T ; °r' ~i\ ~~ Jit p

''■j

This equation expresses the very remarkable result that the change in
vapor pressure is to the total vapor pressure as the change in the exter-
nal pressure upon the solid or liquid phase is to the pressure which that
phase would exert if it should behave as a perfect gas. This may be
presented in different form by the aid of equation ( 1) as,

,//', ;,,-qa)

/> / D 7»\

(f)

In attempting an explanation of these relations let us consider first a
e in which the change in " . is negligible compared with the change in
ft„. That is, jl: changes while •<■ remains constant Since the attractive
forces remain constant, it might be predicted a priori that the vapor pres-
sure would be proportional to /•'. . the thermal pressure, especially if the
kinetic point of view is adopted; for the vapor pressure is believed to be

LEWIS. — THERMAL PRESSURE. 153

dependent upon the chance of any one molecule to escape from the phase
in question and upon the number of molecules per second which share
this chance. This chance for each molecule will depend (1) upon its
momentum and (2) upon the various influences that retard its motion
outward. The latter are the various attractive forces that have been
included in the quantity u.2. Therefore in any isothermal change in
which a, is coustant neither the momentum * nor the retarding influences
vary, and the vapor pressure then will be proportional to the number of
molecules coming to the surface per second, and therefore proportional to
the thermal pressure. Hence,

P1 = kfo <*,*£ = &. (8)

•* 1 P 2

Moreover in the general case, when both /?2 and a2 change, if we postu-
late, as on page 152, that equal changes in (32 and a2 produce equal and
opposite effects on the vapor pressure, then we should expect that the
change in vapor pressure would be to the total vapor pressure as the
effective change in thermal pressure is to the total thermal pressure ; if
"effective change" in /32 is used to mean the change in /?2 over and above
that required to compensate for change in a2 . That is,

dPi _ d(p-2-o2)
Pi ~ ' & " W

Comparing this equation, derived from kinetic considerations, with
the one which has been proved thermodynamically, namely, equation (7),

dPx _rf(&-a2)

Pi

G?)

it is evident that the only assumption necessary to make the two identi-
cal is the following: The thermal pressure of any phase is equal to the
pressure which the substance would exert if under the same conditions,
it should behave as a perfect gas.

Objection to this assumption cannot be made on the ground that it is
not sufficiently simple : but is it too simple ? What has become of the
correction " b " of van der Waals, to say nothing of all the complications
that may exist in a liquid or solid phase ? It must be confessed that the
above assumption seems, at first sight, absurdly simple and quite improb-
able. I shall attempt to show, however, that this assumption is not only

* Whether this momentum is quite constant will be considered later.

154 PROCEEDINGS OF THE AMERICAN ACADEMY.

not opposed to any facts, hut is capahle of explaining many facts besides
those which we have already discussed. On the other hand, this assump-
tion seems entirely irreconcilable with one of the accepted principles of
the kinetic theory, but it is directly deducihle from this theory if the
latter is modified in the way which will now be proposed.

In the kinetic theory of gases there are two quantities of fundamental
importance: one is the kinetic energy of a molecule, and is represented
by I mu2, where m is the mass, and u the velocity of the molecule; the
other is that which has been called thermal pressure, and is proportional
to mu X n, the product of mu, the momentum of the molecule and n, the
number of molecules whose centres of gravity pass in one second through
unit area. In the perfect gas n is proportional to u ; the kinetic energy
and the thermal pressure in a perfect gas are proportional to each other
ami to mu2. In substances, however, which deviate from the condition
of a perfect gas the kinetic energy will still be proportional to mit1, but
(ui'i)n will be proportional to mu- only when n is proportional to zi,
and this will be the case only when the molecules behave on collision
as perfectly elastic mathematical particles, that is, when there is no
correction corresponding to the quantity b of van der YVaals.

In the kinetic theory of gases the temperature is shown to be measured
by mw2, and in every attempt hitherto to extend this theory to less sim-
ple conditions of matter the fundamental assumption has been that the
kinetic energy of progression of the molecule is proportional to the
absolute temperature. I now propose to reject this assumption entirely,
and to substitute the assumption that the temperature is in all cases
measim-d by the quantity (mn) n or by the thermal pressure; more ex-
plicitly, instead of assuming that the kinetic energy of a molecule of any
substance is the same as if it were a perfect gas, while the quantity
(mu)fl may vary in any way, it will now be assumed that (mu) n in any
substance is the same a* it would be it' the substance Bhould behave as a
perfed gas and that at one temperature the average kinetic energy of
progression of the molecule may vary.

This proposition appears less revolutionary if it is borne in mind that
in the only case in which the kinetic theory is entirely satisfactory]
namely, the perfect gas, the two assumptions become identical, and.
therefore, the change in no way affects the previous kinetic explanation
of all the phenomena ol The only other application of the kinetic

theory thai ha- met with any degree of Buccess, the equation of van
der Waal 8, will be discussed later in it> relation to this m\v kinetic
eption.

LEWIS. — THERMAL PRESSURE. 155

In the meantime it is necessary to consider what advantage in theory
may be gained by adopting the proposed assumption. According as we
deal with a pure substance or a mixture, the quantity (rnu) n will be a
measure of the total thermal pressure, or of the partial thermal pressure
of any one molecular species, if we so designate that pressure which may
be conceived to be exerted on either side of an infinitesimally thick
membrane in the interior of a homogeneous phase by the molecules of
that particular molecular species. The principle of thermal pressure
offered on page 153 may be put in the form of an equation, as,

/» = n-^, (10)

where /3 is the partial or total thermal pressure, as the case may be, of
some one molecular species ; n is the number of gram molecules of this
species ; R is the gas constaut ; T, the absolute temperature ; and V, the
total volume occupied.

This idea could be otherwise expressed by an extension of the rule
of Avogadro, as follows : All substances at the same temperature and
the same thermal pressure have the same number of molecules in unit
volume.

Equation (10), which, if correct, represents a universal principle of
nature, must be capable of very wide application. In the next section it
will be shown that by assuming the correctness of this equation, and with-
out any other hypothesis, it is possible to derive all the laws of dilute
solutions. It is well to emphasize that while equation (10) was shown
to be consistent with the kinetic theory in order that it might appear more
probable, still, having once assumed this equation, it is unnecessary
henceforth to adopt any kinetic view whatever.

III.

A Theory of Solutions.

Notwithstanding the simplicity of the phenomena of solutions, they
have as yet received no entirely adequate explanation. In fact, in some
explanations assumptions have been necessary that are inherently improb-
able. In the theory of solutions here presented it will be unnecessary
to assume either that the solute does or does not combine chemically
with the solvent in any way. It will only be necessary to suppose that
when n gram molecules of any substance are dissolved in a solvent, that

156 PROCEEDINGS OF THE AMERICAN ACADEMY.

in the solution there are still n gram molecules of something different
from the Bolvent itself.*

All the general laws of solutions may be derived thermodynamically
from either of two empirical equations, each of which, moreover, may be
derived from the other. The first is the law of van 't IIotF,

n R T

n = -pr-, CD

where Fl is the osmotic pressure of a solution containing n gram molecules
of a solute in a volume V.
The second maj be written,

ii^*i= " . (i2.)

or. more strictly,

^ = ^. (12 1,,

In < 12 a) // ami /// represent the number of gram molecules of solute and
solvent respectively; ij/x and if/a denote the escaping tendencies of the
solvent from the pure solvent and the solution respectively; in i 12 In d\p
represents the change in the ex-aping tendency of the solvent, due to -In
gram molecules of the solute. In order to make the equations entirely
definite it is necessary to give the function \p, or the escaping tendency,
a meaning less vague than that which sulliced on page l Is, by defining
the actual value of ij/ for some one condition of each substance. There-
lore the escaping tendency of a perfect gas will be defined as equal to its
gas pressure, and the escaping tendency from most actual gaseous phases
will be approximately the partial gas pressure.

Equation (12b) Bimply unites in one general equation the law of
Raoult for the lowering of vapor pressure, the law of Nernsl for the
lowering of solubility, and less directly the law for the depression of the

* It is probable that, in some cases, " association " with the solvenl takes place
and in others thai it does not A crystal containing water of crystallisation seems
in mi way different from any other double salt, and since it lias often been shown
that double salfa nch In sniuiioii.it i> probable that water enters into

many molecular COmpOUnda m solution. Thui it -e. in- probable that when a Ball

forms a number of solid hydrates, all these compounds and others whose solubility
i- neater are present In solution In proportions varying continuously with concen-
tration and temperature. <>n the other hand, it ii • ictr< mely unlikely that when a
substance like hydro lissolves in water s chemical combination of any kind

takes pis

LEWIS. — THERMAL PRESSURE. 157

freezing point. It may be stated iu words thus : The "relative" diminu-
tion in the escaping tendency of the solvent upon the addition of an infin-
itesimal amount of a solute is equal to the ratio of the number of gram
molecules of solute and solvent. It is probable that equation (12 b),
besides being more general than the equation of Raoult, is also more
accurate; for it will appear likely from the theory here developed that
Raoult's law is exact only when the vapor of the solvent follows the gas
law, but that equation (12 b) represents a universal law.

Our problem is now to show that equations (11) and (12) are directly
deducible from the idea of thermal pressure contained in equation (10).
Since the reasoning would be the same whether our goal is the general
equation (12) or the special form of Raoult, for the sake of concreteness
it will be convenient to develop first the latter equation, and in the
simple case of a solution in a solvent to whose vapor the gas law may be
applied. Let us determine theoretically in this case the influence of the
solute on the vapor pressure of the solvent. This effect may be divided
into two which are entirely independent; the first is the effect of the
thermal pressure of the solute on the condition of the solvent (this may
be pictured kinetically as the influence of the mere motion of the solute
molecules) ; the second is the effect of the attraction or repulsion of the
solute for the particles of the solvent. It will therefore simplify the
discussion if we study these two influences separately, beginning with
the latter.

In a hypothetical case in which we may imagine the particles of the
solute to be evenly distributed through a solution, and to have no effect
except by an attraction for the solvent particles, the only practical effect
of the presence of the solute will be to increase the attractive pressure of
the solvent inward. Then, in order to maintain equilibrium, according
to equation (4),

P=(3-a,

since the external pressure is unchanged, the total volume will decrease
on account of the new attraction until the thermal pressure of the solvent
is increased by the same amount as the attractive pressure, and we may
write, therefore,

rf|3 = rfa; or, dp — da = 0.

Comparing this with equation (5) or equation (7), it is evident that the
attraction of solute for solvent is not the cause of the lowering of vapor
pressure in a solution, and that a mere attraction or repulsion between
the solvent and solute does not change the vapor pressure of the solvent,

L58 PROCEEDINGS OF THE AMERICAN ACADEMY.

b, cause the change in the attractive pressure is always compensated by
:in equal change in the thermal pressure, and these two changes produce
equal and opposite effects upon the vapor pressure. This conclusion will
simplify the discussion of the second influence of the solute, the effect of
thermal pressure ; for Bince the attraction or repulsion of solute for solvent
is \\ ithout effect, we may consider with perfect generality the case in which
tlii-. attraction or repulsion is zero. In such a case if (1 and a represent
the thermal and attractive pressures of the pure solvent, when the solute
is added in general a change in volume occurs in which (3 aud a change
to /i + d/3 and a + da. The total attractive pressure of the solution
is a -f da; but the total thermal pressure of the solution includes the
partial thermal pressure <>f the solute, which may be designated by «
The equation of the solution is, then.

P =-- (P + df3 + d&) - (a + da). (13)

Combining this with the equation of the pure solvent,

P = P-a,

we obtain

dp-da = -dp. (14)

Now the vapor pressure depends on the attractive pressure and the ther-
mal pressure of the solvent alone, in accordance with equation (7),
which may be written,

d_Px _d(p-u)

Pl mRT '

V
Substituting equation (14) and writing from equation (10),

we obtain,

(rfw) R T

d P, V dn j

/', m R T ' m

V

<l I\ being negative for a decrease in P,. This is a statement of the law
of Raoult, which is thus shown to 1"' a direct consequence of the princi-
ple of thermal pressure expressed in equation (10). Perhaps a more
intimate understanding of the way in which the thermal pressure of the
solute affects the vapor pressure of the Bolveni may be obtained from
another point of vi< w which affords a simple but somewhat less rigorous
demonstration.

LEWIS. — THERMAL PRESSURE. 159

In any pure solvent, according to equation (5), the tendency to escape
into some other phase is dependent, not upon the actual values of the
thermal and attractive pressures, but only upon their difference, which is
the external pressure. At constant external pressure, therefore, the
escaping tendency of the solvent is constant under all circumstances as
long as the external pressure represents the difference between the ther-
mal pressure of the solvent and the attractive pressure. If in a solution
the thermal pressure of the solvent were equal to the total thermal
pressure, the difference between this and the attractive pressure would
be equal to the external pressure, and the escaping tendency would be
the same as that of the pure solvent under the same external pressure.
In reality the total thermal pressure is the sum of the partial thermal
pressures of the solvent and solute ; therefore, other things being equal,
the escaping tendency should be to that of the pure solvent as the partial
thermal pressure of the solvent is to the total thermal pressure. That is,

where ft is the partial thermal pressure of the solvent, ft' that of the
solute; but from equation (10),

therefore,

From this equation,

P — r/ ' P — tz

if/2 m

{pi — i/f2 n

\j/x m -\- n

which is equation (12a).

From this it appears that the lowering of vapor pressure, the lowering
of solubility, and the lowering of the freezing point are all due to the
fact that the solute shares with the solvent the support of the external
and attractive pressure. Other things being equal, the greater the ther-
mal pressure of the solute the less that of the solvent, and thus an
increase in the amount of the solute diminishes the escaping tendency of
the solvent, whether towards the gaseous phase, the solid phase, or
another liquid.

An analogy may illustrate this explanation of Raoult's law, and at the
same time introduce the consideration of osmotic pressure. If we con-
sider m gram molecules of a gas A under a constant external pressure,

1G0 PROCEEDINGS OP THE AMERICAN ACADEMY.

its tendency to escape is measured by that pressure. If n gram mole-
cules of another gas B are now introduced, while the external pressure
remains constant, the partial pressure of A is diminished by the fraction

- of its original value and its escaping tendency will be diminished

m -f a

in the Bame ratio. Therefore gaseous solutions also obej equation (12),

and in this case the explanation is obviously the one which has been
given here for solutions in general.

It" now w<- imagine the external pressure to be produced by the pas A
outside, while within the pressure is borne by A and B, aud if we imag-
ine the outside and inside connected by a membrane permeable to A
alone, then the gas A baving a lower partial pressure inside will pass in
from the outside, and equilibrium will not be established until the partial
pressure of A is the same inside and outside ; that i-. until the total
pressure inside is greater than that outside by the partial pressure of B.
This i- an exact analogy to the osmotic pressure in solutions, and in tact
tin- Bame explanation can he given for both.*

Let us consider a solution and the pure solvent at the same external
pressure originally, connected by a semipermeable membrane. The
tendency to pas-; from solution to solvent is less than tin- tendency to
pa-- from the solvent to solution, as we have seen. The solvent will
therefore flow into the solution until, in some way, this tendency is bal-
anced. Let us suppose that it is balanced by an increase in the external
pressure on the solution. Then, since the function \p of the solvent has
been decreased by the thermal pressure of the solute, according to equa

lion i 1 I i it must lie restored to its original value by an external pressure
equal to that thermal pressure. In other words, the osmotic pressure of
a dilute solution must be equal to the partial thermal pressure of the

BOlute and. j, m

n= ;. .

which is the law of van 't Hoff. This may be shown mathematical lj as

follows : —

By generalizing equation (18) we obtain, a- the equation of the
solution, when tie external pressure upon the solution is variable,

/' • dP - (3 + (fft + f//{ - « - <!•>. (15)

- in-,- thii paper »;i- written :m exactlj similar statement <>f tliis analog}
between the osmotic pressure and the pressure which a mixture of gases would
-i.ow under the above conditions has been given by [keda, Zeit Phya Chem.,
XXXII

LEWIS. — THERMAL PRESSURE.

161

For the solvent,
Combining these we obtain,

P — a.

dP=d/3-da + dp.

For equilibrium, when the escaping tendencies of the solution and solvent
are the same,

d/3 — da = 0.

Hence the equation for osmotic equilibrium is,

dP= dp.

But dP ' = dU, the osmotic pressure, and

(dn) R T

dp =

V

from equation (10) ; hence

dn

(dn) RT

the equation of van 't Hoff.

The conclusion that a certain osmotic pressure, and an equal change in
the external pressure, together have no effect upon the tendency of a
solvent to escape into some other phase may be verified in an interesting
way.

In Figure 4 let A represent a pure liquid X ;

B, a solution in X, whose osmotic pressure is II ;

C, the vapor of X ; and D, an inert, insoluble
gas exerting a pressure equal to II. M and
M' are two membranes permeable to X alone.
Since the gas pressure in D is equal to the
osmotic pressure II, X will not pass through the
membrane M, therefore none of X will distil
from solution to solvent or vice versa, for such
a distillation would form a cyclic process con-
tradicting the second law of thermodynamics.
Hence the vapor pressure over a solution is

the same as that over the pure solvent when the solution has an additional
external pressure applied, equal to its osmotic pressure. The effect of
the thermal pressure of a solute upon the vapor pressure of a solvent
may be regarded, therefore, as due to a stress upon the surface of the
solution acting like a diminution in external pressure of the same
magnitude.

vol. xxxvi. — 11

162 PROCEEDINGS OF THE AMERICAN ACADEMY.

IV.

If further investigation and further accordance with observed facts
prose that the principle of thermal pressure, which has been here shown
to possess a high degree of probability, is in fact an exact and universal
law of nature, then there are few physico-chemical phenomena in which
thermal pressure is not one of the most important determining factors ;
yet it would be premature to attempt at present any general application
of the new theory. Bui there is oue field in which the view here advo-
cated is in direct conflict with a well-known and fruitful theory, — the
theory of van der Waals. We shall, therefore, consider briefly the rela-
tions of thermal pressure to the equation of condition of liquids and gases
and the van der Waals formula.

liquations (4) and (10) give at once a general equation of condition
for all substances,

P = —1} a, or P= a, (16)

where n is the total number of grain molecules in the vol nine J', and a,
the attractive pressure, is the only quantity which is not immediately
determined. It is probably, in most cases, a complicated function of
volume and also of temperature.

Equation (10) is directly opposed to the theory of van der Waals, in
that it does not recognize the influence upon the pressure of the so-called
covolume, but, on the other hand, it in no way contradicts the validity of
the equation of van der Waals considered as a purely empirical formula;
for if that equation is regarded merely as a statement of experimental
observation, it may be written as well in some other form. For example.

instead of

UT _„

v — b V
the second term might 1"' expanded, and the equation written,

/'= + R / j+ s+ j • • • ) - ...

V \ r- v* V* J r-

It would then be in the lorm of equation 1 16), iii which

// h h

/."/■( u

/■ I r' / V

il

This form of a would be of significance only in case the equation of
van der Waals were perfectly accurate ; as a matt< r of fact, no oue has

LEWIS. THERMAL PRESSURE. 163

claimed for it absolute accuracy, and it is probable that, except for its
plausible theoretical basis, it would already have given place to a more
accurate, purely empirical equation. If equation (16) is used in place of
the formula of van der Waals as the equation of condition of gases and
liquids, then the form of the function a must be found for each substance
separately from its empirical equation ; and on account of the presence of
this function of unknown form, it is obviously impossible at present to
test by experiment the validity of equation (16).

The curve of attractive pressure may be easily plotted in diagram on
the PV plane. On this plane the rectangular hyperbola which, in the
older theory, is only of ideal significance as the limit towards which the
equation of condition tends as the substance approximates the perfect
gas, in our present theory always represents a real physical quantity, — ■
the thermal pressure. If from the ordinates of this hyperbola are sub-
tracted the ordinates of the actual equation of condition, then, according
to equation (4), these differences may be drawn as ordinates of the curve
of attractive pressure. A general survey of many such curves will be
necessary in order to show what general laws govern the variation of the
quantity a. This task must be reserved for the future, but it may be
confidently predicted that interesting and useful relations will be found
which will give to equation (16) a specific value which it now lacks on
accouut of its great generality ; for the study of coincident conditions
shows that from the behavior of one unassociated liquid or gas we may
predict the behavior of any other, and therefore it seems eminently prob-
able from equation (16) that since /? has the same form for all substances,
the form of the substance a for any simple liquid or gas will be found to
be closely related to its form for any other simple liquid or gas.

Notwithstanding the indefiniteness given to equation (16) by our ig-
norance concerning the quantity u, in one respect the equation is explicit,
and is again in direct antagonism to the equation of van der Waals. It
follows directly from our equation that the pressure of any substance is
greater than, equal to, or less than the corresponding pressure of a perfect
gas, according as the attractive pressure is less than, equal to, or greater
than zero. In other words, whenever the volume of any phase is greater
than the volume that it would occupy as a perfect gas, a is negative,
and the total resultant force between the molecules must be repulsive.
Regarding such a repulsion, which finds no place in van der Waals'
theory, it is true that there is little positive evidence, but what evidence
there is seems to indicate decidedly the existence in some cases of some
kind of a repulsive force. If a liquid is cooled at constant pressure, its

L64 PROCEEDINGS OP THE AMERICAN ACADEMY.

volume does i lot appear to tend to become zero at the zero of tempera-
ture, but rather to approach aa a limit Borne definite volume. As the
kinetic tore- become less there must be some other force which enters
to oppose the attraction between tin ■ molecules. If it is permissible
tu consider the limiting case where the motion of the molecules cea
there must exist at the absolute zero a condition in which the total ex-
ternal pressure and all the attractive forces betweeu the molecules are
ther balanced by some sort of outward force which is equal to their
sum. This would be greater than the attractive forces alone and the
difference would depend upon the external pressure. In other words,
there would !><■ a resultant repulsive force equal to the external pressure.
A- to whether this force is of the nature of elasticity, or of some action
at a distance, it would he presumptuous to speculate. From these con-
siderations, which must be admitted to be very hypothetical, it would
seem that at ordinary temperatures there should be analogous conditions
in which the repulsive forces would he greater the higher the pressure.
According to equation 1 16), in all liquids the resultant attractive pressure
diminishes with increasing external pressure, and finally changes Bign at
the poiut where mi the P V diagram the equation of condition cuts the
hyperbola of thermal pressure; that is. at the point where the volume is
the same as it would he if the substance were to behave as a perfect gas
under the same pressure. Similarly, at high pressures probably all gas< a
have a greater volume than corresponds to the gas law. and according to
our theory their particles repel each other under these conditions. At
at spheric pressure, on the other hand, almost all ga8e8 have too small

a volume, but hydrogen still has a volume which corresponds in our
theory to an intermolecular repulsion. It is interesting, therefore, to
oote that in the experiments of Joule and Thomson, while other ga
Bhowed an increase oi internal energy on expansion, hydrogen showed a
sliL.ditdecre.a-c-. Helium is in all probability another gas which has too

: a volume: and it ha- been shown by Donnan * \\- his experiments

on the effusion of gases, that probably helium also ha- a beating effect on
free ion. like hydrogen. Such a beating effect can be explained

in no other way so simply a- by a-sumiiiLr that there i- a repulsion
between the molecules in both helium and hydrogen. Finally, a similar
repulsion could explain the phenomenon observed in the experiments
oi Ramsay \ on the distribution of hydrogen between two Bpaces, one of
which contained hydrogen alone, the other hydrogen and nitrogen, lie

I'lal. !.l\ I2i Phil Mag., XXXVII] 206 L894)

LEWIS. — THERMAL PRESSURE. 165

found that, if the two spaces were connected by a semipermeable mem-
brane, the partial pressure of the hydrogen was less in the space contain-
ing nitrogen. This could be easily explained by assuming a repulsion
between a molecule of hydrogen aud one of nitrogen, and this explana-
tion is asain in full accord with the observation of Joule and Thomson
that hydrogen mixed with nitrogen greatly decreased the cooling effect
of the latter when expanding through a porous plug. These bits of
evidence, accumulated, point decidedly to the truth of equation (16).

It was stated above that in general a is a function of the temperature;
for if either in equation (16) or in the equation of van der Waals the
term representing intermolecular attraction should be independent of the
temperature, it would follow immediately from thermodynamics that
the specific heat at constant volume of a liquid or a vapor should be
independent of its volume, and also that a liquid and its vapor should
have the same specific heat at constant volume.* All our evidence, both
experimental f and theoretical,! is opposed to this conclusion. There is
little doubt that the specific heat at constant volume, even in gases,
changes in all cases, and sometimes considerably, with an isothermal
change in volume ; and it may be noted that while the equation of van
der Waals is incompatible with any deviation in the specific heat at con-
stant volume, such a deviation is a direct consequence of our present
theory. For, since the specific heat depends upon the energy required
to raise a substance from one temperature to another, and since that
energy, according to any kinetic theory, is partly consumed in increasing
the kinetic energy of the molecules, the specific heat must depend in
part upon the kinetic energy of the molecules at the two temperatures.
According to the accepted kinetic theory the energy of progression of
the molecules must be the same at any one temperature under all condi-
tions. The energy required to increase the progressive energy of the
molecules from one temperature to another would depend, therefore, on
the two temperatures, and would be independent of all other circum-
stances. On the other hand, according to the views expressed on pa^e
154, the kinetic energy of the molecules may vary at any temperature
according to the other conditions of a substance. Let us consider a gas
in which there is no attractive or repulsive force between the molecules.
According to the theory hitherto accepted, if the volume occupied by the

* Cf. Nernst, Theor. Chem., p. 234 (1898).

t Joly, Phil. Trans. Roy. Soc, 182 A, 73 ; Proc. Roy. Soc, XLVII. 218 ; LV. 390.
t Lewis, These Proceedings, XXXV. 1 (1899) ; Zeit. Pins. Chem., XXXII. 364
(1900).

166 PROCEEDINGS OF THE AMERICAN ACADEMY.

molecules* of the gas were appreciable, then the progressive energy of
the molecules would be the same as if it were a perfect gas, but the pres-
sure would be greater than tin- pressure of a perfect gas. According to
our new theory the pressure in this case would be necessarily the same
a- that of a perfect gas, and therefore tlie kinetic energy of the molecules
must be less than that of the molecules of a perfect gas. In general,
then, we should expect the internal kinetic energy to diminish with the
volume.

This is an important consequence of the new theory, and it is evident
that the volume of the molecules must be as important in our theory as
in that of van der Waals; hut in the former the quantity "£" concerns
energy relations, while in the latter it concerns pressure relations. We
are thus led to a consideration of the total change in internal energy in
an isothermal change of volume of a liquid or a gas, and in the change
from a liquid to a vapor.

It has frequently been assumed that the energy change in such a process
is a measure of the attractive forces which oppose or assist the process.
In an earlier paper t I have shown that this is the case only when the
specific heat at constant volume remains the same ; and, in fact, it is
obvious that the change in potential energy which is a measure of attrac-
tion is in general only one factor of the total change in energy which
includes also any change in the internal energy of the molecules as well
as the change in progressive energy of the molecules, which is assumed
in our present theory.

These three factors will be designated as follows : The change in
potential energy . or free Knergy, which is the measure of intermolecular
attraction, will be represented by dX; change in the progressive motion
of tin- molecules by dE; change in the internal energy of the molecules
by '//• If '/ C is tli<' total change in internal energy,

dU=dX+dE+dl (17)

In the paper} already referred to I have developed the general ther-
modynamic equation of condition,

(18)

,. = *1- >•„,,- ..''/-, rf^.rr.

v dv J r„ I dv

* This i » J i r:i I in conformity tu usage. The quantity "6" of van der

Waali may be defined more generally and less hypothetically as a quantity depend-
ing on the difference between the time in which two molecules approach, collide,
and separate, and the time which two mathematical particles would require for
the same process.

■ Lewis, I c. I Ii>i'i

LEWIS. — THERMAL PRESSURE. 167

where F ' (v) is an unknown function of v, and cv is the molecular heat at
constant volume, T0 some arbitrary temperature. Comparing this with
equation (16), we obtain

m = rVT+%-Tfi*£*T. (19)

Now, substituting d Z7from (17), and bearing in mind that a = —=—, we
obtain

dv dv J j'0 1 dv w

This is the closest insight that we can obtain at present into the general

form of — 1- t-j but it is sufficient to show that in general the total

dv dv s

change in energy is not identical with the change in potential energy,

dcv .
at least when -7- is not zero.
dv

If we could pass continuously from the liquid to the gaseous state, and
equation (16) were assumed to hold good continuously throughout the
process, the total work of the process could be found, and would be equal
to the actual work done in the evaporation of the liquid. From equation
(16),

jpdv— I dv — jadv,

or

/"a v, Cu-

pdv = RT\n~— I adv;

dX 1 , •

'or, since a = — — , the total work is
d v

5 _ r

V2 J r,

ETlu-- ~dX.

v2

The work done in evaporation at the same temperature is P'(yl — v2),
where P' is the vapor pressure. We may write, therefore,

P\Vl -v2) = R Tin ^ - [''dX.

Now, if d X were always equal to d U, we could write

E Tin - = (Ux - U2) + P'iv, -v2) = L, (21)

where L is the common heat of evaporation.

L68 PROCEEDINGS OF THE AMERICAN ACADEMY.

Now, as a matter of feet, /? T In is always less than L, and this is

precise!; wliat would be predicted from our theory; for in passing from
a liquid to a vapor energy must be used ool only in overcoming the

intermolecular attraction ( the quantity that would be equal to R T^n — ),

hut also in providing for the increase in the progressive energy of the
molecules, which was shown on page 105 to be a necessary consequence
of our theory.

In concluding this brief treatment of some of tlie consequences of our
theory, we may mention one that is important in the study of homoge-
n< "ii- equilibrium. The isothermal, according to our present view, is not
identical with the line of constant internal energy of progression of the
molecules, which might also be called the line of constant molecular
velocity; the latter line must, therefore, l>e of considerable independent
importance ; for if we regard the mass law kiuetically a- an expression
of the law of probability, the chance ol any two molecules reacting must
depend upon their momentum at impact, and this would be constant along
the line of constant molecular velocity, and not along the isothermal.
The mass law, then, disregarding other disturbing factors, should hold
along the former line rather than the latter. On account of an entire
lack of data that would give any experimental evidence on this point, this
_ >tion oi a possible future modification in the theory of chemical
kinetics must suflice.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 10. — September, 1900.

INTERNATIONAL ATOMIC WEIGHTS.

By Theodore William Richards.

INTERNATIONAL ATOMIC WEIGHTS.
By Theodore William Richards.

Received August 18, 1900.

Near the close of the year 1897, the German Chemical Society
appointed a committee to select values of the atomic weights for common
use in Germany. The confusioD arising from the use of different values,
and especially of different standards of reference, had become unbearable-
After nearly a year's deliberation, they announced their conclusion that
it is expedient to call oxygen exactly 16, and to refer other elements to
this standard. They published a carefully considered and conservative
table of values,* which immediately gained wide acceptance, partly be-
cause of its intrinsic merits, and partly because it was vouched for by
such eminent men as Landolt, Ostwald, and Seubert. At the close of
the remarks accompanying this table, the three members of the committee
expressed their hope that the matter might be clinched by international
agreement. The hope was strengthened by the fact that the two other
modern tables, those of Clarke and of Richards, differed but slightly from
the table presented in Germany. Time has strengthened this hope still
further, for the two subsequent yearly editions of the three tables have
steadily tended toward the elimination of earlier differences, until now
they are even more alike than they were at first. |

On March 30, 1899, having heen encouraged by the favorable recep-
tion of their work, the German committee issued to all important asso-
ciations interested in chemistry throughout the world a general invitation
to appoint delegates to an International Committee, t The number of
delegates was not determined, and the outcome was the appointment of
fifty-seven men from among the most eminent chemists of eleven nations.
As representatives on this International Committee the American Chem-

* Ber. d. deutsch. chem. Gesell., 31, 3761 (1898).

t See Journ. Am. Chem. Soc, 22, 78 (1900) ; also, These Proceedings, 35, 621
(1900), and Ber. d. deutsch. chem. Gesell., 33, 1 (1900).

J Ber. d. deutsch. chem. Gesell., 31, 2949 (1898) ; 33, 1847 (1900).

17:2 PROCEEDINGS OF THE AMERICAN ACADEMY.

ical Society appointed Professors F. W. Clarke, J. W. Mallet, E. W.
Morlev, T. W. Richards, and E. F. Smith. Shortly afterward the
American Academy of Arts and Sciences added Professors Wolcott
Gibhs and Ira Remsen to the list of American delegates. Besides these
seven men, the present International Committee contains fifteen from
Germany, eleven from Austria, eight from England, five from Belgium,
three each from Switzerland and Italy, two from Japan, and one each
from Holland, Russia, and Sweden. It is much to be regretted that
Denmark, France, and Norway have as yet made no appointments.

Having thus received very general support, the German committee, in
October, 18'J9, decided upon another step. They forwarded to each
member of the International Committee a circular letter containing three
questions, which were speedily answered by nearly all of the delegates.
A literal translation of these questions follows : —

" 1. Shall O = 16 be fixed as the future standard for the calculation
of atomic weights ?

" 2. Shall the atomic weights be given with so many decimals that the
last figure is certain withiu half a unit, or what other procedure shall be
adopted ?

" 3. Is it desirable that a smaller committee should be formed, which
should undertake the continual revision of the yearly atomic weight table
and its publication ? In case of agreement upon this point, it is proposed
that each association name a single delegate to this smaller committee."

The forty-nine answers to these questions are highly interesting.* As
regards the first question, only seven chemists (one American and six
Germans) were decidedly in favor of retaining hydrogen as the standard,
while forty were decidedly of the opposite opinion. Two were willing
to accept either or both standards of reference. In spite of the fact that
men as eminent as Professors Mallet, Volhard, Winkler, and Wislicenus
are in the minority, the majority is so overwhelming that we must look
upon this point as settled for a long time. Even it the probable delegates
from the unrepresented countries Bhould all vote in the negative, the

majority must remain in favor of O = 16. TIiii- the new term, "in-
ternational atomic weLdit." is perfectly clear and unequivocal in its
meaning as to the Standard of reference, and an important step has been
made.

It may not be irrelevant here to enumerate the advantages and disad-

• Her. .1. ileutsch. clieni QflSSlL, 33. Heft 12, p ]-'.".!'. ) The answers are

published in full.

RICHARDS. — INTERNATIONAL ATOMIC WEIGHTS. 173

vantages of this international standard. In the first place, it is evident
that the question is a practical one, not a theoretical one. If Prout's
ancient hypothesis seemed at all probable, there would indeed be a strong
reason for assuming hydrogen as unity ; but Prout's hypothesis cannot
now claim serious consideration, at any rate in its original form. No
other theoretical reason for calling hydrogen exactly 1.000 is known to
me. What, then, are the relative practical advantages to be gained by
taking hydrogen or oxygen as the standard ? In the first place, the
precise quantitative analysis of compounds containing hydrogen is a very
difficult matter, and water is the only one which has been adequately
studied. Hence nearly all atomic weights must be referred to hydrogen
through the medium of oxygen ; and if the ratio H : O is found to be
even a little in error, all other values must be recalculated. Morley's
work on this ratio is indeed magnificent, and it is not likely that his
accuracy can be surpassed for a long time ; nevertheless the principle
still remains. Oxygen, on the other hand, has been directly compared
with many metals, as well as with potassic chloride and similar salts
obtainable from the chlorates and their analogues. Hence from the
point of view of directness of comparison, oxygen is to be preferred.
Silver might be even better, as Erdmann and Volhard point out in their
replies to the circular letter ; but the question does not concern the start-
ing of an entirely new system, but rather the choice between two old
ones.

Another point to be considered is the effect of the decision upon the
data contained in the past literature of chemistry. Any change which
might confuse the understanding of the work of the past would be
indeed a grievous one; and a change to O = 15.879 could not but have
this effect. Little or nothing has been written with the assumption of
this standard, while a great bulk has been written with the assumption
O = 16. The confusion caused by the inaccurate value O = 15.96 is
quite bad enough, without the introduction of a new stumbling-block.
Moreover, in the gas constant and a multitude of other physico-chemical
constants the value 0 — 16 enters, and a change in this standard would
complicate the use of a great mass of valuable literature of this kind.

Another, although somewhat trivial, reason why oxygen should be
taken as 16 is because in that case a somewhat larger proportion of the
atomic weights approximate whole numbers than would be the case
otherwise.

The chief objection to the proposed standard is a pedagogical one. It
is claimed that confusion is caused in the mind of the elementary student

174 PROCEEDINGS OF THE AMERICAN ACADEMT.

by the use of the number 1.0075.* The German committee points out
in its last report that this difficulty may be avoided by giving the ele-
mentary student only the round numbers (which suffice amply for his
purpose), accompanied by the statement that these are rough approxima-
tions. There is obviously another way of avoiding the confusion, and
that is by doing away with hydrogen as a standard of specific gravity.
The difficulty of preparing this gas in a pure state and its great lightness
are arguments against it. in any case. Moreover, in my experience the
simplicity of the relationship between the specific gravity referred to
hydrogen and the molecular weight is quite as likely to be a stumbling-
block as an assistance. Many a beginner learns by heart the statement
that the specific gravity is twice the molecular weight; for he does not
pause to think about it and see that he has inverted the ratio. If, on
the other hand, the specific gravity of oxygen is taken as the standard,
the adverse arguments disappear, and even a dull student can hardly
forget the reason why the specific gravity of the gas X referred to oxy-
gen must be multiplied by 32 to give the molecular weight. For several
years I have used this method with large classes, and find that it gives
no trouble. The only data needing recalculation are the specific gravi-
ties of the gases, and that is a simple matter. It seems to me, by the
way. that t he use of 2 instead of R in physico-chemical formulae has the
same pedagogic fault of obscuring the source and nature of the symbol.

The answers to the second international question, which seeks to de-
termine the number of decimal places to be given, support the German
committee in its position with a majority almost as overwhelming as in
the case of the first question. The minority of eight consists of three
Americans, three Germans, and two Japanese, all the others desiring to
omit all figures which are not certain to within half a unit. The com-
raittee, in Bumming np the opinions upon this subject, states thai its

desire 1S to propose a table for common use, and that the minority, whieh
desires the retention of one uncertain decimal place, has rather had in
mind the requirements ot work of the greatest precision. Undoubtedly

the curtailed table will answer for most purposes, but it. seems to me that
the nature of the decimal notation causes an unfortunate incoinplet) D

in it. Although, in the f;ice of so great a majority, this matter, like the

Other, must he considered a-, settled, I am tempted to Call attention to

this incompleteness in relation to numerical data of all kind-.

• Much ii.i- in i n « rittcn upon both lidei of this question. Besides the articles
already referred to, many references may be found in two papers l>y Kttsterand
Brainier (Zeil anorg Chem 14, 261 and 267 respective!}

RICHARDS. — INTERNATIONAL ATOMIC WEIGHTS. 175

Let us consider a concrete example, — the case of nitrogen. Few
would be willing to contend that the last figure in the number 14.04 is
certain. For the sake of argument, let us assume that the value of this
atomic weight may really be as low as 14.034 or as high as 14.046.
According, then, to the rule which has just been adopted by the Inter-
national Committee, this figure 4 should be omitted, and nitrogen should
be called 14.0. Such an omission causes an error far greater than the
uncertainty which leads to the dropping of the figure. The uncertainty
named above is only 0.04 per cent, but the minimum error in the value
14.0 is 0.24 per cent, if the lower value given above is supposed to be the
lowest possible. Clearly one must either record uncertain figures, or
else omit figures which have a real significance. The dropping of a
decimal place at once reduces tenfold the ability of a number to express
slight changes of value ; but numerical results may have any degree of
accuracy, and cannot be classed strictly into classes separated by gaps so
wide. For this reason it is a well-known practice in scientific calcula-
tion to retain during the calculation one uncertain figure, while a final
result is sometimes relieved of this uncertain figure. According to this
rule, the table of atomic weights should always give an uncertain figure
in each value ; for atomic weights are simply data for further calculation.
If this is done, the user has all the truth, and may reject as much of it as
his occasion permits. Such a table of atomic weights seems to me to be
the best, because it is capable of fulfilling all uses.

Quite another point of view should be adopted in making a table solely
for common use. Here we are concerned not with the number of cer-
tainly ascertained figures which may have been determined by Stas, but
rather with those figures which will have an influence upon the work in
hand. Usually the error in such a value is important not on account of
its absolute magnitude, but rather on account of its relation to the value
itself. In short, the percentage error is that which ought to be considered
in constructing a table of atomic weights for common use. A majority
of chemists would probibly decide that a table in which the values were
within 0.1 per cent of their true values would serve all ordinary purposes.
Most common methods are not able to attain as great a degree of accu-
racy even as this, but the admission of a wider range of inaccuracy in
a table of atomic weights might by summation cause an appreciable
error. Silver might be called 108.0, chlorine 35.5, bromine 80.0, iodine
126.9, potassium 39.2, sulphur 32.1, and lead 207.0, without seriously
affecting the results of most quantitative work, and indeed every one
frequently uses such approximate values. Even in this approximate

L76 PROCEEDINGS OP THE AMEBIC AN ACADEMY.

table hydrogen should be 1.008, and not 1.01, if the percentage rule is to
be followed.

The last question asked by the German committee, suggesting the
appointment of a small standing committee, was answered affirmatively
by every one. The original plan of having a representative from each
society would evidently result in the formation of an unwieldy body ;
hence the German committee has wisely concluded that a small number,
perhaps three, one each from Germany, England, and the United States,
should be elected from among those who have had especial experience in
the matter of atomic weights.

The matter is by no means finished. The German committee asks
any one who has anything new to say upon the questions under consider-
ation to send a brief statement of his views to Professor Landolt before
November 15 ; and some new ideas may have been advanced in the
discussion in Paris at the end of July. The balloting for the election of
the smaller International Committee is already in progress.

The German Chemical Society, as well as the members of its commit-
tee, is greatly to be congratulated on the success of the undertaking, not
only because of the immediate gain to chemistry, but also because of the
manifest advantages of the growth of scientific cooperation between men
of all nations.

Mi Desert, Maim., August C, 1000

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 11. — October, 1900.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE
MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE,
UNDER THE DIRECTION OF E. L. MARK. — No. 113.

PERIPHERAL DISTRIBUTION OF THE CRANIAL
NERVES OF SPELERPES BILINEATUS.

By Mary A. Bowers.

With Two Plates.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE
MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE,
UNDER THE DIRECTION OF E. L. MARK. — No. 113.

No. 11. — PERIPHERAL DISTRIBUTION OF THE CRANIAL
NERVES OF SPELERPES BILINEATCS.

By Mary A. Bowers. .

Presented by E. L. Mark, June 14, 1899. Received August 20, 1900.

Contexts.

A. Introduction

B. Methods

C. Eye-muscle Nerves, III., IV., and VI. .

The Oculomotor

The Trochlearis

The Abducens

D. Trigeminus

1. Roots

2. Branches

a. R. ophthalmicus trigemini . .

b. R. maxillaris trigemini ....

c. R. mandibularis trigemini . . .

E. Facialis and Acusticus

1. Roots

2. Branches

a. Ophthalmicus superficial VII. .

b. Buccalis VII

c. Acusticus

(I. Palatine

e. R. mandibularis internus VII.

(r. alveolaris)

179

180

1-1

181

181

181

182

182

1«3

183

184

184

185

185

185

1-5 S

185

186

186

1S6

/. R. mandibularis externus VII.

(r. mentalis) 186

g. R. hyomaudibularis VII. ... 187

h. R. hyoideus 188

Glossopharyngeus and Vagus .... 188

1 Roots 188

2. Branches 189

a. R. lateralis 189

b. R. supratemporalis 189

c. R. auricularis 189

(I. R. communicans IX. ad VII 190

e. R- pharyngeus 190

/. R. Ungualis IX 190

g. R. branchialis 190

h. R. visceralis vagi 191

First and Second Spinal'Nerves . . ■ 191

1. First spinal 191

2. Second spinal 191

a. Hypoglossus 191

A. Introduction.

The study of the distribution of the cranial nerves of Spelerpes biline-
atus was undertaken merely as introductory to an intended investigation
of the central origin of the nerves ; but it has seemed advisable to make
this study more careful and detailed than was at first contemplated, and
as a result the nerve roots have as yet been but superficially examined.

By studying the "components'" of the individual nerves, following in

W> PROCEEDINGS OF THE AMERICAN ACADEMY.

some degree the method adopted by Stroug ('95), it has seemed possihle
to make out more accurately the homologies of Urodelan and Auuran
nerves than has before been done.

The color scheme used by Strong has been followed as closely as
practicable ; roots and ganglia, however, have been left neutral from
lack of knowledge of the exact proportions of the different components
passing through the ganglia. Strong's nomenclature has also been
adopted, since it proves quite as applicable to the Urodelan as to the
Anuran type. The confusion which would have arisen from the use of
two or more sets of terms is thus avoided.

My work has been carried on in the Radcliffe College Laboratory at
the Museum of Comparative Zoology, Cambridge; and it is with great
pleasure that I acknowledge here my indebtedness to the Director of the
Laboratory, Professor E. L. Mark, who has assisted and advised me, and
helped me in many ways. I also wish to express my sincere thanks to
Dr. Harris II. Wilder and Dr. B. F. Kingsbury, who have kindly pre-
served and sent to me all my Spelerpes material.

B. Methods.

Several methods of preservation and staining were tried, but the mate-
rial preserved in four per cent formalin and stained with Heidenhain's
iron haematoxylin was the most satisfactory. The decolorizing process
was stopped at the moment when the other tissues had given up their
stain but the nerves still retained the deep blue color imparted to them
by the haematoxylin. It was in this way that the series which served
for the reconstructions were made. The sections were cut parallel to the

ttal plane through the left half of the head of a larva 23 mm. long,
and were 20 //. thick. Figures 1 and 3, representing the nerves of the
left side as projected on the sagittal plane, were made by outlining the
accurately superpiiM-d images of the successive sections given 1>\ an Abbe"
camera. To insure the proper Buperposition, direction planes were em-
ployed. Before sectioning, the paraffine block wascul in prismatic form,
the bounding planes being perpendicular to the prospective plane ol
sectioning : thi of the prism wire painted with a mixture of lamp-

black and turpentine; the block was then qaickly immersed in rather -.'ft
paraffine and agaiu trimmed before sectioning. The rim of black around

b section afforded a satisfactory means of accurate superposition.
Figures 2 and 1, representing projections on the frontal plane, were
constructed from the Bame series, by plotting on millimetre paper the

itiona of the nerves as they occurred in the successive sections.

BOWERS. CRANIAL NERVES OF SPELERPES BILINEATUS. 181

C. Eye-muscle Nerves, III., IV., and VI. (PI. 2, Figs. 3, 4).

It was my intention to confine my work to the fifth, seventh, ninth,
and tenth nerves, but in my examination of the ophthalmic branch of the
trigeminus it became necessary to make a study of the eye-muscle nerves,
from which some facts of interest were established. I therefore give
a brief account of these nerves, — the third, fourth, and sixth cranial
nerves.

The oculomotor (III.) arises, as usual, from the floor of the mesen-
cephalon, passes through the brain wall, and then, lying between that
wall and r. ophthalmicus V. ( V. opt.), immediately gives off its r. supe-
rior (PI. 2, Figs. 3, 4, III. rt. sit.) to m. rectus superior. It then passes
under r. ophthalmicus V., — which takes a more median position (Fig. 4),
— and here lies for a short distance in close contact with a branch of the
abducens (VI.) In some series of sections the two nerves, while indis-
tinguishable on one side of the head, were clearly separable on the other.
There is, then, no real fusion of the two.

A cluster of ganglion cells (cL gn.) is constantly found enveloping that
portion of the oculomotorius that lies directly ventrad to the optic nerve
(see Figs. 3, 4). In the 23 mm. stage these ganglionic cells are grouped
into a compact mass, but in older stages (40 mm.) they are somewhat
scattered along the nerve. The oculomotorius follows closely m. inferior
rectus, giving fibres to it, aud ends in m. inferior obliquus.

The trochlearis (IV., PI. 2, Figs. 3, 4) is very minute, and its whole
course was not traceable in the 23 mm. stage ; but in older larvae an
interesting condition was observed. At the posterior margin of the eye-
ball this nerve joins a dorsal branch of r. ophthalmicus ( Va), and the
two run along the median dorsal surface of the eyeball to near its ante-
rior margin as one nerve. Here the fibres of the trochlearis (IV. ob. su.)
pass ventrad and are distributed to the superior oblique muscle ; this
ophthalmic branch of the fifth ( Va) then passes forward and dorsad to
innervate the skin in the regions in front of the eye. Since IV. joins the
dorsal branch of V. on its dorsal side and separates from it on the ventral
side (Fig. 3), there must be a crossing of fibres; this crossing I found
clearly shown in sagittal sections of one of my older embryos. Gaupp
('97, p. 136) speaks of this relation of IV. and V. in the frog, and pre-
sumes that the branch given off to m. superior obliquus is composed of
fibres from IV., but he did not actually observe the crossing of fibres.

For the study of the abducens (VI.), Spelerpes proved to be especially
advantageous. In this species the abducens does not enter the Gas-

182 PROCEEDINGS OF THE AMERICAN ACADEMY.

serian ganglion, as it does in some of the Amphibia, but (Fig. 3) comes
into contact with the ventral side of r. ophthalmicus V. ( V. opt.) a short
distance anterior to its emergence from the ganglion. Owing to this
condition it is much easier to follow the fibres of the two nerves in
this species than in those where the two nerves emerge from the gan-
glion together. llerrick ('94, p. 200) describes for Amblystoma two
branches of the sixth nerve, one of which goes to m. rectus externus,
the other to m. retractor bulbi, but he thinks the branch to the latter
muscle ought really to be assigned to the trigeminus, and so colors it in
his figures. Although in Spelerpes the sixth lies in immediate contact
with p. ophthalmicus V., yet I was able to trace its fibres with accuracy
in several series of preparations, and am certain that there is no ventral
branch given otf from V., but that VI. divides into two branches, one
( VI. rt. ex.) going to m. rectus externus, the other ( VI. ret. bl. ) torn,
retractor bulbi. On the latter branch was found, in the region indicated
by cl.gn'. (PI. 2, Figs. 3, 4), a distinct aggregation of ganglionic cells,
similar to those enveloping the oculomotorius. I have found no refer-
ence to such cells connected with the sixth nerve, except in one sentence
of Strong's article ('95, p. 134), where he says: "There seem to be
ganglion cells in connection with it [abducens] (703), although these
may belong to the oculomotor nerve." I have not yet tried methods
to show the connection of these ganglionic cells with the nerve fibres.
Oidy two roots were found, as in Necturus ; these emerge from the
ventral side of the medulla in about the same transverse plane as the
most anterior roots of IX. + X.

D. Trigeminus.

(1) Roots. — The little study which I have hitherto given to the roots
of the cranial nerves in Spelerpes has shown that there is apparently
a close correspondence to the condition found by Kingsbury ('95.) in
Necturus.

Fibres forming the trigeminus root are: (1 ) a large ascending bundle
(Plate 1, Fig. 2, V. >:r.) of mostly small fibres lying ventrad to the
eighth nerve, presumably tin- ascending tract of tin- fifth, though they
were not traced back into the dorsal column of the spinal cord. It is
possible that fibres from a sensory nidus (terminal) may lie associated
with these, as in Necturus; but none uric distinguished; (2) two small,
presumably motor, bundles of large, deeply staining fibres ( V.rx.mot.),
which arise from tin- (lorn- of the metencephalon ami join the ventral
Bide of the ascending tract of the fifth, just before its exit from the brain ;

BOWERS. — CRANIAL NERVES OF SPELERPES BILINEATUS. 183

(3) coarse fibres from the roof of the mesencephalon (not shown in the
Figures), which curve ventrad and caudad and pass out of the brain
with the other roots of the fifth. After emergence from the brain, the
trigeminal fibres run obliquely cephalad and laterad, as indicated in
Figures 1 and 2 (PL 1), and can be traced through the ventral half of
the Gasserian ganglion into rami ophthalmicus, maxillaris, and man-
dibularis. The fibres are of medium size and in iron haematoxylin take
a grayish blue stain, in sharp contrast to the large fibres from the seventh
nerve, which take a deep blue stain. The latter run through the dorsal
part of the ganglion, from which they emerge as r. ophthalmicus super-
ficialis VII. and r. buocalis VII.

(2) Branches. — (a) Ramus ophthalmicus trigemini (Figs. 1-4, V.
opt.) leaves the anterior mesial part of the Gasserian ganglion, runs
directly cephalad, and comes in contact with III. and VI., as already
(pp. 181 and 182) described. Just posterior to the eye, in the transverse
plane in which the optic nerve emerges from the cartilage of the brain
wall, the ophthalmicus trigemiui gives off a large dorsal branch ( Va,
Figs. 1, 3, 4). From this branch arise two small branches Va1 and
Va2 (Fig. 3). The more posterior and lateral of the two branches
( Va1) follows m. rectus superior to its insertion on the eyeball.

The other branch ( Va2) goes to the skin of the dorsum ; it divides,
sending one branch cephalad and another caudad. The main branch
( Va) curves along the dorsal median surface of the eye (Figs. 1, 3, 4)
in connection with IV. (Fig. 4), as previously described, and is dis-
tributed to the skin in the region in front of the eye.

Ramus opthalmicus V., after giving off the dorsal branch ( Fa) just
described, takes its* usual course forward, above the optic nerve (Fig. 3),
close beside m. rectus internus, and divides into three branches (Figs.
1-3). The most ventral of these ( V. I. na.) curves around in front of
the eye to the skin of the external nares and cheek ; it corresponds to
Gaupp's r. lateralis narium. The middle one of the three branches curves
ventrad and anastomoses (corns.) with r. palatums VII. This condition
agrees with that found by Herrick in Amblystoma. Strong ('95, p. 122)
also finds an anastomosis of these two branches in the tadpole, but it takes
place farther cephalad than in Spelerpes. The anastomosing nerves
continue cephalad after their union, but could be traced only a short
distance in the loose tissue in the roof of the oral cavity, just median to
the internal nares. The most dorsal of the three branches ( V. m. na.) of
the ophthalmicus (Gaupp's r. medialis narium) runs above the olfactory
and innervates the skin at the tip of the nose.

184 PROCEEDINGS OF THE AMERICAN ACADEMY.

(//) Ramus maxillaris trigemini emerges from the Gasserian gan-
glion in the same transverse plane with r. buccalia and r. ophthalmicus
Bnperficialia VII., and the three appear | Fig. 2) as one nerve for a short
distance, as they pass laterad and cephalad between m. temporalis and
m. masseter. While ramus ophthalmicus Buperficialis VII. separates from
the others and curves mesiad, rami maxillaris V. and bnccalis VTI. pass
cephalad along the dorso-lateral surface of the masseter muscle. In
many of the specimens examined the two appear like one nerve, hut the
difference in the size and distribution of the fibres was always noticeable,
and in the series of sagittal sections used for reconstructions they were
clearly separate. R. maxillaris ( V. mx.) gives off a small general
cutaneous branch, not shown in the drawings, to the dorsum immedi-
ately after leaving the ganglion, and its final distribution is to the skin
of the cheek. It does not anastomose with ramus palatiuus VII., as is
the case in the tadpole of Rana (Strong).

(c) Ramus mandibularis trigemini ( V. md.) leaves the ganglion
directly below ramus maxillaris, passes ventro-laterad, gives nil first a
branch | J'/i) to m. masseter, next, two more small brandies (not
figured) to the same muscle, and then a fourth { J y), which runs
dorsad to m. temporalis. In the same transverse plane with I y there
aii>es a cutaneous branch ( 1 7,), which runs ventro-laterad and is dis-
tributed to the skin posterior to the angle of the jaw. It comes into
very close connection with the more anterior branch of ramus man-
dibulars externus VII. ; in some series of sections it even appears to
anastomose with this branch of VII. ; but other sections show conclu-
sively that it does not. Another general cutaneous branch ( JY) is
given (iff from ramus mandibularis V. to the angle i>f the jaw, a small
twig from it turning caudad; the main branch | I'd. however, runs
forward, close above ramus mandibularis externus VII. to the skin of
the lower lip. A division of the main oerve into nearlj equal parts
| I. nut. i, and J '. md. ex.) soon occurs. A musculo-cutaneous branch
( I '. md. i ). Gaupp'a ramus mandibularis internus V., passes ventrad
between os dentale and .Meckel's cartilage to the lower Bide of the jaw,

where it innervate^ m. mylohyoideus and the skin superficial to it.
The other branch | I'. ///'/. ex.), which is purely cutaneous, remains on the

upper -id>- of the jaw and is distributed to the skin of the lower lip. It
corresponds to Claupp's ramus mandibularis externus V.

BOWERS. — CRANIAL NERVES OF SPELERPES BILINEATUS. 185

E. Facialis and Acusticus.

1. Roots. — Apparently the roots of the facial and acoustic nerves
(PI. 1, Figs. 1, 2) agree with the condition in Necturus as described by
Kingsbury ('95) ; at least the relation of the several components as they
emerge from the brain is the same. Most caudal and ventral is the
motor root of the facial ( VII. rx. mot.),* formed from two rootlets of
deeply staining fibres, which arise from the median ventral region of the
medulla. Dorsal to this emerges the one large root of the acoustic
( VIII rx.), and close to the dorsal side of the latter the fine unstained
fibres of the fasciculus-communis root of the facial ( VII rx. fas. com.).
At a little distance dorsad to these three roots and a little more cephalad
emerge the very coarse, deeply staining fibres of the lateral-line root
of the facial ( VII. rx. In. /.).f This is the "dorsal VII." of Strong.
The fibres of this root are not grouped into two distinct rootlets, as in
Necturus, though from the appearance of some sections there seems to
be a tendency toward such a grouping. Immediately on leaving the
brain the fibres of this component unite with those of the three roots
ventral to it, forming a continuous dorso-ventral sheet (Fig. 1).

The dorsal root (VII. rx. In. I.), however, quickly divides, and its
more ventral fibres curve veutrad and laterad to enter the acustico-
facialis ganglion (gn. ac.-fac), while its more dorsal fibres pass directly
cephalad along the mesial surface of the ear capsule in the form of a thin
dorso-ventrally expanded sheet (compare Fig. 2), and enter the Gas-
serian ganglion.

This dorsal part of the lateral-line root of VII. contains ganglionic
cells, which lie dorsal to those of the Gasserian ganglion proper, with
which, however, they become fused into a common mass. Nevertheless,
the fibres of this portion of the lateral-line root are clearly traceable
in their passage through the Gasserian ganglion, from which they emerge
as two.

2. Branches. — (a) R. ophthalmicus superficialis ( VII. opt. suf.)
and (b) r. buccalis ( VII hue ) of the seventh nerve. Both of these
branches are distributed to lateral-line organs, the former giving off
numerous twigs throughout the whole of its course from the ganglion to
the tip of the nose. Ramus buccalis VII. has already been referred to
as following closely the course of r. maxillaris V. (see Fig. 2), though

* By an oversight this is lettered in Fig. 1 as VI. rx. mot.

t Not lettered in Fig. 1. In this figure it is almost directly below the lettering

186 PROCEEDINGS OF THF. AMERICAN ACADEMY.

it .-xtends farther anterior than the latter, (c) TJie oensticus. Three of
the brandies of the eighth nerve have been represented in the drawings
in neutral tint, — raraulus ampullae anterioris (rml. amp. a.), ramulus
ampullae externae {rml amp. ex.), and ramulus ampullae posterioris
{rml. amp. p.). These were plotted simply to show their relation
to the seventh nerve. Fibres to the macula ac. sacculi and macula
ac. neglecta were observed, but these were only slightly differentiated
at the stage here reproduced. They could be clearly traced in a larva
10 mm. long. The anterior portions of VII. and VIII. remain con-
nected with each other until after r. palatums VII. is given off; the
acoustic then curves dorsad and enters the ear capsule, while the faci-
alis curving ventrad and ectad passes out through the cartilage of the
skull.

The first of the branches to be given off from the ventral part of the
facialis is the (//) r.palatinus, whose fibres are derived exclusively from
the fasciculus-communis component. The ganglion of this root, lying close
to the point where r. palatums emerges, is fused with that of the eighth
nerve and "represents probably the geniculate ganglion " (Kingsbury, '95,
p. 1ST). The fibres of this component are always easily recognizable,
since they are very fine and with iron haematoxylin take only a faint
stain, or sometimes none at all. The greater portion of the fibres of the
fasciculus communis goes to form this branch (r. palatums), which passes
directly cephalad, giving a few fine branches to the roof of the pharynx,
and anastomosing, as already stated, with r. ophthalmicus V.

The few remaining fibres of the fasciculus communis pass out through
the ventral wall of the ear capsule with the hyomaiidibular portion of the
facialis, run through the ganglion of the lateral-line component of
the facialis, which lies beneath the cephalic end of the otic capsule, and
emerge a- ('•) /•. mandibularit internus !//.(! //. //"/. i. ), or /-. alveolaris
I'll., which Strong ('95, p. 187) homologies with tin' chorda tympani

of higher vertebrate.. In Spelerpes ibis branch is very minute, and in
tip' sagittal section, used for reconstruction could be traced only a short
inC8 anterior to tin' articulation of the jaw (Fig8. 1. '-' » ; but in trans-
verse sections it could be traced along the inside of the jaw to it.s termi-
nation in the ventro-lateral wall of the pharynx. It does not. in any
pari of its course, enter a canal, agreeing in this respect with siren la*
crttina as described by Wilder t'9i).

(/'.) /,'. mandibularis extemtu VII. { VII. md. ex.) , or r.mentalia VII.,
tin- ventral part of the lateral-line c ponenl of tic seventh, passes ven-
trad and cephalad from ii jlion, which lie- under the otic capsule,

BOWERS. — CRANIAL NERVES OF SPELERPES BILINEATUS. 187

and soon divides into two nearly equal branches, both of which first run
ventrad into the lower jaw, then cephalad, nearly parallel to each other,
to the tip of the chin, supplying the lateral-line organs along their
course. One of these ( VII. md. ex.') is more ventral and median than
the other ( VII. md. ex.). Strong ('95, p. 134) says of this branch in
Amblystoma : " It is a cutaneous nerve ; and probably, as in the tad-
pole, supplies the lateral sense organs." Its ganglion is " composed of
large ganglion cells." Herrick ('94, p. 199), however, found no gan-
glion in the course of the r. mandibularis externus in Amblystoma.

Strong ('95, p. 163) has called attention to the fact that the r. buc-
calis of von Plessen und Rabinovicz is misnamed, that it corresponds to
r. mandibularis externus VII., and is derived from the ventral division of
the lateral [i. e. lateral-line] root. Apparently a further interpretation
of their designations must be made, for their " Begleiter des R. hyoideo-
mandibularis " (their ramus hyoideo-mandibularis and by inference
its " Begleiter" are distributed to the skin of the cheek and the lower
jaw) also clearly belongs to the lateral-line component. It is probable
that, while their r. buccalis corresponds to one of the two branches (the
more lateral, VII. md. ex.) of r. mandibularis externus VII. as found in
Spelerpes, their "Begleiter" corresponds to the other (the more nearly
median, VII. md. ex.') branch.

Herrick ('94, p. 199), following too closely von Plessen und Rabin-
ovicz, has repeated for Amblystoma their mistake concerning this " Be-
gleiter." He describes the ramus buccalis and the accessory ramus
hyomandibularis ("Begleiter") as following the ectal aspect of the jaw
bone to the tip of the lower jaw, and as being distributed to the skin. He
says : "The fibres of the ramus buccalis are mainly, if not wholly, derived
from the dorsal root, as Strong has pointed out. The nerve which I have
called the accessory hyomandibular seems to be the same as Strong's 'small
branch to lower jaw,' which he derives from the fasciculus communis
and considers the representative of the chorda tympani of higher forms."
But obviously the homologue of Strong's " small branch to lower jaw "
is not Herrick's accessory hyomandibular, but his r. alveolaris.

(g.) R. hyomandibularis VII. in Spelerpes is very short, as com-
pared with its length in the tadpole. In the latter, it is prolonged, as it
were, almost to the anterior margin of the eye before its first branch, r.
hyoideus VII., is given off. The condition in Spelerpes corresponds
more nearly to that in the adult frog. Presumably the 23 mm. larva
of Spelerpes undergoes much less alteration in the course of its further
development than does the tadpole, so that the two stages are not

lss PROCEEDINGS OF THE AMERICAN ACADEMY.

directly comparable. Evidently the name r. hyomandibularis must be
restricted in Spelerpes to tin - Bhorl trunk of the ventral part of the
Beventh, which is included between the region of its separation from the
eighth and that of the giving off of r. hyoideus. Though much shorter
than in the tadpole, it is made up of the same components, except that
it docs not embrace the general cutaneous. This constituent is repre-
sented exclusively by the communicating branch from IX. to VII.
Since in Spelerpes this r. communicant IX. ad VII. is not received by
the hvomandibular trunk of seveu, but by r. hyoideus after its separa-
tiou from the trunk (Fig. 2), it is probable* that no general cuta-
neous fibres are included in the hyomandibularis proper. This, however,
is not a fundamental difference, being merely a question of the earlier
or later accession of r. communicans to branches of the seventh. In
one point, only, is a greater importance to be noted: r. mandibulars
ext. in Spelerpes probably does not contain, as in the tadpole, a genera]
cutaneous component.

The (ft) r. hyoideus after receiving r. communicans from IX.-fX,
curves latero-ventrad and is distributed to m. digastricus and m. mylo-
hyoideus posterior and to the skin ventral to them. The main branch
(VII. hoi.) is figured iu neutral tint, since the proportion of the two
components was not accurately ascertained.

F. Glossopharyngeus and Vagus.

1. Hoots. — The roots of this group, like those of the seventh and
eighth, show apparently a close correspondence to the condition in
NectuniB, though a careful comparison will be necessary to determine
this with certainty. The most cephalic and most dorsal rout is the
lateral-line component (PI. l, Figs. 1, 2, /A".1+2); it resembles dorsal
VII. in appearance and position. In this case, however, the rout is
composed of two bundles. It is equivalent to Kingsbury's /A.'-' and
tn Strong's first root of IX.-f X.

The second rool (IX. M) emerges one section (20 ^ thick) caudad
and Blightly ventrad of the first root ; it is composed of < I ) the char-
acteristic line colorless fibres of the fasciculus-communis component and
(2) a ventral bundle, presumably motor. It corresponds to Kingsbury's

/ X. ' and t'i Strong's BeCOnd root.

The third rool (A.1), the equivalent of Kingsbury's X.\ emerges

• It i- cf course (?) thai tome fibre* from the r communicans take a

centripetal course in the r. byoidetu and thus reach the hyomandibular trunk of

VII

BOWERS. — CRANIAL NERVES OF SPELERPES BILINEATUS. 189

at some distance caudad of the preceding and is made up of three com-
ponents: (1) most dorsal, fine fasciculus-communis fibres, (2) more
ventral and caudad of fasc. com. a large bundle of coarser fibres (ascend-
ing X.), which correspond to those of ascending V., and (3), in this
differing from Necturus, a ventral bundle (motor ?). The third root
in the tadpole has, according to Strong, the same triple composition, but
in that animal the fasciculus-communis fibres emerge ventral to those of
ascending V., not cephalad of them, as in Spelerpes.

The remaining four roots are small and could not be traced in the
sagittal series of the 23 mm. larva ; but they were plotted on the
frontal reconstruction (Fig. 2) from frontal sections of another 23 mm.
specimen, and the results were checked by the study of other series
of (transverse) sections. These four roots all appear to be motor, for
they arise in the same horizontal plane from ventral fibres, which turn
cephalad after emerging from the medulla. Kingsbury's X.'2, X.3,
and X.b are in his opinion motor, but he says that X4 is probably
sensory, and that its fibres accompany those of ascending V.

2. Branches. — The coarse fibres of the first root in Spelerpes can be
easily traced through the upper part of the ganglion IX.+X., and all
but a few of them, which are given off dorsally (see Fig. 1, rm. su'tp.),
pass out at the posterior end of the ganglion as the lateral-line nerves —
(a) ramus lateralis — to be distributed to the lateral-line sense organs of
the body.

The remaining coarse fibres form a branch (rm. su'tp.) which passes
dorsad and ectad and divides into two small branches which innervate
sense organs just posterior to the ear. Following Strong's nomenclature,
this may be called (b) ramus supratemporalis, though it does not curve
cephalad, as in the tadpole. It may be noted in passing that the " ectad
tendency" of the cranial nerves in Spelerpes, as compared with the
" cephalad tendency " in the tadpole, is a noticeable difference between
the two forms, and is probably due to the more anterior position of the
gills in the tadpole.

There is a bundle of general cutaneous fibres (rm. aur.), which leaves
ganglion IX.+X. in company with r. supratemporalis, from which,
however, it immediately separates and runs dorso-cephalad to the skin
above the ear capsule. It corresponds in composition and distribution
to the branch called by Strong (c) ramus auricularis, and known in the
frog as r. cutaneus dorsalis.

From the cephalo-lateral portion of the ganglion there pass out three
branches, which for a short distance are united iuto a single trunk. The

i'.tO PROCEEDINGS OF THE AMERICAN ACADEMY.

tirst branch to separate from the trunk is (d) ramus communicans ad
facialem {rm. comn, IX. -VII.), which follows the latero-ventral surface
of the ear capsule till it unites with r. hyoideus VII., as already de-
scribed. Strong ("95, p. 130) gives reasons for considering this branch a
general cutaneous ; its relation to VII. in Spelerpes would seem to add
evidence of the correctness of this view.

The second branch, (e) ramus phargngeus (rm. phg.), given off from
the main trunk is very small, and is composed of the fine unstained
fibres of the fasciculus-communis component ; it was traced to the roof
of the pharynx, where fibres were seen to pass down to the end buds.
This branch was traced in transverse series, and likewise in sagittal
series of larger heads, as far cephalad as the separation of r. palatums
VII. from its ganglion, but in the sagittal sections of the small individual
used for reconstruction it could be followed only as far as indicated in
Figures 1 and 2.

The third branch of this group, (/) ramus lingualis (IX. rm. lug.),
passes ectad, gives off a small motor branch to m. cerato-hyoideus
externus, then runs ventrad to the under side of the first epibrancbial
bar, then curves cephalo-mesiad and is traceable in transverse sections,
though not in sagittal ones, to the sense organs of the dorsum of the
tongue. The fibres have the same appearance as those of all the Other
branches of the fasciculus-communis group.

Another branch of fasciculus-communis fibres (</) ramus branckialis
(rm. brn.), leaves ganglion IX.+X. immediately caudad of the com-
mon trunk of the three branches last described. It appears to give
some fine branches to blood-vessels soon alter its emergence, but this
could not be determined with certainty. The trunk soon divides; the
more anterior branch goes to the first and second gills, the posterior to
the third gill. It was very difficult to follow the fibres, and their dis-
tribution could not be satisfactorily worked out. Minute branches \-

n off daring their coarse to the gills, but could not be traced to their
terminations! A few motor twigs to gill muscles could be distinguished,
a-; they were more deeply stained, and from the posterior branch general
cutaneous fibres were seen to pass to the skin. The fine fasciculus-
communis fibrec were followed to the under Bide of the gill bars, where
they lie close to large blood vessels, but they could not be traced further.
Though their ultimate distribution could not be learned, it seems to me
allowable to homologize them with the rami branchiales of the tadpole.
Their caudad rather than cephalad course is accounted for by the more

iterior position of the gills in Spelerpes.

BOWERS. — CRANIAL NERVES OF SPELERPES BILINEATUS. 191

The last branch from the IX. -f X. complex to be described is (k)
ramus visceralis (X. rm. vsc). It has been left in neutral tint to the
point of branching, for the same reason that was given for r. hyoideus.
It leaves the latero-caudal angle of the ganglion ventral to the lateral-
line nerves, and curves ventrad. It is composed of deeply staining and
of unstained fibres. Part of the latter are given off in the first branch
(X. rml. vsc. oe.), which was followed caudad to the region of the oesoph-
agus. Other branches, not plotted in the drawings, were given off
mesiad from the region in which the nerve separates into two motor
branches (Fig. 1). One of these motor branches (rm. lar.) curves
cephalad, followiug closely the course of the hypoglossus for a short dis-
tance. It is distributed in part to a transverse sheet of muscles below
the pharynx, which Wilder ('91, p. 188) interprets as resulting from a
fusion of digastricus pharyngis and dorso-laryngeus. Spelerpes, being
a lungless form, has no use for the laryngotracheal muscles as such.
Wilder maintains that they are employed to form this pharyngo-oeso-
phageal sheet. The distribution of the remaining portion of this branch
is to the constrictor arcuum branchiarum. This branch would therefore
seem to be r. laryugeus, although it issues from the ganglion with r.
visceralis.

The other motor branch (rm. scap. ?) was traced caudo-ventrad to the
region of the shoulder girdle, and is perhaps the r. scapularis of Fiir-
bringer.

G. First and Second Spinal Nerves.

(!) The first spinal nerve in Spelerpes bilineatus arises by two ventral
roots, which run cephalad a short distance after they emerge from the
brain, then pass through the cranial cartilage and divide into two
branches, — a dorsal (rm. d. spi-i), which goes to m. longissimus dorsi,
and a ventral (rm. v. s/n'.i), which curves caudad, ectad, and then ventrad,
finally corning into such close relation with the hypoglossus that the two
appear, even when cut transversely, as one nerve. The two remain
together for a short distance, then the first spinal curves mesiad and
cephalad ; it was traced to the mesial surface of m. sterno-hyoideus, but
fibres were not seen to actually enter that muscle.

(2) The second spinal nerve was found to have a ventral root (rx. v.
spi.2) similar to that of the first, and also a well-developed dorsal root
(rx. d. spi.2). Kingsbury ('95. p. 149) says, "The hypoglossus in Am-
phibia is generally described as formed by the union of the ventral trunks
of the two nerves arising immediately caudad of the vagus group, to

192 PROCEEDINGS OF THE AMERICAN ACADEMY.

which dorsal roots are wanting." He refers, however, to the fact that
an indication of a dorsal root for the first spinal was found by Kingsley
('92, p. 678) in Amphiuma and by Mrs. Gage ('93, p. 275) iu the larva
of Diemyctylus.

In Spelerpes the two roots of the second spinal nerve unite into a
ganglion, from which were traced a dorsal (rm. d. spi.2) and a ventral
branch (h'gls.). The former has been represented as entirely motor,
because no general cutaneous fibres were actually demonstrated, although
the nerve could always be followed to the narrow space betweeu the
dorsal muscles and the overlying skin. The ventral branch is the hypo-
glossal nerve, which runs cephalad and supplies m. sternohyoideus and
m. genio-hyoideus.

Von Plessen und Babinovicz ('91, p. 20) describe the hypoglossus in
Salamandra as a branch of the first cervical. No dorsal root was found,
but they describe and figure a ganglion connected with the dorsal branch.
Kingsley ('92, p. 678) mentions the persistence of the dorsal ganglion
as a noticeable feature in connection with the twelfth nerve.

BOVVERS. CRANIAL NERVES OF SPELERPES BILINEATUS. 193

Papers referred to.

Gage, S. P.

'93. The Brain of Diemyctylus viridescens from Larval to Adult Life, and Com-
parisons with the Brain of Amia and of Petromyzon. The Wilder Quaiter-
century Book, Ithaca, N. Y. pp. 259-299. 8 pis.
Gaupp, E.

'97. A. Ecker's und R. Wiedersheim's Anatomie des Frosches. Abth. 2, Halfte 1.
Lehre vom Nervensystem. Aufl. 2. 234 pp. 62 Fig.
Herrick. C. J.

'94. Studies from the Neurological Laboratory of Denison University. XI. The
Cranial Nerves of Amblystoma. Journ. Comp. Neurol., Vol. 4. pp. 193-207.
Pis. 19, 20.
Kingsbury, B. F.
'95. On the Brain of Necturus maculatus. Journ. Comp. Neurol., Vol. 5. pp.
139-203. Pis. 9-11.
Kingsley, J. S.

'92. The Head of an Embryo Amphiuma. Amer. Nat., Vol. 26. pp. 671-680.
Plessen, J. von, und J. Rabinovicz.
. '91. Die Kopfnerven von Salamandra maculata Ira vorgeriickten Embryonalsta-

dium. Miinchen. J. F. Lehmann. 20 pp. 2 Taf.
Strong, O. S.

'95. The Cranial Nerves of Amphibia. A Contribution to the Morphology of
the Vertebrate Nervous System. Journ. Morph., Vol 10. pp. 103-230. Pis.
7-12.
Wilder, H. H.
'91. A Contribution to the Anatomy of Siren lacertina. Zool. Jahrb., Abth. f.
Anat u. Ontog., Bd. 4. pp. 653-696. Taf. 39, 40.
Wilder. H. H.
'96. Lungless Salamanders. Second Paper. Anat. Anz., Bd. 12, No. 7. pp.
182-192. 7 figs.

VOL. XXXVI. — 13

. .

Bowers :rves Spelerpes.

Plate 2.

£

— — —9—

MAB.de;

EXPLANATION OF PLATES.

All figures are reconstruction drawings from sagittal sections of Spelerpes bi-
lineatus.

The colors employed follow the scheme adopted by Strong, viz.:
Black for lateral line (also for optic n.), Pink for fasciculus communis,
Yellow for general cutaneous,. Blue for motor.

//. . .
///. . .
///. 06. inf.
III. rt. i.
III. rt. inf.

III. rt. su.

I V. ob. su.
Va . .
Va\ . .
Va* . .
V0 . .
Vy . .
T5 . .
Ve . .
V.l.n. .

V. I. na. .
V. m. na.
V. md. .
V. md. ex.
V. md. i.
V. mx. .
V. opt. .
V. rx.

V. rx. mot.

VI. . .
VI. ret. hi.
VI. rt. ex.

VI. rx. mot

VII. buc.
VII. hoi.

VII. hoi-md.
VII. md. ex
VII. md. ex

List of Abbreviations.

Opticus.

Oculomotorius

Branch of III. to m. obliquus inferior.

" " " rectus internus.

•' inferior.

" " " " superior.

Trochlears to m. obliquus superior.
Dorsal branch from ophthalmicus trigemini.
Small branch of ]'a. (follows m. rectus superior).

" " " (to skin of dorsum).

Branch of r. mandibularis V. (to m. masseter).

" " " " (to m. temporalis).

" " " (to skin posterior to angle of jaw).

" (to skin of lower lip).
Erroneously engraved in Fig. 1 for V. I. na.
Ramus lateralis narium.

" medialis "

" mandibularis trigemini.

" " " externus trigemini.

" " " internus "

" maxillaris trigemini.

" ophthalmicus trigemini.
Ascending root of the trigeminus.
Motor
Abducens.
Branch of VI. to m. retractor bulbi.

" " " rectus externus.

Erroneously engraved in Fig. 1 for VII. rx. mot.
Ramus buccalis facialis.

" hyoideus "

" hyomandibularis facialis.
Branch of ramus mandibularis externus (or mentalis) facialis.

EXPLANATION OF PLATES.

VII. mil. i.
VII. opt
VII. pat. '.

\ 1 1 . i t. fas. I ■

VII. rx.ln.L

VII IT. tliOt.

VIH.rx.

IX--. .

IX +* . .

IX. im. In'/.

A'1 — A

A

A mi. v* i"

A / ID •

II

cl. rjn.

rl. ;,„.' . .

corns .

I X A
an. ac-fac

i/ii. (ins. . .

gn. -/•>.. . .

fi'i/ls. . . .

/.'....

Ill

/'

i in. our. . .
i in. brn. .

omn. IX - I'll

!. spi.y .

Tin. d. tpi

i m. I. .
i ml. amp. a.
rml. 11111/1 i i
i ml. amp. /■.

rin. Inr. .
i in. /ih i/. .

mi. pry. .

/ m tu'tp

i in. •■ tpi.y
■
pi. i .

Ramus mandibular^ internua (or alveolarie) facialis.
ophthalmicus superficialis facialis.
palatums facialis.

Pasciculus-communis runt of the facialis.

Lateral-line root of the facialis.

.Motor root of the facialis.

Root of auditory ner

Anterior root of the ninth cranial nerve.

Posterior

Ramus lingualis glossopharyngei.

Roots ol the tenth cranial nerve.

Branch of visceralis vagi to the oesophagus.

Erroneously engraved in Fig. 1 for X rml. vsc. or.

Ramus risceralis vagi.

Anterior.

Ganglionic cells of oculomotorius.
" " " abducens.

Commissure between the r. palatinus VII. and the r. ophthal-
micus V.

Ganglion of the glossopharyngeos and the vagus nerves.

Ganglion acustico-facialis.

Gasserian ganglion.

Ganglion of second Bpinal nerve.

Bypoglossus nerve.

Lateral

Median.

Posterior.

Ramus auricularis.
branchial is.
" communicans glossopharyngei ad facialem.

Dorsal branch of first Bpinal. Notk. — In Pig. 2 the dotted
line has been omitted ; it should have run to the left and
downward from the letters rm.

Dorsal branch of second spinal.

Ramus lateralis.

Hamulus ampullae anterioris.
" " externae.

" " posterioriB.

laryngeus.
" pharyngeui

Erroneously engraved in Pig. 1 (or rm.phy.

Ramui icapularii ( ! i
■upratemporall

Ventral branch oi Brat spinal.

I h i ial rout ol lecond spinal.

Root of lirM spinal.

Ventral root ol lecond ipinaL

Bowers - Cranial Nerves Spelerpes.

For the sake of convenience in comparing Fig. 1 with Fig 3, and Fig. 2 with Fig. 4, Plates 1
and 2 have been bound in facing er.ch other.

PLATE 1.

Fig. 1. View of the roots, ganglia, and branches of the cranial nerves of the left
half of the head, as if seen from the left side projected on to the sagittal plane.
The ear capsule and the eyeball are represented in outline, and the contour of
the brain by a line accompanied by a tint on one side of it. X 72.

Fig. 2. Dorsal aspect of the roots, ganglia, and branches of the cranial nerves of
the right half of the head projected on to the frontal plane. The drawings were
made from reconstructions of the left half of the head, but in engraving were
reversed for the sake of readier comparison with Fig. 4. Outlines of eye, ear,
and brain represented as in Fig. 1. X 72.

PLATE 2.

Fig. 3. View of the branches of the cranial nerves in the vicinity of the left eye,
seen from the left side, projected on to the median plane, to show especially
the eye-muscle nerves in relation to the II. and certain branches of the V. and
VII. cranial nerves. X 72.

Fig. 4. Dorsal view of the nerves shown in Fig. 3, but from the right side of the
head, projected on to the frontal plane. Compare also with Fig. 2. The super-
ficial ophthalmic VII. has been only faintly shaded, instead of being printed in
black, in order to allow the course of the underlying nerves to be seen. The
distal end of IV. ob. su. has not been colored blue, owing to a mistake of the
lithographer. X 163.

Proceedings of the American Academy of Arts and Sciences,
Vol. XXXVI. No. 12. — November, 1900.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

ON CERTAIN DERIVATIVES OF ORTHOBENZO-

QUINONE.

By C. Loring Jackson and Waldemar Koch.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

ON CERTAIN DERIVATIVES OF ORTHOBENZOQUINONE.*

By C. Lokixg Jackson and Waldemar Koch.

Presented May 9, 1900. Received October 17, 1900.

Attempts to prepare the orthobenzoquinone

have been mentioned, so far as we can find, in only two papers. The first
is that by Zincke,t in which he describes tetrabroniorthoquinone and
tetrachlororthoquinone, but says that a number of attempts to make the
unsubstituted orthoquinone have failed to produce the desired result.
The second is by Hinsberg and Himmelschein,| who obtained ortho-
dioxydiphenylsulphone by the action of potassic dichromate on a mixture
of pyrocatechine and benzolsulphinic acid, and draw the inference that
the formation of this substance probably was preceded by the formation
of orthoquinone. They add that all their experiments to isolate ortho-
quinone have failed because of its strong tendency to condensation and
polymerization.

It seemed to us that one probable cause of the ill success of these
experiments was the sensitiveness of the orthoquinone to oxidizing agents ;
this view was confirmed by the observations of Cousin, § that trichlorpy-
rocatechine and dibrompyrocatechine give only red resinous products with

* The work described in this paper formed part of a thesis presented to the
Faculty of Arts and Sciences of Harvard University for the Degree of Doctor of
Philosophy by Waldemar Koch.

t Ber. d. chem. Ges., XX. 1776 (1887).

1 Ber. d. chem. Ges., XXIX. 2025 (1896).

§ Annates de Chiui. et de Phys., [7] XIII. 485.

198 PROCEEDINGS OF THE AMERICAN A< ADEMY.

nitric acid, whereas the triclilor- or tribrom-homopyrocatechine gives under
the same conditions an orthoquinone; and also by some experiments
undertaken by II. A. Torrey and on • of us on the action of hydric diox-
ide or electrolytic oxygen on pyrocatechine, which Led to unpromising
black products. Accordingly wo searched lor a method which would
produce a quinone without the use of mi oxidizing agent, and found one
in the action of iodine on the lead -ah of the dioxybenzol. After having
shown that the method converted hydroquinone into (juinone, we applied
ii rhino; hut our first experiments led to no result, because
we used alcohol as our Bolvent, which attacks the orthoquinone as soon
as it is Formed. Upon replacing the alcohol by chloroform we succeeded
in obtaiuing a dark garnet red solution, in which the presence of ortlm-
benzoquinone v tblished by a number of t' sts given later in this

paper.

From this solution we next tried to isolate the solid orthoquiuone, but
after trying every method we could devise, which gave even a remote
promise of >u ee. ■>-, we have failed to obtain a trace of solid orthoquinone,
and are forced to the opinion that it cannot exist in the Bolid -t ;itt-, or
even for any length of time in solution. The products obtained in place
of the orthoquinone in every case were a black substance insoluble in
chloroform, and a brown substance, or mixture of substances, which dis-
solved in chloroform : this brown product we have not attempted to inves-
• gate, hut we have studied the black insoluble product, and can give a
fairly complete account of its nature in spite of its unmanageable proper-
ties. It begins to form in the chloroform solution of the orthoquinone in
less than an hour, and is liest obtained by allowing the solution to Btand in
a corked flask for not more than thirty hours. A combustion of the bud-
stance gave results corresponding to the formula Ci8H80:, ; the analysis of
a lead Ball prepared in a somewhat different way indicated the same for-
mula, and this was confirmed by two analyses made by II. A. Torrey ol
his product from the electrolytic oxid ition of pyrocatechine, which Beemed

to be identical with ours, and by an analysis of the benzoyl derivative of

our black Bubstance; we have, therefore, five analyses of four different

samples, all of which agree with each other, and leave no doubt that the
itance is a definite compound, and has the formula (', I UK. This

formula indicate, that the compound was for d by the union of two

molecule- of orthoquinone with the introduction of an additional hydroxy]
group, a view which La confirmed by the analysis of the had Ball

I II OjPb, in which there are evidently two benzol rings to each atom

of 1- el. When we consider the manner in which the-,, two benzol ri

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 199

are united, two theories seem especially worthy of attention : first, the two
rings may be joined directly, so that the compound is a substituted di-
phenyl, (I.) 02(HO)H2C6-CGH3(OH)2 ; or second, they may be united by
two atoms of oxygen forming a substituted pyrocatechine ether according
to this formula : (II.) (HO)H8C6=02=C6Ha(OH)2* The first (diphenyl)
formula is rendered much more probable than the second by the anal-
ogy between our orthoquinone and the orthonaphthoquinone, which
is converted by dilute sulphuric acid into dinaphthyldiquinhydrone,t
O2rl5Cl0"C10H5(OH)2, a black powder like our alteration product; this
has been proved by Korn £ to be a dinnphthyl body by conversion into
a-a dinaphthyl and diphthalylic acid, (COCO)(C6H4COOH)3. The
support of the diphenyl formula by this analogy is so strong that we
should feel it was hardly worth while to consider the second (ether)
formula, if it were not for the fact that we have observed an oxygen
attachment similar to that in Formula II. in a derivative of tetrabrom-
orthoquinone, described later in this paper. To decide between Formula
I. and Formula II. we studied the action of reducing agents on the black
substance, as, if it were an ether (Formula II.), it should be reduced to
a mixture of pyrocatechine and a trioxybenzol ; whereas a diphenyl com-
pound (Formula I.) would only take up two atoms of hydrogen without
decomposition^ We selected as the reducing agent sodium amalgam
and water, because this had proved effective in breaking up the pyrocate-
chine ether containing bromine already alluiled to ; but, even after long
continued action on our black alteration product, not a trace of pyrocate-
chine could be detected ; the only reduction product was a black substance
evidently as complex as the original body. This result, therefore, de-
clares decisively in favor of the diphenyl structure (Formula I.).

The action of benzoylchloride on the black substance in presence of so-
dic hydrate converted it into a tribenzoyl derivative, C12H50203(COCGH5)3,
thus proving the existence of three hydroxyl groups in the substance ;
and this benzoyl derivative, and also the original black compound, gave
new substances with phenylhydrazine, which contained nitrogen, as was
shown by the analyses. Unfortunately the percentage of nitrogen found
in both cases was far below that required by the probable formulas, which

* In both these formulas the third hydroxyl group may be attached to the other
benzol ring.

t Stenhouse and Groves, Ann. Cliem. (Liebig), CXCIV. 205.

% Ber. d. chem. Ges., XVII. 8020.

§ Korn did not succeed in splitting dinaphthyldiquinhydrone into simpler bodies
by reduction. — Ber. d. chem. Ges., XVII. 3021.

200 PROCEEDINGS OF TIIE AMERICAN ACADEMY.

is undoubtedly accounted for by the difficulty in purifying these amor-
phous and infusible substances; our analyses, therefore, amount only to
qualitative proofs of the presence of nitrogen. The introduction of the
nitrogen, however, makes it highly probable that both the original
substance and its tribenzoyl derivative contain oxygen atoms in the
quinone position. We feel justified from the results of the foregoing
experiments in assigning to the black alteration product the formula
ilii! I ( ', ( ,1 [,(OH u in which two of the hydroxy! groups are prob-
ably in the ortho position to each other, since the substance forms a stable
bad salt when treated with plumbic acetate.

We have made no progress toward determining the position of the third
hydroxy! group, which may even be attached to the other benzol ring.
The introduction of this third hydroxyl group is not without analogy,
since the dinaphthylorthoquinone CiqHgCVCioIIsO.i >s converted bj the
action of the air on its solution in weak alkali into the dioxy compound
I ll.uIIO. CioH4OHOa.* In our case, however, the oxygen of the air
took no essential part in the reaction, as the same alteration product wa6
formed in an atmosphere of carbonic dioxide; and this is not surprising,
as we have bad frequent occasion to observe that the orthobenzoquinone
i< a much stronger oxidizing agent than the orthonaphthoquinone, so that
one portion of it might easily oxidize the other during the polymerization.
The portions of the orthoquinone which gave up oxygen during the
formation of the black body were undoubtedly converted into the brown
Bubstances soluble in chloroform, which made up about half of the total
product.

Although, as has been already stated, we could not obtain the ortho-
quinone in the Bolid state, many of its properties can be determined
from the study of its solution in chloroform; thus, there can be little
doubt that it has an intense red color and little if any odor, as the
solution smelt of nothing but chloroform, whereas a smell as strong as

that of paraquin could have been detected even in presence of

the amount of chloroform used. Its very slight stability has been

dwelt on at sufficient length already, but we Bhould add that it Beems
to be decomposed at once by even -mall quantities of water or alcohol.

w turn next to the reactions of the Bolution on which is based our
inference that it contained orthoquinone. Hydrochloric acid converted
it into monochlorpyrocatechine, which, however, was mixed with a little
pyrocatechine, as would be expo ' Wichelhausf has Bhown that

• Korn, ft r. .1. chi m. Gei . WII 8021.

• l: • r d. chem. Ges.. \n. 1504

JACKSON AND KOCH. — DERIVATIVES OF OBTHOBENZOQUINONE. 201

the action of hydrochloric acid on parabenzoquinone consists in the
formation of hydroquinone and chlorine, which subsequently react to-
gether, and in this corresponding case part of the pyrocatechine formed
conld well have escaped the substituting action of the chlorine. Re-
ducing agents like amnionic sulphide or sulphurous dioxide produced
pyrocatechine from this chloroform solution. Bromine formed with it
tetrabrompyrocatechine, the reduction of the orthoquinone being achieved
in this case by the hydrobromic acid formed in making the substituted
quinone.

When the chloroform solution was added drop by drop to a solution
of benzolsulphinic acid also in chloroform, an orthodioxydiphenylsul-
phone, C6H5S02C6H3(OH)2, melting at 153° was formed, which should
be identical with that obtained by Hinsberg and Himmelschein * by
oxidizing pyrocatechine in presence of benzolsulphinic acid, but the
properties of the two substances show they are different. Hinsberg
and Himmelschein's substance contaiued water of crystallization, and
melted at 117°, if this were not driven off; at 143° to 145° after heat-
ing in the steam-bath; and at 164° if completely dried by previous
fusion; our compound contained no water, and melted at 153°, whether
dried at ordinary temperatures or by previous melting. We have not
succeeded in finding the conditions under which Hinsberg and Him-
melschein's compound is formed, so that we were unable to compare the
two substances, and must consider with our present knowledge that they
are isomeric.

With aniline the chloroform solution gave a dianilidoquinoneanil,
C6H2(C6H5NH),O(C0H5N), which melted at 203°, and by this and
a careful comparison of its other physical properties was proved to
be identical with the body prepared by Zincke and von Hagen f from
paraquinone by the action of aniline and acetic acid. As this sub-
stance has previously been made only from paraquinone, it has always
been supposed that it contains the oxygen and anil group in the para
position to each other, but after our preparation of it from ortho-
quinone it is also possible that it is a derivative of orthoquinone ; and
some weight is given to this view by the fact that it is formed easily
from the orthoquinone by aniline alone, whereas it can be prepared
from the paraquinone only with difficulty, and by the combined action
of aniline and acetic acid. On the other hand, it seems hardly prob-

* Ber. d. chem. Ges., XXIX. 2025.
t Ber. (1. chem. Ges., XVIII. 785

202

PROCEEDING OF THE AMERICAN ACADEMY.

able that a derivative of orthoqninone would be so stable as this; and
the beta-naphthoquinone forma with aniline and alcohol a correspond-
ing anilidonaphthoquinoneanil, I : H60(C H6NH)(C II.-X). which is
andoubtedly a derivative of alpha-naphthoquinone. Zincke* also has
obtained recently from the azimidodichlororthoquinone by wanning it
with aniline an azimidoanilidomonochloroxyparaquinone, and similar
■ if conversion of orthoquinones into derivatives of para-
quinones are nut rare. We are so strongly of the opinion, therefore,
that our dianilidoquinoneanil i- a para body with the constitution
i ). 1. ( II X. I. | ( . II XII i ,,2.5, that we have not thought it worth
while to submit the question to an experimental investigation, which
would of necessity occupy much time. Assuming that the constitu-
tion given above is correct, the body would be formed from the ortho-
quiuone by the following series of transformations, which are modelled
after those worked out by C. Liebermann f for the corresponding case
of beta-naphthoquinone : —

I.

II.

III.

IV,

C6H6NH

O

C .II-XII

MK' II,

C0II6NH

NC6H6

O

NIK .11

6"5

N'MI;,

The replacement of the two hydrogen atoms of the orthoquiuone by
the anilido groups to form II. take- place under the oxidizing influence
of another portion of the orthoqninone, which takes up hydrogen to
form pyrocatechine (always formed in our process) ; II. then passes by
isomerization into III., which i- converted into IV. by the further action

of anil':

Upon treating the solution of orthoqninone with orthophony lene-
diamine no phenazine was formed, bul the reaction seemed to coi

Only in the oxidation of the diamine.

\\ ■ bave also devoted some attention t" the Btudy of the tetrabrom-
orthoquinone, and have obtained from it two derivatives. The first of
these was prepared by treating tetrabromorthoquinone with tetrabrom-
pyrocatechine dissolved in acetic acid and water; it was a beautiful red

• Ann chem. (Liebig), CCCX] 276.
Bi r. d. chem. <■■ i., XIV. II I

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 203

body, which was soluble only in warm nitrobenzol. Its formula was
C12Br604, and it was formed by the following reaction :

C6Br4(OH)2 + C6Br402 = C12Br604 + 2HBr,

as was shown by the deteetiou of hydrobromic acid as one of the prod-
ucts of the reaction, and the fact that a good yield was obtained only
when the two reagents were used in equal amounts. This formation
of the substance is most naturally explained by supposing that the two
benzol rings are united together by two of the atoms of oxygen forming
a substituted pyrocatechine ether, C6Br4=0./C6Br202, but it is also possi-
ble that the union took place directly between the two benzol rings
forming a diyhenyl compound with this formula, 02Br3C6-C6Br302, the
two atoms of bromine removed in this reaction combining with the
atoms of hydrogen of the pyrocatechine, and thus oxidizing it so that
the body contains two orthoquinone groups. To decide between these
two formulas the substance was submitted to reduction with sodium
amalgam, as under these circumstances the pyrocatechine ether should
be split into two molecules of pyrocatechine, while the diphenyl com-
pound should be reduced without splitting the molecule. The princi-
pal product of this reduction was a chocolate brown substance, which
contained an amount of bromine nearly identical with that of the origi-
nal body, and was converted into it again by nitric acid. With it there
was always formed a certain amount of pyrocatechine, and this seemed to
us strongly in favor of the ether formula, but it could not be considered
conclusive, as the amount of pyrocatechine was much less than that of
the brown product. In the hope of increasing the yield of pyrocatechine
we submitted the brown product again to the action of sodium amalgam,
but found it was entirely unaffected ; we accordingly oxidized it to the
original red body, — a reaction which takes place quantitatively, — re-
duced this again by sodium amalgam, obtaining fresh quantities of pyro-
catechine and the brown body, and continued these alternate reductions
and oxidations until the original specimen of the red body had been
completely converted into pyrocatechine. Although we could wish that
this conversion had taken place by reduction alone, the use of the oxida-
tions cannot be held to vitiate the argument, as they gave a quantitative
yield of the original substance, and we feel justified, therefore, in infer-
ring that the two benzol rings are united by two atoms of oxygen, and
that the substance is a pyrocatechine ether.

The brown product formed by the action of sodium amalgam was also
obtained by the action of tribrompyrogallol on the tetrabromorthoqui-

204 PROCEEDINGS OF THE A.MEBICAN ACADEMY.

none, and of hydrochloric acid at 1G0° to 17.3° on the red pyrocatechine
ether. There seems no doubt, therefore, that it is formed from the red
body by the conversion of two orthoquinone oxygen atoms into liydroxyl
groups ; in other words, that it is the pyrocatechine corresponding to the
red orthoquinone ; and this view is confirmed by the oxidation of the
brown substance to the red by nitric acid. The reaction of the hydro-
chloric acid is especially instructive in regard to the constitution of these
two bodies, as it is the familiar conversion of a quinone to a hydroqui-
none by this reagent. The red color of the original substance also points
to the presence of orthoquino oxygen in it, whereas the lighter color
of the reduction product (brown when crystallized, but purplish white
when amorphous) confirms the view that it is the hydroxy compound.
We therefore assi<m the following formulas and names to these suit-
stances, — red compound, Br4C6=On (',, Br8< ).., hexabromorthoquinopyro-
catecbine ether; the brown reduction product, Br4CVOo=CcBr2(OII |s,
hexabromdiorthoxypyrocatechine ether. The only observed fad not in
accordance with this last formula is that the substance does not dissolve
in a solution of Bodic hydrate, as we should expect it to do considering
that it contains two phenol bydroxyls; this may, however, be explained
by the very slight solubility of the substance (soluble only in warm
nitrobenzol). which might prevent the solution of sodic hydrate from
comin" into sufficiently close contact with it to react.

The hexabromorthoquinopyrocatechine ether was formed when acetic
acid or alcohol was used as the solvent, but none of it was obtained
when the tetrabromorthoquinone and tetrabrompyrocatechine were mixed
in solution in chloroform or ether. In this latter case the residue left
after the evaporation of the solvent was a reddish white mass, apparently
a mixture of the unaltered reagents, which contained none of the hexa-
bromorthoquinopyrocatechine ether. This is strange, as Zincke by mix-
in- the two Bubstances in ethereal solution obtained a black crystalline
itance. An attempt will be made in this laboratory nexl year to find
the conditions under which this lda«;k compound is formed.

W< have Buccei led also in obtaining a reaction Bimilar to thai jnst
ribed from tribromresorcine and tetrabromorthoquinone. The other
phenols tried acted -imply as reducing :i_r<'nt>. The product proved to
be i ditribrommetoxyphenyldibromorthoquinophenylene ether,

HBr.OHO < Br,Oa.

Thi instance melts at 217 , and like the reduction product just de-
scribed, does not di-solvc in a Bolution of Bodic hydrate, in Bpite of the two

JACKSON AND KOCH. — DERIVATIVES OP ORTHOBENZOQUINONE. 205

phenol hydroxyl groups it contains, according to its most probable for-
mula. We should explain this anomaly in the same way as before,
although this substance is much more soluble than the reduction
compound.

With glacial acetic acid the tetrabromorthoquinone C6Br402 formed a
white compound, to which we have given much attention, but of which
we can only give a very imperfect account. Ziucke * in hie paper on
tetrabromorthoquinone noticed that it was attacked by glacial acetic acid,
but did not isolate the product ; on the other hand he obtained a white
body by crystallizing tetrabromorthoquinone from ether and ligroin, which
we have not succeeded in encountering. When tetrabromorthoquinone is
evaporated to dryness with glacial acetic acid, a small amount of the red
hexabromorthoquinopyrocatechine ether is formed, but the greater part
of the quinone is converted into a white crystalline body melting at 230°,
the analysis of which led to one of the two following formulas, C14H2Br805
or C14Br805. The first would be formed by the union of two molecules
of the tetrabromorthoquinone with one of acetic acid, one molecule of
water being eliminated, thus : —

2C6Br402+C2H402= C14H2Br805+H20.

The second would be formed from the first by removing two atoms of
hydrogen, which could be done by the oxidizing action of another portion
of the tetrabromorthoquinone, and the tetrabrompyrocatechine thus
formed would then give rise to the red body, which always accompanied
the white product ; but the amount of this red compound always fell far
below that which should be formed, if it had taken such an important
part in the reaction (10 grams of tetrabromorthoquinone should give 5
grams of the red product, whereas the usual yield was only 0.8 gram),
and, as the amount of hydrogen found in the white compound was dis-
tinctly larger than should have been given by a body free from that
element, we have decided to adopt provisionally the formula C14H2Br805.
We had hoped to settle this question by studying the decomposition prod-
ucts of this substance, but so far it has either been unaffected or com-
pletely decomposed by all the reagents we have tried. We hope that in
the coming year the study of this body will be completed and the work
carried on in various other directions in this Laboratory.

* Ber. d. chem. Ges., XX. 1777.

•j1"'' proceedings of the american academy.

Experimental Part.

Action of Iodine on the Lead Salt of IFi/droquinone.

-•■ experiments were tried to test the efficacy of this method of
makin" qu'mones. An aqueous solution of ten grama of hydroquinone
was treated with the calculated amount of Bodic hydrate in a flask, to
which the air had access only through a tube containing an alkaline
solution of pyrogallol ; to the Bodium Bait thus formed an aqueous solution
of plumbic acetate was added, which gave a white precipitate of the lead
salt. A.8 this blackened rapidly on exposure to the air, no attempt was
made to filter it. but an alcoholic solution of live grams of iodine was
added directly to the precipitate suspended in the liquid, wdien plumbic
iodide was formed, as shown by the appearance of a yellow color. After
the reaction seemed to have come to an end. the insoluble substances
were filtered out, and the filtrate, which had a strong odor of quinone,
extracted with ether. This ether extract left on evaporation yellow
}tals with a Btrong odor of quinone, which under the microscope
showed the characteristic form of this substance, and after purification by
sublimation melted at 11 2° -11 3°, which, although distinctly below the
melting point of quinone, 115°. 7, is near enough to it to prove in con-
junction with the other properties that quinone was formed.

To prove that this quinone was formed by the action of the iodine on
the lead salt and not by the oxidizing effect of the air. a specimen of the
salt mixed with water was exposed to the air for several hours, until it
was thoroughly blackened ; upon filtration a red solution was obtained
which had no odor of quinone. Extraction with ether did not remove
the red color, and on the evaporation of the ether, crystals of hydroqui-
npm obtained, recognized by the melting point, but no quinone.

Upon adding an alcoholic solution of iodine to a similar specimen of
the lead Bait blackened by exposure to the air, the sharp odor of qui-
nine was observed at once, and a considerable amount of this body was

obtained from the filtrate. The method, therefore, is applicable to the
production of quinones, but naturally would be osed only in cases where
oxidation is inadmissible.

We also tried some experiments with iodine and the had salt of resor-
cine, but as it was evident that die purification of the product would he
i matter of great difficulty, we have postponed the continuation of this
work until we have finished the Btudy of pyrocatechine.

JACKSON AND KOCH. — DERIVATIVES OP ORTHOBENZOQUINONE. 207

Orthoquinone.

