\.

PROCEEDINGS

OF THK

A.MERICAN ACADEMY OF ARTS AND SCIENCES.

PROCEEDINGS

OF THE

AMERICAN ACADEMY

OF

ARTS AND SCIENCES.

NEW SERIES.
Vol. XVllI.

WHOLE SERIES.
Vol. XXVL

from may, 1890, to may, 1891.

SELECTED FROM THE RECORDS.

BOSTON:

UNIVERSITY PRESS: JOHN WILSON AND SON.

1891.

*3o)

^ ^^ ^

15 L S

CONTENTS.

Page
I. The Prehistoric and Kiowa County Pallasites. By Oliver

Whipple Huntington 1

II. Preliminary Notes on the Species of Doassansia, Cornu. By

William Albert Setchell 13

III. On some Theorems which connect together certain Line and

Surface Integrals. By B. O. Peirce 20

IV. The Quantitative Determination of Arsenic, hy the Berzelius-

Marsh Process, especially as applied to the Analysis of Wall
Papers and Fabrics. By Charles R. Sanger ... 24

V. On the Structure and Development of Choreocolax Polysiphonice,

Reinsch. By Herbert Maule Richards 46

VI. On the Matrical Equation <^i2 = £2 0. By Henry Taber . 64

VII. On the Products obtained by the Action of Nitric Acid upon
Bromtrinitrophenylmalonic Ester. By C. Loring Jackson
and W. B. Bentley 67

VIII. Note on Trihrommononitrobenzol. By C. Loring Jackson

AND W. B. Bentley 98

IX. 0)1 a Kephir-like Yeast found in the United States. By

Charles L. Mix 102

X. Dampening of Electrical Oscillations on Iron Wires. By

John Trowbridge 115

XT. Contributions to American Botany. By Sereno Watson 124

vi CONTENTS.

Page

XIT. Descriptions of New Plants, chiefly Gamopetake, collected in

Mexico by C. G. Pringle in 1889 and 1890. By B. L.
Robinson « i64

XIIT. Concerning the Life-History of Saccorhiza dermatodea. By

William Albert Setchell 177

XIV. On some simple Cases of Electric Flow in Flat Circular

Plates. By B. O. Peirce 218

XV. A Revision of the Atomic Weight of Copper. Fourth Paper.

By Theodore William Richards 240

XVI. The Action of Acetoacetic Ether on Quinones : Synthesis of

Benzofurfuran Derivatives. By M. Ikuta 295

XVII. Note on the Variation of Molecular Pressure. By Carl

Barus 313

Proceedings 327

Memoirs: —

Henry Jacob Bigelow 339

Charles Otis Boutelle 351

Alfred Hostiier . . 354

George Bancroft 355

Jaliu.s Erasmus Hilgard 370

Christian Heinrich Friedi'ich Peters 373

Charles John Maxiraowicz 374

Karl Wilhelm von Naes^eli 370

Eduard Schiinfeld 381

List of the Fellows and Foheigv Honorary INIembers . . 384
Index 391

Plate I.

Fig. 1. — Prehistoric Pallasite before Cltting.
[Closely natural size.]

Fig. 2. — Kiowa County Pallasite in Harvard Cabinet.
[Reduced to one fifth in linear dimension,]

PROCEEDINGS

OF THE

AMERICAN ACADEMY

OF

ARTS AND SCIENCES.

VOL. XXVI.
PAPERS READ BEFORE THE ACADEMY.

I.

THE PREHISTORIC AND KIOWA COUNTY
PALLASITES.

By Oliver Whipple Huntington, Ph. D.

Presented April 8, 1891.

In the Harvard Collection of Meteorites there is a small specimen of
a pallas iron which is very highly prized as having the oldest authentic
record. It is the main portion of a specimen which was found by
Prof. F. W. Putnam in 1883 on the altar of Mound No. 4 of the
Turner Group, in the Little Miami valley, Ohio. Therefore it is
entered in the Harvard Catalogue as " Prehistoric." Ever since the
acquisition of this Prehistoric specimen, specialists have been interested
in trying to identify the main mass from which the smaller individual
must have come. Figure 1, Plate I., shows the specimen of nearly its
natural size as it came from the mound.

When the meteorite was first placed in the Harvard collection, the
author of this paper made a careful study of the character and arrange-
ment of the various constituents of the mass, thinking it possible that
it might be identified with some of the pallas irons from the desert of
Atacama, South America, which it appeared to resemble, thus indicat-
ing that the old builders of the mounds had visited that part of the
world at some period in the remote past, and had brought away the

VOL. XXVI. (N. 8. XVIII.) 1

2 PROCEEDINGS OF THE AMERICAN ACADEMY

meteorite as a sacred object sent from Heaven. As is well known,
meteorites have been worshipped in very early times, and, since this
identical specimen was found on an altar, it must have been highly
prized, if not an object of adoration. The result of the examination,
however, showed that the Prehistoric iron differed in some of its most
essential characters from all the South American specimens. It did, on
the contrary, resemble most strikingly the famous Siberian meteorite
from Medvedeva, Krasnojarsk, which has given the name of its finder,
Pallas, to that class of meteorites in which the iron forms a continuous
network enclosing grains of transparent green or yellow olivine. Not-
withstanding the close resemblance of the Prehistoric and Pallas irons,
as the localities were so widely separated, it seemed improbable that
they could have come from one and the same original mass.

In 1880 a meteorite was found in Carroll County, Kentucky, and,
since it contained iron and olivine, it was at once described as the prob-
able origin of the Prehistoric mass, since no other olivine meteorite
had up to that time been found in this country.* The diagrams, how-
ever, published at that time showed that the two irons did not even
belong to the same class, since the Carroll County consisted of olivine
surrounding small masses of iron, while the Prehistoric was a true
jjallasite, consisting of iron enclosing olivine ; and in the further de-
tails of the specimens there appeared to be no resemblance whatever.
This fact has since been admitted ; and more recently, since some
very remafkable pallas irons were found in Kiowa County, Kansas,
these have been claimed to be the original masses from which the
Prehistoric came.f

One of the Kiowa County specimens was obtained from Prof. F. W.
Cragin of Washburn College, Iowa, for the Harvard Cabinet. On a
hasty examination, the resemblance of the mass to the Prehistoric
iron appeared very striking. A further study, however, brought out
certain features which are quite unique and worthy of description,
and at the same time led to the conclusion that the resemblance to
the Prehistoric meteorite was not so remarkable as it at first ap-
peared ; while, on the other hand, a further comparison of the Pre-
histoi'ic with the Pallas iron again showed the two to be almost
identical, as already stated.

The mass from the Kiowa County find now in the Harvard col-
lection weighs one hundred and twenty-six pounds. It has a more

* Am. Jour. Sci., Vol XXXIII . March, 1887, p. 228.
T ycience, Vol. XV. No. 384, June 13, 1890, p 359.

OF ARTS AND SCIENCES. 3

or less ragged outline, as shown in Figure 2, Plate I., which is a repro-
duction from a photograph, much reduced; and the striking outline of
a human profile on the right-hand upper corner suggests the possibility
that it might have been taken for an image of a god. Its dimensions
are, length fifteen and a half inches, height twelve and a half inches,
thickness ten inches. The exterior is deeply pitted, and shows signs
of fusion, as if the specimen had reached the earth as a complete indi-
vidual. Its internal structure consists throughout of a continuous net-
work of iron, enclosing grains of more or less transparent green olivine.
In certain portions, as would be expected, the olivine has become some-
what altered by weathering. One spur of the mass has been sliced, and
the slabs show in the unexposed portions beautiful crystals of trans-
parenf green olivine. The surface exposed by sawing has an area of
about ninety square inches, and exhibits some strikmg variations in its
different parts. Some of the olivine appears in two distinct zones, the
outer portion being so dark colored that at first sight it appears by
reflected light to be black, and on the large section just mentioned this
apparent dark olivine occurs most abundantly around the outer edges
of the section, extending in some cases an inch or more into the interior
of the mass. But it is still more noticeably distributed along a crack,
which extends irregularly through the mass and divides the large
cut surface into nearly equal halves. This crack is followed through-
out its entire length, a distance of ten inches, by an abundant deposit
of the dark olivine, the grains being separated from one another
by deposits of troilite, while at a short distance from the crack on
either side occurs transparent green olivine, wholly distinct from the
dark variety, and here the troilite is less abundant. In the original
description of the Kiowa County meteorites the peculiar appearance
of the olivine is described as follows : " Many of the olivine crystals
are in two distinct zones, — the inner half a bright transparent yellow,
the outer a dark brown iron olivine. In reality this dark zone is an
intimate mixture of the troilite and the olivine, as the analysis of Mr.
Eakins and a microscopical examination of the crystals by Mr. J. S.
Diller of the United States Geological Survey fully proved."* This
description, however, does not notice what is perhaps the most strik-
ing feature of the dark olivine, namely, that it is so strongly magnetic
that lumps of considerable size will readily jump to an ordinary horse-
shoe magnet. Since olivine is not attracted by the magnet, and most
troilite only feebly so, and pure troilite not at all, it seems hardly pos-

* Science, Vol. XV. No. 384, June 13, 1890, p. S61.

4 PROCEEDINGS OF THE AMERICAN ACADEMY

sible that a simple mixture of the two should become as magnetic as
magnetite.

On referring to the analysis made by ^Ir. L. G. Eakins in the labo-
ratory of the United States Geological Survey, it appears that the
composition of the dark olivine somewhat resembled that of hyalosi-
derite, — a variety of olivine which might be strongly magnetic. The
question then arose as to whether the zone of dark olivine owed its
magnetism to the composition of the olivine, or whether it was due
to an admixture of a magnetic variety of sulphide of iron.

It was very difficult to obtain pieces of the dark olivine free from
veins of troilite, but with sufficient care quite large fragments could
be picked out, which under the microscope showed a somewhat resinous
lustre, and a color varying from black and opaque to transparent red
and yellow or colorless, but with no signs of any characteristic grains
of troilite. After grinding the material to powder, the darker specks
could be separated by the magnet, leaving the light transparent part.
The magnetic portion thus selected would gelatinize with acid, but
also invariably would give the reaction for sulphur, although showing
no other characteristics of troilite even when examined in a thin sec-
tion under the microscope. The same material was found in the
Pallas and Prehistoric specimens, thoui>h of a still darker color and
giving a far stronger sulphur reaction. When it became evident
that the sulphur was a constant accessory of the magnetic olivine, it
precluded the possibility of the olivine being a distinct variety, like
hyalosiderite.

The distribution of the dark olivine mainly near the exterior of the
mass and along the crack, with only occasional patches in the interior,
would suggest its being an alteration product. This appeared still
more probable after examining a thin section of one of the dark olivine
crystals. It was seen at once that the crystals of olivine were intact
except for a dark deposit along the cleavage cracks. "Where the dark
portion bordered on the green, the olivine was somewhat discolored,
red or yellow, but evidently it had not undergone any change from
weathering. It seemed more as if the dark portion had been fused
and drawn into the cracks of the olivine. Experiments were there-
fore made to see what the effect would be of fusing the olivine and
troilite of the Kiowa County pallasite, and it was found that, if non-
magnetic troilite and crystals of transparent green or colorless olivine
were heated together out of direct contact with the air, and the tem-
perature raised to near the melting point of the olivine, the troilite
would turn black, become strongly magnetic, and permeate all the

OF ARTS AND SCIENCES. 5

cracks of the olivine, while portions of the olivine near the troilite
changed to various shades of red and yellow, the darker portions being
strongly magnetic. Otherwise, the olivine retained all its former char-
acteristics of crystalline form and cleavage. Moreover, portions of
the artificial product could be selected which would so closely resemble
similar specimens from the meteorite, that, after they had been once
mixed together, they could not be distinguished. This seemed to show
conclusively that the dark outer zone of olivine is a mixture of troilite
and olivine only in the sense above indicated.