Action of Iodine on the Lead Salt of Pyrocatechine, Preparation of a

Solution of Orthoquinone.

To prepare the lead salt of pyrocatechine two grams of pyrocate-
chine dissolved in 150 c.c. of water were added gradually, with constant
stirring, to a hoiling solution of 7.5 grams of plumbic acetate in 250 c.c.
of water. The precipitated lead salt was allowed to settle, filtered out,
and washed five times with hot water. It was then dried at 100°, and
finely pulverized. It is not advisable to use more than two grams of
pyrocatechine at a time in making the lead salt, as with larger amounts
it was found to be hard to wash out all the impurities.

Ten grams of this lead salt were thoroughly moistened with chlo-
roform in a 500 c.c. glass-stoppered Mask. The salt must be absolutely
dry, and the chloroform free from water and alcohol, as' these sub-
stances decompose orthoquinone. A boiling solution of five grams of
iodine in 200 c.c. of chloroform was then added to the lead salt and
the whole shaken well for five to ten minutes ; after which, upon fil-
tering through a dry Gooch crucible, a garnet red solution of orthoqui-
none was obtained. The amount of iodine used in this experiment is a
little more than half that required by the theory (one molecule to each
molecule of the lead salt). This large excess of the lead salt was used
to avoid as far as possible the presence of iodine in the chloroform
solution, and in this we were fairly successful, as a sample of it, when
tested with a fresh specimen of the lead salt of pyrocatechine, showed
only a trace of plumbic iodide.

This chloroform solution contained orthobenzoquinone, C6H402 1.2, as
is proved by the experiments described later in this paper, and it must
have been a nearly pure solution of this body, since the only other sub-
stance which could have been present was iodine, and the amount of
this was very small, as shown in the preceding paragraph. It had a dark
garnet red color, and no odor could be perceived except that of the
chloroform. Within an hour from the time of its preparation it began to
show signs of alteration by depositing a black precipitate, and after
twenty-four to thirty hours all the orthoquinone had disappeared from
the solution. We have made numerous attempts to isolate the solid qui-
none from its solution in chloroform, of which the following is a brief
summary : Evaporation spontaneously at ordinary temperatures ; evap-
oration at — 12° ; evaporation in vacuo of a solution in pure chloroform
made from chloral, so as to avoid all traces of alcohol which decomposes

208 PROCEEDINGS OP THE AMERICAN ACADEMY.

the orthoquinone; evaporation by a stream of dry air, or dry carbonic
dioxide passed through the solution cooled by a freezing mixture ; pre-
cipitation with ligroin at— o ; precipitation with tetrachloride of carbon at
—20°; we bad great hopes of this la>t attempt, as we had not succeeded
in making any of the orthoquinone, when tetrachloride of carbon was
substituted for chloroform in the preparation, and we inferred from this
that the quinone probably was insoluble in this substance ; but this, as
well as all the other experiments mentioned above, did not lead to the de-
sired result. The product in every case was the black alteration product
already mentioned, the nature of which will be discussed presently, so
that we are forced to the conclusion that the orthobenzoquiuoue cannot
exi.>t for any length of time, if at all, in the solid state.

Study of the Alteration Product of Ortiiobenzoquinone.

This substance was most conveniently obtained by allowing the chloro-
form solution of orthobenzoquiuoue, prepared as already described, to
stand at ordinary temperatures in a corked flask from twenty-four to
thirty hours.* The black precipitate thus formed was not easy to purify,
for although it was soluble in alcohol and some other solvents, all our
attempts to crystallize it have failed; we finally decided to depend for its
purification on its insolubility in chloroform and benzol. Accordingly,
after filtering it out from the chloroform in which it was formed, we
washed it thoroughly with chloroform, and afterward extracted it three
times with hot benzol. About 40 per cent of the orthoquinone used
was converted into this black product. In calculating this yield we
assumed that the amount of orthoquinone in the chloroform solution cor-
responded to the amount of iodine used in making it. The black sub-
Btance, after it had been purified by extraction with benzol, was dried til
;■</<•,/,, and analyzed with the following results: —

0.2" 11 gram of the substance gave on combustion 0.4G84 gram of car-
bonic dioxide and 0.0696 "ram of water.

i f'.r Calculated for «._. .

I ii 0 "ii ',ii ■ ■

Carbon 66.67 62.07 62.59

Hydrogen 8.70 8.45 8.79

Thee numbers would indicate thai the Bubstance has been formed b\
the union of two molecules of orthoquinone with the introduction of a

• l in- solution ihould not be allowed to itand more than thirty hours, as in tlmt
case the product is contaminated with a brown Imparity.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 209

hydroxyl group, but it is obvious that the compositiou of a body like this,
which is amorphous, and has no definite melting point, can be determined
only by the analysis of several samples prepared by different methods.
Accordingly we undertook the preparation and analysis of the lead salt,
which should be insoluble and stable, if the substance contains two
hydroxyls in the ortho position to each other, as would be the case if it
is formed from two molecules of orthoquinone. Further, as we thought
that the additional hydroxyl indicated by the analysis might have been
introduced by the action of the oxygen of the air, we repeated the prep-
aration in an atmosphere of carbonic dioxide. The solution of ortho-
quinone was made from the lead salt of pyrocatechine and iodine in a
flask filled with this gas, and was separated from the plumbic iodide by
means of a rose filter with an asbestos film without exposing it to the
air. Upon standing over night still in an atmosphere of carbonic dioxide,
a black precipitate was formed identical in appearance with that obtained
by the ordinary process. This was purified in the way already described,
and then converted into its lead salt as follows : 0.3 gram of this black
body dissolved in 40 c.c. of alcohol was added to a boiling solution of one
gram of plumbic acetate in 400 c.c. of water ; this large amount of water
is necessary to prevent the alcohol from precipitating plumbic acetate,
which, becoming imprisoned in the lead salt, could hardly be removed
by washing. The jet black lead salt thus obtained was washed five
times with hot water, dried at 100°, and analyzed with the following
result : —

0.2352 gram of the salt gave 0.1645 gram of plumbic sulphate.

Calculated for Calculatpd for VmnA

C12HG0202l»b. Cl2H50,OH02Pb. iouna.

Lead 49.17 47.39 47.76

This result shows that the oxidizing agent, which introduces the ad-
ditional hydroxy], is not the oxygen of the air, but the orthoquinone
itself, and this view is confirmed by the fact that the chloroform solution,
from which the black compound has separated, contains brown substances
probably formed from the orthoquinone in oxidizing the insoluble prod-
uct, but which were so unmanageable that we were unable to determine
their composition.

We would add here two analyses made by H. A. Torrey some years
ago in this laboratory with a body agreeing with our black compound in
its properties, which he had prepared by the electrolytic oxidation of
pyrocatechine. They gave the following results : —
vol. xxxvi. — 14

Calculated for

1 Jl 1

i.

Carbon

62.07

62.88

Hydrogen

3.45

3.80

210 PROCEEDINGS OF THE AMERICAN ACADEMY.

I. 0.1275 gram of the substance gave on combustion 0.2910 gram of
carbonic dioxide and 0.0436 gram of water.
II. 0.0608 gram of the Bubstance gave 0.1401 gram of carbonic dioxide
and U.U230 gram of water.

Found.

II.

G2.85
4.20

The substance is obviously the same as that previously analyzed, although
not bo pure.

We bare, therefore, four analyses of three samples of this substance
prepared or purified in different ways, and, as these agree, there can be
no doubt that the black alteration product of the orthoquinone is a deli-
nit, compound, and has the composition Ci2IIs05.

Properties of the Black Alteration Product of Orthobenzoquinone
( Orthodioxyphenyloxyorthoquinone,* ( ,1 [8| < )11 (..CV.ILOIIOj. — It forms
an almost black amorphous powder, which does not melt, but decomposes
gradually at temperatures above 170°. It is easily soluble in cold alco-
hol, ether, or glacial acetic acid; insoluble in benzol or chloroform,
whether cold or hot. Strong acids have no apparent action on it, but
alkalies dissolve it.

The fallowing experiments were undertaken to determine the constitu-
tion of the black alteration product.

Reduction of the Alteration 1 'rod act.

The formula of the substance C',. I U< >- indicates that it contains two
benzol rings, and these may cither be united directly, forming a diphenyl
compound, or they ma) be united h\ two atoms of oxygen, forming a
substituted pyrocatechine ether, [(',-. I b<>l I ]Os[C6l I,.<<)1 1 1, ]. The di-
phenyl formula, (HO)08H8C9-CeH8(OH)a, is rendered probable by the
formation of the dinaphthyldiquinhydrone "f under Bomewhal similar cir-
cumstances from beta-naphthoquinone, which has been proved to be a
dinaphthyl compound;! the pyrocatechine ether formula, on the other
hand, is in harmony with a derivative of tetrabromorthoquinone, which is
ribed later in this paper. Reduction Beemed the easiest waj to de-

• This name and formula n>r the substance are established by the «.>rk

bed later in tlii* paper.

. m. <Ii-ii.. (I.i. in i. i V [V 206
I Ibid. 206. K'.m Ber. d. cbem. Ges., XVII 8020.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 211

cide between these formulas, as the ether should be split into pyrocate-
chine and a trioxybenzol, while the diphenyl product would be reduced
without splitting apart the two benzol rings. Upon treating the black
alteration product with sodium amalgam and water in an atmosphere of
carbonic dioxide for several days, an uninviting black solid was obtained
on acidification, which was soaked in water, and this water as well as the
acidified filtrate from the reduction extracted with ether, but not a trace
of pyrocatechine was obtained. As the bromether, which we have ob-
tained in another part of our work, gives a large quantity of pyrocate-
chine by this treatment, we infer that the black alteration product is a
diphenyl derivative.

Action of Benzoylchloride on the Black Alteration Product.

Half a gram of this substance was dissolved in 10 c.c. of a solution of
potassic hydrate (1:5), and to the warm solution benzoyl chloride was
added drop by drop, until the product of the reaction rose to the surface
as an oily mass, leaving the liquid clear. This was then treated with
sodic hydrate, in which it did not dissolve, and thoroughly washed with
water. The tarry mass thus obtained was dissolved in chloroform, the
solution filtered, and the filtrate mixed with an excess of alcohol, which
preciptated out the substance. After drying in vacuo it was analyzed,
with the following result: —

0.2210 gram of the substance gave on combustion 0.5933 gram of carbonic
dioxide and 0.0744 * gram of water.

Found.

Carbon 70.91 72.79 73.21

Hydrogen 3.64 3.68 3.74

The substance is therefore a tribenzoyl compound, and the original black
body must have contained three hydroxy! groups. This benzoxyorthoquino-
orthodibenzoxydiphenyl, C6H2(OCOC6H5)02C6H3(OCOC6H5)2, f forms a
light yellow amorphous powder without a definite melting point. It is
easily soluble in chloroform ; insoluble in alcohol or water. Alkalies do
not dissolve it.

The fact that only three benzoyl groups are taken up by the black
alteration product indicates that two of its five atoms of oxygen do not

* A little water was lost in this analj-sis.

t The distribution of the benzoxy groups in this substance has not been
determined.

Calculated for
C12HG03(OCOCcH5)2.

Calculated for
Ci2H30,(0C0CsH5)3

70.91

72.79

3.64

3.68

212 PROCEEDINGS OF THE AMERICAN ACADEMY

form part of hydroxy] groups : from the formation of the substance these
two atoms of oxygen must either belong to a <|uinone, or be the connect-
ing atoms of a pyrocatechine ether, ami. as our work on the reduction < t
the alteration product baa Bhown that it i- probable tiny air nol pyro-
catechine ether oxygen atoms, there can lie little doubt that they occur in
the quinone configuration, which indeed was the more probable supposi-
tion of tin- two. We tried to confirm this inference by the formation of
hydrazones both from the original black alteration product and its ben-
zoyl derivative, but without satisfactory results. When phenylhydrazine
in excess was added to a solution of 0.5 gram of the black alteration
body dissolved in 10 CC. of alcohol, the dark blackish color of the
lution was discharged, and on pouring the product into acidified water a
golden yellow precipitate was formed; but an analysis of the substance
gave 6.39 per cent of nitrogen, whereas the most probable formula.
I II i kNNI K',.11.-,. requires 8.70 per cent of nitrogen. The tribenzoyl
derivative of the black sulistance was dissolved in chloroform, and boiled
with an excess of phenylhydrazine for several minutes ; a new com-
pound was formed, which was obtained by allowing the chloroform to
evaporate spontaneously, treating with an acid to remove the excess of
phenylhydrazine, dissolving the residue in cold alcohol, and precipi-
tating the substance from this solution with water. After drying in
vacuo it was analyzed, and 3.19 per cent of nitrogen found instead of
1.42 per cent of nitrogen required by the formula of the hydrazone,
< II ' i NMIC, II..)(OCOC6Hj)3. It is evident from these results that
phenylhydrazine acts on both of these substances, and that the product in
each caBe contains nitrogen. The results, therefore, as far as they _
confirm our inference that there are two quinone atoms of oxj gen in the
black Bubstance, but they are of little value, because the exceedingly unman-
ageable nature of both of these substances prevented us from purifying
them sufficiently to determine their composition with certainty.
The work just described establishes the following points with tolerable
linty. The black alteration product is a diphenyl compound, and
contains three hydroxy] groups and an orthoquinone group ; it may there-
fore be called orthodioxyphenyloxyorthoquinone, and receive the formula
f II (OH ' II OH)0„ although the three hydroxy] groups may be
attached to the -one benzol ring, as we have not determined the position
of the third hydroxy] experimentally.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 213

Experiments with the Solution of Orthobenzoquinone.

Action of Hydrochloric Acid on the Solution of Orthobenzoquinone.

Dry hydrochloric acid gas was passed through a fresh chloroform
solution of orthoquinone (prepared from iodine and the lead salt of py-
rocatechine), until it fumed strongly. After filtering out a slight tarry
precipitate, a large part of the chloroform was recovered by distillation,
and the concentrated solution thus obtained allowed to stand over niaht in
an open vessel. The residue, which was usually oily, was heated on the
steam-bath to drive off the last traces of the solvent and a slight impurity
of iodine, and then extracted with a hot mixture of three parts chloroform
with one of ligroin, which left behind a large quantity of a black oil.
The solution became milky on standing, and deposited an oily substance,
followed, when it had cooled, by the formation of crystals. As soon as
these crystals began to appear the clear solution was decanted from the
oil and allowed to crystallize in a separate vessel. The residual oils by
a similar treatment with chloroform and ligroin yielded an additional
amount of the crystals. The crystals were dissolved in benzol, and
treated with bone-black, after which they were purified by repeated crys-
tallizations from the mixture of chloroform and ligroin. They showed
the melting point 84° to 85°, which was not quite constant, indicating
that the substance was not perfectly pure ; but in spite of this it was dried
in vacuo, and analyzed with the following result : —

0.2005 gram of the substance gave by the method of Carius 0.1865 gram
of argentic chloride.

Calculated for „ .

C6H3C1(0H)2. Found-

Chlorine 24.57 23.00

This result indicates that the body is monochlorpyrocatechine, but
that it contains some pyrocatechine, as would be expected if the reaction
ran in the same way as that of hydrochloric acid on parabenzoquinone,
since Wichelhaus * has shown that in this case hydroquinone and
chlorine are formed in the first instance, and that the substitution com-
pounds, which form the final product, are made by the action of the
chlorine on the hydroquinone. We have not attempted to purify the
chlorpyrocatechine more thoroughly, as much time and labor would be
necessary to obtain sufficient material for this purpose, and the forma-

* Ber. d. chem. Ges., XII. 1504.

214 PROCEEDINGS OF THK AMERICAN ACADEMY.

tion of chlorpyrocatechine under tliese conditions is confirmed by the
following analysis of the lead salt, made by adding plumbic acetate to an
aqueous solution of the product of the reaction of hydrochloric acid on
the solution of orthoquinone.

67 gram of the salt gave .03306 gram of plumbic sulphate.

Calculated for -. . ,

C,H8C10,Pb. Found-

Lead 58.98 58.55

Tlie monochlorpyrocatechine in the slightly impure state in which we
obtained it crystallized in white pearly plates, which are easily soluble
in water, alcohol, ether, benzol, chloroform, or acetone; insoluble in
ligroin. With an aqueous solution of plumbic acetate it gives a had
salt. It blackens quickly if exposed to the air after treatment with
an alkali. The melting point observed by us is of course of no value,
on accouut of the presence of the impurity of pyrocatechine.

Behavior of the Solution of Orthobcnzof/ttinono with Reducing Agents.

Amnionic Sulphide. — Some of the solution of orthoquinone in chloro-
form was shaken with about one third of its volume of the ordinary
laboratory solution of amnionic sulphide. The red color of the ortho-
quinone disappeared, and the aqueous solution took on a dark color.
The two liquids were separated by means of a drop funnel, and the
chloroform was extracted twice with water, the extracts being added to
the ammonic sulphide solution at first obtained. The mixed aqueous
portion was then heated on the steam bath, acidified with hydrochloric
acid, allowed to stand until the precipitated sulphur had settled, filtered,
and the clear dark colored filtrate extracted with ether. The residue
left on the evaporation of the ether was recrystallized twice from benzol,
wh'-n it. was found to melt constant at 103°-1<>I .and to show all the
oilier properties of pyrocatechine. In order to be certain that the pyro-
catechine, detected in the product from the action of ammonic sulphide

on the orthoquinone solution, was fun 1 from the orthoquinone, it was

necessary to prove that this solution contained no free pyrocatechine;

for this purpose a -mall amount of it taken before the reduction was
■diaken with water, the water evaporated, and the residue extracted with

benzol; the benzol extract was evaporated to dryness, the very Blight

residue dissolved in water, and tested with plumbic acetate, when it
• no precipitate, showing thai no ])■>■<■ pyrocatechine was contained in
the chloroform solution.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 215

Action of Sulphurous Dioxide. — Dry sulphurous dioxide was passed
through the chloroform solution of orthoquiuone until it was saturated ;
the principal product was a black unmanageable precipitate, probably a
compound of orthoquinone and pyrocatechiue corresponding to chiuhy-
drone ; but the filtrate from this left on evaporation crystals, which were
recognized as pyrocatechiue by their properties and the formation of the
lead salt.

Action of Bromine on the Solution of Orthobenzoquinone.

An excess of bromine was added to the solution of orthoquinone in
chloroform, the chloroform allowed to evaporate spontaneously, and the
crystalline residue, after being heated gently to drive off the excess of
bromine, was spread on unglazed porcelain to remove oily impurities.
After recrystallization from benzol it showed the constant melting point
191°, which is near enough to 192°-193°, that given by Zincke* for
tetrabrompyrocatechine to leave no doubt as to the nature of the sub-
stance. For greater certainty the substance was analyzed with the fol-
lowing results : —

'&

I. 0.1835 gram of the substance gave by the method of Carius 0.3210
gram of argentic bromide.
II. 0.1509 gram of the substance gave 0.2637 gram of argentic bromide.

Calculated for Found.

C6Br4(OH)2. I. II.

Bromine 75.11 74.44 74.39

These numbers are sufficiently near to show that the product of the
action of bromine on the orthobenzoquinone is tetrabrompyrocatechine.

Behavior of the Solution of Orthobenzoquinone with Benzolsulphinic Acid.

The chloroform solution of orthoquinone was added drop by drop to a
solution of benzolsulphinic acid, f also in chloroform, until the color began
to disappear less rapidly, and the solution of the sulphinic acid had as-
sumed a decided reddish tint. After standing over night, the liquid was
filtered, the filtrate evaporated to dryness, and the residue allowed to
solidify. The solid residue was next treated with hot benzol, which

* Ber. d. chem. Ges., XX. 1777.

t If ttie conditions were reversed, and the solution of benzolsulphinic acid was
added to the solution of orthoquinone, none of the desired sulphone was obtained;
we ascribe this difference in behavior to the oxidation of the sulphinic acid by the
excess of orthoquinone.

216 PROCEEDINGS OF THE AMERICAN ACADEMY.

dissolved all but a slight black residue. On allowing the benzol solu-
tion to evaporate a crystalline substance was obtained, which was pari-
tied by crystallizations with the aid of bone-black at first from water
mixed with a little alcohol, finally bom water alone, until it Bhowed
the constant melting point 153°, when it was dried at 100", and ana-
lyzed with the following results: —

I. 0.2o5o gram of the substance gave on combustion 0.4931 gram of
carbonic dioxide and ".<iSG2 gram of water.
II. 0.1454 gram of substance gave by the method of Carius 0.1418
gram of baric sulphate.

Found.

II

Carbon

Calculated for

.',11 Oil .<0,C,;Uj.

57.60

i.
57.14

1 lxdrogen

4.00

4.07

Sulphur

12.80

13.39

As this body must be a substituted pyrocatechine, it should give a
lead Bait, and upon trying it with a solution of plumbic acetate we
obtained one, which was analyzed with the following result: —

0:2601 gram of the substance gave 0.1729 gram of plumbic sulphate.

Calculated for _ ,_,

CclI3O.,t>uS0,»', II... *0aUd-

Lead 45.50 45.39

There can be no doubt, therefore, that the substance is a pyrocate-
chinesulphone or orthodioxydiphenj lsulphone.

Properties of Orthodipxydiphenyhidpkone, C6Hg(OH)2SOsCatT5. — It
crystallizes from alcohol and water in short, thick, well-formed, white
prisma terminated by two planes at an obtuse angle to each other.
It contains no water of crystallization, and melts at 158°. It is
ily Boluble in alcohol, ether, chloroform, or glacial acetic acid;
\.-rv Blightly soluble in cold water, or benzol, freelj Boluble in either
of these liquids when hot; essentially insoluble in ligroin, either hot or
cold. Alkalies dissolve it with a yellow color. Ferric chloride gives a

bluish green color with it, which changes to red on tin- addition of

>odic carbonate. A- has been already mentioned, its aqueous solution
gives a precipitate with plumbic acetate.

This Bubstance has the Bame composition a- the orthodioxydiphenyl-
Bulphone made by Hinsberg and Eiimmelschein * by oxidizing pyrocate-

• Ber. <1. dam Gei \\l\ 202

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 217

chine in presence of benzolsulphinic acid, and we should expect that the
two hodies would be identical, but, strangely enough, they seem to be
isomeres, since Hiusberg and Himmelscheiu's compound contains water
of crystallization, and melts in presence of this at 117° ; when dried iu
the steam oven, at 143°-145° ; when previously fused in the air, at
164°; whereas our substance contains no water, and melts at 153°,
whether dried at ordinary temperatures or by previous fusion. We made
five attempts to prepare their body according to Hiusberg and Himmel-
schein, but did not succeed in finding the conditions under which it is
formed, which must lie within very narrow limits ; and, as in doing this
we had devoted more time to the subject than we could afford, we were
obliged to abandon our intention of making a comparative study of the
two compounds.

Action of Aniline on the Solution of Orthobenzoquinone.

The solution of orthoquinone in chloroform was heated with an excess
of aniline for five minutes on the steam-bath ; the chloroform was then
allowed to evaporate, and the residue, after having been freed from ani-
line by treatment with hydrochloric acid and washing with water, was
purified by recrystallization from a mixture of two parts of alcohol to one
of benzol, until it showed the constant melting point 203°, when it was
dried at 100°, and analyzed with the following result : —

0.2909 gram of the substance gave 28.2 c.c. of nitrogen at a tempera-
ture of 22° and a pressure of 774 mm.

Calculated for „ ,

CcH^CoIIsNH^CelLjNO. *ound-

Nitrogen 11.51 11.27

The substance is therefore a dianilinoquinoneanil. When the aqueous
wash waters obtained in the preparation of this substance were extracted
with ether, pyrocatechine was obtained, recognized by its odor and
lead salt.

Properties of the dianilinoquinoneanil, C6Ho(C6H5NH)2C6H5NO. —
Crystallized from alcohol and benzol it forms bronze-colored small
needles melting at 203°, although they begin to draw together some-
what at 198°. The melting point of this substance is identical with that
(202°-203°) given by Zincke and Hagen * for the dianilinoquinoneanil
made from paraquinone by the action of aniline and glacial acetic acid.

* Ber. d. cliem. Ges., XVIII. 787.

218 PROCEEDINGS OF THE AMERICAN ACADEMY.

and, as a comparison of the two substances showed th< \ crystallized in
the same form, there cau be no doubt of their identity. The formation
of a paraquinone derivative from our orthoquinoue has been already
explained in the introduction to this paper.

In the hope of obtaining a phenazine the solution of orlhoquinone was
treated with orthophenylene diamine j a reaction took place, the principal
products of which consisted of uninviting black Bubstances and pyrocate-
chine recognized by the formation of its lead salt. From the appearance
of the pyrocatechine we inferred that the reaction consisted principally
in the oxidation of the diamine by the orthoquinoue, and as even after
trying the experiment under several varying conditions no more prom-
ising results were obtained, this line of work was abandoned. Phenol
also gave with the solution of orthoquinoue such an uninviting product
that we did not attempt to study it. Sodic hydrate gives with the ortho-
quinoue solution a green coloration similar to that obtained by the action
of sodic hydrate on tetrabromorthoquinone.

Dkrivatives of Tktrabromorthobenzoquinone.

f

The tetrabromorthoquinone used in this work was prepared as follows:
20 grams of pyrocatechine were dissolved in CO c.c. of glacial acetic
acid, and 37 c.c. of bromine added gradually from a burette. The mix-
ture was allowed to stand over night, after which the product was puri-
fied by recrystallization from 200 c.c. of glacial acetic acid. In this way
forty grams of pyrocatechine yielded 118-12G grains of reervstallized
tetrabrompyrocatechine ; that is, between 7o and 80 per cent of the theo-
retical yield. To convert this tetrabrompyrocatechine into tetrabrom-
orthoquinone thirty grams of it were dissolved in 300 c.c. of glacial acetic
acid by heating on the water-bath ; the solution was then cooled to 16°,
or until the glacial acetic acid began to solidify, and eleven grams of
fuming nitric- acid of Bpecific gravity 1..") diluted with 60 c.c. of glacial
acetic acid added rapidly, the mixture being kept cool and stirred vig-
isly during the addition; after Btanding for five minutes 300 c.c. of
water were added, and stirred in thoroughly. If the process had run
successfully, the tetrabromorthoquinone settled to the bottom in a glis-
tening mass of dark red crystals, which were filtered off) and dried. The
product thus obtained melted at 1 !•'» -l 17 instead of 150° 151°, but
was pure enough for our work, which was fortunate, as recrystallization
from glacial acetic acid is attended In a great loss ,,f material. Thirty
grams of tetrabrompyrocatechine gave eighteen grams of tetrabromortho-
quinone, which amounts to about 60 per cent of the theoretical yield.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 219

As has just been stated, the crystallization of tetrabromorthobenzo-
quinone from glacial acetic acid cannot be effected without considerable
loss, which we found was due to the conversion of the tetrabromorthoqui-
none into two new, bodies, one of which was red, the other white.* A
more careful study of the conditions under which the red body was
formed showed us that it was produced by the action of the orthoquinoue
with tetrabrompyrocatechine, either existing as an impurity in our tetra-
bromorthoquinone, or formed from it by reduction. After this the study
of this red product was carried on with comparatively little difficulty.

Preparation of Hexabromorthoquinopyrocatechine Ether, C12Br604.

Eleven grams of tetrabromorthoquinone and ten grams of tetrabrom-
pyrocatechine were dissolved in 470 c.c. of glacial acetic acid by heating
the mixture on the steam-bath, 180 c.c. of water were then added, which
produced no immediate precipitate, and the mixture allowed to stand in
a warm place for twenty-four hours. During this time a red precipitate
was formed, which was filtered out, and washed with hot alcohol to re-
move unaltered tetrabromorthoquinone and pyrocatechiue. In this way
10.7 grams were obtained; that is, somewhat over 60 per cent of the
theoretical yield. The purification of the substance offered at first some
difficulty because it was insoluble in all the common solvents, but later
we found that it dissolved to a limited extent in hot nitrobenzol, and pro-
ceeded as follows : Two grams of the crude red body were dissolved in
150 c.c. of nitrobenzol by heating the mixture on the steam-bath (at
higher temperatures more of the substance will dissolve, but the solution
is attended with decomposition). Upon cooling, the solution deposited
about 0.7 gram of crystals, which were submitted to a second similar
crystallization from nitrobenzol at 100°, after which the product was
extracted several times with hot alcohol to remove adhering nitrobenzol,
dried at 100°, and analyzed. As the substance does not melt, we had
no means of determining whether this treatment had been an efficient
purification.

I. 0.1809 gram of the substance gave on combustion 0.1374 gram of

carbonic dioxide and 0.038 gram of water.
II. 0.1008 gram of the substance gave by the method of Carius 0.1637
gram of argentic bromide.
III. 0.1129 gram of the substance gave 0.1843 gram of argentic bromide.

* Zincke (Ber. d. chem. Ges., XX. 1777) observed that acetic acid acted on tetra-
bromorthoquinone but did not isolate the products of the action.

220 PROCEEDINGS OF THE AMERICAN ACADEMY.

Calculau-J for

i.

Found.

11.

Carbon

20.98

20.71

Hydrogen

0.

0.23

Bromine

69.76

GO. 13

111

G9.49

Properties of BexabromorthoquinopyroccU/echine Ether. — The sub-
Btance crystallizes from warm nitrobenzol in Bhort slender prisms of a
pure vermilion red color; it does not nail, and is insoluble in all the
common solvents; nitrobenzol dissolves it especially when warm, but, it'
the solution is beau- 1 above LOO . decomposition sets in, as shown by a
cbange of color. The ordinary strong acids have no effect on it iu the
cold, but fuming nitric acid dissolves it with decomposition; alkalies do
not dissolve it, but the substance turns black after standing with an alkali
for a short time. The red color of the substance indicates thai it con-
tains two orthoquinone oxygen atoms, ami this view is confirmed by its
behavior with hydrochloric acid and reducing agents described later.

The red substance is probably formed from tetrabrompyrocatechine
and tetrabromorthoquinone by the following reaction : —

( , r.r1OII)2+CcB.^O,=CcBr40:,CcBr202+2IIBr.

To test this a new preparation was made as' described above, and. after
the mixture bad stood twenty-four hours, it was diluted with five times
it- volume of water, the reddish white precipitate formed filtered out, the
aqueous filtrate acidified with nitric acid, and treated with argentic nitrate,
when a heavy precipitate of argentic bromide was thrown down, showing
the presence of hydrobromic acid among the products of the reaction. It
was also found that the best yield was obtained when t he two reagents
were used in the proportion of equal molecules, as required by the reac-
tion. The besl Bolventto be used in the preparation was glacial acetic
acid and water, as already described, but the red body was also formed
iii dilute alcohol (the proportions used were "."' .'ram of each reagent,
40 r.r. of alcohol, and 1<> <\c. of water). On the Other hand, no reaction

was observed when the solvenl was ether or chloroform. One gram of
each of the reagents was dissolved in 50 <■■>•. of ether (care was taken

that the Substances wen: free from even a trace of BCBtic acid i. and

allowed to Btand twenty-four hours. As do precipitate had formed, the

ether was allowed to evaporate spontaneously, when a reddish white res-
idue was left, which looked like a mixture of the unaltered reagents, and

»lved completely in bol chloroform. This showed that none of the
hexabromorthoquinopyrocateohine ether had been formed) as it is insola-

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 221

ble in chloroform. Another portion of the residue from the ethereal
solution was dissolved in glacial acetic acid, and slightly diluted, when a
copious deposit of the red body appeared on standing. The fact that we
observed no action when the tetrabromorthoquinone and tetrabrompyro-
catechine were mixed in ether is strange, since Zincke* states that by the
action of these two bodies in ether or benzol he obtained black shiuiucr
needles, which changed color at 1 10° and sintered together to a brownish
liquid. We are unable to explain this difference in the results of the
experiments, but hope that future work may lead us to the conditions
necessary for preparing the black substance.

If the substance is formed by the reaction given above, it is probable
that the two benzol rings are connected by two of the atoms of oxygen ;
but this is not necessarily the case, as the product might also be a deriv-
ative of diphenyl, in which both rings contain orthoquinone groups. To
settle the mode of union of the two benzol rings, we submitted the sub-
stance to reduction with sodium amalgam as follows : Ten grams of the
red substance were covered with water in a flask, from which all the air
was excluded by a stream of carbonic dioxide, 3 per cent sodium amal-
gam was then added in large excess, and the reaction was allowed to
run at ordinary temperatures in the atmosphere of carbonic dioxide for
several days. At the end of this time the red substance had been con-
verted principally into a purplish white body, which seemed to resist the
further action of the sodium amalgam ; it was, therefore, filtered out, and
washed with dilute acid and alcohol ; the aqueous filtrate from this sub-
stance gave no precipitate on acidification, but on subsequent extraction
with ether we obtained from the extract a considerable quantity of a white
residue, which after several crystallizations from benzol gave the melting
point 100° instead of 104°, the true melting point of pyrocatechine ; from
the form of its crystals, however, and as it gave the lead salt and the
characteristic color with ferric chloride, there can be no doubt that it was
pyrocatechine. As has been stated, the principal product of this reduc-
tion was the purplish white insoluble substance, the weight of which
amounted to 6 grams from 10 grams of the red body; it was obvious,
therefore, that no safe inference in regard to the constitution of the red
substance could be drawn from the appearance of this small proportion
of pyrocatechine. Accordingly, in the hope of increasing our yield of
pyrocatechine, this purplish white product was submitted to reduction
with sodium amalgam under the conditions used in its formation, the

* Eer. d. chem. Ges., XX. 1778.

222 PROCEEDINGS OF THE AMERICAN ACADEMY.

action in this case being allowed to continue for a week, but, as we bad
reason to expect from our observations in our first reduction, the .-ab-
stain•• remained essentially unaltered, not more than a trace of pyrocate-
chine being formed. Meanwhile we had discovered that the purplish
white product was converted by nitric acid into the original red substance,
and, therefore, we oxidized the unaltered body from the last reduction,
and on reducing this again with sodium amalgam, obtained a fresh amount
of pyrocatechine with the purplish white product again. 15y repeating
these alternate oxidations and reductions we have succeeded in convert-
ing a -ample of the red substance completely into pyrocatechine ; a result
which, it seems to us, prove- that tins red body is a pyrocatechine ether
and not a diphenyl derivative.

The purplish white product of the reduction of hexabrornorthoquino-
pyrocatechine ether by sodium amalgam mentioned frequently in the
preceding paragraph proved to have the same solubilities as the mother
substance; that is, it was insoluble in all the common solvents except
uitrobenzol. It was accordingly recrystallized three times from boiling
nitrob z 1, when, as it does not have a definite melting point, it was
assumed to be pure. Alter extraction with alcohol to remove adher-
ing uitrobenzol it was dried at 100°, and analyzed with the following
result : —

0.1215 gram of the substance gave by the method of Carius 0.1974 gram
of argentic bromide.

Calculated f<>r p„„„ i

CltBr6(0B FounJ-

Bromine 69.56 69.16

Properties. — The Bubstance crystallizes from hot nitrobenzol in very
small chocolate brown crystals, which appear to be short, rather blunt
needles. As obtained from reduction it has a purplish white color and a
characteristic silky lustre. It does uol melt at 800°. It i- insoluble in
all the common solvents except nitrobenzol. An aqueous solution of
Bodic hydrate does not dissolve it: fuming nitric acid convert- it back
into the red Bubstance, from which it was mad'-, as is proved by the
following experiment : —

Three grams of the chocolate colored crystals of the reduction product
were mixed with 20 <•.<•. of glacial acetic acid, and •"» c.c. of fuming nitric
acid were added. The -olid immediately tuned red, and. after the mix-
tore had been thoroughly Btirred, it was poured into water, and washed.

JACKSON AND KOCH. DERIVATIVES OF ORTHOBENZOQUINONE. 223

The oxidation was then repeated in the same way, and the dried product
was found to weigh three grams ; the conversion from the chocolate-
colored substance to the red body, therefore, takes place quantitatively.
The red product was recrystallized from warm nitrobenzol, freed from
the excess of nitrobenzol with alcohol, dried at 100°, and analyzed with
the following result : —

0.1317 gram of the substance gave by the method of Carius 0.2140 gram
of argentic bromide.

Calculated for ut„™,»

C12Br604. Found-

Bromine 69.76 69.15

There can be no doubt that this substance is the hexabromorthoquin-
opyrocatechine ether described above.

As to the formula of the chocolate-colored product of the reduction
there is a certain amount of doubt. Our analysis shows only that its
composition does not differ to a marked extent from that of the red
mother substance. Its formation by reduction and its conversion to
the original body by nitric acid would indicate that it was the hydroxyl
compound corresponding to the red quinone, but, on the other hand, its
insolubility in sodic hydrate is hard to reconcile with this view. As
this may be due, however, to the very slight solubility of the substance
in all solvents, which might prevent the sodic hydrate from coming
sufficiently in contact with it to react, we are inclined to consider that it
is the hexabromorthodioxvpyrocatechine ether, and to assign it provision-
ally the formula C12Br602(OH)2. We hope that the study of this sub-
stance may be finished in this Laboratory in the near future.

Our attempts to decompose the hexabromorthoquinopyrocatechine
ether with sulphuric acid led to no result, as no action was observed
even after boiling it for several hours with sulphuric acid of 1.44 or
even 1.6 specific gravity; but hydrochloric acid reacted upon it, as is
shown by the following experiment : 5 grams of the red body were
sealed in a tube with 25 c.c. of strong hydrochloric acid, and heated
from 160°-175° for five hours ; the product, after washing with water
and alcohol, showed a strong resemblance to the purplish white body
obtained by the reduction with sodium amalgam. On treating it with
fuming nitric acid diluted with glacial acetic acid it turned red at
once, so that there can be little doubt of the identity of the two
substances. This observation can be satisfactorily explained on the
theory already adopted, that the purplish white substance is the dihy-

224 PROCEEDINGS OF THE AMERICAN* ACADEMY.

droxv derivative from the red orthoquinone body, as then this action is
the familiar reduction of quinone oxygen by hydrochloric acid. This
theory is still further supported by the reduction of tetrabromortho-
quinone to the purplish white body by tribronipyrogallol described in the
next section.

Behavior of Tetrabromorthoqu'nonc with Trihromresorcine.

After we had Found, as just described, that tetrabromorthoquinone
reacted with tetrabrompyrocatechine, it became of interest to try the
action of the tetrabromorthoquinone with other diatomic phenols, and,
as we found that unsubstituted phenols of this class reduced the quinone,
— pyrocatechine, for instance, reducing it completely to tetrabrompyro-
catechine, — we turned our attention to the bromphenols. Of these,
ti-trabromhydroquinone in alcoholic solution acted oidy as a reducing
agent, giving bromani] and tetrabrompyrocatechine. Tribrompyrogallol
also acted only as a reducing agent, giving tetrabrompyrocatechine, which
combined with unreduced tetrabromorthoquinone to give the hexabrom-
orthoquinopyrocatechine ether, and this was converted by further action
of the tribrompyrogallol into its purplish white reduction product. With
trihromresorcine, on the other hand, a new compound was obtained.

8.4 grams of tetrabromorthoquinone and 7 grams* of trihromresor-
cine f were dissolved with the aid of heat in GOO c.c. of glacial acetic
acid mixed with 400 c.c. of water, and allowed to stand for twenty-four
hours. At the end of this time a dingy pink precipitate had formed,
which was filtered out, washed, and recrystallized from benzol, until
it showed the constant melting point 217°, when it was dried at 100°,
and analyzed with the following results: —

I. 0.1389 gram of the substance gave by the method of Carina 0.2173
gram of argentic bromide.
II. 0.1370 gram of the Bubstance gave by the method of Carius 0.21 I I
gram of argentic bromide.

Calculated for Found.

I Br B0B0 0 Hr.O,. I II-

Bromine 66.94 66.58 CG.fi 1

* Tin- two reagent* were Died in the proportion of equal molecules, whereas
the analysis of the product showed that two molecules of the resorcine should
have been taken for each molecule of the quinone.

I The trihromresorcine was prepared by the method of Benedikt (Monatsh. f.
:i [V. 227.) by mixing the calculated amount of bromine with resorcine dii
solved in glacial acetic add. The product was purified by crystallization from
benzol and ligroin.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 225

The substance, therefore, is evidently formed by the union of two
molecules of the resorcine with one of the quinone with elimination of
two molecules of hydrobromic acid, and may be called the ditribrom-
metoxyphenyldibromorthoquinophenylene ether.

Properties. — It crystallizes from benzol in slender lemon-yellow prisms
with square ends, and melts at 217°. It is soluble in ether; slightly
soluble in cold alcohol or chloroform, easily soluble in either of these
solvents when hot; nearly insoluble in cold benzol, much more soluble in
hot benzol, which is the best solvent for it. It does not dissolve in sodic
hydrate, which throws some doubt on the formula we have ascribed to
it, as this contains two hydroxyl groups.

Action of Glacial Acetic Acid on the Tetrahromorthoquinone.

As has been already stated, tetrahromorthoquinone, when crystallized
from glacial acetic acid, is partially converted into a white substauce *
which we found it convenient to prepare as follows: 10 grams of
tetrahromorthoquinone were dissolved in 80 c.c. of glacial acetic acid and
the solution slowly evaporated to dryness on the steam-batb, the reddish
white residue was moistened with 10 c.c. of glacial acetic, and the slow
evaporation on the steam-bath repeated, in order to act on any tetra-
hromorthoquinone which might have escaped the previous reaction ; the
residue was then extracted with 200 c.c. of hot alcohol, and filtered boiling,
when the red hexabromorthoquinopyrocatechine ether previously described
was left behind, as it is insoluble in alcohol. This substance seemed to
be a constant product of the reaction, but not a principal one, as the
amount was only about 0.8 gram. The alcoholic solution was allowed to
cool, and filtered again ; then it was diluted with 200 c.c. of water, which
precipitated out the white product, amounting to from 5-6 grams.
This was purified by crystallization from glacial acetic acid, until
it showed the constant melting point 230°, when it was dried at 100°,
and analyzed with the following results : —

I. 0.2658 gram of the substance gave on combustion 0.1903 gram

of carbonic dioxide and 0.0179 gram of water.
II. 0.1373 gram of the substance gave by the method of Carius 0.2306
gram of argentic bromide.

* Zincke (Ber. d. chem. Ges. XX. 1777) converted tetrahromorthoquinone into
a white substance by crystallizing it from a mixture of etlier and ligroin. On re-
peating the experiment we did not succeed in finding the conditions under which
the white product is formed.
vol. xxxvi. — 15

226 PROCEEDINGS OF THE AMERICAN ACADEMY.

III. 0.301.1 gram of the substance gave on combustion 0.2090 gram of

carbonic dioxide and 0.0223 gram of water.

IV. 0.1265 gram gave 0.2122 gram of argentic bromide.
V. 0.1279 gram gave 0.2147 gram of argentic bromide.

VI. 0.1 138 gram gave 0.101 1 gram of argentic bromide.
VII. 0.122'.' gram gave 0.2074 gram of argentic bromide.

IV. V. VI. VII.

Found

I.

ii.

III.

Carbon

19.53

is. 90

I Ivd;

0.75

0.82

Bromine

71.45

71.39 71.44 71.48 71.83

Of these analyses I. and II. were of samples prepared and purified as
described : III. to VII. were of specimens prepared earlier in our work,
and in some cases purified by slightly different, and perhaps less effective,
methods.

Two determinations of the molecular weight were made by the method
of freezing, the solvent being benzol.

I. 0.2G2 gram of the substance dissolved in 16.58 grams of benzol

produced a depression of 0.099°.

II. 0.594 gram of the substance in 16.58 grams of benzol produced a

depression of 0.219°.

I. ii.

Molecular Weight 782 802

The moit probable formula to be derived from these data is C14IIoBrs06,
formed by the union of two molecules of tetrabromorthoquinone with one
of acetic acid, one molecule of water being eliminated. The result s,
however, agree almost as well with C14lirb08, although the percentage of
hydrogen found is high for a substance which is free from this element.
A compound with the formula Culh\06 would he formed by the removal
of two atoms of hydrogen from the preceding compound, a reaction which
might easily be brought about by the oxidizing action of a portion of the
tetrabromorthoquinone; the tetrabrompyrocatechine thus formed in turn
acting on the unaltered quinone would give the red body which, as
already stated) i^ always produced with the white Bubstancej but it
should lie remembered that the quantity of the red body formed is far
below that which would be expected from such a reaction.

JACKSON AND KOCH. — DERIVATIVES OF ORTHOBENZOQUINONE. 227

Calculated for
C14H2Br805.

i.

Found

II.

Calculated for
CuBr805.

Carbon

18.88

19.53

18.92

Hydrogen
Bromine

0.22
71.95

0.75

71.45

72.07

Molecular Weight

890

782

802

888

It is obvious that the decision can be made between these formulas only
by the chemical study of the substance, but unfortunately we have not
succeeded as yet in decomposing it by any of the reagents we have tried,
so that we must leave this question undecided for the present, with the
statement that we consider the formula C14H2Br805 the most probable.
The study of the substance will be continued in this Laboratory next
year.

Properties of the Substance C14H2Brs05. — When crystallized from
glacial acetic acid it forms small, almost square rhombic plates with a
pearly lustre, which melt at 230°. It is easily soluble in ether, benzol,
chloroform, acetone, or ethyl acetate ; almost insoluble in glacial acetic
acid, but its solubility is increased by heat, as the boiling acid takes up
from two to three per cent of the substance. Its solubility in ethyl alco-
hol is similar to that in glacial acetic acid ; curiously enough, when first
formed, the crude mass is rather freely soluble in alcohol, but, as the
purification continues, it becomes less and less so, until, when pure, it is
nearly insoluble in cold alcohol, as already stated. It is essentially in-
soluble in ligroin or in water, hot or cold. The best solvent for it is
glacial acetic acid. None of the strong acids dissolve it when cold ; nor
does even a hot solution of sodic hydrate show any perceptible action
with it.

This substance is decidedly stable, as is shown by the following exper-
iments, which were tried in the hope of decomposing it into simpler
bodies, in order to throw some light on its nature: 0.5 gram of the
white substance were treated with sodium amalgam and water in an atmos-
phere of carbonic dioxide for eight days. At the end of this time the
substance was recovered unaltered, as shown by its melting point. The
filtrate was acidified, and extracted with ether, but yielded no pyrocate-
chine. 0.2 gram of the substance were dissolved in boiling glacial acetic
acid, and mixed with 5 c.c. of fuming nitric acid also diluted with glacial
acetic acid ; no visible reaction took place, and upon dilution with water
the original substance was obtained, as shown by the melting point.
0.2 gram were dissolved in cold chloroform, and 3 c.c. of bromine added ;
after standing for twenty-four hours the chloroform was evaporated,

228 PROCEEDINGS OF THE AMERICAN ACADEMY.

leaving the original substance, as shown by the melting point. 0.2 gram
of the Bubstance were dissolved in benzol, and an excess of aniline added ;
after Btanding for twenty-four hours the benzol and aniline were re-
moved, when the unaltered substance was obtained, as shown by tlie
melting point. Phenylhydrazine gave a similar negative result. No
action could be observed when the substance was treated with an alka-
line solution of potassic permanganate, or was boiled with sulphuric acid
of specific gravity 1.44 or 1.6, or heated to 150° in a sealed tube
with hydrochloric acid. If it was heated to 250° with hydrochloric acid,
the compound was completely destroyed, and complete destruction also
ensued when it was treated with fuming nitric acid, or a solution of
chromic anhydride, or when it was fused with potassic hydrate. In con-
tinuing this work next year renewed attempts will be made to obtain
decomposition products of the substance, and efforts will also be made to
throw light on its constitution by preparing allied compounds.

It should also be mentioned that aeetacetic ester converts tetrabrom-
orthoquinone into tetrabrompyrocatechine, which then forms hexabrom-
orthoquinopyrocatecbine ether with unaltered tetrabromorthoquinone.
Chloral showed no action with tetrabromorthoquinone even at 100°.
The study of the little-known and very reactive tetrabromorthoquinone
will be continued in various directions in this Laboratory.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 13. — November, 1900.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

ON THE ACTION OF SODIC SULPHITE ON

TRIBR OMDINITR 0 BENZOL AND

TRIBE OMTRINITR OBENZOL.

By C. Loring Jackson and Richard B. Earle.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF

HARVARD COLLEGE.

ON THE ACTION OF SODIC SULPHITE ON TRIBROM-
DINITROBENZOL AND TRIBROMTRINITROBENZOL.

By C. Loring Jackson and Richard B. Earle.

Presented May 9, 1900. Received October IT, 1900.

In the course of an extended study of the action of various reagents on
tribromdinitrobenzol (Br3 1.3.5. (N02)2 2.4.) it became of interest to
determine whether this substance could be converted into a sulpho-acid
by the method of Strecker,* that is, boiling the halogen compound with
an aqueous solution of normal sodic sulphite, and experiments in this
direction were tried rather early in the work, with negative results.
Upon returning to the subject later we did not find any conditions under
which an aqueous solution of sodic sulphite acted on tribromdinitrobenzol,
but when we substituted alcohol for water as the solvent, we obtained a
reaction. This, however, took a different direction from that which we
had expected, as under these conditions the sodic sulphite acts as a re-
ducing a^ent, converting the tribromdinitrobenzol into dibromdinitro-
benzol by replacing one of the atoms of bromine by hydrogen. The
atom of bromine replaced is the one between the two nitro groups, since
the dibromdinitrobenzol formed melted at 117°, which is the melting
point of the dibromdinitrobenzol f Br21.3.(N02)24.6. That the product
had this constitution was confirmed by its conversion into the bromaui-
lidodiuitrobenzol, which showed the melting point 157°. t Tribromtri-
uitrobenzol undergoes a similar reduction with an alcoholic solution of
normal sodic sulphite, the product being dibromtrinitrobenzol, which can
have only this constitution : Br2 1.3. (N02)3 2.4.6. It is a new sub-
stance, and melts at 135°.

* Ann. Chem. (Liebig), CXLVIII. 90.

t Koerner, Gazz. Chim., 1874, 305.

\ Jackson and Cohoe, These Proceedings, XXXVI. 75.

232 PROCEEDINGS OF THE AMERICAN ACADEMY.

In the hope of finding an easier method for making this dibromtrinitro-
benzol the action of dibroradinitrobenzol (Bra 1 . 8 . (NOa)a 4 . 6) with
nitric and sulphuric acids was studied. This gave the desired dibromtrini-
trobenzol, hut the product was mixed with bo large a proportion of tribrom-
dinitrobenzo] that the method was not adapted to the preparation of the
substance. This tribromdinitrobenzol is a new substance, nu Iting at 150°.
There can be no doubt that its constitution is Br3 1 . 2.3. (NOa)a4.6,
Bince it is made bv introducing a third atom of bromine into the dibrom-
dinitrobenzol Bra 1 . 3 . (NOa)a 4 . 6, and, if this occupied the only other
vacant place (5), the product would be the common tribromdinitro-
benzol, which melts at 192°. The formation of this tribromdinitrobenzol
during this reaction is strictly according to the analogy of the behavior
of the tribromdinitrobenzol, which, when boiled with a mixture of fum-
ing nitric and sulphuric acids, gives rise to tribromtrinitrobenzol and
tetrabromdinitrobenzol. *

As neither of the new substances just described can he prepared easily
on a large scale, we have confined our work on them to the study of the
principal action of sodic ethylate on each. This reaction runs in a nor-
mal way. giving styphnic <lieth\ lether,

C6H(NOa)8(OCaHfi)a, ((OCaH6)a 1.3. (NOs)8 2.4.6),

with dibromtrinitrobenzol, recognized by its melting point 121° ; and
with tribromdinitrobenzol a new bromdinitroresorcine monoethylether
melting at 78°. The subject did not promise to be of sufficient interest
to make it worth while to study the secondary products of these reactions.

Experimental Part.

Action of Sodic Sulphite in Alcoholic Solution on Tribromdinitrobenzol.

Five grams of symmetrica] tribromdinitrobenzol ( Bra 1 .3.5. (N02)2
•_' . 1 ) were dissolved in 70 c.c. of alcohol mixed with 20 c.c. of benzol,
li\ e gram- of normal sodic Bulphite added, and the mixture boiled for five
hours iii a flask with a return condenser. The red liquid thus obtained
was filtered from the insoluble portion, evaporated to dryness, and the
residue washed with water, until the yellow color Mas removed. The
part insoluble in water was then recrystallized from alcohol with the aid
of bone-black, until it showed the constant melting point 117°. This
melting point indicated that it was Koerner's dibromdinitrobenzol t

* Jackson and Wing \.\I1 189

izz. Cliira , 1874, 805.

JACKSON AND EARLE. TRIBROMTRINITROBENZOL. 233

(Br2 1 . 3,(N02)24 . 6.), and this was confirmed by treating it with
aniline in the cold, when it was converted into the bromanilidodinitro-
benzol, * CGH2Br(C6H5NH)(N02)2, recoguized by its melting point 157°.
For greater certainty the dibromdiuitrobenzol was dried at 100°, and
analyzed with the following results : —

I. 0.3277 gram of the substance gave 24.6 c.c. of nitrogen at a tempera-
ture of 18°. 5 and a pressure of 760.3 mm.
II. 0.3195 gram of the substance gave by the method of Carius 0.3707
gram of argentic bromide.

Calculated for Found.

C6H2Br2(N02)3. I. II.

Nitrogen 8.59 8.65

Bromine 49.38 49.08

The sodic sulphite, therefore, has removed the atom of bromine stand-
ing between the two nitro groups in the tribromdinitrobenzol, and
replaced it by hydrogen.

That this was not the only action of the sodic sulphite was shown by
the marked yellow color of the products of the reaction. The yellow
substance was soluble in water, but, in spite of its strong color, was present
in such small quantity that we could not isolate it from the other sub-
stances soluble in water, which consisted of sodic bromide and unaltered
sodic sulphite. It was undoubtedly the sodium salt of a phenol formed
by a secondary reaction.

Action of Sodic Sulphite in Alcoholic Solution on Tribromtri>iitrobenzol.

Ten grams of symmetrical tribromtrinitrobenzol (Br, 1.3.5. (N02)3
2.4.6) were dissolved in 120 c.c. of common alcohol with the assistance
of 30 c.c. of benzol, ten grams of normal sodic sulphite added, and the
mixture boiled for four hours in a flask fitted with a return condenser.
The reaction seemed to run as in the case of the tribromdinitrobenzol,
except that it took place more easily, since the red color appeared more
quickly. The red solution was filtered, the filtrate evaporated to dryness,
and the residue washed with water, after which it was recrystallized from
a mixture of alcohol and benzol with the assistance of bone-black, until it
showed the constant melting point 135°, when it was dried at 100°, and
analyzed with the following results : —

* Jackson and Cohoe, These Proceedings, XXXVI. 75.

234 PROCEEDINGS OF THE AMERICAN ACADEMY.

I. 0.2384 gram of the substance gave 25.1 cc. of nitrogen at a temper-
ature of 2o° and a pressure of 753.3 nun.
II. 0.2887 g • tm of the substana by the method of C'arius 0.2910

gram of argentic bromide.

1 fur

1,11.

i.

Nitrogen

11.36

11.65

Bromine

13.10

Found.

II.

12.89

The substance is, therefore, dibromtrinitrobenzol ( Brsl. 3.1 N( > ,2. I. 6 I,
formed by the replacement of one atom of bromine in the tribromtrini-
trobenzol by hydrogen, and the Bodic sulphite bas behaved like a reduc-
ing agent in this case just as it did with the tribromdinitrobenzol.

Properties of Metadibrom-s-trinitrobenzol. — It crystallizes from ben-
zol -or better, a mixture of benzol and ligroin — in broad prisms termi-
nated by two planes at a slightly acute angle to each other, or in slender,
blunt-ended prisms, or in rhombic crystals. It has a pale yellow color with
a slight greenish tinge, and melts al 135°. It is freely Boluble in benzol
or ether; soluble in chloroform, acetone, or cold alcohol, freely soluble
in hot alcohol ; slightly Boluble in hoi ligroin; insoluble in water. The
best solvent for it is a mixture of alcohol and benzol, or of benzol and
ligroin ; strong hydrochloric or Bulphuric acid has no action on it. either
hot or cold. Strong nitric acid due-, not dissolve it in the cold, hut
form- a yellow solution with it when hot, which deposits the unaltered
substance on cooling. Sodic hydrate solution gives a pale yellow solu-
tion when hot: no perceptible action when cold. Alcoholic sodic
hydrate gives a deep red color.

The aqueous wash waters obtained in the preparation of dibromtri-
nitrobenzol by this process, although showing a Btrong red color, con-
tained so little organic matter that we did not Bucceed in isolating any
new compound from them. The residue left mi evaporating them
consi-:rd chiefly of Bodic bromide and unaltered sodic Bulphite.

Preparation of Dibromtrinitrobenzol from Dibromdinitrobenzol,

In the hope of finding an easier method of preparing the dibromtrini-
trobenzol we next tried to make it from dibromdinitrobenzol. The
dibromdinitrobenzol was mad'- by the method of Jackson and Cohoe, *
which, however, we modified Bomewhat in converting the bromacetanilid

* These Proce< dUi _■-. \\\vi. 77.

JACKSON AND EARLE. — TRIBROMTRINITROBENZOL. 235

into dibroiuacetanilid, — the monobromacetanilid was mixed with enough
glacial acetic acid to convert it into a semi-liquid mass, to which the cal-
culated amount of bromine was added from a burette. It was then
warmed on the steam-bath, until it was transformed into a clear,
dark-red liquid, which was poured into water, filtered, and the pre-
cipitate ground in a mortar with a solution of sodic hydrate to remove
the excess of bromine. The dibromdinitrobenzol used for our work was
recrystallized until it showed the correct melting point 117°.

To convert the dibromdinitrobenzol into dibromtrinitrobeuzol we pro-
ceeded as follows. Twenty grams of dibromdinitrobenzol were added to
500 c.c. of fuming nitric acid of specific gravity 1.50 and 200 c.c. of
sulphuric acid of specific gravity 1.86, aud the mixture was boiled
violently for about three hours in a flask closed with a porcelain
crucible. Toward the close of the operation a reddish oil separated,
after which the process was continued only so long as the liquid
boiled freely, since decomposition and darkening of the color resulted
from too long continued boiling. After the mixture had been boiled
for a sufficiently long time, it was allowed to cool, poured into ice water,
and filtered. The precipitate was recrystallized from a mixture of
alcohol and benzol, until it showed a constant melting point, but, as
this stood at 150°, the substance was not the expected dibromtrinitro-
benzol, which melts at 135°. On examining the mother liquors from
this substance, however, we found the dibromtrinitrobeuzol, which, after
purification by crystallization from a mixture of benzol and ligroin, was
recognized by its melting point 135°, and the following analysis of the
substance dried at 100° : —

0.2174 gram of the substance gave by the method of Carius 0.2203
gram of argentic bromide.

C6HBr„(N02)3.

Bromine 43.10 43.10

The dibromtrinitrobenzol is therefore formed by this process, but so
much of the other product melting at 150° is formed at the same
time that this is not a convenient method of preparing it, and the
dibromtrinitrobenzol remains a not easily accessible substance.

Tribromdinitrobenzol, C6HBr3(N02)2. (Br, 1. 2. 3. (N02)2 4. 6.).

The substance melting at 150°, prepared by the action of nitric acid
and sulphuric acid on dibromdinitrobenzol as described in the preceding
section, was dried at 100°, and analyzed with the following results : —

236 PROCEEDINGS OF THE AMERICAN ACADEMY.

I. 0.2 tm of substance gave according to the method of Carius

0.4138 gram of argentic bromide.
II. 0.2987 gram of substance gave 0.4150 gram of argentic bromide.

C&lcnUted for

Found.

C Hi:. \o,),.

i.

11.

59.26

59.24

59.13

Bromine

The substance is, therefore, a tribromdinitrobenzol) and its appearance

i> not unexpected, since J. F. Wing and one of us* found that tetra-
bromdinitrobenzol was always formed during the preparation of tribrom-
triuitrobeuzol from tribromdinitrobenzol. There can be no doubt about
the constitution of this tribromdinitrobenzol, because it is made from the
dibromdinitrobenzol (Bra 1. 3. (N08)2 4. 6) by the introduction of an
atom of bromine, wbicb can enter only in two places, — between the
nitro groups, or between tbe atoms of bromine. If tbe substitution
takes place between the nitro groups, there must be formed symmetrical
tribromdinitrobenzol, which melts at TJ23 ; there is left, therefore, for
our substance melting at 150° only the constitution Br3 1. 2.3. (NOs)a4.6.
Properties of the 1.2.3. Trihrom- 4. 6. dinitrobenzol. — Yellowish
white rectangular plates bevelled on tbe sides, when crystallized from a
mixture of alcohol and benzol. Melting point 150°. It is freely
soluble in benzol, or ether; soluble in chloroform, glacial acetic acid, or
;oue; soluble in cold alcohol, freely soluble in hot alcohol; slightly
soluble in cold ligroin, soluble in hot ; insoluble in water. The best
solvent for it is a mixture of alcohol and benzol. Strong hydrochloric
or sulphuric acid gives no visible action hot or cold. Strong nitric acid
appears not to act in the cold, but dissolves it when hot, depositing the
unaltered substance as it cools. A solution of sodic hydrate does not
act on it apparently either hot or cold, but in presence of alcohol gives a
light yellow color.

Action of Sudir Ethylate >>n Dibromtrinitrobenzol.

L.5 grams <>f dibromtrinitrobenzol (Bra1.3.(NOs 2.4.6) were
lived in 1 •"» c.c. of benzol, and mixed with sodic ethylate in the pro-
portion of three molecules of the ethylate to each molecule of the dibrom
compound. The Bodic ethylate was prepared by adding the calculated
niit of sodium to •'!" c.c. of absolute alcohol. The liquid .-it once
took on an intense red color, and became turbid. To make sure of com-
pleting the reaction tbe mixture was allowed to stand for twelve hours al

ordinary temperature in a cork flask, alter which it was lillered, and the

• These Proceedings, Will. 1:39.

JACKSON AND EARLE. TRIBROMTR1NITROBENZOL. 237

filtrate allowed to evaporate spontaneously ; the residue was washed with
water until the washings were colorless, and then recrystallized from
alcohol until it showed a constant melting point, which was found to be
121°. The melting point of styphnic ethylether,

C6H(OC2H5)2(N02)3 (OC2H5)2 1.8. (N02)3 2.4.6,
is given as 120°. 5, and it is described as forming long plates, which
quickly turn orange-brown in the light ; our substance crystallized in
prisms connected by their longer sides, and turned brown by exposure to
the light. It contained no bromine. There can be no doubt, therefore,
that the two bodies are identical, and that this styphnic ether formed by
replacing each bromine by an ethoxyl group is the principal product of
the reaction. It was not, however, the only product of the action of sodic
ethylate on dibromtriuitrobenzol, as the reaction product insoluble in the
organic solvents but soluble in water gave tests for nitrite as well as
bromide, and the aqueous washings were colored ; but the amount of these
secondary products was so small that they could not have been studied
without preparing the mother substance on a very large scale, and the
importance of the subject did not warrant the great expenditure of time
which would have been necessary for this purpose.

Action of Sodic Ethylate on Adjacent Tribromdinitrobenzol.
Seven grams of the tribromdinitrobenzol (Br3 1 . 2 . 3 . (N02)2 4. 6)
were dissolved in 70 c.c. of benzol, and mixed with sodic ethylate in the
proportion of three molecules of the ethylate to each molecule of the tri-
bromdinitrobenzol. The sodic ethylate was prepared by adding the cal-
culated amount of sodium to 70 c.c. of absolute alcohol. A deep red
turbid solution was formed immediately, which was allowed to stand in
a corked flask at ordinary temperatures for two days, and then filtered,
the filtrate allowed to evaporate spontaneously, and the residue washed
with water, until the washings were colorless. The portion insoluble in
water after recrystallization from alcohol showed the constant melting
point 58°, and crystallized in long white needles turning yellow in sun-
light, but was formed in such small quantity that there was not enough
for analysis ; we, therefore, sought the principal product of the reaction
in the portion soluble in water. The precipitate formed during the
action of the sodic ethylate was dissolved in water, and mixed with the
wash waters, which had previously been concentrated to a convenient
bulk. The solution was then acidified with acetic acid, which threw
down a black tarry precipitate. This was filtered out, washed, sus-
pended in water, and dissolved by adding sodic hydrate to neutralization ;

IJ11HJJJU1J.

Calculated for
[CoIIBrCNOjJjlOCjH^OLBa.

i.

Barium

18.33

17.91

Nitrogen

7.47

238 PROCEEDINGS OF THE AMERICAN ACADEMY.

the Bolation was filtered, and treated with baric chloride, which produced
at once a mass of shining yellow needles. These were filtered out,
washed with water, and rcerystalli/.ed several times from alcohol; after
which they were dried at LOO . and analyzed with the following results :

I. 0.363G gram of the substance gave 0.1108 gram of baric sulphate.
II. 0.3071 gram of the Buhstance gave 21.2 c.c. of nitrogen at a tempera-
ture of 22°.5 and a pressure of 7t'>.~i mm. The Bubstanee was
mixed with a large excess of a mixture of eight parts of fused
plumbic chromate with one of potassic dichromate. It was neces-
sary to heat very gradually to avoid too rapid an evolution of the

Found.

II.

7.8G

The substance is, therefore, the barium salt of a bromdinitroethoxy-
phenol. We have not determined the constitution of this substance
experimentally, but there can be little doubt that it is a derivative of
resorciue, as the two bromine atoms in the meta position to each other
are also in the para and ortho positions to the nitro groups, and this has
been shown to be the position most favorable to replacement ; whereas
the third atom of bromine, which is in the meta position to the two nitro
groups, would be replaced with difficulty according to all previous work
on this subject; we have no hesitation, therefore, in calling this substance
the bromdinitromonoethylether of resorciue.

Properties of (he Barium Salt — It crystallizes from alcohol in fine
yellow needles. It is nearly insoluble in water, whether hot or cold,
but soluble in alcohol. It is -table at 100°, but if heated suddenly to a
high temperature it explodes. Acetic acid does not decompose it.

The tree bromdinitromonoethylether of resorciue was obtained by
heating the barium salt with an acid : after crystallization from alcohol it
melted constant at 78°. It form- long white feathery needles, which
turn yellow on standing.

The experiments just described make it highly probable that the Bub-

nce melting at 58 and insoluble in water is the bromdinitroresorcine-
diethylether. There were al.-o other products of the reaction, as the
soluble Baits obtained gave a test for nitrite as well as for bromide, but
they were formed in such small quantity that it did not seem to us worth
while to undertake the study of them.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 14.— January, 1901.

FALSE SPEQTRA FROM THE ROWLAND CONCAVE

GRATING.

By Theodore Lyman, Ph.D.

With a Plate.

Investigations on Light and Heat made and published wholly or in part with Appropriations

from the rumford fund.

FALSE SPECTRA FROM THE ROWLAND CONCAVE

GRATING.

By Theodore Lyman.

Presented October 10, 1900, by W. C. Sabine. Received October 26, 1900.

It is proposed to show in the following paper that among the spectra
formed by the Rowland concave gratings there are spectra not accounted
for by the ordinary theory of the grating ; that such spectra are common,
and at times fairly strong and of excellent definition ; that these spectra
are diffraction spectra, of much less dispersion than the ordinarily recog-
nized spectra, and that the errors of ruling to which they are due are not
local, but general to the whole surface of the grating. Finally, it is pro-
posed to explain an experimental method by which these false lines can be
sorted out from the regular and calculable overlapping spectra. These
lines are especially dangerous in series spectra work, giving a some-
what systematic reproduction of strong lines and groups, which reproduc-
tions in actual vibration frequencies do not exist. There is probability
and some evidence that such errors have been committed in the past, and
it was in the presence of this danger that the false spectra were here
discovered.

In the year 1893, Schumann showed that the ultra-violet spectrum
could be extended to the neighborhood of 1000 jx. He used a spectroscope
fitted with a fluorite prism and fluorite lenses. The apparatus was so
constructed that it could be enclosed in a case from which the air could
be exhausted. According to his investigation, it was chiefly the absorp-
tion of the air for light of very short wave lengths that had prevented
other investigators from extending the spectrum below wave length
1800 ft. Owing to the nature of a prism spectrum, it was difficult for
Schumann to determine accurately the wave lengths of the lines which
he discovered. Since his paper no attempt seems to have been made to
measure these lines. It was in an effort to measure these wave lengths
that the false spectra above referred to were first observed.

It seemed probable that if a concave grating could be used to obtain
these ultra-violet lines, an accurate determination of their wave length
vol. xxxvi. — 16

242 PROCEEDINGS OF THE AMERICAN ACADEMY.

would be possible. Accordingly, some two years ago, Mr. E. II. Col pitta
and the writer undertook the in \ «--t i - :tt i. >n. The methods used and the
results obtained were briefly as follows : —

The concave ditfraction grating, >lit, and photographic plate were all
enclosed in an air-tight cast-iron box. Light was admitted to the box
throngb a fluorite window. The ><>urce of light was a powerful electric
spark, obtained from a transformer run from the commercial circuit ; a
capacity equivalent to lour large Leyden jars was placed in parallel with
the spark gap; magnesium was used for terminals. Following Schu-
mann's work, the air was exhausted from the box. The spark was
placed close to the fluorite window in order that the column of air
between the source and the plate should be as short as possible. Under
these circumstances a photograph was obtained Bhowing lines apparently
down to wav.- length 924 /x, a value lower than Schumann's lowest esti-
mate'! wave length. Terminals of aluminum, cadmium, cobalt, nickel,
and copper were tried. All showed lines helow 1.300 /a; some showed
lines as tar down as position 950 /x.

During these experiments, the air pressure iu the apparatus did not
always have the Bame value. Some plates were taken with the pressure
as high as 1 cm. ; others with it as low as .2 cm. It was observed that
this difference in pressure did not have any effect on the strength of the
hues.

Up to this time the first spectrum had been used; now it was desired
to find these lines in the second spectrum. In spite of the greatest care,
however, no trace of any line below 1900 /a could he found iu the second

Bpectrum.

Tin- absence of the Hues in the second spectrum, together with their
behavior under -m ill differences of air pressure, raised the suspicion that
either the lines were not due to lightofa very Bmall wave length, or that
light of very short wave length was not absorbed by the air to the extent
indicated 1>\ Schumann. To test this idea the spark was first removed
to a metre's distance from the fluorite window. This air column did not
reduce the intensity of these new lines. Next, air was admitted to the
box, and -till the lines were unchanged in intensity, even through a col-
umn of air, now five met re- long. Evidently the lines were not absorbed

by air.

Lest the special brand of plate employed was particularly sensitive to

the extreme ultra-violet, a number of different plate- were tried. In each

:t good photograph was obtained, -bowing practically all the lines.

A piece ofwindow-gla placed between the Blil and Bpark, and no

LYMAN. — FALSE SPECTRA FROM THE ROWLAND GRATING. 243

trace of the lines could then be found. Since the light causing these
lines is absorbed by glass, and is not absorbed by air, it would seem to
have a wave length between 3500 /x and 2000 /x, according to the accepted
absorbing power of these two media.

It was now observed that the groups at 2790 /x, 1873^, and 924 jx
were, even to casual inspection, strikingly characteristic and identical in
general appearance. Further, on measuring their dispersions, they were
found to be proportional to the displacements of the grorps from the
slit. Therefore, these groups formed either a remarkable series repro-
duction or a heretofore unobserved diffraction phenomenon. Thus the
issue was definitely raised whether these lines were real first spectrum
lines, produced by light of wave lengths 924 /x, 1873 /x, and 2790 /x,, and
arranged in al, 2, 3 series, or whether they were all produced by light
of one wave length and owed their position to some diffraction phenom-
enon connected with the grating.

To distinguish between these alternatives, recourse was had to the
relative refrangibility of these lines through quartz. To this end a
small angle quartz prism was placed between slit and grating and a
short distance from the slit. The effect of this prism was to produce
virtual images of the slit whose displacements were nearly proportional
to the refrangibility of the light. This device resulted in the displace-
ment of each line in the grating spectra by an amount proportional to
the refrangibility of the light producing that line.

If the lines at 1873 /x and 924 /x were caused by light of the same
wave length, they would suffer equal displacements. If they were
caused by light of different wave lengths, they would be displaced by
different amounts.

The experiment was tried, and it was found that the groups at 2790 p,
1873 /x, and 924 /x all suffered about equal displacements. From 2790 /x
down to the region of 1873 /x, the displacement of the groups increased
as the wave length decreased. There was then a discontinuity in the
rate of increase of the displacement before 1873 /x and a second discon-
tinuity between 1873 /x and 924//.

The most obvious conclusion from these data is that the lines at 1873 /x
and 924 fx are not due to light of wave lengths corresponding to the
positions of the lines in the spectrum. They seem to be caused by
light having a wave length near 2800 /x ; they seem to be reproductions
of some part of the true spectrum, — curious phantom liues due to some
property or imperfection of the grating. An examination of the lines
below 1900^ taken with aluminum terminals fully carried out this

-II PROCEEDINGS op THH AMERICAN ACADEMY.

theory. Before leaving this first grating, which we will call No. 1. it
is important to note that although the groups in magnesium at '27'.*" u,
1878 /tt and 924 /<. closely resemble each other, their wave Lengths are
not exactly in the 1:2:8 ratioj ■_' < 924 = 1848, which differs from
1873 by _'•"..

It was proved that these phantom lines were not due to any local
variation in ruling by Bhielding all hut a narrow strip of grating surface,
ami then moving the screen so as to expose each strip in turn. In this
way a series of photographs were ohtained, each from a different strip
of grating Burface. They all showed the phantom lines with about
equal intensity.

In addition to the grating with which the investigation was begun,
three others were examined. Two of these were of larger radius, —
21 ft.: the other was similar in type to the first instrument. Of the
two large gratings, one showed the phantom Hues distinctly. The
two Btrong groups formerly at 1873 /x and 924 /i were now found at
positions corresponding to 17:.'*/;. and 1079 /t. The other large grat-
ing did not show tin- lines at all. The last small grating showed
the reproduction of group 2790/4 at positions of about 1870 //. and 924/x.
In addition, however, there were also numerous less distinct repro-
ductions scattered between these two positions.

We may now sum up the facts arrived at in these i'vw pages. Of
four concave diffraction gratings examined in this investigation, three
showed the phenomenon of phantom lines. These lines vary in
position with the grating by which they are produced, hut the vari-
ation cannot In- connected either with the variation of radius of curva-
ture of the grating or with the number of lines to the inch. The most
prominent of these phantom lines closely resemble some group in the

true spectrum. This repetition of a true group occurs m08l strongly
twice. The width of these phantom repetitions is proportional to their
apparent wave length. The wave lengths of the true group and its
phantom repetitions bear no simple relation to one another. The
photographs obtained with gratings 1 ami 1 .show a \rvy large number
of taint repetitions of a group or s,-t of groups. These repetitions form

t of background to the main spectrum between 1800 /u and 900 /x..
It may he safely s:(i.l that a Btrong line in the Bpectrum is repeated a
great number of times, and that these phantoms are distributed appar-
ently at random. We have no reason to believe that this repetition is
confined to a certain proportion of the ultra-violet Bpectrum. Everj line
in tl brum is probably reproduced. These reproductions are '-on-

LYMAN. — FALSE SPECTRA FROM THE ROWLAND GRATING. 245

cealed by the strong real lines of the spectrum in places where strong
lines occur. They are chiefly noticeable in the extreme ultra-violet,
where there are no real lines. In parts of the spectrum where gaps
occur, however, these phantoms may be visible and may there be mis-
taken for real lines. Such, in brief, are the results arrived at. A
more detailed account of the apparatus and of the numerical results of
the investigation will now be given. The consideration of the causes
which produce these phantom lines is left for the end of the paper.

Figure 1.

The first grating investigated was one of 180 cm. radius. The plate
was prepared by Brashear and ruled by Rowland's engine in the year
1897 ; it had 14,438 lines to the inch. To economize space a mounting
somewhat different from that of Rowland was used. The grating and slit
were fixed in position at a distance apart equal to the radius of cur-
vature of the grating. The plate-holder was carried by an arm pivoted
at a point on the line connecting slit and grating and midway between
the two. The length of the arm was, of course, half the radius of
the grating. The plane of the grating was so adjusted that the prin-
cipal image fell about 1 cm. to the right of the slit. This was done to
avoid reflection from the polished jaws of the slit. The plate-holder was
constructed to take a plate five inches long and one inch broad. The
form of the box in which the apparatus was enclosed is not material to
this investigation. The spark which served as a source of light was
placed within 3 mm. of the fluorite window. In the first experi-
ments the width of the slit was about .1 mm., and the time of exposure
was about one hour. Later, however, good photographs were obtained
with exposures of from twelve to fifteen minutes. In making tests for
the absorption of the air a quartz lens was placed between the source
and the slit.

In developing photographs of this type, great contrast is the chief
object. The kind of developer used thus becomes important. After
several trials, an Ortol-Soda developer was found to give the best results.

The apparent wave lengths of the different lines which come under
consideration were determined by comparison with the normal sun spec-

.

PROCEEDINGS OK THE AMEBICAN ACADEMY.

tram. A dividing engine with a J mm. screw was used. This screw,
when used to measure lines on a six-foot grating possesses for a run of
20 cm. an ao uracy better than one one-hundredth of an Angstrom unit.

y means of a stationary microscope fitted with
a four-inch objective and micrometer eye-piece. The value of one cen-
timeter of plate length in terms of Angstrom units was determined by
measurements takm between two given Bun lines, located l>v Rowland's
map. Iu order to get ;i photograph of the sun Bpectrum and the maf-
nesium spectrum on the same plate, a shutter was placed before the slit
and so adjusted that the upper or lower half could he exposed at plea-
By this means two narrow Bpectra were obtained, one above the
other. The method is preferable, when working with the firsl spectrum,
to the usual mode of protecting the plate itself by a swinging screen.

When it became necessary t<> test the nature of the light
producing the lines al 1873 /u and '.•:_' 1/j. in the magnesium
Bpectrum, the following arrangement was adopted : The
quartz pri-.ni of an angle of five degrees was placed some 30
cm. distant from the slit, but not on the straight line joining
slit and grating. Thus when the source was in position A.
the light did not fall upon the prism ; but when it was in
position 15, the light passed through the prism. In the Iir.«t
case, the upper half of the slit was closed by the shutter;
in the second, the lower half was closed. The result was,
of course, to produce two spectra, one above the other; the
one due to light which had passed through the prism was
thus shifted toward the red. The general appearance may
be seen from Plate No. I.

It now becomes necessary to consider more in detail the
behavior of the light of these ultra-violet Bpectra when
d through the quartz prism. Tin- results tor magne-
sium may be seen from the following table. The lettering
of the groups on the plans .,!' the spectra is, of course,
arbitrary. The chief or real groups are marked with an
Unsubscripted letter. The reproductions are marked with
letters bearing a subscript The plates are from drawings
made from the original photographs. Beginning with mag-
nesium Plate II., six principal groups of lines are to be
I: they extend from the pair of lines marked Aj, —
length 1966 /a, — to the Btrong group (',. — wave
length 924 /a. B< id< tl rong lines, there are six fainter groups

/>' ./

LYMAN. — FALSE SPECTRA FROM THE ROWLAND GRATING. 2-17

distinctly visible, — A2, th, C2, A3, b2, and Cs. A2 resembles group
Ai, C2 group Cx; C3 resembles group C4, A3 group A4. It is to be
noted that while A2 and C2 lie on the right or ultra-violet side of Ax
and Ci, C3 and A3 lie on the left or red side of C4 and A4. In addition
to the simple groups of lines already mentioned, the whole spectrum,
from one end to the other between Ax and C4, is filled with a number
of fine, faint lines. These lines are not regularly arranged so as to pro-
duce fluting, but rather present the appearance of a reproduction of
groups of lines.

MAGNESIUM.

ALUMINUM.

Group.

Apparent
wave length.

Displacement.

C

2790 m

42.2 mm.

I

2713 ju

43.4 mm.

II

2682 n

43.9 mm.

Ax

1966 fi

42.1 mm.

A2

1924 M

bi

1914 ft

Ci

1873 n

42.2 mm.

c2

1834 p

42.3 mm.

D2

1475 ix

39.8 mm.

E2

1266 m

40.2 mm.

A4

909/4

41.9 mm.

c3

964 ft

b2

944 ix

c4

924/1

42 2 mm.

Group.

I

Bi
Ci

Do

E2
B3
Cs

Apparent
wave length.

2376 n

2075/x

1890 m

1300 m

1184 m

1023 m

933 m

Displacement.

38.8 mm.

36.1 mm.

36.7 mm.

34.9 mm.

35.3 mm.

36.0 mm.

36.6 mm.

If we examine the displacements produced in these groups when the
light is allowed to pass through the quartz prism before falling upon the
grating, the following facts become evident : A steady increase in dis-
placement is observed in the successive lines going from the strong group
at 2790 (C) towards the ultra-violet until we get in the neighborhood of
group At. At this point the amount of displacement suddenly returns
to about the value it had at 2790 fx. Between groups Ax and D2 the dis-

l!ls PROCEEDINGS OF THE AMERICAN ACADEMY.

placement suffers a second discontinuity. Thus Ax gives 42.1 mm., D3

.8 mm. The displacement of C4 is almost exactly that of CV and
very nearly that of group C, — i'7'.'U /t. The figures are 42.25 mm.,
42.25 mm., 12.21 nun. As Btated hefore, the most obvious conclusion
from these data is that the lines A] to C'4 are not due to light of wave
lengths corresponding to the positions of the lines in the spectrum. They
Beem t<> be reproductions of some part of the true spectrum.

( )n the aluminum plate there are present six strong lines, — Bj, CV D2,
1 . I! . (' . — with wave lengths from 2U""» /< to '.».').'! u. There are faint
reproductions of these groups similar in arrangement to those observe. 1
with magnesium. There is a background of fine, indistinct lines. On
allowing the groups to be displaced by the prism, we find group D2 less
displaced than group C\. and group Bj less displaced than a real group
called I.

We may now consider the false nature of the lines as establish, d.
It remains to investigate the causes which produce this diffraction
phenomenon.

It is perfectly evident that the phantoms are not due to a local varia-
tion in the rate of ruling, such as produce ghosts; for we have seen that
the fault is not local in the grating surface. Moreover, we know that a
ghost always oeeurs close to the line of which it is a reproduction.

The phantoms are not due to a false source of light, since the dispersion
of the various reproductions is not the same as that of the real group.
We may discard the hypothesis that the mounting of the grating or the
position of the source of light has any effect. The lines occur with
various forms of mounting and with various positions of lens and source,
as will be seen later. We have to deal with a diffraction phenomenon,
with an inherent property of the diffraction grating itself.

I: seems probable that in developing the theory of the grating some
tmptions have been made which are not according to fact. It is in
the error of such assumptions that we find the solution of our problem.

It is generally assumed in treating the grating thai the lines of the
ruling are of equal width and are separated by equal Bpaces. In the very

nature of things, it is evident that this cannot be the Case in view of

the v.-ry minute distances involved and the almost inconceivable rigidity
which would be necessary. These variations, Blight in absolute amount,
may be a considerable fraction of the distance from line to line. It is
ir, moreover, on experimental as well as <>n theoretical ground, that
these variations are not local, but extend over the whole surface of the
grating.

LYMAN. — FALSE SPECTRA FROM THE ROWLAND GRATING. 249

It appears, then, that the departure from equal spacing assumed in the
theory of the concave grating is not the random departure which is next
in theoretical simplicity, but a more or less systematic departure which
has emphasized disproportionately certain of the subordinate maxima
called for in the conventional theory. If the position of the false maxima
occurred in the ratio 1:2:3, or in any other simple ratio, the phenomena
would invite an analytic discussion. As an experimental fact, however,
our lines are not harmoniously placed with respect to the slit, nor are
they arranged according to any discoverable system. Moreover, they are
different in character and position when obtained by different gratings
ruled by the same engine. It seems most likely that the phantom lines
are due to a number of superposed first spectra of varying dispersions ;
and that these false spectra owe their existence in some manner to the
variation in width and separation of the grating lines. Under these con-
ditions, an analytic discussion seems less profitable than the presentation
of the experimental data.

In order to substantiate the statements just made, it becomes the
next object of the investigation to observe carefully the relative posi-
tions of the phantom lines, not only in one, but in several different
gratings.

We must observe, in looking over the tables already given, that al-
though several groups bear a strong resemblance to each other, and though
the apparent wave lengths of these groups are nearly in one, two, three
ratio, this ratio is never exact.

In aluminum, the line C3 seems to resemble the line C1? but
2 x 933 = 1866, not 1890. The group B3 resembles Bx, but 2B3 =
2046, not 2075. There certainly is no exact 1, 2, 3 relation, although
lines corresponding roughly to 3C3 and 3B3 in wave length may be found
for aluminum. The difference between 2046 and 2075 = 29 Angstrom
units cannot possibly be due to the error of observation. It is perfectly
safe to say that the measurements of wave lengths in the table are correct
to 1 Angstrom unit.

Before examining another grating, there are one or two more points
which need attention. A concave grating produces two spectra of the
first order, which in this form of mounting lie on each side of the slit.
The spectrum on the left had been investigated ; it was now necessary
to observe the spectrum on the right of the slit. The grating was there-
fore turned upside down, by which process the spectrum formerly on the
right of the slit now fell upon the left. Photographs taken showed the
lines Ax, Cu A4, C4, etc., in magnesium as clearly as before. The grat-

250 PROCEEDINGS OF THE AMERICAN ACADEMY.

in^ was next turned about a verticil axis until the direct ima^e of the
slit fell upon the plate. No lines below group (', wen- revealed by this

process, • was tin- position of the lines greatly changed. It must be

noted, however, that the turning of the grating will somewhat distort the
tra and change the value of /u to cm.
The second grating to be investigated was <»f 21 ft. radius, l 1.438
lines to the inch. It was arranged on Rowland's mounting. The in-

tensity of the m tra given by this grating was verv inferior to that

obtained with the 6 ft. instrument. Tin' first Bpectrum was selected to
work with, the magnesium -park was used as ;i source. It was found
necessary to give an exposure of an hour, and to use a slit ."_' nun. in
width. A good deal of care was taken to have the exposure of the
proper length : if too short, the lines themselves would not appear ; if too
long, the- lines would be obscured by fog. With this 21 ft. grating, a
verv little fog would quite shield any very fainl line. After some trouble,
the characteristic reproductions "f group 2790 were found. They now
irred, however, at positions corresponding to wave lengths of
1728 /x, 1079/1. The widths <>f the groups were 4.8mm., 2.8mm.,
1.8 mm. Thus, though these phantom lines have different apparent wave
lengths when measured on different gratings, their dispersion seemed
proportional to their wave length. It must be noted that with grating
No. 2 only the lines called A,. (',, and A4, C4 are visible, the other
fainter lines, as well as the llu ed effect in the background, being totally

absent. This is well accounted for by the feebleness of the spectrum

obtained from grating No. 2. The apparent wave length of these groups

obtained by the same method as that previously employed. Owing
to the great distance between the head group (2790) C and its reproduc-
tions, a cathetometer was used. The results are correct, however, to
better than - A in units. The widths of the groups were measured

on the dividing engine and are correct to within .03 mm. The value of
1 cm. of plate length was obtained from measurements between two sun
lint

I feel certain that groups 2790 /x, (', and (', a- obtained with grating
No. 2 are not in 1:1': .'! ratio.

The next grating Investigated was one of 21,000 line, to the inch and

_'l ft radius. Here the illumination was even more feeble than with

ting No. 2. Great trouble wi erienced from fog. The lines in

magnesium, C] and C4, COuld not be found. The last grating examined

tnilar to No. I. It had a radiuB of 183 cm., 14,438 lines to the
in<h; it was prepared by Brashear; i; was ruled by Rowland's engine at

LYMAN. — FALSE SPECTRA FROM THE ROWLAND GRATING. 251

Johns Hopkins in the year 1894. The mounting was the same as that
of grating No. 1.

The character of the photographs obtained with grating No. 4 was
quite different from those obtained with grating No. 1. Lines A1? C1?
and A4. C4 were present, and with about the same apparent wave length
as before. They were, however, very much less distinct than when ob-
tained with grating No. 1. The fluting effect was not very sharp ; but,
on the other hand, the space between positions 1878 /a and 924 /x was
filled with reproductions of groups Ax and Cl5 of such clearness as to be
cpiite unmistakable. The wave lengths of some of these phantom lines
was determined in the same manner as before. Groups corresponding in
appearance to the family marked Au A2, A3, etc., were found at 1970 /z,
1020 /x, 993 /x, and 97u^. The groups called Cl5 C2, etc., were found at
1878^ and 924 1/. very cleanly, and at more than eight other positions
in a less distinct form between 1878 and 924. In fact, the spectrum in
this region is so tilled with groups of the A, C, type that the lines of one
family overlap lines of the other in many cases.

Such, then, are the experimental data connected with this phenomenon.
Before coming to a statement of conclusions, however, the following facts
are of interest : —

Some of the reproductions belonging to the set of lines A2 and E2 in
aluminum occur under the real spectrum, and may be observed if proper
care is used.

Gratings of large radius do not show the false lines with as much clear-
ness as gratings of small radius. So feeble, indeed, are the false lines
obtained with the 21 ft. grating, that it is improbable that they would
be noticeable in ordinary spectrum work. With a six-foot grating,
however, the false lines are so strong that they might easily cause errors
in series spectrum investigations. We have seen, moreover, that the
false lines from grating No. 1 are much stronger than those from grating
No. 4, and we have reason to suspect that gratings exist which pro-
duce false lines even stronger in character than those obtained with
grating No. 1. Gratings 1 and 4 differ little in radius, and have exactly
the same number of lines to the inch, yet their false spectra are entirely
different in character. We cannot apparently predict the nature of the
false spectra of one grating from the false spectra produced by another
of the same ruling and radius.

The conclusions arrived at in this paper may be stated as follows : —

1. The first spectrum produced from the Rowland concave grating
is not pure, but is complicated not only by underlying spectra of

252 PROCEEDINGS OF THE AMERICAN ACADEMY.

higher order, but by the presence of a number of spectra of lower
dispersion.

2. The number and dispersion of these false spectra obey no simple
law, but present the appearance of a reproduction of some real group or
groups distributed at random.

3. The number, dispersion, and clearness of these spectra differ with
different gratings.

1. The lines due to the false spectra are most clearly visible in the
extreme ultra-violet, owing to the absence of strong real lines, and are
so pronounced as to screen any feeble real line that may occur in this
region.

Jefferson Physical Laboratory, Harvard University,
Cambridge, Mass., October, 1000.

Lyman. — False Spectra.

PLATE I.

A 02790 Displaced Spectrum

x^_ : ..■ir n_ in

A 275o Normal Spectrum
MAGNESIUM

PLATE II.

__ A, A2 b, C, C; D; Ej A3 A^b£4

Ql;! i 'iLj ~T I I

1873 924

MAGNESIUM

PLATE III.

A, C,

1 I

Displaced Spectrum

1

tz n !

HIS

1

A

t

MAGNESIUM

A, c,

Normal Spectrum

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 15. —January, 1901.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
CASE SCHOOL OF APPLIED SCIENCE.

INVESTIGATIONS ON THE COMPOSITION OF

PETROLEUM.

By Charles F. Mabery.

XXXV. On the Composition of California Petroleum. By Charles

F. Mabery and Edward J. Hudson.
XXXVI. On the Chlorine Derivatives of the Hydrocarbons in Cali-
fornia Petroleum. By Charles F. Mabery aud Otto
J. Sieplein.
XXXVII. On the Composition of Japanese Petroleum. By Charles
F. Mabery and Shinichi Takano.

Aid in the Work described in these Papers was given by the Academy from the
C. M. Warren Fund for Chemical Research.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
CASE SCHOOL OF APPLIED SCIENCE.

INVESTIGATIONS ON THE COMPOSITION OF

PETROLEUM.

By Charles F. Mabery.

Received November 5, 1900. Presented November 14, 1900.

No. 35. — ON THE COMPOSITION OF CALIFORNIA

PETROLEUM.

By Charles F. Mabery and Edward J. Hudson.

In its composition as a natural product, California petroleum is the
most interesting of any of the numerous petroleums that have come under
our observation. When its constituents are more fully understood they
will doubtless shed new light on the origin of petroleum, from the fact
that they have been subjected to less rigorous agencies during, or subse-
quent to, their formation than petroleums from other well-known fields.
In fields not yet developed, such as those of South America and Japan,
similar unstable oils are found. It is useless to attempt to separate from
these peculiar oils their individual constituents by the ordinary method
of fractional distillation under atmospheric pressure. Not only on account
of the high boiling point", but because the presence of air causes such
decomposition, hydrocarbons boiling above 200° cannot be separated in a
pure form without the aid of a vacuum. With no further knowledge of
the constituents of the crude oils, this explains the numerous ineffectual
attempts during the last thirty years to obtain acceptable refined prod-
ucts from California petroleum. By excluding air, as in vacuum distilla-
tion, with the reduction of boiling points, it is possible to distil without
decomposition all the hydrocarbons of the crude oils. With some modi-
fications of the apparatus employed in vacuum distillation of the higher
fractions of Pennsylvania oil, the fractional separation may be carried on
with little more delay than in distillation under atmospheric pressure.

PROCEEDINGS OF THE AMERICAN ACADEMY.

Provision for drawing in fresh fractions without loss of vacuum mav be
made by sealing on a second tube to the neck of the flask a little lower
than the exit tube for the escaping vapors. Then, as in the ordinary
application of the Hempel principle, the cooling space in the neck of the
flask may be increased by filling it with broken glass. This requires a
long-necked flask, and the beads may be supported on a piece of glass
rod resting on the bottom of the flask, the upper end flattened so as
nearly to till the neck of the flask as a support to the broken glass. The
chief hindrance to the use of beads is the condensation of liquids with high
boiling points in the neck of the flask, due to the cooling effect of the
outside air. The 111 ;ck of the flask must, therefore, be packed an inch or
more thick with asbestos. With this arrangement, and with a flat perfo-
rated ring burner for heating, the distillation may be conducted with a
fairly rapid fractional separation, without decomposition even of the
constituents that distil above 350 , 50 mm.

This examination of California oil was undertaken with the intention
of ascertaining the series of hydrocarbons which constitute the main body

the crude oil, as well as the principal members of the series, as shown
by percentage composition and molecular weights.

Tic first specimen of crude oil was received through the kindness of
Mr. C. A. Black, chemist in an oil company in Ventura County. The
ind Bpecimen came from a well of the Coalinga < HI Company in Fresno,
through the secretary of the company. Mr. Henderson llayward. These
two specimens will be referred to respectively as Ventura oil and Fresno
oil. Another specimen was received from the Puente Hills in Los An-
geles County, through the kindness of the Puente Oil Company at China
This oil will be referred to as Puente oil. Four other specimens were
received from the Torrey wells and the "Los Angeles" wells in the
Sespe district (Scott's Hill) in Ventura County, through the kind:
Of the Onion Oil Company at Santa Paula. These samples will be
designated respectively a> Adams' Canon. Bardsdale, Scott's Hill, and
Torrey oils.

Another specimen is of especial interest, since it came from a well
sunk beneath the Pacific Ocean at bigh tide. It was sent by Mr. .'. B.
Treadwell from a well at StTmmerland, Santa Barbara Count v. at a
point where the oil strata outcropped on a north and Boutfa line at t he
top of an elevation, and again d< scending on the south side, disappeared
under the ocean. The well was driven to a depth of -\~> feet, just
within the shore line at high tide, where it reached the bottom of the
oil -and. This oil will be referred to as Santa Barbara oil.

MABERY. — COMPOSITION OF PETROLEUM. 257

This collection of specimens seemed to offer a fair average of the oils
from the California fields, representing four different counties, and deposits
that yield oil in considerable quantities.

Since these crude oils differed essentially in their appearance and in
their physical properties, in ascertaining their composition it seemed
advisable to examine them individually.

In the spring of 1897 an examination of the Ventura oil was begun,*
and besides giving the composition of the crude oil, the fractional sepa-
rations were carried far enough in vacuo to identify the principal constit-
uents boiling below 175°. The several hydrocarbons homologous with
benzol, which have been found in petroleum from other sources, were
found to constitute a large part of the unpurified distillates ; but the
chief constituents were shown by their composition and specific gravity
to be methylenes, for the first time recognized as essential constituents of
American petroleum.

Petroleum from Fresno County.

A brief account of the composition of a specimen of crude oil from
Fresno County was given in the preliminary paper referred to above.
When this oil was subjected to distillation, it was evident that decompo-
sition could only be prevented in the fractions above 1.30° by collecting
them in vacuo. The portion collected below 150° atmospheric pressure
— 2090 grms. — was submitted to long continued fractional separation
within 1° until 50 grms. collected below 55°, and only 5 grms. between
55° and 68°. At 68°-70° the distillates amounted to 60 grms., which
was separated for the most part at 08°. The specific gravity of the
crude distillate at 68° was found to be 0.G913, and after thorough agita-
tion with fuming sulphuric acid, 0.6844. A combustion of the purified
oil gave the following percentages of carbon and hydrogen : —

0.1692 grm. of the oil gave 0.5234 grm. C02 and 0.2314 grm. H.,0.

Calculated for
C8H10. csn18.

Found.

c

85.70 83.72

84.38

H

14.30 16.28

15.30

The loss in this analysis is doubtless due to an escape of a very small
amount of gaseous decomposition products without complete combustion.
Combustion of such a volatile oil requires the closest attention in manip-

* Am. Chem. Journ., XIX. 790 (1897).

VOL. XXXVI. 17

258 PROCEEDINGS OF THE AMERICAN ACADEMY.

illation to avoid loss in transferring to the combustion tube, and as will
appear later, the highest temperature that the hardest tube will stand is
necessary lor complete combustion. These results seem to indicate that
tli«- nil was a mixture of hexane and bexamethylene. The slight change
in specific gravitj after tr< atment with tinning sulphuric acid and potassic
hydrate shows that the benzol bad been almost completely removed.
The fractions at 68 appear to be the last in California oil containing
members <>t the series C II 8. Since .small quantities of distillates
collected in the viciuitj of 60 . 38 .and in one specimen at 30°, it is
inc hable that the very light gasoline from California oil has the same
composition a- that from Pennsylvania oil.

Tin- specific gravity of hexane. 0.6630 at lb' . and of bexamethylene,
0.690, also indicate that the fraction t.b°-70o is a mixture of these two
hydrocarbons.

lit itami.tiiylkm:. C71Iu.

The distillate collected at 9G°-98° amounted to fifty grams. Before
purification it gave as it< specific gravity 0.7499, and after purification,
H.7 US. The specific gravity of heptane boiling at this point is 0.64 tO
r2" ..")). Beilstein and Kurbatoff gave 0.742 at 18° as the specific grav-
ity of beptanaphtene. A combustion gave the following percentages of
carbon and hydrogen : —

(Mil:; grm. of the purified oil gave 0.4142 grm. C02 and 0.1814

'_rnn. II..O.

Calculated for (\,H.j„.

Found.

C

85.7<)

8o.71

11

14.30

14.3.3

A determination of the vapor density of the fraction 9G°-98° by col-
lecting the vap..r over memirv gave a value corresponding to the molec-
ular weight of heptamethylene. The determination was made by the
method of Y. Meyer, in vacuo, essentially according to the modification
of Lunge and Neuberg:* —

l.i| grm. of the ,,il gave 22.9 C.C. under a tension of 373 mm. and
at 2 1 .

rc:iiu. Pound

3.40 8.40

• Ber. deutach. Gesell., 1891, 7-J'.'.

MABERY. — COMPOSITION OP PETROLEUM. 259

The high specific gravity of this hydrocarbon corresponding to that of
heptanaphtene, the composition given by analysis and the molecular
weight, leave no doubt that this body is a heptamethylene, and that
heptane is not contained in California (Fresno County) petroleum.

Benzol.

The distillates in the vicinity of 80°, especially 79°-81°, gave the
characteristic odor of benzol. 20.5 grms. of these fractions shaken be-
low 60° with a mixture of nitric and sulphuric acids, gave a heavy nitro-
product ; after washing with water and distilling, the distillate weighed
13 grms., showing that 35.5 per cent of this oil was benzol. Some of the
same fractions treated with fuming nitric acid deposited needle-shaped
crystals that melted at 89°-90°, the melting point of dinitrobenzol.

Toluol.

After the twelfth distillation, 60 grms. collected at 109°-110° with
the penetrating odor of toluol. 30 grms. of this distillate was shaken
with a mixture of nitric and sulphuric acids, and the oil not nitrated
separated from the nitro-product by distillation. The distillate weighed
13 grms., showing that 54 per cent of the original distillate was toluol.
Upon agitating this product with fuming sulphuric acid and distilling,
1.6 per cent more was removed. The remaining oil was dried over
sodium for analysis.

The unpurified distillate gave the following values on analysis : —

0.1584 grm. of the oil gave 0.5090 grm. C02 and 0.1663 grm. H20.

Toluol.

Calculated for

C«H2».

Found.

c

91.31

85.70

87.64

H

8.69

14.30

11.73

The change in composition by treatment with acids was very
marked : —

0.1725 grm. of the purified oil gave 0.5387 grm. C02 and 0.2282 grm.
H20.

Calculated for CkH2».

Found.

c

85.70

85.20

H

14.30

14.71

Evidently the small quantity of oil remaining after the treatment with
acids was octonaphtene, that had not been completely separated by
distillation.

2l>0 PROCEEDINGS OF THE AMERICAN ACADEMY.

( )( TtiN U'll 1 INI . C8H16.

After the twentieth distillation, 30 grms. collected at 118° -120°,
which, with no further purification than drying over sodium, gave 0.7615
a> its specific gravity at 20 , and, on analysis, values required for
octonaphtene : —

0.14.37 grin, of the oil gave 0.45GG grm. C02 and 0.1922 grra. 11,0.

C 85.47

II 14.71

After thorough treatment with fuming sulphuric acid the specific
gravity was reduced to <).75.'!2, at 20° (octonaphtene, 0.7582 at 17°. 5,
BeilsteiD and Kurbatoff), but the composition was scarcely changed : —

I. ".1 171 grm. of the oil gave 0.4618 grm. CO, and 0.1895 grm. H.,0.
II. 0.1334 grm. of the oil gave 0.4185 grm. C02 and 0.1762 grm. ILO.

Calculated for

ill

i.

Found.

II.

c

85.70

85.62

85.54

II

14.30

1 L32

14.G7

After the eighteenth distillation, 25 grms. remained persistently at
124 -125 . which was not affected by fuming sulphuric acid in the cold,

but at 100° the acid became colored and gave off much SO.,. On pour-
ing into water, the solution gave a peculiar odor resembling turpentine.
and a black powder separated that contained nitrogen. Analysis of the
distillate after the treatment with the acid gave values required for

1 1 : —

0 1391 grm. of the oil gave 0. 1354 grm. CO... and 0.1875 grm. 1LO.

Required for <\,n._.„. Pound.

C 85.70 85.36

II 14.30 11-98

Sine- no other hydrocarbon than normal octane has been recognized
with this boiling point, this distillate musl be a mixture of octonaphtene
with a higher body, but evidently normal octane is not present in appre-
ciable quantity.

Nonon \ i-i 1 1 1 m . ( ',11 .

A considerable quantity of distillate collected at 184°-185 , that gave

Be gravity 0.8175. After purification with fuming Bulphuric

MABERY. — COMPOSITION OP PETROLEUM. 261

acid its specific gravity was reduced 0.7591. 28 grms. of the crude
distillate gave 13 grms. of the purified oil with a loss of 53 per cent. The
purified oil gave by the action of fuming nitric acid a nitro-derivative
that, after crystallization from glacial acetic acid, melted at 85°. The
melting point of dinitromesitylene is 86°. A combustion of the puri-
fied distillate gave values for carbon and hydrogen required for
nononaphtene : —

0.1382 grm. of the oil gave 0.4318 grm. CO, and 0.1827 grm. H20.

Calculated for C9H18. Found.

C 85.70 85.21

H 14.30 14.68

Xylols.

Since para- and meta- xylols have frequently been found in petroleum,
it did not seem advisable to separate the isomeric xylols, although the
distillates 137° -140° gave the strong penetrating odor of these bodies.
At the end of the thirteenth distillation, 630 grms. collected within these
limits. On treating 125 grms. of this product with a mixture of nitric
and sulphuric acids, and agitating below 60°, washing with sodic
hydrate and water, drying and distilling, only 50 grms. came over
below 145°, showing that the xylols formed 60 per cent of the original
distillates and had been removed as nitro-products. At the low tem-
perature of the nitration only the aromatic hydrocarbons could have
been affected. The change in composition by the nitration is shown
by .the following analyses: —

0.1469 grm. of the unpurified oil gave 0.4770 grm. C02 and 0.1495
grm. H20.

Calculated for Xylol. Found.

C 92.31 88.53

H 7.69 11.31

Analysis of the purified distillate gave the following results on
analysis : —

I. 0.1382 grm. of the oil gave 0.4318 grm. CO., and 0.1827 grm.

H20.
II. 0.1386'grm. of the oil gave 0.4318 grm. C02 and 0.1837 grm.