On an exammation of the slabs of this iron by etching, certain other
features appear which in part connect it most closely with the Pallas
and Prehistoric irons, and in part are peculiar to the Kiowa County
iron alone.

The olivine crystals are in the first place surrounded by a deposit
of what is probably the purer iron. This border is of varying thick-
ness, not generally exceeding one or two millimeters, and occasionally
wholly disappearing. Parts of the border are most beautifully marked
by innumerable Neumann lines, microscopically fine, and so numerous
as to give it a silvery appearance and a brilliant lustre that at once
strike the eye. Between this iron border and the olivine come
masses of troilite, and these fill the space between the olivine crystals,
and thus have the same kind of outline as the iron network, and
appear as a continuation of the network. Troilite, however, also
occurs in small nodules in the iron itself, and sometimes asfain as
the central portion of an olivine crystal. Such a quantity and dis-
tribution of the troilite appear in no other meteorite except the
original Pallas. A further resemblance between these last named
meteorites is brought out in the occurrence of schreiberseit. This
mineral is at once distinguished by the eye from the troilite, on account
of its brighter lustre, granular structure, and more silvery color. Fur-
thermore, it is harder, and strongly magnetic. The schreiberseit oc-
curs in close connection with the troilite, usually in patches coming
between it and the iron, but frequently portions of schreiberseit project
into the iron itself, sometimes in elongated masses reaching a lensfth
of half an inch or more. In such cases the iron borderinff on the
schreiberseit has the same characters as when in contact with the
olivine crystals, as just described.

The main portions of the iron show a most perfect crystallization,
which is very beautifully brought out by etching. Crystal plates start
out from the iron border already mentioned, and reach back through
the whole extent of the interior iron. The figures thus produced are

6 PROCEEDINGS OP THE AMERICAN ACADEMY

finer and sharper than those of any other specimen in the Harvard
coUectiou. They consist of thin plates closely packed together, not
exceeding half a millimeter in thickness, but sharply defined by their
border lines of bright uickeliferous iron. These plates as seen in
section exhibit beautiful Neumann lines, the same as the border iron
previously described, and appear to be of a piece vpith it. A slight
suggestion of a similar crystallization is given by the Pallas iron, but
the specimen at tlie author's disposal is not large enough for a thorough
comparison. In cutting the large Kiowa County mass, the saw passed
through portions of iron of considerable area, and in etching one of
these nodules the plates were brought out in their greatest perfection.

The appearance of the Widmanstattian figures is best shown by
Figure 3, Plate II., which is printed directly from the iron. Unfor-
tunately, the i)rinting does not bring out the Neumann lines on the
border iron, though they are suggested in some parts of the illustra-
tion, but the ireneral character and distribution of the Widmanstattian
plates is very fairly shown. One peculiarity, however, does not exist
in the slab from which the plate was printed, and though it only occurs
in two or three places, yet, as it has not been observed in any other
meteorite, it is worthy of note in this connection. Ordinarily in the
pallasites the olivine is surrounded by a layer of iron, as if the iron
had been depo^^ited on the olivine as a nucleus ; but in some of the
specimens of the Kiowa County the plates of iron which form the
AVidmanstiittian figures actually project into the olivine crystals, as if
the two had solidified simultaneously. It is common here, as in other
pallasites, for little plates of iron to separate two crystals of olivine ; but
in the specimen under discussion there are cases where the Widman-
sttlttian plates cross the natural boundary line of the iron and project
out into the olivine crystals, and intersect each other at the octahedral
angle apparently wholly independent of the presence of the olivine.
Occasionally, too, a little plate of iron is seen isolated from the rest,
and in the very centre of a nodule of olivine. No cases were ob-
served where the Widmanstattian plates actually cut through the
olivine crystals so as to connect with the iron on the opposite side,
but they projected into the crystals several millimeters.

Finally, by far the most striking and characteristic feature of the
Kiowa County pallasite is the abumlant occurrence of chromite. This
mineral is easily confused with the dark olivine described in an earlier
part of this paper, since the mode of occurrence is exactly the same,
and on a polished surface the only difTerence is that the chromite has a
more metallic lustre, and is more opaque. Chromite is widely dis-

OP ARTS AND SCIENCES. 7

tributed through the meteorite intimately mixed with olivine, although
in some cases large crystals of clear green olivine are to be found em-
bedded in the ehromite. In several places masses of nearly pure
chromite of more than an inch in diameter appear, intersected by the
network of iron with its accompanying troilite and schreiberseit, the
chromite largely replacing the olivine. This mixture of chromite and
olivine appears as a whole perfectly black and opaque, breaking with
a subconchoidal fracture, and having a brilliant submetallic lustre.
It is strongly attracted by the magnet, and differs most markedly from
the dark olivine before described in being perfectly opaque, and the
powder is dark brown or black where in the former case it was gray.

Section of Olivine Crystal showing Chromite.

A perfect octahedron with an axis of about two millimeters was broken
out from one of these masses, but in most cases the chromite took the
form of the olivine. The cause of this will be evident by reference to
the accompanying figure, which represents a microscopic section of a
black perfectly opaque and strongly magnetic lump of this material,
the diagram having been drawn from the microscope by means of a
camera lucida. It will be seen at once that, instead of its being a
homogeneous substance, it proved to be a section of a transparent
crystal of olivine which had had all its cleavage cracks well filled with
chromite. Under the microscope, the chromite is still perfectly opaque
even in such a thin section, and has a noticeably metallic lustre, with
no gradual change from dark to light, as was the case in the other
variety of dark olivine.