H20.
III. 0.1441 grm. of the oil gave 0.4492 grm. C02 and 0.1901 grm. H20.

262 PROCEEDINGS OF THE AMERICAN ACADEMY.

I.

n.

nr.

i-.n.',..

c

85.21

84.96

85.00

85.70

II

1 1.69

14.72

14.66

14.30

The chief portion of the nitrated oil was shown to be meta xylol, since
the nitro-compouncl distilled at 240°-260°, and with fuming nitric acid
it was converted into a crystalline powder insoluble in alcohol ; crystal-
lized from glacial acetic acid, it melted at 176°, the melting point of
trinitrometaxylol. Probably the fifty grams not nitrated was for the
most part nononaphteue.

NONANE.

Since considerable quantities of distillates collected in the vicinity of
150 .alter the seventeenth distillation, attempts were made to ascertain
whether nonane is a constituent of California petroleum. This seemed of
especial importance since nonane forms such an important constituent of
Pennsylvania petroleum. Without purification a combustion showed the
following proportions of carbon and hydrogen: carbon, 86.11: hydro-
gen, 12.67. The specific gravity was ( > . s J 17. After agitating several
times with fuming sulphuric acid, the composition was materially
changed: —

0.1719 of the oil gave 0.5372 grm. CO, and 0.2206 grm. 11,0.

Calculated for CnH2n. Found.

C 85.70 85.20

H 14.30 14.26

The principal constituent of this fraction was therefore a naphtene or
a mixture of naphtenes.

Df.kaxai'iitene, C1(,1I,o-

At the end of the seventeenth distillation, the quantity of distillates
Within on.- degree limits between 159 ' and 1 «*.;» amounted to 150 grms.
By farther distillation these came together for the mosl part at 160°-
161°. The specific gravity of the un purified distillate at 20° was
0.8272. When agitated in the cold with fuming sulphuric acid, this

distillate developed great heat: alter treatment several times, until the

acid was no Longer much colored, the product was washed with Bodic

hydrate and water and dried for analysis. The original distillate lost

75 p<T cent of its weight by this treatment The aromatic hydrocarbon

MABERY. — COMPOSITION OP PETROLEUM. 263

doubtless consisted, for the most part at least, of ethyl toluol, boiling
point 161°-16"2°, but no further attempts were made to identify it.

The specific gravity of the residual oil was 0.7841, practically the
same as the specific gravity of dekanaphtene, 0.783, separated by
^Tarkownikoff and Oglobine from Baku petroleum. Its odor is pre-
cisely similar to that of the corresponding hydrocarbon which one of us
(Mabery) has separated from crude Baku oil.

Analysis of the unpurified distillate gave the following values : —

I. 0.1462 grm. of the oil gave 0.4724 grm. C02 and 0.1524 grm. H20.
II. 0.1496 grm. of the oil gave 0.4844 grm. C02 and 0.1564 grm. H20.

Found.

C 88.10 88.28

H 11.57 11.61

After treatment with the acid the composition was materially changed ;
the results correspond closely to the composition of dekanaphtene : —

0.1481 grm. of the oil gave 0 4646 grm. C02 and 0.1932 grm. H20.

Calculated for C10H20. Found.

C 85.70 85.55

II 14.30 14.50

A determination of the molecular weight gave a value required for
dekanaphtene.

0.8850 grm. of the oil and 24.22 grms. benzol gave, by the Beckman
method, at the freezing point, a depression of 1°.28.

Calculated for CJ0H20. Found.

140 140

After the twenty-fourth distillation, 50 grms. collected at 169°-170°,
which gave as its specific gravity without purification, 0.8358. After
thorough agitation with fuming sulphuric acid and washing, the specific
gravity was reduced to 0.7749, but the larger portion of the oil was
removed by the acid, indicating that the fraction consisted for the most
of a benzol homologue. The small quantity remaining after the treat-
ment was doubtless dekanaphtene, since its specific gravity was nearly
the same.

The presence of a naphtene in Russian oil, boiling at 182°, led us
to believe that the same body should be found in California oil. But

264 PROCEEDINGS OF THE AMERICAN ACADEMY.

the small quantity of distillate collecting here, only 10 grms. after long
distillation, excluded this hydrocarbon in any appreciable quantity.

Dndekanaphtene, CnH22, 195°.

Between iss an,l ][)\ , .">() grms. came together at the end of the
tenth distillation, which gave the following results on analysis: —

0.1423 grin, of the oil gave 0. t550 grm. CO., and 0.1617 grm. 11,0.

Calculated for CuHn. Found.

C 85.70 87.20

II 14.30 12.63

The specific gravity of this distillate before treatment was 0.8386.
After vigorous agitation with fuming sulphuric acid, and washing, the
specific gravity was reduced to 0.8044.

A distillate separated from Russian oil at 197° gave (Markownikoff
and Oglobine) as its specific gravity at 11, 0.8055, and it was assigned
the formula C12H24. The quantity of oil remaining after separation of
durol, which no doubt was the benzol hydrocarbon in this distillate, was
small. That it was a naphtene, as shown by its specific gravity, is con-
firmed by analysis : —

0.1459 grm. of the oil gave 0.4575 grm. CO, and 0.1886 grm. ILO.

Calculated for CltHM. Found.

C 85.70 85.70

II 14.30 14.40

Determinations of the molecular weight of this fraction by the Beck-
man freezing point method using benzol gave the following value: —

•01 grm. of the oil and 24.68 grms. benzol gave a depression of
1.1249 grm. of the oil and 25.90 grms. benzol gave a depression of 1°.36.

Calctilati-i] f«,r md.

C„B I. II.

154 155 157

It appears from these results that the hydrocarhon in California petro-
leum boiling at 195 tsnol dodekanaphtene, CijH24, but undekanaphtene,
C .II . This conclusion is further confirmed, as will appear later, by the

composition of the distillate 21 6% which appcai'3 to be dodekanaphtene,

MABERY. — COMPOSITION OP PETROLEUM. 265

and still further by the composition of the monochlornaphtene obtained
from the hydrocarbon boiling at 195°, which gives as its formula
CnH21Cl.

In the vicinity of 200°, 50 grms. of a distillate collected which gave
as its specific gravity 0.8684. After agitation with fuming sulphuric
acid, the specific gravity was diminished to 0.8202. An analysis before
treatment gave 87.84 per cent of carbon and 11.91 per cent of hydrogen.
After treatment with the acid, the analysis gave 85.99 per cent of carbon
and 13.97 per cent of hydrogen, showing that some of the aromatic
hydrocarbon still remained ; the small amount of naphtene was doubt-
less undekanaphtene, boiling at 195°; specific gravity, 0.8010, as given
above.

DODEKANAPHTENE, Ci2H24.

A considerable quantity of distillate collected at 208°-210°, correspond-
ing to 216°, which corresponded in composition by analysis and molec-
ular weight to dodekanaphtene : —

0.1496 grm. of the oil gave 0.4677 grm. C02 and 0.1962 grm. H20.

Found.

85.26
14.57

Determinations of the molecular weight of this product by the Beck-
man method at the freezing point, gave results corresponding to the for-
mula of dodekanaphtene.

1.0141 grm. of the oil and 25.61 grms. benzol gave a depression of 1.17.

Calculated for C,,nM. Found.

166 166

It is therefore evident that California petroleum differs from Russian
oil in containing dodekanaphtene, boiling at 216°, as well as undecanaph-
tene, boiling at 195°.

Pcente Oil.

The specimen received from the Puente Oil Company was somewhat
thicker than the Fresno oil. Its specific gravity at 20° was found to
be 0.892. It contained 0.80 per cent of sulphur. Two determinations
of nitrogen by the Kjeldahl method gave, (I.) 0.564, (II.) 0.587 ; and
by the absolute method, measuring the volume of nitrogen, (I.) 1.18, and

Calculated for C„Il2„.

c

85.70

H

14.30

206 PROCEEDINGS OF THE AMERICAN ACADEMY.

EL) 1 .22. A combustion of the crude oil pave, 84.06 per cent of carbon
and 11.96 percent of hydrogen. Two determinations of bromine absorp-
tion pave, fl.) 18.8, (II.) 18.3 percent.

A distillation of 1346 grms. of the crude oil under atmospheric pres-
sure gave the following weights : —

-100° 150 JOO0 200°-25<P

167 17" 132 130 Grms.

The small proportions distilling below 250° made it evident that the

chief constituents could only be separated in vacuo. The specific

gravity of these distillates at 20° were ascertained by weighing on a
\V, stphal balance : —

1CKP 100 -150 150 2

0.7642 0.81-Vj 0.8538

Ten litres of the Puente crude oil, after collecting below 150° at
atmospheric pressure, gave the following weights under 50 mm : —

150 At. it. -120' 120°-150D 150"-17:. 176 L8f 200D

590 800 827 990 160 195 dins.

The residue above 200° was next divided into fractions of convenient
>i/e without noting temperatures: —

1. 2. 3.

4.

655 1010 G70

430 Grms.

Further distillation of the fraction -150 At. Pressure gave

the follow-

ing weights: —

78°-452° 88 ^-nS' 1"'.> 110°

11^ 120

10 1<> Ki 27 40 120 60

65 Grms.

Smaller weights collected at intermediate temperatures.

II II' I \ Ml. I II VI. KM . ( I IIt.

No attempts were made t'> ascertain the composition of the individual
fractions below 96 from Puente oil. Those portions will receive
further attention with the corresponding distillates from the other oils.
Without purification tin- distillate 96 - g ve a- it- specific gravity
at 20 . 0.7499. After repeated agitation with finning sulphnric acid
and potassic hydrate, the specific gravity was scarcely changed, 0.74 I":

MABERY. — COMPOSITION OF PETROLEUM. 267

specific gravity of heptauaphtene, 0.742 at 18°. Analysis I. was made
of the untreated oil, analysis II. of the portion after treatment: —

I. 0.1409 grm. of the oil gave 0.4438 grm. C02 and 0.1683 grm. H20.
II. 0.1396 grm. of the oil gave 0.4383 grm. C02 and 0.1792 grm. H20.

Calculated for Found.

C7H14. I. II.

C 85.70 85.90 85.60

H 14.30 13.27 14.26

The difference in analysis I. is doubtless due to oxygen and nitrogen
compounds, since agitation of the crude distillates with potassic hydrate
separates from most of the distillates from California petroleum heavy
oils, with an odor resembling that of creosote. A description of these
oxygenated compounds is reserved for a later paper. The specific
gravity of the purified fraction, 0.7440, is essentially the same as that
of heptauaphtene separated from Russian oil, 0.748, by Beilstein and
Kurbatoff.

OCTONAPHTENE, C8H16.

The fraction 11 8° -120°, without further treatment, gave as its specific
gravity at 20°, 0.7615, which was reduced to 0.7540 by agitation with
fuming sulphuric acid. Specific gravity of octonaphtene, 0.7552 at
17°.5. Analysis I. was made of the crude distillate, and analysis II. of
the oil after purification with the acid.

I. 0.1406 grm. of the oil gave 0.4424 C02 and 0.1760 grm. H20.
II. 0.1503 grm. of the oil gave 0.4692 grm. C02 and 0.1941 grm. H20.

Calculated for Found.

C8Hl6. I. II.

C 85.70 85.81 85.20

H 14.30 13.91 14.35

No further examination was made of the fractions from 120° to
148°, but it seemed of interest to ascertain whether nonane formed any
part of the distillate 148°-150°, since it forms such a large proportion
of the Eastern oils. The crude distillate 150°-151° gave as its specific
gravity 0.7910, and after purification with sulphuric acid, 0.7730.
Analysis of the treated oil gave percentages of carbon and hydrogen
that still showed the presence of the benzol homologue.

2G8 PROCEEDINGS OF THE AMERICAN ACADEMY.

0.1449 grm. of the oil gave 0.4.380 gnu. C02 and 0.1807 grm. ILO.

Calculated for ''„Hj,- Found.

C 85.7" 86.37

II 14. .in 13.57

From the results of this analysis it is evident that nonaue forms no
appreciable proportion of Puente oil.

Dekanaphtene, Cj„irM.

50 grms. of a distillate collected at 160°— 162°, with smaller quantities
on either Bide. A determination of its specific gravity at 20° gave
0.7966, and after agitation with sulphuric acid, 0.7745. The composi-
tion of this distillate after treatment with the acid was shown by
analysis: —

ii. 1 7 17 grm. of the oil gave 0.5515 grm. CO. and 0.2185 grm. IL.O.

Calculated for C^Ug,. Found.

C 85.7' » 8G.09

II 14.30 13.90

Fraction 1G8°-172°.

In all the California oils examined, as stated above, considerable
quantities have collected at 1G8°-172°. Since large proportions of
the benzol hydrocarbons appear in all these oils, the distillate 1G8°-
172° might consist to a large extent of pseudoeumol, boiling point 1G8°.
The high specific gravity of the fraction before treatment with the acid,
0.8063, and the lower value after treatment, 0.7874, indicates a con-
siderable proportion of the benzol homologue. Even after the acid
treatment, the percentages of carbon and hydrogen showed the presence
still of the same constituent: carbon, s5.:>I ; Indrogen, 13.G1. It
Beems probable, therefore, that the benzol homologue constitutes the
greater pari of this distillate, and the residue after treatment is doubtless
higher or lower constituents. Hut the absence of a decane corresponding
to the decane in Pennsylvania petroleum, boiling at 172°, is assured.

Fraction 180°-182°.

Only 10 grms. collected al 180 -182°. with Bmaller quantities on either
Although this quantity of the nil was insufficient for complete

examination, its specific gravity before and after analysis was ascertained,

and an analysis made after treatment with acid. Before treatment the

MABERY. — COMPOSITION OF PETROLEUM. 269

value obtained was 0.8116, and afterward 0.7955. A combustion gave
8G.27 per cent of carbon, and 13.79 per cent of hydrogen. Evidently,
like the Fresno oil, Pueute oil has no constituent with this boiling point,
except perhaps a benzol homologue.

Undekanaphtene, CnH22, 195°.

A naphtene should be expected between 190° and 200°, since undeka-
naphtene, CUH22, is a constituent of the Fresno oil. The 20 grams
collecting at 190°-192° gave as its specific gravity at 20°, 0.8196, and
after treatment with fuming acid, 0.8046 (0.8055 at 14°, Markownikoff).
A combustion of the oil after treatment gave 85.90 per cent of carbon
and 13.85 per cent of hj'drogen.

The series CJion+2 forms no part of the Puente oil, at least above 95°.
The low percentage of hydrogen seems to indicate a series still poorer in
hydrogen than the naphtenes, especially in the higher distillates, possi-
bly a series C„H2„_2, containing two naphtene rings. Whether such hy-
drocarbons are really present, or whether benzol homologues have not
been completely removed, will appear in more prolonged distillation of
these oils boiling above 200°.

»

Bardsdai.e Oil.

The specimen marked Owens No. 2, sent by the Union Oil Co. from
the Bardsdale district, was quite thick and heavy. Its specific gravity at
20° was 0.8923. It contained 1.25 per cent of nitrogen, as shown by the
volume of nitrogen collected. A Kjeldahl determination gave 0.50 per
cent. A Carius determination gave 1.5 per cent of sulphur, showing it
to be a high sulphur oil. A combustion gave 84.17 per cent of carbon
and 12.15 per cent of hydrogen. 108 grms. of the oil distilled under
atmospheric pressure gave the following weights: —

-150° 150°-310° 3KP-4100

9 24 7 Grms.

0.7433 0.8170 Specific gravity.

In distilling 5088 grms. under 50 mm., except -150° the following
weights were obtained : —

-150° At. Pr. 150°-200° 200°-25(P 250°-290°

250 555 374 1165 Grms.

The two fractions below 200° were distilled within 5°, and then
within 2°, which brought together the following weights with smaller

J o © O ©

quantities at neighboring temperatures : —

270 PROCEEDINGS OF THE AMERICAN ACADEMY.

96M00' 116-120"' 130-140 L48°-160° 15-< -102^ 1GS-1723 190M94D

34 25 85 W CO 20 40 Grms.

The lower portions of this oil began to come over at 45°, but only
small quantities collected below 6.0°. At 6oc-70°, 40 grms. collected,
which will be examined in connection wiih the similar distillates from
the other oils to ascertain whether the principal constituent is hezane
or hexamethylene. The high specific gravity of this oil, even after
treating thoroughly with fuming sulphuric acid, 0.7017, would seem to
exclude hexane; specific gravity, 0.6671.

Kept AM ETHYLENE, CTII14.

The unpurified fraction 96°-100° gave as its specific gravity at 20°,
0.7395, and after treatment with fuming acid, 0.7384. The slight change
in Bpecific gravity indicates that this fraction consists for the most part
of one hydrocarbon, and analysis with the boiling point shows that the
hydrocarbon is heptamethyleue ; specific gravity, 0.7420 : —

0.1 173 grm. of the oil gave 0.4G20 grm. CO., and 0.1879 grin. 11,0.

Calculated for C;II,4. Found.

C 85.70 85.54

II 14.30 14.18

OCTONAPHTENE, C8H16.

The specific gravity of the fraction 116°— 120° was 0.7('2">. and after

treatment with fuming sulphuric acid, 0.7."iGfi, which indicated that this

product contained very little of the benzol homologies. An analysis

r purification gave proportions of carbon and hydrogen required for

octonaphtene : —

0.1780 grm. of the oil gave 0.5572 grm. CO.. and 0.2287 grm. II..O.

I ilculated for C8III0. Pound.

C 85.70 85.39

II 11.30 14.28

Fraction 150 -152 .

Tie- distillates above 120° received no further attention, except to

rtain that the xylols were present in large proportion, as in the other

oils. To ascertain whether uonane is contained in this oil. the Bpecific

ity of the. fraction 150 152 was ascertained and found to be

MABERY. COMPOSITION OP PETROLEUM. 271

0.7756, after treatment with the fuming acid. Its composition after
treatment with the acid corresponded to that of C„H2;t : —

0.1434 o-rm. of the oil have 0.4480 grm. C02 and 0.1839 grm. H20.

Calculated for CftH2„.

Found.

c

85.70

85.20

H

14.30

14.25

The lower specific gravity of nonane and the results of analysis show
that nonane is not a constituent of this oil.

Dekanaphtene, C10H20.

The specific gravity of the unpurified fraction 1G0°-162° at 20° was
0.7966, and after purification with fuming sulphuric acid and sodic
hydrate was 0.7905 (C10H20, 0.7950 at 0° Markownikoff) .

It gave the following results on analysis : —

I. 0.1471 grm. of the oil gave 0.4625 grm. C02 and 0.1907 grm. H20.
II. 0.1475 grm. of the oil gave 0.4477 grm. C02 and 0.1813 grm. H20.

Calculated for

C]0II.,0.

i.

Found.

II.

c

85.70

85.50

85.86

H

14.30

14.37

14.17

On account of the small amount of the distillate 167°-170°, no

examination was made of it. It was doubtless chiefly a benzol

homologue.

Fraction 176°-178.

To ascertain whether a naphtene boiling at 182° is present in this oil,
the fraction 17G0-178° was treated thoroughly with fuming sulphuric
acid and sodic hydrate. Its specific gravity before treatment was 0.8116 ;
after treatment, 0.7955. It gave on analysis 86.01 per cent carbon and
13.31 per cent hydrogen, indicating still the presence of a benzol homo-
logue. But the small quantity remaining showed the absence of any
other constituents with this boiling point in this distillate.

Undekanaphtene, CnH22.

The fraction 190°-194° gave 0.8186 as its specific gravity before
treatment with fuming sulphuric acid and sodic hydrate, and afterward,
0.8046. Markownikoff found 0.8119 at 0° as the specific gravity of
dodekanaphtene from Baku oil.

•2~-2 PROCEEDINGS OF THE AMERICAN ACADEMY.

The percentages of carbon and hydrogen given by analysis supported
the formula of undekanapliteue : —

0.1466 gnu. of the oil gave 0.4588 grm. C02 and 0.1909 grm. II20.

Calculated for l',,!!*.. Found.

C 85.70 85.37

II 14.30 14.47

Some of the Bardsdale distillates, 140°-220°, 50 mm., turned pink
soon after they were collected, changing to a darker red on standing, and
after some time they deposited a dark oil. The colored distillates alter-
nated with the others, leaving sharply lined uncolored oils beside those
depositing the insoluble heavy oils. The following distillates were
colored : —

154 -156 . 1.">3°-1G0°, 168°-170°, 180°-182°, l98°-206°, 212°-
216 •

The distillates between these were nearly or quite colorless. This
coloration \v;n caused by the phenol bodies, which are easily extracted
by alkaline hydrates. The bodies will receive further attention in

another paper.

Adams' Canon Oil.

The Adams' Canon Oil from the Ex-Mission district was thicker and
heavier than the oils previously described. The specific gravity of this
specimen at 30° was 0.9212. It contained a large percentage of nitro-
gen. 1.46, determined by the volume of nitrogen. The Kjeldahl method
gave 0.58 per cent. Two determinations of sulphur gave, (I.) 0.92,
(II.) <>.s7. This oil contained a very small proportion of the lower
hydrocarbons. 100 grms. distilled under atmospheric pressure gave
- jnns. below 150°, and 27 grms., 150°-300°. The specific gravity
of the lower distillate was o.7<'>7.">. and of the higher, 0.8457. In
vacuum distillation, 9382 grms. of the crude oil gave the following

weights

—

V. IV.

-120

120 160

160

1 23

L045

525

v arly three-fourths of the original oil remained above 250 .
Continuing the distillation of the lower distillates within 5° and 2°,
the following weights were collected: —

116 120 168 168 170 180 182

87 10 12 20 10 Grms.

MABERY. — COMPOSITION OF PETROLEUM. 273

In continuing the distillation of the lower fractions nothing came over
below 60°, and only in quantities of a few grams below 95°.

Heptamethylene, C7Hu.

The fraction 98°-100° unpurified gave as its specific gravity at 20°,
0.7444, and after treatment with fuming sulphuric acid and sodic hy-
drate, 0.7414. A combustion of the purified oil gave the following per-
centages of carbon and hydrogen : —

0.1525 grm. of the oil gave 0.4797 grm. C02 and 0.1976 grm. H20.

Calculated for C7H14. Found.

C 85.70 85.53

H 14.30 14.40

As in the other oils, much benzol was contained in the fractions near
80°, and toluol in those near 110°.

j

OCTONAPHTENE, C8H16.

The fraction 118°-120° gave as its specific gravity unpurified at 20°,
0.7632, and after treatment with fuming acid and sodic hydrate, 0.7600.
Specific gravity of octonaphtene, 0.7582. This distillate gave after puri-
fication, 86.62 per cent of carbon, and 14.60 per cent of hydrogen,
which with its boiling point showed it to be octonaphtene.

To ascertain whether nonane is a constituent of Adams' Canon oil, the
distillate 148°-150° was determined, and found to be 0.7858. After
treatment with the fuming acid and sodic hydrate its specific gravity was
but slightly changed, 0.7800. A combustion then gave 86.10 per cent
of carbon, and 13.91 per cent of hydrogen.

Since nonane requires 84.37 per cent of carbon and 15.63 per cent
of hydrogen, it is evident that this hydrocarbon is not a constituent of
Adams' Cano'n oil.

Dekanaphtene, C10H20.

The unpurified distillate 158°-160° gave as its specific gravity at 20°,
0.7972, and 0.7904 after treatment with the fuming acid and sodic hy-
drate. A combustion of the purified oil gave the following percentages
of carbon and hydrogen : —

0.1455 grm. of the oil gave 0.4583 grm. C02 and 0.1804 grm. H20.

Calculated for C]0H20. Found.

C 85.70 85.91

H 14.30 13.78

VOL. XXXVI. 18

274 PROCEEDINGS OF THE AMERICAN ACADEMY.

i

The low percentage of hydrogen and high percentage of carbon in-
dicates that the benzol homologue was not completely removed. In
some of these distillates, the benzol hydrocarbon is present in such large
proportion that long continued action of the fuming acid is necessary to
remove it completely.

The small amount of the distillate 180°-182° gave as its specific
gravity at •_'') , 0.8154, and after treatment with the fuming acid, 0.8097.
A combustioD of the treated oil gave 85.56 per cent of carbon and 13.95
per cent of hydrogen. No further examination was made of the frac-
tions collected uuder atmospheric pressure. Like the Bardsdale oil,
part of the distillates between 140° and 220° under 50 mm. turned
pink soon after they were collected, and on standing deposited a dark
colored oil.

While tin- possibility of hydrocarbons poorer in hydrogen, such as con-
densed naphtene, is suggested by the low percentage of hydrogen, on
account of the very large proportion of aromatic derivatives of benzol
contained in Adams' Canon oil, it seems more probable that these results
are due to benzol derivatives not wholly removed, especially as the boil-
ing points of the double ring naphtenes undoubtedly are much higher.
Such a double ring naphtene apparently formed as one of the products
by the action of sodium on monochlorheptamethylene boiled in the
vicinity of 240°.

Torrey Wklls Oil.

The specimen of petroleum from the Torrey wells was much lighter
than that from Adams' Canon ; its specific gravity at 20° was 0.8837.
It contained 0.49 per cent of sulphur, and gave 0.88 per cent of nitrogen
by the Kjel.lahl method and 1.15 per cent 1>\ volume. A combustion
• 8G.00 per cent of carbon, and 12.48 per cent of hydrogen; 100
grm8. distilled under atmospheric pressure gave 11 grms. with a specific
gravity 0.7519 below 150°, and 29 grms. specific gravity 0.8226 between
150° aid 250. In distilling Torrey oil in vacuo, 98G0 grms. gave the
following weights : —

100 \' I'r.

-126

126 175°

176

•^: 2:;,

790

1175

900

700 Grms

Further distillation of the lower fractions gave, after the fourth dis-
tillation, the following weights, with smaller proportion-, at temperatures

between : —

MABERY.

— COMPOSITION OP PETROLEUM.

27S

96D-98°

116°-120°

1363-14(P 158°-162° 168°-171°

178 ^-1823

78

40

45 37 43

35 Grms,

The Torrey oil contained more of the lower distillates than any other
of the specimens examined. Several grams collected below 40°, and at
65°-70°, 30 grams came together. The un purified distillate 65°-70°
gave as its specific gravity at 50°, 0.6981, which was reduced only to
0.6926 after agitation with fuming sulphuric acid. The specific gravity
of hexane boiling at 68° is 0.6630.

Analysis of the purified oil gave the following results : —

I. 0.1551 grm. of the oil gave 0.4768 grm. C02 and 0.2138 grra. H20.

II. 0.1995 grm. of the oil gave 0.6146 grm. C02 and 0.2769 grm. H2G.

III. 0.1369 grm. of the oil gave 0.4248 grm. C02 and 0.1888 grm. H20.

c

Calculated for

85.70 83.72

i.

83.85

Found.
II.

83.93

in.
84.63

H

14.30 16.28

15.32

15.41

15.33

In analysis III. every precaution was taken to avoid loss of the oil
after weighing, and the temperature of the combustion was maintained as
high as the tube would stand. Probably the coincidence in the. per-
centages of hydrogen is due to retention of unsaturated hydrocarbons
in the sulphuric acid.

Apparently, as in the Fresno oil, the hydrocarbon at 68° is a mixture
of hexane and hexamethylene.

Heptamethylene, C7Hu.

The fraction 9 6° -98° had the specific gravity at 20°, 0.7496, which
was scarcely changed by treatment with fuming sulphuric acid, 0.7430.
Analysis I. was made of the crude distillate, and analysis II. of the
purified product : —

I. 0.1222 grm. of the oil gave 0.3810 grm. C02 and 0.1602 grm. H20.
II. 0.1530 grm. of the oil gave 0.4801 grm. C02 and 0.1974 grm. H,0.

Calculated for

Found.

C7H]4.

I.

II.

c

85.70

84.97

85.60

H

14.30

14.56

14.34

This product is, therefore, fairly pure heptamethylene.
The fractions in the vicinity of 110° were shown to be composed for
the most part of toluol by the formation of nitro-products.

270 PROCEEDINGS OF THE AMERICAN ACADEMY.

Octonai'iitlm:, C8H16.

The specific gravity of the crude distillate 118°-120° was 0.7598,
and after purification with faming Bulphuric acid and sodic hydrate,
0. 7.330. A combustion of the anpurified oil gave 86.21 per cent of
carbon and 13. 3d per cent of hydrogen.

A combustion after purification indicated octonaphtene, although the
percentage of hydrogen is somewhat too high, probably on account of
accidental moisture : —

0.1470 grm. of the oil gave 0.4598 grin. C02 and 0.197G grm. 11,0.

Calculated for C„Hie. Found.

C 85.70 8o.34

II 11.30 11.93

The distillates 135°-140° consisted largely of the xylols.

D i; K A N A 1 • 1 1 1 1 ■: N E, C juHa,.

The specific gravity of the crude fraction 158°-1G0° was 0.7742, and
after purification with the fuming acid and sodic hydrate, 0.7742. Car-
bon and hydrogen were determined in the purified oil : —

0.1455 grm. of the oil gave 0.4583 grm. C02 and 0.1804 grm. 11,0.

Calculated for C10 ll20. Found.

C 85.70 85.91

H 14.30 13.78

The specific gravity of the crude distillate 1G8°-170° was 0.7928,
and 0.7*40 after treatment with Coming acid and sodic hydrate. A
combustion of the purified oil gave 85.98 per cent of carbou and 13. G9
per cenl of hydrogen.

The fraction 17s -1*2 had the specific gravity before treatment,
106, and afterward, 0.7924. A combustion of the oil after treatment
gave s"'.'.i7 per cent of carbon and 13.51 per cent of hydrogen. The
fraction 198 200 had the specific gravity of 0.8183, and after treat-
ment, 0.8069. The following percentages of carbon were given by com-
bustion: carbon, 86.51 ; hydrogen, 13.54.

Tin' low proportions of hydrogen and high proportions of carbon
alluded to before are especially apparent in the Torrey oil. Whether
this be due to a higher Beries or to benzol homologues not easily

MABERY. — COMPOSITION OP PETROLEUM. 277

removed will appear later in -the composition of the higher vacuum
distillates.

The Torrey distillates l40°-220°, 50 mm., showed a larger proportion
of oxygen and nitrogen compounds than any other of the crude oils.

Scott's Hill (Sespe District) Oil.

The specimen of petroleum from Scott's Hill was lighter than the
Torrey oil ; its specific gravity at 20° was 0.8782. It contained 1.25
per cent of nitrogen. Two determinations of sulphur gave (I.) 0.38 and
(II.) 0.49 per cent. 8260 grms. of the crude oil gave the following
weights on distillation in vacuo : —

-150° At. Pr. -160° 160D-208D 208°-275°

1220 495 930 880

The distillate -150° was collected under atmospheric pressure, the
others under 70 mm. The portion -150° was subjected to further dis-
tillation collecting at first within 5°, and twice within 2°. The fractions
collected in larger quantities within the following limits : —

f

66D-70° 86°-90° 96VL(XP 1163-12(P 133^-14(P im°Am° 170°-174° 178°-182° 190°-192°

26 38 35 37 38 62 41 38 41

Continuing the distillation of the higher fractions in vacuo, after the
third distillation within limits of 2°, the larger quantities collected
within the following limits with smaller quantities between : —

144°-148° 164D-1C8° 182:'-1860 210°-214c'

60 55 65 23

In distilling corresponding fractions from other fields in vacuo a
tendency to collect within the same limits was observed. After agita-
tion with fuming sulphuric acid, the fraction 66°-70° atmospheric pres-
sure gave as its specific gravity at 20°, 0.6984; the specific value much
higher than the specific gravity of hexane, 0.6630(17°5). A combus-
tion gave percentages of carbon and hydrogen correspondiug to a mix
ture of hexamethylene and hexane : —

C6H12.

Calculated for

C6H14.

Found.

c

85.70

83.72

84.67

H.

14.30

16.28

15.15

278 PROCEEDINGS OF THE AMERICAN ACADEMY.

The distillates 65 -70 J from Torrey, Scott's Hill, and Fresno oils were
put together and carefully distilled a number of times in order to separate
so far as possible the hydrocarbon boiling at 68° or 69° from other
admixtures, and the fraction 68°— 69° was thoroughly treated with rain-
ing sulphuric arid, warming gently and allowing it to stand with the acid
over night. Before treatment the Bpecific gravity at 20° was 0.7005,
and alter treatment, 0.6929. The following results were obtained by
analj Bis : —

I. 0.2 122 grin, of the oil gave 0.7596 grm. CO., and 0.3234 grm. H20.

i il'-ulated for
C«H„. CHM.

Found.
I.

c

85.70 83.72

85.50

H

14.30 16.28

14.84

In this analysis the combustion tube was filled with oxygen before the
oil was volatilized, and the temperature was kept as high as the tube
would stand. There seems therefore to be little doubt that the hydro-
carbon in California petroleum boiling at 68°-69° is composed chiefly
of hexamethylene.

Since small quantities of distillates remained in the vicinity of 90°-91°,
it suggested the possibility that isoheptane might form a part of this
product But its high specific gravity, 0.7303 at 20°, isoheptane 0.6819
(17°. 5), and the composition showed by analysis, excluded isoheptane in
any considerable quantity.

0.1273 grm. of the oil gave 0.3987 grm. CO., and 0.1665 grm. ILO.

Calculated for
C,HW. en,,.,

Found.

c

85.7d 84.00

85.1"

11

14.30 16.00

14.53

Especial precautions were taken in this analysis to have the tem-
perature of the Combustion as hot as possible, and the tube was tilled
with oxygen before the oil volatilized. The proportions of carbon and
hydrogen indicating the absence of isoheptane could not have been due
to the presence of benzol, since the oil was treated several times with
fuming Bulphuric acid. The volatile portions of California petroleum,
therefore, contain at raosl very small proportions of the hydrocarbons,
(II .,. and these if present consist almosl exclusively of members

In-low normal lmxane. Further continual ion of these formulae is given

by the chlorine derivatives, which will be described in another paper.

MABERY. — COMPOSITION OP PETROLEUM. 279

Heptamethylene, C7Hu.

Tlie fraction 96°-98° gave as its specific gravity at 20°, 0.7479, and
after purification with fuming sulphuric acid, 0.7436. A combustion of
the purified oil gave the following results : —

0.1530 grm. of the oil gave 0.4801 grm. C02 and 0.1974 grm. H20.

Calculated for C7HU.

Found.

c

85.70

85.60

H

14.30

14.34

These values leave no doubt that this constituent was heptamethylene.

OCTONAPHTENE, C8H16.

A determination of the specific gravity of the fraction 11 8° -120° at
20° gave 0.7628, and after agitation with fuming acid and sodic
hydrate, 0.7569. Determinations of carbon and hydrogen were made
both before and after purification.

I. 0.1644 grm. of the unpurified oil gave 0.5196 grm. C02 and
0.1975 H20.
II. 0.1470 grm. of the purified oil gave 0.4600 grm. C02 and 0.1976
grm. H20.

Calculated for

Found.

^8"16-

i.

II.

c

85.70

86.21

85.34

H

14.30

13.35

14.93

Although all these oils were dried over sodium, the hydrogen in
analysis II. is somewhat too high, probably on account of a trace of
moisture.

No further examination of the higher fractions of this oil below 160°
was made, except of the portions collected at 135°— 140°, to show that
large proportions of the xylols were contained as in the other specimens
examined.

Dekanaphtene, C10H20.

The specific gravity of the fraction 158°-160° was found to be
0.7848, and after agitation with the fuming acid, 0.7751. The propor-
tions of carbon and hydrogen were determined by analysis : —

280 PROCEEDINGS OF THE AMERICAN ACADEMY.

I. 0.1444 grin, of the oil gave 0. 1575 grin. CO., and 0.1885 grm. HnO.
II. 0.1451 grin, of the oil gave 0.4533 grin. COa and 0.1 $48 grm. HoO.

ulat*d for Found.

' H . I. II.

C 85.70 85.65 85.24

II 14.30 14. 22 14.33

I " N I > I.K ANAPHTENE, CnIL._>.

The specific gravity of this fraction without purification was 0.8093 at
20°, and after treatment with the fuming acid and sodic hydrate, 0.7952.
A combustion gave the following percentages of carbon and hydrogen: —

0.1458 grm. of the oil gave 0.4593 grm. C02 and 0.1817 grm. U.,0.

Calculated for C'uIIj,. Found.

C 85.70 85.90

II 14.30 13.85

Assuming, which is probable, that the benzol homologue was com-
pletely removed, these proportions of carbon and hydrogen point to the
presence of a hydrocarbon of a series containing less hydrogen.

A similar result was obtained with the fraction 194°-196°, specific
gravity crude, 0.8145, and after treatment, 0.8022. Combustion of the
purified oil gave the following percentages: —

i. ii.

C 85..".i' 85.85

II 13.97 13.92

"While the percentages of carbon in these analyses are satisfactory for
the formnla C„II..„. the percentages of hydrogen are less satisfactory.
The deficiency of abont one half of one percent in the sum of the carbon
and hydrogen his appeared in many of the analyses of products that
could be reasonably accepted as to their formula. With the greatest
care in the details of analysis, we have also found difficulty in obtaining
the total carbonic dioxide evolved by combustion of the oil in the ordi-
nary method of analysis. This subject has received attention in another
paper on this method of analysis.

Tridi B w \ ii i i i vr. c, I I .

After the separation of distillates iii vacuo from the ernde oil-, the
fractions may be distilled at temperatures which would cause decomposi-

MABERY. — COMPOSITION OF PETROLEUM. 281

tion of constituents of the crude oil. It is therefore possible to continue
the distillation of the hydrocarbons even as high as tridekanaphtene
under atmospheric pressure after the first distillation. In continuing the
distillation, 35 grams collected at 230°-232°, specific gravity 0.8511.
After purification with fuming sulphuric acid it gave as its specific grav-
ity 0.8134 at 20°. A combustion gave the following results : —

0.1511 grm. of the oil gave 0.4733 grm. C02 and 0.1937 grm. H20.

Calculated for C13H26. Found.

C 85.70 85.41

H 14.30 14.24

This formula was also verified by a determination of its- molecular
weisrht : —

1.2712 grm. of the oil and 25.24 grms. benzol gave a depression of 1°.370.

Calculated for C13H26. Found.

182 180

A determination of the index of refraction gave 1.4745, and the molec-
ular refraction 60.254 ; calculated for C13H2G, 59.839.

Tetradekanaphtene, C14H28.

A considerable quantity of distillate collected at 144°-146°, which
gave as its specific gravity before purification 0.8428, and after purifica-
tion 0.8154. A combustion gave percentages of carbon and hydrogen
required for the formula C^H^.

0.1462 grm. of the oil gave 0.4600 grm. C02 and 0.1824 grm. H20.

Calculated for C,,!!^. Found.

C 85.70 85.81

H 14.30 13.87

A determination of molecular weight at the freezing point gave a value
required for tetradekanaphtene : —

1.3356 grm. of the oil and 24.22 grms. benzol gave adepression of 1°.394.

Calculated for C14H23. Found.

196 194

The index of refraction of this hydrocarbon was found to be 1.4423,
and the molecular refraction 63.75 ; required for C14H28, 64.44.

282 PROCEEDINGS OF THE AMERICAN ACADEMY.

Pentad i'.kanai'htknk, C^H^o.

From each of the California oils, distillates in vacuo collected at
160°— 162 . 50 mm., corresponding nearly to 260°— 262°, atmospheric
pressure. The specific gravity of the fraction from Scott's Hill oil was
0.8600, and after purification with fuming sulphuric acid, 0.8171. A
combustion gave the following values for carbon and hydrogen : —

0.1454 grm. of the oil gave 0.4558 grm. C02 and 0.1829 grm. ILO.

Calculated for C^IIm. Found.

C 85.70 85.47

II 1 1.30 13.97

The percentages of carbon and hydrogen in the analysis of the two
hydrocarbons last described indicate a falling off in the proportions of
hydrogen. The deficiency in the analysis is more probably due to loss
of carbonic dioxide than of water. Similar variations have been noted in
connection with some of the lower hydrocarbons in other crude California
oils. The deficiency in hydrogen may indicate, as mentioned heretofore,
the presence of hydrocarbons composed with more than one methylene
ring, which would require prolonged distillation for their complete
removal. From the heavier California oils composed, it appears, largely
of asphaltic hydrocarbons, the falling off in the proportion of hydrogen
and consequent increase in carbon indicates the presence of hydrocarbons
with the formula CnII2„_2. Such differences do not appear in determina-
tions of molecular weights, but are shown by analysis.

There are wide variations in the specific gravity of the distillates of
California oil above 230° from different sources. This may indicate a
certain proportion of hydrocarbons of a lower series than C„II ... and the
higher percentages of carbon and lower percentages of hydrogen in some
of these oils indicate the possibility of hydrocarbons Cjr,„_2.

For faithful assistance in this work the following gentlemen should
reci lil : Messrs. Shaw. Ames. Richards, Cushing.

From this examination of California petroleum, the following conclu-

18 may be drawn : —

An essential characteristic is the relatively small proportions of the
distillates below 225 . The main body of the crude oils from the prin-
cipal fields distilling below 225 is composed of methylenes which resem-
ble those identified in Russian oil, in boiling points and in specific gra\ ity,
■ I » r andekanaphtene, (MII .. dodekanaphtene, C II.,. and trideka-
naphtene, Cl:jII20, which differ in boiling points. The proportion of the

MABERY. — COMPOSITION OF PETROLEUM. 283

aromatic hydrocarbons is much larger, apparently, in California oil. The
homologues of benzol form a considerable proportion of the distillates,
especially of those with lower boiling points. In the distillate 221°-
222° from Puente oil so much naphtalene was present that the distillate
became solid at 0°.

California petroleum differs totally from the Eastern oils, — Pennsyl-
vania, Ohio, Canadian, etc., — and also materially from Russian oil, in
not containing members of the series C,^,,,^. In this respect, and in
respect to the large proportion of aromatic hydrocarbons, California
petroleum is unlike any other petroleum that has been examined in this
Laboratory. Incidentally it may be mentioned that California petroleum
differs from other petroleums hitherto examined in the large proportions
of oxygen and nitrogen compounds which it contains. These bodies are
under investigation in this Laboratory.

Study of the portions with high boiling points, which is now in prog-
ress, will have an especial interest, since, when they are separated with-
out decomposition, they form the most valuable constituents of lubricating
oils and asphalts that have been separated from petroleum. In some of
the high distillates, such as those from Summerland oil, hydrocarbons of
the series C„H2n_2 and the series QJIon^ have been identified.

No. 36. —OX THE CHLORINE DERIVATIVES OF THE HYDRO-
CARBONS IN CALIFORNIA PETROLEUM.

Bt Charles F. Mabery and Otto J. Sieplein.

In further confirmation of the composition of the hydrocarbon 68°-70°
described in the previous paper (Mabery and Hudson), the chlorine de-
rivatives were formed by exposing the hydrocarbon over water to the
action of chlorine, in ordinary daylight. After washing and drying, the
chlorine product was fractioned under atmospheric pressure until it col-
lected at 125° -130°, and for the most part at 126°. It distilled con-
stant under normal conditions with the mercury all in the vapor at
125°. 5.

The specific gravity at §£ was 0.9255; at g£, 0.9239 ; at ^, 0.9143;
and at fj£, 0.9044. The coefficient of expansion calculated from the
average of these determinations is 0.000918.

284 PROCEEDINGS OF THE AMERICAN ACADEMY.

A chlorine determination gave the following percentage: —
0.1572 grm. of the oil gave 0.1872 grm. AgCl.

Calculated for C0H„C1. Found.

CI 29.92 29.50

The molecular weight was determined at the freezing point of benzol..
1.1611 grm. of the oil and 1'.). !>2 grm. benzol gave a depression of
•_' .482.

Calculate l for (', H,,C1. Found.

118.5 118

The index of refraction at 20° was found to be 1.41 G, and the molecu-
lar refraction, 33.29. Required for ( I !,.''!. 32.54. Hexamethylene is,
therefore, the principal hydrocarbon with this boiling point.

In distilling the portions of California petroleum below 100°, it has
always been observed that a distillate collected at 90°-91°. To ascer-
tain whether a hydrocarbon were really present with this boiling point,
distillation of the fractions ,s.V'-l<»i> was continued through a tall
Hempel column until a larger portion collected at 89°-90°. The specific
gravity of this fraction without purification was 0.720.3, HJ-| . After
thorough treatment first with common sulphuric acid, then with fuming
acid, the specific gravity was not changed, 0.7295. The index of re-
fraction of this hydrocarbon at 20° was 1.41 1. and the molecular refrac-
tion 33.35 ; calculated for C7IIi4. 32.22. The molecular weight at the
freezing point was found to be as follows: —

0.8G09 grm. of the oil and 17.02 grms. of benzol gave a depression
2 .547.

Calculated for <'.Ult. Found.

98 99

With the mercury column all in the vapor, Bar. 745.3 nun., this hy-
drocarbon distilled completely at 90 . 1.

The chlorine derivative of this hydrocarbon was formed by the action

blorine over water. After washing, drying, and distillation under

atmospheric pressure) the chloride came together at 1 15 '-150°, for the

ino-t purt at 117.

A determination of chlorine gave a value required for the mono-
chloride : —

MABERY. — COMPOSITION OF PETROLEUM. 285

0.1605 grm. of the oil gave 0.1 GOG grm. AgCl.

Calculated for CTH13C1. Found.

CI 26.77 26.13

The specific gravity of the chloride at §£ was 0.9332 ; at f£, 0.931 G ;
at f1,-,, 0.9231 ; at f^o, 0.9138. The coefficient of expansion for one de-
gree calculated from these results is 0.000973.

A determination of the molecular weight at the freezing point gave
the following value: —

o

0.9318 grm. of the oil and 19.98 grms. of benzol gave a depression of
1°.744.

Calculated for 07H13C1. Found.

132.5 131

The index of refraction at 20° was found to be 1.441, and the molec-
ular refraction, 37.57; calculated for C7H13C1, 37.11.

These results are sufficient to establish the formula for this hydro-
carbon as C7H14. It is probably dimethylpentamethylene. It differs in
its properties from methylhexamethylene, boiling point 99° -100° ; its
chloride boils at 147°, while methylhexamethylene chloride boils at
141°-142°.

Metiitlhexamethylexe Chloride, C7H13C1.

To ascertain the correct boiling point of heptamethylene, the dis-
tillates 95°-l00° were carried through a series of distillations until the
greater portion collected at 98°-100°, and this product distilled for the
most part at 99°-100°, Bar. 745°. 3, with the mercury column wholly
in the vapor.

The empirical formula of this hydrocarbon has been ascertained in the
distillates from various specimens of California oils, by analysis and de-
terminations of molecular weight. The boiling point of hexahydrotoluol,
prepared by the addition of hydrogen to toluol, was given as 97°. Mark-
owuikoff found that the same hydrocarbon separated from Russian
petroleum, and also the synthetically prepared heptanaphtene, boiled
at 101°.

For further identification of our product it seemed advisable to study
its derivatives. The chlorine derivative was first formed by passing
chlorine into the hydrocarbon over water until the greater part was
converted into the chloride. This reaction takes place very readily in

286 PROCEEDINGS OF THE AMERICAN ACADEMY.

ordinary daylight, and with a large generator the action of chlorine
may be made continuous, saving much time in the chlorination. This
method seems to be more advantageous than that formerly used by as
in which dry chlorine was allowed to act on the vapor of the dry oil.
The chlorinated oil was washed, dried over calcium chloride, and frac-
tiooed. After five distillations through a Ilempel column, a consider-
able portion collected at 141°-14"_'J. which proved to be the monochloride.
and more remained above 160°, which was doubtless the dichloride.

A determination of chlorine in the fraction 141°-142° by the method
of Carius gave a value required for C7Hi:iCl: —

0.1863 grm. of the oil gave 0.2025 grm. AgCl.

Calculated for C7H13C1. Found.

CI 26.77 2G.88

The molecular weight was also determined by the Beckman freezing
point method: —

1.1960 grm. of the oil and 25.8960 grms. benzol gave a depression of
1°.701.

Calculated for C7II13C1. Found.

132.5 133

The specific gravity of the monochloride at 20° was found to be
0.9310. For further confirmation of the formula, the index of refrac-
tion was determined with the aid of a Pulfrich refractometer, and from
the density and molecular weight the molecular refraction was calcu-
lated. The index of refraction found was 1.441, and the molecular
refraction : —

Calculated rorC7HuCl. Foui, I

37.11 37.57

The theoretical value was calculated on the assumption that all the
carbons are singly connected.

In order to ascertain whether the chlorine atom enters the ring or side
chain, thenitril was formed by heating the chloride with alcoholic potassic
cyanide for several hours. There was an abundant separation of potas-
sic chloride, aid on diluting tin- solution the nitril separated as an oily
liquid above the water. This oil had the characteristic «><lor of the
nitril-. Ik specific gravity al 20 was 0.9258, and it- index of refraction,
1.45. Tli'- molecular refraction, assuming the molecular weight, was

MABERY. COMPOSITION OF PETROLEUM. 287

35.78, calculated from the elements, assigning to the cyanogen group
the value 5.33, which was determined from ethyl cyanide, C2H5CN.
The theoretical molecular refraction for the nitril calculated on the
same basis is 36.45.

The nitril was saponified by heating it to 110° with concentrated
hydrochloric acid. The acid formed had the odor of alphatoluic acid,
but the amount obtained was not sufficient for complete examination.

When heptamethylene chloride was brought together with metallic
sodium, a vigorous action soon set in with the evolution of oreat heat,
sufficient to cause the decomposition of the products unless it was con-
trolled by cooling. This reaction was best carried on by dissolving the
chloride in ether, adding the sodium in slight excess over the calculated
amount, and keeping the solution cold. In two hours the reaction was
complete. The products included an unsaturated hydrocarbon, boiling
point 97°, and another hydrocarbon, boiling point approximately 220°-
230°. That the hydrocarbon boiling at 97° was unsaturated was shown
by the formation of a bromine addition product ; bromine added readily
in the cold with no escape of hydrobromic acid. A Carius determina-
tion of the bromine in this product gave the following result : —

0.2045 grm. of the oil gave 0.3023 grm. AgBr.

Calculated for C7HI2Br,. Found.

Br 62.5 62.9

The specific gravity of the dibrom-derivative at 20° was 1.648. A
determination of its molecular weight was made : —

0.9233 grm. of the oil and 25.3 grms. of benzol gave a depression of

0°.725.

Calculated for C7H12Br2. Found.

256 247

The odor of the unsaturated hydrocarbon was very sharp and pene-
trating, very different from that of the paraffine and methylene hydro-
carbons, but resembling the olefines. In further proof that it contained
doubly bonded carbon, it was titrated with the Hiibl reagent, an alcoholic
solution of iodine and mercuric chloride. The amount of iodine absorbed
was approximately equivalent to two atoms of iodine for each molecule
of the hydrocarbon. When heated with hydriodic acid, the monoiodide
was readily formed. The specific gravity of the unsaturated hydrocarbon

PROCEEDINGS OF THE AMERICAN ACADEMY.

at 2<> was (i.7 172. Its molecular weight was ascertained by the freezing
point method : —

0.7116 grm. of the oil and 25.19 grins, of benzol gave a depression of
1.420.

Calculated firi'-II,,. Pound.

96 98

The index of refraction with sodium light was found to be 1.416.

The molecular refraction calculated from the density and molecular

weight was : —

Calculated for Found

1 ! I ' 1 1 . C);IIjrt CKg.

30.06 31.77 32.34

These values indicate that the unsaturated hydrocarbon contains a
double bond between the side chain carbon and a carbon atom in the
ring, confirming the position of the chlorine atom in the side chain which
is indicated by the ease with which it is replaced in the reaction with
potassic cyanide. The form that the chlorination takes doubtless de-
pends on the facl that the hydrogen in the side chain is Less firmly bound
than the hydrogen atom- in 'he methylene ring. The unsaturated con-
dition C8H10=CH2 is shown by the action of halogens and haloid acids,
and confirmed by the molecular refraction, which corresponds to the sum
of the atomic refractions assuming the double bond.

This constituent of California petroleum is therefore identical with
methyl hexamethylene, which maybe formed by the addition of hydrogen

to toluol.

The hydrocarbon with a boiling point 220 •_,"I(»'D, formed by the

action of sodium on methyl hexamethylene chloride, doubtless contains

methylene rings: — ( '.-. IIn( 'I I ,. ( 'II ..(',-, II,,. The quantity of this

product formed was too small for identification, bul it evidently affords

-■in- for building up the higher methylene hydrocarbons containing

more than one methylene ring. That it was a condensed hydrocarbon

shown by it- high specific gravity, 0.8872.

1 inn .'i in I ill \ \mi i HI i.i m < Ihloridi . ' 1 1, -CI.

The formula of dimethylhexamethylene was ascertained, as shown in
the pre\ion- paper, hy analysis and determination of its molecular weight
The formula was -till further verified by the formation of the chlorine
derivative. As in the chlorination of methyl hexamethylene, chlorine

MABERY. COMPOSITION OF PETROLEUM. 289

was allowed to act on the hydrocarbon over water in ordinary daylight
until it was nearly all converted into the chlorine derivative. After
washing and drying, the product was distilled under atmospheric pressure ;
after several distillations the monochloride collected in larger part at
168°-170°. The specific gravity of this product at 20° was found to be
0.9358. A determination of chlorine gave the following result: —

0.2034 grm. of the oil gave 0.1973 grm. AgCl.

Calculated for C8H15C1. Found.

CI 24.21 24.01

The molecular weight of the chloride was determined by the freezing
point method.

0.7293 grm. of the oil and 25. G6 grms. benzol gave a depression of
0°.982.

Calculated for C8H15CL Found.

146.5 142

The index of refraction in sodium light at 20° was 1.455, and the mo-
lecular refraction : —

Calculated for C8H15C1. Found.

41.69 41.60

That the chlorine enters a side chain in this reaction, as in the case of
methylhexaniethylene, appears from the ready formation of the nitril by
heating the chloride with alcoholic potassic cyanide. On diluting the
solution the nitril separated as an oil, with the characteristic odor of the
nitrils. It was saponified by heating with aqueous potassic hydrate. On
acidifying the solution, a solid was precipitated, with an odor characteristic
of the alpha-toluic acids; but the quantity obtained was not sufficient for
identification. It was probably meta-methyl, alpha-toluylic acid.

TRniETHYLHEXAMETHYLENF. CHLORIDE, C0H17C1.

This chloride was also formed, washed, dried, and fractioned under
atmospheric pressure. It came together in larger quantity at 186°-188°;
its specific gravity at 20° was 0.9380. The percentage of chlorine was
determined.

0.2300 grm. of the oil gave 0.2041 grin. AgCl.

Calculated for C9H17C1. Found.

CI 22.10 21.94

vol. xxxvi. — 19

290 PROCEEDINGS OF THE AMERICAN ACADEMY.

The molecular weight was also ascertained at the freezing point.

0.8672 gnu. of the oil ami 25.03 grins, henzol gave a depression of
1°.090.

Calculated for i',,II17CL Found.

160.5 1 ."if.

In further support of the formula of the chloride, the index of refrac-
tion was determined, L.462, and the molecular refraction calculated.

Calculate! fort',II,;('l. Found.

46.3 47

"When sodium was allowed to act on trimethylhexamethylene chloride,
a vigorous reaction set in that could be controlled by cooling. By carry-
ing on the reaction in an ethereal solution, on standing over night in
water the reaction was complete. The product of the reaction consisted
for the most part of an unsaturated hydrocarbon, and a smaller quantity
of a heavy oil, doubtless formed by the union of two methylene rings.
The boiling point of the unsaturated hydrocarbon was L3o°— 140°. Its
unsaturated condition was shown by the readiness with which it united
with iodine in Iliibl's reagent, absorbing approximately two atoms of
iodine. The specific gravity of the unsaturated hydrocarbon at 2U° was
0.7762. The molecular weight at the freezing point was determined.

0.8-372 grm. of the oil and 22.45 grms. benzol gave a depression of 1°.57 1 ,

Calculated for C,jII10. Found.

I _' I 120

The index of refraction was found to be 1.431, and the molecular
refraction : —

Calculated for Fn„n,i

c,nT(Ciis),. c ii . cir.,1.. rii.. f0Unu-

39.26 40.:»7 11.46

Dekanaphtene Chloride, Ci0H19C1.

In confirmation of the formula of dekanaphtene, the index of refraction
was determined and the molecular refraction calculated. With the den-
Bit) 0.7898, and the index 1.1325. the molecular weight gave the
following value : —

Galon! "•■ i IbrC B Pound.

i«;.l0 ir..00

MABERY. — COMPOSITION OF PETROLEUM. 291

The chlorine derivative of this hydrocarbon was formed by passing in
chlorine over water. After washing, drying, and fractioning in vacuo,
the chloride collected in larger quantities at 105°-110° (50 mm.) An-
other portion collected at 140°-145°, probably a dichloride. The specific
gravity of the monochloride at 20° was 0.9470. A determination of
chlorine gave the following value : —

0.1840 grm. of the oil gave 0.1476 grm. AgCl.

Calculated for C10H,9C1. Found.

CI 20.33 19.80

The molecular weight of the chloride at the freezing point of benzol
was also determined : —

1.2536 grm. of the oil and 21.29 grms. benzol gave a depression of

1°.6S9.

Calculated for C,0H19C1. Found.

174.5 171

A determination of the index of refraction of this chloride gave 1.468.
The molecular refraction calculated as follows : —

Calculated for C10H19C1. Found.

50.89 51.34

On account of the small quantity of the dichloride obtained, it was not
possible to purify it completely by distillation. But a determination of
chlorine gave 31.53 per cent; required for C10H18C12, 33.90. A deter-
mination of molecular weight gave 199 ; required for the dichloride, 209.
The molecular refraction calculated from the index was 56.98 ; calculated
from the formula, 55.81. On account of the differences between the
theoretical values for the mono- and di-chlorides, these values are sufficient
to show that the dichloride was formed and separated in nearly a pure
form.

Undekanaphtene Chloride, CnH2iCl.

The index of refraction of this hydrocarbon was found to be 1.4403,
and the molecular refraction : —

Calculated for CnH22. Found.

50.70 50.63

This determination was made in the distillate 190°-192°, which was
used for all work on this hydrocarbon.

292 PROCEEDINGS OF THE AMERICAN ACADEMY.

The chlorine derivative was prepared in the same manner as those of
tin- hydrocarbons previously described. On fractioning it in vacuo at
35 nun. it collected for the most part at 125o-130o. Its specific gravity
was found to be 0.9583 at 20°. A chlorine determination supported the
formula of the monochloride.

0.1560 grm. of the oil gave 0.1179 grm. AgCl.

I culated for i',,II„n. Found.

18.81 18.G9

The molecular weight was determined at the freezing point of benzol.
O.GOoo grm. of the oil and 17. 37 benzol gave a depression of 0°.890.

Calculated for C„H31C1. Found.

188.5 L92

index of refraction was found to be 1.47G, and the molecular
refraction : —

Calculated for <', ,1^,01. Found.

55.48 54.32

The composition of its chloride, together with its molecular weight and
molecular refraction, all show that undecane has for its boiling point
194 -196.°

DODEKANAPHTENE ClILORIDE, C12Ho3Cl.

The index of refraction of dodekanaphtene was found to be 1.4G49,
and the molecular refraction: —

Calculated f"r C, ,11 ,. ind.

55.38 55.24

The chlorine derivative was formed, washed, dried, and fractioned in
van,,,, it collected for the most part, at 130 L35 (17 mm.) Its spe-
cific gravity at 20 was 0.9616. A determination of chlorine gave the
follow ing result : —

0.1657 grm. of the oil gave 0.1 L53 grm. AgCl.

Calculated for <', UMC1. Pound.

CI L7.52 17.20

The molecular weight was also determined: —

MABERY. — COMPOSITION OF PETROLEUM. 293

0.7172 grin, of the oil and 19.04 grms. benzol gave a depression of 0.925.

Calculated for C12HMC1. Found.

202.5 204

The index of refraction was 1.480, and the molecular refraction : —

Calculated for C12H23C1. Found.

60.77 59.9G

Tridekanaphtene Chloride C13H.25C1.

The index of refraction of this hydrocarbon as determined is 1.4745,
and its molecular refraction : —

Calculated for C13n28. Found.

60.25 59.84

The chlorine derivative was formed by the action of chlorine, dried
and fractioned under 17 mm. ; it came together for the most part at
140°-145°. Its specific gravity at 20° was 0.9747; a Carius deter-
mination gave the following percentage of chlorine : —

0.1674 grm. of the oil gave 0.1137 grm. AgCl.

Calculated for C13H25C1. Found.

16.38 16.78

In the chlorine derivatives of the higher hydrocarbons the weight of
oil that can be taken for analysis is limited, since as in the last analysis
the pressure of the large volume of gases formed is more than the tube
can stand. The molecular weight of tridekanaphtene chloride was
determined at the freezing point of benzol : —

0.4447 grm. of the oil and 17.88 grms. benzol gave a depression of 0.763.

Calculated for C^H^Cl. Found.

216.5 217

Tetradekanaphtene Chloride, CuH27C1.

Tetradekanaphtene chloride was readily formed by passing chlorine
into the hydrocarbon over water in ordinary daylight ; if the action was
stopped with a small portion of the hydrocarbon unacted on, the product
consisted for the larger part of the monochloride. After washing and
drying, the chlorine derivative was separated by continued distillation in

20-1 PROCEEDINGS OF THE AMERICAN ACADEMY.

vacuo under 13 mm. ; it came together for the most part at 150°-155°.
The specific gravity of the monochloride was 0.97 18 U ; at |j 0.9730 ; at

. 0.9661; and at \[' ■ 0.9579. The mean coefficient of expausion,
within 20 -40°, from these results, is 0. 178.

A determination of chlorine gave a value required for the mono-
chloride : —

0.1559 grm. of the oil gave 0.0989 grm. AgCl.

Calculated for CuII,;i'l. Found.

CI 15.39 15.80

As mentioned above, the decreasing proportion of chlorine in these
derivatives with the increasing molecular weight, and the consequent
larger volumes of gases, beyond the strength of the ordinary Carius
tubes, gives a smaller weight of silver chloride than could be desireil.
but the accuracy of the method permits of sufficiently reliable result-.
even with the small weights.

The molecular weight of the chloride was determined at the freezing
point of benzol : —

0.8923 grm. of the oil and 17.76 grms. benzol gave a depression of 1°.094.

Calculated for CI4II..7C'. Found.

230.5 225

The index of refraction was found to be, 1.493, and the molecular
refraction CM. 07 ; calculated for C14H27C1, 69.87.

The boiling point of this chloride cannot be accurately determined
under atmospheric pressure because it is rapidly decomposed at the
higher temperature and in presence of air. although it may be distilled
indefinitely in vacuo. Probably the boiling point under atmospheric
pressure is not far from 275°.

l'i \ i \i»i k wAiim.Ni: Chloride, Ci6HS9C1.

This chloride was prepared from i" ntadekanaphtene, which had been
well fractional in vacuo. The-.- higher chlorides are forme,] as readily
as the lower ones. After washing and fractional distillation under
II mm. this monochloride came together for the most part at 170 -175°
without decomposition. It- boiling point under atmospheric pressure is
probably near 800 . Its specific gravity at £j was 0.9771; at jjj 0.9758;

and at ','! , 0.9714 ; and l.9(i 13. The coefficient of expansion calcu-

MABERY. — COMPOSITION OF PETROLEUM. 295

lated from these values is 0.000576. A Carius determination gave the
required percentage of chlorine.

0.1536 grm. of the oil gave 0.0923 grm. silver chloride.

Calculated for C15HMC1. Found.

14.51 14.85

A determination of the molecular weight supported the same formula.
0.9073 grm. of the oil and 18.93 gnus, benzol gave a depression of 0°.984.

Calculated for ClBHMCl. Found.

239 244.5

The index of refraction was 1.493, and the molecular refraction, 72.90;
calculated for C15H29C1, 73.97.

Attempts will be made to form the chlorides of the higher hydro-
carbons in connection with the study of the composition of these
bodies.

No. 37. —ON THE COMPOSITION OF JAPANESE PETROLEUM.

By Charles F. Mabery and Shinichi Takano.*

The oil fields in Japan are the most promising and are under the most
rapid development of any recently discovered oil territory. The output
in 1891 was 56,000 bbls. annually; in 1899 it was 1,000,000 bbls.
The oil territory in Japan is contained in the province of Echigo, at
least to the extent of 90 per cent, on the northern coast of the Sea of
Japan. This province is surrounded in part by a mountain chain which

* Mr. Takano has spent three years in the study of petroleum in this country,
having been sent for this purpose by the Japanese Government. He is thoroughly
familiar with the geology of the oil territory in Japan, having given especial at-
tention to this subject in a thesis which he presented for a degree in the University
of Tokio. Besides a thorough knowledge of the composition of petroleum from a
chemical point of view, which he has gained during the two years he has spent
exclusively in the study of this subject in this Laboratory, Mr. Takano spent one
year as a laborer in different refineries, where he took charge by actual manipula-
tion of every process in the preparation of commercial products. The work pre-
sented in this paper formed the subject of a thesis by Mr. Takano for the degree
of M.S.

296 PROCEEDINGS OF THE AMERICAN ACADEMY.

encloses a section of country 140 miles long by 1^0 miles broad, rising
gradually from the sea to the mountains with a varying altitude from
L 50 feet to 500 feet. Oil is found chiefly in the Upper Eocene of the
Tertiary formation, where it is held under pressure between impervious
layers of shite and sandstone.

Nodules of calcite are frequently found embedded in the sandstone,
and crystals of calcite twinned, resembling those found in the Ohio Tren-
ton-limestone, oil-bearing rock. The oil strata are full of sea-shells, a
good indication as to the origin of the oil formations. There are great
variations in the depths of the wells. For example, the wells of the
Amaze held are approximately 2000 feet in depth, while those of the
Miyagawa field, only six miles distant, are 700 feet deep ; the wells of
the Niitsu field are still shallower. 600 feet. The Nagaoka wells are
800 feet in depth. Nevertheless the oil strata at the different depths
are essentially of the same formation. The sandstone in which the
Nasaoka oil is found is coarser than that of the other fields. On the
surface of the Niitsu oil territory are immense beds of peat which are
used for fuel.

Although the oil fields of Japan are situated in close proximity, speci-
mens of oil from the different fields differ as essentially in composition
as the variations in specific gravity indicate. As will appear later,
the Amaze. Miyagawa, and Ilirei fields yield paratline in considerable
quantities. While the hydrocarbons above (din do not include mem-
bers of the serie> ('11 .,, at least so far as it appears from analysis of
distillates prepared on a laboratory scale, the lower distillates do con-
tain members of the parafline series.

The parafline hydrocarbons with low boiling points, such as the
butane,, pentanes, and hexanes, have an agreeable sweetish odor that is
easily recognizable, and quite different from the pungent harsh odor of
the methylenes. The odor of the paraffines is more apparent in the
Miyagawa oil than in any other of the Japanese oils we have seen, and
somewhat less in the Amaze oil. When these hydrocarbons with low
boiling points are present, the crystalline parafline hydrocarbons ( M
are usually to he found.

( onsiderable quantities of gas were formerly obtained from the Amaze
oil territory, but the Bupply now Beems to be exhausted, at least -.. far as
wells have been bored; Bome gas • from the shallow Miyagawa

well-, more from the shallow wells of the Niitsu territory. As wells are
sunk deepei in this territory it is probable that lighter oils will lie found.
On accounl of the triable condition of the oil rock and the danger of

MABERY. — COMPOSITION OF PETROLEUM.

297

clogging the wells by loose material, " shooting " the wells after boring
is not practised. The flow from the wells is irregular; frequently it
will amount to 300 barrels a day for a week, and then stop, and the oil
must, be pumped.

The oil territory in Japan includes the following fields: — Amaze,
Nagaoka, Niitsu, and Hiyama. The different sections of the oil terri-
tory may be classified as follows: —

Amaze \

Miyagawa \ green oil

Gendogi )

Hirei )

Kitatani > dark oil

Katsudo )

Koguchi )

Kusodsu )

Hiyama { Hiyama } dark green oil

Amaze

Nagaoka

Niitsu

The Amaze field is the oldest ; the Hiyama field has been most re-
cently developed; both these oils contain paraffine.

The crude oils from the different fields differ essentially in their prop-
erties, as shown by the different specific gravity, the different percentages

Amaze.

Hirei.

Katsudo.

Kitatani.

Koguchi.

Kusodsu.

Miyagawa.

Specific Gravity . . .

0.8245

0.8622

0.8771

0.8952

0.9435

0.9210

0.8911

Per cent Sulphur . .

0.23

0.41

0.82

0.61

0.49

0.37

0.32

Per cent Nitrogen . .

0.35

0.74

0.97

0.75

1.34

1.23

0.55

Iodine Absorption . .

0.0

0.84

7.66

0.82

9.79

0.62

1.63

Coefficient of Expansion

82.5

78.5

76.5

79.5

67.5

61.5

74.5

Percentages of Distillates.

Amaze.

Hirei.

Katsudo.

Kitatani

Koguchi.

Kusodsu.

Miyagawa.

- 150°

22.8

21.8

22.2

14.

0.0

0.0

15.

150°-300°

40.2

38.3

38.8

38 8

25.0

25.0

36.8

+ 300°

37.

39.9

39.

47.2

75.

75.

48.2

_ ^ PROCEEDINGS OF THE AMERICAN ACADEMY.

of sulphur, nitrogen, and the different proportions in which they distil at
different temperatures. There are also marked differences in iodine ab-
sorption, and in the coefficients of expansion. The latter were determined
by ascertaining the specific gravity at <3°, 10°, 20°, and 25°, and divid-
ing the differences in specific gravity at the different temperatures hy
five and multiplying the quotient hy 100,000, the method described in
Redwood's treatise on petroleum.

In the development of oil territory hitherto, no attempts have been
made to ascertain the series of hydrocarbons which compose the main
body of the crude oil. Beside the work of Pelouze and Cahours,
Schoelemmer, and Warren on the Pennsylvania and Canadian oils, and
the work carried on in this Laboratory, no attempts have been made to

ermine the form of the hydrocarbons in the lower distillates of
American oils, and nothing whatever beside the unpublished work of
this Laboratory <>n the determination of the composition of the portions
with higher boiling points. Beside the work of Markownikoff on the
Russian oils and the work of Warren and Storer on Rangoon petroleum,
very little lias been done in this direction on oils from other fields. On

account of the ease in the preparati >f commercial products from the

lighter oils of Pennsylvania and Ohio, the ultimate composition was of
less importance than it is now becoming in the development of oil fields
that yield heavier crude oils, such as the oil territory in California,
Texas, South America, Japan, and numerous other fields recently dis-
covered. The methods that must be applied to these heavy oils are
essentially different from the methods that have been universally in nse
Bince the beginning of the oil industry. In Japan, the promoter- of
those oil fields will have the advantage not only of all former experiei
in oil refining, but the further advantage of a knowledge of the hydro-
carbons which form the main body of the crude oils. Japanese petro-
leum apparently differs from other heavy petroleums in that it contains

Bmaller amount- of the benzol homologies. Benzol and its bomolosrues

were found in the Amaze oil, and Bome of the other crude oils, but
fuming Bulphuric acid failed to reduce materially the specific gravity of

eral of the distillates that should yield benzol hydrocarbons, if tl
were pr< sent.

Constituents oi Petroleum from mi Amaze Field.

The lightest oil from the Japanese fields is found in the Amaze terri-
tory. It consequently contain! the largest proportion of more volatile

MABERY. COMPOSITION OF PETROLEUM. 299

constituents. A combustion of Amaze crude oil gave 84.66 per cent
carbon and 13.22 per cent hydrogen. After several distillations under
a Hempel column, fractions collected in larger quantities at 68°, 98°,
119°, 135°, 62°, 96°, 216°. No attempts were made to ascertain the
form of the hydrocarbons below 68°. The distillate collected at 68°
gave as its specific gravity 0.7343, which indicated that hexane was not
present in any considerable quantity. But nothing further was done
toward identifying hexamethylene, which no doubt is the principal
hydrocarbon in this distillate.

The distillate collected at 98° -100°, after purifying with fuming sul-
phuric acid, gave as its specific gravity at 20°, 0.7450.

A combustion gave the following percentages of carbon and hydro-
gen : —

0.1347 grm. of the oil gave 0.4221 grm. C02 and 0.1721 grm. H20.

Calculated for C-H,,. Found.

C 85.70 85.45

H 14.30 14.20

Its index of refraction was found to be 1.4174, and the molecular
refraction, 33.14; calculated for C7IIi4, 32.22. This hydrocarbon was
therefore methylhexamethyleue.

The distillate 118°-120r, after purification with fuming sulphuric
acid, gave 0.7621 as its specific "gravity at 20°. A combustion gave the
following percentages of carbon and hydrogen : —

0.1406 grm. of the oil gave 0.4378 grm. C02 and 0.1812 grm. H20.

Calculated for C8HIl3. Found.

C 84.70 84.94

H 14.30 14.33

A determination of the index of refraction of this oil gave 1.4256, and
the molecular refraction, 37.68 ; calculated for C8H16, 36.82. This
hydrocarbon was therefore dimethylhexamethylene.

The distillate collected at 134°-135° was purified with fuming sul-
phuric acid and analyzed.

0.1528 grm. of the oil gave 0.4805 grm. C02 and 0.1950 grm. H20.

Calculated C0H13. Found.

C 85.70 85.76

H 14.30 14.29

300 PROCEEDINGS OF THE AMERICAN ACADEMY.

It give as its specific gravity at 20°, 0.7787.

The refractive index of this oil was found to be 1.4348. and the
molecular refraction, 12.27 ; calculated for C.,II18, 41.42. This con-
stituent was therefore trimethylhexametbylene.

A distillate collected in considerable quantity at 1G0°-1G2°, which
was purified with filming sulphuric acid and analyzed: —

0.2055 grm. of the oil gave 0.6445 grm. CO, and 0.2556 grm. 11,0.

Calculated for C10H20. Found.

C 85.70 85.."."

II 14.30 13.82

This "il gave as its specific gravity at 20°, 0.7002. Its index of
refraction was 1.4418, and its molecular refraction, 4G.04 ; calculated
for CxoH-jo, 46.03.

The distillate at 100°-102°, after purification with fuming sulphuric
acid, gave as its specific gravity at 20°, 0.80G1. Its composition was
determined by analysis : —

0.1G12 grm. of the oil gave 0.5046 grm. CO, and 0.2025 grm. 11,0.

Calculated for '',,11^. Found.

C 85.70 85.35

II 1 1.30 13.96

The molecular weight of this hydrocarbon at the freezing point of
benz<>l was 156; calculated for ('nil,,. 154. The index of refraction
at 20 was 1.4482, and the molecular refraction 51.24; calculated for
C„II,,. 50.63.

The distillate 212°-214°, purified with fuming sulphuric acid, gav<
its specific gravity at 20°, 0.8165. A combustion gave the following
proportions of carbon and hydrogen: —

0.1875 grm. of the oil gave 0.5870 grm. CO, and < >.J 1 "> 1 grm. 11,0.

Calculated (brC„HM. Found.

C 85.70 's-"'-"»l

II 14.80 14.52

A determination of it- molecular weight at the freezing point gave
172; calculated for C,..!!.,,. 168. fa index of refraction was L. 4535, and
the molecular refraction, o<~>. 70; calculated for CigHM) 55.23.

MABERY. COMPOSITION OF PETROLEUM. 301

Composition of the Portions of Japanese Petroleum with

High Boiling Points.

Japanese petroleum from different sources differs materially in its com-
position. From such oils as the Hirei no crystalline solids can be sepa-
rated, even at low temperatures. But from others, such as the Amaze,
Miyagawa, aud Hiyama, crystalline solids separate from the higher
fractious. The fraction 310°-315° atmospheric pressure, from Amaze
crude oil became solid on cooliug. The solid portion was separated by
cooling and filtration; it was washed, pressed, and warmed with gasoline
which removed all color. Melting point, 68°. The fractions above 225°
were collected in vacuo under 30 mm. The purified solid from 225°-
230° melted at 70°, that from 250°-275° at 73°, and that from 260°-
265° at 74°. The solid from 22o°-230° gave, by combustion, values
showing it to belong to the series C„H2n+2.

0.1535 grm. of the oil gave 0.4890 grm. CO, and 0.1787 grm. H20.

Calculated for r ,

^\v^w

CnH2«

ruuuu.

c

85.14

85.70

85.29

H

14.86

14.30

14.99

The fraction 250° -260°, 30 mm., from the Amaze oil gave the follow-
ing percentages by combustion, also showing the series CnH;>n+2 : —

0.1654 grm. of the oil gave 0.5161 grm. CO, and 0.2233 grm. H20.

C 85.03

H 15.00

A combustion of the fraction 265°-270° gave values required for a
hydrocarbon of the series C„H2«+2 : —

0.1750 grm. of the oil gave 0.5469 grm. C02 and 0.2346 grm. H20.

C 85.21

H 14.81

In determining the molecular weight of this hydrocarbon at the boil-
ing point of benzol, the following result was obtained : —

0.4190 grm. of the oil and 24.3 grms. benzol gave a rise in boiling point
of 0.126.

Calculated for C^H^. Found.

367 367

302 PROCEEDINGS OF TIIK AMERICAN ACADEMY.

I >i terminations by the boiling point method of hydrocarbons with such
high molecular weights are of necessity somewhat uncertain on account
of the slight rise in boiling point. The question of the molecular weights
of these solid hydrocarbon* will receive more attention with the constitu-
ents of Pennsylvania and California petroleums with high boiling point-.

The specific gravity of the solid 250°-260° is nearly the same as that
of the corresponding hydrocarbon from Penusylvania petroleum: —

Japa

250 mm.

lvnn-\ Ivaoia,'
292 295 . 50 una

60

0.7977

70

0.7943

0.7950

80

0.7920

0.7943

90

0.7918

The specific gravity of the -Japanese solid at 60° could not be deter
mined, because it was not liquid at that temperature.

.Mivacaw \ I'll ROLEUM.

Although this oil is from a field situated only >ix miles from the
Amaze field, it «lilTers essentially in its specific gravity, and in the pro-
portion- in which it distils, from the Amaze oil. The crude oil was
distilled under atmospheric pressure, and the distillation of the lower
portions continued until they came together in larger quantities at tern
peratures at which distillates were collected from the other oils. The
specific gravity of the distillates after treating with concentrated sulphuric
acid is as follow- : —

98 100°. 11- 120°. 184 -186°. 160 162°. 194 196°. 212°-214°. 228 i

0.7364 0.7631 0.7772 0.8088 0.8493 0.8674 0.8770

Tin- last three fractions were also treated with fuming sulphuric acid and
the specific gravitj determined: —

194 212 -214 . 228 -

5412 0.8650 0.8720

The slight change after the thorough treatment with fuming sulphuric
arid shows that no benzol homologues were present in these portions,
v t the specific gravity of the hydrocarbons above 196 is considerably
higher than of those from Amaze oil or (rum Hirei oil. The distillates

Unpublished data.

MABERY. — COMPOSITION OF PETROLEUM.

303

collected under atmospheric pressure were thoroughly treated with
fuming sulphuric acid, washed, dried, and their molecular refraction
determined : —

Distillate.

Refractive Index.

Molecular Refraction.

Calculated.

Required.

98o_100°

118°-120°

134°-136°
160°-162°
194°- 106°

214°-216°

1.4117
1.4163
1.4261
1.4463
1.4605
1.4706

33.14
3692
41.51
46.26
50.27
51.34

32.21
36.82
41.42
46.03
50.03
55.23

Hirei Petroleum.

A combustion of Hirei crude oil gave 82.28 per cent of carbon and
13.19 per cent of hydrogen. In the distillation of the Hirei oil, fractions
collected within the same limits of temperature as those from the other
crude oils. The distillate collecting in the vicinity of 98°, after treat-
ment with sulphuric acid, gave 0.7412 as its specific gravity. Its molec-
ular weight at the freezing point of benzol was 98; required for C7IIU.
98. Its index of refraction was 1.409-3. and its molecular refraction,
32.77 ; required for C.H14, 32.22.

The fraction 11 8° -120° gave as its specific gravity 0.7523. Its mo-
lecular weight was 114; required for C8H16, 112. Its index of refraction
was 1.4151, and its molecular refraction, 37.34; required for C8H1G,
36.82.

The fraction at 135° gave as its specific gravity 0.7676. Its molecu-
lar weight was 128; required for C9H13, 126. Its index of refraction
was 1.4372, and its molecular refraction, 42.12; required for C9II1S.
41.42.

The fraction 162° gave for its specific gravity 0.7887. Its molecular
weight was found to be 137; required for C10H20, 140. Its index of
refraction was 1.4372, and its molecular refraction, 46.60; required for
C10H.20, 46.03.

The specific gravity of the fraction at 196° was 0.8192. Its index of

304 PROCEEDINGS OF THE AMERICAN ACADEMY.

refraction was 1.4516, and its molecular refraction, 50.78 : required for
C„Ho.,, 50.63.

The specific gravity of the fraction 216° was 0.8327, and the molecu-
lar weight, 172; required for C12IIo4, 168. Its index of refraction was
1.4599, and its molecular refraction, 55.22 ; required for Ci2H24. 55.23.

The specific gravity of these distillates from Ilirei oil is somewhat
higher than was found in the corresponding distillates from Amaze oil,
and the differences increase with increasing molecular weights. It was
at first thought that this was due to incomplete removal of benzol hydro-
carbons, but still further treatment with fuming sulphuric acid failed to
diminish these values. It is probable that the oil contains hydrocarbons
with more than one methylene ring.

The results of this examination show that Japanese petroleum is 'din-
posed for the greater part of hydrocarbons of the series CnH2», — the
methylene hydrocarbons. Probably the very heavy oils contain hydro-
carbons with two or more methylene rings, of the series CnII2n_2 or
1 1 I2„_4. Some of the oils contain solid paraffine hydrocarbons, others
do not. The proportion of benzol derivatives in the oils examined is
relatively much smaller than in California petroleum. The proportion
of nitrogen compounds and of sulphur compounds is quite variable. In
some of the oils the percentages were nearly as large as any found in
California petroleum, in others the amounts were much smaller.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 16. —January, 1901.

THE ECLIPSE CYCLONE AND THE DIURNAL

CYCLONES.

Results of Meteorological Observations in the Solar Eclipse op

May 28, 1900.

By II. Helm Clayton.

THE ECLIPSE CYCLONE AND THE DIURNAL

CYCLONES.

Results of Meteorological Observations in the Solar Eclipse of

Mat 28, 1900.

By II. Helm Clayton.

Presented November 14, 1900. Received December 12, 1900.

The path of total solar eclipse in the United States on May 28, 1900,
was visited by a number of experts and trained observers, who took
meteorological observations as a part of the program on the day of the
eclipse. These included Mr. A. Lawrence Rotch and Mr. S. P. Fer-
gusson at Washington, Ga. ; Mr. O. B. Cole at Centerville, Va. ; Mr. G. W.
Pickard at Virginia Beach, Va. ; and myself at Wadesboro, N. C. These
observations were sent to the Blue Hill Meteorological Observatory and
given to me for discussion. Besides these I obtained records from a
number of well-equipped observatories in North America. These in-
cluded the Toronto Observatory, the New York Central Park Observa-
tory, the Blue Hill Observatory, the Belen College Observatory of
Havana, the McGill College Observatory of Montreal, the Meteorological
Station at Cornell University, the City Engineer's Office at Providence,
R. I., and observations by Mr. Eddy at Bayonne, N. J. These obser-
vations were all within the area of partial eclipse, and the data were
furnished by the kindness of the directors.

The details of the discussion of the observations are published in a
Bulletin of the Blue Hill Meteorological Observatory.* The results
embody certain conclusions of general interest which I am permitted to
present to the Academy.

The meteorological changes due to the eclipse were separated from
other changes of greater length, such as the diurnal and cyclonic, by
interpolating a uniform change between the beginning and the end of the
eclipse and subtracting this from the observations. For example, in
Figure 1 is plotted the observations of temperature at Wadesboro, N. C.

* Annals of the Astron. Observatory of Harvard College, XLIII. No. 1.

r,os

PROCEEDINGS OF THE AMERICAN ACADEMY.

The outside vertical lines 1 5 and K show the beginning and end of partial
eclipse, and the central vertical lines T show the times of total eclipse.
The dotted straight line connects the observed temperature at the begin-

7 A.M

8 A.M.

9 A.M.

0 A.M.

80:

70c

60c

3

T

Z^-^

E ^^-

Fig u ke 1.

ning and end of the eclipse, and represents the interpolated uniform
change. The observed temperatures are shown by the unbroken curved
line, and the departures of these from the values represented by the dotted
line is assumed to be the depression of temperature arising from the
eclipse. The pressure, humidity, and vapor tension were treated in the
same manner.

In order to obtain the eclipse wind in velocity and direction, the
observations were treated in the following manner.

In the accompanying diagram, Figure 2, let A B represent the direc-
tion and velocity of the wind prevailing independent of the eclipse, and

Figure 'J.

A.C the wind observed at any momenl daring the eclipse; then com-
pleting the parallelogram of forces, A I) will represent the eclipse wind
in direction and velocity. The prevailing wind was derived from the
mean of the wind- immediately preceding and following the penumbra,
or, what was found to be practically the Bame thing, from the mean wind
direction during the passage of the penumbra, since the eclipse wind was

CLAYTON. — THE ECLIPSE CYCLONE. 309

from opposite directions during this time. The penumbra is used to
indicate the area of partial eclipse, and the umbra to indicate the area of
total eclipse. The mean wind and the eclipse wind were at first deter-
mined graphically for all the stations ; then as the results seemed to be
of importance, they were rigidly computed for all the stations where the
observations were sufficiently accurate to warrant it. These results,
when plotted, indicate very clearly an outflow of wind from around the
umbra, and an inflow around the borders of the penumbra.

But there are certain irregularities due to the normal irregularities
of the wind. In order to diminish the effect of these, I smoothed the

observations by the formula — • These winds were plotted

at their proper places on maps of the United States for 8.15 a.m., 75th
meridian time, when the umbra was about to enter the American conti-
nent from the Pacific, and also plotted for 9 a.m., when the umbra had
passed off the coast of the United States on to the Atlantic. These maps
are shown in Figure 3. The position of the umbra is shown on each
map by a dark circular area. The depressions of temperature by the
eclipse are shown by numerals on the maps, and isotherms are shown by
dotted lines. The weather conditions are indicated by symbols, and the
direction and velocity of the eclipse wind are indicated by the direction
and length of the arrows. The winds were practically reversed in direc-
tion as the umbra moved from one side of the continent to the other,
but both charts show a distinct anticyclonic circulation and an outflow of
air extending from the umbra, or central area of the eclipse, to a distance
of about fifteen hundred or two thousand miles. In the 8.15 a.m. chart
the outer limit of the outflow appears to be in New York, beyond which
there is an inflow. In this chart the stations of observations are so far
iu advance of the central area of the eclipse that no appreciable depres-
sion of temperature is shown ; but in the 9 a.m. chart, which coincides
with the greatest depression of temperature at Wadesboro, Washington,
and Virginia Beach, there is a central area shown by the isotherms where
the depression of temperature exceeds 8° F. This area of greatest cold
lags behind the umbra about five hundred miles.

The charts in Figure 3 show only a portion of the eclipse area, or
penumbra, which was about five thousand miles in diameter. Heuce the
charts do not give an idea of the winds on the outer area of the penum-
bra, or the successive changes which occurred at any one station as the
eclipse passed over it. A view of these changes is obtained by plotting
the winds, temperature, etc., at given stations when they were successively

310

PROCEEDINGS OF THE AMERICAN ACADEMY.

OCLEAR;eFAIR; ©CLOUDY.

I'l-.t 1:1

CLAYTON. — THE ECLIPSE CYCLONE.

811

in different portions of the eclipse area. The eclipse shadow travelled
with a velocity somewhat greater than two thousand miles an hour. By
placing the stations at their proper distances from the path of the umbra
and plotting the successive fifteen minute observations at intervals of about
five hundred miles, a synoptic chart is obtained showing the conditions
observed at any given station or group of stations when they were in dif-
ferent portions of the eclipse area. In this way Figure 4 was constructed.
In this diagram the direction and width of the path of the umbra is

Figure 4.

shown by parallel lines forming a long arrow. The central shaded area
shows the umbra, and the outer unbroken circle shows the outer limit of
the penumbra. The data north of the path of the umbra are derived from
the mean of the observations at Ithaca, Toronto, and Blue Hill ; the data
along the path are derived from the mean of observations at Washington,
Ga., and Wadesboro, N. C. ; the data south of the path are from Havana.

312 PROCEEDINGS OF THE AMERICAN ACADEMY.

Figure 4 indicates distinctly an anticyclonic circulation of the wind
around the centre of the eclipse extending out to a distance of about
fifteen hundred miles from the umbra. Outside this area there is an
equally distinct cyclonic circulation about one thousand miles in width
extending out to and beyond the edge of the penumbra. Beyond this
there are indications of another ring of outflowing winds. The inner
circle of broken lines in the diagram represents a probable ring of low-
air pressure. The outer circle of broken lines surrounding the penumbra
represents a probable ring of high pressure. The isotherms are shown
by dotted lines. They show an elliptical area of cold air central about
five hundred miles in the rear of the umbra. The greatest depression of
temperature is north of the track of the umbra. This was chiefly due
to the continental effect. The difference may also have been due in part
to the fact that the sky was partly cloudy at Havana. On comparing
stations similarly situated as regards the eclipse, it was found that the
depression of temperature due to the eclipse was less at stations where it
was cloudy, and that it also diminished with height above the sea. This
indicates that the cooling is chiefly in a thin stratum of air very near the
earth's surface. The analogy to the diurnal change of temperature
would also indicate that this must be true. The shape and position of
the areas showing the humidity departures are so similar to those of tem-
perature that it is not deemed necessary to reproduce them. The chief
difference is that in one case the departures are plus and in the other
minus. In other words, during the eclipse there is a rise of absolute
and relative humidity and a fall of temperature.

The observations indicate very clearly a lowering of the air pressure
(luring the eclipse, the minimum of pressure occurring soon after the
minimum of air temperature. This is shown by records made at "Wash-
ington, Ga., at Toronto, and at Blue Hill. The accompanying diagram

In. I 1:1 6

shows a record made li\ an " aerograph," or air barometer, al Toronto.
This barograph, devised by F. Napier Denison, has its air-chamber
buried eighl feel below the Burface «»t the ground t" protect it from

CLAYTON. — THE ECLIPSE CYCLONE. 313

external changes of temperature. The curve is traced from the record,
and is given on the same scale without correction in any way. The
eclipse began at Toronto about 7.47 a.m. and ended about 10.18 a.m.
A straight dotted line is drawn through the curve connecting the pres-
sure recorded at the beginning and end of the eclipse. It is seen that
the pressure was generally below the dotted line throughout the eclipse,
but there was an upward swell between 8 and 9 a.m., shortly preceding
the middle of the eclipse. Immediately preceding and following the
beginning and end of the eclipse the curve rises above the dotted line,
indicating a ring of high pressure surrounding the umbra, and thus
agreeing perfectly with the distribution of pressure demanded by the
wind circulation. I find that when the changes of pressure observed
during previous eclipses are separated from the normal diurnal changes,
they show changes very similar to those given by the curve for Toronto,
except that the rise in pressure near the middle of the eclipse is greater
for stations in the path of total eclipse. This central rise of pressure is
due to the increased density of the air from cold, and on it depends the
outflow of air surrounding the umbra. Hence in normal eclipses there
is a central area of relatively high pressure ; surrounding this is a ring of
minimum pressure, and beyond this, outside the edge of the penumbra,
is a ring of maximum pressure.

The low temperature, the circulation of the winds, and the form of
the pressure curve accompanying the eclipse of May 28, 1900, all pro-
claim the development by the eclipse of a cold-air cyclone, the theory
of which has been so well worked out by Ferrel that no better descrip-
tion of it could be given than in his own words. Ferrel maintains from
theoretical considerations that cyclones necessarily have an inner area of
low pressure, surrounded by a ring of high pressure, which Professor
Davis has named a pericyclone. Ferrel further maintains that a cyclone
may have its origin either in a high temperature increasing toward a
central area, or in a low temperature decreasing toward a central area.
The one he calls a cyclone with a warm centre, the other a cyclone with
a cold centre. Of cyclones with a cold centre he says : —

" If for any reason the central part of any given portion of the atmos-
phere of a somewhat circular form is maintained in any way at a lower
temperature than the surrounding parts, and the temperature gradient on
all sides is somewhat symmetrical, we have approximately the conditions
which give rise to a cyclone. In this case it is readily seen that there
must be a vertical circulation, as in the ordinary cyclone, but that it is
reversed, out from the centre below, and in toward the centre above,

314 PROCEEDINGS OF THE AMERICAN ACADEMY.

with a gradual settling down of the air in the interior to supply the out-
ward current beneath. This vertical circulation, as in the case of the
ordinary cyclone, gives rise to a cyclonic motion in the interior and an
anticyclonic in the exterior part of the air under consideration, but in
this case the gyratory velocity is greatest above and is less at lower alti-
tudes, diminishing down to the earth's surface, where it is least. In the
anticyclonic part the reverse takes place, the gyratory velocity being
least above and greatest down near the earth's surface. The distance
from the centre at which the gyratory velocity vanishes and changes
sign, is greatest above and gradually becomes less, with decrease of alti-
tude down to the earth's surface, where it is nearest the centre. . . . The
conditions of a cyclone with a cold centre which are the most nearly
perfect are those furnished by each hemisphere of the globe, as divided
by the equator, in which the pole is the cold centre, and the temperature
gradient from the pole toward the equator is somewhat symmetrical in
all directions from the centre. . . . The easterly motions in the higher
latitudes and the westerly ones in the lower latitudes, in the one case,
correspond to the cyclonic in the interior and the anticyclonic in the
exterior part, and the belt of high pressure near the tropics to that of
high pressure in the case of any cyclone with a cold centre. . . . The
centre of a cyclone with a cold centre may or may not have a minimum
pressure, according to circumstances. A certain amount of temperature
gradient, and of pressure gradient which is independent of the gyratory
motion, as explained in the case of the general circulation of the atmos-
phere, is necessary to overcome the friction in the lower strata and to
keep up the vertical circulation, upon which the cyclone depends ; and
the pressure gradient, which depends upon the temperature gradient and
is independent of the gyrations, may be such that the increase of pressure
in the central part due to this cause may be greater than the decrease of
pressure arising from the cyclonic gyrations, especially where surface
friction is great." *

The eclipse cyclone is of especial interest from a theoretical point of
view, because its origin, clearly connected with the fall of air tempera-
ture attending the eclipse, is Ut^'d from all questions of condensation
of vapor or of the dynamic effects due to the meeting of air currents
whose possible influence complicates the question as to the origin of the

ordinary cyclone. The eclipse may he Compared to an experiment by

Nature in which all the causes thai complicate the origin of the ordinary
• A Popular Treatise on the Winds, pp. 387-389.

CLAYTON. — THE ECLIPSE CYCLONE. 315

cyclone are eliminated except that of a direct and rapid change of tem-
perature. The results derived from the observations by eliminating the
influence of other known phenomena give quantitatively the effects of
a given fall of temperature near the earth's surface in a given time.
They show that a fall of temperature is capable of developing a cold-air
cyclone in an astonishingly short time, with all the peculiar circulation
of winds and distribution of pressure which constitute such a cyclone.
They show, furthermore, that a fall of temperature of the air does not
act primarily to cause an anticyclone but a cyclone, and the anticyclone
is a secondary phenomenon, or rather a part of the cyclone.

The eclipse cyclone shows no apparent lag or dynamic effect due to
the inertia of the air. To keep pace with the eclipse shadow moving
about two thousand miles an hour the eclipse cyclone must continuously
have formed within the shadow and must have dissipated in the rear
almost instantly. In this way its motion may be considered to have a
certain analogy to wave motion. Any given particle of air moving with
the velocity of the eclipse winds could not have moved more than five
miles as a maximum during the passage of the eclipse. Hence all the
changes of pressure must have been derived from the deflective influence
of the earth's rotation acting on air moving this distance.

In brief, the meteorological effects of the eclipse are important —

(1) Because they confirm so fully Ferrel's theory of the cold-air
cyclone ;

(2) Because they show the wonderful rapidity with which cyclonic
phenomena can develop and dissipate in the atmosphere ; and

(3) Because they show that cyclones do not necessarily drift with the
atmosphere, but move with their originating cause, which iu the eclipse
had a progressive velocity of about two thousand miles an hour.

The Diurnal Cyclones.

The discovery that the brief fall of temperature attending a solar
eclipse produces a well-developed cyclone which accompanies the eclipse
shadow at the rate of about two thousand miles an hour, suggests that
the fall of temperature due to the occurrence of night must also produce
or tend to produce a cold-air cyclone. Since the heat of day produces or
tends to produce a warm-air cyclone, there must tend to occur each day
two minima of pressure, one near the coldest part of the day, and an-
other near the warmest part of the day, with areas of high pressure
between them due to the overlapping of the pericyclones surrounding the

31G PROCEEDINGS OF THE AMERICAN ACADEMY.

cold-air and the warm-air cyclones respectively. These causes must
produce entirely or in part the si ell-known double diurnal period in air
pressure. At any rate, in view of the faet that an eclipse causes a cy-
clone over half a hemisphere, it will be necessary before rejecting such
a theory to show that the fall of temperature at night does not product- a
cyclone, or that this cyclone and the corresponding warm-air cyclone of
the day do not appreciably influence the barometer.

The points in favor of the theory that the double diurnal period in
pressure is due to two diurnal cvclones, one developed by the cold of
night and the other by the heat of day, may lie stated in brief as follows.
The theory is based on well-known physical laws. The possibility of a
cold-air cyclone under conditions similar to the diurnal cyclone is con-
firmed by the eclipse cyclone. The theory explains the annual oscilla-
tion of the time of maxima and minima of pressure iu the diurnal period;
and explains the occurrence of a third maximum in high northern lati-
tudes in winter. The theory also explains why the warm-air cyclone is
well developed over continents, and on clear days, and causes a marked
fall in the barometer during the afternoon, while the morning minimum
of pressure over continents does not attain an excessive development as
compared with that over oceans where there is slight retardation of
the air movements on which the (all of the barometer in the cold-air
cyclone depends.

The diurnal cyclones move from east to west, contrary to the motion
of ordinary cyclones in temperate latitudes. Their velocity of motion is
about one thousand miles an hour at the equator, and diminishes toward
the poles. The two charts in Figure <"> indicate the circulation of the
surface winds and upper currents in the diurnal cvclones. In tin se
charts the ordinates represent the hours of the day. and the abscissas
represent distances from the equator. The data for the surface Winds
are derived from observations at Blue Hill. hit. 42 1."-' N., long. 71° 7'
W.. and Cordoba, Argentina, hit. 31° 25' S.. long. 0-1° 1_'' W* The
directions of the arrows represent in the usual way wind directions, and
the position <>f the arrow shows the time of maximum frequency of each

wind. Thus the greatest diurnal frequency of BOUtherly winds occurs at

Cordoba at 7 i.m., and at Blue Hill between 7 and 8 p.m. There is

also a second maximum frequency of BOUtherly winds at Blue Hill about
10A.M. The wind arrow- at Cordoba ami Blue Hill are. in general,

• Annals of tin Astron Observatory of Harvard College, XX3 Pt. iv., 415 and

ir.'.

CLAYTON. — THE ECLIPSE CYCLONE.

317

CJ

— *

o

cj

3

>

«

t

CL
CD

:*

J, X

*
#

i

t

•

rvl

'••1 ^

-Hi

>

J

CO

o

-7 CVJ

-<r

/ >».

2

•

Mi O

1

1

<

u.

D

in CO

3

» #

CO

si

<

VO

*

«t"

•

\ _ i

**-->

i :

<vf

>* #

/ /

1

0
*
*
m

cvl

— >

•2

Of

111

~

— *

o

* .«

,->

00

*

r
»
i
i
r

f

co *

Ld

h-

111 *

It

z *

o
o

O «i

3

> *

c

1

o
1

\

1

1

1
1
/

o -

o

o

CVJ

O

*

•

0

n

O

Z

>' t

ER Wl

5 A.M. 1

; 1

<?x

0_

D

j

i

*

i

cvl

*

CJ

>

uJ

u

_J

_j

O

o

a

0.

I

UJ

o

CL

o

&
O

318 PROCEEDINGS OF THE AMERICAN ACADEMY.

in opposite directions, and distinctly indicate a circulation of the wind
around two cyclonic centres passing along the equator, and an outflow
from high pressures half-way ln-tween them. The lower chart, beaded
• Upper Winds.1' shows the hours of greatest frequency of each wind
direction in the upper air between 2,o00 and 10,000 meters. These
times were determined by observations of clouds at Blur Hill, and from
hourly wind records on the Siintis in Switzerland. Cloud strata at three
different levels between 3,000 and 10,000 meters above Blue Hill each
gave a result similar to the other. Tins is indicated by heavy arrows in
the chart.* The observations on the Siintis at an elevation of 2,500
meters are indicated by the light arrows in the same diagram. There
are no observations available at these heights south of the equator, but
the observations north of the equator indicate a circulation very different
from that at the earth's surface. There is apparent at this height only
one cyclonic and one anticyclonic circulation. The low pressure in the
cold-air cyclone of night persists at these levels, and probably with in-
creased intensity, while the low pressure in the warm-air cyclone of day
has been replaced by a high pressure and an anticyclonic circulation.

* Annals of the Astron. Observatory of Harvard College, XXX. l't. iv., 415 and
419.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 17. — January, 1901.

CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF THE
MASSACHUSETTS INSTITUTE OF TECHNOLOGY.

L.— AN APPARATUS FOR RECORDING ALTERNATING

CURRENT WAVES.

By Frank A. Laws.

With a Plate.

CONTRIBUTIONS FROM THE PHYSICAL LABORATORY OF THE
ASSACHUSETTS INSTITUTE OF TECHNOLOGY.

L. — AN APPARATUS FOR RECORDING ALTERNATING

CURRENT WAVES.

By Fkank A. Laws.

Presented May 10, 1900. Received December 15, 1900.

FiG I

The apparatus forming the subject of this communication was con-
structed at the Rogers Laboratory in 1898, and has proved of sufficient
value to merit a short description. In brief, the arrangement gives us a
modification of the "contact method," by which the record is rendered
continuous and traced photographically.

The necessary electrical connections are shown
in the diagram. Kx and K2 are two rigidly
connected contact wheels of ebonite. Into the
periphery of each wheel are set four brass blocks.
These are accurately placed 90° apart. Upon
each wheel a brush and collector ring give per-
manent contact with the blocks. Another brush
resting on the periphery of the wheel completes

electrical connection as the blocks pass under it. The brushes are so
placed that contact is made and broken at K2 before Kx closes. The
contact wheels are driven by a synchronous motor, which gives one rev-
olution for four complete alternations of the E.M.F. G is a dead-beat
galvanometer, and C is an adjustable condenser. The leads a and b are
carried to the points between which the P.D. is to be investigated. By
inspection of the diagram it will be seen that once on each wave and at a
definite point the condenser C is charged to the potential existing between
a and b. As the charge is determined by the breaking of the contact,
the blocks may be of sufficient width to eliminate the effect of the jumping
of the brushes. Also the resistance at the contact will not be of suffi-
cient magnitude to prevent complete charging of the condenser.

VOL.

xxxvi. — 21

322 PROCEEDINGS OF THE AMERICAN ACADEMY.

The function of Kj is to discharge the condenser through the galva-
nometer after K2 has hroken circuit. The instrument would ordinarily
experience a constant deflection, but Ka and K2 are rigidly connected
and mounted on a radial arm, which is geared to the shaft so that it
moves very slowly. The effect is to gradually move the contact point
over the wave. The deflection of the galvanometer will at any instant
be proportional to the P.D. between a and b at the instant of breaking
at K2, or in other words, the deflection follows the wave form.

The actual arrangement is shown in Figure 2 (see Plate), where the
contact device, the synchronous motor, and the direct current motor used
for starting the apparatus will be seen. By use of worm gearing the
wheel train necessary for moving the brushes is made very compact ; the
reduction for the instrument shown is 7200 to 1.

I have found Sullivan's Universal Galvanometer to be a most satisfac-
tory instrument for use with the apparatus. This galvanometer, of the
D'Arsonval type, has a carefully balanced coil, so that it is not very
susceptible to mechanical disturbances ; also the magnetic damping is
most carefully adjusted. The instrument is not of great sensitiveness,
but owing to the stiff suspension the zero is perfectly definite.

The camera used for recording the curves is shown in Figure 3 (see
Plate). The plate is contained in an ordinary plate holder. This is
moved vertically by a fine wire which is wound on a drum, seen in Fig-
ure 2, just in front of the lower worm-wheel. This drum can be thrown
in at pleasure by a pin clutch. The slide of the plate holder is held sta-
tionary by a pin, so that the plate is exposed as the holder is drawn up.

The front of the camera, shown removed, is provided with a narrow
slit about , ,', ,, of an inch wide. In front of it are projecting lips 9 inches
long and | of an inch apart. They are blackened within and serve effect-
ually to shut out extraneous light, and thus prevent fogging of the plate
The spot of light used was the sharply focussed image of the filament of
an incandescent lamp. An alternative arrangement is to use a plate of
ground glass in the holder, and to have a straight-edge fastened across
the guides. It is then easy to keep the point of a pencil in contact with
it and upon tin- spot of light.

Tin- arrangement described is of course a device for obtaining the
average wave, and unsuitable for recording transient phenomena. Tin'

time taken in recording a wave at l"-'" cycles per second is about 1 '
minutes.

Tin- adjustable condenser allows one to adapt the apparatus to vary-
ing conditions, so the B.M.F. curves may be taken directly, ami the

LAWS. — Recording Alternating Current Waves.

Figure 2.

Figure 3.

LAWS. — RECORDING ALTERNATING CURRENT WAVES.

323

Figure 4. — Potential difference between the carbons of an enclosed arc, in
series with reactive coil.

The curves were taken about one minute apart with a view to testing the con-
cordance of the readings.

Figure 5. — A, potential difference between the terminals of a small alternat-
ing current motor. B, current through motor.

Figure 6. — Curve A is E.M.F., and B is current in one phase of aquarter-
pliase synchronous motor running idle with the field adjusted for minimum current.
Source of power was a three-phase dynamo with phasing transformers arranged
on the Scott system.

current curves by the use of a drop wire, as indicated in Figure 1. In
starting the arrangement it is very easy to determine when the proper

o24 PROCEEDINGS OF THE AMERICAN ACADEMY.

speed for synchronism has been attained by watching the spot of light,
or by listening to a telephone which is inserted in place of the galva-
nometer. "With the latter one hears slow beats as the contact moves over
the wave. Opposite are given some examples of the records obtained
with the device.

ROGEKS L.ARORATORY OF PllYSICS,
J unt, l'JOU.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 18. — January, 1901.

SUGGESTION CONCERNING THE NOMENCLATURE OF

HEAT CAPACITY.

By Theodore William Richards.

SUGGESTION CONCERNING THE NOMENCLATURE OF

HEAT CAPACITY.

By Theodore William Richards.

Received December 19, 1900. Presented January 9, 1901.

The word " calorie " has come to have so many significations as to liave
lost definite meaning, unless qualified by an explanatory phrase. For
this unfortunate condition of affairs, which is due primarily to the varia-
bility of the specific heat of water, the best remedy seems to me to be
the general use of a new standard for the measurement of heat capacity.

The growing tendency to refer all energy measurements to the centi-
meter-gram-second basis makes it fitting that heat also should be meas-
ured directly in these terms. It is not unusual to do this, as far as
heat-energy is concerned ; but the practice is hampered by the fact that
the standards of temperature and heat-capacity have no rational relation
to the so-called absolute units. One calorie equals about 42,000,000
ergs, or 4.2 joules. Would it not be a convenience to arrange the
standards of heat measurement so that the direct product of heat ca-
pacity and change of temperature would be expressed in joules? Nearly
ten years ago Ostwald pointed out some of the advantages of such a
practice,* but the suggestion does not seem to have received the attention
which it deserves.

One way to accomplish the desired result would be to construct a new
scale of temperature with degrees about ^f the size of the present cen-
tigrade degrees, and to retain the specific heat of water at a definite
temperature as the unit of capacity. This course retains one of the disad-
vantages of our present system, and is open to several other objections,
the chief of which would be the variability of the degree with each new
measurement of the value of the mechanical equivalent.

Another obvious course is to retain the centigrade degree, measured
by the hydrogen or helium thermometer, as final, and to choose for the
unit of capacity that capacity which is warmed 1 degree centigrade by

* Zeitschr. phys. Chem., 9, 577 (1892).

328 PROCEEDINGS OF THE AMERICAN ACADEMY.

1 joule (1 watt-second, or 107 ergs). Here the unit of capacity will
vary with each new increase in accuracy in the determination of the
mechanical equivalent; hut capacity is a less tangible dimension than
temperature, and its variation would cause less instrumental confusion.

This convenient unit of capacity would be nearly represented by the
heat capacity of a gram of magnesium at low temperatures (—50°), or
that of a gram of aluminum at high temperatures (about 290° C). At
ordinary temperatures an alloy of zinc and magnesium containing about
■').'> per cent of zinc would probably have the desired capacity.

Specific heats are frequently spoken of in terms of calories, thus con-
founding heat capacity with energy. As a matter of fact, the idea of
specific heat is mathematically nothing but a simple ratio like specific
gravity, — a pure number without physical dimensions. The unit sug-
gested here is rather to be compared with density; it has the definite
dimension of energy divided by temperature.

It seems to me that a name for this unit would greatly assist the
beginner to discriminate between energy and capacity. Would not the
name "mayer," in honor of the unfortunate Julius Robert Mayer, our
of the discoverers of the first law of energy, be a convenient and fitting
term for the centimeter-gram-seeond -f- centigrade unit of heat capacity ?

On this basis the heat capacity of a gram of water at twenty degrees
centigrade is about 4.181 mayers, and that of a gram of liquid mercury
is .0333 X 4.18 = 0.139 mayers. The gas constant becomes 8.32
mayers, if the atomic weight of oxygen is taken as 1 G ; and the Duloug
and Petit constant or gram-atomic heat capacity becomes about 26.5
mayers on the same basis. These numbers are all of convenient magni-
tude. For larger values, such as the heat capacities of solutions used in
thermochemistry, the kilomayer is a convenient unit. For instance, the
capacity of IIC1 -f 100 ILO is 7.41 kilomayers, while that of a similarly
dilute solution of 40 grains (a mol) of sodic hydroxide i- 7. 12 kilomayers.
The solution produced by mixing these two lias a rapacity of 1.3.02 kilo-
mayers. In order to show how convenient these figures are as a basis
of calculation, it i> only necessary to point out that this difference of 0.19
kilomayer between the capacities of factors and product indicates that the
heal of neutralization will vary 0.19 kilojoule* for each degree of tem-

* ( tetwald hat pointed out tin- convi nience <>f the kilojoule as a unit in thermo-
chemistry, in tin- latest edition <>f the "Grundrisa der allgemeinen Chemie." It
seema to me that it would be well to represent this useful unit by kj , in analogy
to km and kg., rather than by J, which might be mistaken as an abbreviation
for joule. Kilomayer may !><■ abbreviated to kmy.

RICHARDS. — NOMENCLATURE OP HEAT CAPACITY. 329

perature, according to the well-known equation CT — CT = —?- ~,

where capacities are represented by C, heats of reaction by U, and
temperatures by T.

Since entropy has the dimensions of heat capacity, it too may be
measured in mayers. This application of the new name may lend
concreteness to an idea which has been to some a stumbling-block.

The greatest sain to be derived from the consistent use of the " abso-
lute" unit of heat capacity is to be found in the field of electrochemistry.
Here even technical men have used for several years the admirable
system of units resting upon the centimeter-gram-second basis. The
increasing use of both the thermodynamic and osmotic equations of
electrochemistry will make the ready application of these units to heat
and gas energy almost a necessity.

Cambridge, Mass., October 31, 1900.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 19.— January, 1901.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

SYMMETRICAL TRIIODBENZOL.

By C. Loring Jackson and G. E. Beiir.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

SYMMETRICAL TRIIODBENZOL.

By C. Loking Jackson and G. E. Behr.

Presented December 12, 1900. Received December 20, 1900.

So far as we can find, three triiodbenzols have been described in the
chemical literature, but the constitution of only one has been established
thoroughly ; this is the un symmetrical (1.2.4) isomere melting at 76°
obtained by Kekule * from the action of iodine and iodic acid on benzol.
More recently Hantzsch f has also made it from the diioddiazobeuzol
iodide prepared from diiodaniliue and found that it melts at 77°.

The other two isomeric triiodbenzols were made by Istrati and
Georgescu t by heating together benzol iodine and concentrated sulphuric
acid. Unfortunately their original paper is not at our disposal, so that we
know their work only from the abstracts of it in the " Jahresbericht liber
die Fortschritte der Chemie " and Beilstein's Handbuch (third edition).
From these it appears that they ascribed the unsymmetrical constitution
(1 . 2 . 4) to their compound melting at 85°, and the adjacent structure
(1.2.3) to that melting at 182° to 184°, but there is no statement that
they determined these constitutions experimentally. The experiments
of Kekule and Hantzsch just cited prove that the unsymmetrical isomere
melts at 77° (or 76°), and, therefore, their compound melting at 85°
cannot have this structure.

Under these circumstances it seemed worth while to prepare the symmet-
rical triiodbenzol and to prove its constitution by experiment ; accordingly
we made the triiodaniline by the method of Michael and Norton, § that
is, treating chloride of aniline with monochloride of iodine ; and replaced
the amido group in this body by hydrogen by means of the diazo reaction.

* Ann. Chem. (Liebig), CXXXVII. 164 (1866).
t Ber. d. chem. Ges., XXVIII. 684 (1895).

t Buletinul d. Soc. d. Sciinte Fiz. d. Bucuresci, I. 62. Jarhresb. Chem. 1892,
1063. Beilstein's Handbucli (3d Edition), III. 73.
§ Ber. d. chem. Gee., XI. Ill (1878).

334 PROCEEDINGS OP THE AMERICAN ACADEMY.

The triiodbenzol thus obtained melted at 181°, which suggested that it
was identical with the substance made by [strati and Georgescu melting
at 182 to 184°, and this was proved to be the case by treating it with
fuming uitrie acid, which converted it into a triioddinitrobenzol melting
at 210 , whereas [strati and Georgescu* under similar conditions ob-
tained from their substance a dinitro derivative melting at 210°-212°.
That our triiodbenzol had the symmetrical constitution (1.3. 5) was
very probable, because the trichloraniline and the tribromaniline made
under similar conditions are symmetrica] compounds; but, as we felt it
was necessary to give an absolute proof of its structure, we treated the
triioddinitrobenzol witli aniline, when it formed a trianilidodinitrobenzol
melting at LSI . The symmetrical trianilidodinitrobeuzol made by Palmer
and one of us f from symmetrical tribromdinitrobenzol melted at 179°;
in spite of the difference of two degrees in these melting points there can
be no doubt of the identity of the compounds, and it follows, therefore,
that the triiodbenzol melting at 181° has the constitution I8 1 . 3 . 5.
This would leave only the structure 1.2. -'! for Istrati and (ieorire>eu's
triiodbenzol melting at 85°; but in view of the fact that this compound
was isolated from a mixture of the two triiodbeuzols and two tetraiod-
benzols, it seems to us possible that it may be the unsymmetrical isomere
(1.2. h (uniting at 77°) contaminated with some of these less fusible
substances. We had hoped to prepare several derivatives of the triiod-
benzol, but were unable to do so in the time at our disposal, and, as
we cannot continue our work together, think it well to publish this
account of the results so far obtained.

Preparation op Triiodanilink.

The triiodaniline was prepared by the method of Michael and Norton t
slightly modified. Ten grams of aniline were dissolved in dilute hydro-
chloric acid, and diluted with water at 15° or a higher temperature to
the volume of seven litres. This solution was then treated with a rapid
stream of air charged with the vapor of monochloride of iodine by pass-
ing it througha flask containing somewhal more than the required amount
of this siih-tanee heated to aboul 60 by immersion in warm water.

Alter thirty or forty minute- the action had come to an end, and the

liquid was filter, d a- quickly a- possible to remove a heavy black precipi-

• Bui Soc s,i. Fiz . i 66

I Tin se Proceedings, \ XIV in.

t Ber. ■! .•hem. Gea , XI. Ill

JACKSON AND BEHR. SYMMETRICAL TRIIODBENZOL. 335

tate which had formed; the filtrate was allowed to stand for several
hours, when it deposited a flocculent buff-colored precipitate, the weight
of which varied from 10 to 20 grams in different operations ; that is, 20 to
40 per cent of the theoretical yield. An additional amount of the triiod-
aniline could undoubtedly have been obtained from the black precipitate,
which was filtered out at first, but this was so impure that it seemed
more time would be lost in purifying it than in preparing fresh quanti-
ties of the second buff-colored precipitate, which without further treat-
ment was pure enough for the manufacture of the triiodbenzol. That it
was the triiodaniline was shown by an analysis of the substance purified
by crystallization from glacial acetic acid and alcohol with the aid of
boneblack* which gave 81.25 per cent of iodine instead of the 80.87 per
cent required by the formula. Its melting point was 185°. Michael and
Norton give 1 85|° .

The monochloride of iodine used in this work was made by the action
of chlorine on iodine according to one of the methods given by Hannay,*
and used later by Bornemann.f Bunsen's $ method, which consists in
boiling iodine with aqua regia, gave a less good result. Distillation of
iodine with potassic chlorate, recommended by Schutzenberger § and
Hannay,|| was not tried. Forty-two grams of powdered iodine (the amount
needed for ten grams of aniline) were treated with dry chlorine in a flask,
until they had gained 12 grams; in addition to the reddish brown liquid
monochloride a considerable amount of the brilliant yellow crystalline
trichloride of iodine was formed, which was converted into the mono-
chloride by adding a small excess of iodine and heating the flask gently
on the steam bath under a short air condenser, until the yellow crystals
had disappeared. The monochloride of iodine solidified in fine gray
crystals, if cooled below 10°, and in this solid state could be kept for
some time without decomposition. Its tendency to solidify made it
necessary in the preparation of triiodaniline to use the solution of chloride
of aniline at a temperature of at least 15° in order to avoid the danger
of having the tube stopped up with monochloride of iodine. Although
this substance melts at 24°. 7 (Hannay), we found no trouble from solidi-
fication with a solution at 15°, which can be explained by the marked
tendency of the monochloride to remain liquid after it has been melted.

* Proc. Lond. Chem. Soc., XXVI. 815 (1873).

t Ann. Chem. (Liebig), CLXXXIX. 184 (1877).

t Gmelin-Kraut's Handbuch, I. 2, p. 416.

§ Zeitschr. d. Chem., VI. 1.

II Proc. Lond. Chem. Soc, XXVI. 815 (1873).

33G PROCEEDINGS OF THE AMERICAN ACADEMY.

Symmetrical Triiodbenzol, C6II3I3.

This substance was prepared as follows: Ten grams of the buff-
colored triiodauilinc, made as described in the previous section, were
boiled with 125 c.c, of benzol and 25 c.c. of alcohol, and disregarding
the undissolved portion, 5 c.c. of commercial sulphuric acid were added.
and then 5 grams of finely powdered sodic nitrite were sifted into the
liquid as quickly as possible. As soon a- the evolution of nitrogen had
slackened sufficiently, the liquid was boiled for some time on the steam
hath, until a large part of the solvents had jtassed off; during this boil-
ing a bright yellow solid, which formed on the particles of sodic nitrite
as they eutered the liquid, changed its color to a light grayish brown.
After the solution had been boiled, it was allowed to stand over night,
and then the deposit was filtered out and washed, first with alcohol and
afterwards with hot water, to remove the inorganic salts. The amount
of this crude product rarely fell below 50 per cent of the theoretical yield.

The product, obtained as described above, was next sublimed from a
large watch-glass heated on the sand bath, and covered with a funnel
which stood on a piece of filter paper above the substance. If this sub-
limation was carried on slowly enough, as much as so p. r cent of white
glistening crystals was obtained, but if it was urged too fast, the impuri-
ties also sublimed, and the crystals were yellow, or even in extreme
cases light brown and sticky. When proper care was used, white
crystals were obtained by this sublimation, even from verj impure
products, such as those recovered by evaporating the liquid portion of the
product of the diazo reaction. The white sublimed crystals were not,
however, pure, and to remove from them a persistent impurity we found
repeated recrystallizations from alcohol were necessary, which finally
raised the melting point to 181°, where it remained constant. The bud-
stance was then dried at 100°, and anal\ zed with the following results : —

I. 0.1187 gram of the substance gave bj the method of Carius 0.1838

gram of argentic iodide.
II. 0.1453 gram gave 0.2249 gram of argentic iodide.

Calculated f<>r 1 oand,

I H i I. II.

Iodine ,l 83.f»C 83.G0

Properties of 1 .3.5 Triiodbenzol. — This substance crystallizes from
alcohol at first in radiating groups composed of a few slender prisms,
which develop later into long white prisms frequently tapering, but

JACKSON AND BEHR. — SYMMETRICAL TRIIODBENZOL. 337

always terminated by a single plane at an oblique angle to the sides of
the prism. It melts at 181°, and sublimes easily. It is freely soluble in
benzol, or carbonic disulphide even in the cold ; in chloroform it is
moderately soluble in the cold, freely soluble when hot ; in ether it is
moderately soluble, whether hot or cold ; in glacial acetic acid or ethyl
acetate it is somewhat soluble in the cold, freely soluble when hot ; in
alcohol or acetone it is slightly soluble in the cold, moderately soluble
when hot ; it is apparently insoluble in water, hot or cold. The best
solvent for it is alcohol. Strong hydrochloric acid has no apparent
action on it, even when hot ; strong nitric acid also has little or no
action on it, but fuming nitric acid acts on it in the way described in the
next paragraph. Strong sulphuric acid, when heated with it to its melt-
ing point, causes a partial decomposition, taking on a dark color ; at higher
temperatures the triiodbenzol sublimes out of the mixture. A strong
solution of sodic hvdrate seems to have no action on it, even when
boiling.

An attempt was made to prepare triiodnitrobeuzol ; for this purpose
five grams of symmetrical triiodbenzol were boiled with 140 c.c. of a
mixture consisting of four parts of fuming nitric acid with one part of
common nitric acid. On cooling light yellow crystals appeared, and an
additional amount of the product was obtained by pouring the acid liquid
into about a litre of water, when a pale yellow flocculent precipitate was
formed, which was filtered out, after it had stood some time, and with
the crystals weighed 5.3 grams. This was purified by crystallization
from a mixture of four parts of alcohol with one of water, until it showed
the constant melting point 210°. This indicates that the substance is
triioddinitrobenzol, since Istrati and Georgescu * obtained a triioddini-
trobenzol melting at 210o-212° from their triiodbenzol melting at
182°-184°. For still greater certainty the substance was dried at 100°,
and analyzed with the following result : —

0.2049 gram of the substance gave by the method of Carius 0.2G44
gram of argentic iodide.

Calculated for CBHI3 (NO.,),. Found.

Iodine 69.77 69.76

The substance, therefore, is a dinitro compound, although obtained
from fuming nitric acid somewhat diluted with common nitric acid. We

* Bui. Soc. Sci. Fiz., I. 66.
vol. xxxvi. — 22

338 PROCEEDINGS OF THE AMERICAN ACADEMY.

did not have time to try whether the nionouitro compound could be
obtained with a still more dilute nitric acid.

The agreement between the nicking points of our triiodbenzol, 181°,
and triioddinitrobenzol, 210 . with those obtained by Istrati and
1 ■ .rgescu, * 182° L8 I and 210 --J12°, establishes the identity of these
substances beyond a doubt. The triioddinitrobenzol also gives us the means
of proving their constitution. For this purpose 0.5 gram of the triioddi-
nitrobenzol were heated with 0.5 gram of aniline; this proportion gives
about six molecules of aniline to each molecule of the dinitro compound.
The beating was carried on for two hours on the steam bath, and the
product, a dark red solution, was treed from the excess of aniline by
washing with dilute hydrochloric acid, when it Formed a dark red sticky
mass, which was purified by working it well with a rod under dilute
hydrochloric acid and then crystallizing it from alcohol. It showed the
constant melting point 181°, which agrees sufficiently well with 179 .
that given by Palmer and one of usf for the trianilidodinitrobenzol
made from the tribiouidinitrobenzol Br. 1.3.5. (N02)9 2 . 4. This
proves, therefore, that the triiodbenzol melting at 181° has the iodine in
the position 1 . 3 . 5, as would have been inferred from its preparation
from triiodaniline, which according to the analogy of the chlorine and
bromine compounds should have the symmetrical constitution.

* Bui. Soc. Sci. Fiz., I. 62, G6.
t These Proceedings, XXIV. 111.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 20. — March, 1901.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

A STUDY OF GROWING CRYSTALS BT INSTANTANEOUS

PHO TOMICR 0 GRAPH Y.

By Theodore William Richards and Ebenezer Henry Archibald.

With Three Plates.

Investigations on Light and Heat, made and published wholly or m part with Appropriations

FROM THE RUMFORD FOND.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF

HARVARD COLLEGE.

A STUDY OF GROWING CRYSTALS BY INSTAN-
TANEOUS PHOTOMICROGRAPHY.

By Theodore William Richards and Ebenezer Henry Archibald.

Received January 7, 1901. Presented January 9, 1901.

Countless observers have watched the growth of crystals under the
microscope. As long ago as 1839 attempts were made to study also the
birth of crystals, in order to determine in what mariner the new phase
makes its entrance into the system. With a microscope magnifying 600
diameters, Link * thought he could detect the formation of minute glob-
ules at the moment of precipitation, — globules which soon joined and
assumed crystalline form. Schmidt, f Frankenheim, $ and especially
Vogelsang, § made similar observations some years later, and several
more recent accounts of this phenomenon have appeared. Modern investi-
gators have been more concerned with the speed of separation from super-
saturated or supercooled liquids than with the form of the first separation.!

Ostwald, in 189 1, accepted the interpretation of these data, which as-
sumes that crystallization is always preceded by the separation of an
initially liquid phase, consisting of a supersaturated solution of the former
solvent in its former solute.**

This explanation is indeed a plausible one, and undoubtedly holds true
in cases like those studied by Schmidt and Vogelsang, where a sub-
stance separates at a temperature not far below its melting point, and
often where a substance soluble in one liquid is precipitated by the

* Link, Pogg. Ann., 46, '258 (1839).

t Schmidt, Lieb. Ann., 53, 171 (1845).

| Frankenheim, Pogg. Ann., Ill, 1 (1860).

§ Vogelsang, Die Krystalliten (Bonn, 1875). See Lehmann, Molecularphysik,
I. p. 730 (1888).

|| Gernez, Compt. Rend., 95, 1278 (1882); Moore, Zeits. phys. Chem., 12, 545
(1893) ; Friedliinder and Tammann, ibid., 24, 152 (1897) ; Tammann, ibid., 25, 441 ;
26, 307, 367, 28, 96; Kiister, ibid., 25, 480, 27, 222; Bogajavlensky, ibid., 27, 585.

** Lehrbuch, I. 1039 (1891).

342 PROCEEDINGS OF THE AMEHICAN ACADEMY.

addition of a consolute liquid in which the substance is insoluble. For
examples, phenol always separates from aqueous solution in the form of
a liquid, and mangauous sulphate forms at first two liquid phases when
alcohol is added to its aqueous solution. On the other hand, the separa-
tion of a high-melting salt like baric chloride from its solution in pure
water is much less likely to take place in this way. The admixture of
water necessary to lower a melting point from 900° to 25° would be so
large as to make the new phase, a solution of water in baric chloride,
supersaturated to an improbable extent. It is hard to guess where the
line between probability and improbability should be drawn.

Ostwald has shown that an exceedingly small particle of solid is capa-
ble of starting crystallization,* — a fact which may not be wholly foreign
to the present discussion.

In any case, the matter seemed worthy of further experimenting. Ost-
wald says, "Die erste Bihlung der Krystalle liisst sich bei Salzlosungen
and dergleichen microscopisch nicht verfolgen, weil gewohnlich im Ge-
sichtsfelde an einer bislang gleichformigen Stelle plotzlich ein Krystall-
chen erscheint." While this is true as far as the human eye is concerned,
instantaneous photography, an art unknown in Link's time, seemed
peculiarly fitted for the unprejudiced recording of the circumstances at-
tending the genesis of crystals. An attempt in this direction is described
below.

The problem resolved itself into the taking of a number of successive
instantaneous microphotographs of a suitable mixture at the point of crys-
tallization. This problem presented some difficulties, however. In order
to secure a sufficiently brief exposure, very great illumination is needed.
The greater the magnifying power of the lenses of the microscope-camera,
the more intense must be the source of light. The difficulty is increased
by the fact that most crystals are so transparent as to absorb but little
light, and reflection is possible only in certain directions. Hence it is
hard to obtain a distinct image even in a strong liiiht. Moreover, the
machinery necessary for shifting the plates must be so frictionless in con-
struction, and so firmly fixed, as to impart no vibration to the camera or
the mobile subject of study.

These difficulties were at least partially overcome by two different ar-
rangements, the first of which caused the successive impression of a brighl
image in a dark field, and the second registered dark images in a succes-
sion of bright fields. Obviously the former was the more economical as

* Ostwald, Zeits. pbys. Chi m . 22. 289 (1897).

RICHARDS AND ARCHIBALD. — GROWING CRYSTALS. 343

regards expenditure of sensitized film, and the more simple in execution ;
for when the field is dark, successive images can be obtained by a very
slight motion of either object or film, while, when the field is light, the
whole previously exposed surface must be replaced by a fresh surface
before each exposure.

The apparatus consisted of a good compound microscope fitted above
with a vertical foldiug camera, which was supported by two massive steel
pillars on the heavy stand. It was, in short, the regular photomicro-
graphic outfit made by Bausch and Lomb. Between the microscope and
camera, in a suitable light-tight box, was placed a revolving shutter, which
allowed an exposure equal to one fifth of the time of its revolution. Thus,
when the shutter made two revolutions in a second, the exposure was one
tenth of a second. A Henrici hot air motor, combined with speed-reduc-
ing double pulleys, enabled the experimenter to use any rate of revolution
desired. The rate was reasonably constant, but no attempt was made to
make it absolutely so. The sensitive plate or gelatine film was held above
in a suitable holder, which was put in the place of the ground-glass plate
used for focusing just before each series of exposures.

In carrvincr out the first of the two methods it was found more conven-
ient to move the crystallizing solution than to move the photographic
plate. For this purpose the slide bearing the drop of liquid was attached
by a wire to a point just below the centre of a segment provided above
with saw-teeth. The segment was moved gradually by the oscillating
motion of a connecting-rod, fastened by a crank to the revolving shutter
at one end and playing into the saw-teeth on the other. In order to
make the motion certain, the stroke of the connecting-rod slightly ex-
ceeded the distance between the saw-teeth. The segment was suspended
in such a way that its centre of gravity coincided with its point of sup-
port, and the friction of its bearings was so adjusted that it would move
easily, and yet remain stationary during the return stroke. The distance
through which the observed object was moved was easily varied bv alter-
ing the relative lengths of the lever-arms ; distances varying from one
tenth to one fiftieth of a millimeter were usually used. The shutter was
so arranged that during the exposure the segment and slide were at rest,
the shift in position being effected during the four fifths of the revolution
through which the shutter was closed. The accompanying diagram will
make the arrangement clearer.

344

PROCEEDINGS OF THE AMERICAN ACADEMY.

FlGUBE 1. Diagram of PnoTOGRAPnic APPARATUS. (', natural size)

A, sensitive plate or film holder
I'., box containing Bhutter.

C, pulley attached to axle of Bhutter to communicate power from motor.
I), light r- »<l moved by crank attached to Bameaxle; D is guided by a stout
support in which it moves loosely.

E, segment provided with ratchet-teeth; moved gradually by rod D.

F. microsc >pe

'. slide for object, moved by wire running to II

II, holes to n amplitude of object's motioa

I, weight, balancing segment.

J, horizontal projection of revolving shutter in detail.

Tin- diagram represents the apparatus an Instant before an exposure begins

RICHARDS AND ARCHIBALD. — GROWING CRYSTALS. 345

As a source of light any ordinary combination of incandescent electric
lights proved to be inadequate. A good Auer von Welsbach light with
a powerful reflector was more satisfactory, but the best results were ob-
tained with the help of sunlight directed by a suitably arranged mirror
and condensed by reflectors and lenses. The chief, though not serious,
difficulty of this arrangement was the great heat caused by the converging
rays, a difficulty which was obviated partially by an absorbent screen in
some later experiments.*

The first photographs were taken by reflected light, the drop of solu-
tion being placed upon a ruby-colored slide. As soon as the crystalliza-
tion had begun upon one edge of this drop, the very sensitive plate was
uncovered and the shutter and segment were set in motion. The expos-
ure was stopped after fifteen or twenty revolutions, so as to avoid confus-
ing superpositions. Even with the strongest light the images were very
faint and unsatisfactory; it is not worth the space to reproduce them
here.

Another mode of obtaining light images on a dark ground, applicable
to all except the isometric system of crystals, is the use of polarized light.f
A Nicol prism was placed in the barrel of the microscope, and another
just below the stage. The main body of the light was thus intercepted
by the crossed prisms, and only that which had been deflected by the
crystalline structure was allowed to emerge. It is true that this method
could not in all probability decide the chief point at issue ; for the pre-
natal globular condition of crystals would probably have no effect on
polarized light. Definite optical structure is of course necessary to pro-
duce the required deflection of the plane of polarization, and such definite
structure might not be possessed by the globules. Nevertheless, the idea
seemed well worth a trial.

The images were now much more clearly defined and striking, and with
a magnification of 30 diameters, ten sharp impressions, each exposed 5^
second, could be obtained in a second. For this low power the eyepiece
was removed from the microscope and an objective with long focal dis-
tance alone was used to give the image. The degree of enlargement was
obtained by actually measuring the image of a micrometer scale divided
into j1^ millimeters. The rapidity of exposure was so great that many
plates were sacrificed, for it was difficult to find the precise moment when
nascent crystals were in the field of view. In most cases the crystalliza-

* Ilutchins has shown that pure water is as good as a solution of alum for this
purpose (Am. J. Sci., 143, 526 [1892]).

t This suggestion was kindly made by Professor E. C. Pickering.

34G PROCEEDINGS OF THE AMERICAN ACADEMY.

tion was already well started wheu the exposure began, as in Figure 4;
but in some nothing but blank negatives were obtained. The best method
is so to arrange circumstances as to have the crystallization begin upon one
edge and spread slowly over the drop. Another difficulty was the attain-
ing of the exact actinic focus, which differed slightly from the visual
focus. It was found that a definite fraction of a revolution in the line
adjustment of the instrument could be relied upon to cover this difference,
wheu experiment had once found the right spot.

Among other substances sodic nitrate, baric chloride, cupric sulphate,
and ferrous amnionic sulphate were found to give satisfactory results. A
few photographs chosen as being typical examples of many negatives are
given here. (Plate I. Figs. 2, o, 4, 5.)

All the images recorded on these plates are perfectly sharp and regular
when in focus ; but the magnifying power was too low to give important
evidence concerning the birth of the crystals. The crystals always lirst
appear as points, indicating a diameter of less than ^^^ millimeter. The
regularity of growth of those already well started is worth a passing
mention.

The next objective used gave a magnification of 110 diameters. With
this power the light was so much diminished that exposures of less than
^ second became too pale. Three examples from among these negatives
are given below. It will be noticed that in all cases the crystals have
their regular forms when they first appear upon the plate. Another
point worthy of attention is the fact that the growth in diameter at first
is more rapid than it seems to be subsequently. This rapid growth of
small particles has already been noticed by Ostwald;* it is treated more
fully in the following pages. The crystals of sodic nitrate grew faster
than those of baric chloride or cupric sulphate, and the two latter sub-
stances evidently appeared at first in very thin plates. It is interesting
to note that the thickening of these plates caused a corresponding change
in the quality of the emerging light, and hence the crystal-images show a
rhythm of dark and light. I Plate I. Figs. 6 and 7, Plate II. Fig. 9.)

At this point the whole method of procedure was changed on account
of the probability that a globular condition, if it existed at all, would not
lie visible through the crossed Nieols. The apparatus was now arranged
for the exposure of successive portions of a film to unpolarized sunlight.
emanating from a bright field, upon which the crystals appeared as dark

Bpots. The slide and crystallizing solution w<re allowed to remain sta-
• Ostwald, Zeita phys Chein . 22, !

ETCHARDS AND ARCHIBALD. — GROWING CRYSTALS. 347

tionary, and the gelatine film was moved, as it is in the common film-
cameras. The 2^ inch Eastman cartridge film was found to answer the
purpose. At first the turning was effected by an automatic electro-mag-
netic arrangement which received its current from a make-and-break
contact attached to the shutter. Since a current of ten amperes was
needed to secure a sufficiently forcible and speedy action, the operation
of this device was somewhat troublesome, and when the exposures were
not much more frequent than one a second the film was rolled by hand.
A suitable signal attached to the shutter axle, which was still turned by
the Henrici motor, gave the necessary indication of the proper moment
for renewing the sensitive surface. With this apparatus it was of course
possible to obtain photographs of isometric crystals, which could not be
examined with the preceding arrangement.

At first a power of 100 diameters was employed, and very satisfactory
pictures of the growth of crystals of potassic iodide were obtained. One
of these negatives is reproduced here as an example. (Plate II. Fig. 9.)
Thev showed nothing new, however; hence a much higher power of
580 diameters was applied by combining a "2 inch" eyepiece with an
"■^ inch" objective. With this contrivance the light was of course far
less intense, and the definition less sharp. Even with the brightest sun-
light, concentrated by mirrors and an Abbe condenser, the exposure could
not profitably be made less than | second. These plates have been
enlarged by usual processes to over seven times their original size, so
that a total enlargement of over four thousand diameters has been at-
tained. Since these larger images are not much more clear than the
smaller ones, while they occupy much more space, the plates herewith
given are all from the original negatives.

A number of good impressions of crystallizing potassic iodide were taken
under these circumstances, but many other rolls were sacrificed. The
chief difficulty, as before, was to secure the right moment ; and this diffi-
culty was of course much augmented by the limited expanse of the field.
Prints from a few of the successful negatives are given below. In order
to give a clear impression in printer's ink, these were much intensified by
successive photographic printing and intensification ; but of course no at-
tempt was made to remove the imperfections of the successive plates, for
which allowance may easily be made. (Plate III. Figs. 10-15.)

The study of these photographs reveals several interesting points. In
the first place, it is noticeable that no image is wholly without evidence
of crystalline structure. The most doubtful cases are those in Figures 9
and 11 ; but the elongated shape of these doubtful images seems to

348 PROCEEDINGS OP THE AMERICAN ACADEMY.

indicate a solid. A globule of a new liquid phase, ,,,',,,-, millimeter in
diameter, would have left an unmistakably circular image on the highly
magnified plate, for its index of refraction could not have been identical
with that of the aqueous solution. The fact that we could not find such
a globule of course does not prove that a globule cannot exist, either for
an infinitesimally brief period of time, or of an infiuitesimal magnitude
beyond the reach of microscopic observation. Nevertheless, so many
BCOres of photographs were taken as to diminish considerably the proba-
bility that Buch globules can ever be seen with substances possessing a
high melting point.

A striking fact to be noticed in nearly all the most highly magnified
records is the ill-defined appearance of the smallest crystals. This ap-
pears to be due, not to a lack of structure, but rather to the rapid growth
in diameter which is manifest in the young crystal. The initial rapidity
is so great that the fifth of a second appears to include several different
stages oi growth, and hence a blurred impression results. It is easy to
obtain some idea of the rapidity of this initial growth by comparing the
sizes of the first two or three appearances of each crystal.

With this object a few of the series were measured by means of an
accurate micrometer; but the conditions of the experiments are too un-
certain to give the very precise measurements much value. Perfect
constancy of temperature and evaporation, as well as in the rate of the
revolving shutter, involving grave complications in the apparatus, should
of course be maintained if great accuracy is attempted. Measurements
with a fine millimeter scale afford all the precision which it is worth
while to attain under present conditions. A typical case (Figure 9, largest
crystal) gave the following measurements in millimeters for the successive
longest diameters : 2.0; 2.6; 3.0; 3.2; 3.3; 3.5; 3.7; 3.!); 4.1; 4.2;
4.4. The actual sizes of the crystals were of course only a hundredth of
these measurements, since the enlargement was loo diameters.

In spite of the inexact nature of such measurements, it is possible to
use them as a means of defining approximately the law regulating the
changes in speed. The following table details six series of corresponding
diametric measurements of six crystals taken at random. The measure-
ments were taken directly from the photographs, in millimeters.

RICHARDS AND ARCHIBALD.

GROWING CRYSTALS.

349

Diameters of Successive Images.

First appearance.
Second appearance.
Third appearance.

Crystal 1.

Crystal 2.

Crystal 3.

Crystal 4.

Crystal 5.

Crystal 6.

1.0

1.6
1.7

2.0
2.6
3.0

3.0
5.7

6.8

3.0
6.0
7.0

4.0
6.7
7.8

2.5
3.5
4.1

These all show greater growth in the first interval than in the second.
In order to reduce them to one standard, the diameter of the third ap-
pearance was taken in each case as unity. The table then becomes : —

Diameters of Successive Images.

First appearance.

Crystal
1.

Crystal
2.

Crystal
"3.

Crystal
4.

Crystal
'5.

Crystal
6.

Average.

0.59

0.67

0.44

0.43

0.51

0.61

0.54

Second appearance.

0.94

0.87

0.84

0.86

0.86

0.85

0.87

Third appearance.

1.00

1.00

1.00

1.00

1.00

1.00

1.00

At the time of the first appearance, the average age of the crystal
must be about half the time intervening between two exposures ; for the
crystal must have been formed since the last exposure, and it is as likely
to come near the beginning as near the end of the interval. Thus in
Figure 11 the crystals evidently started immediately after the previous
exposure, while in Figure 9 they were registered while still very young.
The averaging of a much larger number of observed diameters led to the
slightly different values given below, corresponding to the accompanying
times : —

Time = t.

Diameter = D.

0 interval

Diameter 0

0.50 interval

Diameter 0.57

1.50 intervals

Diameter 0.87

2.50 intervals

Diameter 1.00

350

PROCEEDINGS OF Till-: AMERICAN ACADEMY.

These data are plotted in Figure 16.

l 2
Time from birth of crystal

Figure 16. Diagram Representing Rate <»k Increase in Crystal

1 >i \ mi ri:i:.

] represents point from average of crystal* taken at random.
O represents point from average <>f selected crj Btals

represents probable actual rate of growth.

represents equation I >'• let.

represents equation /'■' - kt.

The unit of time is the time of one revolution of the Bhutter, or 1 ,25 Beconds.
The Bubstance w m - potassic iodide.

The inspection of the figure shows al once thai the curve is similar in
eral shape to one represented l»v the equation P" li. where l> h

RICHARDS AND ARCHIBALD. — GROWING CRYSTALS. 351

the diameter of the crystal, t the time from the birth of the crystal, and
k a constant. The only question is as to the magnitude of n. The
curves which result from the assumptions n = 2 and n = 3 are given
above, for comparison with the experimental curve. It is clear that the
curve with the latter value, n = 3, is the nearest, possessing the same gen-
eral curvature, and deviating from the average less than the individual
measurements do. This is equivalent to saying that equal increments of
lime correspond to equal increments of volume, instead of equal increments
of surface, as one might have supposed. Of course a law based upon such
merely approximate data cannot be considered as definitely settled ; but
clearly the general character or tendency of the curve is established. It
is probable that under the necessarily ill-defined conditions of our experi-
ments the growth follows no one law with precision ; supersaturatiou,
convection, diffusion, and evaporation must all influence the result. The
crystal which seems to have deviated most widely from the average is
that depicted in Figure 10; this crystal grew at first less rapidly than
usual, and finally came almost to a standstill. It is possible that an in-
creasing solubility due to increasing temperature may have caused this
delaying tendency

It is interesting to compare this average, calculated on the assumption
that the crystal starts in the middle of the dark interval, with a few sin-
gle cases which appear to have begun to crystallize very near the begin-
ning or end of the interval. In these cases, the first image of the crystal
will appear either almost as large as the second image, or very small
compared with it. It will appear almost as large as the second image
when the preceding exposure has just not caught the beginning of the
crystal, which has thus had a whole interval for growth ; or very much
smaller than the second image when the first impression has registered a
crystal only a very small fraction of a second old. Marked examples of
the former case are to be found in Figure 11, and of the latter in the largest
crystal in Figure 9, and the smallest crystal in Figure 15. The times of
revolution represented by Figures 9 and 11 are the same, 1.25 seconds, and
the other conditions also were identical, hence we may compare these with
accuracy. Careful measurements of the sizes in Figure 9 showed the first
large impression of the crystals to be about eighty per cent of the diam-
eter of the next impression, and approximately the same relationship ap-
pears in Figure 11. In order to find if this relationship corresponds with
the equation D3 = Id, the larger diameter is assumed to be 0.93, the the-
oretical value corresponding to two intervals of time, if that corresponding
to two and one half intervals is taken as unity. Hence the smaller one

352 PROCEEDINGS OF THE AMERICAN ACADEMY.

becomes 0.75, corresponding to one interval of time, a value, marked in
a circle on t lie diagram, which is Burprisingly near the cubic curve.
Hence the equation D3 = kt is confirmed. That the same curve holds
approximately for the further growth of the crystal is manifest by a
quantitative Btudy of Figure 9 (Plate II.).

In this connection it is interesting to note that the crystal seems often
to grow at first in the Bame proportion in all directions. Even the very
minute image in the centre of the second exposure, given in Figure 9, shows
itself under the microscope to he elongated like the crystal which grows
from it. In the next exposure this crystal had the proportions 0.02 mm.
X 0.012o mm., and after four more exposures it still had almost exactly
the same proportions, being 0.035 mm. X 0.022 mm. After two or three
more seconds the form given in Figure 9 began to change slightly, the
crystal becoming slightly less elongated in shape ; but by this time the
neighboring crystals had grown so much as to approach it, and hence
to alter the conditions. A similar constancy in proportion may be
observed in many other series here given.

The diagram shows how exceedingly fast the diametric growth of the
crystal must be in the first tenth of a second of its existence. Hence we
have an explanation for the suddenness of its appearance to the eye of
an observer, and for the blurred edges of its photographic image. It is
true that another cause may contribute to the blurred effect; namely, the
irregular refraction caused by the convection of the lighter solution which
has just deposited part of its load; hut the speedy growth alone is capa-
ble of explaining the observed indistinctness.

Interesting as the rapid initial growth in diameter may be, it places a
serious bar in the way of more precise study of the birth of crystals. One
clearly needs not only high magnifying power, but also great speed; and
these two together require very intense light. Whether or not we shall
be able to obtain more positive knowledge with the present apparatus, is
a questionable matter. The same phenomenon casts a measure of doubt
over some of tli<- observations of Link ami his followers. Is it not possi-
ble that the subjective effect of the rapid ly growing crystal might be mis-
taken for that of a globule of liquid? Even upon the photographic plate
there is a slight resemblance, and in one or two cases deliberate Btudy is
needed to detecf evidence of structure in the smallest cry-tab.

In conclusion, the report of the foregoing pages may be summarized a>
follows. It baa been found possible t<> take very frequent photomicro-
graphs of crystals during their birth ami growth. An enlargement of
over four thousand diameters was obtained, and both common and polar*

RICHARDS AND ARCHIBALD. — GROWING CRYSTALS. 353

ized light were used. Only substances with high melting points were ex-
amined, and the crystallization was always from aqueous solution. No
properly focused image on any of the plates seemed to be devoid of crys-
talline structure. The growth in diameter during the first second of the
crystal's life was found to be vastly greater than during the subsequent
period. Not the diameter itself, but a power of the diameter, was propor-
tional to the time under the conditions used in our experiments. This
exceedingly rapid initial diametric growth accounts for a lack of definition
noticed in the first images, — a lack of definition sufficient to have misled
the eye, but not enough wholly to obscure the photographic evidence of
crystalline structure.

Hence we may conclude that whatever theoretical reason there may be
for believing that crystals always develop from a transitory liquid phase,
the present experimental evidence is inadequate to prove that these
globules attain a size visible in the microscope, except in the case of
substances which melt at temperatures not far from the temperature of
crystallization. The present paper is to be regarded rather as the sug-
gestion of a mode of study than as a finished treatment of the subject,
however.

The apparatus might be used to obtain a series of kinetoscopic pictures
of insects or other small animals or plants, and is now being used for the
study of the change in structure of steel at high temperatures.

Cambridge, Mass., October, 1898, to October, 1900.

vol. xxxvi. — 23

Richards and Archibald — GrowinguCrystals

Plate

Figure 2. Baric chloride ; 30 diam.;
exposure 0.04 sec.

Figure 4. Sodic nitrate : 30 diam.
exposure 0.10 sec.

Figure 6. Baric chloride; 110 diam.;
exposure 0.10 sec.

Figure 3. Baric chloride ; 30 diam. :
exposure 0.08 sec.

****** stiT,

I •. I 1 » * . ■ \

Figure 5. Sodic nitrate; 30 diam.
exposure 0.12 sec.

Figure 7. Cupric sulphate; 110 diam.;
exposure 0.12 sec.

Richards and Archibald — Growing Crystals

Plate

'I 1 \

«* »«B Egg I 1

Figure 8. Sodic nitrate ; HOdiam.; exposure 0.12 sec.

Figure 9. Potassic iodide ; lOOdiam. ; exposure 0.25 sec.

Richards and Archibald — Growing Crystals

Plate III

^

^

Figure 10. Potassic iodide; 580 diam. ; exposure 0.17 sec.

Figure 11. Potassic iodide ; 580 diam.; exposure 0.25 sec.

Figure 12. Potassic iodide ;[ 580 diam. ; exposure 0.17 sec.

Figure 13. Potassic iodide ; 580 diam. ; exposure 0.17 sec.

Figure 14. Potassic iodide; 580 diam. ; exposure 0.17 sec.

Fig re 15. Potassic iodide ; 580 diam. ; exposure 0.17 sec.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 21. — March, 1901.

DESIGN AS A SCIENCE.

By Dexman W. Ross, Ph.D.,

Lecturer on the Theory of Design in
Harvard University.

DESIGN AS A SCIENCE.
By Denman W. Ross, Ph. D.

Presented January 9, 1901. Received January 12, 1901.

Art may be defined as the expression of Life, or, more specifically, as
excellence in the matter of expression ; and excellence, in this case, may
be defined as consistency, a consistency in forms of expression. Consist-
ency has many manifestations, but they fall under three principal heads :
Balance, which is a consistency of oppositions (antitheses) ; Rhythm,
which is a consistency of association (joint action or movement) ; and
Harmony, which is a consistency of character (likeness). If art is con-
sistency in forms of expression, Balance, Rhythm, and Harmony are its
principles. They are also the principles of Beauty. We have no other
definite conception of beauty. It is a perfect relationship or connection
of parts in one organic whole. We find this unity in nature, when we
seek it, and we find it in the art of man, homo additus naturae. Wher-
ever and whenever we find it, we have the perception of beauty.

The idea which the Greek philosophers had of art was as nearly as
possible the one which I have given as the major premise of this argu-
ment, and it is in the art of the Greeks that I have found its most perfect
illustration. See Plato, in the Gorgias (§ 504) : "The artist brings all
things into order, making one part to harmonize and accord with another,
until he has constructed a regular and systematic whole; this," Socrates
says, " is true of all artists." Aristotle expresses the same idea when,
in his Poetics, he speaks of poetic imitation having " as its subject, a
single action, whole and complete, with a beginning, a middle, and an
end. It will thus," he says, " resemble a living organism and produce
its proper pleasure." See Poetics, xxiii. 1.

While consistency in forms of expression may be regarded as a defi-
nition of art, it is not, of course, a definition of what is significant or im-
portant in art. We can take a few lines and put them together so that
they shall be absolutely consistent, expressing one idea, unmistakably.
The result is a work of art, but the work is unimportant. It is an easy
thing to do. The work of art is important in proportion to the number

358 PROCEEDINGS OF THE AMEBICAN ACADEMY.

and variety of elements which are reconciled and united in its idea.
Consider, for example, Raphael's Dispute of the Sacrament. Think of
the number and variety of the elements united in that great composition.
Such a design is an achievement representing intellectual power of the
very highest order. So we discover, as the principal factor in art, the
mind of the artist, and the measure of this is observed in his ability to see
in many things one idea, and to express in one idea many things. Beyond
this power of the mind to grasp and express many things in single ideas,
a power which we can analyze, understand, and appreciate, lies some-
thing which defies analysis, something which we may appreciate, but
which we cannot understand. This is the strictly personal element
which goes into the work of a man, which stamps it as his, which dis-
tinguishes it from the work of other men. This personal element, when
ii i~ important, we call genius. The genius of the artist in his art is con-
stantly mistaken for the art itself. It seems to me that the genius of
tin- artist is something which lies beyond his art. His art is simply the
technique in which, and through which, his genius finds expression. In
speaking of art, therefore, I am speaking of the technique of expression
and nothing more than that. That is a matter of precise definition and
analysis. There is a passage of Plato in the Philebus (§ 55), where
Socrates says, "If arithmetic, mensuration, and weighing be taken from
any art, that which remains will not be much." In talking about art and
its principles, I mean art in this definite sense. There is a passage in
the eleventh canto of the Inferno of Dante which is significant in this
connection : "If you read your physics attentively [Dante refers here to
the physics of Aristotle], you will discover, after not many pages, how
your art follows that [physical science] just as far as it cau, as the
disciple follows the master."

Then; are many arts, the different modes and forms of expression:
gymnastics (including dancing); music, speech (including poetry); con-
Btruction (including architecture); modelling (including sculpture); and
painting (including design). Tin' particular art to which your attention
i~ called in this paper is the art of painting, in its highest form. Design.
Painting may be defined as expression by spots of paint, paint being in
this case :niv coloring material, no matter what it is, that may be used.
Design is painting with particular reference to the principles of art. We
have painting ae Design and painting a- Representation, which is the defi-
nition of visual impressions, a description of things seen, remembered, or
known, and we have Design in Representation. Design in which there
i- no representation, or in which the elements of representation are not

ROSS. — DESIGN AS A SCIENCE. 359

considered as such, may be called Pure Design. This may be defined as
the arrangement or composition of spots of paint for the sake of balance,
rhythm, and harmony; for the sake of consistency, unity, beauty. Pure
Design appeals to the eye just as music appeals to the ear. The term of
expression in music is the sound; the term of expression in design is the
spot of paint.

The spot of paint is three things : it is a tone, a measure, and a shape.
By tone I mean the pigment material used in drawing the measure of
the spot and its shape. By measure I mean the area covered by the spot,
its size. By shape I mean its outline, or contour. Put a spot of paint
upon a piece of paper, then change (1) its tone alone ; (2) its measure
alone ; (3) its shape alone ; (4) its tone and measure, leaving its shape
unchanged ; (5) its measure and shape, leaving its tone uuchauged ;
(6) its tone and shape, leaving its measure unchanged ; (7) change its
tone, its measure, and its shape, producing an altogether different spot.

Taking the spot of paint as the subject of my investigation, I will con-
sider, first, the element of tone, then the element of measure, and, lastly,
the element of shape. In order to study the element of tone we must
eliminate all differences of measure and of shape, which might be confus-
ing. Producing as many different tones as we can, in circles of half an
inch diameter, we find that we can produce a very great number and a
very great variety. Looking over the tones we have produced, we ob-
serve that every tone is relatively light or dark. It has what is called
value. It is a measure of light in the white-to-black scale. Observe,
also, that every tone has a color. It is red, or green, or violet, or some
other color, and the color which it has is relatively intense or neutral, or
it may be quite neutral. We shall find it convenient to regard the neu-
tral as a color. It is the color of white, or gray, or black. Tone means,
according to these observations, two things, — value and color.

We will consider, first, the element of value, afterwards, the element of
color. In considering values alone we must eliminate all differences of
color which might be confusing. Take the neutral pigments, white and
black, and see how many neutral values you can produce in circles of half
an inch radius. You can produce seventeen certainly, and perhaps a few
more. You will observe that in producing as many as seventeen neutral
values you are nearing the limit of visual discrimination, the limit of dis-
tinct definition, or expression. Observe that every value which you have
produced is a force drawing attention to itself. Observe that different
values exert different degrees of attractive force, that this force is deter-
mined in each case (other things, measure, shape, and color, beiug equal)

360 PROCEEDINGS OF THE AMERICAN ACADEMY.

by its contrast with the ground-tone upon which it has been drawn. If
the ground-tone is white paper, the value having the greatest attractive
force is black; if the ground-tone is a half-tone between white and black,
the forces of white and of black are equal. What is the result of all
these forces of attraction, as they act upon the eye? The eye is held at
rest at their centre of equilibrium. Where is that centre? In order to
answer this question, we must bring the values into a scale-relationship
upon a common ground-tone, otherwise we have no means of measuring
their respective contrasts, or the forces of attraction which depend upon
their contrasts. Make a scale of seventeen values, exclusive of white and
black, in seventeen circles of half an inch radius, in a straight line, half
an inch apart, and upon a ground-tone of the middle value. Be sure that
the values are at equal intervals of equal contrasts. In order to get
them into the perfect scale-relation which this implies, establish the ex-
tremes lirst. then the mean between the extremes, then intermediates, until
the scale is complete. The interval or contrast between value and value
may be great or small; the scale may be central in pitch, high in pitch,
or low. It is central in pitch when its middle value is at the half-point
between white and black. Considering the scale of values which you have
produced, you observe what you have observed before, that each value is
a force of attraction, that this force, other things being equal, depends
upon the contrast with the ground-tone. The only value which has no
force of attraction is the central one of the scale, the value which coincides
with the ground-tone and cannot be distinguished from it. Looking at the
scale again, pick out the values which have the same force of attraction.
They will be those at equal distances from the half-tone, which is the
ground-tone, making equal contrasts with it. In order to distinguish the
different values of the scale, we will call the middle value zero (<>).
The values above the middle value we will call 1, 2, 3, etc.. above. The
values below the middle we will call 1, 2, 3, etc., below. The values
above can lie written thus: ]_, 2, 3, etc.; the values below thus: T, 2, 3,
etc. The values basing the same force of attraction are, then, th

1 2 •'!
having the game numbers: ,■ "• .,- etc. The numbers are the measures

of the contrasts, and of the forces of attraction depending upon the con-
trasts. If now we Bcatter our seventeen values over the ground-tone of
the middle value we shall be able to discover the centre of equilibrium of
their forces, thai is to say. the poinl where the eye is held by them. We
have simply to remember the familiar principle of balance; that equal
attractions balance al equal distances on a line connecting their centres;

ROSS. — DESIGN AS A SCIENCE. 361

while unequal attractions balance in the same way, but at distances which
are inversely proportional to them, as attractions. Measures, shapes, and
colors being equal, values alone differing, values 4 and 1 balance on value
0, at distances 1 and 4 respectively. If the ground-tone were 2 instead
of 0, 4 and 1^ would balance on 2, at distances 1 and 2 respectively. In
this explanation of the balance of values we find the principle upon which
the designer proceeds when he wishes to create such a balance. He may
prefer to depend upon his visual feeling, but his feeling must be guided
by the law of balance whether he thinks of the law or not.

The scale of values is not merely a scale of visual attractions to be bal-
anced, it is also a rhythmic movement of values. The scale-relationship
is not, properly speaking, a relationship of opposition or antithesis ; it is
one of association or joint action. The values of the scale combine to lead
the eye in a movement from light to dark, or from dark to light, and
this movement is easy in proportion to the perfection of the scale. If
the scale is imperfect, if the intervals are not equal intervals of equal
contrasts, we have the same discomfort that we have in walking on the
irregularly placed sleepers of a railway track. We all know how tire-
some it is to do that. Not only is the eye led in the scale of values
from dark to light and from light to dark, but if the values be squeezed
together the eye is led quickly or abruptly ; if they are pulled apart the
movement is comparatively slow or gradual. By changing the direction
or the shape of the scale of values the eye may be led in different direc-
tions, and its movement may take a variety of shapes. A few simple dia-
grams would show the rhythmic character of the scale of values in these
several aspects. Values are in harmony when they are in the same scale,
and when the relations of the scale can be felt, visually. The least
contrast of the scale is a factor of the greatest, and when this relation is
distinctly felt we have a perception of harmony. The most perfect har-
mony is that of corresponding values.

Tone, as we have seen, means two things: value and color. We have
been considering the element of value. We will now consider the other
element, color. In order to do that satisfactorily we must eliminate all
differences of value. Producing as many differences of color as we can,
all in the same value (the half-tone between white and black), and all in
the same measure and shape (the circle of half an inch radius), we shall
find that we can produce perhaps twelve differences of color, and in each
color a certain number, perhaps eight differences of intensity. In order
to study color without being confused with the differences of intensity, let
us put all the colors not only in the same value, but in the same degree

362 PROCEEDINGS OF THE AMERICAN ACADEMY.

of intensity, the greatest intensity possible to the pigments on our pal-
ette. This being done, we shall observe, at once, that the colors have a
natural order or connection with one another. Red passes into green
through yellow ; yellow passes into blue through green ; and blue
passes into red through violet. There is, in other words, a natural
relationship for all the colors we can produce in the same value and
intensity. There is a natural scale of colors, as there was a natural
scale of values. This scale of colors is, of course, the scale of the
spectrum. The spectrum which I have followed in this investigation, is
the normal spectrum, the spectrum of the grating, not the spectrum
of the prism. The difference is explained by Rood, by Lommel, and
by other writers. Reading the spectrum from the red end towards the
violet end, the colors follow one another, approximately at equal inter-
vals of equal contrasts, as follows : red, suggesting Chinese vermilion ;
yellow, suggesting aureolin ; green, suggesting emerald green ; blue, sug-
gesting cobalt with a little emerald green in it; violet, suggesting ultra-
marine with a little rose madder in it. Beyond the violet end of the
spectrum we may observe the color which we call purple. It suggests
rose madder witli a little ultramarine in it. This color does not belong
in the spectrum series. It is due to the overlapping of the red and
violet ends of the spectrum. It is, however, a color which we must use,
and if the primary or important colors of the spectrum are red, green,
and violet, purple exists for us as an intermediate between red and
violet, just as yellow is intermediate between red and green, and blue
between green and violet. Between the six colors, red, yellow, green,
blue, violet, and purple, come intermediates and the intermediates of
intermediates up to the limit of visual discrimination. Setting the nor-
mal spectrum upon the circumference of a circle, with purple as a con-
necting link between the ends, the interval between any two colors can
be described as an interval of so many degrees. We have an interval
of 60°, the interval separating red from yellow, yellow from green, green
from blue, blue from violet, violet from purple, adjacents in the scab' of
six colors. We have the interval of 120° between the adjacents of a
scale of three colors, — red, green, and violet, I'm- example; and we have
the interval of 180° between the adjacents in the scale of two colors, —
red and blue, for example. This is the greatest possible interval. Ir is
the interval between colors opposite one another in the circle, the colors
which we call complementaries. In order to study the various intervals
of the BCale of colors with certain conclusions, we must eliminate all
differences of value and all differences of intensity. Wc can then

ROSS. — DESIGN AS A SCIENCE.

363

which intervals give the greatest satisfaction to the sense of vision.
The attempt to reach conclusions on the question of color-contrast, by
comparing colors in different values and of different intensities, is per-
fectly futile.

The greatest possible interval in the color scale is, as I have said, the
interval of 180°, the interval between opposite colors of the circle.
These colors are known as complementaries. In the scale of six colors
there are three pairs of complementaries: red and blue, purple and
green, yellow and violet. For the sake of brevity we will indicate the
colors by their initial letters: R for red; Y for yellow; G for green;
V for violet ; B for blue; P for purple; N for neutral. For the scale
of neutral values we have already a terminology. Complementary
colors, when tones of the same value and intensity are mixed together,
neutralize one another, approximately. The relation of the complemen-
tary colors may, therefore, be stated in this form : —

Y

— N —

V

R

— N-

B

P

— N-

•G

Observe that the complementaries balance, one against the other, on
the intermediate neutral, when of the same value and intensity. The
degree of intensity may be represented by the distance or space between
the sign of the color and the sign of the neutral which separates it from
its complementary. The greater this distance the greater the intensity.
In the statement which follows, yellow and violet and purple and green
are all equally intense, but the red and the blue are twice as intense : —

Y — N — V

E N B

P — N — G

The complementaries balance on the intermediate neutral, other things
being equal, at equal distances from one another on the straight line

But in the arrangement

connecting their centres

R — N

B

the neutral being the ground-tone, the red being only half as intense as
the blue, it will have to be moved to twice the distance, unless its

364

PROCEEDINGS OF THE AMERICAN ACADEMY.

measure or quantity is doubled. I shall, presently, speak of measure
as an element of balance.

When it comes to the consideration of color intervals we have to
think not only of the interval between one color and another in the
spectrum scale, but, also, of the interval between each color and the neu-
tral in which it disappears and is lost to vision. Take red, for example,
in its greatest possible intensity, an intensity limited by the pigment
material which we possess. This red is contrasted not only with its
neighbors in the scale of colors, purple on the one hand and yellow on
the other, but it is contrasted with itself in various degrees of neutraliza-
tion. Establishing the greatest possible intensity of red on the one hand
and a perfect neutrality on the other, both in the same value, and using
for measure and for shape the circle of half an inch radius, make a scale
of nine tones of red, the extremes of intensity and neutrality being
included in the scale. What has been said of the other scales may be
said of this one ; a repetition is unnecessary.

We have considered the scale of values and the scale of colors sepa-
rately. Now let us put the two scales together. The values being
neutrals in every case, we can set complementary scales of colors on
the right and left of the scale of values. That will give us two scales of
colors, and between them the scale of values, as follows : —

Light

White

G 8

P

Y 7

V

R 6

B

P 5

G

V 4

Y

B 3

R

G 2

P

Y 1

V

R 0

B

P 1

G

V 2

Y

B :;

i;

G 4

P

Y 5

V

R 6

B

P 7

(i

V 8

Y

Black

Darkne

ss

Following this diagram, put in the place of the value numbers values
of neutral color, moving from the central neutral up towards light i white)

ROSS. — DESIGN AS A SCIENCE.

365

and down towards darkness (black). Then, on a ground-tone of the
central neutral, alongside of the values and in the values, set the colors
in spots of paint, all in the same intensity: if you can. You will imme-
diately discover that you cannot do this. The colors in the light values
are inevitably neutralized by white, and colors in the dark values are
inevitably neutralized by black or some equivalent dark neutral. It
is only towards the centre of the scale of values that you can get to
any considerable intensity of color. If you consider the matter you
will understand that this neutralization of the colors in light and in
darkness is as it should be. It is exactly what happens to the colors
in nature as they occur between light and darkness. Color is observed
in its greatest intensity at the half-point between the light, whatever it is,
and the darkness, whatever that is. It is evident that the form in which
I have described the relation of the color and value scales needs to be
modified. The colors as they approach the half-point between light and
darkness must become more and more intense, the greatest possible inten-
sity being reached at the half-point, exactly. We have seen how the
measure of intensity can be indicated, diagramatically, by increasing or
diminishing the space between the complementary colors in any value
and the intermediate neutral ; so all we have to do in order to describe
the law of increasing and decreasing intensities is to pull the color scales
apart at the half-point between the extremes of light and of darkness.
This has been done in the diagram which follows : —

V

B

G

Intensity R

Light
White

8

Y7 V

ROB

P 5 G

4

3

2

1

0

1

2

R

V
G

r> Complementary
Intensity

B

G

R

4

5

R 6 B

P7 G

8

Black
Darkness

V

3GG

PROCEEDINGS OF THE AMERICAN ACADEMY.

It will be found that this system works just as well if we turn it
upside down. In doing this the relation of the color scales to the value
scale is reversed, as in the following diagram: —

B

<;

Intensity R

B

Light

"White

P7 G
R 6 B

Y 5 V

4
3

2

1
0
1
•2
3
4

R

J t. Complementary

Y Intensity

R

G

R 6 B

V 7 V

8

Black

Darkness

The diagrams which have been given show systems in which the colors
red and l>lue are the dominant colors, having the greatest intensity.
Predominance might be given to yellow and violet, or to purple and green,
or we might bring the point of intensity b< tween two pairs of comple-
mentaries. In that case we should have four colors as dominants, all
equally intense. Every change of the dominants means, of course, a
change of the whole system. The system may lie changed in other
ways. We can raise or lower its pitch within the extremes of white

and black, thus : —

ROSS. — DESIGN AS A SCIENCE.

367

White

/\
V

Black

White

/\

V

Black

White

/\
\/

Black

We can extend or contract the scale of values, thus : —

White

/\
\/

Black

White

0

Black

White
Black

We can increase or diminish the degree of intensity, thus : —

White

White

White

O

0

Black

Black

Black

By these various modifications an infinite number of specific forms
of the system can be developed, all consistent with the system in its
abstract idea.

If necessary, a system of half lights, half darks, and half intensities

368

PROCEEDINGS OF THE AMERICAN ACADEMY.

of color can be used in connection with the oue described, in the manner

shown in the following diagram

Light

White

8

Y7 V

ROB

P 5 G

V 4 Y

B B 3 R R

G G 2 P P

Intensity

Y Y 1 V V

R R 0 B B

P PIG G

V V 2 Y Y

B B 3 R R

G 4 P

Y 5 V

R 6 B

P 7 G

8

Black
Darkness

Complementary
Intensity

As we increase the number and variety of tones to be kept all at
equal intervals of equal contrasts, all in perfect rhythm and balance,
the problem of consistency, which is the problem of art, becomes more
and more difficult.

There is a possible objection to the system of color-values or tones
which I have described. In the spectrum the colors are all equally
intense. They differ only in value or luminosity. In the system just
described this equality of intensity is ignored, and as to the luminosities,
if they are observed on the hot side of the spectrum, they are ignored on
the cold side, and if they are observed on the cold side they are ignored
on the hot side. A different .arrangement of values and colors is pos-
Bible. [f we distribute our range; of light into five registers, — a register
of half-tones, a register of lights, a register of darks, a register of high
lights, and a register of extreme dark-, — we can consider each one of
these five registers as a potential spectrum, and we can arrange the colors
in each register according to their several values or luminosities and have
them all equally intense. The increase or diminution of intensities is
not, then, from color to color but from register t<> register. In the scale,
of five registers, the middle one will be the register of greatest intensities.
The registers above it and below it will be registers of less ami of least
intensity. This Bystera is described in the following diagram : —

ROSS.

DESIGN AS A SCIENCE.

369

Light

White

YN

G

Hi

?h N B Lights

P

V

Y

N
Register of

G

R

N

B

P

Lights

Y

N

V

Register of Half-Tones

G

II

N

B

P

Intensities

Y

N
Register of

G

V

R

N

B

P

Darks

N

Y

G

V

Extreme N R Darks

P

NV

Black

Darkness

This system would seem to be a particularly natural and proper arrange-
ment of the values and the colors, for, if you throw the spectrum on
white paper in sunlight the colors are seen all pale in the white light,
equally intense so far as you can see them, but with the differences of
value or luminosity which are indicated in the diagram. If you throw
the spectrum on white paper in shadow (half light), you see the colors in
equal intensity and in the greatest intensity, with the same differences of
luminosity. If you throw the spectrum on black paper in shadow, you
will observe the same equal intensity of colors, so far as you can see
them, and the same differences of luminosity, but the whole spectrum
is disappearing in neutral darkness. This system of color-values or tones
vol. xxxvi. — 24

370 PROCEEDINGS OF TUE AMERICAN ACADEMY.

in which we have a spectrum to a register, in which the colors in each
register arc all equally inteuse, but in values representing their several
natural luminosities, cannot, of course, be turned upside down, because
that would reverse the luminosities; but the system admits of the other
changes which I have described, — the chauges of pitch, the extension
or contraction of the value scale, and the extension or contraction of the
intensities.

When it seems desirable, the middle register of greatest intensities
may be left out of the system. The register of lights and the register
of darks can then be brought close together, just above and just below
the central neutral. Then the lights and the darks are all equally in-
tense, and the first diminution of intensity is found in the register of high
lights and in the register of extreme darks. This arrangement may be
used both in Pure Design and in Representation. It is a system which
ought to give great satisfaction to the colorist because of the number and
variety of the colors, all equally intense, which it allows him to use.

If you take your palette and, following any of the diagrams which I
have given, work out an illustration of the system, taking the central
neutral as ground-tone, and putting the tones in circles of half an inch
radius, you will observe that you have in the relationship of the tone- a
relation of balance, of rhythm, and of harmony. The system, whichever
system it is and whatever form of the system is followed, is an illustra-
tion of Pure Design. Again, I am tempted to quote a passage of Plato
in his .Symposium (§ 187), in which the physician Eryximachns Bays
that "harmony is composed of differing notes of higher or lower pitch
which disagreed once but are now reconciled by art." In these various
sj stems of color-values or tones, we have a reconciliation of many differ-
ing element-, harmonized by the art of design. Observe how the rhythms
of the different scales are so disposed that they balance in a perfect
equilibrium, and how by the principle of equal intervals of equal con-
trasts the many elements of each system are all perfectly related.

Now we must take up .and consider the second element of the spot of
paint. — measure. In order to do this without confusion, take one tone,
black on white paper, and one .shape, the Square. Thus eliminating all
differences of tone and of shape, you can vary the measure and study it
in all possible \ aiiations. Take some white paper and draw on it five
black si i ua res of different sizes. Observe thai you have harmony of tones
because the BquareS are all black, and you have harmony of shapes be-
cause the shapes are all square, but you have no harmonv of measure.

There is do connection between your measures, unless you have made

ROSS. DESIGN AS A SCIENCE. 371

j
one, intentionally. It would not happen by accident. Now draw five

squares in a scale, so that they shall be as 1 to 2, to 4, to 8, to 16, to 32,
in the proportions of their measures. This is easily done by drawing the
second on the diagonal of the first, the third on the diagonal of the sec-
ond, and so on. Observe the difference between the five related and the
five unrelated measures ; the harmony of the related measures. Arrange
the related measures in a row at equal intervals apart, the smallest first
and the largest last, and observe how you have in your arrangement not
only a harmony of measures which the scale-relationship gives, but you
have, also, in the connection of the measures, a rhythmic relationship.
The eye is led from measure to measure, just as it was led in the scale of
values from value to value. By rearranging the rhythm of the measures,
the movement can be made to change its direction and also its shape. By
bringing the squares close together the movement becomes abrupt.
Separating the squares by a larger interval, you can make the move-
ment more gradual.

There is another point of view from which the measure must be con-
sidered. Every measure is a force of attraction, and the amount of this
attraction is determined (other things being equal) by the measure
itself. A large measure attracts more attention than a small one. The
measure of two attracts twice as much attention as the measure of one.
We have in our scale of measures, therefore, a scale of visual attractions
proportioned as 1 to 2, to 4, to 8, to 16, to 32. Break up the scale and
scatter the squares over your paper and observe that the eye is no longer
led in a rhythm, but is held at rest by the opposition of attractions at the
point which is their centre of equilibrium. When the problem is, to find
this point, we must remember the law of balance : that equal attractions
(measures in this case) balance at equal distances on a straight line con-
necting their centres, and that unequal attractions balance in the same
way but at distances inversely proportional to them. In balancing tones
we considered the element of contrast, measures being equal. In balanc-
ing measures we consider what they amount to respectively. The centre
of equilibrium may be indicated by a point, or more satisfactorily by a
symmetrical outline enclosing all the balanced measures and having
with them a common centre. When the' measures are accidental and
unrelated, as they were before we brought them into scale-relationship,
they are nevertheless attractions which hold the eye at their centre, and
the centre can be found, approximately, by means of a small unit of
measurement taken as a common divisor. The centre can be approxi-
matelv ascertained by visual feeling, but we are talking about a scientific

372 PROCEEDINGS OF THE AMERICAN ACADEMY.

basis for design, to be a verification or correction of visual feeling. The
part which visual feeling plays in design is well enough understood.

The third element of the spot of paint, the one which we have not yet
considered, is shape. To study shape alone we avoid all differences of
tone and measure. For tone we may take black on white paper, and for
measure the square of an inch. Then we must vary the shape in every
possible way without varying either the tone or the measure. It is a
little difficult to vary the shape without varying the measure, but we can
do it. approximately, with the help of an underlay of small squares put
under a tracing paper upon which we draw. The power of estimating
the measure of the shape, no matter how irregular it is, is a power which
every draughtsman, every painter, every designer must have. Make as
many different shapes as you can, all black on white paper, and all in the
measure of the square of an inch. Observe that some of the shapes are
rhythmical, suggesting a joint action or movement of parts, that others
are symmetrical, suggesting opposition or contradiction of parts, while
others show both rhythmic and symmetric elemeuts. Shapes are in
harmony when they have the same or a similar character. Straight lines
go together in harmony. Curved lines have in common their curvature,
and fall into classes, circles, spirals, etc. Sipiare spots harmonize as
squares, and round spots as rounds. Angles go together in scale-rela-
tions based upon degrees.

Observe, however, in this connection as in others, that a little difference
is more disturbing than a large difference, when there is no sufficient
reason for any difference at all. when the repetition of the same shape-
character would be as satisfactory. Most perfect harmony exists, of
course, between shapes which have one and the same character, so in
design we prefer a repetition of similar elements to any composition of
insignificant differences. We are, however, apt to have differences of
character given to us in the terms or conditions of our problem. What we
have to do is to make the hest of these conditions. In such cases we can
make up for any lack of harmony in shapes by harmony in other than
shape-relations. Shapes are in harmony when they have the Bame mi
me (harmony of measure). They are in harmony when they have the
-one tone (harmony of tone). They maj have the same value without
having the same color, and the Bame color without having the same value.
They may have the same color without having the same intensity, so that
there are many ways of achieving harmony when there is no harmony of
the Bhapes themselves.

Shapes having the ame measure are in balance when they are n rereed

ROSS. DESIGN AS A SCIENCE. 373

and set side by side so as to contradict one another. A perfect balance or
antithesis of shapes is what we call symmetry. Symmetry is, accordingly,
a specific form of balance. It is shape-balance, and as such it must be dis-
tinguished from tone-balance and measure-balance. The only perfect
balance of shapes is the balance of similar shapes set in reverse, one
against the other, and having the same measure ; but we may have a
partial balance in the reversion and opposition of similar shapes when
they have different measures.

When two or more shapes are arranged so as to suggest a joint action
or movement, we have what may be called a rhythm of shapes. This
rhythm may be straight or curved in its character, or it may combine
both curvature and straightness. As the eye moves more rapidly upon a
straight line than upon any other, a rhythm showing many straight lines,
all having the same direction, will give to the eye the sense of rapid move-
ment, and this sense of rapid movement is lost in a rhythm which shows
many curves or angles upon which the eye moves more intricately and
therefore more slowly. There is another element to be considered in
connection with the rhythmic composition of shapes ; that is the sugges-
tion of a possible resistance. The idea of resistance does not lie in the
shape of the spot of paint, in the shape itself, but in a mental association.
If we wish to produce the sense of rapid motion we must be sure not to
suggest any opposition or resistance. Rhythms set in contrary motion
tend to balance one another, and in the measure in which they balance
one another they bring the eye to the rest of equilibrium.

I have now described the spot of paint in its three elements, tone,
measure, and shape, and I have shown, or tried to show, how each of
these elements may follow the principles of balance, of rhythm, and of
harmony, which, as we have seen, are principles of order and of beauty.
In the practice of Pure Design, which is the composition of spots of paint
for the sake of order and beauty, we begin with a few tones, measures,
and shapes, and try to bring them into the relations of balance, rhythm,
and harmony ; in other words, into an idea of beauty. When we have
achieved this, in the composition of a few elements, we do the same thing
with a larger number, proceeding, thus, from comparatively simple to more
and more difficult problems. The elements which we use in any problem
are not necessarily simple. We may take simple tones, measures, and
shapes, or we may take compositions of them. Then the problem takes
the form of a composition of compositions. In order to rise to anything
important in design, the designer must be able to think freely and easily

374 PROCEEDINGS OF THE AMERICAN ACADEMY.

in the terms of his art. The designer must be able to think in tones,
measures, and shapes precisely as the composer of music thinks in the
sounds of voices and of instruments. The measure of his ability as a
designer is then revealed in his power to think of many things in single
ideas, and to express in single ideas many things. At first a somewhat
painful effort has to be made to bring the composition of tones and meas-
ures and shapes into the lawful relationship of a single idea; but, by
degrees, the designer comes to think of his tones, measures, and shapes in
lawful forma only. He is then a master, and he will follow the sugges-
tions of his imagination as it leads him into the world of tone-, measure-,
and shape-ideas. This world must be as wonderful as the world of
musical sounds. We know something of that in the revelations which
the great composers of music have given us in their compositions. Of
the possibilities of Pure Design, we can only guess what they may be.
Then, when it comes to Design in Representation, and we have in addition
the lawful composition of tones, measures, and shapes, the expression of
visual knowledge in the form of true ideas, we rise to still higher possi-
bilities in the connection and relationship of Beauty with Truth. As we
rise above the accidents of vision or of memory to the knowledge of things
seen in their ideas or ideals, we discover that our knowledge of nature or
life is a knowledge of Nature's consistency, of her balances, her rhythms,
her harmonies, her order, her incomparable beauty. In other words, as
science rises from particulars to what is general and universal, as she rises
to the understanding of principles and laws, causes and sequences, she
comes to a conception of nature as pure design. The statement of sci-
entific truth becomes an illustration of pure design, and art and science
become one. "At last the vision is revealed to him of a single science,
which is the science of beauty everywhere." (Plato, Symposium, ^ 210.)

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 22. — April, 1901.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

THE OCCLUSION OF MAGNESIC OXALATE BT CALCIC

OXALATE, AND THE SOLUBILITY OF

CALCIC OXALATE.

By Theodore W. Richards, Charles F. McCaffrey, and

Harold Bisbee.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF
HARVARD COLLEGE.

THE OCCLUSION OF MAGNESIC OXALATE BY CALCIC
OXALATE, AND THE SOLUBILITY OF
CALCIC OXALATE.

By Theodore W. Richards, Charles F. McCaffrey, and

Harold Bisbee.

Received Feb. 21, 1901. Presented March 13, 1901.

For many years it has been known that magnesic oxalate is carried
down with calcic oxalate in the ordinary course of quantitative analysis.
This is only one case of the very general phenomenon of concomitant
precipitation, or occlusion, sometimes explained with the help of van't
Hoff's conception of "solid solution,"* and not widely understood, no
matter what name may be used. It is of considerable interest, both prac-
tically and theoretically, to obtain evidence concerning the mechanism
of this class of analytical irregularities.

In a foregoing paper, f one of us has shown that occlusion is probably
the distribution of an undissociated substance between the solution and
the nascent solid. If this is the case, the amount of material occluded
should be directly proportional to the concentration of the undissociated
part of the substance in question. The application of this idea to the
present case of calcium and magnesium seemed capable of furnishing
a further clue to the theory of the general problem, as well as of pro-
viding a more satisfactory analytical method in this particular case.
Since the magnesium comes down with the calcium in the form of
oxalate, it is the concentration of the undissociated magnesium oxalate
in the solution which will determine the amount of the occlusion, if this
hypothesis be true. According to the law of mass action and the dis-
sociation hypothesis, this concentration may be increased by adding
either constituent ion in excess. It may be diminished by adding a
large excess of any other partially ionized substance, % especially those

* Schneider, Z. phys. Chem. 10, 425 (1892).

t Richards, These Proceedings, 35, 377 (1900).

X Compare Arrhenius, Z. phys. Chem. 31, 198 (1899).

378

PROCEEDINGS OF THE AMERICAN ACADEMY.

which form a complex ion or undissociated substance in solution involving
either magnesium or oxalic acid.

One of the most effective causes diminishing the concentration of the
magnesic oxalate, and therefore the occlusion, should be the hydrogen
ion, for this tends to remove the oxalic ion, and hence to cause the
ionization of undissociated magnesic oxalate.

Another effective cause should be an increase in the concentration
of amnionic salts present, which not only exert the effect of any other
partially dissociated salts, but have also the well-known property of
forming complex compounds with magnesium. This formation naturally
removes magnesic ions and hence magnesic oxalate from the solution.

A third obvious means of diminishing the concentration of the mag-
nesic oxalate is by diluting the solution. By this process, the actual
amount of undissociated oxalate is diminished, and the concentration of
the undissociated part is thus diminished even more rapidly than in the
ratio of the changing volumes. All of these tendencies except the sec-
ond apply to the calcium as well as to the magnesium, although to a less
extent, for calcic oxalate is far less soluble than magnesic oxalate. Cal-
cium has not so great a tendency to form complexes with amnionic salts

as magnesium.

These relations are partially expressed by the following scheme, in
which no attempt is made to express the exact nature or the ionization
of the magnes-ammonium complex : —

MgCl2 lz^

+
x NH4C1

A I

Mg(Nii4La+2(?)

2C1- + Mg+ +

+ +

211+ + C204=i-; II2C204

2IIC1 MgCaO* in solution

S I

MgC304 + CaC204

Precipitate

Reviewing the older work upon this subject, one finds that most of
the known facts Bupport the hypothesis. Presenilis, in the few experi-
ments which are recorded in the end of the second volume of bis
"Quantitative Analysis,"* showed that the weight of the calcic pre-
cipitate obtained was less when the dilution was greater; that it was

* Fresenius. Quantitative Analyse, 2. s'Jl ilsTT-lss;).

Richards, McCaffrey, and bisbee. — calcic oxalate. 379

diminished by the presence of an excess of either amnionic hydrate or
amnionic chloride. When acetic acid was present, too little precipitate
was obtained, and a very large excess of oxalate in alkaline solutions
gave too much precipitate. But Presenilis attacked the problem merely
in an empirical fashion ; the guiding hypothesis of the present day had
not yet been suggested. It did not occur to him to study the effect of
gradual precipitation in a strongly acid solution.

Most writers of handbooks upon quantitative analysis have accepted
Fresenius's method of double precipitation, and the recent literature
upon the subject is unusually scanty. In a paper, of which we have
seen only an abstract,* Hefelmann calls attention to the necessity of
using dilute solutions in order to attain satisfactory results ; but other
important references to the subject in periodical literature could not be
found. A good method devised by H. P. Talbot, not to be found in the
periodicals, will be mentioned later.

For our experiments calcic chloride and magnesic chloride or sulphate
were prepared separately in a state of great purity. By the repeated
crystallization of the nitrate, the calcium material was freed from its
usual impurities, and the carbonate was precipitated from this nitrate by
means of pure ammonic carbonate. Standard solutions of this material
in hydrochloric acid were made from time to time, and the concentration
of these solutions was determined with the utmost care by precipitation
with ammonic oxalate, according to the method to be described. Mag-
nesic chloride and sulphate were carefully purified by repeated crystalli-
zation, and experimental proof was obtained of the absence of calcium
from them. The magnesium solutions were always of a strength approx-
imately equivalent to the calcium. Ammonic chloride was made by
passing pure ammonia gas into freshly made hydrochloric acid; oxalic
acid was especially purified ; and all the ammonia used was freshly pre-
pared in platinum vessels.

In the first place, the worst possible results were obtained in order to
show how great an improvement was possible by the successive intro-
duction of the modifications suggested above. To neutral solutions of
a mixture of 25 c.c. each of the magnesium and calcium solutions, made
up to 200 c.c, was added an excess of ammonic oxalate, but without the
addition of either ammonic chloride or acid. The two precipitates thus
formed were ignited at a very high temperature until constant in weight.
Below are recorded the results : —

* Zeitschr. anorg. Chem. 18, 401 (1898).

:;>()

PROCEEDINGS OF THE AMERICAN ACADEMY.

Precipitation i\ Absence of Precaution.

No. of
Experiment.

\\"t. of Ignited
Precipitate.

W't Of CaO taken

(determined in
Parallel Portions).

Difference

MgO present.

Error.

1
2

0.27C:
0.2597

0.2358
0.2358

0.0385
0.0239

+ 10.4%
+ 10.1%

Average error, + 13.3%

The next step was to test the effect of hydrochloric acid — or rather
of the concentration of ionized hydrogen — upon the occlusion. Since
in this case from au acid solution all of the calcium could not be pre-
cipitated, a larger amount was used, but the concentration of the mag-
nesium was the same as before. 25 c.c. of the magnesium solution,
7o c.c. of the calcium solution, and 10 c.c. of normal hydrochloric acid
were diluted up to 200 c.c. To this solution, heated to boiling, were
added 27 c.c. of normal oxalic acid, to which had previously been added
10 c.c. of normal hydrochloric acid for the sake of diminishing its
dissociation. Over half of the calcium was precipitated in the form of
fine white crystals of the oxalate. It is well known that the substances
iii- ■ — t capable of easy supersaturation are those which form the largest
crystals; and conditions which tend to promote the solubility of a pre-
cipitate in general tend to promote its ease of supersaturation. Hence
precipitates formed from solutions in which they are somewhat soluble,
are more crystalline than those formed from liquids in which they are
insoluble. Calcic oxalate is no exception to this general rule. The
easily handled precipitate was collected, thoroughly washed, ignited to
constant weight at a bright yellow heat, dissolved, reprecipitated, re-
ignited, and weighed again. The results below .-how bow comparatively
small was the occlusion of magnesium in this precipitate.

Precipitation from Acid Solutions.

ol
Experiment.

\\ . i $h( ol 1 hi
Oxides.

Weigh) ol
(2d).

■*] 1 1 prea nt.

Krror.

8

4

o U86
0.4128

o lion
0 l"-7

0.0036

11

1.00%

Average error,

0.94

Richards, McCaffrey, and bisbee. — calcic oxalate. 381

The average error attained only one-fourteenth of its previous magni-
tude. Thus the presence of acid has a remarkable restraining effect upon
the occlusion ; a fact in accordance with the prediction of the theory.

In the next place, the action of amnionic chloride was studied. A
neutral solution of 25 c.c. each of the calcium and magnesium solutions
was made up to 200 c.c. and treated with two grams of amnionic chloride
and the same amount of ammouic oxalate as in the first experiment.
Two trials of this process yielded the following results : —

The Effect of Ammonic Chloride.

No. of
Experiment.

Weight of Oxides
found.

Weight of CaO

taken.

MgO present.

Apparent
Error.

5
6

0.2376
0.2364

0.2358
0 2358

0.0018
0.0000

0.77%
0.25%

Average error, 0.51 %

About 1.1 milligrams more of calcic oxide was recovered from each of
these solutions by the addition of much more ammonic oxalate. This
precipitate, which fell after some time, was redissolved and reprecipitated
before weighing in order to free it from magnesium. The total amount
of magnesic oxide in the first precipitate was thus about one per cent,
or somewhat more than in the precipitates from an acid solution.

These experiments show that ammonic chloride diminishes very much
the tendency of the magnesic oxalate to be precipitated, but that it exerts
also a similar although much less considerable effect upon the calcic
oxalate. The more ammonic chloride is added, the greater concentra-
tion of oxalate ion is necessary completely to precipitate the calcium, but
the more effective is the retention of the magnesium in the solution.
The limit to the advisable amount of ammonic chloride depends upon the
subsequent method to be used for the determination of the magnesium ;
but for ordinary purposes an equivalent-concentration ten times as great
as that of the magnesium present should answer.

This action may be explained, as has already been stated, partly by
an effect which would be caused by any electrolyte aud partly by the
additional formation of an unstable complex. The existence of this
complex is abundantly confirmed by the other reactions of magnesium in
the presence of a large amount of ammonic chloride. It is well known
that many of the common reactions fail, and that other reactions, such

:;sj

PROCEEDINGS OF THE AMERICAN ACADKMY.

a- i lie precipitation of amnionic magnesic ])hosphate, require more time
for their completion, when much amnionic chloride is present.*

If, as a matter of fact, the occlusion is proportional to the concentra-
tion of the uudissociated magnesic oxalate, additional amnionic oxalate,
even in the presence of amnionic chloride, ought to increase the weight
of the precipitate. The following experiments, similar to the last except
that in each case three grams instead of one of amnionic oxalate were
ii.M-d, were made to test this point.

The Effect of Excess of Ammoxic Oxalate.

No of
Experiment.

Weight of

Oxide-.

Weight <.f CaO

taken.

MgO present.

Error.

7
8

0.2390
0.2303

0.2358
0.2358

0.0032
0.0035

130%
1.48%

Average error, 1 12%

Thus a trebling of the amount of oxalate present increased the error
by about a third of its previous value. At first one is surprised that the
increase is not greater; but it must be remembered that the oxalate was
added rather slowly, so that most of the precipitate was formed before
a large excess of oxalate was present. It is chiefly the concentration of
the magnesic oxalate present at the instant of precipitation, not the sub-
Bequent amount in contact with the precipitate, which influences the
distribution. When the precipitate has once appeared as a solid, the
action must be confined to the surface; lor diffusion into solids is ex-
ceedingly slow because of their rigid structure.

A number of analyses were made in the hope of combining all the
circumstances which tend toward complete separation, and of eliminating

all those which oppose it : but yet further difficulties arose. It seemed
probable that by gradual neutralization of an acid solution the calcic
oxalate might be precipitated in an environment containing as little mag-
nesic oxalate as possible, and thus be as \\'*^' as possible from this impur-
ity. Tlic mode of procedure was a- follows. To a mixture containing
it. 200 (■.,-.. 25 c.c. each of the calcic and magnesic solutions, were added
three grams of amnionic chloride. I 6 grams of oxalic acid, am! enough

• Compare Ostwald, Scientific Foundations <>f Analytical Chemistry (Mac-
millan, 1895), p 186.

richards, McCaffrey*, and bisbee. — calcic oxalate.

383

hydrochloric acid to keep the calcic oxalate in solution. Subsequently
strong ammonia was poured very slowly into the liquid, with continual
stirring, until the solution contained an excess of ammonia. Methyl
orange was found to assist materially the exact neutralization.

Precipitation by Concentrated Ammonia.

No. of
Experiment.

Weight of
Mixed Oxides.

Weight of CaO
taken.

Weight of
MgO.

Error.

9

0.2373

*

0.0015

0.6%

10 •
11

0.2383
0.2384

-0.2358-

0.0025
0.0026

1.1%

1.1%

12

0.2387

■

0.0029

1-2%

Average error

, 1-0%

Some calcium was found in the mother liquors upon the addition of
more amnionic oxalate ; but this is included above. Evidently no mate-
rial <*ain in accuracy is effected in this series, and the reason is not hard
to find. The ammonia was so strong that it caused instant neutralization
of the acid in its neighborhood ; and hence the idea of the method was
defeated, for the design was to effect a gradual neutralization, giving
time for the supersaturated calcic oxalate to separate.

In the next series twice as much oxalic acid was used, but the am-
monia added to effect the precipitation was far less concentrated.

Precipitation by Dilute Ammonia.

No. of
Experiment.

Weight of
Mixed Oxides.

Weight of CaO
taken.

Weight of
MgO.

Error.

13

0.2375

1

0.0017

0.7%

14

02374

0.0016

0.7%

15

0.2369

■ 0.2358-

0.0011

0.5%

16

0.2380

0.0022

0.9%

17

0.2379

■

•

0.0021

0.9%

Average error

0.74%

384

PROCEEDINGS OF THE AMERICAN ACADEMY.

Thus, diluting the ammonia had the beneficial effect which was ex-
pected. In the presence of so much oxalate the solution was, of com m .
practically free from calcium, hence this result indicates a distinct
improvement.

It has been already noticed that dilution of the original solution has
beeu found by others to lessen the amount of magnesium carried down.
This fact might have been easily predicted by the hypothesis that the
phenomenon is regulated by the Distribution Law. It is further verified
by two experiments given below, in which the method of Experiments
13-17 was repeated, except that the volume was 800 c.c. instead of
200 c.c.

The Effect of Dilution.

No. of
Experiment

Weight of
Mixed Oxides.

18
19

0.230'J
0.2371

Weight of CaO
taken.

0.2358

Weight of
MgO.

0.001 1
0.0013

Error.

0.42%
0.54 ;

Average error, 0. 1-

Ilere also no calcium could be recovered from the mother liquor, hence
it is clear that the dilution was really of service, reducing the magnesic
oxide from 0.74 per cent to 0.4S per cent. The theoretical diminution
to less than a quarter of the former value was not to have been expected,
for the addition of the ammonia in finite dilution introduces an irregu-
larity for which it is impossible to make quantitative correction.

At this stage in the work it was found that a method of precipitation

entially identical with that just given had already been published by
Professor II. 1'. Talbot in his admirable treatise on the elements of
Quantitative Analysis.* So far as we know, no account of it is to be
found elsewhere. Since no examples accompany this publication, the
preceding work, which was wholly Independent, affords useful confir-
mation "I his method as an approximate one for rapid work.

The chief difference between the method of Talbot and ours lay in the
respective amounts of oxalate, a much larger amount having been used
in our work in order to insure the total precipitation of the calcium.

It i- clear from the above account that this large initial 6XCe88 of the

• Talbot, Quantitative Analysis (Macraillan), 3d ed p 12 (1899).

richards, McCaffrey, and bisbee. — calcic oxalate. 385

oxalate, although necessary so far as the calcium is concerned, must in-
crease the amount of precipitated magnesium oxalate. But according to
•our original hypothesis, this excess is chiefly harmful when it is present
at the moment of precipitation, although it is not really needed to prevent
the solution of traces of calcic oxalate until the precipitation is practically
finished. Hence the oxalic acid also should be added gradually, or at
least in two portions, the first to supply enough oxalate to combine
with the bulk of the calcium, and the second to diminish the solubility of
the last traces. This plan was followed in the following work.

Another point was as yet undecided, — the length of time needed for
the essentially complete separation of the calcium. Hence five determi-
nations similar to the above were made with solutions each capable of
yielding 0.1496 grams of lime and an equivalent quantity of magnesia.
The only variation in these experiments was in the time elapsing between
precipitation and filtration.

The Effect of the Time between Precipitation and Filtration.

No. of
Experiment.

Weisht of
Precipitate.

Error.

20

Filtered after standing \ hour

0.1488

-0.53%

21

" " f hour

0.1492

-0.27%

22

«• " " 24 hours

0.1501

+ 0.33%

23

" " " 48 hours

0.1507

+ 0.73%

24

" " " 80 hours

0.1511

+ 1.00%

Amount of calcic oxide present,

0.140G

.

The filtrates from the first two of these (Nos. 20 and 21) deposited
traces of calcic oxalate ou standing for two hours, but the others did
not deposit a trace in several days. From these five experiments two
conclusions may be drawn : first, that several hours are needed for a
separation of the last weighable traces of calcium, even when much
ammonic oxalate is used ; and secondly, that after the calcium has all
been precipitated, magnesium oxalate is absorbed by the precipitate at
a fairly constant although very slow rate, from a solution which will
of itself deposit no solid. The magnesic oxalate is slowly occluded by
vol. xxxvi. — 25

380

PROCEEDINGS OF THE AMERICAN ACADEMY.

the precipitate or deposited upon it even after the precipitation of the
calcium salt.

This points to another flaw in the earlier work, — namely, the solutions,
which had all been allowed to stand for at least sixteen hours before
filtering, had been left too long in contact with their precipitates.

In a case of this kind great accuracy is to be obtained only by a succes-
sion of approximations ; hence it seemed worth while again to make a
series of precise analyses, embodying all the advantages which had been
found up to this point, in order to discover from their possible variations
if there might be still another cause of error as yet undetected.

Newly prepared very pure solutions were used in these analyses,
which were made two years after the ones previously detailed. All the
precautions suggested by the foregoing pages were heeded, and two
series of results were obtained, one from calcium solutions only, and the
other from solutions containing precisely the same amount of calcium with
an equivalent amount of magnesium. These are given below in parallel
columns : —

Precipitation of Pure Calcic

I ».\ALATE.

(Volume at precipitation = 200 c.c.)

No. of
Anal) bis.

Time of
Digestion.

Weight nf
CaO.

25

3.6 hours

0.30G3

'20

3.5 "

0.3064

'_' 7

3.5 "

0.30G5

28

Q c u

• ) .)

0.3064

Average,

0.8064

Precipitation of Calcium in

Presence of Magnesium.

(Volume at precipitation = 500 c.c.)

No. of

A 1 1 : 1 1 > sis.

Time of
Digestion.

Weijrht of
CaO.

29

2.5 hours

0 3000

30

2.5 "

0.3060

31

3.5 "

0.3064

32

3.5 "

0.3066

33

:;.:» "

0.3058

Averag

re of last three

0.8063

The agreement between the two averages is very striking; it is to
some extent due to a compensation of errors. Even when we have
made due allowance for these, however, it furnishes Btrong evidence in
favor of the hypothesis which led to the before mentioned precautions!
Experiments :.".* and •">". not included in the average, show that at the

Richards, McCaffrey, and bisbee. — calcic oxalate. 387

end of two hours and a half there probably still remained a trace of
calcium in the solution.

While the possibility of separating calcium from magnesium with con-
siderable completeness by a single operation had thus been demonstrated,
one point still remained to be studied. In the course of the washino- of
these precipitates of calcic oxalate, it was noticed that the wash-waters
always gave a faint opalescence with neutral argentic nitrate, an opales-
cence which dissolved in nitric acid. Unlimited washing seemed not
to free the precipitate from this substance, hence the substance must
have been calcic oxalate itself. In short, calcic oxalate appeared to be
soluble in boiling water to an extent sufficient to affect a precise analysis.
Since no account of accurate determinations of the solubility of calcic
oxalate in boiling water could be found, the next problem was to deter-
mine this solubility.

The calcic oxalate was precipitated in the usual manner and washed
with exceeding thoroughness. A solution, not necessarily saturated, but
closely resembling one which might be obtained in the process of wash-
ing, was made by stirring the powder for fifteen minutes with water in
a platinum dish kept at the desired temperature. Of the three usual
methods of analyzing such a solution, — by weighing, titration, and the
measurement of electrical conductivity, — volumetric determination

Solubility of Calcic Oxalate. Series I.

Temperature = 90° .

No. of
Aualysis.

Volume of Permanganate
required by 100 c.c. of Sol.

Weight of Calcic
Oxalate corresponding.

Remarks.

34

35
36

1.80 C.C.

1.81 c.c.
1.81 c.c.

0.00115

0.00112
0.00112

Fresli filter, thoroughly
moistened.

Same filter as above.

Same filter as above.

Temperature = 25°.

37
38

1.10 c.c.
1.06 c.c.

0.00068
0.00060

Fresh filter.

Same filter as preceding.

:>S PROCEEDINGS OF THE AMERICAN ACADEMY.

seemed the best suited to the present purpose. The filtered solution was
therefore titrated with a solution of approximately hundredth normal
permanganate, of which one cubic centimeter corresponded to 0.00062
gram of calcic oxalate. The amount of the solution, 0.12 c.c, which
was required to impart the usual pale pink color to the assay, was
always subtracted from the total volume run out. In each analysis ex-
actly a decilitre of calcic oxalate solution was used.

These results arc typical of a number of similar determinations which
were made in this way. The first portions running through a filter
paper always seemed to contain slightly more calcic oxalate than the
later portions, probably owing to the escape of fine particles which are
retained by the filter when it has become somewhat clogged. The later
values are probably the more reliable, hence it is clear that hot water
will dissolve easily over a centigram per litre of calcic oxalate, while
water at the ordinary temperature dissolves nearly seven milligrams
per litre.

This solubility is altogether too great to be passed without heed in
precise work. Its magnitude was such as to make desirable a more
exact determination of a more nearly saturated solution.

The first problem to be solved in this connection was the retention of
even the finest particles. The best filter papers of Schleicher and
Schiill, Dreverhoff, and others were tested, all with the result already
described; namely, that the first filtrate always yielded slightly more
calcic oxalate than the following ones. Finally it was decided to use
four layers of paper, anil to reject at least half a litre of filtrate before
beginning to collect for analysis. The filtration was effected by means
of a platinum inverted filter devised by J. P. Cooke.* and a simple
arrangement of tubes and stopcocks made the rejection of the first and
the collection of the subsequent portions of filtrate an easy matter. The
filtered liquid appeared perfectly char on inspection in strong light.
The inverted filter made it possible to maintain the filtering liquid at
the desired temperature. The time of digestion at the higher tempera-
tures was an hour; but at 25° an hour and a half was allowed. A
ver\ large platinum dish Berved as the vessel for digestion, and the liquid
was suitably protected from the products of combustion of illuminating
.j:,- and other impurities. A diagram of the apparatus will furnish all
further necessary explanation. A- before, a hundred cubic centimeters

Of the solution Berved \'>>r each titration.

» These Proceedings, 12. 124(18'

richards, McCaffrey, and bisbee. — calcic oxalate. 889

Figure I. Apparatus for Filtering.

A is the inverted filter, clad four-fold with filter paper. B is the collecting
flask, which fills only when C is open and D closed. E serves to retain the re-
jected liquid.

It is not certaiu that these solutions were wholly saturated ; but the
difference in concentration between these and the earlier solutions, which
had been digested during much less time, made it highly probable that
any error from this cause would be less thau the necessary errors of
titration. Hence no further determinations were made.

After these determinations were finished, Kohlrausch and Rose's *
determinations by means of electrical conductivity at lower temperatures
were consulted, and were found to furnish agreeable confirmation of these
independently obtained results. At 18° Kohlrausch and Rose's figure
was 5.9 milligrams per litre, while our extrapolated value is 6.0 milli-
grams; and at 40° the respective figures are 8.0 and 8.4 milligrams.
Our slightly larger values may be due to the fact that we determined all
the calcic oxalate in solution, while Kohlrausch and Rose measured only
that part which is dissociated. According to the table upon page 200

* Zeitschr. fur phjs. Chem., 12, 234 (1890).

300

PROCEEDINGS OF THE AMERICAN ACADEMY.

Solubility of Calcic Oxalate. Series II.
Temperature = 95°.

No. of
Analysis.

Volume of Permanganate
required1 by 100 c.c. 8ol.

Weight of CaC,04
in 100 c.c. Solution.

39

2.19 c.c.

0.00136 gram.

40

2.18 "

0.00135 "

41

2.35 "

0.00146 "

42

2.28 "

0.00141 "

Average,

0.00140 gram.

Temperature = 5f

1°.

43

1.55 c.c.

0.00096 gram.

14

1.56 "

0.00097 "

45

1.53 "

0.00095 "

46

1.52 "

0.00094 "

Average,

0.000955 gram.

Temperature = 2a

°.

47

1.09 c.c.

0.00068 gram.

48

1.11 "

0.00069 "

49

1.08 "

0.00067 "

Average,

0.00068 gram.

of Kohlrausch and Ilolborn's book (1898) this interpretation might
account for the difference, but in any case they make no pretensions
to great accuracy. In view of Ostwald's recent work on the surface
tension of Bolids,* a more precise determination than ours would be of
no service unless the diameters and shapes of the solid particles were
defined.

« Ostwald's Zeitsclir pliys. Cliem., 34, 495 (1900).

richards, McCaffrey, and bisbee. — calcic oxalate. 391

In the light of all these facts, there can be no question that calcic
oxalate is soluble euough to demand further precautions in washing than
are usually taken. The obvious means of diminishing this solubility is
to wash the precipitate with a dilute solution of amnionic oxalate, instead
of with pure water. With the idea of testing the efficacy of this pre-
caution, as well as with the purpose of determining the strength of a
new solution of calcic chloride, the following series were made.

In the first place three portions, removed with a very exact 25 c.c.
pipette, were individually precipitated by degrees in our usual fashion
from a solution having a volume of two hundred cubic centimeters.
They were each allowed to remain four hours before filtering. The
weights of calcic oxide resulting upon ignition to constant weight at a
bright yellow heat were respectively 0.3479, 0.3480, and 0.3480 gram.
The last two of these were now dissolved in hydrochloric acid and repre-
cipitaied with the utmost care by means of amnionic oxalate. The
weights were now, after ignition, 0.3474 and 0.3475 gram respectively —
an average loss of 0.00055 gram or 0.16 per cent. If pure water had
been used for washing, the loss would have been much greater. This
loss may have been due to incomplete precipitation, or to a slight solu-
bility in the wash-water in spite of the presence of amnionic oxalate, or
to the mechanical passage of exceedingly small particles through the
filter. Whatever may be the cause, this deficiency is clearly a quantity
which must apply to most if not all of the foregoing work. Its applica-
tion does not cause sufficient change in the results to affect the con-
clusions which have been drawn ; in many cases of comparison each of
the numbers compared is affected equally.

It became now an interesting question to carry out in the presence of
magnesium the operations just described. Except for the degree of
dilution, the other conditions were in every respect like those named.
In two experiments where the precipitation took place in a volume of
300 c.c, the weights of impure calcic oxide were respectively 0.3491
and 0.3489 gram, a gain of a milligram, or 0.29 per cent. In two other
cases where the volume was GOO c.c, the weights were 0.3486 and
0.3483 gram respectively, a gain of 0.13 per cent.

This last gain is almost exactly equal to the inevitable loss during
filtration, so that the occluded magnesia almost exactly replaced the lime
which is not collected.

The slightly impure specimens of lime of the last two experiments,
weighing respectively 0.3486 and 0.3483 gram, were now dissolved and
reprecipitated as oxalate, which was washed with ammonic oxalate as

392 PROCEEDINGS OF THE AMERICAN ACADEMY.

before. The resulting specimens of oxide weighed respectively 0.3469
and 0.347-4 gram, an average of U. 34715 gram, while the weights ob-
tained by a similar double precipitation in the absence of magnesium
averaged 0.3-1 7 -15 gram, or 0.3 milligram mure. This difference is too
small to have much significance, hence the conclusion is allowable that
as much of the lime as is practicable had been precipitated.

It is clear from these experiments that the double precipitation usually
practised, while certainly eliminating the magnesium, inevitably involves
the loss of calcium. The single precipitation minimizes this last risk,
and by the compensation of two errors very small in themselves, gives
a result which is very near the truth.

The conclusions attained in this paper may be summarized as
follows : —

1. All those conditions which tend to diminish the amount of undis-
sociated magnesic oxalate in a solution tend to diminish the amount of
this impurity in calcic oxalate precipitated from the solution.