8 PROCEEDINGS OP THE AMERICAN ACADEMY

On examining the specimens of the Prehistoric and Pallas irons
in the Harvard collection, no chromite was found, though of course
the absence of it in those particular specimens cannot be taken as con-
clusive evidence of its absence in all ; but it may be safely said that it
would be impossible to select equally large specimens from the Kiowa
County iron which would be free from this mineral.

In comparing the three pallasites, Prehistoric, Krasnojarsk, and
Kiowa County, we obtain the following results.

First. All three have the dark olivine, strongly attracted by the
magnet, appearing near the outside of the specimen, and frequently
surrounding the clear green variety.

Secondly. All have a striking border of iron surrounding the oli-
vine, showing a silvery sheen from the innumerable Neumann lines.

Thirdly. All show signs of Widmanstattian figures in the ground-
mass of the iron.

Fourthly. The Krasnojarsk and the Kiowa County both have a
large quantity of troilite between the crystals of olivine, and also
patches of schreiberseit between the troilite and the iron, and occa-
sionally included by the iron. This same character appears, but in a
much less degree, in the Prehistoric.

Fifthly. The Krasnojarsk and Kiowa specimens further show a
much larger proportion of iron than the Prehistoric, though closely
resembling each other in that respect. Figure 4, Plate II., shows a
slab of the Prehistoric, printed directly from the iron.

Sixthly. The Kiowa County iron shows a very striking and far
more perfect crystallization than any other pallasite heretofore de-
scribed, so that if AVidmanstiittian figures can be used at all as a means
of distinguisliing irons of different falls, then the Kiowa County is
distinct from any meteorite thus far described.

Lastly. The Kiowa County pallasite contains large quantities of
chromite distributed through it, completely permeating large masses of
the olivine, but no chromite is to be found in the Prehistoric iron or in
the specimens of the Pallas meteorite in the Harvard collection.

In the description of the Kiowa County iron already referred to, the
analysis and specific gravity of the olivme of the Prehistoric are com-
pared with those of the Kiowa County as a proof of the identity of the
two ; but a glance at the following table will show that the analysis
of the Kiowa County olivine shows a still closer resemblance to the
olivine from the Pallas meteorite, and also that from Mount Etna.

OP ARTS AND SCIENCES.

!f

SiOj

ALO3

Fe^Oa

FeO

NiO

MnO

MgO

u.o

Total.

Sp. Or.

Kiowa County

Etna

Pallas

Vesuvius . . .
Mt. Somma .
Prehistoric . .
Atacama . . .
Antuclo, Chili

40.70
41.01
40.83
40.35
40.08
40.02
36.92
40 70

tr.

0.64

tr.

0.18

0.18

10.79
1006
11.53
12.34
15.26
1406
17.21
19.60

0.02
0.20

0.14

0.29

0.48
0.10
1.86

48.02
47.27
47.74
46.70
44.22
45.60
43.90
39.70

1.04

99.85
100.22
100.39

99.39
100 24

99.78

99.89
100.00

3.376
3.334
3.334

3.336
3.330

Furthermore, since the specific gravity of olivine only varies from
3.33 to 3.56, it is not surprising that the specific gravity of the Kiowa
County and Prehistoric olivines should be so nearly alike ; but it will
be seen by the above table that it differs from the Prehistoric more
widely than the latter differs from the other volcanic olivines. The
above table could be largely extended, but enough is given to show
that the olivine from the Kiowa County meteorite more closely re-
sembles some modern volcanic products and the original Pallas me-
teorite, than it resembles the Prehistoric. Furthermore, when the
difficulty of obtaining an average sample of meteoric material for
analysis is borne in mind, it will be seen how useless it is to compare
the analyses as a proof of the identity of any meteorites, and particu-
larly the olivine ones. But the same difficulties lie in the way of
giving undue weight to the etched figures or structure of a small speci-
men. For instance, the well known meteorite that fell at Esther-
ville, Emmet County, Iowa, is made up largely of olivine, the iron in
most specimens not forming a continuous network ; yet portions can be
selected almost entirely free from iron, and others where iron forms
much the larger part. Figure 5, Plate IL, is pi-inted directly from an
iron nodule taken from the Estherville meteorite, and placed here for
comparison with the similar mass shown in the Kiowa County slab.
It will be seen at once that the etched surfaces of the two irons show
utterly different Widmanstiittian figures. Yet the Estherville speci-
men does not exhibit any of the peculiar characteristics which we are
in the habit of associating with that iron, and the slab from which
Figure 5 is printed might be mistaken by an expert for at least half
a dozen typical Widmanstattian irons, while the olivine iron to which
it really belongs would be one of the last to be compared with it.

A similar feature appears even more strikingly on comparing the
specimens found in Kiowa County. The largest of these has through-
out the structure typical of the pallasites, but several of the smaller

10 PROCEEDINGS OF THE AMERICAN ACADEMY

masses consisted wholly of iron exhibiting a highly developed crystal-
line structure, and in one at least, which passed into the possession of
Mr. Howell of Rochester, N. Y., one end of the mass was pure iron
whila the rest was pallasite. We give a figure of an etched slab of
this last mass. Figure 6, Plate III., which is printed directly from the
iron. When first found, one was inclined to believe that the irons and
the pallasites so closely associated over an area hardly exceeding sixty
acres could not have come from the same fall, and this opinion seemed
supported by the greater coarseness of the Widmanstiittian figures on
the sections of the isolated irons ; but such association as is exhibited in
Figure 6, Plate III., makes the intimate connection evident, and gives
evidence in favor of the theory which regards meteorites as resulting
from intensely violent volcanic outbursts on the surface of a planet so
far cooled that the still melted nucleus was coated with an earthy
crust, through which the surface water as it condensed percolated to
the molten interior. By the resulting violent eruptions, of which we
can form only a faint conception from terrestial volcanoes of the
present day, this crust would be fissured on long lines of least resist-
ance, and volcanic bombs thrown into space beyond the sphere con-
trolled by the planet's attraction. Tiiose bombs which came from the
zone of contact of the melted iron with the crust would naturally have
the structure of pallasites, mixed with masses consisting wholly or
chiefly of metal. The specimen from which Figure 6 is printed
seems to furnish the link needed to connect the stony with the iron
meteorites, and if the iron portion be compared with the iron in
Figure 3, Plate II., from the same fall, it will be seen of how little
value the appearance of the etched surfaces would be in identifying
selected portions from the two slabs.

The result of this discussion merely shows the impossibility of iden-
tifying these pallasites, which at first sight appear so much alike. If
the Kiowa County specimens are accepted as identical with the one
from the mounds, then both must be the same as the one from Krasno-
jarsk, Siberia, without further question ; but the striking occurrence
of chromite in such unusual quantity in the Kiowa County mass
would seem sufficient to place that by itself, while leaving the close
resemblance between the Prehistoric and Krasnojarsk pallasites as yet
unexplained.

The question naturally arises as to the possibility of the mound
builders having actually brought the Prehistoric specimen from Siberia
as a sacred object. This is scarcely probable. It is well known from
the writings of various authors that the inter-tribal traffic of the

OF ARTS AND SCIENCES. 11

American Indians was very <:freat, and tliut tlioy occasionally made
journeys completely across this continent, as, for instance, the journey
of IMoncacht-Ape,* together with others fully as remarkable and per-
haps more authentic. Moreover, Margry writes : " 11 nous disoit, dans
la Bibliotheque du Roy, a. leu M. Thevenot et a moy, qu'il estoit dans
la mesme opinion, et cela d'autant plus que le Pere Martini luy avoua,
en la Chine, qu'il avoit coni'esse en espagnol uue femme Mexitjuaine,
qui, ayant este enlevee esclave au Mexique, estoit arrivee de pays en
pays, de nation en nation et d'esclavage en esclavage, en la Chine,
par terre, sans avoir passe qu'un petit dctroit de mer, et cette histoire
est rapportt^e dans le cinquiesme volume in-8° en italien, avec figures
de Giro del Mondo, du docteur Gemelli, Napolitain, arrive depuis uu
an a Naples, d'ou je me suis fait venir ce livre en six volumes." t

If this could be believed, it might be possible that the Prehistoric
iron had found its way across Behring Straits till finally it was col-
lected, with the other relics from all parts of this country, on the altar
of Mound No. 3 of the Turner Group in Little Miami Valley, Ohio.
Unfortunately for this solution, Prof. F. W. Putnam considers that
the people who built the mounds came fi'om the south. That they may
have at one time had communication with China seems probable from
the frequent occurrence of specimens of jadeite among their imjjlements,
and from the fact that in most cases the jadeite implements have been
subsequently cut up into ornaments, while the nephrites have been left
intact, showing that the former must have become more and more
rare and highly prized as they were passed down through successive
generations.

Provided that the original owners of our Prehistoric meteorite had
associations with China, still there is no reason for supposing that they
had any communication with Siberia. Nevertheless, can it be consid-
ered much more remarkable that a Siberian meteorite should be found
on an altar in an Ohio mound, than that a pipe of the red indurated
clay, found only on the Pipe Stone Branch of the Little Sioux River
of the Missouri, should be picked up on the banks of the Rio de la
Plata in South America, and several more in New England ? J

The only other explanation of the close resemblance between the
Prehistoric iron and the original Pallas would be that they were two
portions of the same meteoric outburst which fell at remote distances

* Proc. Am. Antiquarian Soc, April 25, 1883.

t Decouvertes de I'Amerique Septentrionale, 1614-1754, Vol. VI. p. 173.

t Long's Expedition, Vol. I. p. 31.

12 PROCEEDINGS OF THE AMERICAN ACADEMY

from each other, possibly in two hemispheres, though as yet the most
distant places in which the same meteorite has been identified are
Mexico and Kentucky.*

The subject is an interesting one, and could be discussed at much
greater length than the limits of this paper will allow, but the only ob-
ject of this discussion has been to show that there is a wonderfully close
resemblance between the Prehistoric and Pallas iron, and that, though
the Prehistoric resembles the Kiowa County far more closely than
it does the Carroll County, yet there is no reason for regarding them
as identical.

In this discussion the Harvard mass of the Kiowa County find has
been compared with another specimen from the same, described in a
paper previously quoted as if the two were identical, and, in closing,
the author would mention that he has had the opportunity of examin-
ing the various specimens in mass, and also the cut slabs placed side
by side, and they are unquestionably the same.

* Proceedings of this Academy, Vol. XXIV. p. 30, October, 1888.

Plate II.

Fig 3. — Kiowa Cokxty, Pkixtkd uikectlv from the Slab.

■■*' '■" '■' ''ill

Fk;. 4. — I'ltKllISTOUK',
I'lUNTED FRO.M THK IkOX.

Fi(i. "). — 1'hixtki> KHO.M

ESTHEKVILLE >."oUL Lli.

I'r.ATK III.

*1\! :»M¥«r .::■■ 'LILT

Fio. C — Suii ..1- Ki,,n A r.iiMV 1j:..n iieceivei. i bom Wain, asu li.raEi.i..

OP ARTS AND SCIENCES. 13

II.

PRELIMINARY KOTES ON THE SPECIES OF
DOASSANSIA, Cornu.

By William Albekt Setchell.

Presented March 11, 1891.