2. Hence evidence is furnished supporting the hypothesis that occlu-
sion is a distribution of an undissociated substance between the liquid
and the " nascent " solid phase.

3. Magnesic oxalate is precipitated upon the calcic oxalate after this
is deposited, although far more slowly than while it is being deposited.
For this reason the filtration should not be too long delayed.

4. Calcic oxalate is sufficiently soluble in pure water to cause grave
inaccuracies in precise quantitative work.

5. This solubility may be diminished, but not wholly prevented, by
an excess of ammonic oxalate in the wash water.

6. By heeding all these relations, a reasonably precise separation of
magnesium and calcium may be made in a single precipitation. The
details are as follows : The magnesium in the solution should be not
much Btronger than fiftieth normal. About ten times the equivalent
amount of ammonic chloride, and enough oxalic acid to combine with
all the calcium, should be added to the mixed solution. It is well to
diminish the dissociation of the oxalic acid beforehand by the addition of
two or three times its equivalent amount of hydrochloric acid. To the
boiling mixture, colored with a drop of methyl orange, should be added
slowly a very dilute solution of ammonia, with continual stirring and
occasional pauses. The final stages of neutralization should not be
reached in less than half an hour.

When the neutralization has been effected, a very large excess of
ammonic oxalate should be added, and the mixture should be allowed to

Richards, McCaffrey, and bisbee. — calcic oxalate. 393

stand for four hours. The precipitated calcic oxalate should be thoroughly
washed with water containing amnionic oxalate. The filtrate will con-
tain all but about 0.1 or 0.2 per cent of the magnesium, and the precipi-
tate will contain all but about the same proportion of the calcium.

7. The precipitate, having been formed slowly from a solution in
which it was somewhat soluble, is distinctly crystalline and far more
easily handled than if it had been suddenly precipitated. This is a
general effect, which might render service in other cases.

Cambridge, Mass.
October, 1897, to February, 1901.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 23. — March, 1901.

CONTRIBUTIONS FROM THE CRYPTOGAMIC LABORATORY OF
HARVARD UNIVERSITY. — XLV.

PRELIMINARY DIAGNOSES OF NEW SPECIES OF
LABOULBENIA CEAE. — III.

By Roland Thaxter.

CONTRIBUTIONS FROM THE CRYPTOGAMIC LABORATORY OF
HARVARD UNIVERSITY. — XL V.

PRELIMINARY DIAGNOSES OF NEW SPECIES OF
LABOULBENIACEAE. — III.

By Roland Thaxter.

Received February 26, 1901. Presented March 13, 1901.

During the summer of 1900 I took occasion to visit Berlin in order
to examine the collections of insects at the Museum of Natural History
there in search of Laboulbeniaceae, and through the courtesy of the
director, Professor Moebius, and of the staff of the Entomological Labo-
ratories, to whom I desire in this connection to express my great obliga-
tions, was successful in obtaining numerous new and interesting forms
from all parts of the world. To the great kindness of Dr. David Sharp
of Cambridge, England, I am further indebted for the privilege of ex-
amining, for a similar purpose, his magnificent collections, especially of
Staphylinidae and Gyrinidae, as well as the large series of singular Cara-
bidae brought from the Hawaiian Islands by Mr. Perkins. A consider-
able number of new or peculiar forms have moreover been added to the
American flora since my return, for a portion of which I am indebted to
Mr. Charles Bullard, who has very kindly placed his material at my
disposal. The number of new forms from all sources thus combines
to make a notable addition to the family as a whole, and indicates that
my former estimates of its numerical importance were by no means
exaggerated.

Among the most interesting of these novelties are those which have
been derived from dipterous insects, since they not only enlarge our
systematic knowledge of new or little known genera, but illustrate in
a striking manner the curiously variable relation of these parasites to
such soft-bodied hosts. That so considerable a number of species were
found on Diptera is in a great measure due to Professor Dahl, who
called my attention to the fact that some of the small flies collected by
him at Ralurn in New Pomerania, near New Guinea, were parasitized,
and I was thus led to make a careful examination of the whole collection

398 PROCEEDINGS OF THE AMERICAN ACADEMY.

with the results hereafter indicated. These dipterous parasites alone are
included in the present paper ; but descriptions of the remaining novel-
ties will be published shortly. A set of duplicate preparations of all the
material found at Berlin has been made, and will be deposited there as
soon as the illustrations for the Supplement to my Mouograph, which is
in preparation, are completed.

A majority of the following forms belong, as will be observed, to
Stigmatomyces, a distinctly dipterophilous genus, which must certainly
prove very large and widely distributed. That it is at the same time
very difficult from a systematic standpoint is evident from the material
studied, and few single characters, even of the appendages, seem to be
wholly reliable. In describing the species the two superposed cells
above the foot are regarded as constituting the receptacle, the upper
bearing the stalk-cell of the perithecium terminally, and that of the ap-
pendage laterally, in mature individuals ; the appendage proper consist-
ing of a more or less distinctly differentiated basal cell, which may or
may not bear antheridia like the series of fertile cells superposed above
it; the antheridia, though often single, are more often asymmetrically
paired. The American species were obtained from small flies in part
collected for me by Mr. W. T. Clarke at Berkeley, California, and in
part by myself at Kittery Point, Maine, or in the vicinity of Cain-
bridge during September, for the determination of which I am indebted
to the kindness of Mr. D. W. Coquillett. The remainder, with the ex-
ception of the Bingle African species on Diopsis, were all obtained from
the Ralum collections of Professor Dahl already referred to.

Stigmatomyces rugosus nov. sp.

Venter of the perithecium dark amber brown, roughened by about ten
transverse more or less irregular and sometimes anastomosing darker
ridges formed by irregular wart-like elevations; evenh oval or elliptical,
and abruptly distinguished from the rather stout neck, which is usually
brut outward and about equal to it in length or somewhat shorter, dis-

tally distinctly enlarged, especially posteriorly; the tip beyond this en
largement abruptly somewhat narrower; the apex asymmetrical, the
three posterior lip-cells forming three corresponding projections, rounded
or bluntly pointed and more prominent than the bilobed papilla formed
below them by the anterior lip-cells. Stalk-cell of the appendage small,
subtriangular, amber brown, abruptly prominent below the relatively

large dark brown basal cell, which, though narrower, nearly equals it

in Bize, may <>r may nol bear antheridia, and has a well-marked annular

THAXTER. NEW LABOULBENIACEAE. 399

thickening on the inner side of its wall at the base ; the fertile cells above
it, four or five in number, bearing the rather large antheridia in pairs ;
the series becoming obliquely lateral or external, the free necks strongly
curved outward. The cells of the receptacle nearly equal, or the upper
larger; the basal cell tapering to the foot and distally slightly broader
than the subbasal cell. Spores about 40 X 4 //.. Perithecium : venter 72
X 45 jx ; neck 62-72 X 15-18 /u. Appendage proper 60-70 fx, stalk-
cell 18 (i. Receptacle 90-100 X 20 fi. Total length to tip of perithe-
cium 250-290 fi.

On the legs, thorax, and abdomen of a minute fly. Berlin Museum,
No. 1296. Ralum, New Pomerania.

A form from the same locality occurring on a small blackish fly, and
also characterized by a roughened perithecium, differs in several points
from that above described ; but a description of this, as well as of a
closely allied form from Kittery Point, Maine, is withheld for the present.

Stigmatomyces Diopsis nov. sp.
Colorless or slightly yellowish. Venter of the perithecium long-oval
or elliptical, pale reddish amber, rather abruptly distinguished from the
paler neck, which tapers but slightly, except at its base, is straight or
slightly bent, and traversed by four broad longitudinal ridges which are
corrugated by about six successive elevations and depressions ; a seventh
distal elevation, larger and more prominent than the rest, is present just
below the tip, which is abruptly narrower and slightly curved ; the apex
asymmetrical, the posterior lip-cells forming a tripapillate prominence,
the middle papilla larger and more prominent ; the anterior lip-cells form-
ing two small lateral papillae placed side by side in such a position that
the apex appears to be laterally notched. Appendage erect or some-
what divergent, straight or slightly curved backward, the stalk-cell more
than twice as long as broad, and more than half uuited to the subtriangu-
lar stalk-cell of the perithecium, distally constricted at its junction with
the well differentiated squarish amber-brown sterile basal cell of the ap-
pendage proper; the eight or nine fertile cells above bearing for the
most part two antheridia each, the series of antheridia external in the
mature types. Spores about 40-45 X 5 fi. Perithecium : venter 80-87
X 50 [x ; neck 72-82 X 18 /x. Appendage proper 70-75 /x, stalk-cell
25-30 ju. Receptacle 75 fx. Total length to tip of perithecium 270-

290 fx.

On Diopsis sp., Berlin Museum, No. 860. Bismarkburg, Togo, West
Africa. On the upper surface of the abdomen near the tip.

[00 PROCEEDINGS OF THE AMERICAN ACADEMY.

Stigmatomyces Scaptomyzae nov. sp.

Venter of the peritliecium becoming reddish amber brown, nearly iso-
diametric, becoming distally enlarged; the nearly hyaline neck very ab-
ruptly distinguished, .slender, straight, or curved, its lower third sometimes
narrower; the tip hardly or not at all differentiated; the apex asymmet-
rical, the anterior lip-cells forming two lateral papillate slightly divergent
protrusions, the posterior lip-cells forming two similar protuberances
above them, between which a slight projection may or may not be pres-
ent. Stalk-cell of the appendage elongate, very abruptly broader than
the very small deep brown squarish infertile basal cell; the fertile cells
u>ually five in number, the antheridia with short curved divergent necks
and produced in pairs, except the terminal one, which is conspicuously
spiniferous, the whole series usually obliquely external. Receptacle
hyaline, the basal cell mostly larger, longer, tapering below. Spores
about 36 x 3.5 /a. Perithecium: venter 90-100 x 36-44 /a; neck
108 X 10-15 fx. Appendage 47-55 /i, the stalk-cell 25-30 /x. Receptacle
65-100 /m. Total length to tip of perithecium 300-325 fi.

I in the abdomen and legs of Scaptomyza graminum Fallen. Kittery
Point, Maine, vicinity of Cambridge, Mass., Berkeley, California.

Stigmatomyces Limnophorae nov. sp.

Venter of the perithecium relatively small, amber brown, the wall-cells
becoming powdered by a darker maculation and separated by u corre-
Bponding number of well-defined unmodified longitudinal ridges which
run somewhat obliquely and end, not abruptly, at the base of the neck:
tin- latter generally slender, strongly bent throughout or even recurved,
abruptly differentiated, sometimes of less diameter than the tip, which is
distinguished from the rest of the neck by an abrupt enlargement more
prominent anteriorly ; the apex (in the not wholly mature types) un-
modified, blunt, slightly oblique. Stalk-cell of the appendage rather
prominently rounded externally, but not protruding abruptly below the
basal cell, which nearly equals it in length and is slender. Blightl) larger
distally, if- base hyaline, its wall, which is dark amber brown above, be-
coming gradually thicker, so that the lumen of the cell is attenuated below,

distally bearing tWO antheridia ; the fertile cells above it. usuallv seven

in number, forming a series outwardly recurved, the terminal cell appar-
ently sterile, the two -mall cells 1m 'low it bearing each a Bingle antheridium,
while the remainder bear two: the antheridia with short, broad, slightly
recurved necks. Receptacle relatively large, hyaline, the basal cell ta-

THAXTER. — NEW LABOULBENIACEAE. 401

pering slightly downward, the subbasal cell slightly longer and much
broader distally. Perithecium: venter 55 x 30 ti; neck 75 X 10/x.
Appendage about 75-80 tt, stalk-cell 28 /x. Receptacle 110 x 25 p. Total
length to tip of perithecium 250-275 p..

On the inferior surface of the abdomen and at the base of the posterior
legs of a species of Limnophorus. Berkeley, California.

Stigmatomyces constrictus nov. sp.

Venter of the perithecium dark amber brown, subrectangular, or more
or less inflated ; the short stout neck about equal to it in length, very
abruptly distinguished beyond the four rounded elevations which mark
the distal ends of the brown wall-cells of the venter, subcorneal, with a
considerable submedian enlargement often more prominent posteriorly ;
the tip often tapering to the five-papillate apex, the middle posterior papilla
blunt and more prominent, the other four nearly symmetrical. Stalk-
cell of the appendage often suboblong and externally prominent through-
out its length ; the basal cell narrower and longer, separated from it by a
rather deep consti-iction and bearing one or two antheridia distally ; while
above it the two remaining fertile cells are very small, each bearing two
antheridia; the series surmounted by a spiniferous antheridium, all the
antheridia relatively large and almost free. Receptacle hyaline, its basal
cell more or less elongate, tapering to a narrow base, a rectangular distal
thicker walled portion separated by a thin incomplete septum ; the sub-
basal cell much shorter, more or less abruptly and prominently inflated
at its base, sometimes slightly also at its distal end, and having a more or
less well defined median constriction, below which the inflated base may
be separated by a thin partial septum. Perithecium : venter 54 X 30-
40 /j.; neck 44-55 X 18 /x. Appendage 43-50 /x, the stalk-cell 18 tt.
Receptacle 70-90 X 22 p. Total length to tip of perithecium 200-300//,
(those on the tips of the legs much smaller, 180-200 p.).

On the legs and abdomen of a small fly. Ralum, New Pomerania.
Berlin Museum, No. 1294.

Stigmatomyces humilis nov. sp.
Venter of the perithecium amber brown, slightly inflated through' nit
and slightly asymmetrical ; the neck rather abruptly distinguished, con-
colorous, but paler distally, generally shorter than the venter, stout,
tapering to the blunt, hardly differentiated apex; about one third of its
length taken up by the tip, which is distinguished from it by a slight
broad constriction ; the outer basal cells subequal and irregularly prumi-
vol. xxxvi. — 26

402 PROCEEDINGS OF THE AMERICAN ACADEMY.

nent. Appendage relatively rather Blender, very long, sometimes extend-
ing nearly to the middle of the neck of the perithecium, the stalk-cell
separated by a Blight constriction from the basal cell, which is relatively
large, the annular thickening about the base on the inner side of its wall
unusually well developed, amber brown, bearing two antheridia ; the sub-
basal cell almost a- large, bearing two antheridia, the two successive cells
above it smaller and bearing each a single antheridium ; the series com-
pleted by a Bingle terminal antheridium; the antheridial necks rather
slender, and tapering, Bomewhat appressed. Receptacle short, stout, the
cells subeipial. Spores about 28 x 3 ft. Perithecium: venter 46-55 X
32-37 /t; neck 15-47 /a. Appendage 05-75/*, the stalk-cell 18 fx. Re-
ceptacle 55 /x. Total length to tip of perithecium 175 /x.

On tin' superior surface near the tip of the abdomen of a muscid some-
what larger than the other hosts from Ralum, New Pomerauia. Berlin
Museum, No. 1287.

Stigmatomyces dubius dot. sp.

Amber brown with the exception of the receptacle and the stalk-cell
of the perithecium. Venter of the perithecium slightly inflated, rela-
tively small, not abruptly differentiated from the broad neck, winch
gradually enlarges distally below the rather abruptly tapering, slightly
bent tip; the middle of the three posterior projections from the lip-cells
larger and longer than the others and bent over so as to overlap the
anterior lip-cells, which are curved abruptly toward it; the two 'atrial
posterior projections prominent beyond the base of the middle one, rather
Blender, and slightly curved inward. Stalk-cell of the appendage distally
darker, abruptly prominent below the basal cell, which is small. squarish,
and deeper brown; the rest of the appendage, which is unusually long,
apparently proliferous above the spiniferous cell, extending beyond the
venter of the perithecium, is made up of about eight cells, which bear
rather long antheridia in pairs, their necks appressed usually in a lateral
series. Receptacle relatively large, hyaline, the subbasal cell much
longer and broader than the basal cell, which tapers but slightly to the
small foot Spores 30 ■ •">.">„. Perithecium: venter 58 X 40 p ; neck
110 ■ 25 fu Append:,-,- -ii 95 ,,. Btalk-cell 25-32 p. Receptacle 145-
185 • 25 -80ft. Total 1 mgth 350-875 p..

I >n a fly with monstrously developed anterior legs resembling those of
Ochtheria mantis. Ralum, New Pomerania. Berlin Museum, No. 1281
and 1298. On the head and at the base of the posterior legs.

THAXTER. NEW LABOULBENIACEAE. 403

Stigmatomyces gracilis nov. sp.

Form long and slender. Venter of the perithecium amber brown,
relatively large above its narrow base, more or less inflated, often more
distinct distally ; the neck usually straight, sometimes curved, nearly hya-
line as a rule, and abruptly distinguished ; the tip abruptly but slightly
narrower above a prominent and usually symmetrical inflation ; the me-
dian posterior projection of the lip-cells erect, larger, and slightly more
prominent than the two lateral ones, which diverge slightly and are nearly
symmetrical with the two anterior ones. Stalk-cell of the appendage
slightly prominent distally below the dark amber-brown basal cell, which
may be more than half as large, bearing one or two antheridia ; the re-
maining cells four in number, relatively large, except the fourth, which
bears a large, curved, conspicuous spine below the base of the terminal
autheridium ; the antheridia in pairs, lateral or obliquely external, the
necks short, becoming pointed and slightly divergent. Receptacle usu-
ally rather long and slender, straight or curved, hyaline ; the two cells
about equal, or the upper larger and distally often broader than the com-
bined diameters of the cells above it. Spores 45 X 3.5 //,. Perithecium :
venter 85-90 X 30-40 //. ; neck 100-110 X 16 fi (the enlargement X
20 /a). Appendage 70-75 fx, stalk-cell 18-25//.. Receptacle 90-125 x
18-20 fi. Total length to tip of perithecium 250-360 fi.

On the same host with S. dubius. Ralum, New Pomerania. Berlin
Museum, No. 1298. Near the tips of the posterior legs.

Stigmatomyces proboscideus nov. sp.

Venter of the perithecium amber brown, sometimes more than twice
as long as broad, usually but slightly inflated, often more so distally ; the
neck lighter brown, rather abruptly distinguished, relatively very stout,
elongate, nearly isodiametric, usually curved throughout ; the short tip
abruptly somewhat narrower, the apex broad and blunt without well de-
veloped elevations. Stalk-cell of the appendage brown, relatively small
and short, slightly prominent distally ; the basal cell broader than long.
the five fertile cells above it rather short and stout, the series curved
sidewise, the antheridia lateral in pairs. Spores about 30 X 3 fx. Peri-
thecium : venter 75-95 X 32-36//; neck 135-185 X 18-22/a. Appen-
dage 55-72 /*, stalk-cell 18 /x. Receptacle 110-125 X 29 fx. Total length
to tip of perithecium 400 fx.

On the abdomen of a small fly. Ralum, New Pomerauia. Berlin Mu-
seum, No. 1288.

4U-4 PROCEEDINGS OF THE AMERICAN ACADEMY.

Stigmatomyces Hydrelliae now sp.

V. -liter of the perithecium umber brown, oval, the wall-cells becoming
separated by well-defined, slightly oblique longitudinal broad ridges,
which become broader distally where they end abruptly; the neck pale,
well distinguished, its middle third prominently inflated, more so poste-
riorly, ami separated from the usually abruptly bent tip by a constric-
tion ; the apex rounded, one of the (lateral?; lip-cells forming a slender,
bluntly pointed, well-defined free projection. Stalk-cell of the appendage
sub-triangular, somewhat prominent below the basal cell, which nearly
equals it in length, sterile ; tin- fertile cells above it nearly equal, bearing
rather large, apparently single, antheridia, with stout, straight necks, the
series ending in a terminal .--piniferous antheridium. Receptacle hyaline,
the two cells Dearly equal in length, the lower tapering below, the upper
broader inflated, its diameter greater than the base of the perithecium
ami stalk-cell combined, bo that the latter region appears to be con-
Btricted. Spores 28 • _//. Perithecium : venter 50— 55 X 33-40 /& ; neck
40— 43 /a. Appendage .j<»/x, the stalk-cell 18 ft. Receptacle 55-G5 X
I'U-L'^u. Total length to tip of perithecium 150— 185 /x.

On the superior surface of tin- abdomen, sometimes on the legs of
HydreUia sp. Kittery Point. Maine. Occurring in scattered groups.

Stigmatomyces purpureus nov. sp.

Hemming wholly Buflused with purple. Venter of the perithecium
inflated toward the base, tapering distally ; the four wall-cells separated
by a corresponding number of prominent longitudinal ridges, rounded in
■ hi, which run spirally, making in well-developed individuals a whole
half turn about tin- venter, and becoming sometimes lobulated through
tin- presence of successive constrictions and enlargements ; neck not
abruptly distinguished, except by the abrupt elevations which form the
terminations of the longitudinal ridges of the venter, rather Blender, an
abrupt posterior subterminal elevation preceded bj a slight constriction,
the tip distally quite hyaline ] the apex becoming furcate through the
presence of an anterior (shorter) and a posterior projection. Stalk-cell

of the appendage relatively small, but sliglnh prominent below the basal

Cell, which ifl nearly as long, Sterile, and, as a rule, followed bv three
cells hearing antheridia singly or in pair-, the terminal one Bpiniferous.

1; eptaele usually straight, the cells nearly equal or the upper larger.

■ ,i. Perithecium: venter 80-100 X 15 50 /ttj neck 80-

- : „. Appendage 55 ft, the stalk-ceil is,,. Receptacle 100-120 /a.

1 i . h to tip ot perithecium 200 325 u.

THAXTER. — NEW LABOULBENIACEAE. 405

On all parts of Scatella stagnalis Fallen. Kittery Point, Maine, and
vicinity of Cambridge, Mass., September. Fully developed individuals
with the typical structure are uncommon, a majority of the numerous
specimens examined having the color dull or paler purplish, the ridges
less well defined, without lobulations and with less than a half twist ; the
neck and apex hardly, if at all, modified. The same host is infested by
an amber-brown form which may prove a mere variety of that above
described, being scarcely distinguishable structurally from the less well-
marked individuals of this species, the type form of which is, from its
remarkable color and the structure of its perithecium, one of the most
peculiar members of the genus.

Stigmatomyces spiralis nov. sp.

Venter of the perithecium relatively long and slender, flask shaped, or
more often but slightly if at all inflated, the granular wall-cells distin-
guished by a corresponding number of abrupt, narrow, longitudinal prom-
inent ridges, which become minutely roughened, and are spirally twisted
so as to describe a full half turn ; the neck concolorous, distinguished by
the abruptly elevated and abruptly broadened terminations of the longi-
tudinal ridges of the venter, as long as or slightly shorter than the venter,
slightly curved or sometimes straight, nearly cylindrical or slightly taper-
ing ; the tip slightly but abruptly narrower, relatively short, somewhat
asymmetrical ; the apex nearly symmetrical, four papillae being arranged
about a somewhat more prominent central projection. Appendage rather
short and stout, distinctly broadened in the middle, the stalk-cell stout,
the basal cell half as large, or less, and fertile ; the series of six to eight
fertile cells above it surmounted by a single antheridium, and distin-
guished by slight successive constrictions, broad and much flattened, each
bearing a single antheridium, the fifth furnished with a very sharp spine ;
the antheridia forming a usually lateral series, their necks becoming
strongly curved. Receptacle elongate, slender, becoming brownish or
yellowish, the upper cell often more than twice as long as the basal.
Spores 22 X 2.0 fi. Perithecium: venter 90-165 X 35-47 fx ; neck 90-
160 X 17 ft (the tip 25-30^). Appendage 40-50 /i, the stalk-cell 15 ,/.
Receptacle 100-250 X 15 fi. Total length to tip of perithecium 35<i-
600 ^ (average 500-550 /x).

On Hydrina sp., Kittery Point, Maine. Usually on the upper sur-
face of the thorax, less ofteu on the legs and elsewhere.

406 PROCEEDINGS OP THE AMERICAN ACADEMY.

Stigmatomyces Limosinae nov. sp.

lVrithecium amber brown, the venter slightly inflated, the neck not
abruptly distinguished, tapering slightly; the tip usually abruptly nar-
rower, the posterior lip-cells forming an inconspicuous irregular truncate
or rounded bilobed projection somewhat more prominent than a similar
projection funnel by the anterior lip-cells; basal cells relatively very
large, forming a short, well-delined stalk, hyaline or colored above, often
carrying the base of the perithecium beyond the tip of the appendage,
and consisting of an inner cell next the appendage and two superposed
outer nii'-. the lower of which (secondary stalk-cell) is smaller; the stalk-
cell below these wholly united to the stalk-cell of the appendage, rather
stout and short, separated from the cells above it by a horizontal septum,
which maj be Blightly oblique or (as in the California variety) strongly
oblique, in which case the secondary stalk-cell extends downward beside
the stalk-cell bo that only the lower third or quarter of the latter is free
externally. Stalk-cell of the appendage relatively large, as long as or
often longer than that of the perithecium and about half as broad, usually
bulging externally, its outer margin usually curved symmetrically from its
base to thr base of the basal cell ; the latter relatively small, deep amber
brown, half as long as broad, pointed distid ly between the antheridium
which arises from its inner side and the base of the first fertile cell above it,
which, with the other fertile cells, are large and prominent, thick walled,
much flattened, and obliquely superposed, distinguished by rather deep
constrictions, seven to ten in all, or rarely more (seven to fourteen in the
California^ form), the original number being increased by the terminal
proliferation of the appendage; the antheridia borne on the inner side of
the appendage, their very long but not abruptly differentiated necks ex-
tending obliquely upward, appressed in a double series; the upper anthe-
ridia often infertile, becoming septate and irregularly swollen. Receptacle
relatively short, the two cells nearly equal. Spores 28 x 3 /x. Peri-
thecium: venter 50-90 < tO-54/t; neck 90-125 X 15-1. S^ ; stalk (basal
.•elk only | 72-100 x 25-35 ,». Appendage 60-100 M, stalk-cell 30-4."> ,,.
eptacle 70-75 < 22 ,/. Total length to tip of perithecium 250-300 fi.
Specimens on legs often much smaller.

< hi Limonna fontinalis Fallen. Kittery Point, Maine, vicinity of
Cambridge, Mass., Berkeh y. California. Fsually in a dense tuft on the
side or near the tip (inferior) of the abdomen and near the base of the
posterior pair of legs. The Californian material, from two specimens of
the boat, differs constantly from the abundant New England material as

THAXTER. NEW LABOULBENIACEAE. 407

noted in the description, as well as from the fact that the venter of the
peritheeium is longer and less distinctly inflated, while its apex shows no
perceptible modification of the lip-cells.

Stigmatomyces Papuanus nov. sp.

Venter of the peritheeium dark amber brown, relatively small and
rather prominently inflated, oval to elliptical ; usually not abruptly dis-
tinguished distally from the hyaline or yellowish neck, which in well-
developed specimens is very elongate, tapering very gradually, in others
shorter and stouter ; the tip clearly distinguished (abruptly so in the
shorter forms), subconical, the posterior lip-cells forming a narrow, sub-
truncate, slightly recurved apical projection beyond the two laterally
placed, papillate, slightly divergent projections of the anterior lip-cells ;
the basal cells forming a short, stout stalk, separated from the stalk-cell
by an oblique septum. Appendage relatively small, resembling that of
the S. Limosinae in general form, the fertile cells not more than five
or six in number, the upper ones separated by constrictions which may
be obsolete between the lower ones. Receptacle relatively short, the cells
subequal, yellowish. Spores about 20 X 2 (i. Peritheeium : venter 50-
55 X 40 /j. ; the neck 90-290 X 20 //. ; the stalk 35-45 x 33-36 p.
Appendage, 35-45 /a, the stalk-cell 22-30 X 14-17 /x. Receptacle 55-
72 /a. Total length to tip of peritheeium 400-485 /j. A few specimens
on the legs much smaller.

On three small flies of different species allied to Limosina. Ralum,
New Pomerania. Perhaps a variety of S. Limosinae.

Arthrorhynchus Cyclopodiae nov. sp.

Becoming tinged with brownish yellow except the hyaline stalk-cell of
the peritheeium. Peritheeium nearly straight and symmetrical, slightly
inflated, usually distinctly constricted in the region of its very small basal
cells just above the very large hyaline stalk-cell, which may nearly equal
it in length and diameter and is often somewhat enlarged distally : the
venter comprising the lower two-thirds, not clearly distinguishable from
the neck, which tapers slightly and almost symmetrically, the tip fairly
well distinguished above a more or less distinct enlargement, from which
it is separated by a slight constriction ; the apex consisting of a crown
of four nearlv symmetrical, distinctly tridentate, erect, or very slightly
divergent projections, which are subtended by a corresponding number of
slight elevations, the middle lobe of each projection more prominent than

408 PROCEEDINGS OF THE AMERICAN ACADEMY.

the lateral and like them bluntly rounded. Receptacle consisting of two
Bmall cells, tin- lower twice as large as the upper, which gives rise distally
to the Btalk-cell and bears the free appendage laterally; the foot an
unmodified cell which penetrates the host, dividing below into a very
copiously branched system of Blender, sinuous, rhizoidal hyphae. Ap-
pendage consisting of a dumbbell-shaped, free Btalk-cell, the basal half-
rounded or Battened, brownish, somewhat larger than the distal portion,
which is deeper brown, flattened and inflated, connected by a narrow
hyaline isthmus (the lumen of which may become almost obliterated) with
the lower half, and mostly broader than the hase of the basal cell of the
appendage, which is infertile, Bubrectangular, or somewhat inflated, slightly
longer than broad, the lower half of the walls becoming conspicuously
modified bj a prog • thickening from above downward, the thick-

1 portion deeper brown; the remaining cells of the appendage three
to four in number, brownish, successively smaller from below upward,
giving the organ a characteristically tapering habit; the two lowest of
these cells usually relatively Bhorter, and hearing each three to four
antheridia Bide bj side, distally and externally; those above relatively
longer and naiTOW< r and producing fewer antheridia. the terminal one
Bpiniferous. Antheridia with sh-nder curved necks. Spores 60-65 x

,. Perithecium: venter 325-350 x 70-90 /*; the Btalk-cell 220-
250 • 75-80/t. Appendage, 100-1 10 /x, the stalk-cell 35-40 X 30-35/*

(the Upper half X 28-30 /' I. Receptacle o."j-7o X 45-50 /A.

On the abdomen of Qyclopodia macrant Speiser. New Pomerania.
Berlin Museum, No. 85 1.

The original name given to this genus in l.s."»7 by Kolenati is here

iued in preference to the much later one applied to it by Peyritsch

in I - nee however absurd and scientifically worthless the original

zoological descriptions of these forms may be, there has never been the

slightest question as to the generic identity of the organisms studied by

■ two authors. Neither the descriptions nor the figures given by

Kolenati and Diesing are, however. Bufficient to render a specific deter-
mination possible, so that the name given by Peyritsch to the European
<.| tin- '." nib. although it is undoubtedly a synonym of A. Die-
;<i Kol. or A. Wutrwnbii K"|.. or more prohahlv of both, may

properly be retained. The new forms here described are very closely
allied, differing chiefly in the details of structure in the appendage and

the tip of the perithecium. lint are very different from Arthrnrln/itrlnis

M •■ rial in my possession obtained from specie- of .Xi/c-

teribia, from Europe, must, I think, he referred without question to the

THAXTER. — NEW LABOULBENIACEAE. 409

last named species ; although the conformation of the tip of the perithe-
cium has apparently been incorrectly reproduced in Peyritsch's plate.

Arthrorhynchus Eucampsipodae nov. sp.

Hyaline throughout. Perithecium straight or distally slightly curved,
tapering gradually from the middle, or lower, to the broad tip ; the apex
consisting of a slight median projection surrounded by a crown consisting
of four slightly shorter, broad, blunt, distinctly divergent projections, which
show indistinct marks of lobing and are symmetrically placed ; the stalk-
cell about one half as long as the perithecium or less. The basal cell
of the appendage constricted in the middle as in the preceding species,
the lower half irregularly rounded and four or five times as large as the
upper half, which is very small, colorless, and less than half as wide as the
cell above it ; the fertile cells three in number, the lower bearing four or
(?) five antheridia, the upper three in addition to the terminal one, which
is furnished with a short hyaline basal spine ; the necks of the antheridia
large, tapering, divergent. Receptacle as in the preceding species.
Spores about 45-50 X 4 /x. Perithecium: venter 250-325 X 65-75 it;
the stalk-cell 110-150 X 55 it. Appendage, 75-90^, the stalk-cell
35 X 25 (the upper half X 10^).

On the abdomen of Eucampsipoda Hyrtli Kol., Egypt. Berlin
Museum, No. 855.

Rhizomyces gibbosus nov. sp.
General habit more or less sigmoid. Perithecium amber brown, con-
colorous, with its relatively large basal cells, from which it is hardly dis-
tinguished, asymmetrically inflated, bent, and tapering somewhat distally ;
a subterminal abruptly rounded enlargement, beyond which the short
asymmetrical tip is clearly distinguished, bearing a large two-celled out-
growth posteriorly, the lip-cells being otherwise unmodified : stalk-cell
hyaliue, variably, sometimes greatly elongated, separated from the basal
cells by a more or less distinct constriction. Appendage nearly hyaline,
except the small deep brown sterile basal cell, the remaining cells, three
to seven in number, bearing short one- to two-celled brunches distally
and laterally on which the free flask-shaped antheridia are borne singly
or several together. Receptacle short and stout, the upper cell several
times as large as the basal cell, which appears to penetrate the host di-
rectly by means of a rhizoidal apparatus. Spores about 35 X 3//. Peri-
thecium, including basal cells, 85-108 X 30-36 /x; the stalk 60-160 X
18-20 fi. Appendage 65-110^. Total length to tip of perithecium
180-325 p.

410 PROCEEDINGS OF THE AMERICAN ACADEMY.

< >n the upper mii far,- near the tip of the abdomen of a species of
Berlin Museum, No. 850. Tanga, Africa.

CERAIOMYCES nov. gen.

Structure of perithecium as in Laboulbenia, its stalk-cell united to the

the free stalk-cell ol the appendage, which hears a well differ-

entiated basal cell terminally, from the end of which are borne antheridial

branches, the successive cells of which produce terminally either succes-

ondary branchlets or antheridia or both, much as iu Laboulbenia.

\i • ptacle two-celled.

Ceraiomyces Dahlii nov. sp.

Perithecium large, blackish brown, with an olive shade, becoming opaque,
usually Blightly curved, tapering gradually to the slender undifferentiated
tip; the anterior lip-cells forming two appressed hyaline-tipped finger-
likr projections; the base very broad, translucent, dull brownish, bulging
jpicuously below the venter, especially on the left side; the stalk-cell
small, nearly isodiametric, united on it- inner side to the base of the stalk-
cell of the appendage. The latter free, though often in contact with the
base "f the perithecium, dull blackish olive, outwardly inflated, narrower
terminally where it bears the characteristically differentiated basal cell of
tli<' appendage, which becomes almost opaque and is Bomewhal flask- or
bottle-shaped with a rounded extremity, from which, typically, two diver-
gent branches arise which in turn may branch one to three times suh-
dichotomously ; the long Blender flask-shaped antheridia borne, one to two
ther, distally from the successive cells. The basal cell of the recep-
tacle nearly Bpherical, penetrating the host by a long filament which is
Blender except for an enlargement immediately below the integument of
the host, Bimple at first but becoming more or less copiously branched ;
the upper cell very large and elongate. Spores about 30 x 3//, Peri-
thecium 275— 310 • 55 60/*; the base, including the stalk-cell, 68— 72 X
- i. Appendage 75-85// (the basal cell L8 x 12 /a), the stalk-cell
10-45 • 18 -22 fu Receptacle 175-240 x 3o H. (the basal cell 20-22//).
I Lai length to tip of perithecium I11" 675 /t, average 550 /t.

On various parts of a small flower fly. Ralum, New Pomerania.
in Museum, Nos. 1283 and 1298. Occurring more often on the head.
where it might be mistaken for a dipterous antenna.

Dimeromyces coarctatus nov. sp.
MaU Individual. Receptacle Dearly hyaline, consisting of usually
ip rposed cells the npper separated by a dark-colored constriction

THAXTER. — NEW LABOULBENIACEAE. 411

from a short, simple, two- to three-celled hyaline or brownish appendage.
The antheridia usually two, seldom three, borne singly from the succes-
sive cells of the receptacle, from which they are separated by a small basal
cell ; the venter having an external depression and not abruptly distin-
guished from the stout curved neck. Receptacle 35-45 X 6-7 p. Ap-
pendage 25-50 fj,. Antheridia 18 X 5 /x.

Female Individual. Receptacle consisting of a large basal cell about
twice as long as broad, bulging so as to form a rounded base which pushes
the small brownish-black foot to a lateral or sublateral position ; the re-
maining cells, usually eight or nine in number, separated by horizontal
septa and superposed in a simple series ; the lower cells greatly flattened,
those above somewhat less so, the series ending in a somewhat abruptly
narrower terminal cell, which is more than twice as long as broad, subcy-
lindrical, its extremity rounded symmetrically and bearing a short, simple,
usually four-celled terminal brownish appendage, which is distinguished
by a constricted dark basal septum and terminated by a somewhat inflated
lighter larger cell, which becomes characteristically disorganized on one
side, so that the appendage appears to end in a slender curved projection.
The remaining cells of the receptacle producing single appendages or peri-
thecia, except the basal and sometimes a subbasal cell. The uppermost
of these secondary appendages arises from the inner side of the subcorneal
subterminal cell of the receptacle, occupying a position in the median line
between the primary appendage and the base of the first perithecium, and
consists of a short subcorneal basal cell, from the narrow extremity of
which the simple, several-celled terminal portion is distinguished by a
constricted dark septum ; the remaining appendages laterally divergent
on opposite sides in such a way as to appear paired, usually three on each
side, each consisting of a rather long basal cell inflated along its upper
side so as to appear more or less geniculate, concolorous with the recep-
tacle, its narrower extremity suffused with dark brown, distinguished
without constriction by a dark septum from the simple terminal portion,
which is usually five-celled, more or less strongly recurved, brown, its ter-
minal cell becoming inflated and undergoing gelatinous degeneration on
the lower side, which causes it to appear split in two, the hook-like upper
half of the cell alone persisting in some individuals. Perithecia yellow-
ish, distally brownish, one, rarely two, in number; the first always arising
from the cell immediately below that which bears the upper secondary
appendage, the second, when one is present, replacing one of the append-
ages lower down; consisting of a symmetrically inflated venter, which
tapers gradually downward, passing into the short stalk ; a short neck

41- PROCEEDINGS OF THE AMERICAN ACADEMY.

rather abruptly distinguished, deeper brown below, its tip inflated below
four terminal projections, three or two of which are in the form of rounded
papillae of unequal size, and one or two of which are pointed and much
more prominent. Spores 12 X 3.5 fi. Perithecium, including the stalk,
which is continuous with it, 125 X 2<>-.'!5 «. Receptacle to base of pri-
mary appendage 50-75 u. Secondary appendages about 75 fj.. Total
length t<i tip of perithecium 150— 180 /a.

I)iii-el\ crowded on the inferior surface of the abdomen or rarely on
the legs <>f a small pale fly, remarkable for a prominent black spur-like
bristle on the posterior legs. Ralum, New Pomerania. Berlin Museum,
No. L282.

Dimerornyces rhizophorus now sp.

Male Individual. Receptacle consisting of a basal cell which penetrates
the ho>t directly without a differentiated foot, and two to three super-
posed cells above it. each of which usually bears an antheridium, the
upper terminated by a short, pointed, slender cell. The antheridia
rather short and -t< »u t . with short, stout necks. Receptacle about 50 X
v a. Appendage 12 x 3.5 jx. Antheridia 25 x 9 /x.

Female Individual. More or less deeply tinged with amber brown.
Receptacle amber brown, consisting of six superposed cells, the small
basal cell, hardly visible above the integument, penetrates the host
directly by means of a very large, abruptly furcate rhizoid, the two cells
above it similar, broader than long, bearing each an appendage consist-
ing of a basal cell bent toward the receptacle, darker and narrower
distally, and separated by a dark septum from the three-celled terminal
portion, which is straight or slightly curved, larger toward the middle, the
smaller terminal cell becoming partly disorganized. The next (fourth)
cell of the receptacle bears the single perithecium ; the distal terminal
cell longer and narrower, and terminated by a short, pointed, one- some-
times two-celled primary appendage (similar to that of the male in-
dividual), from which it is separated by a constriction; the subterminal
C6 1 narrower distally, producing on its inner side an appendage similar
lo those below it, but straight and somewhat shorter. Perithecium with
a short Btoul stalk rapidly expanding into the asymmetrically inflated
deeper brown venter of the perithecium; the neck very short and

abruptly distinguished | the tip relatively large, four-lobed, inflated with
two lateral papillate OUtgTOWths, above which the lips form a Bubconical

projection.
Spores about 25 • '■'> u. Perithecium including stalk 70-90 x 20-

THAXTER. — NEW LABOULBENIACEAE. 413

25^. Receptacle about 45 x 12 p. Primary appendage 12 (i ; second-
ary appendages 35-40 p. Penetrating rhizoidal branches 150-184 x
10-12 (i. Total length to tip of perithecium 90-110 /<.

On the inferior surface of the abdomen of a small fly. Raluuu, New
Pomerania. Berlin Museum, No. 1295.

Dimeromyces crispatus nov. sp.

Male Individual. Receptacle consisting of four superposed hyaline
cells, the basal one much longer than the rest combined ; the upper bear-
ing distally a two-celled terminal appendage, the lower cell of which is
small, the upper elongate, brownish ; the two remaining cells of the
receptacle producing each a single antheridium. Antheridia superposed,
the stalk-cell, neck, and venter well distinguished, the latter symmetri-
cally and considerably inflated, the neck slightly curved. Receptacle
50 X 8 fi. Antheridia 33 X 8-9 ^. Appendage 36 /x.

Female Individual. Receptacle consisting of usually five superposed
cells, the basal cell very elongate, slender, and hyaline ; while of the four
remaining cells the two lower are much flattened, broader than long, and
separated by oblique septa, the two upper unlike and narrower. Of
these four cells the second from below gives rise to the stalk-cell of the
perithecium, while the others by successive proliferation produce each
a branch consisting of eight or ten obliquely superposed cells ; while each
of these cells in turn jaroduces a single simple branchlet from its upper
side, originally terminal, but becoming lateral through the further pro-
liferation of the cell which bears it; the branchlets distinguished by a
slight constriction and a broad dark septum at the base, brown, curved,
distally helicoid, slightly enlarged and paler. The primary terminal
appendage thus appears as the lowest of the upper series of branchlets,
from which it does not differ in structure. Perithecia one to three in
number, the first lowest, and always formed from the second from below
of the four distal cells of the receptacle, others sometimes arising from
each of the two upper distal cells ; the stalk hyaline, long and slender, the
venter small, narrow, not distinguished from the stalk, becoming brown-
ish, distally slightly inflated, the neck short and well but not abruptly
distinguished ; the tip well differentiated, hyaline, symmetrical or nearly
so, shovel shaped or spatulate, swollen at its base, and tapering to the
broad, bluntly rounded or nearly truncate apex. Spores about 30 X
3.5/,.. Perithecium : 70-75 x 18 ft, the stalk 50-125 x 15 ft. Recep-
tacle, basal cell 185-250 X 18 p, the distal portion about 50 jx. Total

1M PROCEEDINGS OF THE AMERICAN ACADEMY.

length to tip of perithecium 5 /•■ Lateral cell series or branch) b

about 50 u long, their branchleta to tip of helix about "»0 p.

On the lega :m<l Buperior Burface of the abdomen of the same host
parasitized l»y D. coarctatu*. Ralum, New Pomerania. Berlin Museum,

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 24. — April, 1901.

CONTRIBUTIONS FROM THE ZOO LOGICAL LABORATORY OF THE
MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE.
E. L. MARK, DIRECTOR. — No. 119.

A STUDY OF VARIATION IN THE FIDDLER CRAB
GELASIMUS PU GIL AT OR Latr.

By Robert M. Yerkes.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE
MUSEUM OF COMPARATIVE ZOULOGY AT HARVARD COLLEGE.
E. L. MARK, DIRECTOR. — No. 119.

A STUDY OF VARIATION IN THE FIDDLER CRAB
GELASIMUS PUGILATOR LATR.

By Robert M. Yerkes.

Presented by E. L. Mark, February 13, 1901. Received March 4, 1901.

Contents.

Page

I. Introduction 417

II. Problems 418

HI. Methods and Material 419

IV. Results 423

1. Tables of Means, Modes, Standard Deviations, and Coefficients

of Variability 423

2. Tables of Occurrence of Right- and Left-Handedness 425

3. Polygons of Frequency 425

4. Table of Correlation of Meropodites of Chelae 438

V. Discussion of Results 440

VI. Summary 441

Bibliography , 442

I. Introduction.

This study was undertaken to get light on some minor problems of
variation, and to furnish records for future reference.

The matter of future reference seems to me of special importance, for
it is evident that with records of the variations of a race during stated
intervals available, the course and progress of evolution may be more
or less accurately traced. Davenport ('99) has emphasized the need of
the collection and filing of statistics of variation, especially of Place-
Modes ; and it was in response to his plea for this kind of work that I
undertook to determine some place-modes for one of the Fiddler Crabs
(Gelasimus pugilator) of a certain region.

In order to make the records of measurements together with my com-
putations and interpretations available for any who may, in the future,
vol. xxxvi. — 27

418 PROCEEDINGS OF THE AMERICAN ACADEMY.

desire to use them, they will be placed iu the Library of Harvard Uni-
v< rsity, subject to the order of either the librarian or the writer.

Through the kindness of Mr. Alexander Agassiz and the authorities
of the United States Fish Commission Station at Wood's Hole, Massa-
chusetts, I was permitted to occupy, during a part of the summer of
L899, one of the tables in the Fish Commission Laboratory controlled
by the Museum of Comparative Zoology. I desire to thank them for
the privilege, and also Professor II. C. Bumpus, Director of the Labo-
ratory, for facilities in collecting material and for many courtesies which
made work pleasant To Professor E. L. Mark I am indebted for
suggestions and assistauce iu the arrangement of this report; to Pro-
t' ssor C. B. Davenport I owe the knowledge and interest which led me
to the problems, and to him I am also indebted for invaluable assistance
in the prosecution of the work.

II. Problems.

The original purpose of this work was the determination of place-
modes for Gelasimus pugilator of West Falmouth, Mass. Examination
of the material for measurement and consideration of the probable value
and significance of various dimensions led to the inclusion of several
other problems. For it seemed advisable to select for the determina-
tion of place-modes such measurements as would be likely to contribute
data for problems of adult form, of development, and of function.

The male Fiddlers are either rinht- or left-handed, one chela beinjr
enormously developed, the other small. The proportions of right- and
left-handed animals will be considered later. In the females the chelae
are marly equal. One immediately asks: (1) What is the meaning of
this condition in the males? (2) What relation does the size of the
chela bear to the other body measurements ? (3) What determines
whether an animal shall be right- or left-handed ? These are some of
the problems which, it was hoped, might be at least partially solved by
a quantitative study of variation; for unquestionably this method can
supplement in important ways histological and physiological studies.

1 1 was my purpose to measure equal numbers of the two species of
Fiddlers, Gelasimus pugilator and (\. pugnaz, found in the vicinity of

W I'a Hole,* selecting equal numbers of male.' and females from each

of the two Bpecies. Bui the task of measuring was slow, and the time
then available for the work was exhausted before its completion. This

• <j. min. ix doefl not OCCUr in the vicinity.

YERKES.

VARIATION IN THE FIDDLER CRAB.

419

report, therefore, represents only a fragment of what was planned, and
what the complete solution of the suggested problems demands.

III. Methods and Material.

The following measurements were made : —

1. Frontal Breadth, the distance between the anterior marginal points
of the carapace (Figure 1, FB).

B

Figure 1. Dorsal aspect of carapace. BA, right marginal length. FB,
frontal breadth. ML, median length.

2. Median Length, the maximum length of the carapace in the median
plane. (Figure 1, ML).

3. Right Marginal Length, the distance from the right anterior mar-
ginal point of the carapace to the posterior end of the lateral ridge, just
above and between the fourth and fifth legs (Figure 1, BA).

4. Left Marginal Length.

5. Meropodite of the Right Chela, the distance between the articula-
tions on the anterior face. The proximal point of measurement, i. e.
the articulation between ischiopodite and meropodite, is on the anterior
ventral surface, and therefore does not show in Figure 2.

6. Carpopodite of the Right Chela, the distance between the anterior
articulations (Figure 2, CL).

7. Propodite of the Right Chela, the distance from anterior articula-
tion with carpopodite to articulation with dactylopodite (Figure 2, PL).

4:20 PROCEEDINGS OF THE AMERICAN ACADEMY.

s. Meropodite of Left Chela.

'.'. ( !arpopodite of Left Chela.

10. Propodite of Left Chela.

1 1. Meropodite of Second Right Leg ; points the same as for meropo-
dite of chela,

12. Meropodite of Second Left Leg.

frn'prt.

dac'pd.

i

mer'pd. carp'pd. pr'pd.

Pioi u -. ])'.r>.il surface of Right Chela, dac'pd., dactylopodite; pr'pd., pro-
podite ; carp'pd., carpopodite \ mer'pd., meropodite; ba'pd., baaipodite ; CL, length
irpopodite ; LP, length of propodite.

The above measurements are similar to those made for Carcinus
maenas by Weldoo ('Jo), but in no case identical with his. lie has,
furthermore, expressed ill measurements in terms of total carapace
length, bo comparison of our results is impossible. Dr. II. W. Rand*
has repeated Weldon's measurements on C. maenas at Wood's Hole.
His results, which have not as yet been published, will furnish, in con-
on with this study, interesting data for the comparison of Carcinus
and ( relasimus.

The measuring apparatus used consisted of a pair of spring dividers,
which could lie adjusted easily and accurately by means of a thumb-

•v. In measuring, the end of one arm of the dividers was fixed upon

the point marking one limit of the dimension to be measured and the

other arm was adjusted by means of the thumb-screw until its point coin-
cided with the other limit of the dimension. The dividers were then
transferred to a Bteel millimeter scale, above which a pocket lens was SO

inged as to give a magnification of about twelve diameters, thus en-
abling one readily to read Prom the Bcale the distance between the points
of the dividers. The Bcale was ruled to fifths of a millimeter, and by

erminillg whether the fine point Of the arm re-ted in or between
and description "f measurements, in manuscript.

YERKES. VARIATION IN THE FIDDLER CRAB.

421

marks it was possible to read to tenths. Although in case of these
measurements readings were made to tenths, accuracy to fifths is all that
can be claimed, because of the tendency to judge in favor of the marks
on the scale.

Over a thousand individuals of Gelasimus pugilator were collected,
without any attempt at selection, from a large colony on the shore of
West Falmouth Harbor, between July
24th and August 22d. The accompanying
map of the harbor and vicinity shows the
location of the colony ; two small circles
mark the limits of the ground used for
collecting.

Weak alcohol, about 20%, gave satis-
factory results as a killing agent. Sub-
stances which act very quickly, such as
hot water, formol, or strong alcohol,
cause autotomy.

Twelve measurements were made for
each of four hundred right-handed ani-
mals, and for the same number of
left-handed individuals. For these two
groups of measurements the means,
modes, standard deviations, and coeffi-
cients of variability have been found.

To avoid the possibility of misunder-
standing as to the meanings of these
terms, I shall briefly define them.* The
Mean is the sum of the products of each
class into the number of individuals in the
class, divided by the total number of individuals. The Mode is the class
containing the greatest number of individuals. The Standard Deviation
"is found by adding the products of the squared deviation-from-the-mean
of each class multiplied by its frequency, dividing by the total number
of variates, and extracting the square root of the quotient, "f The
Coefficient of Variability is one hundred times the quotient of the standard
deviation divided by the mean.

Figure 3. Map of West Fal-
mouth Harbor. Material was
collected from the region between
the small circles.

* The terms and methods given by Davenport ('99").
t Davenport ('99", p. 15).

422

PROCEEDINGS OF THE AMERICAN ACADEMY,

>

u.

:

-

-
-
—
:

- Z

—

a
%

a

'

—

ea

cc

—

1

o\

::

o

_

1

.-.

r*

/.

9 H 00

t- cn o o o

• 7 r- o cc cc

o •*; cn i-H co

as o o c» cs

a

=

7i
?i

O

r

US

—

-

•—

-

©
o
o

CO

err

r- •
CI

=

i-

i-

X

J.

--.

c

o

— ■ CO co c:

M fH <— l~

I- uO DO —

CO ci CO 00

CN

I-

-

-

CO

o

::

-

I--

CO

c;

co

O

O

CO

■

tc

—

o

CO

• 7

-r

i-

1-

1 -

CI

—

1 -

71

eo

—

77

O

ci

co

71

r.

i-

1-

•7

CO

CM

CO

r~

C:

CO

f~

7.

e
—

q
i

s

CO

i 2

CO r-i

© CN

CO CO

CO t>-

• 7

-o

71

::
::
n
I-

C5

r.
■ -.
/
—

71

c?

"*

x

lO

CO

^—

CN

<w

CN

CI

CD

co

2

H

'7

IC

'7
-

'7
77

'7

• C

3J

>7
1 -

77

1-

U3

—

1-

/

CN

71

CO

i?

oa

.7

si

o

•7

'7

'7
BO

'7

::

■7
l~

i

■7

i7

CO

'7
O

'7

J.

■7
1-

'7

l -

C5

'7

CN

CN

1-

©

V

-
-
I

I
O

-

-

i- -t

-.: /

Cl C4

7 1 JO

CO CN

I-

CO

id

oa

77

::
i -

CN

T I / '7 C7

■ 7 I- 77 '- 71

K / c cr. '7

co' cb ci co" i-^

■-

i
:i
i-

•7

'7
—
-I

71

—
CO

i

—

/

o

I-H
I-

co

—

71

ci

CO o o o
■- ia ^h co

O O I- o

O CO

CN

71
I-

—

rt

4) a> r: zz

Bd

= — -r

- y, M - hh

I S O

5 i ~

E Z

o
o

■77
O
Q,

C

•-. 4

c

E
-7?

L
&

«-,
o
o>

-9
o

o

o

-r
o
A
o

Pi

~ «

r «

o ^^
0)

r

^-

o

T

o
c

bj

-

71

0)

■3

o

c-
c

u

YERKES. — VARIATION IN THE FIDD ER CRAB.

423

IV. Results.

1. Tables of Means, Modes, Standard Deviations, and Coefficients

of Variability.

Table I. contains the results of all calculations so arranged as to facili-
tate comparison of right and left-handed animals. The first of each pair
of columns gives the values for the right-handed individuals, — indicated
at the top by "Rights;" the second, for the left-handed, — indicated by
" Lefts."

It is to be noted that no identical measurements for Rights and Lefts
(for sake of brevity we shall hereafter speak of the two types as Rights
and Lefts instead of right-handed and left-handed) are strictly comparable,
except frontal breadth and median length, for in every other case there
is marked asymmetry, and the measurements for the left side of one type
have to be compared with those for the right side of the other.

TABLE II.
Means and Coefficients of Variability arranged for Direct

Comparison.

Gelasimus pugilator (Males).

Frontal breadth

Median length

Lateral margin, great chelar side . .
Lateral margin, small chelar side . .

Meropodite of great chela

Carpopodite of great chela ....

Propodite of great chela

Meropodite of small chela

Carpopodite of small chela ....

Propodite of small chela

Meropodite of 2d leg, great chelar side
Meropodite of 2d leg, small chelar side

Mean.

Rights.

mm.
15.728
12.245
8.842
8.400
8.780
7.115
9.248
5.563
3.050
2.710
7.589
7.127

Lefts.

mm.
15.658

12.237

8.716

8.285

8.650

6.878

9.038

5.312

2.969

2.733

7.523

6.965

Coefficient of
Variability.

Rights.

mm.
7.022

7.26G

8.845

8.649

9.000

10.166

10.215

8.734

9.518

8.313

8.479

8.724

Lefts.

mm.
7.471

7.598

8.880

9.000

9.657

10.472

10.255

9.920

11.139

8.898

9.660

9.1G0

424

PROCKEIUNCS of THE AMERICAN ACADEMY.

Table II. g - the means ami coefficients of variability which are
directly comparable. In every case, with the exception oi the propodite
<>t the small chela, the means for the Rights are larger than those for the
Lefts; the coefficients of variability are without exception greater for
the Lefts, In other words, the right-handed individuals are the larger
and the less variable.

TABLE 111.
Right- and Lei i Bandedni bb.

Q. pugilator.

0. pugnax.

Hales.

Females.

Bights.

Lefts.

Lot from \V. Falmouth, Aug. 18, 1-
Lot bom W. Falmouth, Aug 22, 18

216

286

1-.".
280

00
161

85
110

Totals.
Per cent

601

61.86

of males.

I- 5

18 U
of males.

211

20.00
of all

196

TABLE IV.
Right- and Lbvt-Hahdbdness.

ii. pngfl&tor.

0, pu.'ii.ix

M .lea.

U ilea.

Females.

[fats.

U 1

Bights.

Ixft.M

I .'.! Imin j

W Falmouth, [

.Inly 16, LOO y

1'. r •• tit

|::l

1 .1

106

652

678

..:

60

of II.

Of ii-

1- 16

ill.

18

of n

61 16

of in

of all.

YERKES. — VARIATION IN THE FIDDLER CRAB. 425

2. Tables of Occurrence of Right- and Left- Handedness.

In this connection a statement of the relative frequencies of right- and
left-handedness is important. Table III. offers an analysis of two lots
of pugilator and puguax from West Falmouth. Of 9GG males 51.86%
were right-handed, 48.14% left-handed. During the summer of 1900
additional evidence was obtained on this subject. Over two thousand
Fiddlers from the same colony were examined, with the results shown in
Table IV. Comparison of these two tables (III. and IV.) indicates that
the male Fiddlers are about equally divided into right- and hit-handed
individuals. The number of females captured was always small, prob-
ably because the males remain outside to fight, while the females and
immature individuals scurry into the burrows. According to Table IV.
the percentage of female pugilators captured was four times as large as
that of female pugnax, but for this I am unable at present to offer any
explanation.

3. Polygons of Frequency.

We shall now consider the various measurements individually, and
examine the Polygons (more accurately Rectangles) of Frequency.

The polygons are all constructed thus. Each vertical column repre-
sents a class, and each of the squares in these columns represents five
individuals. The classes are separated by 0.4 mm.* In polygon No. 1 R,
for example, the first class is 13.15 mm. and contains one individual ; the
second class is 13.55 mm. and contains four individuals.

1. Frontal breadth. This is the largest and least variable dimension
of the twelve under consideration. Polygon No. I R represents the
distribution, among the various classes, of the right-handed individuals,
and No. 1L that of the left-handed. Comparison of these two polygons
shows a larger number of individuals in the modal class of the Rights
than in that of the Lefts, indicating less variability among the right-
handed animals. The curves f are both unsymmetrical. This lack of
symmetry is probably due, as polygon No. 1L especially indicates, to the
fact that the curves are really compound. In No. 1 R the modes of the
component curves are 15.55 and about 17.55; in No. 1L, 15.55 and
about 17.15. In both cases the modal class 15.55 contains many more
individuals than the mode at 17.15 or 17.55. Professor Davenport has

* Except in case of the carpopodite and propoditeof the email chela, where they
are separated by 0.2 mm.

t That is, the curve given by a line passing through the tops of the central
ordinate* of the rectangles.

-

PROCEEDINGS OF THE AMERICAN ACADEMY.

ted thai these compound curves probably represent two successive
separated bj an ecdyaifl, and inasmuch aa the animals measured

70

:,o

"-:

-■

'..

s

--

- -

- i

•*»

*

>

to

to
to

- •-

to | to

to

to

to

Co
to

■si

to

to

to

-:
to

- 1
to

to

to

I'nl.YiiON No. ]\{.

70

00

SO

10

20

—

■

^-

C

'-

to

to

k)

1 ' • » I ..... N. Ni i II.

YERKES. — VARIATION IN THE FIDDLER CRAB.

427

were collected at three
different times (75 July
24, 125 August 17, and
200 August 24), of
which the first and last
were separated by over
three weeks, this is not
improbable. But as I
have no data on ecdysis,
it is impossible to settle
this point.

2. Median length.
This is the second larg-
est dimension taken. It
is slightly more variable
than the frontal breadth.
The polygons No. 2R
and 2L represent simi-
lar distributions, with a
greater frequency, how-
ever, at the mode of the
Rights, as in the frontal
Breadth, indicating less
variability than in the
Lefts.

8. Right margin. It
should be remembered
that No. 3R and 3L are
not strictly comparable
because of the asymme-
try of the animals. The
polygons have been
placed together because
they represent identical
measurements. Compar-
ison should also be made
of No. 3R with 4L, the
polygons for the margin
on the side of the great
chela of the Rights and

80

70

GO

50

40

30

20

10

so

M
©

ki

ki

kl

kl

ki

k,
ki

ki
(0

ki

kl

to

kl

to

kl

to

kl
lit.

kl
k.

ki

Ci

so
t-1

to

Ci

kl

Cl

so

Co
bi

--*

ki

Ci

so
tl

to
Si

kl

Polygon No. 2R.

80

■

70

GO

SO

40

30

20

10

so

ki

so

kl
so

kl
kl

kl

kl
—

kl

kl

<0

kl
to

ki

to

kl

Co

ki

k.

ki

ki

Ci

to
Ci

'I

Ci

kl
Ci

Ci

SO

Ci

to
Ci

-1
Ci

kl
c.

Cl

ti

Ci

- i
Ci

ki
Ci

Polygon No. 2L

128

0CEEDJNG8 OF Till: AMERICAN ACADEMY.

Lefts respectively. No. 8B differs from SL in having greater modal fre>

quency and more classes; it differs From IL by much greater modal

frequency. As compared with the left margin the right is somewhat

variable. For both Rights and Lefts the margin on the side of

too

70

no

SO

40

30

20

in

h

•*

-•

-•

OS

as

•-

<0

<o

--

-~

C*

- »

Ci

hi

- 1

too

90

S,l

7"

so

Ml

40

:;n

20

10

Oi

- j

- i

* i

00

*

«o

«

«o

k.

to
t-1

Co

*•>

50

•-:

~J

POL! QOH So K

POLTGOM No. 3L.

the small chela is 5 Bhorter than the other. In 1 1 * « - ease of ili«' Rights,

I • 1-8.842 Right) = .9500, i.e. Left is 95% of Right And

fortheLefl 3.285 Righl 3.716 (Left) = .9505, i.e. Right is 95+ %

of Left. All dimensions are considerably larger on the Bide <>f the

i chela, a fact which will be discussed later in connection with the

il( in ol i .jlit- and left-handednet

YERKES. — VARIATION IN THE FIDDLER CRAB.

429

00

so

70

to

50

40

30

20

10

c>

<a>

^*

■<«

~»

00

Oo

to

•o

to

Co
In

In

to

In

Co
Cn

In

In

to
In

to
In

Cn

90

80

70

00

50

40

30

20

10

^*

^

*■-*

Oo

00

to

to

to

»*4

0

k4

tn

Cn

cn

to

Cn

Co

Cn

■^4

In

In

Cn
In

to

«o

Cn

in

In

Polygon No. 4R.

Polygon No. 4L.

4. Left margin. Polygons No. 4R and 4L. Compare also No. 4R
with 3L. The three dimensions thus far considered are the most valu-
able for place-mode determinations because they are expressive of the
form of the carapace and may be measured with ease and accuracy. As
given in Table I., the modal classes of the Lefts, — in two instances, —
the frontal breadth and right margin, are below those of the Rights; in
both of these cases there is considerable difference in the means.

430

PROCEEDINGS OF THE AMERICAN ACADEMY.

'JO

80

TO

CO

JO

40

30

Co

10

wl

~1

~J

- '

Oo

Oo

to

to

'-a

>**

>-*

C/>

Kl

Cl

Polygon No. 5R.

00

80

7"

80

50

40

30

20

10

1

>•«

h4

!-i

i

•

i
! 1 1 1

* i

tl

5. Meropodite of
great chela. I shall
here consider homol-
ogous organs. With
the exception of the
usual greater fre-
quency of the modal
class of the Rights.
the polygons No. 5 It
and 5L are very sim-
ilar. The difference
in the coefficients of
variability between
the Rights and Lefts
is .657 in favor of the
Rights.

POLTOOK NO

YERKES. VARIATION IN THE FIDDLER CRAB.

431

100

'JO

80

70

CO

50

■10

30

20

10

ex

to

ex

Ci

cj

"s*

■si

^a

Co

Co

50

to

to

to

to

to

1-4

to

to
to

is

to

to
to

•si

to

s3

to

770

706

90

SO

70

60

50

40

30

20

10

to

to

to

C>

Ci

- i

■si

-1

Cc

Oo

<o

to

to
to

sD

to

to
to

■si

to

to

to
to

to

to
to

•si

to

to

Polygon No. 6R.

Polygon No. 6L.

6. Carpopodite of great chela. Here we meet an exception in the
matter of modal frequency, for the Lefts' mode contains more individuals
than that of the Rights. No. 6R and 6L are otherwise similar. In
comparison with the meropodite the carpopodite is much more variahle.

Bateson ('94, p. 79), in his book on meristic variations, states that
terminal members of a series are likely to be most variable. " In such
oases as that of the last rib of Man, and several other animals, the wis-
dom teeth of Man, etc., it is quite true that in the terminal member
Variation is more noticeable than it is in the other members." But Buch
series as those of teeth and ribs, in which there is repetition of similar
organs, are so different from the linear series of joints with which we
are dealing, as to make inferences from one to the other unwarrantable.

132

PROCEEDINGS OF THE AMERICAN ACADEMY.

Bateson does mention (p. 63) substantive variation in the joints of the
leg of the cockroach (Blatta americana), but only in its relation to the
Dumber of joints, no lighl being thrown upon <>nr problem.

In case of the series of joints of the crab leg it would seem probable
that the distal would be the most variable, since they are most likely to
be modified by the use of the leg.

For convenience of comparation the Coefficients of Variablity of
the measurements taken on the chelar joints are here repeated (see
Table [L

Meropodite of greal chela,
Carpopodite of great chela,
Propodite of greal chela,
Meropodite of small chela,
( larpopodite of small chela,
Propodite of small chela,

Rights.

9.000
10.166

Ki.215

8.7;; i
9.518
8.313

Lefts.

9.657
10.472

10.255
9.920

11.139
8.898

Only for the greal chela of the Rights is there a constant increase in
variability from the proximal to the distal joint. In every case, how-
ever, the carpopodite is much more variable than the meropodite.

so

:o

00

zo

to

30

20

in

-..

•

c.

>

Os

U>

»

'•c

>■*

fc*

>■*

k.

k4

Ex

Cl

Cl

Cl

Co

(a

C-i

5.1
i.-.

-

t-1

• i

ti

ti

- »

POLTOOH N" 7B

YERKES.

VARIATION IN THE FIDDLER CRAB.

433

Among the hundreds of crabs examined not a single variation in the
number of joints was noticed.

so

70

60

50

40

30

20

10

Cl

Ci

o

•^

M

•>»|oo

00

sJ-JLU0

l~l

\i

>-*

so

to

-0

K-4

so io

C» | Cj>

k. ti o
to, 1 o> 1 fcl

CO

Polygon No. 7L.

7. Propodite of great chela. Polygons No. 7R and 7L are more
nearly symmetrical than any of those already examined; 7R has a
slightly wider range of variation.

vol. xxxvi. — 28

134

PROCEEDINGS OF THE AMERICAN ACADEMY.

U(

130

ICO

110

100

•.in

so

70

60

00

JO

30

Co

10

.

'-.

c-i

ti

Cl

a

- I

■-.

- /

14f

13(

106

110

100

90

SO

70

GO

50

■10

30

CO

10

Co

-

•fc.

0»

ex

«

c>

<0 I Co
» 0,

Ci

c.

*o
Cl

* *
C-i

I*« H-i '.<>N No. XK.

Polygon No. 8L.

C. Meropodite of small chela. Polygons No. 8R and 8L

YERKES. — VARIATION IN THE FIDDLER CRAB.

435

110

100

90

SO

70

60

50

40

30

20

10

to

to

'0

to

Co

Co

Co

Co

Co

Co

so

Ol

Co

Ct
Ol

Cl

so

IOC

90

SO

70

GO

50

JO

30

20

10

to

to

to

to

Co

Co

Co

I"!

Ct

_Co_

en

Co

so
CI

Co
Ol

Cn

c->

*OJ

Ol

so

CI

Co
Oi

Polygon No. 9R.

Polygon No. 9L.

9. Carpopodite of small chela. Polygons No. 9R and 9L. The car-
popodite is the most variable organ measured, the largest coefficient of
variability being that of the carpopodite of the small chela of the Lefts,
and the second largest, that of the carpopodite of the great chela of the
Lefts.

43G

PROCEEDINGS op THE AMERICAN ACADEMY.

/.■■•'

i:n

lit

U

90

SO

70

CO

60

40

30

20

10

'.

► ;

»0

Co

Co

Co

o.

»»

>~4

fcl

1H

/,.,

00

so

10

GO

-,o

to

30

20

JO

-J

N>

to

fo

Co

Co Co

Co

to

***

Ci

C-i

>»4

Cn Cs

PoLTooa No. in];.

Poi roon No. 10L.

10. Propodite of small chela. Polygons No. LOB and 10L. The
propodite of the small chela of the Rights is absolutely the least variable
of the measurements made on the legs, while the propodite of the great
chela of the Rights (Polygon No. 7R) is the most variable of all the
organs of the Rights. We have here an illustration of Darwin's law,
thai a part (the propodite of the great chela) developed to an extraordi-
nary degree is more variable than a pari (the propodite of the small chela)
of ordinary size. This law of variation holds foi all the measurements
on the great and -mall chelae of the Rights.

YERKES. VARIATION IN THE FIDDLER CRAB.

437

WO

00

so

70

60

50

40 j

30

20

10

Cn

tl

Cs

Ci

■sj

^1

*oi

Cc

Co

so

so

se

Ci
CT

Ci

to

C.T

c-1

Cl

so

C~i

Co

Ci

c-1

s>

so

Cl

700

00

SO

70

60

50

40

30

20

10

ti

0:

0?

•*l

"•»

"*?

30

o>

-

SVi

c-.

■-I

Co

Ci

C-i

Co
Cl

- '
tl

Polygon No. 11R.

Polygon No. 11L.

1 1 . Meropodite of second leg on side of greed chela. Polygons No. 1 1 R
and 11L.

138

PROCEEDINGS OF THE AMERICAN ACADEMY.

7H

:

00

80

70

no

so

40

.W

20

10

ci

t-t

~

Cl

■ i

~l

-*J

oo

■y.

ci
Ci

to

to

Ci

Cl
Cl

Cl

Co

Ci

1

90

80

70

60

50

40

30

20

JO

<■-■<

<ft

Ci

- 1

^1

-1

Oo

03

«o

<o

•c

Cl

Co
Cl

-1

Cl

1-4

Cl

to

Cl

- 1
Cl

Cl

Cl
Cl

Cl

Polygon No. 12R.

Polygon No. 12L.

1 _'. Sferopodite of second leg on side, of small chela. Polygons No.
12B and 12L. Attention may be called to the fact that in both Rights
and Lefts the meropodite on the side of the great chela is larger than tli.it
of the opposite Bide. This apparently indicates that the influences which
determine the development of the great chela affect other parts of the
b "1\ as well.

I TabU of Correlation of Meropodites of Chelae.

In Table V. the correlation <>f the meropodites of the chelae of the left-
handed animals is Bhown. This was determined in order to see whether
there was do! an inverse correlation; for it Beemed ool improbable thai
an unusually large Great Chela Bhould be accompanied by an exceptional
Bmall Small Chela. Such, however, is not the case ; instead there is a
fairly close direct correlation, 0.774 ,

YERKES. — VARIATION IN THE FIDDLER CRAB.

439

■*♦*

«

to

»— 1

0

0

CO

l-H

00

0

r~

^H

uO

CO

CO

O

t~

O

a>

to

00

CO

t— •

■^

l~

t—

t^

CS

CO

00

I-

1-

•1-

i-H

■«

+

+

+

+

+

+

+

+

+

+

+

+

+

+

•

g *> • ~~

CN

<*

CO

t-~

CTi

0

CT2

O

>o

■*

CN

0

CS

.

.2 v q t3 F

uO

1^

OS

r^

co

O

CM

CS

00

0

CO

CS

CD

| 5-|. £S

t"-

CN

I—

CO

00

CO

1— 1

u.i

O

iO

0

O

CS

Ti

CN

CN

,_;

i-H

1 — 1

M

<M

<M

CM

00

a

0

►J

1

1

1

1

1

1

+

+

+

+

+

+

+

+

<

-—*

a

«

CI

0

^

S w .

O

cn

CN

0

as

r~

r-

10

0

t-

1— 1

■>*

O

<3

.2 g>« P

■S 4- -S .4,7« •

<N

O

0

CO

iO

0

vO

^H

CO

0

1 — 1

CM

t—

CO

0

a

CO

O
CM

0

. — ■

O

1— 1

iq

00

O

•>*

CO

I— 1

as

I— 1

CN
CN

CN

1 -

O

a

Q

O ft! '

1

1

1

1

1

1

+

4-

+

+

+

+

+

+

0

4->

*

s

13

■<

a

a

co

rs

CO

CD

os

rH

00

CNI

■*

CO

_

r-

rt

■*

CJ

5fi

H

^55

as

0

0

O

CN

<M

0

<M

■^

kO

0

T— ■

CO

T}<

ft

r— 1

I-H

T—l

I-H

I-H

0

ft
O

o

Q

]

1

1

1

1

1

+

+

-X.

+

+

+

+

+

O

oa

a >

0

CN

iO

iO

CM

O

"*

CO

0

CS

Ol

OO

01

uO

ft
o

*2 ° £

CO

CM

O

id

O

CO

0

0

CO

CO
CO

CO

0

CO

I—

CO

w

£

«.

IJ

CN

co

CM

CO

-r-

O

CO

!>■

CO

0

CN

CO

CO

1 — 1

0

w

K

•si^npiAjvui jo -om

1— 1

CO

0

as

t^

O

<M

CN

^H

0

o

ft
o

•0

CM

O

i-H

0

rH

r^

CM

l-H

00

00

CO

+

w

H

0

CO

-N

I-H

rH

I-H

1^

O

-M

l-H

~r

C

o

CO

i-H
+

CN

ft

o
ft

CO

CO

CM

CO

t^

1— 1

•0

•—1

CN

l-H

ftl

0

3

ft

lO

Tf

CN

1 — 1

T— 1

i — 1

■*

1—

CO

CN

o

id

+

*"" '

CNI

CO

3j

55

O

H

U5

O

l-H

CO

-P

0

O

CN

CN

I— I

r^>

rH

CN

0

-Tl

I-H

T

ft
ft
ft

0

1

1—1

uO

CO

Tf<

-r

CN

0

CO

CM

o

o

•*

1

CN

. 1

t—

1

CO

0

CN

(N

iO

O

CO

CO

O

r>

-*'

1

CM

1-3

0

co

co

CO

co
p-i

1

rH

^

<J

r^

wco £ .a S

O

O

0

O

0

0

0

O

O

0

O

O

O

O

GO

CO
CN

0

l-H

1—

CO

»— '

O

CS

CO

1-

CM

CM

CS

oi

=

EG
3

C rj
v. ?

V "« 3 1 1

0 w 0 '1

1

1

1

1

1

1

+

+

+

+

+

+

+

+

>

"C

c

B

1 B

3^ 43 .

iO

O

O

O

iO

iO

0

iO

i.O

1.0

iO

iO

IO

10

l-H

1 _ri

J es

33 _ « 1

00

CD

t—
CO

I~

OS

CO

06

1^

06

CS

0

CS

0

CS

M
O

i-H

0

O

(. 0 = .2

I-H

i-H

55

440 PROCEEDINGS OP THE AMERICAN ACADEMY.

V. Discussion of Results.

Concerning the original purpose of this study, the determination of
sonic place-modes for the Fiddler Crab, it may lie said that a number of
modes have been given in Table [., which it is hoped will be of value in
the future as helps in the study of racial variation and of the formation
of Bpecies. As before remarked, those measurements which would seem
the most valuable for place-modes are : (a) the frontal breadth, (b) the
median length, " ' the right marginal length, and (<7) the left marginal
length.

It remains to be asked, What answers do the results of this Btudy
enable us to give to the group of questions centring about the Great
Chela condition, with which we Bel out ? (1) What is the significance
of right- and left-handedness ? (2) What relation does the size of the
il chela bear to other dimensions ? < 3 ) What determines right- and
left handedness ?

The chief significance of the gr< at chela, observation leads me to con-
clude, ia in its value (1) as a meaus of defence and offence, and (2) as a
means of burrowing. Bui granting that these are sullicient reasons for
its existence, we have still to ask why it is sometimes on the right Bide,
Bometimes on the left. The fact that approximately equal numbers ol
R ghtS and Lefts are found, seems to indicate that the great chela is not
determined by heredity, or, at least, not directly. For if this were the
case, the probability of an equal distribution between the two types
would be very Blight It is more likely, therefore, that we are dealing
with what is usually called (although improperly) a chance determina-
tion: that is, there are a number of variable factors, only partially
known at present, which throw the greatrchela-developing-tendencj now
ibis way. now that. Other instances of determination of this kind are
the determination of Bex and of the crossing of the optic nerves. IYo-
or < i. II. Parker, who has studied the crossing of the optic nerves
of (idle,, bm whose results have not yet been published, has found
thai many symmetrical Bshes are about equally divided between those
having the right nerve crossing above the left and those having it en
iic_r below the left. Since ii is at present impossible to point out, in bucIi
any uniform cause or group of causes for the condition, we sa\ it is
a matter of chance.

I ii answer to the second question, — What relation does the Bize of the

it chela bear to the oilier dimensions ? _ it uiay be said, that the in

urements which allow of com pan - I the two sides ol the body, namely,

YERKES. — VARIATION IN THE FIDDLER CRAB. 441

the margin and the meropodite of the second leg, are larger on the
side of the great chela. Expressed in percentages, the relations are as
follows: Margin on chelar side in case of Rights, 5% greater than on
opposite side, — the same is true of the Lefts; meropodite of second leg
on chelar side in case of Rights, 6% greater than that of the opposite side ;
meropodite of second leg on chelar side in case of Lefts, 8% greater. The
fact that measurements are greater on the chelar side may be taken in
support of (a) the idea of chance determination, for any advantage of one
side of the body over the other would be likely to affect all organs. On
the other hand it may be held (b) that the development of the great chela
is itself the cause of the greater development of the other organs on the
same side. Or, combining these two views, it might be maintained
(c) that the causes which led to the development of the great chela were
operative also in case of the other organs, but that the development of
the great chela was an additional cause for the unusual size of the other

.->

organs.

As to the difference in size of the Rights and Lefts I can only say that
it is an interesting and surprising fact, for which I have found no expla-
nation. I at first interpreted the fact that the Lefts are the smaller and
the more variable as meaning that the Rights represented the stable
form ; and in support of this I had, as a result of the first summer's work
(see Table III.), evidence that there were more Rights than Lefts. But
this conclusion was not confirmed by the additional observations made
during the summer of 1900 (see Table IV.). It may be, however, that
the proportion of right- and left-handed animals varies in different col-
onies, and that during the interval between the collection of the animals
referred to in Table III. and those of Table IV. migration or some other
change had caused alteration in the relative frequency of Rights and
Lefts ; but this does not seem very probable.

VI. Summary.

1. The place-modes for twelve measurements on the carapace and
extremities of right- and left-handed Gelasimus pugilator arc given in

Table I.

2. The means of Table I. show that the right-handed animals are
larger and less variable than the left-handed.

3 Right- and left handed animals occur in approximately equal
numbers.

1 12 PROCEEDINGS OF THE AMERICAN ACADEMY.

4. Two " stages " appear to be represented by the individuals meas-
ured. Many of the curves are compound, and it is probable that in these
one mode represents those individuals which have recently molted, the
other those which are almost ready for the process.

5. All the measurements are larger on the side of the great chela.

6. The results of this study in variation seem to indicate that right-
and left-handednese is not directly due to heredity, but is caused by
certain Blight variations which give one side of the body an advantage
over the other.

Bibliography.

Bateson. W.

'94. Materials for the Study of Variation treated with special Regard
to Discontinuity in the Origin of Species. London and New
York, Macmillan iN Company. 608 pp.
Davenport. C. B.

'99. The Importance of Establishing Specific Place-Modes. Science,
M.S., Vol. 9, No. 220, pp. 415-110.
Davenport. C. B.

'99v Statistical Methods with special Reference to Biological Variation.
59 pp., h) tables, and 28 figs. New York, John Wiley & Sons.
Weldon. W. F. R.

'93. On Certain Correlated Variations in Carcinus maenas. Proc. Roy.
London, Vol. 51, pp. 318-329.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 25. — April, 1901.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE
MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE,
UNDER THE DIRECTION OF E. L. MARK. — No. 122.

THE DEVELOPMENT AND FUNCTION OF REISSNER'S

FIBRE, AND ITS CELLULAR CONNECTIONS.

A PRELIMINARY PAPER.

By Porter Edward Sargent.

With Two Plates.

CONTRIBUTIONS FROM THE ZOOLOGICAL LABORATORY OF THE
MUSEUM OF COMPARATIVE ZOOLOGY AT HARVARD COLLEGE,
UNDER THE DIRECTION OF E. L. MARK.— No. 122.

THE DEVELOPMENT AND FUNCTION OF REISSNER'S

FIBRE AND ITS CELLULAR CONNECTIONS.

A PRELIMINARY PAPER.

By Pouter Edward Sargent.

Presented by E. L. Mark, March 13, 1901. Received March 15, 1901.

In a recent number of the Anatomischer Anzeiger (Bd. XVII. pp. 33-
44) I have described the occurrence of Reissner's fibre in the canalis
centralis of vertebrates, and I there promised the early appearance of a
paper on its development. My further researches on its development
and cellular connections have led to the discovery of a highly specialized
apparatus, of which the so-called fibre forms a part. The subject has
developed much larger aspects than was anticipated, and the consequent
delay in the publication of my complete results makes necessary a fur-
ther preliminary paper.

As described in the previous paper, Reissner's fibre lies in the cere-
brospinal fluid; it extends through the whole length of the canalis
centralis and the brain ventricles to the anterior portion of the optic
lobes, 'where it passes into the brain tissues; it gives off throughout the
posterior portion of its course fine branches, which enter the tissue of
the spinal cord.

In the present paper I hope to show briefly that there is normally in
the central nervous system of all vertebrates a highly specialized appara-
tus, the cells of which lie in the optic lobes and send their axons into
the optic ventricle and the central canal to form the so-called Reissner's
fibre. I have investigated the occurrence of this apparatus in upwards
of one hundred species, representing all the principal classes and sub-
classes of vertebrates. The development has been studied in about
twenty different species, and has been more or less completely worked
out in representatives of all the chief groups of vertebrates. In this
paper I shall describe the apparatus and its development in a few char-

iV> PROCEEDINGS OF THE AMERICAN ACADEMY.

acteristic forms, reserving the more detailed account of the investigations
for :i later and more extended article.

In Amid, at about the time of batching, there is in the anterior por-
tion of the roof of the third ventricle a differentiation of the neuroblasts.
Some of them have already increased in size, have become more nearly
Bpherical and take the stain more deeply than the surrounding cells
(Figure 1). These enlarged cells are concentrated for the most part
in the median plane at the anterior end of the tectum opticum near the
posterior commissure. During the first stage of differentiation there are
from twenty to thirty of these cells, but by the development of indiffer-
ent neuroblasts they increase rapidly in number, so that by the end of
the fourth day after hatching there are from eighty to one hundred of
them. During the tirst day the tectum is only one cell thick. Each cell
baa a large nucleus — whose diameter is somewhat more than one half
that of the cell — with a single deeply staining nucleolus.

During the first and second days the axons develop from the cells as
emely fine processes, which always appear on the side of the cell
nearest the ventricle, directly towards and into which they grow. Early
in the third <lav adjacent axons, now projecting into the ventricle, come
together in groups and coalesce at their tips (Figure 2), appearing as a
Bingle fibril in their further growth through the cerebro-spiual fluid. In
later stages these fibrils coalesce with others similarly formed, and in
their growth backwards through the fluid of the ventricles and canal
form what has been known as Keissner's fibre (PI. 2, Figure 8.) During
larval development the fibrils which make up Reissner's fibre coalesce
more and more, until by the eighth day there are only six or eight separate
divisions of the fibre in the ventricle, and later these are still further
reduced in number by further coalescence. The cells also give off den-
drite-, which run backward through the substance of the tectum.

In the later development of the tectum opticum these cells undergo an
apparent migration, due to the development of intervening tissue, as a
result of which their relative positions are considerably changed; the

cells bec< • more crowded together. During these changes in position

the axis of the cell may rotate through an arc as great as 180°. It is an

interesting fact that the eccentric nucleus always retains its position at
the Bideoi the cell opposite that from which the axon emerges.

At about the time of hatching one may find in the extreme posterior
end of th<- canalis centralis i Figure 5) a number of small cells, three to

four iniera in diameter, lying in the lumen of the canal and ventriculus

SARGENT. — REISSNER'S FIBRE. 447

terminals. The surrounding walls of the canal are made up of a single
layer of cells, actively dividing. Certain of these cells become separated
from the walls and pushed into the lumen. Of these many atrophy
and disintegrate. Some eight to twelve of them persist and continue to
develop. These cells, at first spherical, become spindle-shaped, as they in-
crease in size. These posterior canal cells (as I temporarily call them)
give off peripherally a number of dendrites which penetrate the tissue
of the cord (Figure 3). From the tapering anterior end of the cell is
given off the axon (Figure 4), which runs cephalad through the lumen,
eventually coalescing with the axon of similar cells (Figure 5) and
then continuing its growth cephalad. About the third day this system
of forward-growing axons meets the system of axons from the cells of
the tectum growing backward and the two systems coalesce in a way
not yet clearly made out, to complete what has been called Reissner's
fibre. This fibre, then, is a nerve tract, composed of axons running in
opposite directions, both cephalad and caudad. The development of this
apparatus as outlined for Amia is typical of all vertebrates.

The condition of the apparatus in Selachii is especially interesting.
In the skate Raja erinacea the cells are of great size (25 /x to 30 /x in
diameter), being the most conspicuous elements in the brain. They are
three hundred to four hundred in number, and lie in two bands, one on
either side of the mediau plane, extending the whole length of the roof
of the optic lobe (PI. 2, Figure 9). At the anterior end these two bauds
meet, forming a large group of closely aggregated cells. The cells are
multipolar, giving off several processes in addition to the large axon, which
is 2 /a to 3 fM in diameter. As in Amia, the nucleus is eccentric, being
always at the side of the cell opposite that from which the axon emerges.
The number of the capillaries in the neighborhood of these cells, supply-
ing them with blood, gives evidence as to the extent of the activity of the
cells.

Although these cells are so conspicuous in the brain of the skate, they
have attracted but little attention. They were first seen by Rohon in
selachians and correspond in position to the " Dachkerne " which have
been described for Amphibians and Reptiles, the function of which has
not been established.

The axons pass dorsad and laterad from their cells, turning either
cephalad or caudad, and form two great tracts of non-medullated fibres of
large size, which extend through the whole tectum opticum lateral and
dorsal to the cells. The fibres running caudad can easily be traced to
the posterior end of the tectum, whence they pass into the cerebellum

448 PROCEEDINGS OF THE AMERICAN ACADEMY.

and come into close proximity to the Purkinje cells. The fibre tract
which runs forward may be traced to the anterior end of the tectum,
where the fibres converge, apparently forming Keissner's fibre, which
emerges from the tectum into the ventricle at this point. Where Reiss-
ue's fibre leaves the brain tissue, the membrane covering of the roof of
the ventricle is continued in a cone-like projection surrounding the fibre
i Figure 9). Still another process given off from these cells passes peri-
pherad nearly to the outer surface of the tectum, where it comes directly
in contact with the endings of the optic nerve fibres. These cells, whose
axons form Keissner's fibre, are, then, in direct connection with the eye
through the optic nerve, and in turn are connected with the Purkinje
cells of the cerebellum, which are known to coordinate muscular move-
ments. The posterior canal cells in the selachians are of relatively large
size and about :i dozen in number.

In tin- ( 'i/clostomala this apparatus is in a more primitive condition. In
P( tromyzon marinus it begins to develop some time after hatching, and
is not fully established until the second month of larval life. In larva?
thirty days after hatching the cells form a small but well-marked group
lying in the roof of the rudimentary optic lobes in the median plane, and
occupying the whole thickness of the tectum (Figure 7). The axons
emerge directly into the optic ventricle and unite in groups to form
Reissner's fibre. These axons may continue backward to the fourth
ventricle in three or four separate divisions, but in later development
these coalesce more and more. At this stage the posterior canal cells are
already developed and are sending their axons cephalad through the
canal, but they have not yet united with the axons of the tectal cells
running candad.

In the adult Pctromyzon Reissner's fibre passes through the canalis
centralis and the fourth ventricle, from which it enters the brain tissue of
the basal portion of the cerebrum and passing through this emerges into
the third ventricle. Here it breaks up into several trunks and continues
forward to the anterior portion of the ventricle, where after further
di\ ision it enti rs the tectum.

The passage of Reissner's fibre for a portion of its course through the
nervous tissue in cyclostomes is especially interesting as tin-owing light
on a little understood structure in Amphiozus. Of the dorsal colossal

nerve cells in the anterior portion of the central nervous svstem of

Amphiozus, the largest ami most anterior, designated by Rohde as the
•■ kolossale ganglien-zelle," is median and lies across the lumen of the

canalis centralis. The cell is in direct connection with the pigment spot.

SARGENT. — REISSNER'S FIBRE.

449

The axis cylinder runs caudad in the median plane just ventral to the
floor of the canal. I venture the conjecture that this cell represents, in
Ainphioxus, the tectal cells of craniotes, and that its giant axis cylin-
der is homologous with Reissner's fibre. In the phylogeny of this appa-
ratus the cyclostomes would then furnish the connecting link between
Amphioxus and the gnathostomes.

A number of experiments have been performed to determine the func-
tions of this apparatus. The best results have been obtained with dogfish
(Squalus acanthias), and small sharks (Charcarias littoralis). In these
animals Reissner's fibre is of such size (10 p, to 20 p. in diameter) that
it is possible to remove portions of it from the brain ventricles or canal
for microscopical examinations. The operation of breaking the fibre

*<&**

fbr n . opt.
eiopl rpc.

Figure A. Diagram of the optic reflex apparatus, cbl., cerebellum ; cl.can. p.,
posterior canal cells ; cl. opt. rfx., optic reflex cells ; cl. P., Purkinje cells of cere-
bellum ; fbr. n. opt., afferent optic nerve fibres ; fbr. R., Reissner's fibre ; vnt. trm.,
ventriculus terminalis ; tct. opt., tectum opticum.

was easily performed. By an incision through the integument, and the
removal of a small bit of cartilage, the tela choroideus covering the
fourth ventricle was exposed. Through this a curved needle was plunged,
and drawn several times transversely across the ventricle to break the
fibre. The wound was then dressed ; the animal, when replaced in the
water, usually recovered from the sbock in from twenty to thirty minutes.
Those individuals in which the operation had been performed without
success in breaking the fibre, served for the control experiments. The
success or failure was ascertained at the autopsy, performed at death or
on the fourth or fifth day after operation. Furthermore, the cord and
medulla of each individual was preserved for microscopical examina-
tion. In about one third of the cases the operation had failed to
vol. xxxvi. — 29

450 PROCEEDINGS OF THE AMERICAN ACADEMY.

break the fibre. These individuals showed a marked difference in be-
havior from those in which the fibre had been broken. The former,
after recovery from shock, were apparently normal in their reactions.
Those in which the fibre had been broken showed a slowness in response
to optical stimuli. For example, such an animal, in swimming about
the tank in which it was confined, would bump into the sides of the cage.
If any obstacle were suddenly placed in front of it while swimming, its
effort to avoid the obstacle would come apparently too late to prevent
collision. Those individuals in which subsequent autopsy and examina-
tion showed the fibre to be intact exhibited much greater dexterity and
quickness in avoiding obstacles. The behavior of the animals operated
on was constantly compared with normal individuals kept with them in
the same enclosure. These experiments could be made entirely con-
clusive only by comparing the time reactions in cases where the fibre was
cut with those where it remained uncut. No attempt has yet been made
V> do this.

The apparatus which is the subject of this paper forms, I believe, a
short circuit between the visual organs and the musculature, and has for
its functiou the transmission of motor reflexes arising from optical
stimuli. This view as to its function is supported by many distinct
lines of evidence, which are here classified and summarized.

(1) Anatomy. There is normally in all vertebrates a group of cells
lying in the optic lobes which by some of their processes are in direct
connection with the central terminations of the optic nerve, and by
others with the Purkinje cells of the cerebellum. Their axons pass by
tin; shortest route through the ventricles and canal to the posterior por-
tion of the nervous system, where they pass into the cord and probahlv
out through the ventral roots to the musculature. These anatomical
connections make it probable that this apparatus is a direct path for
the transmission of motor reflexes arising from optical stimuli. The
connection with the Purkinje cells of the cerebellum, it is equally evi-
dent, is for the coordination of muscular movements.

( ■_' ) Experimental Physiology. When Reissner's fibre is cut or
broken, the animal loses the power of responding quicklj to optical
stimuli. This is not due to the shock result iim- from the operation, for
animals on which the equivalent operation has been performed without
breaking the fibre are Dearly or quite normal in this respect.

(•".i Comparative Physiology. (") In anyone group, as for example
the teleosts, the apparatus ha- its highest development in those animals
which arc most active. In the predatory and rapacious bluefish (Poina-

SARGENT. — REISSNER'S FIBRE. 451

tomus) the apparatus is highly developed, Reissner's fibre having a di-
ameter of 10 (x. In Lophius, which is a much larger fish, but is
usually sedentary and quiescent in its habits, and more dependent upon
tactile than visual sensations for obtaining its food, the apparatus is
degenerate and Reissner's fibre inconspicuous, (b) The corpora quad-
rigemina in higher vertebrates are degenerate organs, having given up
most of their functions to other parts of the brain. In mammals they
are, it is generally believed, concerned only with reflex functions. As-
suming this to be true, the apparatus under discussion, since it has its
centre in the corpora quadrigemina, must have, in the higher vertebrates
at least, a purely reflex function.

(4) Embryology, (a) The apparatus does not reach full development
until just before the animal attains free life. In the active fry of trout
and salmon, which give reflex responses at this early age, it is fully devel-
oped at the time of hatching. In those forms in which the young are
retained within the uterus of the mother until a late stage of develop-
ment, as in mammals and some selachians, the apparatus is not developed
until a relatively much later stage, (b) In sluggish larvae the apparatus
is not established until after the time of hatching; in Amia not until the
fourth or fifth day, in Petromyzon not until the second month. This
corresponds closely with the time at which these larva; begin to respond
promptly to definite optical stimuli from surrounding objects, (c) In
those mammals which are born blind, as the mouse and kitten, the
apparatus at birth is in a very incomplete state, so that it cannot be
functional. In the mouse at birth the axons are just penetrating into
the ventricle, and Reissner's fibre is as yet unformed.

(5) Degeneration. A study of the blind vertebrates of the cave fauna
from the collections of Professor C. H. Eigenmann show that the optic
reflex apparatus is reduced in direct proportion to the degeneracy of the
eye. In those species which are totally blind no trace of this apparatus
is to be found.

Experiments are now in progress to determine the effect of artificial
extirpation of the eye on this apparatus.

Such a " short circuit " for the transmission of optical motor reflexes
must be of great importance in saving time. If the impulse were trans-
mitted from the termination of the optic nerve to the posterior muscula-
ture through the cord, it would involve transmission through a chain of
from two to three neurons. Now, it is well known to psychologists that
the time of transmission of a nerve impulse is delayed by passing through

452 PROCEEDINGS OP THE AMERICAN ACADEMY.

a cell boch\ and by passing from one neuron to another. Wundt found
that the delay in the spinal ganglion cell of man was 0.003 second.
The delay in passing between different neurons is probably greater, say
0.005. So in man, if two cell bodies and two contacts between separ-
ate neurons are to be traversed, the loss of time would be ([0.003 X 2] +
[0.005 X 2]=) 0.016 second. In lower animals, as is well known, the
time reactions are much greater. It is quite probable that in some of the
lower animals this short circuit may mean the saving of a considerable
fraction of a second.

An animal suddenly confronted with some optical evidence of danger
from which it recoils in fear, does so reflexly, calling into operation this
apparatus. When we pause to consider how often in the struggle for
existence the saving of a fraction of a second may be a question between
life and death, we see how important a part this apparatus may have
played throughout the vertebrate series in the survival of the fittest.

EXPLANATION OF PLATES.

Abbreviations.

cbl.,

cl. opt. rfx.,

corns, p.,

e'phy.,

Cerebellum.
Optic reflex cells.
Posterior commissure.
Epiphysis.

fbr. R., Reissner's fibre.
tct. opt., Tectum opticum.
vnt. III., Third ventricle.
vnt. IV., Fourth ventricle.

PLATE 1.

Figure 1. Amia calva, 1st day. Sagittal section (anterior end at right) through
roof of mid-brain. The large cells in the tectum opticum are the
tectal reflex cells, a few of which are just beginning to send out
their axons.

Figure 2. Amia calva, 12th day. Sagittal section (anterior end at left) through
anterior part of tectum opticum. The numerous fibrils entering the
third ventricle are the axons of tectal reflex cells, a few of which
are shown in the section, but the most of which are lateral to this
section.

Figure 3. Amia calva, 6th day. Transverse section of the spinal cord through
the ventriculus terminalis ; posterior canal cell sending its dendrites
through the fluid of the canal into the tissue of the cord.

Figure 4. Amia calva, 1st day. Sagittal section through the posterior end of
the spinal cord and ventriculus terminalis, showing a single posterior
canal cell, its long process directed cephalad.

Figure 5. Amia calva, 12th day. Sagittal section of a portion of the spinal cord
near its posterior end ; posterior canal cells in the ventriculus ter-
minalis and canalis centralis, sending their axons forward to form
Reissner's fibre. Reconstruction drawing from four successive
sections.

Figure 6. Squalus acanthias, embryo 2 cm. long. Longitudinal sagittal section.
Posterior canal cells in the canalis centralis sending dendrites into
the cord and its axon cephalad.

Figure 7. Petromyzon marinus, 30 days. Sagittal sections of the tectum opticum
(anterior end to the right) ; the axons of the developing tectal reflex
cells are just emerging into the third ventricle.

PLATE 2.

Figure 8. Amia calva, 12th day. Sagittal section through brain, somewhat
diagrammatic.

Figure 9. Raja erinacea, 14 cm. long. Sagittal section of the mid-brain, dia-
grammatic.

Sargent.-Reissner's F

IBRE.

Plate 1.

•

•

7.

9.

•:

-

5.

*.

; : li^f I

A - u

6.

a

5.

„.-^iM-^

Sargent -Reissner's Fibre.

Plate 2.

tct.opt.

eHj.

Iint-TIL.

g c! opt.r/j.

coins, f-

8.

9.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 26. — April, 1901.

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY.

New Series. — No. XX.

By B. L. Robinson.

I. Synopsis of the Genus Melampodium.

II. Synopsis of the Genus Nocca.

III. New Species and newly noted Synonymy among the Spermato-
phytes of Mexico and Central America.

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD
UNIVERSITY, NEW SERIES, NO. XX.

By B. L. Robinson.

Presented March 13, 1901. Received March 16, 1901.

J. — SYNOPSIS OF THE GENUS MELAMPODIUM.

In the following key to Melampodium the genus is limited as by Ben-
tham and Hooker in their Genera Plantarum and by Hoffmann in Engler
& Prantl's Natiirlichen Pflanzenfamilien. It will, therefore, be unneces-
sary here to reproduce the generic description or give generic synonymy.
The key is based chiefly upon the material which has accumulated in the
Gray Herbarium, including the recently acquired Klatt collection and
some borrowed material from the U. S. National Museum. The writer
has also in connection with this work been kindly permitted by Mr.
Casimir de Candolle to examine and trace the types in the Prodromus
Herbarium. Much difficulty has been experienced in giving the species
a natural secpience, and after many efforts the hope of securing such an
arrangement has been abandoned. The employment of pubescence in
grouping the species of this genus is new and appears to yield more satis-
factory results than an implicit reliance upon the fructiferous bracts.
The latter, as is well known, often surpass the achenes, forming above
them a cup or hood. This hood is often pointed dorsally at the summit
and the point may be recurved or spirally coiled. Unfortunately, how-
ever, these features, the hood and its appendage, show too great varia-
bility in certain nearly related forms, such as M. sericeum and its
varieties, to yield diagnostic characters of the first rank. However, the
presence or absence of a hood can usually be determined readily, and
the two sections Eumelampodium and Zarabellia may conveniently be
retained.

Bentham and Hooker, 1. c, estimated the species at eighteen, and
Hoffmann, 1. c, accords twenty-five specif-; to the genus. It will be
seen, however, that this number can. with our present knowledge, he
somewhat increased. The genus reaches its greatest development in
Mexico, where, if we include Lower California and Central America, no
less than thirty-one species occur. Of these species three reach the
southern United States (one merely as an introduction), two are found

456 PROCEEDINGS OF THE AMERICAN ACADEMY.

in the West Indies, and two ur three extend to South America. There
are also two species known exclusively from South America. Early in
the nineteenth century a species of Mdampodium {M. diffusum) was
discovered on the island of Luzon in the Philippines. As the genus is
otherwise American, the occurrence of this species in a region so remote
has always been problematic, and it has been a matter of no small inter-
est to find the Philippine plant closely matched by specimens recently
collected by Dr. Kdward Palmer, about Acapulco, Mexico. Tliere can
therefore be scarcely a doubt that the genus is in reality of New World
origin, and that a Bingle Mexican species was accidentally introduced into
die Philippines, where it attracted scientific attention before it was recog-
nized in America. This seems the more likely from the circumstance
thai Mexico and the Philippines were under the same national control,
and early connected by a certain amount of oceanic traffic. This being
the ease, the transference of seed from Acapulco, the most important
Pacific port of Mexico, to the neighborhood of Manila, presents no
inherent improbability. The writer is under obligation to M. Robert
Buser of the De Candollean Herbarium for critical comparison, notes,
and sketches relative to this and related species.

In t li j -v paper the term fruit is applied to the ray-achene and the closely
enveloping bract

§ 1. Ki mi i. ami'ODILM, DC. Inner (fructiferous) bracts of the invo-
lucre exceeding the inclosed achene and developed at the summit into a
cup or hood (this obsolete iu some forms of M. sericeurri). — Prodr. v.
518 I 1836).

■j

• Lower -urfaceof the leaves sparingly pubescent to hirsute, villous, or tomentose,

but not sericeous.

+- South American species : hoods scarcely or not at all appendaged.

— Herbaceous annual : rays conspicuous, 6.5 mm. long, unguiculate.

1. M. paludicola, Taubert in Engl. Jahrb. xxi. 4.">r> (1896). —
Swamps on the Paranahyba River, Prov. Goyaz, Brazil. He, uo. 2978.

\ .i seen i>\ the writer.

— — Suffrutescent : rays very small, inconspicuous.

2. M. . 8UFFRUTIC09UM, Baker in Mart. Fl. Bras. iv. pt. •">. 162 | L884).
( »n the Esmeralda plains of the upper Orinoco in S. Venezuela. A
species omitted from the Index rlewensis. The achenes are crowned by
a -hallow cup. otherwise the plant would be placed next .1/. camphoratum,
to which according to the original description it is presumably related.

■•- *• Species of Mexico and 8. United Stat

~> li'ays -.hurt and inconspicuous: appendage <>f the hood elongated, recurved
tiled: heads usually (but no! always) lubsessile or abort-peduncled,

ROBINSON. — SYNOPSIS OF THE GENUS MELAMPODIUM. 457

3. M. longicornu, Gray, Mem. Am. Acad. v. 321 (PI. Thurb.),
1854, where by misprint longicorne. — S. Arizona, near Ft. Huaclmca,
Lemmon, no. 2777 ; Sonora, Santa Cruz, Thurber, no. 937 (type) ; Chi-
huahua, near the city, Pringle, no. 10 ; San Luis Potosi, Parry & Palmer,
no. 443J.

■*-*■ ++ Ligules longer, exceeding the involucral bracts, conspicuous : peduncles

mostly long.

= Soft-stemmed, strictly herbaceous and annual.

a. Pubescence sliort, scanty : leaves oblong to linear, entire : appendage of the

hood long: involucre gamophyllous about to the middle.