The genus Doassansia was established by Cornu, in 1883,* to
receive the Sclerotium AUsmatis, Nees, on Alisma, and a new species,
D. Farlowii, on Potamogeton. Since then there have been several
additions both of old and of new forms, until at present the number
of species referred to it is twelve. All of them inhabit hosts which
are more or less aquatic in habit, though belonging to widely sepa-
rated families. They are said to differ little" in structure, but to be
distinguished from one another chiefly by the differences of the host
plant.f A careful study of the species distributed in the various
" Exsiccati," as well as of the accessible living material, has shown
that this is not strictly correct. Not only are most of the species
fairly well characterized by peculiarities of structure, but there are
also several types of structure sufficiently diverse to be given sub-
generic or even generic rank. Moreover, by the discovery of sev-
eral new species, additional types of structure have been found, and
have rendered it even more necessary that a careful revision of all
the species should be made. On this account, full descriptions and
figures of all the specias of which material was available have been
prepared, but as there is a delay in publication, it has seemed best
to give a brief summary of the results in the present preliminary
notice.

The spores of the Doassansice resemble those of the species of
Entyloma both in structure and in germination ; but in the former
they are collected and compacted into balls, called by most writers
" sori." The species of the genus Doassansia have in addition a coat,
or "cortex," of sterile cells surrounding the sorus, Cornu certainly

* Ann. Sci. Nat., ser. 6, Tom. XV. p. 285.

t Cf. Schroeter, Pilzfl. Schles., p. 286, 1887. De Toni, Journ. Myc, Vol.
IV. p. 14, 1888.

14 PROCEEDINGS OF THE AMERICAN ACADEMY

considered that the possession of this cortex was the distinguishing
feature of the species of his genus, although, as will be noted under
D. Farlowii, he did not always recognize the true cortex. It seems
best, therefore, to refer to Doassansia all the species of the group of
the Entylomata which have the sorus invested with a cortex of sterile
cells.

The sori of several species referred to Doassansia show, when
thin sections are examined, that this cortex is not present. Such are
the sori of D. NiessUi, De Toni, D. Limosellce, (Kunze), Schroeter,
D. decijnens, Winter, and an undescribed form on Echinodorus ros-
tratiis, mentioned by Harkness under B. AUsmatis* These several
species, together with Entyloma crastophilum, Sacc, and probably
others, form a group intermediate between the simpler species of
Entijloma and those of Doassansia proper ; but the limits of this
group cannot be ascertained with any exactness without a careful
study of all the numerous forms which of late have been referred as
species to Entyloma. They must therefore be left unsettled in
position for the present.

Among the Doassansice, as is the case among the Entylomata in
general, the specific distinctions are not striking. They differ slightly
in habit when dried, yet when fresh most of the species may be
distinguished at a glance by the peculiar distortion or discoloration
of the host plant produced by them. As a rule, the Entylomata do
not produce distortions, but two of the species of Doassansia cause
swellings of considerable size. The structure of the sorus varies
decidedly, and has been made in these notes the basis of generic and
subgeneric distinctions. The germination has been obtained wherever
possible, and has been found to vary in its details among the different
species.

Following is given the arrangement of the species and genera.

DOASSANSIA, Cornu.

Spores resembling those of Entyloma both in structure and in
germination, collected and compacted into sori. Cortex of sterile cells
present.

Subgenus I. Eudoassansia.

Body of the sorus consisting entirely of spores, which are readily
separable from one another at maturity.

* Cf. Proc. Cal. Acad. Sci., ser. 2, Vol. II. p. 231, 1889.

OF ARTS AND SCIENCES. 16

The type of this subgenus is D. AUsmatis, Cornu, which represents
Cornu's idea of Doassansia.

1. D. Epilobii, Farlow.

On leaves of Epilobium alpinum.

United States !

The cells of the cortex of this species are very small and flattened,
and the sori, on this account, very closely resemble those of
D. decipiens, Winter. It may perhaps be looked upon as a form
intermediate between the group represented by D. decipiens and
the group of the Eudoassansice.

2. D. HottonifB, (Rostr.), De Toni.
On leaves of Hottonia palusfris.
Denmark !, Germany !, France ! .

3. D. Sagittarice, (Westend.), Fisch.

On leaves of Sagittaria sagittifoUa^ graminea, variabUis, and Mon-
tevidensh.

Italy !, France !, Germany !, Belgium !, England ! ; Argentine Re-
public ; Canada !, United States ! .

4. D. opaca, sp. nov. Spot orbicular, slightly swollen on both surfaces

of the leaf, lemon-yellow, at length dark brown and opaque. Sori
crowded, indistinct when viewed with a lens, globular to almost
cubical, 200-300 /x by 80-100 /a, light brown. Spores rather
loosely compacted together, nearly spherical, 10-15 /x in diameter.
Cortical cells of various shapes, from brick-shaped to almost
cubical. Germination unknown.
On leaves of Sagittaria variabilis.
United States ! (Newton, Mass., W. G. Farlow !, Medford, Mass. !,

Norwich, Conn. !).
This species was mentioned by Farlow * under the name of Pro-
tomyces Sagittarice, as occurring at Newton, Mass. The same form
has been collected by myself, both at Medford, Mass., and at Norwich,
Conn., in abundance, and has been studied in all stages of develop-
ment. The species differs decidedly from D. Sagittarice in habit
and form of the sorus, as well as in the character of the cortical
cells. Sowings of the spores have been repeatedly made, but the
germination, unfortunately, has not been obtained. The species
is readily detected by holding an infected leaf between the eye and

* Bot. Gaz., Vol. VIII. p. 276, August, 1883.

16 PROCEEDINGS OF THE AMERICAN ACADEMY

the light, when the spots appear as dark, black patches in the leaf
substance.

5. D. Alismatis, (Nees), Cornu.

On leaves of Alisma nutans and Plantago.

Italy!, P^ ranee, Germany!, Finland!, England!; Siberia!; United
States ! .

Subgenus II. Pseudodoassansia.

Central portion of the sorus composed of an irregular-shaped mass
of fine, densely interwoven hyphce. Spores in several layers, loosely
compacted together. Cortex of large, well differentiated cells.

6. D. obscura, sp. nov. Spot none. Sori in lines in the larger inter-

cellular spaces at the base of the petioles and peduncles of the host,
globular to ellipsoidal, 180-220 /x by 200-300 /x, light brown.
Spores almost globular, 8-1 2 /a in diameter, light-colored. Cor-
tical cells obversely conical, conspicuously lobed at the outer,
broader end. Promycelium narrowly cylindrical, about 20 /*
long. Sporidia in whorls of 5 to 7, elongated fusiform, 16-17 jx
long and 1.5-2 /x. thick, producing bunches of secondary sporidia
without conjugation.
On petioles and peduncles of Sagittaria varinhllis.
United States ! (Cambridge, Mass. !, Medford, Mass. !, Norwich,

Conn.!).
This species is very inconspicuous, being detected only upon the
most careful examination. When occurring upon the green portions
of the petioles and peduncles, it causes a very faint yellowish discolor-
ation. It most frequently inhabits the white portions at the very
base of these parts, and then the dark lines of sori show through the
more or less transparent outer tissues. It is abundantly distinct from
all the other species of the genus. The central hyphoe, the loosely
compacted spores, the obconic lobed cells of the cortex, and the method
of germination of the spores, are all characteristic. It seems to differ
so much from the species which cluster about D. Alismatis as to
demand a special subgenus for its reception.

Subgenus III. Doassansiopsis.

Sorus compact, not separating into its component elements at
maturity. Central portion consisting of a compact mass of pai'enchy-
matous tissue. Spores in a single layer. Cortex of small flattened

cells.

OF ARTS AND SCIENCES. 17

7. D. occulta, (Hoffm.).

D. occulta, (HofFm.), var. Farlowii, (Cornu).

Authentic specimens of D. Farlowii, Cornu, do not seem to me to
correspond exactly to Hoffmann's figures * but differ particularly
in having spores which are elongated in a radial direction.
Specimens collected by myself near Norwich, Conn,, agree better
with Hoffmann's figure, and are considered in these notes to
represent the type of D. occulta, while D. Farlowii is for the
present placed as a variety under it. The elongated cells on
the periphery of the sorus in Cornu's figure f are really the
spores, and the real cortex of small, radially flattened cells is not
shown, while the rounded cells in the centre are not immature
spores, but are sterile cells resembling parenchyma cells.

In the ovaries of species of Potamogeton.

Type. Germany ; United States ! (Norwich, Conn. !).

Var Farlowii. Canada (Ottawa, /. B. Fletcher !).

8. D. Martianojfftana, (Thuem.), Schroeter.
On leaves of species of Potamogeton.
Siberia ; Germany, Sweden ! ; Canada ! .

In the specimens from Sweden, distributed by Johanson, $ there
appear to be conidia, almost identical in appearance with those already
known in some species of Entyloma. There seems to be an intimate
connection between the mycelium of the Doassansia and that of the
conidia.

9. D. deformans, sp. nov. Species forming distortions, often of large

size, on all parts of the host. Sori globular, 100-140//, very
light brown. Spores polyhedral, 8-10/^ by 4-8 /x. Cortical
cells polygonal, flattened radially. Promycelium somewhat ob-
conical, about 12 /* long. Sporidia 5 to 6, broadly fusiform, 12 /x
by 4-5 //, conjugating and producing a short germ tube.

In the leaves, petioles, peduncles, pedicels, and ovaries of Sagittaria
variabilis.

United States ! (Norwich, Conn. !, Cambridge, Mass. !) ; Canada
(London, leg. Benrness, comm. J. B. Ellis).

A species forming large distortions on Sagittaria, nearly related to
the other species of the subgenus, but abundantly distinct from all the

* Ic. Anat. Fung., Taf XVI. Fig. 3, 1862.
t Ann. Sci. Nat., Tom. XV. PI. XVI. Fig 6, 1883.

X Cf, Eriksson, Fung Scand. Par., No. 264, 1888. Fazschke, Fung. Eur.,
No. 3602, 1890.

VOL. XXVI. (n. s xviii ) 2

18 PROCEEDINGS OF THE AMERICAN ACADEMY

Other species on the same host. It is to be distinguished from
D. occulta (type) by its method of germination. It is without doubt
more widely spread, but the distortion is probably mistaken for the
work of an insect rather than that of a fungus.

Species Inquirendce.

10. D. Cotnari, (B. & Br.), De Toni.
On leaves of Comaram palustre.
England.

11. D. punctif or mis, ^ 'inter .

On leaves of Lythrum hyssopifolium.
Australia.

12. D. Lijthropsidis, Lagerh.
Od Lythropsi s peploides.
Portugal.

Species Excludendoe.

D. Niesslii^ De Toni.

D. Limosellce, (Kunze), Schroeter.

D. decipiens, Winter.

D. Alismatis, Hark, (not Coruu).

EURRILLIA, gen. nov.

Sorns compact, not separating into its elements on being crushed.
Central portion composed of an irregular mass of parenchymatous
tissue. Spores closely resembling those of Entyloma^ both in structure
and in germination, compacted into several dense rows. Corte x none
or composed only of a thin, irregular ayer of hardened hyphai.

B. pustulata^ sp. nov. Spots irregularly orbicular, often confluent,
light yellow, chiefly hypophyllous. Sori elongated ellipsoidal, at length
bursting through the epidermis, which appears raised in small blisters,
200-350 /A by 150-180 /a, light brown. Spores not separable at ma-
turity, almost globular, 4-6 jx in diameter, germinating while the
sori are in position. Promycelium cylindrical, IS/x. long, bearing 4-5
sporidia in a whorl at the blunt apex. Sporitlia slightly bent,
1 6 /x by 3 /x.