4. M. appendiculatum. Slender, erect, sparingly pubescent an-
nual, 3 to 4 or more dm. high, branched almost from the base : leaves
thin, oblong to linear, attenuate at the apex, scarcely narrowed to the
sessile subauriculate base : obsoletely serrate to quite entire, the larger
ones (near the middle of the stem) 5 cm. long, 1 cm. broad : pedun-
cles 2 to 7 cm. long, erect, slender : involucre saucer-shaped or shal-
lowly cup-shaped, gamophyllous, the limb shallowly 5-lobed; the lobes
rounded or barely and very obtusely pointed, their margins scarious :
pubescence of the peduncles and involucres short and sparing : rays
8 to 10, oblong, yellow, 6 mm. in length, 2-3-toothed at the apex ; fruit
tuberculate, the conspicuous appendage a linear coiled awn from an
ovate-lanceolate somewhat 2-toothed base : pales scarious. — South-
western Chihuahua, Dr. Edward Palmer, no. 245 (collection of 1885).
Type in herb. Gray. This species has the outer involucre of M.
cupulatum, Gray, and the fruit of M. longicornu, Gray, yet it is
clearty distinct from both, differing from the former not only in its long
peduncles and 'well-developed ligules, but in stature and in the size of the
leaves, and from the latter in the presence of a hood and appendage
(both totally lacking in M. cupulatum) and in the subauriculate base of
the leaves.

Var. leiocarpum. Similar in all points but the fruit smooth, striate,
glandular-punctate, not at all tuberculate. — Collected by Dr. Edward
Palmer at Alamos, 16-30 September, 1890, no. 726. Type in herb.
Gray.

Var. sonorense. Involucre deeper, subcampanulate : fruit slightly
roughened : otherwise like the type. — Collected by C. V. Ilarttnan at
Cochuto, Sonora, 2 October, 1890, no. 71. Type in herb. Gray.

b. Pubescence short and stiff: leaves lanceolate, undulate: fruit hooded, but

the appendage shorter or sometimes obsolete.

5. M. arenicola. Decumbent or suberect, branching from near the
base; stems dark purple, covered with stiff white somewhat refluxed

458 PROCEEDINGS OF THE AMERICAN ACADEMY.

hairs : leaves lanceolate from a narrowed auriculate base, undulate to
sparingly and irregularly scabrous-pubescent upon both surfaces, 3 to
5 cm. long, 8 to 1 1 mm. broad: peduncles slender, pubescent, 3 to 7 cm.
long; heads often nodding, 1.2 cm. in diameter (including narrow yellow
entire or bidentate ligules); involucre shallow, saucer-shaped, the 5 divi-
sions united nearly to the middle, broad, scarious and ciliate at the
margin : fruit finely striate, punctate and slightly tuberculate, bearing
a well-developed hood surmounted by a slender recurved hispidulous
appendage not flanked by lateral teeth at the base. — Collected by F. II.
Lamb in sandy soil on Isla Piedra, Mazatlan, Sinaloa, 31 December,
1 89 I. do. 861a. Mr. Lamb's no. 380 also from Mazatlau differs in hav-
ing no tubercles upon the fruit and in the obsolescent appendage, yet it
is probably of the same species. Type in herb. Gray.

Pubescence copious, soft, long, villous: leaves ovate-lanceolate to ovate: ap-
pendage of the hood short : involucre gamophyllous only near the hase.

6. M. LONGIPILUM, Robinson. Involucre externally villous, its
divisions acutish. — Proc. Am. Acad, xxvii. 173 (1892). — San Luis

~i. Pringle, nos. 3G39, 4537.

= = Stems tending toward lignescence : roots at least in part perennial : species
of northern Mexico and southern United States.

Heads rather small, (including the rays) about 1 to 1.2 cm. in diameter:
leaves conspicuously sinuate or pinnatifid: rays thin, short.

7. M. cinkki im, DC. 1. c. (1836). Hood muticous. — Laredo, Texas,
Berhtndicr, who appears to have confused this with the variety ramosissi-
ntinu. so that his numbers cannot be depended upon.

Yar. BAMOSISSJMTJM, Gray. Hood mucronate. — Syn. Fl. i. pt. 2,
239 (1884), in part. M. ramosissimum, DC. Prodr. v. 518 (1836).—
Near Laredo, Berlandier, S. W. Texas and adjacent Coahuila, Palmer,
no.. 556, 557, 558 (coll. of 1880).

Var. ARGorriYLLUM, Gray. Hood muticous: leaves small, tomentose
upon both Mirfaces, canescent above, snowy white beneath. — Gray in
Wats. Proc. Am. Acad, xviii. 104 (1883 without description). —
Coahuila and Nuevo Leon, Palmer, no. 2068 (coll. of 1880).

>. M. in. wnir.M, Torr. & Gray, Fl. ii. 271 (1842). — The com-
mon, 'st form of our southwestern States. Kansas, Hamilton County,
Hitchcock, no. 250; W. Texas, Lindheimer, no. 636, Reverchtm, no.
1880*, Thurher, no. 128, Heller, no. 1632, Pope, Bigelow, Wislizentu ;
New Mexico, Thurber, no. 1105, Wooton, no. 117; Arizona, Rothrock,
do. 827, Palmer, no. 608, Pringhy coll. of 1884; Chihuahua. Pringle.
This plant has of late been generally regarded as a mere form of

ROBINSON. — SYNOPSIS OP THE GENUS MELAMPODIUM. 459

M. cinereum, DC. However, it differs conspicuously in its more entire
erect or ascending thickish leaves, its much larger heads (nearly or quite
twice as broad) and its long thickish, firm, persistent, and veiny rays.
It is a much commoner and more widely distributed plant than M. cine-
reum, and may be conveniently regarded as a specific type.

* * Lower surface of the leaves silky-villous, the pubescence more or less floccu-

lent and tending to be deciduous.

•*- Ligules shorter than or about equalling the fructiferous bracts : heads sessile

or short-peduncled.

9. M. sericeum, Lag. Hoods tipped by a slender recurved append-
age.—" Elench. Hort. Madr. 1805," Gen. et Spec. Nov. 32 (1816) ; DC.
Prodr. v. 518; not HBK. — Mexico, Mendez ; Oaxaca, Pringle, no.
6728; D urango, Rose, no. 3476; Jalisco, Hose, nos. 2819,3561; Espe-
ranza, Duges.

Var. exappendiculatum. Hood destitute of a mucro or appendage,

sometimes itself obsolete. — In mountains near Morales, San Luis Potosi,

Schaffner, no. 271 in part; base of Iron Mountain, Durango, Dr. E.

Palmer, no. 926 (coll. of 1896) ; Guanajuato, Prof. A. Duges, Pringle,

no. 5309 ; Federal District, Pringle, no. 7978 (form approaching M.

hispidum, HBK).

■*- +- Ligules conspicuous, usually much exceeding the fructiferous bracts : pedun-
cles long and slender.

++ Leaves (at least in part) pinnatifid ; segments rather broad.

10. M. americanum, L. Spec. ii. 921 (1753); Rel. Houst. 9, t. 21 ;
DC. Prodr. v. 518. — Vera Cruz, Mexico, Houston. With this clearly
figured plant from Vera Cruz I have been unable to match any speci-
mens from Southeastern Mexico. However, the following specimens
from the western coast probably belong here : Manzanillo, Xantus, and
Colima, Palmer, no. 136 (coll. of 1897), and no. 1172 (coll. of 181)1).

++ t-f Leaves, at least in part, deeply cleft, segments narrow, linear.

= Outer bracts of the involucre pointless, surrounded by a thin yellow somewhat

hyaline border.

11. M. line aril obttji, DC. Prodr. v. 518 (1836). M. sericeum,

Benth. in Oerst. Videusk. Meddel. 1852, p. 86, not Lag. — A well-marked

species represented by the following specimens : Oaxaca, Nelson, dos. 2809,

2339 (pathological) ; Chiapas, Nelson, no. 2949 ; Guerrero, hills near

Iguala, Pringle, no. 9162 ; Nicaragua, Oersted; Sinaloa, Pose, no. 3183.

= = Outer bracts of the involucre not membraneous-margined or colored,
herbaceous to the acuminate apex.

12. M. longipes. M. sericeum, var. longipes, Gray, Proc. Am. Acad.
xxii. 423 (1887).— Erect, 4 to 5 dm. high, widely branched: upper and

460 PROCEEDINGS OF THE AMERICAN ACADEMY.

lower leaves entire, lance-linear, acute at both ends, 4 to 5.5 cm. Ion**.
4 to 5 mm. broad, the middle cauline leaves deeply and pinnately 3-cleft
into liuear acute segments, finely pubescent above, flocculent-sericeous
beneath: peduncles filiform, springing from the forks, 2 to 5 cm. long:
heads 1.2 to 1.4 cm. broad (including the rather numerous well-exserted
narrow bright yellow ligules) : fruit tuberculate, the hood well developed
and passing gradually and without intermediate toothing into a long
slender spirally coiled appendage. — Jalisco, Mexico, on dry hillsides
near Tequila. Dr. Edward Palmer, no. 391 (coll. of 1886), C. G.
Pringle, no. 1598. Type in herb. Gray. This plant although in habit
identical with M. hetcmphylhun. Lag., is strictly herbaceous and annual.
This fact, together with the hooded and appendaged fruit, seems to war-
rant its separation. It is certainly distinct from M. sericeum, Lag.

-h+ ++ +* Leaves undivided.

13. M. Kunthianum, DC. Prodr. v. 519 (1836). M. sericenm, HBK.
Nov. Gen. & Spec. iv. 272, t. 398 (1820), not Lag. — Of this species I
have seen only a single and imperfect specimen in the De Candollean
Herbarium. The leaves are linear, or nearly so, and entire; the fruit is
provided with a well -developed hood but no appendage. This species
also exhibits a suspicious resemblance to M. heterophyllum, Lag., and it
may represent Lagasca's var. ft.

II. M. mi rr mm, Cass. Diet. lix. 238 (1829). M. manillense, Less.
Linnaea, vi. 155 (1831). After examining authentic material of this
species in the Prodromus herbarium I can confidently refer to it Dr.
Palmer's nos. 3 and 281 from Acapulco, Mexico (coll. of 1895). The
ies bas been hitherto recorded only from the Island of Luzon. As
the genus as a whole is American, aud as this species is now found to be

• an American plant, its occurrence iu the Philippines may very likely
1" due to introduction. At all events it seems from the distribution
of tie- other species more likely that this plant has been carried from
Mexico to the Philippines, than the reverse.

Var. lanceolatum. M. lanceolatum, DC. Prodr. v. 519 (1830). —
Fruit with a short hood but no appendage : otherwise closely like the
typical form. — Collected by Nee, but the locality unknown. Nee visited
both Acapulco, Mexico, and tie- Philippine Islands.

2. Zarabellia, DC. Fructiferous bracts not exceeding the in-
1 achenes, nor developed into a cup. hood, or appendage at the
Bummit. — Prodr. v. 519 (1886). Zarabellia, Cass. Diet lix. 240.
• Peduncles lung and Blender: lignlea well exeerted, conspicuous.
<r- Leaves sericeous beneath.

ROBINSON. — SYNOPSIS OP THE GENUS MELAMPODIUM. 461

++ Leaves of diverse forms, partly entire, partly cleft : Mexican.

= Peduncles long : rays conspicuously exserted : sterile flowers numerous : pales

yellow-tipped.

15. M. heterophyllum, Lag. Gen. et Spec. Nov. 33 (181 G) ; DC.

1. c. — Tantoyuca, Huasteca, Berlandier, nos. 741, 2161, Ervendberg,

no. 80; between San. Luis Potosi and Tainpico, Palmer, no. 1103 (coll.

of 1879).

= = Peduncles short or none : rays 3 to 4, minute : sterile flowers 1 to 3 : pales

purple-tipped.

1G. M. Pringlei. Root perpendicular, long, with fibrous branches,
probably annual ; stem copiously branched from the base ; the branches
terete, purple, hirsute : leaves chiefly undivided, narrowly elliptic-lan-
ceolate, 2 to 3 cm. long, 3 to G mm. broad, acutish, finely subappressed-
pubescent above, snowy-sericeous beneath : lower heads short-peduncled
from the forks, the upper subsessile, lateral and terminal, all small and
few-flowered; outer bracts of the involucre 5, lance-linear and acute or
somewhat spatulate, 3.5 mm. long, so narrow as to disclose the young
fruit at an early stage in its development : ray-flowers 3 to 4 ; ligules
very small, oval or sub-orbicular, entire or slightly 2-toothed: pales pur-
ple tipped ; disk-flowers 1 to 3, mostly reduced to a clavate rudiment :
fruit of the ray-flowers obovoid, strongly tuberculate, destitute of hood,
cup, or appendage. — Collected by C. G. Pringle, at Las Sedas, Oaxaca,
altitude 1,850 m., 15 September, 1894, no. 5722. Type in herb. Gray.

++ ++ Leaves all undivided, linear : South American.

17. M. angustifoltum, DC. 1. c. (1836). — Peru, Haenke.

•<- -t- Leaves not sericeous beneath.
++ Slightly lignescent perennial : Lower California.

18. M. sindatum, Brandegee, Proc. Calif. Acad. Sci. ser. 2, iii. 144

(1891). — San Jose del Cabo, Brandegee, no. 302.

++ -w- Mexican and S. American species: annuals except M. montanum.
= Leaves cordate- or auriculate-clasping at the base.

19. M. Rosei. Erect annual, 3 to 4 dm. high ; stem purplish, copi-
ously branched almost from the base, covered with short white retrorse
hairs: leaves oblong-lanceolate, attenuate, sinuately toothed, slightly
narrowed to auriculate-clasping somewhat connate bases, sparsely pubes-
cent upon both surfaces, scabrous on the margin, scarcely paler beneath,
the larger about 7 cm. long, 2.5 cm. broad : peduncles in the forks of the
stem, 7 cm. long, filiform, retrorsely pubescent; outer involucre ."-parted,
the divisions broadly ovate or suborbicular, rounded or barely pointed at
the apex, pubescent upon the back : ray-flowers 8 to 10, ligules obloDg,

4G2 PROCEEDINGS OF THE AMERICAN ACADEMY.

golden yellow, 2-toothed at the tip, about 4 mm. long ; fructiferous bracts
i without hood or appendage) marked with 3 rows of tubercles upon each
lateral surface. — Collected by Dr. J. N. Rose between Rosario and
Conception, Sinaloa, no. 3271. Type in herb. U. S. Nat. Museum.

Var. subintegrum. Leaves clasping at the base but their margins
anlobed, obsoletely crenate-serrate. — Collected by Dr. J. N. Rose at
Rosario. Sinaloa. 7 July, 1897, no. 1568.

20. M. inimulifolium. Dichotomously branched herb; stems pur-
plish-streaked, loosely villous and finely pubescent along a longitudinal
line : leaves lanceolate, acuminate, entire or obsoletely serrulate, 5 to 7 cm.
long, 1.2 to 2 cm. broad, appressed-pubescent upon both surfaces, some-
what narrowed to an auriculate amplexicaul base: peduncles in the
forks of the stem, 4 cm. long, filiform, covered with a fine spreading
pubescence: outer involucre 5-parted; segments ovate-lanceolate, acute,
somewhat accrescent, appressed-villous : ligules about 8, short-oblong,
2 3-toothed : fruit short, broad, somewhat quadrate, without hood or
appendage, compressed, thin, depressed upon the lateral faces, finely
tuberculate dorsally. — Collected by E. W. Nelson in the vicinity of
Totontopec, Oaxaca, altitude 1700 to 2150 m., 15 July, 1894, no. 740.
Type in herb. U. S. Nat. Museum. The foliage recalls that of Stimulus
ringens.

= = Leaves neither cordate nor auriculate-clasping (except obscurely so in

M. gradlt ).

a. Divisions of the outer involucre 3, ovate to ovate-lanceolate, acute or acuminate.

21. M. panicdlatum, Gardn. in Hook. Lond. Jour. Bot. vii. 287
(184*). — Eastern and Central Brazil.

22. M. GRACILE, Less. Leaves rhombic, unlobed or panduriform,
more or less narrowed to an obscurely auriculate or at least obtuse base.
— ■ Linnaea, vi. 407 (1831). — Papantla, Schiede & Dcppe ; Tautoyuca,
Buasteca, Ervendberg, no. 92; Jalapa, C. L. Smith, no. 1605.

_'.;. M. OBLONGIFOLIUM, DC. Leaves oblong to oblong-lanceolate,
acute to attenuate at the base, never lobed. — Prodr. v. 519 (1836). —
Vera Cruz, near Orizaba, Ihtteri, no. 809, Section, no. 461, Cordova,
Bourgeau, no. L628, near Tantoyuca. Berlandier, no. 733; Oaxaca, at
San Felipe, Conzatti & Gonzalez, no. 560: Michoacan, Pringle, no.
1322 MoreloB, Pringle, do. ~-'>'2\ ; Costa Rica, Pittier, no. 6963.
6. Divisions of the outer involucre 6, obliquely acuminate.

24. M. MICEOCEPHALUM, Less. Linnaea, ix. 268 (1834). — This

species is known to me only from Lessinn's characterization and from an

ilent tracing, prepared from the type ;it Berlin by Mr. J. M. Green-

ROBINSON. — SYNOPSIS OP THE GENUS MELAMPODIUM. 463

man. It is evidently close to 31. paludosum in habit and foliage, differ-
ing chiefly, as Lessing himself notes, in its obliquely acuminate instead
of rounded or obtuse involucral bracts.

c. Divisions of the outer involucre obovate, rounded or obtuse.

1. Decumbent perennial with elliptical discolorous leaves and pale yellow rays

(often tinged with purple).

25. M. montanum, Benth. PI. Hartw. 64 (1840). M. Liebman-
nii, Sch. Bip.in Klatt, Leopoldina, xxiii. 89 (1887). — Oaxaca, Graham,
Pringle, no. 4666; Chiapas, Ghiesbreght, nos. 174, 564; San Luis Potosi,
Pringle, no. 3818 ; Cumbre de Estepa and Yavesia, Liebmann, no. 232.

2. Erect annuals.

26. M. texellum, Hook. & Arn. Bot. Beech. 299 (1840). — Aca-
pulco, Sinclair.

27. M. cupulatum, Gray, Proc. Am. Acad. viii. 291 (1870). —
Sonora, Palmer, no. 20 ; Mazatlan, W. G. Wright, no. 1213; Alamos,
Palmer, no. 726 (coll. of 1890). This species may possibly prove iden-
tical with the preceding. Both are distinguished from the following by
their narrow lance-linear or oblong-linear leaves.

28. M. paludosum, HBK. Nov. Gen. & Spec. iv. 273 (1820). M.
divarication, DC. Prodr. v. 520 (1836). M. pumilum, Benth. PI. Hartw.
64 (1840), described from starved specimens. M. copiosum and M.
panamense, Klatt in Engl. Jahrb. viii. 41, 42 (1887), founded upon tri-
fling foliar variations without accompanying floral distinctions. Dyso-
dium divarication, Rich, in Pers. Syn. ii. 489 (1807). D. radiation,
Desf. Cat. Hort. Paris, 1829, p. 182. Alcina ovalifolia, Lag. " Elench.
Hort. Madr. 1805," Gen. et Spec. Nov. 32 (1816). A. ovatifolia, Jacq.
f. Eclog. i. 115, t. 78 (1815?). A. minor, Cass. Diet. lix. 243.
Wedelia ovatifolia, Willd. Suppl. 61 (1813). W. minor, Ilornem. Hort.
Hafn. 855 (1813). — A common weed throughout Mexico, Central
America, and also occurring in the West Indies. Highly variable in leaf
contour, length of ligules, etc., thus passing into many very diverse yet
seemingly unstable forms.

* * Rays short, inconspicuous, exceeded by the involucre : peduncles short or none,
-t- Leaves ovate-lanceolate, rounded at the subsessile base : Panama to Brazil.

29. M. camphoratum [Benth. & Hook. f. Gen. ii. 349 (1873)],
Baker in Mart. Fl. Bras. vi. pt. 3, 161 (1884). M. digynum, Benth. &
Hook. f. 1. c. ace. to Hook. f. & Jacks. Ind. Kew. ii. 188. Unxia cam-
phorata, L. f. Suppl. 368 (1781). U. digyna, Steetz in Seem. Bot.
Herald, 154, t. 30 (1852-1857). — Panama, Seemann, and Llanos de

4G4 PROCEEDINGS OF THE AMERICAN ACADEMY.

Cumaral, Colombia, Andre, no. 1120, to British Guiana and tropical
Brazil, where apparently common.

*- -•- Leaves narrowed to a petiole or an exauriculate base : stems solitary.
++ Leaves rhombic to elliptic-oblong, obscurely toothed, undivided.

30. M. flacciddm, Bentli. Vidensk. Meddel. 1852, 86. M. tenelli/m,

var. flaccidum, Benth. Bot. Sulph. 115 (1844). — Nicaragua near

Granada, Oersted; Costa Rica. San Francisco de Guadalupe, Tondux,

nos. 7187, 841)8; Tepic, Mexico, Hinds, Palmer, no. 1814 (starved

specimens).

*+ *• Leaves narrow, linear-oblong and unlobed or deeply cleft into narrowly oblong

segments.

31. M. mspiDi-M, IIl'.K. Nov. Gen. & Spec. iv. 273, t. 399 (1820).
1/. coronopifulium, Sch. Bip. in Hemsl. Biol. Cent.- Am. Bot. ii. 145

"1 ). without character. — Arizona, Apache Pass, and near Ft. Ilua-

chuea. Lemmon, nos. 331, 2795, Santa Rita Mountains, Pringle ; Souora,

Wright^ no. 1205; Chihuahua, Pringle, no. 297; Durango, Palmer, no.

486 (coll. of 189G) ; San Luis Potosi, Parry & Palmer, no. 444^ ;

Jalisco. Palmer, no. 260 (coll. of 188G), in part; Tacubaya, Bilimek,

no. 593, Schaffner, no. 195. — Except in the nature of the pubescence

this species closely simulates M. sericeum, Lag.

■*- I- +- Leaves obovate, narrowed to an exauriculate base : steins several from the

very base.

32. M. arvense. Prostrate spreading annual; root fibrous; stems
several, 1 to 2 dm. long, more or less branched, purplish, covered all
around with short weak white hairs : leaves obovate, ei.tire or obsoletely
crenate, rounded at the apex, 3-nerved above the acuminate and slightly
connate base, bright green and glabrous or nearly so upon the upper
BUrface, distinctly paler and hispidulous upon the nerves beneath, 1.2 to
2.5 cm. long, 1 to 1.6 cm. broad : heads very small, surrounded by small
ovate to orbicular foliaceous bracts and borne close in the forks of the
stem and also upon such short, lateral cymes as to appear axillary; outer
bracts of the involucre 2, ovate, distinct at the base, obtusely pointed:
ray flowers 1 to 3, disk flowers about equally numerous: fruits semi-
obovate, Btrongly compressed, reticulated upon the sides, more or less
tuberculate dorsally. — Collected by C. G. Pringle in the Valley of
Mexico, Federal District, 19 October, 1896, do. 7827 (type, in herb.
Cray), and in fields near Toluca, 26 September, 1892, do. 5257, also at
an earlier dale by Schaffner in mountains Dear Santa Angela. Nearest

1/ '. bibracU -niiiiii. Wats., but differiog markedly in the contour and cuneate
base of the leaves as well as in its prostrate several-stemmed habit.

ROBINSON. — SYNOPSIS OF THE GENUS MELAMPODIUM. 465

■*- -t- t- t- Leaves rhombic to oblong, narrowed to a sessile auriculate base.
++ Leaves oblong, relatively narrow.
= Outer involucral bracts 2, not accrescent, or scarcely so.

33. M. bibracteatum, Watson, Proc. Am. Acad. xxvi. 140 (1891).
Fields, Del Rio, State of Mexico, Pringle, no. 3230.

= = Outer bracts 4 to 5, obtuse, united at the base into a cup.

34. M. glabrum, Watson. Proc. Am. Acad. xxvi. 139 (1891). —

Guanajuato in valley near Irapuato, Pringle, no. 2821, and Jalisco near

La Barca, Pringle, no. 3863.

++ ++ Leaves broad, mostly obovate or rhombic : outer involucral bracts 5, con-
spicuously accrescent.

= Outer bracts of the involucre lance-oblong, acute, distinct nearly or quite to

the base.

35. M. longifolium, Cerv. ace. to Cav. Anal. Cien. Nat. vi. 303

(1803). M. rhomboideum, DC. Prodr. v. 520 (1836). — San Luis

Potosi, Parry & Palmer, no. 444 ; Valley of Mexico, Bourgeau, no.

868, Pringle, no. 6455, Harshberger, no. 176.

= = Outer bracts of the involucre ovate, obtuse or obtusish, connate toward

the base.

36. M. perfoliatum, HBK. Nov. Gen. & Spec. iv. 274 (1820).—
A common and well-marked weed throughout Mexico, also established
in S" California at Los Angeles, Parish Brothers.

Synonyms and Doubtful or Excluded Species.

M. achillaeoides, Hemsl. Biol. Cent.-Am. Bot. ii. 145 (1881) =Vil-
lanova achillaeoides, Less.

M. australe, Loefl. It. Hisp. 268 (1758) = Acanthospermmn brasihnn,
Schrank, ace. to Hook. f. & Jacks. Ind. Kew. ii. 188.

M. Baranguillae, Spreng. Syst. iii. 619 (1826) \_M. Baranquillae, DC.
Prodr. v. 521 J = Sclerocarpus africanns, Jacq., ace. to DC. Prodr. v.
521.

M.Berlerianum, Spreng. 1. c. An unrecognized and poorly described
West Indian plant, very likely not of this genus.

M. Lrachyglossum, J. D. Smith, Bot. Gaz. xiii. 74 (1888) =Jaegeria
hirta, Less.

31. copiosum, Klatt in Engl. Jahrb. viii. 41 (1887) = M. paludosui/i,
HBK.

M. coronopifolium, Sch. Bip. in Hemsl. 1. c. = M. hispid/an, HBK.

M. digynum, Benth. & Hook. f. Gen. ii. 349 (1873), ace. to Hook. f.
& Jacks. 1. c. = M. camphoratum, Benth. & Hook. f.

M. divaricatum, DC. 1. c. = M. paludosum, HBK.

VOL. XXXVI. — 30

466 PROCEEDINGS OF THE AMERICAN ACADEMY.

^^. Dotnbeyanum, DC. 1. c. 521, is a still doubtful species from Peru.

M. Hiklalgoa, DC. 1. c. = llildalgoa tcrnata, Llav. & Lex.

31. hirsutum, Benth. & Hook. f. I. c. ace. to Hook. f. & Jacks. 1. c.
:= M. camphoratum, Benth. & Hook. f.

.1/. humile, Sw. Prodr. Veg. Ind. Occ. 114 (17 '8S) = Acanthosper mum
fiumile, DC.

M. lanceolatum, DC. Prodr. v. 519 (183G) = 31. diffusum, var.
lanceolatum.

3f. Liebmannii. Sch. Bip. iu Klatt, Leopoldiua, xxiii. 89 (1887) =
31. montanum, Benth.

31. longifolium, Brouss. ex AVilld. Euuni. Hort. Berol. 934 (1809) =
(? i M. longifolium, Cerv.

31. manillense, Less. Liunaea, vi. 155, t. 2 (1831) = M. diffusum, Cass.

.1/. ovatifolium, Reichenb. Ic. Exot. t. 42 (1827) = M. paludosum,
UV,K.

31. /in immense, Klatt, 1. c. 42 = M. paludosum, HBK.

M. pumilum, Benth. PL Ilartw. G4 (1840) = starved 31. paludosum,
HBK.

3L ra?7wsissimum, DC. Prodr. v. 518 (183G) = M. cinereum, var.
ramosissim urn, Gray.

M. rhomboideum, DC. 1. c. 520 (1836) = M. longifolium, Cerv.

M. ruderale, Sw. Fl. Ind. Occ. iii. 1372 (1806) = Eleutheranthera
ovata, Poit., ace. to Hook. f. & Jacks. 1. c.

M. sericeum, Benth. in Oerst. Videusk. Meddel. 1852, p. 86, not
Lai.r. = M. linearilobum, DC.

M. sericeum, var. brevipes, Gra}', Proc. Am. Acad. xxii. 423 (1887)
= typical .)/. sericeum, Lag.

M. sericeum, var. longipes, Gray, Proc. Am. Acad. xxii. 423 (1887)
= M. longipes, Robinson.

.1/. ternatum, DC. ace. to Hook. f. & Jacks. Ind. Kew. ii. 188, =
Bidalgoa ternata, Llav. & Lex.

It must be frankly confessed that among the species here kept up the
following are to the writer still doubtful: —

M americanum, L., which, although the type of the genus, cannot be
matched by any BpecimeD from near the original station.

.1/. microcephalum, Less. Not as yet satisfactorily represented in tho
herbaria examined.

M. paludiro/</, Taub., the description of which suggests a Sclerocarpus,

M. panicvlatum, Gardn., which from description is not clearly sepa-
rable from M. oblongifolium, DC.

ROBINSON. — SYNOPSIS OF THE GENUS NOCCA. 467

II. — SYNOPSIS OF THE GENUS NOCCA.

NOCCA, Cav. Icon. iii. 12, t. 224 (1795) ; Pers. Syn. ii. 498 (1807) ;
La Llave & Lex. Nov. Veg. i. 31 (1824) ; Spach, Hist. x. 40 (1841).
Noccaea, Willd. Spec. iii. 2393 (1804) ; Jacq. Frag. 58, t. 85 (1805);
Spreng. Anleit. ii. 548 (1818) ; Less. Linnaea, vi. G95 (1831), & Syu.
151 (1832) ; not Moench. Lagasca, Cav. Ann. Cien. Nat. vi. 331
(1803). Lagascea, Willd. Enum. Hort. Berol. 941 (1809;; Spreng.
1. c. 549 (1818) ; HBK. Nov. Gen. & Spec. iv. 24 (1820) ; DC. Prodr.
v. 91 (1836); Benth. & Hook. f. Gen. PI. ii. 342; Hoffmann in Engl.
& Prantl, Nat. Pflanzenf. iv. Ab. 5, 212. — Heads 1- to rarely 2-flowered,
aggregated in dense cainpanulate »r subglobose capitate glomerules ;
these subtended by ovate to linear more or less specialized herbaceous
bracts ; proper involucre calyx-like, tubular, gamophyllous, 5-toothed.
Flowers alike, perfect, fertile.- Corolla with narrow proper tube, en-
larged cylindric throat, and 5-toothed limb, yellow to white or reddish
purple, well exserted from the surrounding involucre. Style-branches
long, attenuate ; achenes columnar or attenuate toward the base ; pap-
pus of 2 to several short scales or rudimentary. — Annual herbs or more
often shrubs, probably all natives of tropical America, a single annual
species now widely distributed in the tropics. Leaves chiefly opposite.
The name Nocca (given by Cavanilles in 1795 in honor of Dominico
Nocca, professor of botany at Padua) is clearly the one to be employed
for this genus by those who wish to apply consistently the generally
conservative Berlin Rules. From the definite characterization and ex-
cellent figure given by Cavanilles there can be no doubt as to the iden-
tity of his genus Nocca, and the fact that the name was taken up in the
same sense within fifty years by Persoon, Jacquin, La Llave, and Sweet,
should establish its validity. The form Noccaea, adopted by various
botanists from Cassini to Kuntze, may be regarded as a different spelling
of the same name. Although substantive in form it has no advantage
over Nocca commensurate with the indefiniteness which succeeds any
modification of a name as originally published, and it is preferable there-
fore to take the name in its earliest form. While always reluctant to
change any current generic name like Lagascea, I hesitate the less in
this instance from the fact that this genus has attained no importance
in horticulture or pharmacy, and its nomenclature has accordingly little
or no significance outside technical systematic botany.

468 PROCEEDINGS OP THE AMERICAN ACADEMY.

* Involucres 2-flowered.

1. N. biflora, O. Kuntze, Rev. Gen. i. 354 (1891), as Noccaea.
Lagascea biflora, Hemsl. Diag. PI. Nov. 33 (1879), & Biol. Cent.- Am.
Bot. ii. 13'.', t. 44 (1881). — Mexico, without locality, Parkinson, Salina
Cruz, C. C. Deam.

* * Involucres 1-flowered.

■<- Shrubs or perennial herbs : Mexican,
■w Flowers large: glomerules sessile or scarcely peduncled, even in age more or
less campanulate, the Bubtending bracts ovate to lanceolate, large and conspic-
uous: central and southern .Mexico.

= Leaves cordate-clasping at the sessile base.

2. N. heltanthifolia, Cass. Stem pilose: leaves scabrous above,
Bcabro-hirtellous beneath. — Diet. Sci. Nat. xxv. 104 (1822), as Noc-
caea. Lagascea helianthifolia, HBE. Nov. Gen. & Spec. iv. 25 (1820).
L. macrophyUa, Zucc. ace. to Steud. Nomen. ed. 2, ii. 4 (1841).
Noccaea macrophylla, Zucc. ace. to DC. Prodr. v. 92 (1836). Acapulco,
Mi rico, Humboldt &, Bonpland; Jalisco, Rio Blanco, Palmer, no. G64
(coll. of L886) ; Guadalajara. Pringle, no. 1824, in part.

Var. levior. Stem pulverulent-puberulent, destitute of the long
spreading pilosity characteristic of the typical form: leaves scabrous
above, scabrous-hirtellous beneath. — Colima, Mexico, Dr. Edw. Palmer,
no. 1148 (coll. of 1891) ; Zopelote, Tepic, altitude 615 to 924 m., Lamb,
no. 561.

Var. suaveolens. Stem pilose ; leaves soft-pubescent beneath. —

Lagascea suaveolens, IIBK. 1. c. (1820). Noccaea suaveolens, Cass. I. c.

105 (1822). Nocca latifolia, Cerv. in La Llav. & Lex. Nov. Veg.

i. 31 (1824). Lagascea latifolia, DC. Prodr. v. 92 (1836). — Tamau-

lipas, Santa Barbara, Berlandier, nos. 753, 2173 ; Guanajuato, Duges

(a form resembling the type) ; Oaxaca, Nelson, no. 1804, L. C. Smit/t,

DO. 964 : ConzatH & Gonzalez, no. 558, llolway, no. 3724; 1'uebla,

Conzatti, do. <s61 ; Vera Cruz, Mirador, Liebmann, no. Ill; Orizaba,

Schaffner, Bourgeau, no. 3342, Botteri, no. 499, Gray, Tantoyuca,

Ervendberg, no. 317; Chiapas, Ghiesbreght, nos. 146, 571 (form with

- exceptionally pubescent almost lanate beneath); Guatemala, J. D.

Smith's nos. 2410, 4227.

I. avea Bhort-petioled, ovate, 2 or .1 times as long as broad: flowers red or

purple.

". Indumentum of the stem spreading, dense, uniform in length, partially glandu-
lar: Leaves elliptic, obtusish, remotely denticulate.

8. N. N. BP. [?]. Branches Iignesccnt. covered throughout by a line

dense spreading pubescence, a part of the hairs glandular, others of

ROBINSON. — SYNOPSIS OF THE GENUS NOCCA. 469

equal length without glandular tip, not rubescent: leaves elliptical-ob-
tusish or very shortly acuminate, grayish-pubescent beneath, somewhat
scabrous upon both surfaces, shortly but manifestly petiolate, the upper
leaves much reduced, oblanceolate, acuminate, remotely denticulate,
much shorter than the rather elongated interuodes; secondary nerves
2 or 3, originating some distance above the base : bracts linear, acute,
exceeding the flowers. — Lagascea Moginniana, DC. Prodr. v. 92 (1836)
as to pi. of Haenke not of Mocifio. — Mexico without locality, Huenke,
in herb. De Candolle. For an excellent sketch of this plant and informa-
tion concerning it I am indebted to the kindness of Mr. Robert Buser.

b. Indumentum unknown : leaves ovate, acute, green upon both surfaces, finely
serrulate with approximate purple-glandular teeth (about 10 to the centimeter).

4. N. MogiNNiANA, O. Kuntze, 1. c. (1891), as Noccaea Mociniana.
Lagascea Mocinniana, DC. Prodr. v. 92 (183G) as to pi. of Mocino. —
Mexico without locality, Mocino.

c. Indumentum of the stem widely spreading, composite, consisting of a dense

short gray subglandular puberulence and a long white horizontally spreading
villosity : leaves ovate, serrate, the margins neither purple nor glandular ;
both surfaces grayish-pubescent.

5. N. Pringlei. Stems slightly lignescent, 1 to 2 m. high, brown, in
dried state striate, covered with a fine spreading gray puberulence and
long spreading white villosity : leaves ovate, somewhat rigid, 5.5 to
6.5 cm. long, 3 to 4 cm. broad, serrate with about 2 teeth to the centimeter,
3-nerved from much above the base and pinnately veined, gray-tomen-
tose upon both surfaces, especially so beneath, obtuse or acutish at the
apex, rounded at the base ; petiole 3 to 4 mm. long : branches of the
inflorescence rigid, divergent, with long internodes and spatulate bracts ;
glomerules rather loose, not many-flowered: involucres 1-flowered, cylin-
drical, upwardly canescent-villous, 1 cm. long, cleft nearly to the middle
into 5 erect very unequal teeth : corolla 1.2 cm. long, purple, covered
on the outside with a fine gray tomentum ; the tube very short, slender;
the throat cylindrical, 4 to 5 mm. in length ; teeth 5, ovate, scarcely
acute : mature achene not seen. — Collected by C. G. Pringle on lime-
stone ledges above Iguala, Guerrero, Mexico, 10 October, 1900, no.
8400. Type in herb. Gray.

d. Indumentum of the stem fine, appressed : leaves elliptic-ovate, finely or more

coarsely serrate, but not noticeably purple-glandular on the margin.

1. Leaves green, lucid, and scabrous-pubescent upon both surfaces. ,

6. N. RiGiDA, Cav. Ic. iii. 12, t. 224 (1795) ; Sweet, Brit. Fl. Card,
ser. 2, t. 26. Noccaea rubra, Cass. 1. c. 104 (1822). Lagascea rubra,
HBK. 1. c. 24, t. 311 (1820). — Mexico, Cordillera of Guihilaque,

470 PROCEEDINGS OP THE AMERICAN ACADEMY.

Berlandier, no. 1018; State of Mexico, Bonrgeau, no. 1235, Schaffncr,

no. 2'Jo, Pringle, nos. .".">'>'. ;;,s,j(',j '.Mj'.is ; Cuernavaca, Bourgem no. 1205.

2. Leaves green above, niveous-sericeous beneath.

7. N. Heteuopappus, 0. Kuntze, I.e. (1891) as Noccaea. Lagascea
Heteropappus, Hemsl. Diag. PI. Nov. 33 (1879). — Mexico, Parkinson,
without locality ; hillsides near Morelia, Michoacan, Pringle, no. 4541.

.. Indumentum of the stem spreading, compound, the short glandular hairs
much exceeded by a long non-glandular villosity ; leaves ovate-oblong, attenuate,
dull and gray-pubescent upon both surfaces, even the upper ones exceeding the
internodes.

8. N. tomentosa. Lagascea tomentosa, Rob. & Greenra. Proc. Am.
Acad, xxxii. 43 (1896). — Guerrero, Mexico, between Ayusinapa and
Petatlan, K. 11'. Nelson, no. 2121.

= = = Leaves oblong-oblanceolate, 4 to 5 times as long as broad, narrowed at the

base to a distinct petiole.

9. X. axgustifoija, O. Kuntze, 1. c. (1891) as Noccaea. Lagascea
angustifolia, DC. Prodr. v. 92 (183G). — N. W. Mexico, Seemann ;
Durango, Palmer, no. 853 (189G) ; Jalisco, Palmer, no. 643 (1886),
Pringle, no. 1784.

— *+ Glomerules slender-peduncled, usually raised much above the foliar leaves,
at length subglobose, the subtending bracts mostly small and narrow.

= Teeth of the gamophyllous involucre relatively long and narrow, lance-linear

to subulate.

a. Involucres soft-villous.

10. N. decipiexs, O. Kuntze, 1. c. (1891) as Noccaea. Lagascea
deeipiens, Hemsl. Diag. PI. Nov. 33 (1879), & Biol. Cent.-Am. Bot. ii.
140, t. 11, f. 1-4 (1881). — North Mexico in the Sierra Madre, See-
mann, no. 2056; Southwestern Chihuahua, Palmer, no. 145 (coll. of
1^""'); Sonora, at Guaymas, Palmer, no. 256 (coll. of 1887), Alamos,
Palmer, no. Ml (1890), La Tinaya, Hartmann, no. 249.

b. Involucre hirsute.

11- N. glandulosa. Lagascea glandidosa, Fernald, Bot. Gaz. xx.
584 (1895). — W. Mexico, head of Mazatlan River, II'. G. Wright, no.
1305; Rosario, Sinaloa, Lamb, no. 483.

= = Teeth of the involucre very short (1 to l\ mm. in length), ovate to deltoid-

lanceolate.

<i. Stem and floral branches pubescent to tomentulose.

12. N. Liebrnannii. Lagascea Liebmannii, Sch. Bip. in Klatt, Leo-
poldina, xx. 91 (1884). — Leaves soft-pubescent beneath: involucre

ROBINSON. — SYNOPSIS OF THE GENUS NOCCA. 471

finely villous, 6 to 7 mm. long : corolla 7 mm. long, puberulent, the teeth
ovate. — Pochutla, Oaxaca, Liebmann, no. 250 (type in herb. Bot. Gard.
Copenhagen ; fragment and good sketch in herb. Gray).

b. Stem and floral branches glabrous.

13. N. Palmeri. Slender-stemmed shrub, with terete divaricate gla-
brous branches ; interuodes long, in dried state finely ribbed : leaves
ovate, acute, scabrous both above and beneath with enlarged and indur-
ated bases of trichomes (in the manner of many Borraginaceae), nearly
or quite entire, 1.5 to 2.5 cm. long, abruptly narrowed at the base to
a very short petiole : floral branches opposite ; peduncles slender, 2 to
4 cm. long ; glomerules 2.5 cm. in diameter, subglobose, about 25-headed;
subtending bracts ovate-oblong to narrowly oblong, 6 to 8 mm. in
length, their pubescence similar to that of the leaves ; involucres terete,
10-nerved, 1 cm. long, sparingly villous; segments deltoid-lanceolate,
minutely ciliated: corolla 1 cm. long, essentially glabrous; its teeth ellip-
tic-oblong, 3 mm. in length, 2-nerved: pappus a finely fringed campanulate
cup. — Collected by Dr. Edward Palmer, at Colima, Mexico, 27 to 28
February, 1891, no. 1320.

-i- ■»- Annual: tropics of both hemispheres.

14. N. mollts, Jacq. Frag. 58, t. 85 (1809). Lagasca mollis, Cav.
Ann. Cien. Nat. vi. 333, t. 44 (1803). JSToccaea mollis, Cass. 1. c. 103
(1822). Lagascea Kunthiana campestris, Gard. in Hook. Lond. Journ.
Bot. v. 238 (1846). Lagasca parvifolia, Klatt, Ann. k. k. Nat. Tlofmus.
Wien, ix. 360 (1894). — Western and southern Mexico, tropical South
America, West Indies, also Bengal, etc.

III. — NEW SPECIES AND NEWLY NOTED SYNONYMY
AMONG THE SPERMATOPHYTES OF MEXICO AND
CENTRAL AMERICA.

Dioscorea platycolpota, E. B. Uline in litt. " Glabrous throughout;
stem stout, dextrorsely twisted ; leaves membranaceous, orbicular-cordate,
9 to 11-nerved; $ racemes short, solitary, densely flower-bearing nearly
to the base; flowers pedicellate, disposed in 3 to 5-flowered cymules,
which in turn are closely arranged on the pend;mt rachis ; perianth
greenish yellow, campanulate, equalling or surpassing the slender pedicel;
stamens three, rising divaricately from the somewhat fleshy receptacle,
a little shorter than the ovate segments of the perianth ; anthers extrorse.
Pistillate spikes short, solitary ; perianth very shortly stipitate ; style

472 PROCEEDINGS OF THE AMERICAN ACADEMY.

column Blender, its slender, divaricate branches bifurcate at the apex ;
capsules small and crowded, obovate. leathery in texture ; seed small,
corrugated, with peripheral wing. — Collected by C. G. Pringle on lime-
stone mountains near Iguala, State of Guerrero, Mexico, altitude 1,230
m.. 15 September, 1900, no. 9224. This plant exhibits close affinities
for 1). Pringlei, Rob., the essential floral characters being identical ; but
the striking difference displayed in its much larger and densely crowded
flowers serves to give it an entirely distinct aspect, and seems to justify
a specific description. Leaves 8 to 9 cm. in diameter. Racemes 12 cm.
long, with rather stout rachis. Perianth 5 mm. wide ; segments 4 mm.
long. Capsules 7 to 8 mm. wide."

Calochortus Pringlei. Bulb ovoid, 3.5 cm. long, covered by more
or less thickened somewhat reticulated fibres, and surmounted by a cylin-
drical mass of long linear fuscous scales, from the midst of which rises
the stem; this I dm. high, terete, glabrous, simple or branched, 3-5-
leaved : leaves flat, linear, attenuate, the basal 3 dm. long, the cauline
gradually shorter ; h-acts of the spathe 2, opposite, subequal, 2 to 3 cm.
long, lanceolate, acun mate; pedicels glabrous, 2 to 7 cm. long; flowers
3 cm. in diameter, dark purple or almost black: sepals narrowly obovate,
bluntly pointed or retuse, glabrous except at a small roundish area a
little below the middle on the inner surface: petals broadly obovate,
cuneate, 1.4 cm. long, two-thirds as broad, obtusely pointed, externally
glabrous, internally covered on all parts except the narrowed base by
rather coarse violet or yellow hairs : filaments 5 lines long, glabrous ;
anthers 3.5 lines long, apiculate : ovary glabrous; capsule acute at each
end, 2.5 cm. long, 8 mm. in diameter. — Collected by C. G. Pringle
in thin soil of the top of knobs of the Sierra de Tepoxtlan, State of
Morelos, Mexico, altitude 2.300 m., 15 September, 1900, no. 8435.
Type in herb. Gray. Evidently related to C. flavus, Schultes, f. hut
differing in its dark colored flowers, short and broad fruit, etc.

Mimosa ^canthoi \ur\. Benth., var. desmanthocarpa. Habit,
foliage, and thorns of the typical form : pods unarmed, the valves mi-
nutely fulvous-tomentose. — Collected by E. W. Nelson between San
Cristobal and Teopisca, Chiapas, Mexico, altitude 2,060 to 2,620 m.,
1 December, 1895, no. 8428. Type in herb. I". S. Nat. .Museum; frag-
menl in herb. < rray.

Mimosa eurycarpoides. Branches glabrous, Boft-woody, with

rather large pith, terete or Bubangular, armed by a few small scattered

spines: leaves large, (including the petiole) 1 .2 to 2 dm. long ; pinnae 7

12 pairs, 8.5 to 5 Cm. long, puberulent upon the rachis; leaflets 20

ROBINSON. — SPERM ATOPHYTES OF MEXICO. 473

to 30 pairs linear-oblong, 5 to 8 mm. long, 1.6 mm. wide, acutish, oblique
at the base, finely appressed-pubescent upon both sides : flowers capitate,
heads globose, 1.3 cm. in diameter; peduncles filiform, 1.5 cm. loug,
borne by twos and threes in the axils of the upper leaves : calyx turbinate-
campanulate, scarious, slightly cuspidate-toothed, 0.8 mm. long: corolla
2.3 mm. long, glabrous, not conspicuously ribbed : stamens 8 or 10 :
pod 4.5 cm. long, 1.2 cm. broad, unarmed, glabrous, dark colored, ob-
lanceolate, acute, thickish ; valves not segmented. — Collected by Dr.
J. N. Rose in -the foothills of the Sierra Madre, near Colomas, State
of Sinaloa, Mexico, 21 July, 1897, no. 1805. Type in herb. U. S.
Nat. Museum ; fragment in herb. Gray. To this species I should refer
also Dr. Rose's no. 3157, collected on a road between Acaponeta and
Rosario. While it is more copiously armed it shares all important char-
acteristics with the plant from Colomas. The species is readily distin-
guished by its smooth unjointed fruit and axillary inflorescence.

Mimosa ionema. Shrub or small tree 3 to 5 m. high ; branches
unarmed, terete or nearly so, puberulent ; branchletsi green, angled and
costate, tomentulose : stipules filiform, 8 mm. long, tomentulose under a
lens ; petioles angled, straight, puberulent-tomentulose, 4 to 6 cm. long,
equalled by the foliar rachis; pinnae 4 to 6 subremote pairs, 5 cm. long,
the lower ones commonly alternate ; stipels 1 mm. long ; leaflets about
15 pairs, oblong, obtuse, green and nearly or quite glabrous upon both
surfaces, slightly paler beneath, oblique at the base, excentricallv 1 -nerved,
pinnately veined, 1 cm. long, 3 to 4 mm. broad : flowers loosely spicate,
spreading or somewhat deflexed ; spikes fascicled by 3's and 4's in the
axils of the upper leaves, ascending, 6 cm. long, on penduncles 2 to 3 cm.
in length: corolla white or pale yellow, 2.5 mm. long, cleft to below the
middle, without conspicuous nerves or ribs, glabrous ; segments 4 to "».
lance-oblong, acutish: stamens 8 to 10; filaments violet-tinged: fruit
(immature) narrowly oblong, acute at both ends, thin, glabrous, 5 cm.
long, 7 mm. broad, about 8-seeded. — Collected by C. G. Pringle in the
valley below Cuernavaca, Mexico, altitude 1,230 m., 17 October, 1900,
no. 8377. Type in herb. Gray.

Mimosa Watsoni. Fruticose and probably scandent: stems sub-
terete, tomentulose chiefly along six longitudinal lines, these bearing
minute approximate subequal very short recurved spines: branchleta
grayish-tomentose : petioles (including the rachis) 5 cm. long, rigid,
angled, armed laterally and ventrally by 2 or 3 lines of recurved spines,
tomentulose especially along the armed angles; pinnae 2 pairs; the
lower bearing 1 or 2 pairs of leaflets, the upper with 2 or 3 pairs ;

474 PROCEEDINGS OF THE AMERICAN ACADEMY.

leaflets rhombic-ovate or obovate, lucid and iinely pubescent above, to-
mentulose beneath, destitute of setae and ciliation, piunately veined ; the
lateral 1.5 to 2.5 cm. long, subsymmetrical, the terminal twice as large,
oblique and with excentric miduerve : globular heads (7 mm. in diameter)
white, racemosely disposed on the long spreading unarmed grayish to-
mentose branches of a loose panicle ; pedicels slender, 1 cm. long : corolla
4-parted to the middle, its lobes short-oblong, rounded at the apex :
-; Linens 8: pods 5 cm. long, 7 to 10 mm. broad, glabrous and lucid
upon the disarticulating valves but finely papillose under a strong
lens, dark brown, unarmed except for a few scattered minute recurved
spines upon the tomentulose replum; rectangular segments thin, broader
than long. — Collected by the late Dr. Sereno Watson in the eastern
portions of Vera Paz and Chiquimula, Guatemala, in 1885, nos. 323 in
fruit and 185 in flower. Types in herb. Gray. This species of the sub-
genus Habbasiti is closely related to the South American M. Spruceana,
Benth. From this it differs, however, in the much less dense and less
rufescent pubescence of the stem, more strongly armed petioles, less numer-
ous leaflets of different venation and finally in its much smaller fruit,
which i-> at full maturity less than half as broad as in the Brazilian plant.

Russelia Deamii. Stems becoming l.G m. long, 4-angled below,
6-angled toward the ends, copiously branched, glabrous except on the
smaller branchlets, with slender ribs at the angles: leaves 1.5 to 2 cm.
long, two thirds as wide, ovate, acute, incisely serrate except near the
cuneate base, green, loosely pubescent and sparingly punctate above,
slightly paler and strongly white-villous beneath especially about the
larger nerves and near the base : cymes numerous, 3-flowered : peduncles
Blender, 1 nun. long, pubescent as are also the filiform slightly longer
pedicels; bractlets linear: lobes of the villous calyx lance-acuminate
from an ovate base: corolla scarlet, l.G to 2 cm. long, the nearly equal
lobes suborbicular : fruit not seen. — Collected at Cuernavaca, Mexico,
7 July. 1900, by Charles C. Deam, no. 30. This species differs from
R. sarmentosa, .Tacq., in its villous more incisely toothed leaves and
6-angled branches, from Ji. jaliscensis, Rob., in the form of its calyx-
lobes, from A', polyedra, Zucc, in its larger flowers and glabrous less
Btrongly ribbed stems as well as in the form and sharp dentation of the
leaves.

Russelia trachypleura. Stems 6-angled; angles prominent, rib-
like, pale, roughened by -mall scattered callosities; areas between the
ribs flat, green, glabrous or (especially near the nodes) somewhat hairy ;
branches usually 4-angled : leaves temate on the stem, opposite on the

ROBINSON. — SPERMATOPHYTES OF MEXICO. 475

branches, elliptic-ovate, acute or acutish at each end, short-petioled,
sharply serrate, green, resinous-dotted, and pubescent on the upper sur-
face, slightly paler and pubescent upon the pinnately arranged veins
beneath : flowers in short few-flowered axillary cymes : calyx-lobes ovate,
caudate-acuminate, externally pubescent toward the sharp tip, 4 mm.
long: corolla bright scarlet, 1.2 cm. long, with cylindrical tube and 4
short rounded subequal lobes, the upper one broader and emarginate. —
Collected by C. G. Pringle on the Sierra de Tepoxtlan, State of
Morelos, Mexico, altitude 2,300 m., 11 September, 1900, no. 9445.
Readily distinguished from all the other species by the callosities on
the rib-like angles of its stems.

Piqueria pyramidalis. Stem terete, 2 to 2.5 m. high, puberulent,
green but maculate with elongated dark brown or purplish dots: leaves
alternate (at least the upper ones), petiolate, broadly ovate, shallowly
about 7-lobed, coarsely crenate, 3-nerved from the rounded to strongly
cordate base, green and scabrous-puberulent above, paler and tomentulose
beneath, the larger 1.7 dm. long and about as broad; petioles subterete,
tomentulose : small and very numerous heads in pedicellate racemose
glomerules ; these forming a large leafy-bracted pyramidal panicle : in-
volucral scales oblong, green, about 2-seriate, 2.5 to 3 mm. long, puberu-
lent and covered with minu'e amber-colored particles upon the outer
surface: corolla white, 3 mm. long, with short proper tube and relatively
large throat, also bearing a few amber-colored particles : styles much
exserterl, clavate, purplish or brown ; achenes dark-colored, glabrous,
lucid, 2.5 mm. long. — Collected by C. G. Pringle in shade of cliffs on
mountains above Iguala, altitude 1,230 m., 10 October, 1900, no. 8389.
Type in herb. Gray. This species, although possessing all the technical
characters of the genus, differs considerably in habit from the other Mex-
ican species, being in fact nearer some of the South American.

Ageratum lucidum. Shrub with buff cortex and opposite spread-
ing curved-ascending terete finely striate glandular-puberulent branches :
leaves opposite, ovate, acutish, serrate from below the middle, thin, veiny,
glabrous or early and completely glabrate upon both surfaces and lucid
especially above, 4 cm. long, half as broad, ciliolate upon the margin,
3-nerved from somewhat above the abruptly acuminate shortly petiolate
base, minutely white-dotted beneath and also covered with globular
resinous or glandular atoms : corymbs long-peduncled (often irregularly
compound), 2-6-headed and subtended by reduced opposite lance-oblong
to linear sessile bracts; pedicels 1 to 2.4 cm. long, curved-ascending,
1-headed with or without 1 or more filiform bractlets ; heads campanu-

476 PROCEEDINGS OF THE AMERICAN ACADEMY.

late, 70—1 00-flowered, Dearly 1 cm. in diameter: outer involucral scales
linear-filiform, ciliated, the inner lance-oblong, acuminate, rigidulous, the
3cariou8 margins erose: achenea black, 2 lines long, sharply angled,
glabrous, very finely and transversely striate: pappus a shallow un-
toothed cup without awn-. — Collected by C. G. Pringle, on mossy sides
of conglomerate knobs of the Siena de Tepoxtlan, near Cuernavaca,
State of Morelos, .■Mexico, altitude 2,310 m., 31 October, 1900, no. 83G2,
and previously 15 March, 1899, no. 7851. Well marked. Type in
heil). Gray.

Ageratum rhytidophyllum. Shrub with opposite terete dark red
branches covered with a fine crisped cinereous puberulence: leaves op-
posite, Bubsessile, lance-oblong, 4 to 7 cm. long, 1.2 to 2 cm. broad,
entire or obscurely serrate, revolute at the margin, acute at both end-.
thickish, grayish green, scabrous-pubescent and strongly rugose above;
veins much reticulated and prominulous beneath where covered by
eading white pubescence ; interstices covered by aureous par-
ticles: intlorescence of compound corymbs terminal upon the brandies ;
ir; beads rather small. 5 mm. in diameter, nodding on short
glandular-puberulent pedicels, about 25-flowered; involucral scales un-
equal, pluriseriate, lance-linear, acute, pungent, green, striate, finely and
sparingly pubescent; pales similar but narrower, rather rigid: corolla
.'! mm. long, glabrous but covered with aureous particles; the proper
tube greenish, about equalling the combined length of the whitish throat
and lavender colored limb: achenes glabrous; pappus a short obscurely
toothed cup. — Collected by C. G. Pringle on the Sierra de San Felipe.
Mexico, altitude 2,150 m.. 5 October, 1894. no. 5675. Type in
herb. Gray. Also secured in an immature state somewhat earlier (20
September, 1894) in the valley of Oaxaca by K. W. Nelson, no. 1446,
and a little later (4 November, 1894) in the mountains of Sau Joan
del Estado by Rev. L. C. Smith, no. 277.

Ageratum stachyofolium. Erect perennial with long tough
• nn white roots; Btem 5 to 6 dm. tall, terete, finely striate, purplish,
densely c ivered with fine crisped white hair-: leaves elliptical. e or

nearly so, mostly alternate, crenate, obtuse, pubescent upon both Bur-
t i ses, 3-nerved. <l>-ep green above, paler green and veiny beneath, 3 cm.
long, half as broad: corymbs terminal, simple, regular, 7— 10-headed;
bractlets filiform-spatulate ; pedicels _ to 2.5 cm. long, canescent-tomen-
tulose; heads large for the genua, L.2 cm. in diameter, about 100-flow-

f the involucre linear, BCUte, Btrongly striate, hirsute:

achenea glabrous, sharply angled, dark brown, somewhat tapering toward

ROBINSON. — SPERMATOPHYTES OF MEXICO. 477

the base, 2 mm. long; pappus a short scarious un toothed cup without
awns. — Collected by E. W. Nelson in the vicinity of La Parada,
Oaxaca, altitude 2,310 to 2,620 m., 19 August. 1894, no. 991. Types in
herb. Gray and herb. U. S. Nat. Museum.

Eupatorium anisopodum. Herbaceous perennial : stems terete,
decumbent, flex uous, irregularly branched, puberulent especially near the
nodes and under a strong lens, green or purplish tinged, 3 to 5 dm. high :
leaves ovate or rhombic-ovate, opposite, 2 to 2.5 cm. long, 1.2 to 1.7 cm.
broad, thickish, not punctate, crenate-serrate from the broadest portion
to the acutish apex, covered above with fine short curved hairs, slightly
paler beneath and appressed-pubescent upon the veins, 3-5-nerved, sub-
cuueate at the base to a short petiole (5 mm. in length) : bracts ovate-
lanceolate, acute, 3 to 7 mm. long, the lower petiolate the upper sessile ;
heads small, 25-30-flowered, 5 to 7 mm. in diameter, not numerous, irregu-
larly corymbose on pedicels of very unequal length ; scales of the turbi-
nate-campauulate involucre pluriseriate, very unequal, the outer short,
ovate, acuminate, herbaceous, pubescent ; the inner oblong, acute, pale,
2-3-nerved, ciliolate, sparingly pubescent or glabrous: corolla 2.2 mm.
long, probably white, glabrous except at the limb where under a strong
lens puberulent, the proper tube short, considerably exceeded by the
subcylindric scarcely ampliate throat : achenes columnar, at length black,
hispidulous on and between the five nerve-like angles, 1.5 mm. long;
pappus-bristles 15 to 20, bright white, about equalling the corolla. —
E. pycnocephalum, Coult. in J. Donnell Smith, Enuiu. PI. Guat. ii. 94
(1891), not Less. — Collected by H. von Tiirckheiui at Santa Rosa,
Department Baja Vera Paz, Guatemala, altitude 1,540 m., April, l-< 7,
no. 1177 of Mr. J. Donnell Smith's Guatemalan set. Type in herb.
Gray. While possessing something the habit of E. pycnocepkalum, Less.,
this plant is readily distinguished by its very different pluriseriate
involucre.

Eupatorium araliaefolium, Less. Linnaea, vi. 403 (1831). Add
syn. E. heterolepis, Robinson, Proc. Am. Acad. xxxv. 335 (1900).

Eupatorium Bigelovii, Gray, Bot. Mex. Bound. 75 (1859). All
syn. E. madrense, Wats. Proc. Am. Acad. xxvi. 137 (1891).

Eupatorium coxspicuum, Kunth & Bouche, Ind. Sem. Hort. Berol.
1847, p. 13. Of this, E. grandifolium, Regel, Gartenflora, i. L02, t. 12
(1852), is certainly a synonym.

Eupatorium Coulteri. Stem slender, straight, terete, densely fus-
cous-puberulent : leaves opposite, deltoid-ovate, acute or subcauda
somewhat hastate lobed at the almost truncate base, shallowly dentate,

478 PROCEEDINGS OF THE AMERICAN ACADEMY.

thin, harsh and slightly scabrous upon both surfaces, 4 to 5 cm. long,
half as broad, minutely punctate, slightly pubescent upon the nerves ;
slender fuscous-pubescent petioles 1 cm. long : heads in rounded axillary
and terminal thyreoid panicles; pedicels filiform, fiexuous, covered with
fine spreading purple pubescence; bracts and bractlets subulate, minute;
involucre turbinate, the lower much reduced scales somewhat decurrent
upon the pedicels, the inner scales oblauceolate-oblong, acuminate, erose,
puberulent or granular dorsally, thin, striate, ofteu purplish-tinged ;
flowers about 8 : corolla tubular, slightly and gradually narrowed from
the summit to the base, essentially glabrous, slightly exceeding the bar-
bellate pappus: achenes dark-colored, upwardly hispid upon the prominent
augles. — E. ageratifolium, var. purpureum, Coulter, Bot. Gaz. xvi. 98
(1891). — Collected by H. von Tiirckheim in Coban, Depart. Alta Vera
Paz, Guatemala, altitude 1,415 m., March, 1887, no. 52 of Mr. John
Dounell Smith's sets. This plant differs from E. ageratifolium, DC,
so greatly in pubescence, leaf-texture, involucre, and inflorescence that
intergradatiou does not appear likely.

Eupatoeium dasycakpum, Gray, Ptoc. Am. Acad. xxii. 420 (1886).
Add syn. Stevia rapunculoides, DC. Prodr. v. 124 (1836).

Eupatorium dryophilum. Perennial from a thickish branched
caudex ; stems several, erect, 5 to 7 dm. high, terete, finely striate,
puberulent or tomentulose, slightly scabrous, not glandular, reddish
brown, branched at the summit: leaves opposite or ternate, sessile, ovate
or oval, the lowest obtuse, the upper acute, all 3-nerved, shal lowly few-
toothed, thickish, reticulate-veiny, very scabrous upon both surfaces,
entirely destitute of glands or resinous particles: heads large, 1.4 cm.
in diameter, sleipler-piidicelled, erect; involucral scales thin, green, oval
to oblong, rounded at the apex, 4-5-seriate, striate, glabrous, mealy not
\ i- id : corollas purple, glabrous except at the short-toothed limb: pappus-
bristles sordid, barbellate, very unequal; achenes dark olive, 4 mm.
long, granular. — Collected by Dr. Edward Palmer, on the Rio Blanco,
Jalisco, October. 1886, no. 651, and by C G. Pringle with oaks and
pines "ii roteky hills near Guadalajara, nos. 2171, 2323. This species is
near II. pleianthum, Robinson, but differs from it in the absence of resin-
• .I- globules upon the leaves and tin- presence of mealiness upon the
involucre which is not viscid, also in its shorter darker achenes with fine
trau-verse striatum.

Eupatorium htssopinum, Gray, Proc. Am. Acad. xv. 28 (1880).
Add syn. A'. koeUiaefolium, Greene, Pittouia, iii. 31 (1896). Identity

ROBINSON. — SPERMATOPHYTES OP MEXICO. 479

Eupatoricm Lemmoni, Robinson, Proc. Am. Acad, xxvii. 171
(1S92). Add syn. E. euony mi folium, Greene, Pittonia, iii. 31 (1896),
founded upon a co-type.

Eupatoriuni Gonzalezii. Apparently herbaceous and to the unas-
sisted vision glabrous throughout ; sterns terete, green, finely striate,
lucid, purple at the nodes, slender ; the younger parts microscopically
puberulent : leaves opposite, long-petioled, deltoid-ovate, obtuse, coarsely
crenate-dentate, 3-nerved from the entire abruptly acumiuate base, thin,
bright green and glabrous upon both surfaces, 4.5 to 7 cm. long, 4 to
4.5 cm. broad ; petioles slender, 3 cm. long ; the uppermost leaves ovate-
lanceolate, entire, shorter-petioled : heads numerous, about 18-rlowered,
small, in compound axillary and terminal cymes ; peduncles and pedicels
ascending, obscurely puberulent ; bracts spatulate and mucronate to
linear and acute, small ; involucral scales green, striate, narrow, linear,
obtuse at the scarious erose and often crumpled apex, a few of the outer
considerably shorter, ovate, acuminate, ciliolate : corollas white, glabrous
throughout, the greenish white proper tube somewhat exceeded by the
pure white throat and spreading limb : achene dark, 1.3 mm. long, up-
wardly hispid on the angles ; pappus pure white, sparse ; the bristles
essentially equal, nearly as long as the corolla, slightly connate into a
minute cup at the base. — Collected by Professors C. Conzatti and
V. Gonzalez at El Fortin, Oaxaca, altitude 1,600 m., March, 1807,
no. 387.

Eupatorium leonense. Stem 4 mm. in diameter, slightly lignes-
cent, pithy, yellowish brown, glabrate ; branches opposite, curved-as-
cending, striate, tomentulose, finely pubescent : leaves opposite, ovate to
deltoid-ovate, long-petioled, thin, coarsely and rather bluntly few-toothed,
acutish, abruptly contracted at the entire base to an acuminate attach-
ment to the long petiole, finely pubescent when young, nearly or quite
glabrate at maturity, 3 to 5 cm. long, nearly as broad, 3-nerved from a
little above the base; petioles 3 to 4 cm. long, pubescent: pedicels
slender, flexuous, green, puberulent ; bracts minute, subulate ; heads
rather few in a small round-topped panicle, medium-sized, 8 mm. long,
about 12-flowered; involucre campanulate, loosely imbricated, lance-
linear, attenuate, very acute, green, striate, thin, sparingly white-pubes-
cent, the outer much shorter : corolla narrowly cylindrical, of essentially
uniform diameter throughout its length and without clearly marked
throat or proper tube, glabrous : style-branches strongly clavate, yellow ;
achenes black, glabrous, lucid, 2 mm. long with conspicuous yellow cal-
losity at the base ; pappus white, equalling the corolla. — Collected by

480 PROCEEDINGS OF THE AMERICAN ACADEMY.

C. G. Pringle cm the Sierra Madre near Monterey, Nuevo Leon, Mexico,
16 June, 1887, no. 2277.

Eupatorium Liebmunnii, Sch. Bip. in Klatt, Leopoldina, xx. 75
(1884). From the characters given and from an excellent drawing (in
herb. Klatt) prepared from the original material I cannot avoid the conclu-
sion that this is identical with the earlier E. hirsutum, DC, the type of
which I have recently examined in the Prodroraus Herbarium. The
species is represented by Mr. Pringle's no. 6046 from the foothills of the
Sierra de San Felipe, Oaxaca.

Eupatorium longifolium. Suffrutescent, 1 m. high : stems virgate,
terete, finely striate, covered by a fine spreading purplish and probably
viscid pubescence : leaves opposite, short-petioled, ovate-lanceolate, cre-
nate-serrate, 3-nerved, thin, dark green and strigillose (under a lens)
above, paler and tomeutulose especially upon the veins and veinlets
beneath, 1 to 1.2 din. long, 4 to 5 cm. broad, attenuate to a caudate
apex, rounded and deeply cordate at the base, the sinus narrow: inflor-

acea rounded-corymbose, together forming a large leafy oval or sub-
pyramidal panicle; its branchlets, slender pedicels, and filiform bracts
brown-pubescent ; heads very numerous, 4 to 5 mm. long, about 10-
flowered ; involucral scales linear, attenuate, subeqnal, 3 mm. long,

ered with jointed purple hairs and resinous lucid atoms: corolla
scarcely 2 mm. long, gradually contracted toward the base, nearly
equalled by the simple white pappus : achene dark, minutely pubescent,
1 . ■) mm. long. — Collected by C. G. Pringle in Tamasopo Canon, San
Luis Potosi, .Mexico, 28 November, 1890, no. 3372. This number was
distributed as E. Palmeri, Gray, to which it is obviously related. It
differs, however, both in the nature of its indnment and the form of the
leaves. The latter are rounded at the base in E. Palmeri while in
/-,'. longifolium they are deeply cordate. E. JUicaulc, Sch. Bip., is also a
nearly related species, but its heads are conspicuously racemose, which is
not the case here.

Eupatorium lucidum, Ort. Ilort. Matr. Dec. 35 (1797). An ex-
amination of authentic material of this briefly characterized species
-hows that it is just the plant to which I have recently assigned the
name /•,'. Capnoresbium, Proc. Am. Acad. xxw. 331, a name which must

rdingly sink into Bynonymy.
Eupatorium Luxii. Apparently shrubby : branches subterete, stri-
ate, covered with a line spreading and verv dark pubescence : lea\ < ■-

opposite, elliptic-ovate (the upper ovate-lanceolate), acute to acuminate

eh end, pinuately nerved. •'. to 13 cm. long, half as broad, serrate,

ROBINSON. — SPERMATOPHYTES OP MEXICO. 481

sparingly pubescent when young, quite glabrate except on the nerves
and- somewhat lucid in age, nigrescent in drying, the teeth salient, mu-
cronulate ; petioles 1.8 to 3 cm. long: heads medium-sized, about 25-
llowered, 8 to 10 mm. long, in an opposite-branched corymbose panicle ;
involucral scales 3-4-seriate, ovate to lanceolate, acumiuate, closely im-
bricated externally, nigrescent at the tip, pale at the base, ciliolate,
silvery and lucid on the smooth inner surface : corolla-tube 5 to 6 mm.
long, glabrous, gradually narrowed from the summit to the base, the
teeth very short, sparingly pubescent : achenes 2 mm. long, dark, glab-
rous, lucid, the angles very prominent ; pappus rather copious, white,
nearly equalling the corolla. — Collected by Heyde and Lux at Nebaj,
Depart. Quiche, Guatemala, altitude 2,150 m., April, 1892, being no.
3387 of Mr. John Donnell Smith's sets of Central American plants, and
having been distributed as E. Tuerckheimii, Klatt, a species with much
narrower leaves and linear-oblong involucral scales, glabrous stem, etc.,
well shown by no. 77 of Mr. John Donnell Smith's sets.

Eupatorium hjratum, Coulter, Bot. Gaz. xvi. 96 (1891), is Conyza
lyrata, II BK.

Eupatorium Mariarum. Herbaceous ; stems terete, weak, pithy,
obscurely pulverulent-puberulent and also covered with scattered spread-
ing white trie-homes; branches ascending: leaves opposite, long petioled,
deltoid, acute, coarsely crenate except at the abruptly contracted base,
thin, deep green and bearing a few scattered white trichomes above,
somewhat paler and essentially glabrous beneath, 3-nerved from the
slightly acuminate point of attachment, 5 to 7 cm. long, 4 to 5 cm.
broad ; petioles weak, flexuous, 3 to 5 cm. long, with a double pubes-
cence as on the stem : heads about 25-flowered, in small axillary and
terminal corymbs, pedicellate ; bractlets acute, lance-linear to filiform ;
involucral scales linear-oblong, acute, subequal, thin, mostly 2-nerved,
o-reen, 3 mm. long: corolla (probably wdiite) glabrous, with a slender
proper tube exceeding a much broader well-marked throat and spreading
limb: achenes dark-brown, slightly spindle-formed, hispiduloua toward
the summit. — Collected by E. AV. Nelson on Maria Madre Islam] of
the Tres Marias Group, May, 1897, no. 4244. At first taken for
E. pazcuarense, HBK., but clearly distinct by its long petioles, different
leaf-contour and dentation, also in its pubescence.

Eupatorium pachypodum. Caudex thickish, lignescent, branched;
stems annual, one to several, erect, simple to the inflorescence, 3 to 4 dm.
hio-h, terete, cinereous-tomentulose : leaves opposite, small, much ex-
ceeded by the internodes, ovate to suborbicular, acute, rounded at the
vol. xxxvi. — 31

482 PROCEEDINGS OF THE AMERICAN ACADEMY.

essentially sessile base, serrate, 1.3 to 1.8 cm. long, nearly as broad,
green and sparingly pubescent above, covered beneath by a short soft
ashy more or less deciduous pubescence, and bearing upon botb surfaces
golden resinous particles : panicle flat-topped, its slender ascending
brunches and pedicels (7 to 9 mm. long) grayish-tomentulose; bractlets
linear to filiform; heads about 12-flowered, 7 mm. long; involucral
scales subequal, oblong, soft-pubescent and pale green upon the outer
surface, 3 to 4 mm. long, obtuse : corollas white or at least pale, 4 mm.
long, with slender tube somewhat exceeded by the gradually ampliated
throat, glabrous : achenes 2 lines long, hispid on the angles, pappus
white, equalling the corolla. — Collected by C. G. Pringle on rocky
hills near Guadalajara, Jalisco, 26 May, 1891, no. 3718. Distributed
as E. scordonioid.es ? a species with shrubby perennial stems, deltoid-
peltate leaves, etc.

Eupatorium pansamalense. Stem glabrous, rather strongly
angled, striate, pithy : leaves opposite or ternate, slender-petioled,
rhombic-ovate, caudate-acuminate, acute at the base, mucronate-serrate,
thin, pinnately nerved, deep green and nearly glabrous except on the
midnerve above, paler and puberulent or tomeutulose upon the nerves
and veins beneath, 1 to 1.3 dm. long, half as broad ; petioles 2 to 3 cm.
long: branches of the inflorescence and filiform pedicels fuscous-tomen-
tulose ; heads about 35-flowered, very numerous in a round corymbose
panicle ; scales of the involucre in about 3 series, narrowly oblong, acute
to acutish, in a dried state stramineous except a central dark brown streak,
the more or less narrowed tips thin, erose : corollas slender, without
definitely marked throat, glabrous except near the limb : pappus white,
equalling the corolla; achenes very small, scarcely 1.4 mm. long, gla-
brous. — E. Tuerck/teimii, Coulter, Bot. Gaz. xvi. 97 and in J. D.
Smith, Enum. PI. Guat. ii. 95, as to no. 1342, not Klatt. — Collected
by II. von Yiirckheim at Pansamala, Depart. Alta Vera Paz, Guatemala,
altitude 1,170 m., April, 1888, no. 1342 of Mr. John Donnell Smith's
sets. /,'. '/'"ry/./irii/iii, Klatt, differs in having much narrower oblong-
lanceolate subcoriaceous sbort-petioled leaves which are quite glabrous
upon both sides, paler beneath, and marked by a fine transverse
venulation.

Eupatorium pinabetense. Shrub with angled and striate glabrous
branches: leaves opposite, lance-oblong, acuminate at each end, mucron-
ulate-serrate, smooth, green and glabrous upon both surfaces, 1 to
1.5 dm. long, 2 to 3 cm. broad, pinnately veined from a strong mid-
nerve : smaller branches of the rather dense thyrsoid panicle fuscous-

ROBINSON. — SPERMATOPHYTES OF MEXICO. 483

tomentulose under a lens; heads numerous, crowded, small, about 10-
flowered, short-pedicelled ; scales of the involucre very few, unequal,
elliptical, rounded at the apex, glabrous but erose-ciliolate, in a dried
state brown : corolla glabrous or nearly so even at the limb, without
clearly marked throat, equalled by the white pappus: achenes (young)
pale, glabrous, 1.8 mm. long. — Collected by E. W. Nelson near Pina-
bete, Chiapas, Mexico, 8 February, 1896, no. 3785. This species is
near but clearly distinct from E. daleoides, Ilemsl. and E. tepicanum,
Hemsl.

Eupatorium pleianthum. Doubtless perennial : stem slender,
terete, scabrous-puberuleut, .finely striate, reddish brown, loosely few-
branched above : leaves opposite, ovate, essentially sessile, thickish,
reticulate-veiny, coarsely few-toothed, acute, cordate, 3-nerved, 2.5 to
3.5 cm. long, two-thirds as broad, slightly scabrous upon both surfaces,
covered below by bright resinous globules : heads large, 2 cm. in diame-
ter, sleuder-pedicelled, erect ; involucral scales imbricated in 3 or 4 rows,
oval to oblong, rounded at the apex, thin, glabrous and glutinous, striate,
pale green, scarcely herbaceous : corollas 6 mm. long, gradually nar-
rowed from the summit to the base, about equalled by the copious stiffish
tawny pappus : styles very long and conspicuous, clavate ; achenes 5 to
6 mm. in length tapering toward the base, reddish brown, covered with
resinous globules, not tranversely striate. — E. adenospermiun, var.
pleiant/tum, Gray, Proc. Am. Acad. xv. 26. — Collected by Dr. Ber-
thold Seemann, Western Mexico. Type in herb. Gray. While sharing
many characters with E. adenospermum, Sch. Bip., this plant differs so
markedly in its opposite (not alternate), shorter, ovate not oblong leaves
of differing dentation that intergradation seems very unlikely.

Eupatorium prionobium. Branches ascending, terete, green,
striate, puberulent : leaves opposite, petiolate, deltoid, ovate to ovate-
oblong, cordate with open sinus and hastate tendency, acutish to rounded
at the apex, crenate or crenate-serrate, thickish, bright green but covered
with short scattered white hairs upon both surfaces, scarcely paler be-
neath, 2 to 4 cm. long, 1.7 to 2.5 cm. broad ; petiole puberulent, 6 mm.
long, sulcate upon the upper side : heads small, about 20-flowered, in
round-topped sessile cinereous-puberuleut corymbs ; pedicels 3 to 5 mm.
lon<r; bractlets subulate; involucre green, the scales lance-oblong, acute,
subequal, loosely imbricated, 2.5 mm. long, scarcely half the length of
the flowers : corolla in dried specimen white, 3 mm. long, with a rela-
tively long throat narrowed into a shorter tube, glabrous, equalled by
the bristles of the pappus; these somewhat unequal, barbellate, white

484 PROCEEDINGS OF THE AMERICAN ACADEMY.

but the barbules upon the lower half tipped with deep purple or violet :
achenes 1 .8 mm. long, upwardly hispid ou the angles. — Collected by
E. W. Nelson on the Sierra Madre. State of Chihuahua, Mexico, 29
September, 1899, no. 6499. Type in herb. Gray and herb. U. S. Nat.
Museum. Related to E. ageratifolium, DC, and E. occidcntale, Hook.

Eupatorium prionophyllum. Tree ; appearing glabrous to the
unassisted vision, but covered upon the brunchlets, petioles, veins of the
leaves, and pedicels by traces of a short close fuscous tomentum ;
branches curved; cortex gray: leaves opposite, slender-petioled. broadly
ovate, conspicuously acuminate, usually obtuse at the base, incisely and
often somewhat doubly serrate-dentate nearly from the base to the apex,
thin, green on both surfaces, pinnately veined, 7.5 to 9 cm. long, two-
thirds as broad ; the teeth acuminate, incurved ; petioles 1 to 4 cm. long:
heads 25-30-flowered, in terminal opposite-branched rounded at first
thyrsoid at length more open panicles; bractlets filiform ; scales of the
involucre imbricated in about 3 series, the outer short and ovate,
acute, mucronate with a glandular tip, the inner oblong, obtusish, all
striate fimbriate-ciliolate, externally stramineous becoming purplish or
fuscous in age especially near the tip, silvery and lucid within: corolla
glabrous, gradually narrowed from the summit to the base: pappus
white, barbellate, moderately copious, nearly equalling the corolla :
achenes at maturity dark brown, glabrous, lucid. — E. ixiocladon, Klatt.
Bull. Soc. Bot. Belg. xxxi. 190, not Benth. — Collected by Prof. H.
Pittier on the banks of the river Poros, no. 1705 (type in herb. Gray)
and near the Rancho Flores. no. 1900, Costa Rica. This species has
no close affinity with E. ixiocladon, Benth., a plant well shown by Mr.
J. 1). Smith's no. 7501.

El PATORIUM QUADRANGULARE, DC. Prodr. v. 150 (1836). Of this,
E. thyrsoideum, Moc, notwithstanding its supposed terete stems, is cer-
tainlv a synonym. Nothing beyond the younger branchlets of E. thyr-
toideum appears to be known, while even in the square-stemmed E.
quadrangulare these younger branches are often Bubterete.

Eupatorium viscidipes. Stem Blender, terete, dark-purple, mi-
nii'ely glandular and veryviscid; branches opposite, spreading, curved
upward : leaves opposite, delioiil-o\ ate. caudate-acuminate, crenate-senato
in the middle. 8-nerved from the obtuse or subtruncate base, 2.2 to 3.5 cm.
l<ui'_r. L.8 to 2.5 cm. broad, minutely puberulenl above, slightly paler and
punctate beneath with translucenl dots; petioles 1.2 to 1.7 cm. long: heads

rather numerous, small, d mm. in diameter, about 18-flowered, bome in

a huge Loose corymbose leafy-bracted panicle; pedicels filiform, 3 to

ROBINSON. — SPERMATOCYTES OF MEXICO. 4S5

13 mm. long, glandular and very viscid, covered with extraneous particles ;
involucre turbinate-campanulate ; the scales very unequal, much imbri-
cated, pluriseriate, the outer ovate-oblong, the inner oblong, subscarious,
striate, all obtuse or rounded at the apex, glabrous or slightly glandular,
viscid : corolla glabrous, white, 2.3 mm. long ; the proper tube short, much
exceeded by the scarcely enlarged cylindric throat : achenes nearly black,
columnar, hispidulous on and between the angles ; pappus bright white,
a little shorter than the corolla. — E. pycnocephalum, Coult. in J. Don-
nell Smith, Enum. PI. Guat. iv. 75 (1895), as to no. 3397, not Less. —
Collected by Heyde & Lux at Chicaman, Department Quiche, Guate-
mala, altitude 1,230 m., April, 1892. no. 3397 of Mr. J. Donnell Smith's
Guatemalan sets. Type in herb. Gray. To be distinguished from E.
pycnocephalum, Less, by its viscidity, quite different and much more
imbricated involucre, etc., and from E. anisopodum, described above, by
its viscidity, different leaf-form, rounded instead of acute or acuminate
involucral scales, etc.

Brickellia amblyolepis. Stem large, fistulose, 2 to 2.5 m. high,
glabrous, purple mottled with green : leaves whorled below, opposite
above, alternate in the upper parts of the open inflorescence, ovate, acu-
minate, 3-ribbed from the cuueate entire base, coarsely serrate from
below the middle to the acuminate apex, green and glabrous above, paler
and finely pubescent upon the veins beneath, 7 to 13 cm. long, half as
broad; petioles 1.3 to 3.4 cm. long: heads rather large, broader than
high, nodding in an open leafy panicle ; pedicels bearing a few linear to
spatulate bractlets ; scales of the involucre in 3 to 4 series, spatulate-
oblong, rounded at the apex, green and nearly glabrous except upon the
deep purple toinentulose margin : corollas greenish, 5 mm. long, glabrous,
nearly equalled by the rather copious purple-tinged pappus : achenes
(immature) 5 mm. long, minutely pubescent; styles strongly clavate,
bright orange turning to brown. — Collected by C. G. Priugle on moun-
tains above Iguala, Guerrero, Mexico, 4 October, 1900, no. 8415. A
stately species 2 to 3 metres high, not very closely related to any hitherto
described. Type in herb. Gray.

Brickellia cardiophylla. Herbaceous ; stem terete, glandular-
pubescent; branches opposite, widely spreading: leaves ovate-triangular,
acute or obtusish, entire or nearly so, broadly cordate at the base, mem-
branaceous, 3-nerved, minutely glanduhir-puberulent, green upon both
surfaces, scarcely paler beneath, 5 to 6 cm. long, nearly as broad; peti-
oles slender, 2 to 3 cm. long : heads subumbellate near the ends of the
branches, campanulate, 1 to 1.5 cm. in diameter; pedicels filiform, gland-

486 PROCEEDINGS OF THE AMERICAN ACADEMY.

ular-puberulent, 2 to 3 cm. long, naked or bearing a single minute lance-
olate bract; involucral scales appressed, imbricated in about 5 series,
from ovate to linear obtusish, striate, the outer herbaceous-tipped, the
inner often purplish ; achenes columnar, 2.5 mm. long, minutely rough-
ened or hispid ; pappus bright white, not copious, 7 mm. in length :
corollas in a dried state cream-colored. — Collected by C G. Pringle in
a harranca near Guadalajara, Jalisco, 1 May, 1894, no. 5885a. Type
in herb. Gray. Near B. floribunda, Gray, but heads fewer and consid-
erably larger: leaves subentire and of different contour.

Brickellia hebecarpoides. Erect herb, seemingly perennial but
root not seen ; stem straight, terete, striate, glandular-pubescent, copiously
branched above : leaves opposite (the upper with a slight tendency to be
alternate), ovate to ovate-oblong, acute or obtuse, membranaceous, finely
and regularly serrate (the rameal entire or nearly so), rounded or ab-
ruptly acuminate at the 3-5-uerved often oblique base to the slender
petiole, puberulent above, glandular-puberulent beneath, both surfaces
green, the lower scarcely paler; blanches slender, spreading, leafy:
peduncles solitary in or usually a little above the forks of the
branches, 1 to 2.5 cm. long, filiform, ascending or erect, glandular-pubes-
cent, bractless; involucre 1.2 cm. long, campanulate, its outer bracts
lance-oblong, acute, glandular-puberulent, herbaceous or at least herba-
ceous-tipped, half to two thirds the length of the lance-linear attenuate
striate often purplish inner bracts ; flowers about 30: corolla 1.3 cm.
long, slender, glabrous, purple-tinged at least near the limb : achenes
short, only 2 mm. in length, minutely pubescent; style-branches orange,
strongly clavate. — Collected by C. G. Pringle in a barranca near
Cuernavaca, State of Morelos, Mexico, altitude 1,540 m., 24 June, 1896,
no. 7332. Type in herb. Gray. This species is related to B. hebecarpa,
Gray, but differs in its scarcely pubescent achenes aud much longer
Outer involucral scales, as well as in other ways.

Brickellia petrophila. Compact shrub, 3 to G dm. high ; stems
branched from near the base; branches curved-ascending, becoming
erect, gray or lightish brown, covered with a fine spreading pubescence
i^ting of minute gland-tipped hairs and longer non-glandular hairs,
the latter readily distinguishable to the unassisted vision: leaves oppo-
site, small, reniform, ovate, crenate, obtw-c broadly and shallowly cordate
at tie- base, 7 to 10 (rarely 2.S) mm. long, 1 to 3 cm. broad, somewhat
rugose above, finely pubescent or even tomentulose upon both surfaces,
but scarcely or not at all canescent: heads Bubsessile, terminal on short
or more elongated branchlets either forming a cylindric racemiform or a

ROBINSON. — SPERMATOPHYTES OF MEXICO. 487

broader and more pyramidal inflorescence, the two forms occurring upon
the same individual ; involucre cylindric or in a dried state narrowly
oampanulate, the scales stramineous, striate, imbricated in about four
very unequal series, the outer ovate, obtuse, pubescent or at least ciliate,
the inner oblong-linear, acute, about 8 mm. long ; flowers about 20 :
achenes columnar, 3 mm. long, olive brown, slightly hispid on the ribs.
— Collected by C. G. Pringle on rocky hills near the city of Chihuahua,
13 October, 1885, no. 610 (type, in herb. Gray). To this species I
would refer the following specimens: from San Luis Potosi, Schaffner,
no. 365, Parry & Palmer, no. 351 ; from Guanajuato, Duges, no. 449;
from Zacatecas, Beam, no. 140; from Durango, Palmer, no. 753. This
species is nearly related to and has been generally confused with the
more southern B. veronicaefolia, Gray, a species which is readily distin-
guishable by the extremely fine canescent non-glandular pubescence of
its brauchlets, in which the individual hairs cannot be seen without the
use of a strong lens. The involucral scales of B. veronicaefolia are
more clcsely imbricated and in more numerous rows.

Var. umbratilis. Heads larger, 1.5 cm. in diameter, terminal upon
more elongated branches. — Collected by Dr. Edward Palmer in a shady
ravine at Parras, Coahuila, October, 1898, no. 438, and also a solitary
individual in a ravine at Mapimi, Durango, no. 521.

Brickellia vernicosa. Shrub with striate buff or at length gray
cortex and erect scabrous-puberulent branches : leaves alternate, ovate,
serrate, acutish, bright green, coriaceous, veiny, subsessile, minutely his-
pidulous, glandular-punctate beneath, often vernicose, 1 to 1.7 cm. long:
inflorescences cylindric thyrses 1 to 1.5 dm. long; bracts leaf-like but
very small ; heads subsessile, 1.2 cm. long, on short branches; involu-
cral scales about 4-seriate, varying from ovate to linear-oblong, obtuse,
minutely granular, striate: slender corollas 7 to 8 mm. long, the upper
half violet: achenes slender, columnar, light-colored, 4 mm. long,
upwardly hispid; pappus copious, 6 mm. long; styles yellow, only
moderately clavate. — Collected by Dr. Edward Palmer at Sautiago
Papasquiaro, Durango, "April and August," 1896, no. 57. Related to
B. baccharidea, Gray, but readily distinguished by its much longer
achenes, etc. '

Montaxoa arbouescens, Sch. Bip. (Montagnaea arborescens, DC.
Prodr. v. 565). An examination of the type shows that the tips of the
pales are recurved as in most of the other species of this genus and not
inflexed as unfortunately described in the original characterization.
The leaves are oblong-lanceolate, attenuate at each end, rather sharply

488 PROCEEDINGS OF THE AMERICAN ACADEMY.

and regularly serrate near the middle and relatively narrower than in
the nearly related M. frutescens, Ilemsl. I am unable to find M. arbor-
escens represented in any of the many recent Mexican collections.

Calea Pringlei. Stems terete, 6 to 9 dm. high, in dried state striate,
dull brown, covered with a line spreading or slightly refiexed tomentum
of sordid color, and on young parts somewhat glandular; branches op-
posite : leaves broadly ovate to nearly orbicular, sessile, acute, coarsely
dentate (1 to l1 teeth to the centimeter), rugose and scabrous-puberulent
above, paler green, reticulate veiny, and covered with a fine wdiite crisped
pubescence beneath : heads 8 mm. long, 3 mm. in diameter, about 10-
flowered, borne on pedicels 2 to 5 mm. long in opposite-branched com-
pound umbelliform cymes; scales of the cylindrical involucre pale
yellow, striate, very unequal, rounded at the tip, the outer sparingly
pubescent, small, not f'oliaceous nor even herbaceous ; pales erose, nearly
or (pate glabrous : corolla nearly glabrous but covered with transparent
globules, the proper tube slender and somewhat contracted to the summit
where it suddenly expands into the very short throat and deeply cleft
limb of i\\c lineai- segments : achenes slender, somewhat attenuate
toward the base, dark-colored with a violet tinge, appressed-pubescent ;
pappus scales about 1 mm. long, also purple-tinged. — Collected by C. G.
Pringle on mountains above Iguala, Guerrero, Mexico, altitude 1,230 m.,
24 October, 1900, no. 8373. Type in herb. Gray.

Calka Zacatecmiciii, Schlecht., var. calyculata. Involucre sub-
tended at the base by 2 or 3 herbaceous-tipped bracts; these usually
shorter than but sometimes exceeding the inner scarious scales. —
Mexico. Nuevo Leon in the Sierre Madre near Monterey, 0. G. Pringle,
no. 2221, 16 duly, 1888; also in the same locality, C. & E. Seler, no.
1080, 12 October, 1895.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 27. — April, 1901.

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF
HARVARD UNIVERSITY.

New Series. — No. XXI.

SOME NEW SPERMATOPHYTES FROM MEXICO AND

CENTRAL AMERICA.

By M. L. Fernald.

CONTRIBUTIONS FROM THE GRAY HERBARIUM OF HARVARD
UNIVERSITY, NEW SERIES, No. XXI.

SOME NEW SPERMATOPHYTES FROM MEXICO AND

CENTRAL AMERICA.

By M. L. Fernald.

Presented March 13, 1901. Received March 20, 1901.

Fimbristylis melanospora. Glaucous, tufted from a hard base:
leaves firm, smooth, strongly nerved, 1 to 1.5 mm. wide, tending to
become involute, the longest barely 1 cm. long, with a short deltoid
cartilaginous tip : culms compressed, with thin edges, 2 dm. or less high :
umbel decompound, with very short rays, forming a dense inflores-
cence 1 to 2 cm. broad ; involucre of 2 or 3 very unequal leaves, the
longest not equalling the mature umbel ; spikelets ovoid-oblong, 3.5 to
4.5 mm. long, 2 to 2.5 mm. thick : scales ovate, blunt, slightly carinate,
pale brown with broad white scarious margins : style slender, terete,
smooth, with 2 pubescent branches ; achenes lenticular, obovoid, broadly
rounded above, 1 mm. long, brownish black, minutely muriculate. —
Vera Cruz, low grounds, Vera Cruz, April 23, 1894 (C. G. Pringle,
no. 5773). Superficially resembling some forms of F. polymorpka,
Boeckl., but differing in the paler and smoother leaves, the broader white
margins of the scales, the very different blackish achene, and the slender
(not compressed) smooth style.

F. alamosana. Low tufted annual, about 1 dm. high : leaves flat,
smooth, mostly much shorter than the slender flexuous culms, rarely
longer; the sheaths densely ciliate : umbels decompound, with 3 to 7
slender lux unequal rays, the longest 2 or 3 cm. long, the spikelets
mostly long-pedicelled ; involucre of 2 or 3 unequal leaves, the Ion-
equalling or exceeding the rays; spikelets pale straw-colored or whitish,
ovoid-oblong, acutish, 2.5 to 3.5 mm. long, 1 to 1.75 mm. thick: Bcalea
ovate, acutish, or the lowest mucronate, slightly carinate: style com-
pressed, pubescent, with 2 branches; achene barely 1 mm. long, broadly
cuneate-obovoid, with 10 or 12 obscure longitudinal bands, white, pearly
and finely muriculate all over. — Sonora, Alamos, Sept., 181)0 (Edw.

492 PROCEEDINGS OF THE AMERICAN ACADEMY.

Palmer, no. 699). Differing from F. laxa, Vahl, in its paler smaller
spikelets and smaller white and glossy achenes.

F. Holwayana. Culms slender, 1 to 3 dm. high, much overtopping
the pale greeu narrow (0.75 mm. wide) ciliate leaves : umbel simple,
with 2 (rarely 1) to 4 ovoid-oblong spikelets 6 to 8 mm. long, 2.5 to 3.5 mm.
thick, the middle (or rarely solitary) one sessile, the others on slender
ascendiug rays 0.5 to 2 cm. long ; involucre of 2 or 3 ciliate leaves, the
longest one usually much exceeding the rays: scales from orbicular
to ovate, blunt, castaneous and shining, the green midrib in the outer
somewhat prolonged into a short awn, in the others barely into a cusp:
achene short-stipitate, pearly, broadly obovoid, subtruncate above, half
as long as the scale, 1.25 mm. long, nearly as thick, with about 16 lon-
gitudinal ribs, and many linear transverse markings ; style flat, fimbriate,
exserted. — Jalisco, Chapala, Sept. 18, 1899 (E. W. D. Rolway, no.
3-143). Closely related to F. pentastachya, Boeckeler, but differing in
its narrower spikelets and broader scales, and in the .achenes which are
smooth or with one or two tubercles at the very summit, not tuberculate-
roughened throughout as in F. pentastachya.

F. obscura. Glaucous, loosely tufted from slender branching caudex :
leaves flat, 2 or 3 mm. wide, the longest 1.5 dm. long, smooth, or the
margins ciliolate-scabrous except at the acutish cartilaginous tip: culms
compressed, with thin edges, 4 dm. or less high : umbel decompound,
with 2 to 4 rays 2 or 3 cm. long and several shorter ones; involucre of
2 leaves much shorter than the umbel; spikelets mostly pedicelled, or
rarely 2 or 3 fascicled and sessile, lanceolate, acute, 5 to 8 mm. long, 1.5
to 2 ram. thick : scales dark brown, ovate-oblong, strongly carinate and
mucronate: style slender, terete, smooth, with 3 pubescent branches;
achenes trigonous with obtusish angles, obovoid, barely 1 mm. long,
whitish, with extremely obscure longitudinal markings (seen only under
strong magnification). — Durango, alkaline bottoms near Durango,
June, 1896 /'/"'. Palmer, no. 186) : San Luis Potosi, in the moun-
tains. San Miguelito, 1-S76 {SchafTnrr, no. 558). Resembling F.
autumnalis, R. & S., but clearly differing in its caespitose perennial
habit, and its larger, paler, and duller achenes.

Glyphosperma Palme bi, Watson, Proc. Am. Acad, xviii. 161.
Mr. Pringle calls attention to the fact thai thi- plant, which in Mexico
appears like an introduced -pedes, is identical with the Old World
Asphodelus Jistulosiu, L.

Cologania Deamii. Stems slender, 1.5 to 2 dm. long, from aslen-
der creeping rootstock, ascending and branched, the tips flexuous but

FERNALD. — SOME NEW SPERMATOPHYTES. 493

slightly twining, covered with fuscous retrorse-hispid pubescence : stip-
ules ovate to lanceolate, 4 to 6 mm. long: petioles retrorse-hispid, 1.5 to
2.5 cm. loDg ; leaflets oblong-ovate to obovate, mucronate, slightly strigose
above, pale and hispidulous beneath, 2 to 4 cm. long, the terminal ou a
petiolule (0.5 cm.) thrice as long as those of the lateral leaflets : umbels
3-6-flowered, the principal peduncles 4 to 6 cm. long, retrorse-hispid,
those on the flexuous branches shorter ; pedicels 1 to 1.75 cm. long :
calyx 1 to 1.2 cm. long, hispid: corolla violet, 2 to 2.5 cm. long. —
Morelos, Cuernavaca, July 7, 1900 (C. C. Deam, no. 40). Unique in
its long peduncles.

Platanus glabrata. Leaves from broad-cuneate to subcordate,
unequally 5-9-lobed, the middle lobe and the upper pair of lateral ones
lon°'-acuminate, the lower ones short-acuminate ; the margin entire ; at
first closely cinereous-tomentose beneath, soon becoming glabrate : pedun-
cles 0.5 to 1 dm. long, bearing 1 or 2 heads 1.5 to 2 cm. in diameter.

— P. Lindeniana, Wats. Proc. Am. Acad, xviii. 155, not Mart. & Gal.

— Coahuila, Monclova, Aug., 1890 (Edw. Palmer, no. 1269); Saltillo,
April 5, 1837 ((?. S. Sargent); by streams near Diaz, alt. 216 m.,
April 24, 1900 (O. G. Pringle, no. 8319). Differing much from P.
Lindeniana of southern Mexico, which has the less lobed leaves perma-
nently cinereous beneath, and the heads more numerous.

Croton Palmeri, Watson, var. ovalis. Leaves oval, 1.5 to 3 cm.
broad, not lanceolate, rounded at base: clusters many-flowered. — Nuevo
Leon, Monterey, Oct. 11, 1895 (C. & E. Seler, no. 1047).

Alcoceria, no v. gen. of Euphorbiaceae (Hippomaneae). Monoecious.
Staminate flowers in a terminal ament; the individual flowers slightly
imbedded at base by close dense pubescence. Calyx of staminate flower
with 2 broad fleshy valvate sepals and a third narrow inconspicuous one.
Stamen solitary; filament thick, columnar; anther 2-celled, longitudinally
dehiscent, dorsally affixed toward the sheathing tip of the filament.
Calyx of pistillate flower with three deltoid-subulate teeth and few or no
intermediate small glands. Ovary depressed-globose, subtrigonous, the
three 1-ovulate cells alternating with the calyx-teeth. Style cylindric,
erect, equalling the three recurved branches. Fruit a 3-locular capsule.
Seed pisiform, scarcely carunculate. — Dedicated to Gabriel Alcocer,
acute observer, teacher in the Escuela Preparatoria, and botanist at the
Museo Nacional and the Iustituto Medico Nacional of Mexico.

A. Pringlei. Slender shrub, 3 to 5 m. high, the sprawling branches
herbaceous, striate, the lower nodes remote : leaves pilose or glabrate
beneath, reniform, acuminate, varying from 0.4 to 1.3 dm. Ion-, and from

494 PROCEEDINGS OF THE AMERICAN ACADEMY.

entire to more or less palmately 3-7-lobed, palmately 5-nerved, the
nerves conspicuous beneath ; petioles 1.5 to 4 cm. long, pubescent at tip :
inilorescence axillary or terminal ; the staminate ament linear-oblong,
2 cm. long, long-peduncled, with few strongly reflexed pistillate flowers
at base; pedicel of pistillate ilower glandular at base, thick-clavate, 1 to
2 cm. long: capsule about 1 cm. broad. — Guerrero, limestone moun-
tains above Iguala, alt. 1,230 m., Sept. 20, 1900 (C. G. Pringle, no. 8433).

Related to Dalembertia, Baillon, which differs in the single bract
ead of a 3-lobed calyx) at the base of the anther. Also approaching
Tetraplandra, Baillon, and Maprounea, Aublet, but the former is distin-
guished by its four terminal anthers, and the latter by its sbort staminate
Bpike and two more or less connate stamens.

Euphorbia (Anisophyllum) puberula.1 Branching from the

. the brandies subligneous, ascending, 2.5 dm. or less high, puberu-

lent. the tips canescent and tomeutulose : leaves rhomboidal, 1 to 2 cm.

long, 0.5 to 1 cm. broad, very oblique, glaucous beneath, blunt, crenate-

serrate, sparingly pilose; stipules setaceous: cymes dense, 1 or 2 cm.

broad, terminating the leafy branches: involucre white-pilose, turbinate-

campanulate, 1.5 to 2 mm. long, with deltoid hairy lobes; glands 4,

short-stipitate, with or without narrow appendages ; false gland absent

from the broad shallow sinus : capsule appressed-pilose, subacutely lobed,

1 .5 mm. long : seed pulverulent, 1 mm. long, oblong or ovoid-oblong,

quadrangular, with somewbat broken ridges between the angles. —

Mexico without locality (Coulter, no. 1438 in part): Guanajuato, hills

of Guanajuato, 1895 (A. Duges): Morelos, Puente de Ixtla, July 3,

1900 {Charles C. Beam, no. 26); Huerta de la Hacienda de Miacatlan,

Distr. Tetecala, Dec. 28, 1887 (C. & E. Seler, no. 341): Oaxaca, near

Mitla, Distr. Tlacolula, June, 1888; Tecomavaca, Nov. 15, 1895;

I nellin, Distr. Cuicatlan, Nov. 15, 1895 (C. & E. Seler, nos. 31,

1360, 1376); mountains of Oaxaca, alt. 1,750 m., July-Aug., 1900

i<\ Conzatti & V. Gonzalez, no. 1042): Chiapas, Ocozwpiaubtla,

p. Tuxtla. Feb. 19, 1896 (01 & B. Seler, no. 1952). Nearest related

to E. pilulifera, L., and E. lineata, Watson, differing from tbe former in

r terminal cymes and blunter leaves, from the latter in its rather

habit and more rhomboidal Leaves, and from both in its cinereous

puberulence.

1 In tlie preparation of the descriptions of these Euphorbias the writer lias
been greatly assisted by Professor C. I'. Millspaugh, who lias generously examined
mens, and who baa already pointed out (Bot. (Jaz. xxr. 13) the impor-
tam e of the Lnvolucral appendages in differentiating tbe species of this genus.

FERNALD. — SOME NEW SPERM ATOPHYTES. 495

E. (Anisophyllum) potosina. Ascending annual, branching from
the base : stems glabrous, or the younger parts minutely pilose, 2 dm.
high : leaves 1 to 1.75 cm. long, oblong, oblique, rounded or subtruucate
at base, bluntish at tip. glabrous, pale green above, glaucous beneath,
the margin coarsely appressed-crenulate; stipules deltoid-acuminate, laciui-
ate : flowers few in small mostly terminal clusters: involucre pyriform,
glabrous ; the lobes elongate-triangular, entire or nearly so, the 2 Hank-
ing the broad rounded sinus much larger and lacerate ; glands 4, long-
stipitate ; appendages very narrow or wanting ; false gland aristate,
rising from the base of the sinus: capsule glabrous, ellipsoidal, 2.5 to
3 mm. broad, barely 2 mm. high, the three lobes rounded : seed short-
oblong, subquadrate, 1.5 to 1.75 mm. long, the dorsal angle prominent,
the ventral suppressed, blackish, pulverulent, the angles and the margins
of the unequal honeycomb-pits paler. — San Luis Potosi and adjacent
Tamaulipas. San Luis Potosi, without locality, 1876 (./. G. Schaffner,
no. 856, in part) ; region of San Luis Potosi, alt. 1,850 to 2,460 m.,
1878 {Parry & Palmer, no. 814, in part) : Tamaulipas, near Tula,
Sept. 21, 1898 (E. W. D. Holway, no. 3200). — Related to E. hyperici- _
folia, L., E. Preslii, Guss., and E. brasilieiisis, Lam., but differing in
its fewer flowers, larger capsule, and much larger differently marked
seed.

B. (Anisophyllum) interaxillaris. Perennial, with slightly woods-
base : stems prostrate or slightly ascending, 1 dm. or so long, pilose with
more or less spreading hairs : leaves short-petioled, oblong or obovate-
obloug, oblique, rounded at each end, 0.5 to 1 cm. long, 4 to 7 mm. wide,
minutely crenuhite, mostly purple-tinged or blotched, sparingly pilose on
the margins when young, soon glabrate ; stipules narrowly triangular,
fimbriate : flowers solitary, a single one accompanying each pair of leaves,
on slender pedicels 2 or 3 mm. long : involucre campanulate, 1.5 nun.
long, glabrous outside; the lobes triangular, about equal, hairy within ;
glands 3 or 4, brownish, mostly with suborbicular white appendages
1 mm. in diameter ; the 5th gland replaced by an elongated false-lobe
at the base of the shallow triangular sinus : ovary glabrous. — Morel* »S,
Puente de Ixtla, July 3, 1900 (Charles C. Beam, no. 25). Related to
E. stictospora, Engelm., but less pubescent and with solitary flowers.

B. (Zygophyllidium) muscicola. Caudex tuberiform, globose or
oblong, 3 to 5 cm. long, narrowed above to a cylindric neck ; stem very
slender, branching from near the base, the slender flexuous subsimple or
slightly' forked branches 2 or 3 dm. long, glabrous: lower leaves subor-
bicular to ovate, rounded or blunt; the upper elliptic-lanceolate, 0.5 to

490 PROCEEDINGS OP THE AMERICAN ACADEMY.

1.5 cm. long, mucronate, entire or appressed-serrulate, glabrous or
sparingly pilose-setulose ; the lower equalling the capillary petioles,
the upper short-petioled : flowers few, terminal and axillary, on pedicels
2 cm. or less in length: involucre short-campanulate, 2 mm. long, ap-
pressed-setulose, the deltoid-oblong truncate erose lobes incurved ; glands
1. dark, with pale oblong-lanceolate blunt appendages 1 to 1.5 mm. long;
false gland similar : capsule rugulose, 4 mm. high : seed oblong-ovoid,
minutely punctieulate, gray or gray -green, mottled with olive or brown. —
MORELOS, on mossy ledges of Sierra de Tepoxtlan, alt. 2,310 m., Sept.
12, 1900 (C. G. Pringle, no. 8443).

E. (Cyttarospermum) calcicola. Annual, freely branching, 5 or
G dm. high, the slender stem and branches viscid-setulose with long pale
hairs: leaves broad-ovate, rounded or bluntish at tip, truncate or sub-
cuneate at base, the lower 2 cm. long, the upper much smaller, setulose
on both faces, dark above, pale beneath, about equalling the capillary
petioles: flowers axillary on capillary pedicels 1 to 1.5 cm. long: in-
volucre broad-campanulate, green and yellow, 1 mm. long, the short in-
curved cuneate lobes lacerate ; glands 4 with 3 finger-like appendages ;
the false gland slightly larger, with 4 appendages: ovary glabrous. —
Glkrrero, limestone mountains above Iguala, alt. 1,230 m., Oct. 5,
1900 (C. G. Pringle, no. 8398). Closely related to E. astroites, F. &
31.. but the stem conspicuously setulose, and the pedicels much longer.

E. lancifolia, Schl., var. villicaulis. Stems more or less villous
with crisp jointed hairs : leaves rhombic-ovate, acutish, green and lucid
above, pale and more or less villous beneath : involucre campauulate,
pilose or glabrate, 3 or 4 mm. long; lobes truncate, serrate ; glands 1 or
2 associate, hoof-shaped ; false glands 3, columnar, truncate, one-third
longer than the lobes: capsule (young) pilose, 5 mm. long, with rounded
lobes : seed (young) short-oblong, 2 or 2.5 mm. long, pale brown. —
Chiapas, mountain-slope above Ococingo, March 13, 1896 (C. & E.
Seler, no. 2214). Distinguished from the species by its pubescence, in-
volucral lobes, and false glands. In E. lancifolia the involucral lobes are
fimbriate, equalling or slightly exceeding the subulate false glands.

Pernettya ovata. Branches pale brown, minutely puberulent or
glabrate: leaves ovate, short-petioled, 1 to 1.5 cm. long, pale and mi-
nutely canescent-tomentulose toward the rounded base; margins finely
appressed-serrate, the youngest glandular-ciliate, as are the lower faces
of the young leaves : inflorescences axillary, subsessile, all the parts
glabrous, tin- deciduous bracts Borne times with erose margins: flower-
ing calyx 3 mm. long, cleft half-way to the base into deltoid smooth-

FERNALD. — SOME NEW SPERMATOPHYTES. 497

edged lobes: corolla ovate-campanulate, 6 or 7 mm. long. — Chiapas,
mountain woods between San Cristobal Las Casas and Huitztan, March 10,
1896 (O. & E. Seler, no. 212Gb). Growing with P. ciliaris, Don, but
nearer related to P. coriacea, Klotzsch, from which it differs in its pale
and minutely puberulent branchlets, paler leaves and glaudless calyx.

Arctostaphylos Conzattii. Shrub with smooth reddish bark easily
peeling from the branches : leaves oblong-oblanceolate, acute, 2 to 4 cm.
long, 0.75 to 1.5 cm. wide, entire aud subcuneate below to a short petiole,
sharply serrate above the middle, smooth and sublucid above, sparingly
sordid-pilose beneath : racemes terminal, panicled, 1 dm. or less lono- ;
the rachis, pedicels and short ovate-lanceolate bracts puberulent; pedicels
1-3-bracted, 1 cm. long : calyx puberulent, with ovate bluutish lobes :
corolla urceolate, glabrous, 5 or 6 mm. long. — Oaxaca, Sierra de San
Felipe, alt. 3,000 m., April 7 and 8, 1898 (C. Conzatti, no. 691). In its
inflorescence resembling A. nipestris, Robinson & Seatou, but differing
in its much smaller but more coarsely toothed leaves.

A. glabrata. Shrub with smoothish gray-brown bark; the young
branches reddish-brown : leaves narrowly elliptic-oblong, acute or blunt,

2 to 4 cm. long, about 1 cm. wide, tapering to short thick petioles

3 or 4 mm. long, entire or sharply serrate with short mucronate teeth,
bright green above, paler beneath, ferrugineous-tomentulose when
young, becoming glabrate : racemes simple, terminal or axillary, 2 or
3 cm. long; the rachis, the lance-acuminate bracts, and the short (3 to
5 mm. long) pedicels ferrugineous-tomeutose, rarely a little glandular :
calyx slightly tomentulose, the lobes broadly deltoid : corolla urceolate,
glabrate, 4 or 5 mm. long : berries dark, 6 or 7 mm. in diameter. —
Oaxaca, ledges at the summit of Sierra de San Felipe, alt. 3,500 m.,
Sept. 25, 1894 (C. G. Pringle, no. 5712), Aug. 15,1897 (O. Conzatti &
V. Gonzalez in Exsicc. Conzatti, no. 409) ; mountains about Yalalag, alt.
1,850 m., Aug. 1. 1894 {E. W. Nelson, no. 973). Related to A. argnta,
Zucc, but differing in its shorter leaves, and shorter simple essentially
^landless raceme.

Parathesis chiapensis. Branches stout, gray and smoothish, the
younger parts closely tomentose with short stellate ferrugiueous hairs :
leaves elliptic, 1.5 to 2 dm. long, narrowed above to a short slender
acumen and below to a thick cinereo-ferrugineous petiole 0.5 to 1 cm.
lontr; margin closely and finely crenate-dentate ; upper surface smooth
and dull, the lower ferrugineous-tomentose when young, becoming gla-
brate: panicle terminal, leafy below, in anthesis 1.5 dm. high, pyram-
idal, the spreading branches 3 to 6 cm. long, bearing many-flowered
vol. xxxvi. — 32

498 PROCEEDINGS OF THE AMERICAN ACADEMY.

corymbs, rusty toraentose and finely linear-maculate throughout ; bracts
lanceolate, 5 mm. long ; pedicels thickish, somewhat clavate, becoming
1 to 1.5 cm. long: calyx in anthesis 3 or 4 mm. long, deeply cleft into
5 linear lobes : corolla-lobes lanceolate, 6 mm. long, strongly recurved,
tomentulose within : filaments broad at base, 2 mm. long ; anthers oblong-
lanceolate, 3 mm. long. — Chiapas, in mountain-woods between San
Martin and Ococingo, March 13, 1896 (0. & E. Seler, no. 2226). Re-
lated to P. sessilifolia, Donnell Smith, but differing in its crenate dull
leaves, closer shorter pubescence and larger flowers.

Evolvulus Seleriana. Perennial : stems prostrate or slightly
ascending, 1 to 2 dm. long, slightly branched, silky pilose, the older
becoming glabrate : leaves obcordate, obovate or obovate-oblong, short-
petioled, 5 to 12 mm. long, above appressed long-sericeous or glabrate,
beneath permanently sericeous : flowers axillary, the filiform arcuate
sericeous peduncles bractless, 9 to 13 mm. long: calyx very pubescent,
4 mm. long, with lanceolate lobes : corolla pale (apparently white) some-
what silky without, 8 to 9 mm. high. — Chiapas, Tuxtla, Feb. 19, 1896
(C. & E. Seler, no. 1926). Nearest related to E. nummularius, L., but
with different pubescence, obcordate leaves, longer peduncles, and larger
flowers.

Ipomoea caudata. Slender glabrous vine : leaves on slender
petioles 3 to 5 cm. long; blade 1 dm. or less long, narrowly ovate,
prolonged into a caudate-acuminate tip, the base rounded-sagittate, the
basal lobes varying from blunt to acuminate, dark green above, glaucous
beneath, palmately 7-nerved at base: peduncles slender, 1 to 1.5 din.

jt, simple or once forked at the tip; pedicels 2 to 4 cm. long: calyx
narrowly oblong, 1 cm. long, the unequal oblong blunt lobes mottled :
corolla rose-purple, 0.5 dm. long, the cylindric tube 4 cm. long, the short
spreading limb 2 cm. broad: style filiform, glabrous. — Morelos, on
mossy faces of the knobs of the Sierra de Tepoxtlan, alt. 2,310 m.,
Sept. 5, 19d0 (C. G. Pringle, no. 8448). Ilabitally resembling I. simu-
lans, Ilaiibury, but with longer peduncles and calyx, and more slender
corolla.

Cordia (Sebestenoides) Seleriana. Branches straggling, cov-
ered with brown verrucosa bark : leaves ovate or suborbicular, 1 to 2cm.
long, pulverulent and bispidulous; petioles slender, hispid, 3 or 4 mm.
long: corymbs few-flowered, the parts very hispid; peduncles 2.5 cm. or
less in length : caly* oblong-campanulate, 1 cm. long, with short rounded
lobes : corolla '•'< cm. long, Bomewhal hispidulous below, the thin ovate-
oblong round-tipped crenulate lobes 1 to 1.5 cm. long. — Oaxaca, dry

FERNALD. — SOME NEW SPERMATOPHYTES. 499

woods, Huilotepec, Tehuantepec, Jan. 20, 1896 (G. & E. Seler, no.
1779).

Salvia (Micranthae) ageratifolia. Tall branching annual, the
sharply quadrangular herbaceous stems long-setulose, with somewhat
shorter gland-tipped hairs intermixed : leaves cordate, ovate or deltoid-
ovate, 2 to 4 cm. long, coarsely crenate, bluntish, dark green and minutely
pilose or glabrate above, glaucous and puberulent beneath ; petioles
slender, 1 to 3 cm. long, more or less setulose : racemes simple, elon-
gated, becoming 2 dm. long; verticels 2-4-flowered, the lower 1.5 to
2 cm. apart ; bracts persistent, ovate-acuminate, 2 mm. long ; pedicels
becoming 3 mm. long : calyx at first narrowly later broadly campanulate,
in anthesis 5 or 6 mm. long, 2 or 3 mm. broad, in fruit 8 or 9 mm. long,
4 or 5 mm. broad, setulose, the entire mucronate upper lip slightly
shorter than the narrower acuminate lower lobes : corolla 7 or 8 mm.
long, blue, the tube included, the pilose galea shorter than the lip. — ■
Oaxaca, mountains near Oaxaca, alt. 1,750 m., July, Aug., 1900 (C.
Conzatti & V. Gonzalez, no. 1049). Closely related to S. micrantha,
Vahl, & S. setosa, Fernald (nos. 8 & 9 of the writer's Synopsis :).
From the former quickly separated by its loose elongated raceme and
setulose stem ; from the latter by its cordate leaves, setulose stem, and
scarcely secund racemes.

S. tiliaefolia, Vahl, var. rhyacophila. Stem (especially above),
petioles, and calyces densely villous : leaves more pubescent than in the
species. — Morelos, lava-fields below Cuernavaca, alt. 1,230 in., Oct.
17, 1900 (G. G. Pringle, no. 8381).

S. (Angustifoliae) setulosa. Tall (about 1 m. high) and erect
shrub with stiff ascending branches ; stem and especially the young
branches setulose with long white hairs : leaves rhombic-oblong to del-
toid, blunt, cuneate at base into slender petioles, 2 to 4.5 cm. long, 1 to
2.5 cm. broad, dark green and sparingly pilose above, pale and more or
less pilose beneath, the margin finely crenate-serrate: peduncles 1 to 1.5 dm.
lone, sparingly setulose or glabrate: racemes becoming 2 dm. long; the
ovate long-acuminate puberulent bracts 1 to 1.5 cm. long, long-ciliate
on the margins, tardily deciduous: verticels 6-10-flowered, becoming
remote, the lowest 3 or 4 cm. apart ; pedicels 2 to 4 mm. long, canescent
with appressed pubescence: calyx in anthesis 7 mm. long, hirsute espe-
cially on the tube? the ovate-lanceolate long-subulate lobes about equal-
ling the tube, the upper lip usually tridentate : corolla deep blue, 18

1 Proc. Am. Acad., xxxv. 489-550.

500 PROCEEDINGS OF THE AMERICAN ACADEMY.

to 20 mm. long, the scarcely ampliate tube twice exceeding the calyx ;
the pilose lips subequal : style densely bearded. — Morelos, mountains
above Cuernavaca, alt. 2,460 to 2,620 in., Oct. 13, 1900 (C. G. Pringle,
no. 8403), Sept. 1, 1900 ( C. G. Pringle, no. 9204). Closely related to
S. prunclloides, II BK. (no. 40 of the writer's synopsis), but differing
from that and the other members of the group in its tall erect stem.

S. (Vulgares) igualensis. Tall (1 m. or more high), the strongly
grooved quadrangular stem canescent with close pubescence : leaves
broad-ovate, the lower suborbicular, short-acuminate, rounded at base,
5 to 7 cm. long, 4 to 5.5 cm. broad, coarsely crenate-serate, minutely
pubemlent or glabrate on both faces, pale beneath, on slender canescent
petioles 1 cm. or less in length : racemes long and slender, becoming 1.5
to 2 dm. long ; verticels 10-20-flowered, mostly approximate, the lower
1.5 cm. apart; bracts lanceolate, caducous; pedicels becoming 2 mm.
long: calyx in anthesis about 3 mm. long, short-pilose-hispid on the
nerves, with short deltoid teeth; corolla blue, pilose, 1 cm. long; the
slightly ventricose tube twice exceeding the calyx ; galea and lip sub-
equal : style bearded. — Guerrero, limestone mountains above Iguala,
alt. 1,230 m., Sept. 26, 1900 (C. G. Pringle, no. 8418). Related to
S. brachyodonta, Briq. (no. 59 of the writer's Synopsis) and & Ghies-
breghtii, Fernald (no. 60). From the former distinguished by its coarser
habit, cinereous stem, short petioles and appressed teeth of the leaves,
and shorter pedicels ; from the latter by its coarser habit, and broader
glabrate leaves without the characteristic lauate pubescence of S.
Ghiesbreghtii.

S. (Scorodoniae) Dugesii. Shrub, the branches canescent with
closely matted simple hairs: leaves oblong or oblong-ovate, 2 to 4 cm.
long, rounded at base, acute or bluutish at tip, strongly rugose, green
above, canescent beneath with close simple hairs ; petioles slender, 1.5 cm.
or less long: racemes becoming 1 dm. or so long, simple; verticels
8-12-llowered, the lowest 1 to 1.5 cm. apart; pedicels at most 1 nun.
long: calyx cuneate-campanulate, canescent with short mostly simple
appressed liairs, in anthesis 7 mm. long, with short flaring blunt or
merely short-subulate lobes: corolla 1.5 cm. long, pubescent with short
gland-tipped hairs. — Guanajuato, Aug., Sept., 1900 {A. Duges).
This plant, known as "<1n<t cimarona" is nearest related to S. Icuiantha,
Henth. (no. 1<)1 of the writer's Synopsis), but is distinguished by its
narrow leaves, shorter pubescence throughout, larger calyx, and by the
pubescence of the corolla, which in X lasiantha is pilose and essentially
glandless.

FERNALD. — SOME NEW SPERMATOPHYTES. 501

S. leucantha, Cav., forma iobaphes. Calyx and corolla both deep
violet. — Morelos, on ledges of the Sierra de Tepoxtlan, alt. 2,310 m.,
Oct. 14, 1900 (C. G. Pringle, no. 8402).

S. (Macrostachyae) albicans- Shrub 3 to 5 m. high, the branches
ciuereous-puberulent : leaves elliptic-ovate, acuminate, rounded or sub-
cuneate at base, 4 or 5 cm. long, 2 to 2.5 cm. wide, rugulose, minutely
canescent-puberulent, becoming glabrate ; the margin finely appressed-
serrate ; petiole slender, cauescent, 1 cm. or less long : racemes 3 to
10 cm. long, the verticels about 1 0-flowered, mostly crowded, the lower
becoming 1.5 cm. apart; bracts thick and firm, densely cinereous-tomen-
tose, ovate, acuminate, 1 to 1.5 cm. long, semi-persistent, the earliest
finally deciduous : calyx cuneate-campauulate, densely white tomentose,
in anthesis 8 mm. long, the short broad lobes slightly unequal, the longer
upper one acute, the lower blunt: corolla 14 mm. long, canesceut with-
out ; the narrow galea and the tube whitish, the broad lip deep blue :
style densely bearded. — Guerrero, limestone mountains above Iguala,
alt. 1,230 m., Sept. 14, 1900 (C. G. Pringle, no. 8430). Nearly related
to S. Shannoni, Dounell Smith (no. 128 of the writer's Synopsis), but
the smaller thinner leaves, not tomentose beneath, and the inflorescence
more canescent.

S. Sessei, Benth. Mr. Pringle's nos. 7065 and 7080, from Cuerna-
vaca, cited under S. pubescens, Benth., are not that species, but are better
placed with S. Sessei. Mr. Pringle's no. 8378 is also referable to the
latter species.

Monarda Pringlei. Simple, about 1 m. high, the stem obtusely
angled, minutely and sparingly puberulent above: leaves thin, lanceo-
late, 7 to 9 cm. long, 1.5 to 2.5 cm. wide, subtruncate at base, the lower
half rather coarsely appressed-serrate, the elongate-acuminate upper half
entire, minutely puberulent on both faces, slightly paler beneath ; peti-
ole slender, puberulent, 0.5 cm. long: the glomerule terminal, 12-15-
flowered ; bracts like the foliage-leaves, but smaller and entire; bractlets
filiform, often equalling the calyces : calyx 10 or 12 mm. long, cylindric,
slightly curved, puberulent, with subulate spreading teeth 2 or 3 mm.
lono-, and numerous crisp ascending shorter hairs in the throat : corolla
vermilion, 4.5 cm. long, the slender tube long-exserted, tapering gradu-
ally to the ampliate curved throat; lips subequal, 1.5 cm. long, the nar-
row galea sparingly pilose. — Nuevo Leon, moist wooded canon near
Monterey, June 14, 1888 ( C. G. Pringle, no. 2199). A handsome plant
related to M. didyma, L., but differing in its simple habit, minute puber-
ulence, narrow leaves, and fewer flowers.

502 PROCEEDINGS OP THE AMERICAN ACADEMY.

Physalis puberula. Annual?; puberulent throughout with fine
appressed or incurved white hairs : branches slender, terete below, angu-
late above: leaves ovate, subentire or slightly sinuate, 3 to 7.5 cm. long,
1.5 to 4 cm. broad, acuminate at tip, rounded or subcordate at base;
petioles slender, 1 to 4 cm. long: peduncle U.5 to 1 cm. long, straight but
soon recurved: calyx in antbesis 6 to 8 mm. long, the deltoid-lanceolate
acuminate lobes equalling the tube: corolla 1.5 to 2.5 cm. broad, sulphur-
yellow with brown patches at centre: anthers yellowish: fruiting calyx
ovate, 2.5 to 3 cm. long, obscurely 5-angled, slightly sunken at base, the
counivent deltoid-lanceolate teeth 0.7 to 1 cm. long. — Mexico, Sacro
.Monte, Amecameca, Aug. 31, 1900 (C. G. Pringle, no. 9147) ; Guade-
loupe. Aug. 9, 1865-6 (Bourgeau, no. 724) : Dukango, Duraugo, 1896
(JSdw. /'aimer, no. 348). Somewhat resembling the northern P. phila-
delphica, Lam., but with persistent close pubescence and more ovate leaves.
M r. Pringle's plant was accidentally distributed as P. Pringlei (with
incorrect authority), a name belonging to a very different plant with long
glandular pubescence.

Solandm rostratum, Dun., var. subintegrum. Leaves sparingly
pubescent above, shallowly siuuate-lobed, not pinnatifid : otherwise as in
the species. — Mexico, Tacuba, valley of Mexico, Sept. 18, 1900 (C. G.
Pringle, no. 9268).

Ruellia (Physiruellia) cupheoides. Slender shrub, 1 to 1.5 m.
high, the quadrangular verrucose branches covered with gray bark, the
younger parts pilose: leaves elliptic-ovate or ovate-oblong, 3 to 5 cm.
long, short-acuminate, pilose, on slender petioles 0.5-1 cm. long: peduncles
axillary, divergent, slender, pilose and verrucose, 2 to 6 cm. long, termi-
nated by a single flower or branching and bearing 2 flowers; bracts 2,
linear-spatulate, in age becoming 1.5 cm. long: calyx pilose, 1.5 to 2 cm.
long, cleft two-thirds to the base into lanceolate or ovate-lanceolate acu-
minate ciliate lobes: corolla tubular, 3.5 cm. long, yellow or reddish,
with short erect blunt greenish lobes; the slightly narrowed tube shorter
than the calyx : Btamens subequal, included ; filaments glabrous ; anthers
linear-oblong, 1 or 5 nun. long: style bearded; capsule oblong-obovoid,
short-pointed, equalling the calyx, glabrous ; seeds four, thin, ovate, 0.5 cm.
long, dark, with narrow pale margin. — Guerrero, limestone mountains
above [guala, alt. 1,230 in.. Sept. 27, 1900 (C. G. PringU, no. 8409).

Valeriana retrorsa. Slender, 2 to 1 dm. high, from a small oblong
tuber; Btem densely retrorse-hispid below, glabrous above: leaves 3 to
5 cm. long, the basal Bhort-petioled, the upper sessile ; leaflets 8 to 5 pairs,

lineal- lanceolate, blunt, the lowest pair minute, the terminal leaflet much

FERNALD. — SOME NEW SPERMATOCYTES. 503

larger, often 3 cm. long, mostly entire : inflorescence 3 to 6 cm. high, a
trichotomous loose cyme ; flowers rather crowded at the tips ; floral leaves
linear: corolla funnel-form, 6 mm. long, the lobes barely 1 mm. long:
stamens and style slightly exserted : fruit hirsute. — Morelos, mossy
rocks, Sierra de Tepoxtlan, alt. 2,310 m., Sept. 5, 1900 (C. G. Pringle,
no. 8454). Resembling V. mexicana, DC, but with hispid stem, nar-
rower leaves, and longer corolla.

Lobelia regalis. Stem tall and stout, softly white-tomentose, almost
lanate, above : leaves lanceolate to ovate-oblong, 1 to 2 dm. long, 4 to
8 cm. broad, subentire or minutely denticulate with pale hard teeth, pale
green above, at first tomentulose, later glabrate, densely canescent-tomen-
tose beneath; petioles slender, canescent, 2 or 3 cm. long: peduncles
white-tomentose, axillary, ascending, becoming 5 or 6 cm. long : calyx
hemispherical, canescent, 8 to 12 mm. long, cleft half-way to the base
into remote lanceolate lobes : corolla 3.5 to 4.5 cm. long, bright red,
canescent-tomentulose without, the lips subequal : filaments pilose-hispid;
anthers 6 or 7 mm. long, glabrous except at the long-hirsute tip. —
Oaxaca, Cuicatlan, alt. 555 m., Dec. 2, 1895 (L. C. Smith, in exsicc.
C. Conzatti, no. 105); between TIacolula and San Dionisio, Chichicapa,
Jan. 3, 1896, and bed of the river Tehuantepec, below Totolapam,
Yauhtepec, Jan. 4, 1896 (C. & E. Seler, nos. 1673, 1669). Related to
L. laxijlora, HBK., but differing from that in its much larger petioled
leaves and its denser white pubescence.

L. Nelsonii. Shrub with hard ligneous branches covered by smooth
gray bark ; branchlets bearing irregular areas of leaf-scars ; young leaf-
bearing branchlets smooth or hispidulous: leaves firm, ovate-lanceolate,
4 to 7 cm. long, 1 or 2 cm. wide, mostly long-acuminate, narrowed below
to a short petiole, sharply and irregularly serrate, minutely and sparingly
hispidulous : peduncles hispidulous, 1.5 to 4 cm. long, few at the tips of
the branches : calyx glabrous, hemispherical, with linear-lanceolate lobes
3 or 4 mm. long : corolla red, 2.5 to 3 cm. long, glabrous within and with-
out: filaments glabrous ; anthers 7 mm. long, glabrous except at the bearded
tip. — Jalisco, oak woods at 1,380 to 1,694 m. alt., near Huachinango,
March 4, 1897 (E. W. Nelson, no. 4009). A unique species in Mexico.

L. gruina, Cav., var. conferta. Coarser and generally more
branching than the species : racemes densely flowered, not secund, at first
short and capitate, later becoming loose and elongated. — DuBAKGO,
abundant by a creek, Dos Cajetes, alt. 2,620 m., Nov., 1896 (Edw. Palmer,
no. 810) : formerly collected by Gregg in 1848-49, but without record
of locality (no. 392).

504 PROCEEDINGS OF THE AMERICAN ACADEMY.

Heterotoma Goldmanii. Annual or biennial, freely branching from
the base, 5 dm. high; the slender branches minutely hirsute below with
short hairs : leaves mostly toward the base of the plant, thin, lanceolate
to ovate, irregularly cut and lobed half way to the middle, 2 to 4 cm. long,
on slender petioles one half as long as the blade : racemes elongated,
slender and flexuous, pedunculate, loosely-flowered, becoming 2 to 4 dm.
long; pedicels filiform, spreading, at first 1 cm. long, in fruit becoming
5 cm. long ; bracts subulate, the lowest 1 cm. long, the upper much
shorter : calyx hemispherical, glabrous, in fruit 2.5 mm. long, hardly
equalling the lance-subulate teeth, and with a linear spur 5 mm. long,
bearing 2 subulate awns at the tip : corolla blue, 1.5 cm. long, the poste-
rior portion elongated and adnate to the spur of the calyx ; tube cleft
nearly to the base ; the 3 middle lobes of the limb much larger than the
lateral ones. — Sixaloa, road from Las Flechas to La Rostra, Feb. 22,
1899 (E. A. Goldman, no. 324). Nearest related to H. macrocenteron,
Benth. in Hook. Ic. xii. 68, t. 1177, but with smaller flowers, looser
more spreading inflorescence, and different leaves.

H. stenodonta. Perennial from a slender rootstock : stems slender,
glabrous, 3 dm. or less high, leafy to the top : leaves ovate-lanceolate,
Ions-acuminate. 1 dm. or less long, 1.5 to 3.5 cm. broad, thin and slab-
rous, except for the entire cuneate base, sharply and irregularly serrate
with erect narrow-deltoid callous-tipped teeth : racemes terminal or on
axillary peduncles, becoming 6 or 7 cm. long; pedicels filiform, ascend-
ing, 1 cm. or so long, twice exceeding the subulate minutely toothed
bracts: calyx campanulate, the tube in fruit 3 or 3.5 mm. long, about
equalling the lanceolate teeth : corolla pale bluish, 8 mm. long, merely
gibbous at base: capsule oblong, one-half exceeding the calyx-teeth. —
Chiapas, Chicharras, alt. 923 to 1,847 m., Feb. 6, 1896 (E. W. Xelson,
no. 3758).

Brickellia glomerata. Stems closely cinereous with short more or
less viscid pubescence, freely branched above : leaves opposite, petioled,
ovate to oblong, acute or bluntish, the lower excluding the petioles 8 cm.
long, the upper shorter, strongly rugose, scabrous above, tomentulose
beneath : heads sessile or subsessile in the upper axils or in terminal
leafy-bracted glomerules of from 2 to 5 : involucre oblong, 1 to 1.2 cm.
long, of about 4 series of oblong or oblong-lanceolate blunt greenish or
purplish-brown dark-nerved hardly scarious bracts, the outer and the
margins of the inner pilose or arachnoid : achenes linear, 4 mm. long,
strongly appivssed-pubescent. — GUERRERO, oak woods, higher moun-
tains near Acapulco, Jan., 1895 (Edw. Palmer} no. 497): Oaxaca,

FERNALD. — SOME NEW SPERMATOPIIYTES. 505

between Panixtlalmaca and Juquila, alt. 923 m., Feb. 26, 1895 (E. W.
Nelson, no. 2391) ; in oak- and pine-woods, San Carlos, Yauhtepec,
Jan. 5, 189G, and in mountain-woods between San Carlos and San
Bartolo, Jan. 6, 1896 (C. & E. Sder, nos. 1765, 1655); previously
collected at San Bartolo by Liebmann (no. 399), wbose specimen has
been referred by Dr. Klatt to B. hebecarpa, Gray. The present species,
however, differs from that as from the other related species in its glomer-
ulate heads and barely scarious involucre.

Xanthocephalum megalocephalum. Perennial from slender
running rootstock : stem glabrous, glutinous, 3 or 4 dm. high, bearing
3 or 4 large solitary heads at the tips of long branches : leaves entire,
resinous-punctate, the lower spatulate and petioled, about 1 dm. long,
the upper lanceolate, sessile and shorter : involucre broad-hemispherical,
10 or 12 mm. high; the outer bracts lanceolate, witli herbaceous tips,
the inner oblong and blunt, mostly chartaceous : rays oblanceolate, blunt,
2 cm. long : achenes glabrous; pappus a crown of minute pales. — Chi-
huahua, Mt. Mohinora, Sept. 1, 1898 (E. W. Nelson, no. 4890);
Sierra Madre near Guachochi, alt. 2,100 m., Sept. 27, 1898 (E. A.
Goldman, no. 174). Related to X. Alamanl, Benth. & Hook., but
much larger throughout.

Bigelowia Nelsonii. Shrub with rough dark gray bark; brandies
rigid and ascending, conspicuously glandular-punctate: leaves crowded,
linear, submucronate, conspicuously glandular-dotted, 0.5 to 1.5 cm.
long: heads solitary, subsessile at tips of short lateral leafy branches:
involucre 1 cm. broad, many-flowered, the 3 series of glandular bracts
thin, linear-attenuate: achenes long-silky. — CHIHUAHUA, in the Siena
Madre, Sept. 29, 1899 (E. W. Nelson, no. 6494). Resembling B. brachy-
lepis, Gray, but with much broader heads, and thinner attenuate involu-
cral bracts.

Solidago Pringlei. Stem puberulent or minutely pilose, with few
ascending branches, very leafy: leaves linear or elongate-lanceolate,
narrowed below to a slender sub-petiolar base, and above to an acumi-
nate tip, 0.5 to 1.3 dm. long, 3 to 7 mm. wide, the midnerve prominent,
and two lateral ones sometimes apparent, finely appressed-serrulate or
subentire, with ciliate margins, puberulous above, finely pilose beneath :
panicles elongate, narrow, 1 to 1.5 dm. long, 2 or 3 cm. broad, secund,
the tips slightly recurved: involucre 4 mm. high; bracts glabrous,
linear, the outer acutish, the inner blunt: achenes pubescent. —
Nuevo LEON, in the Sierra Madre, near Monterey, Aug. •>■!, 1889
(C. G. Pringle, no. 2886). Resembling S. rupcsiris, Raf., but with

506 PROCEEDINGS OP THE AMERICAN ACADEMY.

more elongated leaves, narrower more secund panicles and much larger

heads.

Contza ltuata, HBK.. var. pilosa. Branches and younger leaves
long-pilose with pale hairs, almost if not quite lacking the viscid charac-
ter and the glands of the species. — Chiapas, on the border of a lake,
Tonahi, Paredon, Feb. 8, 1896 (C. & E. Seler, no. 1879).

Pectis Lessixgii, Fernald, Proc. Am. Acad, xxxiii. 67, a species
formerly known only from Florida and the West Indies, was collected
at Nicoya, Costa Rica, in Decemher, 1899, by A. Tonduz (no. 13791).

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. No. 28. — April, 1901.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY
OF HARVARD COLLEGE.

THE SOLUBILITY OF MANGANOUS SULPHATE.

By Theodore William Richards and Frank Roy Fraprie.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF

HARVARD COLLEGE.

THE SOLUBILITY OF MANGANOUS SULPHATE.
By Theodore William Richards and Frank Roy Fraprie.

Presented by T. W. Richards, March 13, 1901. Received March 21, 1901.

The solubility of manganous sulphate has been carefully determined
within the past year by F. G. Cottrell,* and the present paper is written
to confirm in some measure the results of his work. He was able to dis-
prove by well-considered experiment the less recent work of Linebarger, f
who published a singular series of results resting ujx>n faulty reasoning
and an erroneous method of experimentation. In a case of this kind,
when authorities differ, the truth is more quickly enforced when it is sup-
ported, hence the publication of the present paper.

The work to be recorded below was done in the spring of 1898, long
before the work of Cottrell. It was part of a still unfinished investiga-
tion which has as its object the study of the transition equilibria of potassic
manganous sulphate. We hope that circumstances may permit the early
publication of the remainder of this work, which has been suspended for
a time.t

It was soon found that a much higher temperature was necessary to
drive off the water of crystallization of manganous sulphate than has
usually been supposed. Linebarger used only 180°, at which tempera-
ture at least one molecule of water remains in the salt if it is surrounded
by air with the usual proportion of aqueous vapor. The handbooks name
210° to 240° as the temperature required, and Cottrell, calling attention
to Linebarger's error, used 280°. We found that traces of water still
remained after heating for half an hour at 350° in an air bath. In this
respect the substance reminds one of cupric sulphate. § The amount thus
retained does not much exceed the tenth of a per cent, and the effect upon

* Cottrell, J. phys. ch., 4, 637 (1900).
t Linebarger. Am. Chem. Journal, 15, 225 (1893).

J A brief statement of the scope and object of this work will be found in 1'roc.
Am. Ass. Adv. Sc., 213 (1898).

§ Richards, Proc. Am. Acad , 26, 240 (1891).

510 PROCEEDINGS OF THE AMERICAN ACADEMY.

Cottrell's results could not have been serious. On the other hand, a
temperature just below redness, perhaps 450°, obtained with a carefully
watched naked flame, and applied after long drying at lower temperature,
was capable of driving off in five minutes so much of the water that sub-
sequent similar heating for a quarter of an hour showed only an average
loss of 0.2 milligram. The product was wholly soluble in water, showing
that no decomposition of the sulphate itself had occurred.

The manganous sulphate was purified with great care, and the crystals
employed were coarsely powdered. The specimen whose solubility was
to be determined was put into a large stout test-tube with a carefully
cleaned rubber stopper, and was kept for at least four hours at the
desired temperature in an Ostwald thermostat before a sample was
taken. The agitation of the mixture was active and continual, being
effected by an apparatus similar in principle to that of Schroder.* The
motive power was a Ileurici hot-air motor. At the close of the ap-
pointed time the sample to be analyzed was removed by an effectual
filtering pipette, somewhat similar to one which has since been used
in van't Hoff's laboratory. f A diagram of the pipette is appended.

B

I

m.

^^m

^zrr

Filtering Pipette.

Tlic filtering attachment, filled with cotton wool ("absorbent cotton," C), Is tem-
porarily attached to the jet A of a 5-c.c. pipette by means of the rubber tube B.

The cotton was very necessary to filter off the fine powder which ap-
peared during the stirring. In order to avoid change of temperature,
the pipette was previously wanned by placing it in a dry test-tube im-
mersed in the thermostat. Only the mouth of the test-tube was allowed
to project above the water of the thermostat, and of course every pre-
caution was taken to obtain a fair sample of the solution.

After filling the pipette and quickly removiug the rubber filtering- jet
attached to it. the clear solution was run into a weighing-bottle and was
quickly stoppered and weighed. The known amount of solution was
washed into a roomy platinum crucible, when it was cautiously evaporated.

ihroder, Zeitschr. phys. chem., 11, 464 (1898); Noyea, ibid., 9, GOG (1892).
P,.r details see Richards and Faber, Am. Chem. Journal, 21, 168 | L899.)
t Van't Hoffand Meyerhoffer, /.. phys. ch., 27, 79 1 1898).

RICHARDS AND FRAPRIE. MANGANOUS SULPHATE.

511

and the residue was ignited in the fashion already described. In order to
make certain that no transition had occurred during the experiment, the
crystal-water contained in the solid phase left over after the saturation
was always determined. Below are the data thus obtained.

The Solubility of MnS04.5H20 at 25°.

Xo. of
Determination.

Time of
Saturation.

Weight of
Solution.

Weight of
MnS04.

MnS04 for
100 gr. Water.

1

hours.
6

grams.
5.0342

grams.

1.9849

grams.
65.09

2

6

6.3405

2.4947

64.89

3

8

4.4318

1.7470

65.07

4

8

5.2465

2.0656

64.94

Averag

e

. 64.99

5

48

3.4514

1.3622

65.20

G

72

3.5349

1.3947

65.17

Averag

'e . . . .' .

. 65.19

Cottrell found 64.78 as a mean of two closely agreeing determinations,
but the time allowed for saturation was only 2.2 hours. The probable
reason for his slightly lower result will be discussed below.

The solubility of the tetrahydrate is recorded on the following page.

The solid material taken from the tubes in Determinations 7 to 10
contained as much as 4.3 molecules of water for each MnS04, but it was
nevertheless probably the designated hydrate containing included mother
liquor. The greater solubility of the pentahydrate at 30° would certainly
involve the solution of any accessible pentahydrate. The salt remaining
from determinations 11 and 12 contained 4.03 molecules of water.

The results at 30.15°, giving an average of G6.38, correspond almost
exactly with Cottrell's figure 66.43 at 30° ; but in this case his time of
saturation was increased to an average of three hours, while ours was not
much lono-er. Hence close correspondence was to have been expected,
and the figures mutually support one another. The slight difference may
be due to a residual trace of water in Cottrell's salt.

512

PROCEEDINGS OF THE AMERICAN ACADEMY.

On the other hand, at 35° Cottrell used only two hours for saturation,
while we used ahout seven times as much time. Hence the difference
between his result, G7.87, and ours, 68.22, may be explained once more
by a difference in the time of mixing.

Until recently the cause of this difference would have been ascribed to
a possible incomplete saturation in Cottrell's case ; and the higher figures,
other things being equal, would have been accepted as the more accurate.

The

SOLDBILITY

of MnS04.4K20 at

30.15° and 35.0°.

No. of
Determination

Temperature.

Time of
Saturation.

Weight of
Solution.

Weight of
MnSO«.

MnSO. for

100 gr. Water.

hours.

grams.

grams.

7

30.15°

4

4.3125

1.7215

GG.4 4

8

30.15°

4

5.7507

2.2943

(',(',.•27

9

30.15°

4

4.5992

1.8339

00.32

10

30.15°

4

5.2322

2 0804

G6.48

Average

11

35.0°

13

3.7009

1.5007

68.21

12

35.0°

14.5

2.7854

1.1296

08.22

Average

08 22

While this may be true, the recent work of Ostwald* on the solubility of
powders and the surface tension of solids has thrown new light on this
matter, and it seems quite possible that neither series of results may be
perfectly definite. Fine powders have a greater solution-tension than
coarse ones, for the same reason that small drops have a greater vapor
tension than large ones.

It has often been Btated that very long agitation is necessary to secure
saturation. t En view of Ostwald's newer work it seems quite possible
that continued active agitation introduces an uncertainty even greater
than the one which it avoids. The crystals of salt in the ever-moving
tube act as mutual millstones, and gradually wear off one another's

* Ostwald, X. phys. oh., 34, 106 (1000).

t For example, see Ostwald's Physicochem. Measurements, Walker (Afacmillan,
p. 176.

RICHARDS AND FRAPRIE. — MANGANOUS SULPHATE. 513

corners, with the production of fine powder. This fine powder must
continually dissolve, because it is more soluble than the larger aggre-
gations. At first it will simply hasten the speed of attaining satu-
ration ; but later, when saturation with respect to the larger particles
has been attained, the fine powder will tend to produce a solution super-
saturated with respect to those larger particles. The experience of
Cottrell and others seems to indicate that supersaturation is harder to
obviate than inadequate saturation. Cottrell, for example, found that
MnS04.H20 attained a concentration of 57.41 grams in 100 grams of
water after two hours of agitation when the salt was added to pure
water, while after twice as long a time the supersaturation in another
tube had only been reduced from 72.5 to 62.3 grams per hundred. The
solubility found after 27 hours, when both methods gave the same result,
was 58.32, or 0.91 grams more than the first and 4. grams less than the
second figure.

The outcome of the matter seems to be that constant results upon
solubility are usually obtained only when the rate of production of the
fine powder exactly balances the rate at which the supersaturation is
relieved, and that constant results thus reached represent only a com-
promise. With cautious agitation, it is possible that a solubility very
near that of flat crystal surfaces might be obtained ; on the other hand,
very active agitation, such as we used, must tend to increase this solu-
bility almost to that corresponding to the fine powder which we always
observed in our tubes.

Everyone will agree with Ostwald in deciding that the solution tension
of flat surfaces, rather than that of sharply curved surfaces, is the quan-
tity which should be determined, if possible. This result would be best
obtained by keeping the solid as free as possible from agitation, and
driving a constant current of the saturating solution over these resting
crystals. A dissolving apparatus which has recently been described, if
assisted by a small turbine and suitably immersed in a thermostat, would
perhaps be the safest apparatus, although complete saturation would re-
quire much time.*

It is evident, as Ostwald points out, that most published determina-
tions of solubility, those of Cottrell and our own included, are subject
to a small uncertainty, but it is also evident that the work of Cottrell
is by far the most complete work upon the solubility of manganous sul-
phate which has been published, and that it wholly overthrows the
erroneous results of Linebarger.

* Richards, Am. Chem. Journal, 20, 189 (1898).
vol. xxxvi. — 33

514 PROCEEDINGS OF THE AMERICAN ACADEMY.

We determined also the solubility of both the dihydrate and the tetra-
hydrate of the double Balphate of manganese and potassium, and found
solutions saturated at 25.0 with both or either hydrate to yield 40.1 grams
of anhydrous solid. Hence this temperature must be near the transition
temperature of the two hydrates. The detailed publication of these
results and many other similar observations must be reserved for a later
communication.

In a few words of recapitulation, the present paper may be said to
conlirm the work of Cottrell and to disprove that of Linebarger, while
a measure of doubt is cast upon the usual methods of determining
solubility.

< vMBKinGE, February, l'JOl.

Proceedings of the American Academy of Arts and Sciences.
Vol. XXXVI. Xo. 29. — June, 1901.

RECORDS OF MEETINGS, 1900-1901.

A TABLE OF ATOMIC WEIGHTS. Br Theodore William
Richards.

REPORT OF THE COUNCIL: BIOGRAPHICAL NOTICES.
Charles Carroll Everett. By Ephraim Emerton.
Nathaniel Holmes. By Jeremiah Smith.
Silas Whitcomb Holman. By Charles R. Cross.
Sylvester R. Koehler. By Charles G. Loring.
John Elbridge Hudson. By James B. Thayer.
John Harrison Blake. By Clarence John Blake.
Charles Franklin Dunbar. By Frank William Taussig.

OFFICERS AND COMMITTEES FOR 1900-1901.

LIST OF THE FELLOWS AND FOREIGN HONORARY
MEMBERS.

STATUTES AND STANDING VOTES.

RUMFORD PREMIUM.

INDEX.

(Title-Page and Table of Contents.)

RECORDS OF MEETINGS.

Nine hundred and seventeenth Meeting.

May 9, 1900. — Annual Meeting.

The Academy met at the Jefferson Physical Laboratory,
Cambridge.

Vice-President Trowbridge in the chair.

A quorum not having been present at the adjourned stated
meeting of April 11, 1900, on the motion of the Recording
Secretary, it was

Voted, That the votes then proposed be now confirmed and
approved.

The Corresponding Secretary read letters from Arlo Bates
and L. H. Bailey, accepting Fellowship in the Academy. He
also read a letter from Seabury C. Mastick, Secretary of the
Committee on the Modification of the Federal Legacy Tax,
enclosing a new petition intended as a substitute for the one
sent out in January and requesting signatures. On the motion
of the Recording Secretary, it was

Voted, That the President be authorized to sign this peti-
tion on behalf of the Academy.

The Chair announced the death of Frederick Edwin Church,
Associate Fellow in Class III., Section 4.

The Annual Report of the Council was read by the Cor-
responding Secretary.*

The Treasurer presented his annual report, of which the fol-
lowing is an abstract: —

* See Proceedings, XXXV. 623.

518 PROCEEDINGS OF THE AMERICAN ACADEMY

General Fund.

Receipts.
Balance from last year $32.61

Investments $5,6-49.80

Assessments 915.00

Admission fees 40.00

Sale of publications 71.23

Miscellaneous 6.30 6.682.33

$6,714.94
Expenditures.
General expenses $2,499.56

Publishing expenses 2,012.73

Library expenses $1,161.77

Expenses of moving 596.78

Furniture 184.60 1,943.15

Balance 259.50

$6,714.94
Rumford Fund.

Receipts.

Balance from last year $3,915.10

Investments $2,566.19

Sale of publications 10.00

Refund of appropriation 100.00

Interest on bank account 71.82 2,748.01

$6,663.11
Expenditures.

Investigations $1,215 00

Medals 327.no

Library 81.28

Miscellaneous 10.00 $1,633.28

Income invested during the year and transferred

to capital account 3,253.90

Balance 1,775.93

$6,663.11
Warren Fund.

Receipts.
Balance from last year $1,149.43

Investments $365.00

Iuterest on bank account 50.27 415.27

$1,56-1.70

OP ARTS AND SCIENCES. 519

Expenditures.

Investigations $500.00

Accrued interest on bonds purchased . . . 70.13 $570.13

Balance ~ ] ~ 994.57

$1,564.70
Building Fund.

Receipts.

Balance from last year $30G.03

Investments $240.00

Income 21.55 261.55

$567.58
Expenditures.

Accrued interest on bonds purchased $28.06

Balance 539.52

$567.58
The following reports were also presented : —

RErORT OF THE RuMFORD COMMITTEE.

At the last Annual Meeting of the Academy, the sum of $1000 was
placed at the disposition of the Rum ford Committee for the furtherance
of researches in light and heat.

The Committee has made the following appropriations from this fund :

October 11, 1899, it was voted that an appropriation of $500 be made
to Professor Edwin B. Frost of the Yerkes Observatory, to assist in the
construction of a spectrograph especially designed for the measurement of
stellar velocities in the line of sight.

January 10, 1900, it was voted that an appropriation of five hundred
dollars be made to Professor Edward C. Pickering of Harvard College
Observatory, for the purpose of carrying out an investigation on the
Brightness of Faint Stars by cooperation with certain observatories
possessing large telescopes.

Furthermore, on January 10, 1900, the Committee voted to ask the
Academy to appropriate one hundred dollars from the income of the
Rumford Fund to Professor Theodore W. Richards of Harvard Univer-
sity, for a research on the Transition Point of Crystallized Salts, which
request was favorably acted upon by the Academy.

And on April 9, 1900, the Committee voted to ask the Academy to
appropriate from the same source the sum of two hundred and fifty

520 PROCEEDINGS OF THE AMERICAN ACADEMY

dollars to Mr. Arthur L. Clark of the Worcester Academy in further-
ance of a research on the Molecular Properties of Vapors in the Neigh-
borhood of the Critical Point. This recommendation has still to be
acted upon.

The attention of the Committee had several times been called to the
fact that for a long period of years there had been no way, except by a
formal vote of the Academy, in which a person other than a member
could purchase a copy of the Life and Works of Count Rumford. Mean-
while occasional applications from reputable persons or libraries had
been received. The Committee therefore voted on April 9, 1900, to
recommend to the Academy that the Treasurer be authorized to make
arrangements for their sale to the public, intending that the whole
matter of the manner of sale, price, and other details should be left to
hi- discretion. Such authorization was duly granted by the Academy.

The Committee was furthermore advised that there is no copy of the
Life and Works of Count Rumford in the Library of the Academy. It
was therefore voted at the meeting of April 9, 1900, to recommend to
the Academy that a set be added to the library.

At the same meeting it was voted, that the Committee recommend to
the Academy the appropriation of one hundred and fifty dollars from the
income of the Rumford Fund for the purchase and binding of the usual
periodicals for the current fiscal year, together with the following:
Deutsche Zeitschrift fiir Elektrotechnik, L'Eelairage Electrique, Fort-
schritte der Llektrotechnik.

It was also voted by the Committee that* the Chairman recommend
that the nsual appropriation of one thousand dollars be made by the
Academy for the immediate needs of the Committee in furtherance of
research.

I have therefore to present the three last-mentioned recommendations
for the consideration of the Academy.

The Committee has considered at much length the subject of an award
of the Rumford Premium. Among the various candidates whose claims
were discussed, there was one who without dissent was considered in
the highest degree worthy of the honor. The numerous researches of
Professor Carl Barus in various branches of heat, carried on for many
years, have uniformly been characterized 1>\ great skill, accuracy, and
originality, and these valuable contributions to science are generally
recognized as of far-reaching importance.

It was unanimously voted by the Committee, for the first time on
January 1". and for the Becond time on April it, 1900, to recommend to

OP ARTS AND SCIENCES. 521

the Academy at the Annual Meeting the award of the Rumford Medal
to Professor Carl Barus, of Brown University, for his various researches
in heat.

In response to the usual request, the Chairman of the Committee has
received reports of progress from the following persons to whom grants
have been made from the Rumford Fund : Professors H. Crew,
E. B. Frost, G. E. Hale, E. H. Hall, F. A. Laws, E. L. Nichols,
E. C. Pickering, B. O. Peirce, T. W. Richards, and W. C. Sabine.

These reports in full or in abstract are presented as an appendix to

this Report.

Chas. R. Cross,

Chairman of the Rumford Committee.

APPENDIX.

Reports of Progress to the Rumford Committee.

May, 1900.

PROFESSOR, HENRY CREW.

Grants of October 26, 1896, $400, for researches on the electrical,
chemical, and thermal effects of the electric arc, and May 3, 1899, $200,
for researches on the spectrum of the electric arc.

Referring to an "air-tight arc" which he has constructed, Professor
Crew says, " This apparatus is well made, and works in a thoroughly
satisfactory manner. With it I have made several series of photographs
of the arc spectra of iron and magnesium, in an atmosphere of pure
hydrogen. These plates I am now measuring up, and hope to be able
to publish the results during the coming summer.

" Contrary to expectations, I do not find the iron spectrum at all
simplified by the introduction of hydrogen. For while many lines are
blotted out, many new ones are introduced by the hydrogen. No less
than sixty-two iron lines, either new or profoundly modified, make their
appearance between A 2753 and A 4415. The origin of these lines I
have not yet succeeded in finding.

" I have spent some time also in trying to produce, in this arc, the
hydrogen series discovered by Professor Pickering in the star Z Puppis ;
but I have not found anything resembling such a series."

PROFESSOR EDWIN B. FROST.

Grant of January 10, 1900, $500, to assist in the construction of a
spectrograph, especially designed for the measurement of stellar velocities
in the line of sight.

522 PROCEEDINGS OF THE AMERICAN ACADEMY

"In reply to your inquiry 'as to the present status of researches
now in progress which have received aid from the Rumford Fund,' I
would state that three remarkably fine prisms have already been pur-
chased, and the other optical parts have been ordered, for the new
spectrograph designed for the determination of stellar velocities in the
line of sight, in aid of which research a grant of $500 was made last
autumn.

" It is hoped that the spectrograph may be completed and the regu-
lar work with the instrument may be begun during the present summer,
and I shall therefore expect to be able to announce some results of the
investigations for your next annual meeting."

PROFESSOR GEORGE E. HALE.

Grant of November 10, 1897, $400, for the purpose of completing
a spi-rtruheliograph.

"The spectroheliograph was first attached to the 40-inch telescope
last December. Preliminary tests in the laboratory had shown it to be
extremely satisfactory from an optical point of view, at least in so far
as the definition and contrast of the photographs of spectra were con-
cerned. As we were without a skilled instrument maker at the obser-
vatory for several months last summer and autumn, I had the moving
plate holder and second slit constructed by an instrument maker in
Chicago. The first tests with the telescope showed the presence of a
mechanical defect, which caused the plates to be striped with lines a
millimeter apart, the pitch of the screw which drives the plate-carriage.
A -eries of experiments showed that these lines were due in large part,
if doI altogether, to the absence of suitable end-thrust bearings for the
W, and to the poor construction of the screw, nut. plate-carriage, etc.
It was then decided to completely reconstruct this part of the Bpectro-
heliograph in our own shop. End-thrust bearings were supplied, together
with a new screw, which was carefully ground. As the result of the
work the millimeter lines did not appear on the photographs taken
after the change had been made. It was found, however, that the
plate was -till covered with line lines about one quarter of a milli-
meter apart. The-.' are shown on the plate which I will send you in
a few day-. In order to remove these lines a large number of experi-
ments were required, bul at last, plates practically i'vi<- from lines, and

r than those obtained with the very satisfactory hydraulic
apparatus of the Kenwood spectroheliograph, were Becured. The diffi-

OF ARTS AND SCIENCES. 523

culty was due in part to errors in the gears, and in part to the theo-
retically correct double Hooke's joint, which has now been replaced with
a belt and pulleys.

"As may easily be imagined, the curvature of the spectral lines in
this spectroheliograph is very marked. If the first slit is straight the
second slit must be curved, to fit the K line, and the resulting image
of the sun is seriously distorted. By dividing the curvature equally
between the two slits, and moving the plate in a direction opposite to
that employed with a straight first slit, the distortion can be wholly
removed. You will notice that the images on the plate sent you are
round.

" As soon as the eclipse work is over I hope to return to the spectro-
heliograph. The preliminary results are such as to lead me to expect
much from the instrument."

PROFESSOR EDWIN H. HALL.

Grant of April 26, 1895, $250, in aid of his investigation on the
thermal conductivity of metals.

"I expect to read by title a paper at the next meeting of the Acad-
emy. I am using some of the money on a side research carried on by
Mr. McElfresh, a graduate student, on the thermo-electric effect of hy-
drogen occluded in nickel. Said effect seems thus far to be lacking."

PROFESSOR FRANK A. LAWS.

Grant of December 13, 1893, in aid of an investigation on the thermal
conductivity of metals.

"In respect to the experiments on Thermal Conductivity I can say
that this constant for the cast-iron specimen has been determined, and
the result calculated. I have not published the method, for the cali-
bration is not satisfactory to me on account of a very peculiar set of
readings which I do not feel justified in rejecting. During the exam-
ination period I hope to straighten the matter out."

PROFESSOR EDWARD L. NICHOLS.

Grant of October 19. 1894, $250, in aid of investigations on the radia-
tions from carbon at different temperatures.

" I bee to report that my investigation of the visible radiation from
carbon, made in part by the aid received from the Rumford Fund, is
nearing completion. Spectrophotometric comparisons of the lighl from

524 PROCEEDINGS OP THE AMERICAN ACADEMY

the two varieties of carbon, black and gray, using as a standard tbe
acetylene flame, have been made up to temperatures above 1600°. I
am at present struggling with the more diilicult range between 1G00°
and the melting point of platinum. The determination of the higher
temperatures offers considerable difficulty as yet, but there are indi-
cations which make it seem worth while to carry the investigation to
as high a point as possible. The research involves three subordinate
investigations, one of which, on the temperatures of the acetylene flame
itself, is completed. Tliis I have made the subject of a paper read be-
fore the Physical Society. It is likewise printed in the April number
of the Physical Review. In my report to the American Academy, I
shall incorporate the results of these measurements in connection with
the main subject. Of the other two subordinate researches, the study
of the thermo elements employed is nearly completed, and a determina-
tion of the distribution of energy in the spectrum of the acetylene flame
by means of the Nichols radiometer is in progress. The latter is being
carried on by one of my assistants, Mr. G. W. Stewart, and we hope
as a result of his measurements to obtain a study of the absolute values,
by means of which spectrophotometry measurements for different wave
lengths can be brought into known relations to each other.

"The work for which the grant was made from the Rumford Fund
has required a much longer time than I anticipated, but in spite of the
numerous difficulties which we have had to overcome, I think the end
is now in sight."

PROFESSOR EDWARD C. PICKERING.

Grant of January 10, 1900, $500 for the purpose of carrying out an
investigation on the Brightness of Faint Stars, by cooperation with cer-
tain observatories possessing large telescopes.

" An appropriation of five hundred dollars ($500) has been made
from the Rumford Fund, to be expended under the direction of Pro-
fessor Pickering, for the purpose of carrying out an investigation on the
brightness of faint stars by cooperation with certain observatories poss. "■
ing large telescopes. This appropriation results from a communication
made to the Council of the American Astronomical and Astro-physical
Society held in New York last January. It was represented that the
mo-t urgent need of astronomy in America was adequate endowment
of the great telescopes of the country so that thej could be kept actively
at work. It was shown that while the two largest telescopes of the
country, and of the world, were Kept constantly al work, the me.ins lor

OF ARTS AND SCIENCES. 525

the reduction and publication of the observations are wholly inadequate ;
while some of the largest telescopes in the country, representing a plant
costing hundreds of thousands of dollars, are nearly idle and therefore
useless. Observations of the greatest value can be obtained with these
instruments at small expense, and it is hoped that the beginning now
made will justify its permanent continuance on a large scale. The
problem undertaken is the determination of the light of faint stars, se-
lected as standards. These will furnish points of reference to which
other photometric measures may be referred. Five photometers have
been constructed in which, by interposing a photographic wedge of shade
glass, an artificial star is reduced in brightness until it appears equal to
a real star, as seen in a large telescope. Thirty-six regions have been
selected in different parts of the sky, in each of which a series of stand-
ards is to be measured. Five stars of about the twelfth magnitude,
five of the fifteenth, five of the sixteenth, and five of the seventeenth
are to be chosen in each of these regions. The faintest stars will be
selected and measured with the Yerkes 40-inch and Lick 36-inch tele-
scopes. Those of the sixteenth magnitude will be measured with the
26-inch telescope of the University of Virginia and perhaps the Prince-
ton 23-inch telescope. The stars of the fifteenth magnitude will be
measured with the 15-inch Harvard telescope. All of these stars will
be compared with the stars of the twelfth magnitude, when absolute
magnitudes will be determined with the 12-inch Harvard meridian pho-
tometer. Their relative brightness will also be determined more accu-
rately with the Harvard 15-inch telescope. After the work is fairly
started it is believed that it can be reduced to a simple routine, by which
great results may be attained with a moderate expenditure. By the
time this report is presented, it is expected that observations with the
Yerkes, Lick, University of Virginia, and Harvard telescopes will be
in progress."

PROFESSOR B. O. PEIRCE.

Grant of December 2, 1892, $200, and April 26, 1895, in aid of an
investigation on the propagation of heat in certain solid bodies.

Professor Peirce writes that he " published last summer in the Pro-
ceedings of the Academy a short paper on the 'Thermal Conductivity of
Vulcanite,' giving therein the results of a long series of experiments.
This paper in a somewhat extended form appeared in the Philosophical

Magazine."

A paper on the specific heats of different marbles is also ready for

publication.

'>-''< PROCEEDINGS OP TnE AMERICAN ACADEMY

PROFESSOR THEODORE W. RICHARDS.

Grants of October 12, 1898, $200, for the construction of a microkine-
toscope, the immediate application of which is to be a study of the birth
and growth of crystals; and January 10, 1900. $100, for a research on
the transition point of crystallized salts.

•• I have ha<l the honor of receiving two grants from the Rumford
I and, the accounts of which have not been closed.

"The first was given for the study of crystal growth by instantaneous
photography. Upon this research I have already made several reports,
[laving pushed the photographic study of the growth of crystals in
a«pieous solutions to the furthest limit which seemed possible without a
very much greater expenditure of money, we have turned our attention
to the study of the change of the crystalline structure of iron and steel at
a red heat. We have been able so to modify our apparatus, with very
slight extra expenditure, that such study seems to be possible.

"The second grant of money was for prosecuting a research on new
fixed temperatures for thermometric standardization. Of this grant only
$27.50 have thus far been spent for materials which have not yet been
exhausted. We have succeeded in showing that the transition tempera-
ture <>f sodic chromate is \i'ry near 19.88 . but the exact point cannot
be fixed until our thermometers have arrived from the Bureau Interna-
tionale. These new thermometers are to be the property of the College.
hence all the remainder of the Rumford grant ^-slUO — $27.50 = $72.50)
will be available for the special purposes of this particular investigation."

PROFESSOR WALLACE C. SABINE.

Grants of January 12, 1898, $400, and .March L5, L899, $200, for

ircln's on ultra-violet radiation.

Professor Sabine states that Mr. Lyman, who is engaged upon the

investigation, "will publish a paper on a by-product of the investigation

which Beems to m<- very interesting and important. In this paper he

proposes to how that among the spectra formed by the Rowland con-

gratings there are spectra not accounted for by the ordinary

theory of tic grating; that Mich Bpectra are c Minn in, and at times fairly

strong and of excellent definition ; that these spectra are diffraction spec-

.! much 1 688 dispersion than the ordinary recognized spectra, and that

the errors of ruling to which they are due are not local but general to

the whole surface of the grating He will also explain an experimental

method of BOrting out these lines from tic regular and calculable spectra.

OF ARTS AND SCIENCES. 527

These false spectra are especially dangerous in series sjjectra work, giv-
ing a somewhat systematic reproduction of strong lines and groups, which
in actual vibration frequencies do not exist. There is some evidence
that such errors have been committed in the past, and it was in the pres-
ence of such errors that the false spectra were here discovered."

Report of the C. M. Warren Committee.

In behalf of the C. M. "Warren Committee, I have to report that Pro-
fessor Mabery, to whom a grant of $500 was voted by the Academy at
the last Annual Meeting, has vigorously prosecuted his researches on the
Composition of Petroleums. During the year he has published also
several papers describing the results of his earlier researches which were
aided by previous grants from the Warren Fund.

Professor H. O. Hofman, to whom grants, amounting in all to $230,
were made several years ago, has in this year published the results of
his work in a memoir entitled ''The Temperatures at which Certain
Ferrous aud Calcic Silicates are formed in Fusion, and the Effect upon
these Temperatures of the Presence of Certain Metallic Oxides."

The C. M. Warren Committee recommends that the sum of six hundred
dollars ($600) from the income of the Cyrus M. Warren Fund be granted
to Professor Charles F. Mabery, of Cleveland, Ohio, for the continuation
of his j'esearches on the Chemistry of Petroleums.

F. H. Storer, Chain/tan.
9 May, 1900.

Report of the Committee of Publication.

The Publishing Committee reports that during the past academic year
there have been issued three numbers of Vol. XXXI Y. and twenty-two
numbers of Vol. XXXV. of the Proceedings, aggregating 579 pages, with
no plates. One number only has been printed at the charge of the Rum-
ford Fund. The expenditure from the General Fund was $2, 01 2.93,
out of an available amount from appropriation and sales of $2,471.23,
leaving an unexpended balance of 0458.30. The Committee desires for
the coming year the same appropriation as for the last, viz., §2,400.

For the Committee,

Samuel II. Scuddeb, Chairman,

528 PROCEEDINGS OF THE AMERICAN ACADEMY

Report of the Committee on the Library.

The most important feature of tlie year was the removal of the Library
from the Boston Atheuaeum building to the third story of the new build-
ing of the Massachusetts Historical Society. Although it was expected
that these quarters would be ready iii April, 1899, it was six months later
before they could be occupied. The books, etc., were moved in good
order between October 17 and November 7, 1899, thanks to the careful
supervision of Dr. Holden, the Assistant Librarian. The opportunity
was taken to send to the binder many volumes which had accumulated
from previous years. The storage-room in the basement of the new
building could not be occupied until April, 1900, and during the first
week of that month the publications of the Academy, which had been
stored in the basement of the Athenaeum, were brought there.

The accessions during the year have been as follows : —

B\ <nft and exchange

By purchase — Gen'l Fund

By purchase — Rumf. Fund

Total 456 2435 328 5 3224

The total number compares with 3284 during the previous year.

25 volumes and 593 parts of volumes were bought with the appropria-
tion from the income of the General Fund at an expense of $191.35.
155 parts of volumes were bought with an appropriation from the income
of the Rumford Fund for $24.28.

1 19 volumes were bound at an expense of $501.55, of which $ 1*7.00
wa> charged to the General Fund, and 817.05 to the Rumford Fund.

202 books have been borrowed by twenty-live persons, including
eighteen Fellows of the Academy. 155 of these volumes were borrowed
between May and October, 1899, and only 47 between November, 1899,
and May. 1900. 22 volume- were not returned before May 2nd, in
accordance with the rule. It will be noticed that the use of the Library
has much diminished during the past, six months. During the preceding
year 235 volumes were borrowed.

Of the appropriation from the General Fund of $1500, $1161.77 has
been -pent. This includes $179.7 1 for incidental expenses of the Assist-
am Librarian, but does not include about $200 for subscriptions to peri-
odicals not yet paid and about $100 for books now in binders' hands. It

Vols.

Parts of vols.

Piims.

Maps.

Tnt'll.

431

L687

328

5

2451

25

593

6 1 8

155

155

OF ARTS AND SCIENCES. 520

is therefore hoped that the uuexpended balance of the $1500, viz. $338.23,
may be reappropriated for next year, making a total of $1838.23.

The appropriation from the Rumford Fund was $120, of which only
$41.33 has been spent, $17.05 for binding and $24.28 for periodicals,
but subscriptions to periodicals aggregating about $65, and some binding
must still be paid for. Therefore it is desired that the balance of the
appropriation, viz. $78.67, be reappropriated for next year in addition
to the $150 recommended by the Rumford Committee. The volumes
needed to complete the file of the " Fortschritte der Pkysik," which it
was voted to purchase, have been ordered.

The cost of moving the Library and belongings, which it was voted to
pay from the funds of the Academy, was $601.78.

Of the special appropriation of $200 for furnishing the Library,
$184.60 has been spent. This includes $30 for the purchase of a type-
writer which has been useful to the Recording Secretary as well as to the
Assistant Librarian.

In conclusion, the need of a new catalogue of the Library is urged.
At present there is only an author-catalogue written on inconveniently
large cards, and the writing is often nearly illegible. So valuable a library
demands a modern catalogue of authors and subjects, either typewritten or
printed on standard cards. The cost of such a catalogue is not known,
but a special appropriation of $200 is asked for, to commence the work.

A. Lawrence Rotch,
Librarian, and Chairman Committee
on the Library.
Boston, Mat 9, 1900.

In accordance with the recommendations contained in the
above reports, it was

Voted, To award the Rumford Medal to Carl Barus, for his
various researches in heat.

Voted, That a set of the Life and Works of Count Rumford
be added to the Library of the Academy.

Voted, To appropriate one hundred and fifty dollars (8150)
from the income of the Rumford Fund for the purchase and
binding of the usual periodicals for the current fiscal year,
together with the following : Deutsche Zeitschrift fur Elek-
trotechnik, L'Eclairage Electrique, and Fortschritte der Plivsik.

Voted, That the sura of one thousand dollars (81000) from

vol. xxxvi. — 34

f,:;il PROCEEDINGS OF THE AMERICAN ACADEMY

the income of the Rumford Fund be placed at the disposal
of the Rumford Committee to be expended in aid of investiga-
tions on Light and heat, payments to be made on the order of
the Chairman of the Committee.

Voted, That an appropriation of six hundred dollars ($G00)
from the income of the Warren Fund be -ranted to Charles F.
Mabery of Cleveland, Ohio, for the continuation of his re-
searches on the chemistry of petroleum.

Voted, To appropriate twenty-four hundred dollars ($2400)
from the income of the General Fund for the expenses of
publication.

Voted, To appropriate eighteen hundred and thirty-eight
and ,v., dollars ($1838.23) from the income of the General
Fund and two hundred and twenty-eight and ffo dollars
($228.67) from the income of the Rumford Fund, in addition
to the amount recommended by the Rumford Committee, for
library expenses.

Voted, to appropriate two hundred dollars ($200) toward
making a new catalogue of the Library.

( )ii the recommendation of the Committee of Finance, it was

Voted, To appropriate two thousand dollars ($2000) from
lli,. income of the General Fund for general expenses.

Voted, That the assessment for the ensuing year be five
dollars.

On the motion of S. II. Scudder, it was

Voted, To meet on adjournment on the second Wednesday
in June.

The annual election resulted in the choice of the following
officers and committees : —

Alexander Agassiz, President.

John Tuowijiiidgk, Vice-President for ('hiss I.

Alpheus Hyatt, Vice-President for Class U.

Augustus Lowell, Vice-Pn \identfor Cl>>ss III.

William M. Davis, Corresponding Secretary.

William Watson, Recording Secretary.

Francis Blake, Treasurer.

A. Lawrence Rotch, Librarian.

OF ARTS AND SCIENCES. 531

Councillors.
Henry Taber, \

Theodore W. Richards, ( of Class I.
Harry M. Goodwin, )

William T. Councilman, \

John E. Wolff, ( of Class II.

George H. Parker, )

James B. Ames, ^

Willtam Everett, >- of Class III.

A. Lawrence Lowell, J

Member of the Committee of Finance.
Augustus Lowell.

Rumford Committee.

Erasmus D. Leavitt, Amos E. Dolbear,
Edward C. Pickering, Arthur G. Webster,
Charles R. Cross, Theodore W. Richards,

Thomas C. Mendenhall.

C. M. Warren Committee.

Charles L. Jackson, Leonard P. Kinnicutt,
Samuel Cabot, Arthur M. Comey,

Henry B. Hill, Robert H. Richards,

Henry P. Talbot.

The following gentlemen were elected members of the
Academy : —

Jay Backus Woodworth, of Cambridge, as Resident Fellow
in Class II., Section 1 (Geology, Mineralogy, and Physics of
the Globe.)

Merritt Lyndon Fernald, of Cambridge, as Resident Fellow
in Class II., Section 2 (Botany).

William Ernest Castle, of Cambridge, as Resident Fellow
in Class II., Section 3 (Zoology and Physiology).

George Mercer Dawson, of Ottawa, as Associate Fellow in
Class II., Section 1 (Geology, Mineralogy, and Physics of the
Globe), in place of the late Sir John William Dawson.

532 PROCEEDINGS OF THE AMERICAN ACADEMY

Melville Weston Fuller, of Washington, as Associate Fellow
in Class III., Section 1 (Philosophy and Jurisprudence).

Rufus Bvam Richardson, of Athens, as Associate Fellow in
Class III., Section 2 (Philology and Archaeology).

Thomas Day Seymour, of New Haven, as Associate Fellow
in Class III., Section 2.

Henry Morse Stephens, of Ithaca, as Associate Fellow in
Class III., Section 3 (Political Economy and History).

William Cawthorne Unwin, of London, as Foreign Honorary
member in Class I., Section 4 (Technology and Engineering).

Sir Archibald Geikie, of London, as Foreign Honorary Mem-
ber in Class II., Section 1 (Geology, Mineralogy, and Physics
of the Globe), in place of the late Carl Friedrich Rammelsberg.

Sir John Murray, of Edinburgh, as Foreign Honorary Mem-
ber in Class II., Section 1, in place of the late Alfred Louis
Olivier Legrand Des Cloizeaux.

Arthur G. Webster called attention to the bill before the
House of Representatives for the establishment of a National
Standardizing Bureau, and on his motion, it was

Voted, That the Academy approves this project.

On the motion of W. M. Davis, it was

Voted, That the committee appointed at the meeting of
April 11, 1000, to consider the propriety of amending the first
chapter of the Statutes be instructed to correct certain clerical
errors in the Statutes.

Clarence J. Blake made some remarks on the scientific
researches of his father, the late John H. Blake, which he
intended to describe more fully in a forthcoming biographical
notice.

John Trowbridge described some results obtained with a
storage battery of twenty thousand cells and exhibited the bat-
tery in operation.

peri men te on the passage of powerful discharges through minute
orifices were described, and proofs of the oscillatory nature of sparks bis

long wen- given. Since these sparks closely represent the main
features of lightning, it, is probable thai most lightning discharges are
also oscillatory. The battery with the aid of large condensers furnishes

OP ARTS AND SCIENCES. 533

powerful quantity discharges which are more interesting from a scien-
tific point of view than discharges of high electromotive force ; for a new
field appears to be opened in spectrum analysis. Photographs of the
spectra of gases can be obtained with one or two discharges with a nar-
row slit.

Perhaps the most interesting results obtained with the battery were
from the methods of exciting the X-rays. Photographs of the usual
subjects treated by these rays can be taken, and they exhibit great con-
trasts ; moreover, there are traces of ligaments and muscles as well as
bones. The members of the Academy visited the battery room and saw
a hydrogen tube excited by the discharge from three hundred Leydeu
jars, and the lighting of an X-ray tube.

The following papers were presented by title: —

Paleontological Notes V. A New Fossil Crab from the Mio-
cene Greensand Bed of Gay Head, Martha's Vineyard, with
Remarks on the Phylogeny of the Genus Cancer. By Alpheus
S. Packard.

On the Thermal Diffusivities of Different Kinds of Marble.
By B. O. Peirce and R. W. Willson.

Paleontological Notes VI. On Supposed Merostomatons and
Other Paleozoic Arthropod Trails, with Notes on those of
Limulus. By Alpheus S. Packard.

On the Continuity of Groups generated by Infinitesimal
Transformations. By S. E. Slocum. Presented by Henry
Taber.

On the Thermal and the Electrical Conductivity of Soft
Iron. By Edwin H. Hall.

An Apparatus for Recording Alternating Current Waves.
By Frank A. Laws.

The Dinitro Compounds of Paradibrombenzol. By C. Loring
Jackson and D. F. Calhane.

On Certain Derivatives of Orthobenzoquinone. By C. Lor-
ing Jackson and Waldemar Koch.

Geometry on Ruled Quartic Surfaces. By F. B. Williams.
Presented by W. E. Story.

On the Action of Sodic Sulphite on Tribromdinitrobenzol and
Tribromtrinitrobenzol. By C. Loring Jackson and Richard B.
Earle.

534 PROCEEDINGS OF THE AMERICAN ACADEMY

Nine hundred and eighteenth meeting.

June 13, 1900. — Adjourned Annual Meeting.,
A quorum was not present and the Academy was not called
to order.

Nine hundred and eighteenth Meeting.

October 10, 1900. — Stated Meeting.

The Corresponding Secretary in the chair.
The following letter was read : —

843 Exchange Btilding, Boston,
September 5, 1900.

Francis Blake, Esq.,

Treasurer of the American Academy of Arts and Sciences.

Dear Sir, — By a letter enclosed in his will, my Father, Mr.
Augustus Lowell, requested his sons to pay " to the American Academy
of Arts and Sciences $10,000." It gives me great pleasure to enclose
to you, in accordance with this request, a check for the amount named.

Very truly yours,

A. Lawrence Lowell.

The Corresponding Secretary also read letters from W. E.
Castle, M. L. Fernald, Jeremiah Smith, and J. B. Woodworth,
accepting Fellowship ; from George M. Dawson, M. W. Fuller,
T. D. Seymour, and H. M. Stephens, accepting Associate Fel-
lowship; and from Sir Archibald Geikie, F. Kohlrausch, Sir
John Murray, and W. C. Unwin, acknowledging election as
Foreign Honorary Members.

Announcement was received of the deatli of D. T. Day,
formerly President of the Buffalo Society of Natural Sciences.

The following deaths were announced : —

Augustus Lowell, Vice-President for Class III. ; John Elbridge
Hudson, of Class III., Section 1., Sylvester R. Koehler, of
Class III., Section 4, Resident Fellows ; James Edward Keeler,
of Class I., Section 1, Jacob Maudes DaCosta, and Alfred Stille,
of Class II., Section 4, William Mitchell, of Class III., Section
1, Associate Fellows; Willy Kiilme, of Class II., Section 4,
Charles Russell, Baron Russell of Killowen, and Henry Sidg-
Wick. (.!' ('lass III., Section 1, Foreign Honorary Members.

OF ARTS AND SCIENCES. 535

The Chair appointed the following standing committees : —

Committee of Publication.

Samuel H. Scudder, Seth C. Chandler,

Crawford H. Toy.

Committee on the Library.

A. Lawrence Rotch, Henry W. Haynes,

Samuel Henshaw.

Auditing Committee.
Henry G. Denny, William L. Richardson.

The Chair appointed from the next retiring Councillors

Theodore W. Richards, of Class I.,
William T. Councilman, of Class II.,
James B. Ames, of Class III.,

a committee to nominate candidates for the offices made vacant
by the death of Augustus Lowell.

Thomas Messenger Drown, of South Bethlehem, was elected
an Associate Fellow in Class I., Section 3 (Chemistry).

J. H. Wright gave an account of " Recent Excavations in
Crete," by Miss Helen S. Boyd, of the American School at
Athens.

Barrett Wendell read an essay on the " Literary History of
America."

W. M. Davis presented some " Geographical Notes on Brit-
tany and Devonshire."

The following paper was presented by title by W. C. Sabine : —

False spectra from the Rowland Concave Grating. . By Theo-
dore Lyman.

Nine hundred and nineteenth Meeting.

November 14, 1900.

The Corresponding Secretary in the chair.

The Corresponding Secretary read a letter from the Ameri-
can Chemical Society transmitting resolutions relative to the
establishment of a National Standards Bureau in connection

536 PROCEEDINGS OF THE AMERICAN ACADEMY

with the United States Office of Standard Weights and Meas-
ures and requesting cooperation in its efforts to secure the
establishment of such bureau.

On the motion of Arthur (>. Webster, it was

Voted, That this matter be referred to a committee to be
appointed by the Chair.

The committee was constituted as follows: —

James M. Crafts, Chairman,
Arthur G. Wtebster,
Theodore W. Richards,
Edwin H. Hall.

On the motion of Barrett Wendell, the following resolution
was adopted : —

Whereas, The Academy has received the sum of ten thou-
sand dollars paid by the sons of the late Augustus Lowell at
the request of their father, therefore be it

Resolved, That the Academy accepts this gift with grateful
appreciation of the generosity of its late honored Vice-Presi-
dent, and that the Corresponding Secretary notify this action
to Air. Lowell's sons.

The following papers were presented by title : —

On the Composition of California Petroleum. By Charles
F. Mabery and Edward J. Hudson.

On the Chlorine Derivatives of the Hydrocarbons in Cali-
fornia Petroleum. By Charles F. Mabery and Otto J. Sieplein.

( >n the Composition of Japanese Petroleum. By Charles F.
Mabery and Shinichi Takano.

James Ford Rhodes read an account of "Sherman's March
to the Sea," of which the following is an abstract: —

After tli^ capture <>f Atlanta, the <|uestion in Sherman's mind was how
I,.- should still further proceed on the offensive. Hood gave him trouble
by severing his communications, hut lie could not he brought to a battle,
nor could he he caught in m pursuit. Sherman resolved to leave Ten-
nessee in the care of Thomas and march through Georgia to the sea.
II,. severed his communications with the North, November 12, 1864,
and from that day to December 14 no direct intelligence from him

OF ARTS AND SCIENCES. 537

reached the North. There was little fighting, but the supply of 62,000
troops in the enemy's country called for foresight and system on the
part of the general ; to prevent the host from degenerating into a law-
less mob required the enforcement of discipline by the general and his
officers. The army foraged liberally on the country in an orderly man-
ner. While there were some abuses, some wanton destruction of prop-
erty, and some pillage, there were no cases of murder or rape. On the
whole the army behaved as well as could have been expected. Sher-
man estimated the damage done to the State of Georgia at §100,000,000.
On the night of December 20, 1864, the Confederates evacuated Savan-
nah, Sherman took possession of the city, and sent his celebrated Christ-
mas-gift despatch to President Lincoln.

H. Helm Chiton read a paper entitled : " The Eclipse Cy-
clone and the Diurnal Cyclone : Results of Meteorological
Observations during the Solar Eclipse of May 28, 1900."

Nine hundred and twentieth, Meeting.

December 12, 1900.

The Academy met at the house of William W. Jacques.

The President in the chair.

The death of Thomas Gaffield, Resident Fellow in Class I.,
Section 3, was announced.

On the motion of the Corresponding Secretary, it was

Voted, That a committee be appointed to make revision of
certain passages in the Statutes and report thereon to the
Academy.

This committee was constituted as follows : —

The President,

The Corresponding Secretary,

C Loring Jackson.

Wallace C. Sabine spoke on "The Influence of Architecture
on Melody and the Development of the Musical Scale."

John E. Wolff described the celebration of the two-hundredth
anniversary of the foundation of the Royal Academy of Sciences,
at Berlin, on the 19th and 20th of March last, which John W.
White and himself attended as delegates from the American
Academy.

538 PROCEEDINGS OF THE AMERICAN ACADEMY

"William Everett gave an account of the life and works of the
late Henry Sidgwick, referring particularly to his eminent
ability and estimable character.

The following paper was read by title : —

"Symmetrical Triiodbenzol." By C. Loring Jackson and G.
E. Behr.

Nine hundred and twenty-first Meeting.

January 9, 1901. — Stated Meeting.

Vice-President Hyatt in the chair.

On the motion of C. L. Jackson, it was

Voted, To defer action on the proposed amendments of the
Statutes.

Voted, That the Academy send a message of congratulation
to Vladimir Markovnikoff, of Moscow, on the occasion of the
40th anniversary of his work on chemistry.

The vacancies occasioned by the death of Augustus Lowell
were tilled by the election of

James B. Thayer, Vice-President for Class III.

Eliot C. Clarke, Member of the Committee of Finance.

Denman W. Ross read a paper entitled, " Design as a Science."

S. C. Chandler gave an account of his "New Discovery con-
cerning the Motion of the Earth's Pole."

The following papers were presented by title : —

" Suggestion concerning the Nomenclature of Heat Capacity."
By T. W Richards.

" A Study of Crowing Crystals by Instantaneous Photomicro-
graphy." By T. W. Richards and E. H. Archibald.

Nine hundred and twenty-second Meeting.

February 13, 1901.

The Corresponding Secretary in the chair.

The Chair announced the death of Charles Hermite, Foreign
Honorary Member in Class I., Section 1.

An invitation to the, Ninth Jubilee Celebration of the Uni-
versity of Glasgow was read. On the motion of the Recording
Secretary, it v.

OF ARTS AND SCIENCES. 539

Voted, That the Academy send delegates to this celebration.

A circular inviting attendance at the Fifth International
Congress of Physiologists was read ; also, a letter from James
B. Thayer, acknowledging his election as Vice-President for
Class III.

The following paper was read by title : —

" A Study of Variation in the Fiddler Crab (Gelasimus pugi-
lator Latr.)" By Robert M.Yerkes. A Contribution from the
Zoological Laboratory of the Museum of Comparative Zoology
at Harvard College. Presented by E. L. Mark.

Nine hundred and twenty-third Meeting.

March 13, 1901. — Stated Meeting.

Vice-President Hyatt in the chair.

The following deaths were announced : —

Charles Carroll Everett, Resident Fellow in Class III., Sec-
tion 1.

George Mercer Dawson, Associate Fellow in Class II., Sec-
tion 1.

The following gentlemen wrere elected members of the
Academy: —

Alexander Wilmer Duff, of Worcester, as Resident Fellow in
Class I., Section 2 (Physics).

Theodore Lyman, of Brookline, as Resident Fellow in Class
I., Section 2.

Lewis Jerome Johnson, of Cambridge, as Resident Fellow in
Class I., Section 4 (Technology and Engineering).

Henry Lloyd Smyth, of Cambridge, as Resident Fellow in
Class I., Section 4.

Frank Shipley Collins, of Maiden, as Resident Fellow in Class
II., Section 2 (Botany).

Ephraim Emerton, of Cambridge, as Resident Fellow in Class
III., Section 3 (Political Economy and History).

Frank William Taussig, of Cambridge, as Resident Fellow in
Class III., Section 3.

Eliakim Hastings Moore, of Chicago, as Associate Fellow in
Class I., Section 1 (Mathematics and Astronomy).

540 PROCEEDINGS OF THE A.MEBICAN ACADEMY

George Ellery Hale, of Williams Bay, as Associate Fellow in
Class I., Section "2 ( Physics).

Edward Leamington Nichols, of Ithaca, as Associate Fellow
in Class I., Section '2, in place of the late William Augustus
Rogers.

Cyrus Guernsey Pringle, of Charlotte, Vermont, as Associate
Fellow in Class II., Section 2 (Botany), in place of the late
George Clinton Swallow.

Franklin Paine Mall, of Baltimore, as Associate Fellow in
Class II., Section 3 (Zoology and Physiology), in place of the
late Alfred Stille*.

Henry Fairfield Osborn, of New York, as Associate Fellow in
Class IP, Section 3, in place of the late Othniel Charles .Marsh.

Charles Otis Whitman, of Chicago, as Associate Fellow in
Class IP, Section 3.

William Stewart Halsted, of Baltimore, as Associate Fellow
in Class IP, Section 4 (Medicine and Surgery), in place of the
late William Alexander Hammond.

William Williams Keen, of Philadelphia, as Associate Fellow
in Class IP, Section 4, in place of the late Jacob Maudes Da-
Costa.

Jules Henri Poincare, of Paris, as Foreign Honorary Member
in Class P, Section 1 (Mathematics and Astronomy), in place
of the late Francesco Brioschi.

Heinrich Muller-Breslau, of Berlin, as Foreign Honorary
Member in Class P, Section 4 (Technology and Engineering).

Hugo Kronecker, of Bern, as Foreign Honorary Member in
Class IP. Section 3 (Zoology and Physiology), in place of the
late Willy Kiihne.

Sir Thomas Pander Brunton, of London, as Foreign Honorary
Member in Class IP, Section 4 (Medicine and Surgery), in place

the late Sir James Paget, Bart.

Roberl Koch,of Berlin,a - Foreign Honorary Member in Class
IP. Section P in place of the laic Louis Pasteur.

Albeit Venn Dicey, of Oxford, as Foreign Honorary Member
in ('lass IIP. Section 1 (Philosophy and Jurisprudence .

William Edward Hearn, of Melbourne, as Foreign Honorary

OP ARTS AND SCIENCES. 541

Member in Class III., Section 1, in place of the late Charles
Russell, Baron Russell of Killowen.

Henry Jackson, of Cambridge, as Foreign Honorary Member
in Class III., Section 2 (Philology and Archaeology), in place
of the late Henry Sidgwick.

Edmondo de Amicis, of Florence, as Foreign Honorary Mem-
ber in Class III., Section 4 (Literature and the Fine Arts), in
place of the late John Ruskin.

On the motion of the Corresponding Secretary, it was

Voted, To meet, on adjournment, on the 10th of April next.

B. L. Robinson gave an account of " Recent Advances in the
General Classification of the Flowering Plants."

R. T. Jackson spoke on " Resorption as a Factor in Growth."

The following papers were presented by title : —

The Occlusion of Magnesic Oxalate by Calcic Oxalate, and
the Solubility of Calcic Oxalate. By Theodore W. Richards,
Charles F. McCaffrey, and Harold Bisbee.

Contributions from the Cryptogamic Laboratory of Harvard
University, XLVI. Preliminary Diagnoses of New Species of
Laboulbeniaceae, III. By Roland Thaxter.

The Development and Function of Reissner's Fibre and its
Cellular Connections. By Porter E. Sargent. Presented by E.
L. Mark.

The Solubility of Manganous Sulphate. By Theodore W.
Richards and F. R. Fraprie.

Contributions from the Gray Herbarium of Harvard Uni-
versity. New Series, No. XX. I. Synopsis of the genus
Melampodium. II. S}rnopsis of the genus Nocca. III. New
Species and Newly Noted Synonymy among the Spermato-
phytes of Mexico and Central America. By B. L. Robinson.

Contributions from the Gray Herbarium of Harvard Univer-
sity. New Series, No. XXI. By M. L. Fernald. Some New
Spermatophytes from Mexico and Central America.

542 PROCEEDINGS OF THE AMERICAN ACADEMY

Nine hundred and twenty-fourth Meeting.

April 10, 1901. — Adjourned Stated Meeting.

Vice-Pbesident Trowbridge in the chair.

The Corresponding Secretary read letters from L. J. Johnson,
Theodore Lyman, II. L. Smyth, and F. W. Taussig, accepting
Fellowship ; and from W. S. Halsted, W. W. Keen, Franklin
P. Mall, E. II. Moore, Edw. L. Nichols, and Henry F. Osborn,
acknowledging election as Associate Fellows.

The chair announced the following deaths: —

Edward Elbridge Salisbury, of New Haven, Associate Fellow
in Class III., Section 2.

Jacob Georg Agardh, of Lund, Foreign Honorary Member in
Class II., Section 2.

On the recommendation of the Rumford Committee, presented
by A. G. Webster, it was

Vbt L To appropriate from the income of the Rumford Fund
the sum of five hundred dollars ($500) to T. W. Richards in
aid of a research on the Joule-Thomson experiment on free
expansion.

Edwin H. Hall presented "An Exposition of the Theory of
Electrons, the Electrical Fragments of Atoms."

The following paper was presented by title: —

The Law of Physico-chemical Change. By Gilbert Newton
Lewis. Presented by T. W. Richards.

Dr. W. G. Farlow was appointed Delegate of the Academy to

celebration of the ninth jubilee of the foundation of the

University of Glasgow and was charged with the presentation of

the address on the following page, signed by the Vice-President

of Class HI. and the Corresponding Secretary.

OP ARTS AND SCIENCES. 548

Viro Illustrissimo Cancellario

Curiae Senatcique Universitatis Glasguensis

Academia Artium et Scientiarum Americana

Quae est Bostoniae Nov.-Anglorum.

S. P. D.

Vix dici potest quauto guadio audiverimus vos in mente habere istam
diem celebrare qua Universitas vestra annum quadringeutesimum exple-
verit, ex quo Pontifex ille benevolus nomen ac jura Universitatis con-
cesserit ; quis enim ignorat longam illam aunoruna seriem orbi terraruru
perpetuis illuxisse virorum operibus, iu omni arte atque in omni scientia
praestantissimorum, nomen semper illustre Caledoniae vel ad majus fasti-
gium efferentibus ?

Quormu ut neque tempus neque charta sufficeret nominibus tantum
percurrendis, nostro officio satisfiat si gratias praecipuas ageutes quod ad
sollemnia celebranda amicissime invitastis socium nostrum viruni ornatissi-
mura Gulielmum Gilson Fallow legaverimus qui istis caeremoniis iuter-
futurus nostram erga vos benevolentiam pro virili parte libentissime
praestabit.

Ilium ergo, viri illustrissimi, prccamur ut benigne excipiatis ut qui
amicitiae vincula inter Scotiam Americamque arctiora dignissime ligave-
rit.

Dabamus Bostoinae Nov.-Anglorum A. D. Non : Maias anno sal-
utis MCMI, atque Rerumpublicarum Foederatarum libertatis vindicatae

cxxv.

544

PROCEEDINGS OF THE AMERICAN ACADEMY

A TABLE OF ATOMIC WEIGHTS
of Seventy-seven Elements.
Compiled in April, 1901, from the most Recent Data.
By Theodore William Richards.

Name.

Symbol.

Atomic
Wright.

Name.

Symbol.

A tomic
Weight.

Aluminium . .

Al

27.1

Molybdenum . .

Mo

96.0

Antimony .

SI,

120.0

Neodymium . .

Nil

143.6

Argon . .

A

39.92

! Neon ....

Ne

19.94

Arsenic . .

As

75.0

Nickel ....

Ni

58.70

Barium . .

Ba

137.43

Niobium . . .

Nb = Cb

94.

Beryllium .

Be = Gl

9.1

Nitrogen . . .

N

14.04

Bismuth

Bi

208.

< ionium . . .

Os '

190.8

Boron . .

B

11.0

Oxygen (standard)

O

Hi.000

Bromine

Br

79.955

Palladium . . .

Pd

106.5

( !admium .

Cd

112.3

1'liospliorus . .

P

31.0

-ium

Cs

L32.9

Platinum . . .

Pt

195.2

Calcium

Ca

40.1

Potassium . . .

K

39.14

< larbon . .

C

12.001

Praseodymium .

Pr

140.5

Cerium . .

Ce

1 in.

Rhodium . . .

Rh

L03.0

Chlorine

CI

35.4V,

Rubidium . . .

Rb

85.44

Chromium

Cr

52.14

Ruthenium . .

Ru

101.7

Cobalt . .

Co

59.00

Samarium ? . .

Sm

150.

Columbium

Cb = Xb

94.

Scandium . . .

Sc

44.

i topper . .

Cu

63.604

Selenium . . .

Se

79.2

" Didymium '

Nd + Pr

142±

Silicon ....

Si

28.4

Erbium .

Er

166.

Silver ....

Ag

107.93

Fluorine

F

19.05

Sodium. . . .

Na

23.05

i ladolinium

Gd

156. ?

Strontium . . .

Sr

87.68

Gallium

Ga

70.0

Sulphur . . .

s

32.065

< iermanium

Ge

72.5

Tantalum . . .

Ta

183.

< rlucinum .

<it=Be

9.1

Tellurium . . .

Te

127.5?

Gold. . .

Au

197.3

Terbium ' . . .

Tb

160.

Helium . .

He

3.96

Thallium . . .

Tl

204.15

Hydrogen .

II

1.0075

Thorium . . .

Th

233.

Indium . .

In

114.

Thulium?. . .

Tu

171. ?

Iodille . . .

I

126.86

Tin

Sn

119.0

Iridium . .

Ir

193.0

Titanium . . .

Ti

18.17

Iron . . .

Fe

55.9

Tungsten . . .

W

184.

Krypton

Kr

81.7

Uranium . . .

U

288.5

Lanthanum

La

188.5

Vanadium .

V

61.4

Lead . .

l'b

206.92

\". noil ....

X

1 28.

Lithium

Li

7.03

Ytterbium . . .

VI,

173.

Magnesium

Mg

24.36

Yttrium . . .

Yt

89.0

Manganese

Mil

55.02

Zinc

Zn

66 10

Mercury .

Hg

•JOIUI

Zirconium . . .

Zr

90.6

OP ARTS AND SCIENCES. 5^5

NOTE CONCERNING THE TABLE OF ATOMIC WEIGHTS.

Three new elements are included in the table, namely, Neon, Krypton,
and Xenon, discovered by Ramsay in the atmosphere. Their molecular
weights cannot be considered as wholly settled, but approximate values are
given in the list. There is a possibility that each of these values, as well as
those given for Argon and Helium, should be halved in order to obtain the
atomic weights ; but for the present it seems safest to assume a monatomic
molecule in each case.

Radium, the singular substance discovered by Madame Curie, is not
included on account of the great uncertainty concerning its atomic weight.
It is possible that some of the work up<*i thorium and uranium is vitiate 1
by radium, and that the presence of this impurity may account for the
inconsistencies of the data. The value given for uranium depends primarily
on recent work by Richards and Merigold, as yet unpublished; but the
matter cannot be considered as settled.

Attention is called to the low atomic weight of iron,* which was included
without comment last year. Investigation upon this important constant
is still in progress.

* Richards and Baxter, These Proceedings, 35, 253. (1900.)

vol. xxxvi. — 35

AMERICAN ACADEMY OF ARTS AND SCIENCES.

Report of the Council. — Presented May 8, 1901.
BIOGRAPHICAL NOTICES.

Charles Carroll Everett
Xathaniel Holmes . . .
Silas Whitcomb Holman .
Sylvester R. Koehler . .
John Elbridge Hudson
John Harrison Blake . .
Charles Franklin Dunbar

Ephraim Emerton.

Jeremiah Smi i ii.
Charles R. Cross.
Charles G. Losing.
James B. Thayer.
Clarence John Blake.
Frank Willi \m Tai S8IG.

REPORT OF THE COUXCIL.

The Academy has lost twenty-one members by death since
the annual meeting of May 9, 1900, as follows : Six Resi-
dent Fellows, — Thomas Gaffield, C. C. Everett, N. Holmes,
J. E. Hudson, A. Lowell, and S. R. Koehler; Seven Associate
Fellows, — J. E. Keeler, H. A. Rowland, G. M. Dawson,
J. M. DaCosta, A. Stille, Wm. Mitchell, and E. E. Salisbury ;
Seven Foreign Honorary Members, — C. Hermite, J. G. Agardh,
W. Kiihne, Baron Russell of Killowen, H. Sidgwick, M. Miiller,
Due de Broglie, and the Bishop of Oxford.

CHARLES CARROLL EVERETT.

Charles Carroll Everett, D.D., LL.D., Busscy Professor of
Theology and Dean of the Faculty of Divinity in Harvard University,
died at Cambridge, October 16, 1900, in the seventy-second year of his
age. Dr. Everett was, through his ancestry on both sides, identified with
the community in and about Boston. His father's family bad long been
settled in Dorchester, and that of his mother in Beverly. His father,
a graduate of Harvard College in 1806, removed to Brunswick, Maine,
soon after his marriage, and entered there upon the practice of the law,
which he carried on with distinction throughout his life. Dr. Everett
was graduated from Bowdoin College in 1850, and seems to have been
drawn first to the study of medicine. He was entered during four years
as a student of the Bowdoin Medical College, and Btodied also with a
neighboring practitioner, as was the custom of the time. His literary
tastes, however, as well as his somewhat delicate health, led him in the
years 1851 and 1852 to spend some time in Europe, chiefly in the Btudy
of modern languages and literature, and on his return, while still regis-
tered in the Medical School, he w:i^ appointed to be Instructor in Mod-
ern Languages at Bowdoin. All accounts agree that his service in this

550 CHARLES CARROLL EVERETT.

capacity was admirable and duly appreciated, but Bowdoin was at the
time going through a financial crisis, which caused its governing board
to be more than ordinarily anxious as to a strict interpretation of their
trusts. A large endowment had recently been secured upon distinct
declarations that Bowdoin was " of the Orthodox Congregational denom-
ination." Dr. Everett had been brought up in the orthodox faith of his
mother, but his father was alreaily known as an avowed Unitarian, and
his own views were inclining more and more in that direction. On this
ground his appointment as Professor, twice made by the Trustees, was
rejected by the Overseers, and thus his bent towards theological study
definitely determined. He entered the Harvard Divinity School, was
graduated there in 1859, and at once entered upon his first and only
pastorate at Bangor, Maine. Ten years later, in 1869, he was called to
be a professor at Cambridge, and in 1878 was made Dean of the Faculty
of Divinity. On the 14th of February, 1871, he was elected a Fellow
of the American Academy.

These are the brief outlines of a life singularly filled with varied
activities, yet passed within the narrow limits of a remote parish and
an academic position. As a pastor, Dr. Everett left upon his parish-
ioners an equal impression as preacher and as man. His first passion
was scholarship, but his was a scholarship that did not cut off its votary
from human sympathy. His preaching, cast always in philosophic
form, was yet in a high degree "practical." It addressed itself to the
highest instincts, and sought to direct these by a principle of reason.
In the trying times of the Civil War his voice gave no uncertain sound,
and he was sought by his fellow-citizens as their orator on many an
occasion of civic discussion or commemoration.

The removal to Cambridge gave to Dr. Everett full scope for the
literary and speculative tastes which had been developing in Bangor.
The "Science of Thought," first published in 18G9, was his introduction
to the world of technical scholarship. It was a fresh and individual
attempt to present, on the lines of the new Hegelian thought, the funda-
mental processes of the human reason. It attracted wide attention, and
became for many their first inspiration to clear and rational thinking.
At Cambridge Dr. Everett's activity was largely absorbed in his regular
duties of teaching anil administration. I lis instruction took naturally
the Conn suggested by the nature of his mind. Never, probably, ua-
then- any teaching of theology less conventional or, in the narrow sense,
"academic" than his. Theology was to him not a system fixed in
forms that could be handed on from master to pupil ; it was the inter-

CHARLES CARROLL EVERETT. 551

pretation, on broad philosophic principles, of that which Christianity has
to contribute to the solution of the universal religious problem. Chris-
tianity could be accepted only as it commended itself to the fundamental
needs of the human soid. First, then, it must be shown what these
needs are, and this Dr. Everett undertook to do in his introductory lec-
tures on the Psychological Basis of Religious Faith. Upon this basis
he then developed his presentation of Christian ideas as rationally cor-
responding to a universal and fundamental demand. These lectures were
his real life-work. They furnished the intellectual nucleus about which
he gathered the results of his ever ripening thought aud his ever wider
reading. It is they that have given him his peculiar hold upon a full gen-
eration of students, and it is they that will form, if their substance can be
restored from the scattered material available, the most complete record
of his intellectual life.

Dr. Everett was not a voluminous writer. His best work was done
in response to some outward occasion which roused him to formulate his
thought upon some given problem. Probably the book most likely to
be remembered outside of strictly academic circles is the volume of
essays on Poetry, Comedy, and Duty, published in 1891, in which his
profound reflection, his wide learning, and his playful fancy found equal
scope. His " Gospel of Paul," in 1893, was an altogether original treat-
ment of the Pauline theology, and was, perhaps, the work of his later
years which appealed most strongly to himself. His contributions to the
" New World." of which he was one of the responsible editors, were
numerous and important.

In losing him the University and the world of scholarship have lost a
unique figure, an interpreter of truth in a transition age, a guide to all
those who were seeking light in the confusion of faiths and iu the
temptations of doubt. His value as a scholar is to be measured only
by his singular quality as a man. After all it was his personality that
impressed, and charmed, and made the way clear for the truth he had to
convey. He satisfied the inquirer, not by the cheap method of com-
promise, but by a suhtle gift of leading men away from the (inessentials
to the realities of faith. His mind had that element of greatness which
consists in the ability to see through the mists of the unimportant and
the transient down to the permanent and the essential in the religious
life and not only to see this but to make it clear to others. One of his
greatest aids in exposition was his ever-present sense of humor. It. as
he has taught us, the essence of the comic is to he found in the Bense of
incongruity, it should follow that the truly humorous mind is keen also

552 NATHANIEL HOLMES.

[iscern the coogruities of things. In the midst of the most serious
development of an idea, it \\a~ often a flash of wit that lit up the obscure
point ><> that it remained clear forever. Quite in harmony with his wide
human interest was also hi.* love of outward Nature, and the use lie
made of it in bis writing. All experience contributed with him to the
interpretation of that religious instinct which in turn gave to experience
its meaning and its purpost ,

Ephraim Emerton.

NATHANIEL Hoi. Ml -

rHAWEL Holmes was born at Peterborough, N. II., July 2,1814.
His paternal grandfather was an emigranl from Antrim County. Ireland:
afterwards a soldier in the Revolution and a Deacon in the Presbyterian
Church. His mother was a daughter of a Presbyterian clergyman, Rev.
David Annan, a native of Fifeshire, Scotland. As a boy he worked in
a machine shop and on his father's farm. Prom 1831 to 1833 he was a
student in Phillips-Exeter Academy j and he graduated at Harvard in
1837. After spending a year in Maryland as a private tutor, be studied
law at the Harvard Law School and in the office of Henry II. Fuller,

.. in Boston. In 1839 he began practice at St. Louis. Mn. In 1 ■
he was appointed one of the Judges of the Supreme Court of Missouri,
ring until 1868. His judicial opinions are contained in Volumes 36
to 42 (inclusive) of the Missouri Reports. In 1868 he gave up his
position on the bench to accept the Royall Professorship of Law in the
Harvard Law School. In ls7l' he resigned the professorship and re-
turned to the practice of law in St. Louis, where he remained eleven

\ears. In ] ss:i he returned to Cambridge, and made bis hi ■ there

until bis death, on February 26, 1901 .

Outside the legal profession, Judge Holmes is best known by his

honk entitled "The Authorship of Shakespeare." a work designed to

re that the plays attributed to Shakespeare were written by Bacon.
This book was first published in 1866, and went through three editions.
One of the ablest opponents of the view taken by Judge Holmes has

utly said of the latt' r's l l< : " This contains the fullest and strongest

presentation of the argument in favor of Bacon's authorship which has

I, and it is also marked for its fairness and candor." < Notes

on the B Shaki leare Question, by Hon. Charles Allen, pp. 1 and 2.)

Judge Holmes also published, after his return to Cambridge, a work

SILAS WHITCOMC HOLMAN. 553

in two volumes, "Realistic Idealism in Philosophy Itself," and another
work entitled ''Philosophy of the Universe.'' In 1889 be delivered the
address at the 150th anniversary of the settlement of Peterborough,

The life of Judge Holmes in Cambridge for the last eighteen years
was a very retired one, but it was neither solitary nor unhappy. He was
a bachelor, but always had some relatives in his household, and never
failed to find adequate companionship in his books. A great reader on a
great variety of subjects, with a tenacious memory, he was lull of
knowledge of all sorts, and was a very interesting talker. Jud re
Holmes was a man of great candor and of remarkable equanimity. He
had had successes and disappointments, but lie was neither unduly elated
by the one nor depressed by the other.

Judge Holmes was elected Fellow of the Academy May 21, 1870.
Having removed from the State, be was elected Associate Fellow May
30, 1876. Having returned to Cambridge, his name was restored to the
list of Resident Fellows November 30, 1889.

Jeremiah Smith.

SILAS WHITCOMB HOLMAN.

Silas Whitcomb HOLMAN was born at Harvard, Massachusetts.
January 20,1856, and graduated from the Massachusetts Institute of
Technology in 1876, having made a specialty of the study of physics
throughout his course. He was thereupon appointed to a position as
assistant in the physical laboratory of that institution, but on account of
illness did not enter upon his duties until a year later. Continuing in
the service of the Institute, he was promoted to more advanced positions,
and was made Professor of Physic- in 1893. Even at this date his health,
never firm, had become much impaired, and a few years Liter it became
necessary for him to relinquish active work. In ]s'.»7 he was made
Emeritus Professor of Physics. He died April 1. 1900.

Professor Holman was elected to membership in this Academy March
14, 1883. His original contributions to science were of high merit, and
give evidence both of great skill in manipulation and of remarkably clear
insight into the choice of methods for conducting a difficult investigation.

The most important of his researches are those upon the viscosit] "t
air and carbonic acid as affected by temperature, which were published
in the Proceedings of this Academy in ls7<'> and L885 ; tin' Brsl "f which
was based upon his graduating thesis at the Institute "f Technology.
These contain by far the- most complete Btudy of this difficult Bubjecl which

554 SILAS WHITCOMB HOLMAN.

had been made up to their date, and the results are still of standard value.
Indeed within the past few years they have played an important part iu
the advancement of the kinetic theory of gases.

In the Proceedings for 1*88 is found a further noteworthy paper
written in conjunction with one of his pupils, upon the determination of
fixed reference points for thermometric measurements at high tempera-
tures, in which several such points are established.

A number of years later, in 1895, appeared auother group of papers,
the last published by him, relating to the thermo-electric measurement of
high temperatures, and a single paper upon calorimetry. which subjects
had occupied much of hia attention for some time previously. Of these,
the one entitled " Thermo-electric Interpolation Formulas " is particularly
valuable for its critique of the various methods of interpolation which have
been employed in dealing with the results of high temperature observa-
tions and that upon the "Melting Points of Aluminium. Silver, Gold,
Copper, and Platinum." published in collaboration with his pupils. Law-
rence and Barr, contains what are undoubtedly the best measurements of
the points of fusion of these metals that had been obtained at the time of
their publication. A third paper contains a description of a novel method
of calibrating the Le Chatelier thermo-electric pyrometer, and the fourth
a new method of applying the cooling correction iu measurements of the
heat of combustion.

The papers of Professor Ilolman thus far referred to have all been
published in the Proceedings of this Academy. Several others of minor
importance have appeared in different scientific journals. An extended
critique upon thermometry of precision presented at the Boston meeting
of the American Association of the Advancement of Science in 1880
unfortunately was never printed.

Besides his published researches, Professor Ilolman was the author of
eral valuable scientific works. The two volumes of " Physical Labora-
tory Note-.'* prepared for the use of his pupils in the Massachusetts
Institute of Technology, embody the results of many years of successful
experience in teaching and form an important contribution to the litera-
ture ..f that subject. They contain much original matter and exhibit a
rare discrimination in the selection and comparison of the methods of
measurement which are discussed. This is particularly th< case with the

volume relating to electrical measurement and testing.

In 1892 he published a treatise upon " The Discussion of the Precision

of Measurements," the basis of which consisted of the notes of lectures
D to his classes. This volume, which is <piite unique in its contents,

SILAS WHITCOMB HOLMAX. 555

contains in convenient form a very compendious and lucid consideration
of the application of the principles of least squares to the theory of obser-
vations,, the calculation of their precision and the choice of proportions in
designing physical apparatus to be used for measurement. Its value as
a text-book has been very jrreat.

The collection of four and five place logarithmic tables prepared in
1896 embodies several features of marked originality, and is prefaced by
a brief but exceedingly useful discussion of the fundamental principles of
computation which contains many useful suggestions for the economizing
of labor.

The last work written by Professor Hoi man, entitled " Matter, Energy,
Force, and Work," appeared in 1898, and is of a character widely differ-
ent from any of those which preceded it. It is a philosophical study of
the fundamental concepts of modern physics in which the subject is
approached from the point of view that matter and energy rather than
matter and force are the primary entities with which physics has to deal,
and that matter itself may be dependent upon energy for its own existence.
"While not technical in its character, and intended especially for the help
of teachers not wholly familiar with modern views, it is distinguished
throughout by great clearness, and is a remarkable presentation of the
newer modes of viewing the subjects which it considers.

Valuable as are his scientific publications, however, Professor Holman's
great work was that of a teacher of young men in the laboratory. From
the besrinninsr of his service as an assistant in the Rogers Laboratory of
Physics, his influence was marked, and by his patient labors, extending
through years, he brought the work which was under his charge to a
high state of development. He possessed great skill in the planning of
apparatus and methods, and rare judgment as to the processes best
suited either for purposes of instruction or for the securing of accurate
scientific results. To the development of the Laboratory of Electrical
Measurements in the Massachusetts Institute of Technology, he gave for
years his best endeavors, and to him is due the success of its work. He
was also placed in charge of the newly-instituted Laboratory of Heat
Measurements, and though prevented by failing health from developing
this as he would have chosen, he laid a solid foundation for those coming
after him.

Professor Ilolman was born a teacher, and never grew wearj in his
profession. His personal relations with his pupils were \. rv intimate.
By that example, which is better than the wisest precept, be impressed
upon them the pre-eminent necessity of thoroughness, accuracy, ami

556 BTLVESTEB II. KOEHLER.

honesty in all work which they might be called upon to perform either
as students or in professional life. He is remembered by them with the
most affectionate regard.

Reference has already l>een made to the interference of ill-health with
the prosecution of the labors of Professor Bolman. In fact, after reach-
ins manhood he was never in good health, and during almost the whole
of his active life a> a teacher be struggled with a painful chronic disease,
whicli gradually, though with some intermissions, sapped his strength.
His cheerful disposition and persistence iu carrying on his work were
Buch thai none but t li< >><• who knew him well wen- aware of the fact that
it was only hi> indomitable courage which prevented him from yielding
to his malady for som<- years before it finally overcame him. In the
sprinir <>f I *'.•<• lie was obliged to discontinue work for a time. He spent
the following year abroad, and came home much improved in health, hut
the relief was only temporary. In 1895 he finally gave up his work of
Instruction. For some years after this, however, though confined to his
chair, and at last even deprived of sight, he continued to labor diligently,
and published the tables of logarithms and the work on Matter and Euergj

mentioned above. Hi~ mind was clear to the last, and his cheerful:
never forsook him. His latest years were his best ones, and his whole
life was a fine illustration of the manner in which a noble spirit may rise
Ruperior to circumstances and produce the best results under conditions
to which an ordinary mind would utterly succumb.

Chas. R. Ckoss.

SYLVESTER R. KOEHLER.

SylveBTI B R. Kor.in.ru was born in Leipsic in 1887. He came to
America at the age of twelve years. Son of an artist, grandson of a
musician, he was destined by inheritance to an artistic career. Its bent
was determined by his moving to Boston in isc.s and entering the estab-
lishment of L. Prang & Co. as technical manager. Tliis position, after
ten years, was given up, that he mi'_rht devote himself exclusively to his
art Btudies. With Charles C. Perkins and William ('. Prime as asso-
ciate editor-, he launched the " Art Review," the most dignified and

scholarly periodical devoted to art thai has 1 n published in the United

States. It was aimed, to quote the preface, "to dwell upon the larger,
more robust, more serious features ofi lern art," but it was in advance

of it- ti — the circle tO which it appealed Was small, and when alter

SYLVESTER R. KOEHLER. 557

two years its publication ceased, its projector modestly claimed that it bad
" quickened somewhat the forces at work iu the healthy development of
art in the United States."

Contributions without number flowed from his pen to magazines and
journals in America, to the " Zeitschrift fur bildende Kunst" ami other
periodicals in Germany, ami to a few of the Loudon publication-.

For a while he held the appointment of Curator of the Section of
Graphic Arts in the United States National Museum at Washington.

When, in 1888, the Trustees of the Museum of Fine Art-, Boston,
found themselves under the necessity of appointing a Curator of the
Department of Prints, Koehler was the expert to whom they turned
without hesitation. His appoiutment to this curatorship gave him the
position for which his previous career had fitted him. To the years
passed in the Prang establishment he owed a mastery of every detail of
the technical processes used in the Graphic Arts. This technical knowl-
edge was supplemented by an artistic temperament, which Bhowed itself
also in his fondness for music, in his love of verse, and his skill, though
a moderate one, with the pencil. Years of study, too, had given him
an intimate acquaintance with the history of his art, and confirmed his
judgment. He was an admirable critic of work, both creative and tech-
nical. These were rare qualifications for the post offered. In it he not
only became the ultimate authority in the land of his adoption, but his
knowledge and judgment were held in gnat esteem in the art centres of
Europe.

A man of strong individuality, of critical mind, interested in all prob-
lems of life and religion, a bold thinker on questions of social reform, a
sharp critic of public abuses, bitterly conscious of the injustices of tin-
world as compared with the ideal life he pictured in verse, he was a
radical in his views of bettering human institutions. Vet he was a sym-
pathetic friend, sociable, of quaint humor, and in the conduct of his
department distinguished for unfailing, unwearied courtesy to all seeker-
for information.

At the Museum his career was one of ceaseless activity. Numerous
exhibitions were held, the catalogues of which offered the opportunity to
impart his knowledge in the introductions and in tin- copious nol
descriptive, explanatory, and critical, of the etched work of Rembrandt,
of Albert Durer, of Blake, Meryon, Seymour-Haden, the Cheneys, and
on various other occasions. Notable among these was thai tor an exhibi-
tion in 1892, " Illustrating the Technical Methods of the Reproductive
Arts from the XV. Century." -with special reference to the phot.,.

558 JOHN ELORIDGE HUDSON.

mechanical processes," for which then- was a steady demand from museums

and collectors in Europe long after the edition was exhausted.

His most important work," Etching," a sumptuous volume with thirty
plates by old and modern etchers, and numerous reproductions, was pub-
lished in New York iu 1885, before his appointment at the Museum.

In 1893 he delivered a course of nine lectures before the Lowell Insti-
tute, subsequently repeated in Washington, on 4t Old and Modern
Methods of Engraving." At other times he lectured before the Art Club,
the Society of Arts, and on various occasions to private classes.

In 1892 Harvard conferred the honorary degree of A.M. lie was
elected Resident Fellow of the Academy May 10, 1893.

The recent transfer to Harvard of the collections deposited with the
Museum, and the sudden acquisition by purchase and bequest of great
numbers of prints a year or two before his death, was a source of anxiety
to him in his feeble health. The end came suddenly, unexpectedly, but
painlessly, following within a year that of his wife. For Ids reputation
one can but regret that his untimely death prevented the completion of a
" History of the Art of Color-printing." for which he had accumulated a
large amount of material, — a difficult task, for which no one was so well
fitted as he to sift the facts and refute prevalent errors.

The large and valuable library which he had accumulated he gave,
with many prints, to the Museum of Fine Arts. A list of his publica-
tions is given in the twenty-fifth annual report of that Museum.

Chas. G. Loring.

JOHN ELBRIDGE HUDSON.

The duty has been assigned to me of communicating to the Academy
some account of our late associate, John Elbridge Hudson, who w;^
elected a Fellow on dune 15, 1892, and was a member of the Council
fro,,. May 8, 1896, to May 10, 1899.

It is a grateful duty, for 1 had known Mr. Hudson long and well, and
had for him a ver\ ir'eat regard. He was a student at Harvard Colli _■
when I first saw him, — a shy. studious, thoughtful boy, at the head of
bis class. A few years later be walked into the office of the law firm
with which I was connected and asked to be received as a student. From

that time to the moment of his death — for half of my life and more
than half of his — I Baw much of him. With few men could he have
talked more confidentially of what most concerned him than he did with
. and certainly with few men did I hold a more intimate friendship.

JOHN ELBRIDGE HUDSON. 559

At the time of his death, — on the first day of October last, Buddenl v. :it
the railroad station in Beverly, five minutes before he was to take his
train for his daily tasks in Boston, — he was a little over sixty-one years
of age, having been born in Lynn August 3, 1839. He was a man of a
very handsome and impressive presence, tall, large, with a massive head,
and an expression in his face of quiet dignity, strength, and composure
which truly reported the quality of the inward man, and attracted a re-
spect and confidence that were never disappointed.

Hudson's ancestry runs back to the very early days of the Massachu-
setts Colony. At his own birthplace of Lynn, the earliest immigrant of
his family had settled, — Thomas Hudson, — about 1G30. It is said
that the first iron works in the country were established ou his land, at
the head of navigation, below the ford, on the Saugus River. Nine years
ago our associate presented to his native city an iron kettle, the first
casting made at these works, just two hundred and fifty years before.

His father was John Hudson, of Lynn, and his mother Elizabeth Chase
Hall Hilliard, of Cornish, New Hampshire. Through her Mr. Hudson
was descended from two clergymen, her great-grandfather and her grand-
father, — -one the Rev. Dr. David Hall, a graduate of Harvard in 17lM.
and for sixty years minister of the First Church in Sutton, Massachusetts,
and the other the Rev. Samuel Hilliard, "a pioneer in Universalism, and
a soldier of the Revolution, serving at Bunker Hill and Bennington."
Mr. Hudson himself was brought up as a Unitarian, and although not a
regular attendant upon any church, was a member of the Unitarian Club
of Boston.

Educated at the public schools of Lynn, Hudson provided himself with
whatever other special fitting for college was needed. He entered Har-
vard in 1858, being a little older than the average of his class: took
distinguished rank as a scholar and graduated in 1862 at the head of his
class; became at once a tutor in Harvard College, where for three years
he taught Greek, Latin, and Ancient History; and for two years of the
same period was a member of the Harvard Law Seliool, where lie gradu-
ated in 1865.

His work as a tutor gave great satisfaction, and he would have been
welcomed as a permanent member of the teaching force at the Coll<
But, with a sound instinct as to the character and reach of hi- own powers,
he chose the world of affairs. As a scholar and a teacher he would, un-

* Memoir by George V. Leverett ; The New-England Historical ami Genealogt

cal Register, LV. 136.

560 JOHN KLDR1DGE HUDSON.

doubtedly, have been distinguished, but be bad a very great and unusual
capacity for business, for shaping large affairs and for influencing men. a
faculty that would have largely missed its opportunity in the quiet life of
a college officer.

lie entered the office of Messrs. Chandler, Shattuck and Thayer, as
a student, in Boston, in 18G5 or I8GG. Admitted to the Suffolk Bar in
( K'tober, 18GG, he soou became managing clerk in the office (if the firm just
mentioned on terms which indicated a high appreciation of bis ability ;
and in February. 1870, on the retirement from the firm of Mr. (ieor^e
0. Shattuck, lie became its junior member. Here be continued, during
some changes in the firm, until its dissolution in 1878. In the interval
between this and the year 1880, when that connection with the telephone
busiuess began which was to last for the remaining twenty years of his
life, he practised law alone, and added some editorial work upon the tenth
volume of the United States Digest. That was a task which it was quite
possible to carry through in a seemingly respectable and yet entirely
perfunctory fashion. But Mr. Hudson took it up as he took up every-
thing, and planned a volume of distinguished merit. On entering the
service of the telephone company be had to leave the completion of the
book to another; but he had begun by making a new analysis and classi-
fication of the titles of the law employed in the Digest, — one which was
so much valued that it forms to-day the model of a recent great under-
taking by a "Western publishing house, known as the Century Digest, now
in general use among lawyers.

Mr. Hudson was thus an active member of the bar for fourteen years.
His preference was for office work ; very seldom could he be induced to
try a case, or to argue a point of law before a court. During the nine
years that we were connected with the same firm, before I went to the
Law School, in 1874. I can speak of bis work from an intimate knowledge
of it. He had the oversight of our accounts, and took charge of a great
pari of the office work, such as the drawing of contracts and wills, and
the preparation of pleadings and court papers. In all this, his work was
admirably well done. The character of it, at a little later period, attracted
the attention of the Chief .Justice of the Supreme Court, who sent for
him and expressed a wish that he would take the place of Clerk of that
court. I recall also the great satisfaction expre 36(3 by a client, a \ en-
able business man in the China trade, who had had occasion to i ■-ult

him about some tangled affairs. "] have not seen," he said, "such a
head for complicated accounts since my earl} experience with John
W. Forbes."

JOHN ELBRIDGE HUDSON. 561

As I have said, there was another department of work in which it was
practically impossible to interest Hudson, that is to say, work in court.
You could hardly drive him to take hold of that. He took no pleasure
in it and showed little capacity for it. He was always a shy person, little
inclined to put himself forward, and absolutely unwilling to appear in
public, if he could avoid it ; and for that reason the breadth and versa-
tility of his extraordinary powers were sometimes overlooked, [ndei d
many of those who knew him well and best appreciated his remarkable
qualities were surprised at the developments of later years.

Early in 1880, Colonel William II. Forbes, the President of what was
soon to become the American Bell Telephone Company, who had known
Mr. Hudson in college and also as a lawyer, invited him to become the
Solicitor of the Company. The invitation was accepted, and Hudsou
entered at once upon his new duties. Five years later he was asked to
become the General Manager. To accept such a place as that was a
serious step. So far he had not left the law, but this new proposition
would plunge him into a career of business, and business of a very engross-
ing sort. He came to talk it over with me ; and I had many misgivings.
He knew, the Company knew, all his friends knew, what he could do in
the law. But this was a new venture. What if all this tremendous,
novel, swiftly developing business should not suit him, or should prove
too much for him ? Was he sure that he could handle it ?

I was greatly impressed by his answer. 01), yes, indeed ; as to that he
had no doubt whatever; he could handle it well enough. As Solicitor
he had got a good insight into the nature of it all, and he had no fears
on that head. This confidence in his own powers of dealing with the
men and the affairs of so great a concern, a confidence fully justified by
the event, opened my eyes to a new side of Hudson's capacity. He took
the office, without relinquishing the place of Solicitor, filled it to tin-
entire approval of the Company, added to it the next year that of Vice-
President, and in 1887 that of President of the American Telephone and
Telegraph Company, then known as The Long Distance Company, and
two years later became President also of the main organization, the
American Bell Telephone Company. These last two offices he held with
great success and distinction up to the time of his death; tor the final
steps in the absorption%of the last named company in the firsl were not
then, and I believe are not yet, completed.

Of the ability which .Mr. Hudson showed in guiding and shaping the

development of this new and complex industry of the teleph , others

vol. xxx vi. — 36

562 JOHN ELBRIDGE HUDSON.

who were associated with him have already Bpoken. Mr. Francis Blake,
one of the directors of the Telephone Company, has -aid that while Mr.
Hudson was Manager and President tin- Dumber of miles of telephone
increased more than tenfold, to over a million miles in 1899, and the
number of exchange connections more than sixfold, reaching nearly
seventeen hundred millions in the same year: and he adds: kt Moreover,
during this period there was conceived and developed a system of long-
distance Bervice which brought more than half the population of the United
States within the limits of telephonic Speech. These statistics,'' he adds,
"emphasize the broad statement that the growth of .Mr. Hudson's busi-
ness capacity not only kept pace with, but kept in advance of the ever
increasing needs of the companies under his control."*

Added to great intellectual capacity, to a remarkable Btrength, grasp,
and tenacity of mind. .Mr. Hudson had another source of power, — his sound
moral quality, lie made no parade of his integrity, but he was thoroughly
hon68t and honorable, and all who dealt with him saw it. In one of the
Company's great law-suits an antagonist thought fit to charge it with
indirection in a particular matter. The counsel of the Company, the late
William <;. Russell, met the charge with a statement by Mr. Hudson.
" And I need not say to the court." he added. " that on a question of tact
within his knowledge, the word of John E. Hudson imports absolute
verity." To a specific reliance upon these personal qualities of the Presi-
dent of the Telephone Company, upon his great intellectual gifts, his
forecasting and shaping power, his sound judgment, cautious arid yet
bold in advancing to meet the great emergencies which he foresaw, upon
his absolute integrity and his power over men, in a word, to personal
confidence in him. may be traced the investment of millions of dollars in
that great corporation.

Undoubtedly Mr. Hudson sacrificed his life to the enormous and ever-
growing requirements of the ollice which he held. It was his habit to
rest by going abroad for a month or two in the summer, and to pa-- the
remainder of that season at the seashore within easy reach of his ollice.
I' i- -aid, and probably with truth, that he lacked somewhat in one of

the qualities of a great administrator, namely, in the power to turn over
work io subordinates. He could not bear to see work imperfectly done,

when he himself could do it -o thoroughly well. His friend- had lotc

urged him to withdraw from these beavj cares, and a few years ago he had

• Memoir by Francia Blake; Proceedings of the American Antiquarian Society,
October 24. 1900.

JOHN ELBRIDGE HUDSON. 563

well-nigh done it. But he was persuaded to remain, arranging at that
time to withdraw in the year 1900. Doubtless, like all strong men, he
enjoyed the exercise of his strength. Moreover, the development of the

business and increasing complications seemed to demand his personal
attention just a little longer, — and so the end came as it did.
seemed, for the most part, to bear it all easily enough, but the strain \
immense, incessant, increasing ; and before he knew, it was too late.

A word or two more should be said as to Mr. Hudson's mental habits,
and his methods of working; aud a word or two also as to the personal
qualities that made him much beloved.

At home Mr Hudson used to have by the side of his plate, as he sat
at table, a pencil aud paper, for he knew the worth of a memorandum
taken at the moment. In his business he was in the habit of causing
be taken and preserved such memoranda of all that took place at each
stage of any particular affair. This full record, perfectly arranged and
indexed, was of the utmost service to him in handling his great business.
He had only to turn to his books to find a record of everything. Often,
indeed, he had no need to turn to his books, unless to convince his inter-
locutor ; for he had an extraordinary capacity of remembering facts, of
visualizing them, and holding them all mapped and co-ordinated in his
mind.

In his private studies he was apt to begin by preparing a chart of the
subject, with names and dates and the order and place of leading facts
aud events, all set forth with extraordinary neatness, open to inspection,
aud speaking volumes to a glance of the eye. These things, thus quickly
visible, passed over into his mind and stood there fixed permanently in a
rational order. It was so with places. London and Paris and all their
streets he saw. He had explored the maps so that he hardly needed them
longer. His mind held the maps.

This faculty gave an extraordinary interest to his conversation. Last
summer I passed several days with him at his house, and he, later, a week
with me at mine. He had been reading Plutarch, and everything ab
him that he could lay hands on. He was trying to place him and his
thought in their true relation to the men and the ideas of the time just
past and just to come. He had been reading also of Alexander, and
reading with equal ease in the Greek and Latin authors as in those in
our own tongue. It was a pleasure of the highest sort to listen to his
talk. The precision and extent of his knowledge, the way in which it
lay in his mind, co-ordinated with whatever related things threw liji

564 JOITN ELBBIDGE HUDSON.

upon it, these and the breadth and illuminating good sense of his own
reflect] >ns were equally instructive and delightful. Alexander and Plu-
tarch stood out before you in their true place in history, and in their true
relations t<> the men, the dates, and the events of their time.

No one could Bee Hudson without guessing at the strength and force
of his character ; and no one who knew him well could fail to see that
under his gentle demeanor there lay qualities of energy and passion that
were not to lie trifled with when once thev were aroused. But he was
an affectionate and charming friend, and one thai women and little chil-
dren and those who were dependent upon him loved.

Mr. Hudson married in 1*71 Miss Eunice Wells Healey, of Hamp-
ton Falls. New Hampshire, who survives him, to bear a most heavy
loss. He left no children. One sister also survives him, the wife of
inel J. I Iollis of Lynn.

He went little into society and but little to the larger clubs.* He was
happiest at home, and there he gave himself up to the refined and simple
pleasures in which he always had his chief enjoyment. He had a large
library, of greal range and variety, to which be was forever adding.
Mosl of all he seemed to like, at the end of the da v. to sit down among
his books and explore his old friends the Creek and Roman classics, —
reading them, as he did, with entire ease in the original. When it came
to the matter of his real tastes and liking, as his associate Mr. Leverett
has happily said, " He was, above all, a scholar, fond of his home."

James 15. Thayer.

• Mr. Hnd8on was a member of many societies and clubs. In the Memoir for
the Historic Gei al Society, already quoted, Mr. Leverett says: "Mr.

Hudson was at the time of liia death a vice-president of this Society. He was alao
a fellow of the American Academy of Arte and Science-, a member of the Corpo-
ration of the husetta Institute of Technology, a member of the American
Antiquarian Society, the American Association for the Advancement of Science,
the British Association for the Advancement of Science. American Geographical
National Geographic Society, the Colonial Socii ty of Massachusetts, the
American Institute of Electrical Engineers, the Virginia Historical Society, the
ciation for the Preservation of Virginia Antiquities, the Bostonian Society,
Selden Society, Hakluyl Society, Lynn Historical Society, the Bat Association of
the i B Jton, and also of the Algonquin, Boston Art. Exchange, National

St. Botolph, Union, University, and other social clubs." It may lie added
that he took much pleasure in one or two social clubs made up of college friends
m temporaries.

JOHN HARRISON BLAKE. 565

JOHN 1 1 AUK ISDN BLAKE.

In reviewing the printed matter, letters, completed manuscripts, and
note-books which furnish the material for a personal memoir of my father,
I have been impressed by the paucity of record of the achievements of an
active life extending over the greater part of a century.

The reason for this is to be found in the character of the life itself,
looking rather to accomplishment than to recognition, and seeking to ex-
press the sense of an obligation for the privilege of living by doing the
work at hand simply and well.

Of published papers there are but few, — the Transactions of this Acad-
emy, of which he was elected a member May 30, 1843, contain none,
other journals in the library of the Academy but five, and the records
of the Boston Society of Civil Engineers, of which he was one of the
founders and its first Secretary, an equally small number.

Of letters, especially family letters, there are many, all clearly written,
containing information as to his travels and treating of the subjects in the
study of which he was most interested.

Of the manuscripts, the majority, in the form of essays, were written
after his retirement from active life with the evident purpose of continu-
ing a companionship which has, to his son, the value both of a precious
memory and a continued inspiration; the persistency with which this
purpose was pursued, under conditions of failing strength and sight, is
shown especially in one of them, begun in ink, continued in black and
then in blue pencil when the blue mark alone was visible to him. and
concluded by sense of touch.

These essays cover a wide range of scientific subjects and bear witness
to an intellectual activity persisting to within the last three years of a

life which ended, as gently and as graciously as it had I n lived, at the

awe of ninety years. Through this life there had run one dominant pur-
pose, — that of usefulness; in it there had always been one keen pleasure,

that of scientific research. The note-books, containing many valuable

records and memoranda, are the transcripts of a variously active profes-
sional life, — one series covers very nearly the whole of the early experi-
mental and constructive history of the manufacture of illuminating gas in
this country, another is a record of researches in the chemistry of arts
and manufacture, and another is devoted to mining, metallurgy, and civil
engineering.

John H. Blake, the youngest son of Thomas and Mary Lowell Bar
nard Blake, was born December 5, 1808, in the house -till Btanding on

5G6 JOHN BARRISON BLAKE.

Washington Street, corner of Union Park Street; he died in his home on

Marlbo _ Si ton, July 5, 1899, in his ninety-first year. Mr.

educated in private schools and in the English High School,

which lie entered in 1821 as a member of it- first class. Later he became

a pupil of the Rev. James Blake 1 1 owe. then rector of the West Parish
Church in Claremont, X. II.. whose daughter, a Becond cousin, he mar-
ried on his return from his explorations in South Ameri

While his Btudies with Mr. Howe were mainly classical and he was
fitted to enter Harvard, hia interest in chemistry and in anatomy was such
' lead him to prefer tfa i lies to those of a collegiate course prin-

cipally literary ami mathematical. It was in pursuit of his chosen
subjects that, after a period of study in a chemical manufacturing estab-
lishment, he became assistant to. and pupil of. Dr. Webster, from whom
be received valuable instruction, which enabled him in tie' year 1827,
when only nineteen years of age, to establish, with money advanced by
. the Norfolk Laboratory, for the manufacture of purr dri _-
and chemicals. This laboratory was situated in Jamaica Plain near
Forest Hills, upon the Dedham Turnpike. One of its products was pure
sulphuric ether, and it was in the larger laboratory of later construction
that the ether used in the first demonstration of the value of ether ana. s-

iia was made under Mr. Blake's personal supervision. In addition to
the commercial work of the laboratory .Mr. Blake carried on a series of
invt »ns into the physiological effects of poisons, the composition of

precious stones, and the production of alloys applicable to the mechanic
arts. At the end of three \ ears of successful operation the buildings u
destroyed by a fire, resulting from the explosion of a carboy of ether,
but were immediately rebuilt on a larger scale j a joint stock company was
formed, arrangements were made with the Rothschilds for the importation
of quicksilver, and with a house in Tuscany controlling what was then the
world's principal supply of boric acid; and Mr. Maximilian Isnard, who
introduced the manufacture of beet root Bugar into France, became an

>ciate.

At that time but little was known of the sources of supply of nitrate

of -oda. of which large quantities were used in the work-, beyond the

name, of the -mall port- on the roast of Peru from which it was shipped.

[gnorance on the subject, the value of the article, and the novelty of

ring and exploring an unknown region were sufficient incentives to

turn Mr. Blake's thoughts in tin- direction rather than alone the beaten

line- of travel in search of the pest and recreation which he needed after

eight year- of anxious labor. Books gave very little information concern-

JOH\ HARRISON BLAKE.

ing the country between the Pacific Ocean and the And. stitutius

the extreme southern part of Peru, the western part of Bolivia, and
northern part of Chili, and all the knowledge that could In- obtaiued was
that it was, for the most part, uninhabited and uninhabitabl tute

of vegetation, and known as the Desert of Atacama ; it was on the sh
of the northern part of this desert that the shipment ports referred to
were situated.

The winter of 183-J and 1836 was very old. New York harbor was
frozen over, and it was not until the 10th of February, 1836, nearly a
month after the time proposed for her departure, that the ship " Factor,"
in which Mr. Blake was a passenger, made her way through a channel
cut in the ice and sailed Eor Valparaiso, where she arrived June !>,
sailing again on the 9th of duly for Arica and Tacna, whence Mr. Blake
prOc :e led by land to Pisaqua and Iquique, arriving at the latter p]
on the 6th of August. The next three months were devoted to surveys
in the province of Tarapaca, and on the 7th of November Mr. Blake left
Iquique with a pack train, two Indians, and dogs to make the first re-
corded exploration of the Desert of Atacama from north to south, arriving
at Valparaiso on the 10th of March, having occupied four months and
three days in a trying passage over an arid and waterless region, in which
all of the animals were lost and the men nearly perished from thirst.

On March 15 Mr. Blake left Valparaiso for Buenos Ayres by San-
tiago, the pass of Uspalato and Mendoza, crossing the Andes and the
Pampas de la Plata, arriving on the 28th of April and making prepara-
tions for immediate departure for the United States.*

At this time Rosas, the then Dictator of the Argentine Republic, was
engaged in strengthening his position by military activity and the pro-
j cted subjection of the Indian tribes to the westward. Mr. Blake \
lined as consulting engineer on fortifications, and was not relea
until the autumn of 1837/j" when he returned to the United St t< - to find

* The only record of this interesting and perilous journey is to be found in
family letters, in the collection of mummies and other objects oi archaeological
value now in the Peabody Museum, Cambridge, ami in the description oi I
collection published from .Mr. Blake's notes in the reports of the Museum, i
carefully kept notebook, containing not only the daily incidents ol travel, but
especially the memoranda of geologic observations, of barometric m<
ami surveys, was stolen after Mr. Blake's return to this country, and m
recovered.

Information to he derived from a traveller who had just crossed the continent
was of value, ami Mr. Blake received a courteous not. that a house adjoin-

ing that of the British Embassy bad been placed at his disposal, and req liim

JOHN HABRI80N BLAKE.

a condition of general financial disaster, in which the Norfolk Laboratory
bad shared aud which made it impossible to take up the grants for the
mining and exportation of nitre BCcured from the Peruvian government.

Under these conditions h<- a pted the management of the Fernandez

copper mines in Santa Clara, Cuba, married, and took up his residence
there, where, in addition to the work in hand, he made observations on
the character and climate of the country. In 1*17 he assisted in the g
logic survey of [sle Roy ale, Lake Superior,* and in 1848 entered into
partnership with Franklin Darracott as a civil engineer, to which work
added that of a consulting chemist and geologist. It was his custom
to make no charge for consultation to individuals seeking to develop new
industries, rega dii 2 this as his contribution to the general \
Among many to whom be gave valuable advice were Goodyear and
Babbitt.

The business of the firm of Blake and Darracott had largely to do with
ojineering and thi truction < a works, and Mr. Blake organ-

ized and was at one time president of live gas companies, his executive
ability and power of control over men making such work a pleasure. In
addition be was interested in iron and gold mining, carrying on the ore
- and blast furnaces and car wheel works at Brandon, Vermont, and

_ inizingand operating the Yahoola River Hydraulic .Mining Company,
of which he was president, in the Dahlonega belt, Georgia. After the
lution of the firm of Blake and Darracott he became interested in
Btreet railway-, building the Middlesex Road and being the President of
the .Metropolitan Road during the period of the Civil War: subsequently
he was President of the Connecticut Arms Manufacturing Company, and

ized and was the first President of the Chapman Valve Manuf'ac-
turi _ I ■ npany.

One of Mr. Blake's latest contributions to manufactures was the
called antique glass. Wishing to carry out previous experiments on
molecular movement in solids at protracted high temperatures, he i
Btructed a crucible furnace in South Boston, and in order to make it pay
i manufactured glass upon the basis of bis earlier analyses i t
precious stones, the result being a glass of great brilliancy and vivid
color.

the Dictator. The invitation was declined, but wai
n an equally courteous note brought by m file of soldiers, the note further
that the Dictator trusted that the invitation would be accepted, as it «
n him to be obliged to provide any narrower accommodation.

promontory at the northeasterly extremity of the islan 1 is called Blake
nt.

CHARLES FRANKLIN DUNBAR.

The mental activity which had stimulated bodily action beyond the usual
term of working years continued to fiud its expression, even after be had
become confined to the limits of his own home, in the manuscripts which
contain those products of his mental laboratory impossible ol record in
busier times ; as might be expected they give an insight into the motive
power of his life and show the strength of a character which looked for-
ward calmly to bodily dissolution as a part of the process of growth to
greater knowledge.

Through the vicissitudes of incessant and protracted work, with the
usual meed of disappointment and much of physical pain, Mr. Blake held
always the cheerful courage born of a simple faith, which counted life as a
primary school and the suffering of his advanced years as a part of its
graduating exercises ; prominent in my memory of him are his fearli
nes's, his kindliness, his love of truth, and his earnest desire not to fail in
doing his part of the world's work, whatever that might honestly be. and
this also, that in fifty-seven years of a dear and close companionship, I
cannot recall a single unkind, unjust, or impatient word.

Clarence John Blake.

CHARLES FRANKLIN DUNBAR.

Charles Franklin Dunbar, Fellow of the American Academy for
twenty-eight* years, Professor of Political Economy in Harvard Uni-
versity for nearly thirty years, was born at Abington, Massach
July 28, 1830. and died at Cambridge, Mass., January 29, 1900.

Professor Dunbar's career divides itself into two very different parts ;
a first, during winch he was editor and guiding spirit of the Boston Daily
Advertiser; and a second, during which he lived the quiet life of the
teacher and scholar.

It was not until he had reached mature manhood that he entered on
his newspaper career. After graduating from Harvard College in 1851,
he en erased for a short time in business ; then, health failing. Bpenl a
year in farming: then studied ai the Harvard Law School and in the
office of the late Justice E. R. Hoar, and was admitted to the Bar in
1858. Meanwhile, contributions from his pen had appeared in the
Advertiser; and finally, in L859, he 1 ame permanently associated with

* Elected January 31, 1

"7" CHARLES FRANKLIN DUNBAR.

that newspaper. At fii <• editor, he became in lSGt sole

tor, and such he remained until he severed his connection
with the Advertiser.

'I'h ■ decade daring which he was thus in charge of the most influential
new in N w England was the most trying and perhaps the most

important in the country's history. Hie position as editor brought
him into contact with leading men in every sort of career in New
England. Both his conduct of the paper, and his association with
men, gradually gave him a position of respect and confidence in t lie
imunity, rarelj ned by those in charm' of ephemeral pub-

lications. He wrote constantly on a great variety of subjects; on
political and military affairs as a matter of course, but with special

and with unusual judgment on the remarkable financial and c
nomic events of the period. His editorials were marked from the outset
l>y the _ and dignity of style which characterized everything that

• from his pen. They showed, moreover, the firm and unwaver-
ing spirit of the man; never abating by a jot the conviction that in spite
of defeat and disaster, in spite of foreign complications and d
disaffection, the war must he carried on unflinchingly until the supremacy
of the Union should be restored. There is not only steadfast faith, but

•i inspiring eloquence, in the editorial pages of the Adver

sor Dunbar conducted them; and not seldom, after a military
failure, his courageous words rang through the community like a bugle
blast.

The financial and economic event? of this period were of the most

.ordinary and varied kind. A huge national debt, anew banking
tem, an imi and complicated system of taxation, a high protective

tarifl he issue of paper money, a wearisome struggle I

the ad\ of paper money and Bpecie, the turmoil of reconstruction

in the South. — such were the phenomena to which the editor of
the Advertiser was compelled to give daily attention. IIi> inborn apti-
tude led him to observe the course of events with 1. gacity, and
gave him a fund of i ince invaluable for his later career. Few
have been bo fortunate in having Keen brought into unremit-
ting contact with the actual affairs of life. Few also have be □
fortunate ii tact with men of all cla d all opinions.
I >.iily there came into the office of the editor of the Advertiser persons of
ort, bringing advice, exhortation, information. A characti ristic
tniit of Profe sor Dunbar's showed itself in th —a

irkable capacity for silent attention. However certain of his own

CHARLES FRANKLIN DUNBAR. 571

ground, he would listen without response to those whose view
different from his own, refrain from stating his ohjections unless the
situation imperatively called for statement, and give bis auditor an
impression, and a true impression, of respectful and sympathetic interi
and yet in due time would follow the course which bis own judgment
dictated as wise. The quality of his mind was eminently judicial. He
saw all sides of a difficult question so clearly that he sympathized with
those who saw perhaps only one side. In the Advertiser office be dealt
with business men, statesmen, soldiers, conservatives, radicals, vision-
aries ; learned something from all, dealt courteously with all, gained tin-
respect of all, and yet never failed to maintain his own sound
independent judgment.

The most active and strenuous years of Professor Dunbar's life, between
the ages of twenty-nine and thirty-nine, were given to the Advertiser.
In his hands its editorship was distinctly a public service: and, cool-
headed and sagacious as he was, uninfluenced by any vapid sentimental-
ism, he so regarded his vocation. But his strength, never very great,
was seriously shaken by these ten years of severe application, and in
1869, when the Advertiser changed bauds, Professor Dunbar was glad to
dispose of his interest and to retire from the paper.

Shortly after, be was offered a professorship of political economy
in Harvard University. This was a career he had never looked for-
ward to, and he doubted bis own capacity for it. Nevertheless, ai
some hesitation, be accepted, on condition that be should have time for
restoring bis strength and adding to his equipment. After two years
spent in Europe in study and travel, he enured in 1*71 on the duties of
the professorship, to which be devoted himself for the rest of his li

Although thus launched on the career of a scholar and teacher, his
abilities were such as to cause him to be enlisted soon in the worb of
guiding and managing the affairs of the University. On the retiremi
of the late Professor Gurney, in 1876, he became Dean of th< Faculty
of Harvard College, and retained thai post until 1882. When the
present Faculty of Arts and Sciences in Harvard University was organ-
ized in 1890, he became its first Dean, and bo acted until 1895. In
addition, he served frequently on commit and was in constant

intercourse with the President of the University, who relied greatlj on
his advice. Repeatedly through bis academic career, be was called upon
to act as judge, as mediator and pacificator, at organizer of new pis
as administrator of new Bystems. All these duties were discharged with
remarkable judgment and succi they were fell b; him to be

572 CHARLES FRANKLIN DUNBAR.

distractions from his chief task as professor iu a great institution of
learuin

Professor Dunbar's career as editor, and bis administrative work in
Harvard University, need to be borne in mind when making an esti-
mate of his work >lar and man of Bcieuce. To those who knew
him well, nothing was more admirable in his career than the* solidity of
lii> scholarly attainments, the breadth of his interests, the maturity of
hia conclusions on his chosen subjects. It might have been expected
that one who bad been a busy newspaper editor, and who remained to
the end keeuly interested in current political happenings, should con-
tinue to deal largely with questions of the day, and take an active part
in current discussion of public issues. Professor Dunbar, however, had
too clear a perception of the ideals and duties of a scholar to give him-
selfto newspaper and periodical writing. For many years he delved in
the literature of political economy at large, and equipped bimself in the
whole range of his subject. Not only the writings of contemporary
economists, but those of earlier days, especially the English and French
authors of the seventeenth and eighteenth centuries, and those of
I; mlo's school, were thoroughly examined. It is characteristic of
1' »or Dunbar that notwithstanding the wide scope of his reading
in tin- theoretic literature of political economy, he published virtually
nothing on this phase of the Bubject : though tin' maturity of the conclu-
sions derived from that reading are unmistakably evident in some of his
vs on the recent phases of economic theory. He regarded these
researcl utial to his equipment as a University teacher, partly
also a- preparation for the inquiries by which he hoped eventually to
contribute to the world's stock of knowledge and thought.

The special subjects on which he planned to publish the results of

arch, and to which he gave nio-t attention in the later years of his
life, were public finance, taxation, currency, banking. It was to t1
that he had given most attention among the economic topics that pre-
sented tin to him as editor of the Advertiser; it was to these

that his own bent mo-t attracted him. Hi- range of information on

them was remarkably wide. I hie. again, his writings L'ive but I
mentary indication of the extent of bis attainments. He was familiar
with the financial history and fiscal experiences of England and Prance
quit ".h as with those of the United States, to which his writ-

chiefly devoted. And not only was he familiar with the
fads; he was singularly skilful in interpreting them. All who had

•c of following his courses of instruction in the University

CHARLES FRANKLIN DUNBAR. 573

fouud before them the conclusions of a sagacious mind, furnished with

ample information on every essential aspect of the situation. It was
Professor Dunbar's undeviating habit to turn to the primary sources of
information, — to the statutes, the official documents, the contem-
porary sources of knowledge. He was never content with information
at second hand, and was frequently able to point out how the conclusions
of authors of repute were overthrown by a careful comparison with the
sources from which their conclusions should have been derived.

Professor Dunbar's writing?, as already intimated, were compara-
tively scanty; they were certainly scanty as compared with what he
was equipped to do. His little book on Banking, brief and unpreten-
tious, is a model, and indeed well-nigh a classic, in its Held. His
essays on the financial history of the United States, published in the
Quarterly Journal of Economies, are also models of their kind. Occa-
sional comparisons, undertaken in these essays, with the financial
experience of other countries, — as, for example, in the essay on Some
Precedents followed by Alexander Hamilton, — give indication of the
wide range of his researches. Similar evidence appears in essays on
some recondite and little understood phases of economic history, such
as those on the Dank of Amsterdam, on the Bank of Venice, and on
Early Banking Schemes. Those who had the privilege of Professor
Dunbar's intimate acquaintance knew that he had pushed his way into
other obscure and difficult places also, lie had given much attention
to the history of the Assignats in France, and to the peculiarities in the
course of depreciation during that remarkable episode in monetary
history. On this subject he had collected, as was Ids habit, a Btore of
contemporary material, and had planned at some time to present the
results of his researches in published form. He had undertaken a
minute and careful study of the financial administration of Alexander
Hamilton, of which the results appeared in print only to a very Blight
extent. He had followed with equal thoroughness the history of bank-
ing operations in the United States, especially from the middle of the
century to the present time ; but here also failing Btrength and an
untimely death prevented the execution of his matured plai

2s'o small part of Professor Dunbar's time and thought w e en in

the later years of his life to the Quarterly Journal of E mmics. That

Journal was established by Harvard University in 1886, Professor Dun-
bar being appointed its editor, and remaining in charge from 1886 to
L896. It was a very different editorial posl from that which he had
held on the Advertiser; but its duties were performed with no |i

574 CHARLES FRANKLIN DUNBAR.

fidelity and skill. From the first a ; indard was set. The Journal

was to In.- a medium of communication for investigators, and took rank
at once as one of the leading Bcholarly repertories on i ts subject. Space
in it v. .lit by eminent writers the world over, and publication in

. .- guarantee of a claim to the attention of the leai
Id. Pi Dunbar always looked back with ju-t satisfaction on

what he had here achieve I, and found in it sou;.' -..lace for his inability
to carry out hi- plans for independent publication.

Professor Dunbar was by nature reserved ; always dignified : in con-
versation, happy in the intuitive selection of the right word; guarded
in expressing an opinion, but sure to express a just one when his
conclusions had been reached. His writings reflected these qualil
They are distinguished by a rounded Btateliness of diction more Bought
for ration ago than in our own day; dignified, yet never sti '

flowing, yet never affected. No more ju-t and delightful tribute has
been paid to a man in hi> own lifetime than is contained in Professor
Dunbar's paper on President Eliot's Administration of Harvard Ohn
. published in the Harvard Graduates' Magazine (for dune. 1894)
of the twenty-fifth year of President Eliot's administration.
Equally sympathetic, and at the Bame time judicial and discriminating,
are his memoirs, in the Proceedings of this Academy, of three men of
very different types, — Henry ('. Carey, Francis A. Walker, and E. W.
( rurney.

It is a singular fact that Professor Dunbar wrote with hesitation, and
often had to nerve himself anew to the task of literary composition,
withstanding many years of experience in rapid writing, he shrank
from taking pen in hand; yet. when the first sentence was written, the
others followed apparently with ease, and certainly in logical sequ<
and with an immediate happy choice of phrase. The present writer has
been bo fortunate a- to examine some of the note-, memoranda, and
unfinished manuscript left by his lamented colleague : and in the briefest
and most fragmentary of these papers he has been repeatedly struck b\
the appositeness of the language, the instinctively systematic arrange-
instant proof of clear and well ordered thought.
In personal intercourse with those who enjoyed hi- more intimate
laintance, Professor Dunbar's habitual dignity and reserve wen' often
broken by flashes of humor. He enjoyed keenly a good story, and
the mirthful side of every subject. Often in solemn meetings the
nkle of ! perceptible only to those who knew him well, Bhoi

ition of the oddities and idiosyncra ies of his contemporai

CHARLES FRANKLIN DUNBAR. 575

As is common with men whose sense of humor is strong, his affection
was deep and lasting; and in the domestic circle the devotion which lie
gave and received was touching. His character, not less than his abilil
and attainments, won from associates in varied walks of life an universal
feeling of esteem and admiration.

F. W. Taussig.

There have been no resignations during the year, but one
Resident Fellow has abandoned his fellowship. One Resident

Fellow, having removed from Massachusetts, has been elected
to Associate Fellowship.

New members elected during the year are : Resident Fellows,
7;, Associate Fellows, 9; Foreign Honorary Members, 9.

The roll of the Academy now includes 197 Resident Follows,
96 Associate Fellows, and 70 Foreign Honorary Members.*

* By election, May 8, 1901, the roll is 198, 98, 74.

American Academy of Arts and Sciences.

OFFICERS AND COMMITTEES FOR igco-igoi.

PRESIDENT.
Al.l XANDER AGASSIZ.

Class I

John Trowbrii

VICE PRESIDENT.
Class II.
A PHI II',

CORRESPONDING SECRETARY.

Wii i jam M. Davis.

RECORDING SECRETARY.
Wii i.i am W \ i son.

treasurer.

Fk •■■■■ Blake.

librarian.

A. L.AWRENI i ROTCH.

Class III.
James B. Tii wii;.

Class I.
Ill m:v TA1

Theodore W. Richards, John E. Wi ilff,
Haj i -. M. Goodwin. i i i. Parker

COUNCILLORS.

Class II. Class III.

William T. Councilman, James B. Amis,

William l \ i rett,

A. Law I 'iii

COMMITTEE OF FINANCE.

Ai r Agassiz, Francis Blake,

Ei roT C. Clarke.

RUMFORD COMMITTEE.

D. I i win, Edward C. Pickering, Charles R. Cross,

A.Mu- I DOLBEAR, ARTHUR G. WEBSTER, THEODORE W. RICHARDS,

Thomas C. Mendenhall.

C. M. WARREN COMMITTEE.

Charli i Jackson, imuel C vbot, Henry B. Hill,

i ■.!!■ P. Kinnicutt, Arthur M. Comey, Roberi ii. Richards,

1 1 1 nicy P. Talbot.

COMMITTEE OF PUBLICATION.

Samuel IF. Scudder, Seth C. Chandli i iwford if. Toy.

COMMITTEE ON THE LIBRARY.

A. I i Rotch, Henry W. Hayni Samuel Henshaw.

AUDITING COMMITTEE.

Ill '. ' , I 1 1 ' .

Wn ii wi L Rich vrds< »n

LIST

OF THE

FELLOWS AND FOREIGN HONORARY MEMBERS.

(Corrected to May 20, 1901.)

RESIDENT FELLOWS. — 198.

(Number limited to two hun<lre<].)

Class I. — Mathematical and Physical Sciences. — 80.
Section I. — 21. Section II. — 21.

Mathematics and Astronomy.

Solon I. Bailey, Cambridge.

Maxime Bocher, Cambridge.

William E. Byerly, Cambridge.

Seth C Chandler, Cambridge.

A. W. Duff, Worcester.

Gustavus Hay, Boston.

Theodore Lyman, Brookline.

Henry Mitchell, Nantucket.

William F. Osgood, Cambridge.

James Mills Peirce, Cambridge.

Edward C. Pickering, Cambridge.

"William II. Pickering, Cambridge.

John Ritchie. Jr., Boston.

John I). Rankle, Cambridge.
T. II. Safford, Williamstown.

Edwin F. Sawyer, Brighton.

Arthur Searle, Cambridge.

William E. Story, Worcester.

Henry Taber, Worcester.

O. C Wendell, Cambridge.

P. S. Yendell, Dorchester.
vol. xxxvi — 37

Physics.

A. Graham Bell, Washington. D.C.

Clarence J. Blake, Boston.

Francis Blake, Weston.

Charles R. Cross, Brookline.

Amos E. Dolbear, Somerville.

II. M. Goodwin, Roxbury.

Edwin II. Hall, Camhridgo.
Hammond V. Hayes, Cambridge.

William L. Hooper, Somerville.

William W. Jacques, Newton.

Frank A. Laws, Boston.

Henry Lefavour, Williamstown.

T. C. Mendenhall, Worcester.

Benjamin O. Peirce, Cambridge.

A. Lawrence Rotcb, Boston.

Wallace C. Sabine, Cambridge.

John S. Stone, Boston.

Elihu Thomson, Swampscott.

John Trowbridge, Cambridge.

A. (;. Webs! Worcester.

Robert W. Willson, Cambridge.

578

DENT FELLOWS.

Hon HT. — 22.
Chemistry.

Samuel Cabot, Boston.

Arthur M. Comey, Cambrii
James M. Cral Boston.

Charles W. Eliot, Cambric1
II ;,]> B. Hill. Cambridge.

Charles L. Jackson, Cambridf
Wall i L. Jennings, Worcester.
L( mard P. Kinnieutt. Worcester.
Charles F. Mabery, Cleveland, O.
Arthur .Michael. Boston.

I). Moore, Worcester.
i rlesE Munroe, Wash'gton, D.C
John l'. Nef, Chicago, 111.

Arthur A. Noyes, ion.

1; i it II. Richards, Bo on.
Theodore W. Richards, Cambridge.
Charles R. Sanger, Cambridge.
Stephen P. Sharpies, Cambridge.
Fram-is II. Storer, Boston.

Henry P. Talbot.
Charles II. Wing,
Edward S. Wood,

Newton.
Ledger, N.C.
Boston.

Section IV. — 16.
Technology and Engineering.
Boston.

Cambridge.
Cambridge.
Boston.
Cambrii

Eliot C. Clarke,
Ira N. IL.llis,
L. M. Johnson,
Gaetano Lanza,
E. D. Leavitt,
William 1!. Livermore, Boston.

Hiram F. Mills, Lowell.

Ceil II. Peabodv. Boston.

Alfred P. Rockwell, Manchester.

Andrew II. Russell, Wash'ton, D.C.

Peter Schwamb, Arlington.

II. L. Smyth, < Cambridge.

Charles S. Storrow, Boston.

Qeorge F. Swain, Boston.

William Watson, Boston.

Morrill Wyman, Cambridge.

Class II. — Natural and Physiological Sciences
Section I. — 13.

ogy, Mineralogy, ami Physics of
the Globe.

II. II. Clayton, Milton.

Algernon Coolidge, Boston.

William ( ». Crosby, Boston.

William M. Davis, Cambridge.

Benj. K. Emere Amher t.

(). W. Huntington, Newport, It. L

Roberl T. Jackson, Cambridge.

William II. Ni! . Cambridge.
John E. Pillsbury, Bo ton.

Nathaniel S. Shaler, Cambrid

ri DeC. Ward, Cambrid

John E. Wolff, Cambrid

.1. B. Woodworth, Cambrid

Section II
Botany.
F. S. Collins,
Geo. F. 1 davenport,
William G. Fallow,
Charles E. Faxon,
Merritl L. Fernald,
George L. Goodale,
II. II. Hunnewell,
John G. Jack,
15. L. Robinson,
Charles S. Sargent,
Arthur B. Sej mour,
Roland Thaxter,

Si i i ion HI.

ZoOlogy and Ph

Alexander A ja i/..
it Amory,

— G4.
— 12.

Maiden.

Medford.

Cambridge.

Bosti 'ii.

Cambridge.

Cambridge.

WeUesley.

Boston.

Cambridge.

Brookline.

Cambridge.

Cambridge.

—24.
Cambridge.

Boston.

RESIDENT FELLOWS.

579

James M. Barnard,

Henry P. Bowditch,

William Brewster,

Louis Cabot,

William E. Castle,

Samuel F. Clarke,

W. T. Councilman,

Charles B. Davenport. Chicago, 111

Harold C. Ernst, Boston.

Edward G. Gardiner, Boston.

Milton.
Boston.
Cambridge.
Brookliue.
Cambridge.
Williamstown.
Boston.

Samuel Henshaw,
Alpheus Hyatt,
John S. Kingsley,
Edward L. Mark,
Charles S. ilinot,
Edward S. Morse,
George H. Parker,
James J. Putnam.
Samuel H. Scudder

Cambridge.

Cambridge.

Somerville.

Cambridge.

Boston.

Salem.

Cambridge.

Boston.

Cambridge.

William T. Sedgwick, Boston.

James C. White, Boston.
William M. Woodworth, Cambi

Section IV. — 15.

.1/. dicine and Surgt ry.

Samuel L. Abbot, ton.

Edward II. Bradford, Boston.

Arthur T. Cabot, Boston.

David W. < heever, Boston.

Frank W. Draper, Boston.

Thomas Dwight, bon.

Reginald II. Fitz, Boston.

Charles F. Folsoin, Boston.

Frederick I. Knight, Boston.

Samuel J. Mixter, Boston.

W. L. Richardson,

Theobald Smith,

O. F. Wadsworth,

Henry P. Walcott

John C. Warren

Class III. — Moral and Political Sciences.— ■> 1

1

CCS. —

Boston.
Boston.
ton.
Cambridge.
Boston.

Section I.

Philosophy and Jui

James B. Ames,
Horace Gray,
John C. (J ray,
G. Stanley Hall,
Geo. F. Hoar.
Francis C. Lowell,
Josiah Royce,
Jeremiah Smith,
James B. Thayer,

isprudence.

Cambridge.

Boston.

Boston.

Worcester.

Worcester.

Boston.

Cambridge.

( lambridge.

Cambridge.

Section II. — 21.

Philology and Archaeology.

William S. Appleton, Boston.
Charles P. Bowditch, Boston.

J. W. Fewkes,
William W. Goodwin,
Henry W. Haynes,
Charles It. Laninan,
David (•. Lyon,
Bennett II. Nash,
Frederick W. Putnam,
Edward Robinson,
F. B. Stephenson,
Joseph II. Thayer,
Crawford II. Toy,
John W. White,
John II. Wright,
Edward J. Young,

Washington.
( 'ambridge.
! in.
( !ambridge.
Cambridge.
Boston.
Cambridge.
Boston,
bon.
Cambridge.
I

( 'ambridge.
Cambridge.
Waltham.

Lucien Carr,'
Franklin Carter,
Joseph T. Clarke,
Henry G. Denny,
William Everett,

Cambrii

Williamstown.
bon.

Boston.
Quincy.

Section III. — 12.
Political Economy and History.
Charles I-\ Adams, Lincoln.
Edward Atkinson, B ton.
Andrew M. Davis,
Ephraim Emerton, * tnbridge.
John Fi < 'aml.i i

580

RESIDENT FELLOWS.

A. ('. Goodell,

Henry C. Lodge,
A. Lawrence Lowell.
James F. Rhodi
I >. ii u inn \V. Ross,
diaries C. Smith,
F. W. Taussig,

Section IV.
Literature and (he
Francis Bartlett,

Salem.

John Bartlett,

Cambridge

Nahant.

Ail. B

Boston.

Boston.

S. Boutwell,

Groton.

1 in.

J. Llliut Cabot,

Brookline.

Cambrid 3

T. W. Higginson,

Cambridge

Boston.

George L. Kittredge,

Cambridge

Cambridge.

Charles G. Loring,

Boston.

— 12.

Percival Low. -11,

Boston.

Charles Lliot N

Cambridge

. 1 lis.

Horace L. Scudder,

Cambridge

Boston.

Barrett Wendell,

Boston.

ASSOCIATE FELLOWS.

581

ASSOCIATE FELLOWS.-98.

(Number limited to one hundred. Elected as vacancies occur.)

Class I. — Mathematical and Physical Sciences. — 37.

Section I. — 1-4.

Mathematics and Astronomy.

Edward E. Barnard, Williams Bay,
S. W. Burnham, Chicago. ['Wis.
George Davidson, San Francisco.
Fabian Franklin, Baltimore.
Asaph Hall, Cambridge, Mass.

George W. Hill, W. Nyaek, X.Y.
E. S. Holden, Washington.

Emory McClintock, Monistown.N.J.
E. H. Moore, Chicago.

Simon Xewcomb, Washington.
Charles L. Poor,
George M. Searle,
.1. X. Stock well,
Chas. A. Young,

Baltimore.
Washington.
Cleveland, O.
Princeton, N. J.

Section II. — 7.
Physics.

Carl Barns,

J. Willard Gibbs,
G. E. Hale,
S. P. Langley,
A. A. Michelson,

Providence, R.I.
New Haven.
William- Bay.
Washington.
Chicago.

Ogden N. Rood,
E. L. .Nichols,

New York.
Ithaca.

Section III. — 8.

Chemistry.

T. M. Drown, So. Bethlehem, Pa.
Wolcott Gibbs, Newport, R.I.
Frank A. Gooch, New Haven.
S. W. Johnson, New Haven.

Charlottesville, Va.

Cleveland, O.

New Orleans.

Baltimore.

J. W. Mallet,
E. W. Morley,
J. M. Ordway,
Ira Remsen,

Section IV.— 8.
Technology and Engineering.

Henry L. Abbot, New York.
Cyrus B. Comstock, New York.[Va.
W. P. Craighill, Charlestown, W.
John Frit/., Bethlehem, Pa.

F. R. Ilutton, New York.

George S. Morison, Chicago.
William Sellers, IM^e moor, Del.
Robt. S. Woodward, New York.

Class II. — Natural and Physiological Sciences. — 31.

Section I. — 14.

Geology, Mineralogy, and Physics of
the Globe.

Cleveland Abbe, Washington.

George J. Brush, New I raven.

T. C. Chamberlin, Chicago.

Edward S. Dana, New Haven.

Walter G. Davis, Cordova, Arg.

G. K. Gilbert,
Clarence King,

Jos 'I'll Le( lonte,

J. Peter Let

S. L. Penfield,

J W. Powell.
K. 1'nii
A. K. C. Sel
Charles D. Wall

Washington.
New York.
Berkeley, < !al
Milton, Mass.
New Haven.
Washington.
New porl , ELL
Vancouver.
Washington.

582

ASSOCIATE FELLOWS.

Section II.
Botany.

L. II Bailey,
J). II. Campbell,
J. M. Coulter,
C. G. Pringle,
John D. Smith,
\V. Treleac

Ithaca.

Palo Alto, Cal.

Chicago.

Charlotte, Vt.

Baltimore.

St. Louis.

Si CTIOH III.— 8.

Zoology and Physiology.
Joel A. Allen, New York.

W. K. Brooks, Lake Roland, Md.
F. P. Mall, Baltimore.

S. Weir Mitchell, Philadelphia.

II. F. Osborn,

A. S. Packard,
A. E. Yerrill,
C. O. Whitman,

New York.
Providence, R.I.

New Haven.
Chicago.

Section IY.— 6.

Mi 'Heine and Surgery.

John S. Billi: New York.

W. S. Halsted, Baltimore.

W. W. Keen, Philadelphia.

William Osier, Baltimore.

Win. H. Welch, Baltimore.

II. C. Wood, Philadelphia.

Class III. — Moral and Political Silences. — 27.

Section I. — 6.

Philosophy and Jurisprudence.
.James C. Carter, New York,
ph II. Choate, New York.
Melville W. Fuller. Washington.
William W. Howe, New Orleans.
Charles S. Peirce, IMilford, Pa.
T. R. Pynchon, Hartford, Conn.

Section II. — 7.

Philology and Archaeology.

Timothy Dwight, New Haven.
B. L. Gildersleeve, Baltimore.
I). C. Gilman, Baltimore.

T It. Lounsbury, New Haven.
Rufus B. Richardson, Athens.
Thomas D. Seymour, New Haven.
A. D. White, Ithaca, N.Y.

Section III. — G.
Political Economy and History.

Henry Adams,
G. P. Fisher.
H. E. von II< ilst.
Henry C. Lea.
Henry M. Stevens.
W. G. Sumner,

Washington.
Nevi Ha
( Ihicago.

Philadelphia.
Ithaca.
New Haven.

Section IV. — 8.
Literature and the Fine Arts.

James B. Angell, Ann Arbor, Mich.
L. 1'. di (Vsimla, New York.
II. II. Furness, Wallingford, Pa.
It. S. Greenongh, Florence.
Augustus St. Gaudens, New York.
John S. Sargent, London.
E. C. Stedman, Bri mxville, N'.
W. It. Ware, New York.

FOREIGN HONORARY MEMBERS.

583

FOREIGN HONORARY MEMBERS. — 74.

(Number limited to seventy-five. Elected as vacancies occur.)

Class I. — Mathematical and Physical Sciences. — 24.

Section I. — 7.

Mathematics and Astronomy.

Arthur Auwers, Berlin.

George H. Darwin, Cambridge.

H. A. E. A. Faye, Paris.

Sir William Huggins, London.

H. Poincare, Paris.

Otto Struve, Karlsruhe.

II. C. Vogel, Potsdam.

Section II. — 6.

Physics.

Ludwig Boltzmann, Vienna.
A. Cornu, Paris.

Oliver Heaviside, Newton Abbot.
F. Kohlrausch, Berlin.

Lord Rayleigh, Witham.

Sir G. G. Stokes, Bart., Cambridge.

Section III. — 6.

Chemistry.

Adolf Baeyer, Munich.

Marcellin Berthelot, Paris.
J. II. van't Iloff, Berlin.

D. Mendeleeff, St. Petersburg.

Sir II. E. Roscoe, London.

Julius Thomseu, Copenhagen.

Section IV. — 5.

Technology and Engine* ring.

Sir Benjamin Baker, London.

Lord Kelvin, Largs.

Maurice \/-\\\ Paris.

II. Miiller-Breslau, Berlin.

William C. Unwin, London.

Class II. — Natural and Physiological Sciences. — 27.

Section I. — 7.

Geology, Mineralogy, and Physics of

the Globe.
Sir Archibald Geikie, London.
Albert Heim, Zurich.

Sir John Murray, Edinburgh.
A. E. Nordenskiold, Stockholm.
Freih. v. Richthofen, Berlin.
Henry C. Sorby, Sheffield.

Heinrich Wild, Zurich.

Section II. — G.

Bold a y.

E. Bornet, Paris.

A. Engler, Berlin.
Sir Joseph D. Hooker, Sanningdale.

W. Pfeffer, Leipsic.
H. Graf zu Solms-

Laubach, Strassburg

Eduard Strasburger, Bonn.

-1

FOREIGN* HONORARY MEMBERS.

Section III.— 8.

Zoology and Physiology.

Sir Michael Foster, I ambridge.

Carl Gegenbauer,

Ludimar Hermann,

A. vmi Kiillikrr.

A. Kovalevsky,

II. Kronecker,

II. de Lacaze-Duthiera

Eliaa Metschnikoff,

1 1. idelberg.

Kbnigsberg.
Wiirzbuxg.
St. Petersburg.
Bern.
Paris.
Paris.

Si CTION IV. — 6.

."/< 'Heine and Surgery.

Sir T. L. B run ton, London.
A. Colli, Home.

R. Koch, Berlin.

Lord Lister, London.

F. v. Recklinghausen, Strassburg.
Rudolph Yirchow, Berlin.

CLASS III. — Mind mid Political Sciences. — 23.

Section I. — 5.

Philosophy and Jurisprudence

Hi ■inrii-h Brnnner, B rlin.

A. V. Dicey,

\V. E. Hearn,

F. W. Maitland,

Sir Frederick Pollock,

Bart., London

Oxford.

Melbourne.

Cambridge.

Section II. — 7.

Philology and Archoeology.

Ingram Bywater, Oxford.

W. Dorpfeld, Athens.
Sir.Tolni Evans, Hemel Hempstead.

II. Jack-on. Cambridge.

J. W. A. Kirchhoff, Berlin.

6. C. C. Maspero, Paris.

Karl Weinhold, Berlin.

Section III. — 4

Political Economy and History.

James Bryee. London.

Herman Grimm, Berlin.

Theodor Mommsen, Berlin.
Sir G. O. Trevelyan,

Bart., London.

Section IV. — 7.
alure and the Fine Arts.

E. de Amicis,
Georg Brandes,

F. Brunetiere,
Jean Leon Ge'rdme,
Rudyard Kipling,

G. Paris
Leslie Stephen,

Florence.
Copenhagen.

Paris.

Par

Rottingdean.

Paris.
London.

STATUTES AND STANDING VOTES.

STATUTES.

Adopted May 30, 1854: amended September 8, 1857, November 12, 1862, I
24, 1864, November 9, 1870, May 27, 187:'.. January 26, 1876, Sum 16,
1886, October 8, 1890, January 11 and A/a</ 10, 1893, May 9 <>W (A7.-W
10, 189-1, J/^-c/j 13, April 10 and May S, 1895, a/id May 8, 1901.

CHAPTER I.
Of Fellows and Foreign Honorary Members.

1. The Academy consists of Resident Fellows, Associate Fellows and
Foreign Honorary Members. They are arranged in three Classes,
according to the Arts and Sciences in which they are severally proficient,
viz.: Class I. The Mathematical and Physical Sciences; — Class II.
The Natural and Physiological Sciences; — Class III. The Moral and
Political Sciences. Each Class is divided into four Sections, viz. :
Class L, Section 1. Mathematics and Astronomy; — Section 2. Physics;
— Section 3. Chemistry; — Section 4. Technology and Engineering.
Class II., Section 1. Geology, Mineralogy, and Physics of the Globe ; —
Section 2. Botany; — Section 3. Zoology and Physiology; — Section I.
Medicine and Surgery. Class III., Section 1. Philosophy and Juris-
prudence : — Section 2. Philology and Archaeology ; - Section '■'<■
Political Economy and History; — Section 4. Literature and the Fine
Arts.

2. The number of Resident Fellows shall not exceed two hundred.
Only residents in the Commonwealth of Massachusetts shall be eligible
to election as Resident Fellows, but resident fellowship may be retained
after removal from the Commonwealth. Each Resident Fellow shall
pay an admission fee of ten dollars and such annua! assessment, nol ex-
ceeding ten dollars, as shall be voted by the Academy at each annual

586 STATUTES OF THE AMERICAN ACADEMY

meeting. Resident Fellows only may vote at the meetings of the
Academy.

3. The number of Associate Fellows shall not exceed one hundred,
of whom there .shall not be more; than forty in either of the three classes
of the Academy. Associate Fellows shall he chosen from persons resid-
ing outside of the Commonwealth of Massachusetts. They shall not be
liable to the payment of any fees or annual dues, but on removing within
the Commonwealth they may be transferred by the Co mcil to resident
fellowship as vacancies there occur.

4. The number of Foreign Honorary Members shall not exceed
seventy-five; and they shall he chosen from among persons most eminent
in foreign countries for their discoveries and attainments in either of the
three departments of knowledge above enumerated. There shall not be
more than thirty Foreign Members in either of these departments.

CHAPTER II.
Of Officers.

1. There shall be a President, three Vice-Presidents, one for each
Class, a Corresponding Secretary, a Recording Secretary, a Treasurer,
and a Librarian, which officers shall be annually elected, by ballot, at
the Annual M seting, on the second Wednesday in May.

2. At the Annual Meeting of 1901, nine Councillors shall be elected
by ballot, one from each Class of the Academy to serve for one year,
one from each Class for two ye irs, and one from each Class for three
years; and at annual meetings thereafter three Councillors shall he
elected in the sam • manner, one from each Class, to serve for three
years ; but the Bame Fellow shall not be eligible for two successive terms.
The nine Councillors, with the President, the three Vice-Presidents,
the two Secretaries, the Treasurer, and the Librarian, shall constitute the
Council. Five members shall constitute a quorum. It shall be the
duty of this Council to exercise a discreet supervision over all nomina-
tions and elections. With the consent of the Fellow interested, they
shall have power to make transfers between the several Sections of the
same Class, reporting their action to the Academy.

3. If any office shall become vacant during the year, the vacancy
shall be filled by a new election, and al the next stated meeting, or at a
meeting called for this purpo

OF ARTS AND SCIENCES. 587

CHAPTER III.

Of Nominations of Officers.

1. At the stated meeting in March, the President shall appoint from
the next retiring Councillors a Nominating Committee of throe Fellows,
one for each class.

2. It shall be the duty of this Nominating Committee to prepare a
list of candidates for the offices of President, Vice-Presidents, Corre-
sponding Secretary, Recording Secretary, Treasurer, Librarian, Coun-
cillors, and the Standing Committees which are chosen by ballot ; and
to cause this list to be sent by mail to all the Resident Fellows of the
Academy not later than four weeks before the Annual Meeting.

3. Independent nominations for any office, signed by at least five
Resident Fellows and received by the Recording Secretary not less than
ten days before the Annual Meeting, shall be inserted in the call for the
Annual Meeting, which shall then be issued not later than one week
before that meeting.

4. The Recording Secretary shall prepare for use, in voting at the
Annual Meeting, a ballot containing the names of all persons nominated
for office under the conditions given above.

5. When an office is to be filled at any other time than at the Annual
Meeting, the President shall appoint a Nominating Committee in accord-
ance with the provisions of Section 1, which shall announce its nomina-
tion in the manner prescribed in Section 2 at least two weeks before
the time of election. Independent nominations, signed by al least five
Resident Fellows and received by the Recording Secretary not later
than one week before the meeting for election, shall be inserted in the
call for that meeting.

•.-•

CHAPTER IV.

Of the President.

1. It shall be the duty of the President, and, in his absence, of the
senior Vice-President present, or next officer in order as above enumer-
ated, to preside at the meetings of the Academy; to summon extraor-
dinary meetings, upon any urgent occasion ; and to execute or Bee to
the execution of the Statutes of the Academy. Length of continuoua
membership in the Academy shall determine the seniority of the Vice-
Presidents.

588 STATUTES OF THE AMERICAN ACADEMY

2. The President, or, in his absence, the next officer as above enumer-
ated, is empowered to draw upon the Treasurer for such suras of money
as the Academy shall direct. Bills presented on account of the Library,
or the Publications of the Academy, must be previously approved by the
respective committees on these departments.

3. The President, or, in his absence, the next officer as above enumer-
ated, shall nominate members to serve on the different committees of the
Academy which are not chosen by ballot.

4. Any deed or writing to which the common seal is to be affixed
shall be signed and sealed by the President, when thereto authorized
by the Academy.

CHAPTER V.

Of Standing Committees.

1. At the Annual Meeting there shall be chosen the following Stand-
ing Committees, to serve for the year ensuing, viz. : —

2. The Committee of Finance, to consist of the President, Treasurer,
and one Fellow chosen by ballot, who shall have charge of the invest-
ment and management of the funds and trusts of the Academy. The
general appropriations for the expenditures of the Academy shall be
moved by this Committee at the Annual Meeting, and all special appro-
priations from the general and publication funds shall be referred to or
proposed by this Committee.

3. The Rumford Committee, of seven Fellows, to be chosen by ballot,
who shall consider and report on all applications and claims for the
Rumford Premium, also on all appropriations from the income of the
Rumford Fund, and generally see to the due and proper execution of
this trust.

4. The C. M. Warren Committee, of seven Fellows, to be chosen by
ballot, who shall consider and report on all applications for appropria-
tions from the income of the C. M. Warren Fund, and generally see to
the due and proper execution of this trust.

5. The Committee of Publication, of three Fellows, one from each
Class, t<> whom all communications submitted to the Academy for publi-
cation shall be referred, and to whom the printing of the Memoirs and
the Proceedings shall be intrusted.

C. Tin' Committee on the Library, of the Librarian ex officio and
three oilier Fellows, one from each class, who shall examine the Library,
and make an annual report on its condition and management.

OF ARTS AND SCIENCES. 589

7. An Auditing Committee of two Fellows, for auditing the accounts
of the Treasurer.

CHAPTER VI.

Of the Secretaries.

1. The Corresponding Secretary shall conduct the correspondence of
the Academy, recording or making an entry of all letters written in its
name, and preserving on file all letters which are received ; and at each
meeting he shall present the letters which have been addressed to the
Academy since the last meeting. Under the direction of the Council
for Nomination, he shall keep a list of the Resident Fellows, Associate
Fellows, and Foreign Honorary Members, arranged in their Classes and
in Sections in respect to the special sciences in which they are severally
proficient ; and he shall act as secretary to the Council.

2. The Recording Secretary shall have charge of the Charter and
Statute-book, journals, and all literary papers belonging to the Academy,
lie shall record the proceedings of the Academy at its meetings ; and
after each meeting is duly opened, he shall read the record of the pre-
ceding meeting. He shall notify the meetings of the Academy, apprise
officers and committees of their election or appointment, and inform the
Treasurer of appropriations of money voted by the Academy. He shall
post up in the Hall a list of the persons nominated for election into the
Academy ; and when any individual is chosen, he shall insert in the
record the names of the Fellows by whom he was nominated.

3. The two Secretaries, with the Chairman of the Committee of
Publication, shall have authority to publish such of the records of the
meetings of the Academy as may seem to them calculated to promote
its interests.

CHAPTER VII.
Of the Treasurer.

1. The Treasurer shall give such security for the trust reposed in
him as the Academy shall recpjire.

2. He shall receive officially all moneys due or payable, and all
bequests or donations made to the Academy, and shall pay such sums
as the Academy may direct. He shall keep an account of all receipts
and expenditures ; shall submit his accounts to the Auditing Committee;
and shall report the same at the expiration of his term of office.

590 STATUTES OF THE AMERICAN ACADEMY

3. The Treasurer shall keep separate accounts of the iucome and
appropriation of the Rumford Fund and of other special funds, and
report the same annually.

4. All moneys which there shall not be present occasion to expend
-hall be invested by the Treasurer, under the direction of the Finance
Committee,

CHAPTER VIII.
Of the Librarian and Library.

1. It .-hall be the duty of the Librarian to take charge of the books,
to keep a correct catalogue of them, to provide? for the delivery of books
from the Library, and to appoint such agents for these purposes as he
may think necessary. He shall make an annual report on the condition
of the Library.

2. The Librarian, in conjunction with the Committee on the Library,
shall have authority to expend such sums as may be appropriated, either
from the General, Rumford or other special Funds of the Academy, for
the purchase of book-', and for defraying other necessary expenses con-
nected with the Library.

.".. To all books in the Library procured from the income of the
Rumford Fund, or other special funds, the Librarian shall can-" a stamp
or label to be affixed, expressing the fact that they were so pro* ured.

1. Every person who takes a hook from the Library shall give a
receipt for the same to the Librarian or his assistant.

•"». Every book shall be returned in good order, regard being had to
the necessary wear of the book with good usage. If any book -hall
be lost or injured, the person to whom it stand- charged shall replace
it b\ a new volume or set, if it belongs to a set. or pay the current
price of the volume or ael to the Librarian ; and thereupon the remain-
der of the set. if the volume belonged to a set. shall be delivered to the
person so paying for the same.

6. All books -hall be returned to the Library for examination at
lea-; one week before the Annual Meeting.

7. The Librarian -hall have custody of the Publications of the

lemy anil shall distribute copies among the Associate Fellows and
i'_'n Honorary Members, at their request. With the advice and con-
sent of the Prea 'lent, he may effect exchanges with other associations.

OF ARTS AND SCIENCES. 591

CHAPTER IX.
Of Meetings.

1. There shall be annually four stated meetings of the Academy ;
namely, on the second Wednesday in May (the Annual Meeting), on
the second Wednesday in October, on the second Wednesday in Jauuary,
and on the secoud Wednesday in March. At these meetings only, or at
meetings adjourned from these and regularly notified, shall appropria-
tions of money be made, or alterations of the statutes or standing votes
of the Academy be effected.

2. Fifteen Fellows shall constitute a quorum for the transaction of
business at a stated meeting. Seven Fellows shall be sufficient to con-
stitute a meeting for scientific communications and discussions.

3. The Recording Secretary shall notify the meetings of the Academy
to each Fellow residing in Boston and the vicinity; and he may cause
the meetings to be advertised, whenever he deems such further notice
to be needful.

CHAPTER X.

Of the Election of Fellows and Honorary Mem

1. Elections shall be made by ballot, and only at stated meetic

2. Candidates for election as Resident Fellows must be propos< d by
two Resident Fellows of the section to which the proposal is made in a
recommendation signed by tliem, and this recommendation shall be
transmitted to the Corresponding Secretary, and by him referred to
the Council for nomination. No person recommended shall be reported
by the Couucil as a candidate for flection, unless he shall have received
a written approval, signed at a meeting of the Council by at least five
of its members. All nominations thus approved shall he read to the
Academy at a stated meeting, ami shall then stand on the nomination
list during the interval between two Btated meetings, and until tin-
balloting. No person shall be elected a Resident Fellow, unless ho
shall have been resident in this Commonwealth one year next preceding
his election. If any person elected a Resident Fellow Bhall lor
one year to pay his admission fee, bis election shall be void j and
if any Resident Fellow shall neglect to pay his annual a ints

592 STATUTES OP THE AMERICAN ACADEMY

for two years, provided that his attention shall have been called to this
article, he shall be deemed to have abandoned his Fellowship ; but it
shall be in the power of the Treasurer, with the consent of the Council,
to dispense (sub silentio) with the payment both of the admission fee and
of the assessments, whenever in any special instance he shall think it
advisable so to do.

0. Tin; nomination of Associate Fellows may take place in the manner
prescribed in reference to Resident Fellows. The Council may in like
manner originate nominations of Associate Fellows, which must be read
at a stated meeting previous to the election, and be exposed on the nom-
ination list dining the interval.

■1. Foreign Honorary Members shall be chosen only after a nomina-
tion made at a meeting of the Council, signed at the time by at least
ii of its members, and read at a stated meeting previous to that on
which the' balloting takes place.

5. Three fourths of the ballots cast must be affirmative, and the
number of affirmative ballots must amount to eleven to effect an elec-
tion of Fellows or Foreign Honorary Members.

G. A majority of any section of the Academy is empowered to pre-
sent lists of persons deemed best qualified to fill vacancies occurring in
the number of Foreign Honorary Members or Associate Fellows allotted
to it ; and such lists, after being read at a stated meeting, shall be re-
ferred to the Council for Nomination.

7. If. in the opinion of a majority of the entire Council, any Fellow —
Resident or Associate — shall have rendered himself unworthy of a
place in the Academy, the Council shall recommend to the Academy
the termination of his Fellowship ; and provided that a majority of two
thirds of the Fellows at a stated meeting, consisting of not less than
fifty Fellows, shall adopt tins recommendation, his name shall be stricken
oil' the roll of Fellows.

CHAPTER XL

Of Amendments of the Statutes.

1. All proposed alterations of the Statutes or additions to them, shall

be referred to a committee, and. on their report at a subsequent meeting,

shall require for enactment a majority of two thirds of the members

present, and at least eighteen affirmative votes.

"_'. Standing VOteS may he passed, amended, or rescinded, at any

OF ARTS AND SCIENCES. 593

stated meeting, by a majority of two thirds of the members present.
They may be suspended by a unanimous vote.

CHAPTER XII.
Of Literary Performances.

1. The Academy will not express its judgment on literary or
scientific memoirs or performances submitted to it, or included in its
publications.

AOL. xxxvi. — 38

594 STATUTES OP THE AMERICAN ACADEMY

STANDING VOTES.

1. Communications of which notice had been given to the
Secretary shall take precedence of those not so notified.

2. Resident Fellows who have paid all fees and dues charge-
able to them arc entitled to receive one copy of each volume or
article printed by the Academy, on application to the Librarian
personally or by written order, within two years from the date
of publication. And the current issues of the Proceedings shall
be supplied, when ready for publication, free of charge, to all the
Fellows and members of the Academy who desire to receive them.

3. The Committee of Publication shall fix from time to time
the price at which the publications of the Academy may be sold.
Put members may be supplied at half this price with volu
which they are not entitled to receive free, and which arc needed
to complete their sets.

4. Two hundred extra copies of each paper accepted for publi-
cation in the Memoirs or Proceedings of the Academy shall be
placed at the disposal of the author, free of charge.

5. Resident Fellows may borrow and have out from the
Library six volumes at any one time, and may retain the same
for three months, and no longer.

G. Upon special application, and for adequate reasons assigned,
the Librarian may permit a larger number of volumes, not
ing twelve, to be drawn from the Library for a limited period.

7. Works published in numbers, when unbound, shall not be
taken from the Hall of the Academy, except by special leave of
the Librarian.

8. Books, publications, or apparatus shall be procured from the
income of the Rumford Fund only on the certificate of the Rum-
ford Committee that they, in their opinion, will best facilil
ami encourage the making of discoveries and improvements which
may merit the Rumford Premium.

'.' Tii- Annual Meeting and the other stated meetings shall be
holden at eight o'clock. 1'. M.

In. .\ meeting for receiving and discussing scientific commu-
nications may be held on the second Wednesday of each month

m.t appointed Med meeting . pting July, A and

Septeml

OF ARTS AND SCIENCES.

595

RUMFORD PREMIUM.

In conformity with the terms of the gift of Benjamin, Count
Rumford, granting a certain fund to the American Academy of
Arts and Sciences, and with a decree of the Supreme Judicial
Court for carrying into effect the general charitable intent and
purpose of Count Rumford, as expressed in his letter of gift, the
Academy is empowered to make from the income of said fund, as
it now exists, at any Annual Meeting, an award of a gold and
a silver medal, being together of the intrinsic value of three
hundred dollars, as a premium to the author of any important
discovery or useful improvement in light or in heat, which shall
have been made and published by printing, or in any way made
known to the public, in any part of the continent of America, or
any of the American islands; preference being always given to
such discoveries as shall, in the opinion of the Academy, tend
most to promote the good of mankind; and to add to such
medals, as a further premium for such discovery and improve-
ment, if the Academy see fit so to do, a sum of money nut
exceeding three hundred dollars.

INDEX.

Ageratum lucidum, 475.

rhytidophyllum, 476.

stachyofolium, 476.
Alcoceria, 493.

Pringlei, 493.
Alternating Current Waves, 319.
Archaeoplax signifera, 7.
Archibald, E. H. See Richards,

T. VV., and Archibald, E. H.
Arctostaphylos Conzattii, 497.

glabrata, 497.
Arthropod Trails, 61.
Arthrorhynchus Cyclopodiae, 407.

Eucampsipodae, 409.
Atomic Weights, 169; Table of, 545.

Barus, C, Award of Rumford

Premium to, 520, 529.
Behr, G. E. See Jackson, C. L., and

Behr, G. E.
Bigelowia Nelsonii, 505.
Biographical Notices, List of, 547.
John Harrison Blake, 565.
Charles Franklin Dunbar, 569.
Charles Carroll Everett, 549.
Silas Whiteomb Ilolman, 553.
Nathaniel Holmes, 552.
John Eldridge Hudson. 558.
Sylvester R. Koehler, 556.
Bisbee, H. See Richards, T. W..
McCaffrey, C. F.,and Risbee, II.
Bowers, M. A., Peripheral Distribu-
tion of the Cranial Nerves of
Spelerpes bilineatus. 177-193.
Brickellia amblyolepis, 485.
cardiophylla, 4b5.
glomerata, 50 1.
hebecarpoides, 486.
petrophila, 486.

petrophila, var. umbratilis, 4*7.
vernicosa, 487.

Calcic Oxalate, 375.
Cancer, Phylogeny of, 1.
Cancer proavitus, 3.

Calea Pringlei, 488.

Zacatechichi, 4SS.
Calhane, D. F. See Jackson, C. L.,

and Calhane, D. F.
California Petroleum, 255, 283.
Calochortua Pringlei, 472.
Ceraiomyces, 410.

Dahlii, 410.
Chandler, S. C, New Discovery

concerning the Motion of the

Earth's Pole, 538.
Chemical Laboratory of Case School

of Applied Science, Contribu-
tions from, 253.
Chemical Laboratory of Harvard

College, Contributions from,

73, 143, 195, 229, 331, 339,

375, 507.
Clayton, H. II., The Eclipse Cyclone

and the Diurnal Cyclones, 305

318.
Cohoe, W. P. See Jackson, C. L.,

and Cohoe, W. P.
Cologania Deamii, 492.
Committee of Publication, Report

of, 527.
Conyza lvrata, var. pilosa, 506.
Cordia Seleriana. 198.
Correspondence, 517, 534, 535, 538,

542.
Council, Report of, 517.
Crab, Fossil, 1.

Cranial Nerves of Spelerpes, 177.
Crew, EL, Report of Progress to

Rumford Committee, 521.
Croton Palmeri, var. ovalis, 493
Cryptogamic Laboratory of Harvard

Fniversity, Contributions from,

395.
Crystals, Study of Growing, 339.
Cyclones, 305.

Davis. W. M.. Geographical Notes on

Brittany and I Devonshire, 5
Design as a Science, ■■

»98

INDEX.

Diffusivitiea, Thermal, of Marbles, 11.
Dimeromyces coarctatus, 110.

crispatus, H3.

rhizophorus, 112.

Dio8Corea platycolpota, 171.
Diurnal Cyclones, 3U5.

Earle, R. B. Sa Jackson, C. L.,

and Earle, It. Ii.
Eclipse Cyclone, 305.
Electrical Conductivity of Iron, 119.
Eumelampodium, 456.
Euj)atoriuin anisopodum, 477.

araliaefolium, 477.

Bigelovii, 477.

conspicuum, 177.

Coulter i. 177.

dasycarpum, 17--.

dryojihilimi. 178.

Gonzalezii. 17!'.

hyssopinnm, 178.

Lemmoni, 179.

leonense, 179.

Liebniannii, lv".

longifolium, 480.

lucidum, 180.

Luxii, 480.

lvratum, 481.

Mariarum, 481.

pachypodum, 481.

pansamalense, 482.

pinabetense, 482.

pleianthum, 483.

prionobium, 483.

prionophyllnm, 184.

quadrangulare, 184.

viscidipes, 484.
Enpborbia calcicola, 496.

interazillaris, 195.

lancifolia, \ar. villicaulis, 496.

muscicola, 495.

potosina, 195.

puberula, 19 1
Everett, W., Tbe Life and Works of

the late Henry Sidgw ick, ■■
Evolvulus Seleriana, l!»^.

Felli • kte, deceased, —

Frederick Edwin Church, 517.
Jacob Maudes I >a< !osta, 53 1
i , orge Mercer I 'a^ son, 589.
James Edward Keeler, '>■'> 1.
William Mitchell, 584.
Henry August us Rowland, 549.
Edward Elbridge Salisbury, ->\2.

Fellows, Associate, deceased, —

Alfred Stille, 534.
Fellows. Associate, elected, —

George Mercer Dawson, 531.

Thomas Messenger Drown. 535.

Melville Weston Fuller, 532.

Charles Ellery Hale, 510.

William Stewart Ilalsted, 5 In.

William Williams Keen, 540.

Franklin Fain.- Mall, 540.

Eliakim Hastings Moon'. 539.

Edward Leamington Nichols, 540.

Henry Fairfield Osborn, 540.

Cyrus Guernsey Pringle, 540.

Rufus Byam Kichardson. 532.

Thomas Day Seymour, 532,

Henry Morse Stevens, 532.

Charles Otis Whitman. 5 In.
Fellows. Associate, List of , 581.
Fellows, Resident, deceased. —

Charles Carroll Everett, 539.

Thomas Gaffield, 537.

Nathaniel Holmes. 549.

John Elbridge Hudson. 534.

Sylvester R. Koehler, 531.

Augustus Lowell. 53 1.
Fellows. Resident, elected, —

William Ernest ( lastle, 531.

Frank Shipley Collins. 539.

Alexander Wilmer Duff. 539.

Ephraim Emerton, 539.

Merritt Lyndon Fernald. 531.

Lewis Jerome Johnson, 539.

Theodore Lvnian, 539.

Henry Lloyd Smyth, 539.

Frank William Taussig, 539.

.lav Backus Woodwortn, 531.
Fellows Resident, List of. 577.
Fernald, M. L., Some New Sper-
matophytes from Mexico and
Central America, 489-506.
Fimbristj lis alamosana, 491.
Holwayana, 192.
melanospora, 491.
obscura, 492.
Foreign Honorary Members, de-
ceased, —

Jacob < reorg Azarth, 542.

I >uc de Br »49

Charles Hermite, 538.

Willy K ii hue. 534.

MaxMiiller, 549.

Baron Russell of Killowen, 684,

1 lenry Sidgwick, 58 1.

\\ illiam Stubbs, 549.

INDEX.

599

Foreign Honorary Members,
elected, —

Edmondo de Amicis, 541.

Sir Thomas Lander Bruiiton, 540.

Albert Venn Dicey, 5-10.

Sir Archibald Geikie, 532.

William Edward Hearn, 540.

Henry Jackson, 5-11.

Robert Koch, 540.

Hugo Kronecker, 540.

Heinrich Miiller-Breslau, 540.

Sir John Murray, 532.

Jules Henri Poincare, 510.

William Cawthorn Unwin, 532.
Foreign Honorary Members, List

of, 583.
Fossil Crab from Gav Head, 1.
Fraprie, F. R. See Richards, T. W.,

and Fraprie, F. R.
Frost, E. B., Report of Progress to
Rumford Committee, 521.

Gelasimus pngilator, 415.

Geometry on Ruled Quartic Sur-
faces, 17.

Glasgow, University of. Ninth Jubi-
lee Celebration, 538, 542.

Glyphosperma Palmeri, 492.

Gray Herbarium of Harvard Uni-
versity, Contributions from,
453, 489.

Groups, Continuity of, 83.

Plale, G. E., Report of Progress to
Rumford Committee, 522.

Hall. E. II., On the Thermal and
Electrical Conductivity of Soft
Iron, 119-141; Report of Pro-
gress to Rumford Committee,
523; An Exposition of the
Theory of Electrons, the Elec-
trical Fragments of Atoms,
.-,11'.

Hardystonite, 111.

Harvard Mineralogical Museum,
Contributions from. 111.

Heat Capacity, Nomenclature of,
325.

Heterotoma Goldmanii, 504."
stenodonta, 50 I.

Hudson, E. J. Set Mabery, C. F.,
and Hudson, E. J.

Hydrocarbons in Petroleum, 2-

Infinitesimal Transformation . 33

International Atomic Weights, 169.

Ipomoea caudata, 108.
Iron, Conductivity of, 119.
Isopod Crustaceans, Tracks of, 67.

Jackson, C. L., and Behr, (i. K.,
Symmetrical Triiodbenzol, 331-
338.

Jackson, C. L,, and Calhane, ]). I'.,
The Uinitro Compounds of I'ara-
dibrombenzol, 533.

Jackson, C. L., and Cohoe, W. P.,
Certain Derivatives of Metadi-
bromdinitroben/ol. 73 82

Jackson, C. L., and Earle, R. B.,
On the Action of Sodic Sulphito
on Tribromdinitrobenzol, and
Tribromtrinitrobenzol, 229-238.

Jackson, C. L., and Koch. \\\,
On Certain Derivatives of Ortli-
obenzoquinone, 195-228.

Jackson, R. T., Resorption as a
Factor in Growth, 541.

Japanese Petroleum, 295.

Koch, W. See Jackson, C. L., and
Koch, W.

Laboulbeniaceae, 395.

Laws, F. A., An Apparatus for
recording Alternating Current
Waves, 310-321 ; Report of
Progress to Rumford Commit-
tee, 523.

Lewis, G. N., A New conception
of Thermal Pressure, and a
Theory of Solutions, 1 13 168;
The Law of Physico-chemical
( lhange, 512.

Librarian, Report of, 528.

Limulus polyphemus, 64; Trails
of, 61.

Lobelia graina. var. conferta. 503.
Nelsonii, 503.
regalia, 503.

Lowell. A., Bequest of, 534, 536.

Lyman. T., False Spectra From the
Rowland Concave Grating, 239-
252.

Mabery, C. F., Investigations on

the Composition of Petroleum,
253 804.
Mabery, C. F . and Hudson, E J.,
On tie- Composition of Cali-
fornia Petroleum, 25S

GOO

INDEX.

Mabery, C. F., and Sieplein, 0. J.,

On the Chlorine Derivatives of

the Hydrocarbons in California

Petroleum, 283 295.
Mabery, C. F.. and Takano, S., On

the Composition of Japanese

Petroleum, 295-304.
Magnesic Oxalate, 375.
Manganous Sulphate. 507.
Marbles, Thermal Diffusivities of,

11.
Mayer, 328.
McCaffrey, C. F. See Richards, T.

W., McCaffrey, C. F., and

Bisbee, H.
Melampodium, Synopsis of, 455.
Melampodium Achillaeoides, 105.

Americanum, 159, 466.

angustifolium, 461.

appendiculatum, 457.

appendiculatuin. var. leiocar-
pum, 457.

appendiculatum, var. sonorense,
457.

arenicola, 157.

arvense, 164.

australae, 165.

Baranguillae, 465.

Berterianum, 465.

bibracteatuin, 465.

brachyglossnm, 165.

camphoratum, 463.

cinercum, 458.

cinereum, var. argophyllum,
458.

cinereum, var. ramosissimum,
! 58.

copiosum, 465.

cornnopifolium, 465.

cupulatum, 463.

diffusum, 160.

diffusum, var. lanceolatum, 160.

digynnm, 465.

divaricatum, 465.

I tombeyanum, 166.

flaccidum, 164.

glabrum, 465.

graoile, 162.

heten phyllum, 160.

Hildalgoa, 166.

hirsutum, 166.

hispidum, 164.

humile, 166.

Knnthianum, 160.

lanceolatum, 166.

Melampodium leuoanthum, 458.

Liebmanmi, 466.

linearilobum, 459.

longicornu, 457.

longifolium, 465, 466.

longipes, 459.

longipilum, 458.

manillense, 466.

microcephalum, 462, 466.

mimulifolium, 462.

montanum, 463.

oblongifolium, 462.

ovatifolium, 166.

paludicola, 456, 466.

paludosum, 463.

panamense, 466.

paniculatum. 462, 466.

perfoliatum, 465.

Pringlei, 461.

pumiliini, 466.

ramosissimum, 466.

rhomboideum, 466.

Rosei, 461.

Kosei, var. subintegrum, 462.

ruderale, 166.

sericeum, 459, 466.

sericeum, var. brevipes, 166.

sericeum, var. exappendicula-
tum, 459.

sericeum, var. longipes, 466.

sinuatum, 461.

suffruticosum, 456.

tenellum, 463.

ternatum, 466.
Melczer, G. See Wolff. J. E.
Merostomatous Trails, 61.
Merostomichnites beecheri, 67.

narragansettenBis, 7<».
Metadibromdinitrobenzol, 73.
Meteorological Observations, '■'>"5-
Mimosa acanthocarpa, var. des-
manthocarpa, 172.

eurycarpoides, 17l'.

ionema, 173.

\\ atsoni, 473.
Montanoa arborescens, 487.
Monarda Pringlei, 501.

Nichols, E. L . Report of Progress to

Etumford Committee, 523.
N'oora, Synopsis of, 467.
Nbcca angusl ifolia, 470.

biflora, 168.

decipiens, 17".

glandnlosa, 470.

INDEX.

601

Nocca helianthifolia, 46S.

helianthifolia, var. levior, 46S.

helianthifolia, var. suaveolens,
468.

heteropappus, 470.

Liebmannii, 470.

Mocinniana, 4G9.

mollis, 471.

Palmeri, 471.

Pringlei, 469.

rigida, 469.

tomentosa, 470.
Nomenclature of Heat Capacity,

325.
Nominating Committee, 535.

Officers elected, 530, 538; List of,

576.
Orthobenzoquinone, 195.

Packard, A. S., A New Fossil Crab
from the Miocene Greensand
Bed of Gay Head, 1-9 ; On Sup-
posed Merostomatous and Other
Paleozoic Arthropod Trails, 61-
71.

Parathesis chiapensis, 497.

Pectis Lessingii, 506.

Peirce, B. O., Report of Progress to
Rumford Committee, 525.

Peirce B. O., and Wilson, R. W.,
On the Thermal Dili'usivities of
Different Kinds of Marble,
11-16.

Pernettya ovata, 496.

Petroleum, 253.

Physalis puberula, 502.

Physical Laboratory of the Massa-
chusetts Institute of Tech-
nology, Contributions from,
319.

Pickering, E. C, Report of Progress
to Rumford Committee, 52 i.

Piqueria pyramidalis, 475.

Plateanus glabrata, 493.

Quartic Surfaces, 17.

Records of Meetings, 517.
Reissner's Fibre, 443.
Rhizomyces gibbosus, 409.
Rhodes, J. F., Sherman's March to

the Sea, 536.
Richards, T. W., Internationa]

Atomic Weights, 169-176; Sug-

gestion concerning the Nomen-
clature of Heat Capacity, 325-
329 ; Report of Progress to
Rumford Committee 526; A
Table of Atomic Weights, 545-
516.

Richards, T. W., and Archibald,
K. II., A Study of Growing
Crystals by Instantaneous Mi-
crophotography, 339-353.

Richards, T. W., and Fraprie, F.
R., The Solubility of Manga-
nous Sulphate, 507-514.

Richards, T. W., McCaffrey, C. F.,
and Bisbee, II., The Occlusion
of Magnesic Oxalate by Calcic
Oxalate, and the Solubility of
Calcic Oxalate, 375-393.

Robinson, B. L., Synopsis of the
genus Melampodium, 455-466;
Synopis of the genus Nocca,
467-471 ; New Species and
newly noted Synonymy among
the Spermatophytes of Mexico
and Central America, 471-488 ;
Recent Advances in the General
Classification of the Flowering
Plants. 541.

Ross, U. W., Design as a Science,
355-374.

Rowland Concave Grating, 239.

Ruellia cupheoides, 502.

Rumford Committee, Report of, 519;
Reports of Progress to, 521.

Rumford Fund, Papers published by
Aid of, 11, 119, 239, 339.

Rumford Premium, 595 ; Award of,
520, 529.

Russelia Deamii, 471.
trachypleura, 474.

Sabine, W. C, Report of Progress
to Rumford Committee, 526;
The Influence of Architecture
on Melody and the Develop-
ment of the .Musical Scale,
537.

Salvia ageratifolia, 499.
albicans, 501.
Dugesii, 500.
igualensis, 500.
leucantha, for. iobaphes, 501.
-'•tidosa, 499.
Sessei, 501.
tiliaefolia, var. rhyacophila, 199.

G02

INDEX.

Sargent, P. E., The Development

and Function of Keissner's

Fibre, and its Cellular Connec-
tions, 1 li 452.
Schefferite, 111.
Sieplein, <). J. See Mabery, C. F.,

and Sieplein, 0. J.
Slocum, S E., On the Continuity of

Groups generated by Infinites-
imal Transformations, S3-109.
Sodic Sulphite, 229.
Solanum rostratum, var. subinteg-

rum, 502.
Solar Eclipse of May 28, 1900, 305.-
Solidago Pringlei, 505.
Solubility of Calcic Oxalate, 375; of

Manganous Sulphate, 507.
Solutions, Theory of, 1 13.
Spectra, False, from the Rowland

Concave Grating, 239.
Spelerpes bilineatus, 177.
Spermatophytes from Mexico and

Central America, 471, 489.
Statutes and Standing Votes, 585.
Stdgmatomyces constrictus, L01.
" Diopsis, 399.

dubius, 402.

gracilis, 403.

humilis, 101.

Ilydrelliao. 101.

Limnophorae, 100.

Limosinae, h'G.

Papuanus, 407.

proboscideus, 403.

purpureus, 104.

rugosus, 398.

Scaptomyzae, 400.

spiralis, 405.
Symmetrical Triiodbenzol, 331.

Takano. S. See Mabery, C F., and

Takano, S.
Thaxter, R., Preliminary Diagnoses

of New Species of Laboulbenia-

ceae, 395 ill.
Thermal Conductivity of Iron, 119.

Thermal Diffusivities of Marbles, 11.

Thermal Pressure, 143.

Trails, Arthropod, 01.

Transformations, Infinitesimal, 83.

Treasurer, Report of, 517.

Tribromdiuitrobenzol, 229.

Tribromtrinitrobenzoi, L'29.

Triiodbenzok, 331.

Trowbridge, J. Results obtained
with a Storage Battery of
Twenty Thousand Cells, 53l'.

Valeriana retrorsa, 502.
Variation in Fiddler Crab, 415.

Warren (C. M.) Committee, Report

of, 527.
Warren (C. M.)Fund, Aid from, 253.
Wendell, B., Literary History of

America, 535.
Williams, F. B., Geometry on Failed

Quartic Surfaces, 17-00.
Willson, R. W. See Peirce, B. O.,

and Wills, ,i], R. W.
Wolff, J. E„ On Hardystonite and

a Zinc Scheff erite from Franklin

Furnace, X. J.; with a note on

the Optical Constants of the

Schefferite by Dr. G. Melczer,

111-118.
Wright, J. H., Recent Excavations

in Crete', 535.

Xanthocephalum
505.

megalocephalum,

Yerkes, P.M., A Study of Variation
in the Fiddler (rah. GelasimUB
pugilator Latr. , 415-4 12.

Zarabellia, 4G0.

Zoological Laboratory of the Mu-
seum of Comparative Zoo
at Harvard College, Contribu-
tions from, 177, 415, 1 13.

New York Botanical Garden Librar

3 5185 00257 8944