On leaves of Sagittaria variabilis.

United States ! (Illinois, leg. G. P. Clinton !, comm. T. J. BurriU,
Wisconsin, W. Trelease !).

The sori of this species resemble in structure the spore balls of
Testicularia Cyperi, Klotsch. ; * but the position of that species is un-

* Cf. Cornu, Ann. Sci. Nat.,ser. 6, Tom. XV. pp. 270-273, PI. XIV. Fig. 1-5,
1883.

OF ARTS AND SCIENCES. 19

certain, and even if it is one of the Uatilagincfe at all, it belongs rather
to the series of the pulverulent smuts, while the present form, both
in spore structure and in germination, is closely related to Doassansia
and to Entrjloma. It is to be distinguished from any of the species
of Doassansia by the lack of a cortex, and from any of the other
Eiitylomata by the presence of the central parenchymatous portion.
In the latter, it resembles the forms described under the subgenus
Doassansiopsis, but it differs from them not only In the lack of a
cortex but also in the possession of several layers of spores. The
genus is named in honor of Prof. T. J. Burrill, of the University of
Illinois, by whom the first specimens were sent.

CORNUELLA, gen. nov.

Sorus hollow at maturity, the interior containing only loose, hard-
ened hypha?. Spores compacted into a firm layer on the outside,
resembling those of Entyloma both in structure and in germination.
Cortex none.

C. Lemnce, sp. nov. Spot none. Sori globular to ellipsoidal, 50-70
-lOOjain diameter, dark brown. Spores not separable at maturity,
often elongated radially, 10-12 /^ by 6-10 /x,. Promycelium some-
what obconical. Sporidia in whorls of 5 to 7, narrowly fusiform, 26 /x
by 2 yu., producing a germ tube after conjugating.

In the fronds of Lemna (Spirodela) polyrrhiza.

United States ! (Cambridge, Mass. !, Newton, Mass. ! ; Belchertown,
Mass., J. E. Humphrey !).

The type of this genus, which I respectfully dedicate to Prof.
Maxime Cornu, of the Jardin des Plantes, is very different from
any described member of the Vstilaginece. The hollow sorus with
only loosely entangled hyphai on the inside, is unique, and yet the
spore structure and germination closely ally it with Doassansia and
Burrillia.

The germination takes place while the spores are in position, and
the whole sorus is covered with a bristly mass of promycelia and
sporidia. Something of the same thing happens also in Burrillia
pustulata and in the species of the subgenus Doassansiopsis.

20

PROCEEDINGS OP THE AMERICAN ACADEMY

III.

ON SOME THEOREMS WHICH CONNECT TOGETHER
CERTAIN LINE AND SURFACE INTEGRALS.

By B. O. Peirce.

Presented May 13, 1891.

In transforminw from one set of curvilinear coordinates to another,
some of the differential expressions which appear in problems in
Hydrokinematics and Electro kinematics, I have found the theorems *
stated below useful.

Theorem. — Let U be any function of the two polar coordinates,
T and d, which, with its first space derivatives, is finite, continuous,

Fig. 1.

and single-valued throughout that part of the coordinate plane which
is shut in by the closed curve T. Let 8 be the angle between the ra-
dius vector, drawn from the origin to any point P on T, and the normal
to T drawn from within outwards at P. Then, if 7' does not include

* London Educational Times, January and February, 1891.

OF ARTS AND SCIENCES. 21

the origin, the line integrals of U cos 8 and CTsin S, taken around T,

are equal respectively to the surface integrals of —^ and ,

taken over the area enclosed by T.

For the element of plane surface in polar coordinates, rAr^O may
be used. Let the radius vector OP, drawn so as to make the angle 6
with the initial line OX, cut T 2n times at points P-^, P^, P3, .... P^n-,
distant respectively r^ , i\, r^, .... r^^ from 0. Let the values of U
at these points of intersection be Ui, U^, V^, .... f^jn? respectively.
Whenever the radius vector cm^s into the closed contour, either +S
or — 8 is an obtuse angle and cos 8 is negative ; whenever the radius
vector emerges from the space enclosed by the contour, either + S or
— 8 is acute and cos 8 positive. The two neighboring radii vectores,
OP and 0P\ which make with each other the angle A^, include be-
tween them the arcs As^, As2, ^Ss? A«4, .... Asjn, cut out of T, and
the arcs r^AQ, r^AQ , ^3^^, .... r2„A^, cut out of a set of circum-
ferences drawn about 0 as centre, with radii r^, r^, r^, r^, .... r^n,
respectively. It is evident that, if A 6 be made to approach zero as a
limit,

+ Limit —
A

cos Sj

= — Limit

^2

• A^

:=

+ Limit

ASg

A5

AS2

• cos 8.^

• cos 8g

= — Limit

■AO

■ cos 84 ~

= + Limit —

A. '2

;n-i ■

^■A0

COSfi^n-

-1

= — Limit

»'2n

^^2„-

•A^

cosSa,

= -1.

If the double integral be extended all over the space enclosed by T,

fS^^^ '''^''^^ "" P^ ^~ '''^' + r,U,-r,Us + .... + r,,U,,-],

where the integration with respect to 6 is to be extended over all
values of the angle for which the corresponding radii vectores cut T.
If now for r-iAO, r^AO, r^AO, etc., — cos Si ■ ds^, + COS82 -ds^,
— cos S3 • dss + ....+ cos San <^^2n t>6 Substituted respectively, we have

I I ^J[S^ — -rdrd9= / [C^i cosSi£?Si + C^cos82C?S2+ •— ^nCos82„],

and this last integral is evidently equal to the line integral of UcosS
taken all around T.

It is to be noticed that, if 0 Were within T, each radius vector would
cut T an odd number of times, and that a negative sign must stand
before the line integral. *

99

PROCEEDINGS OF THE AMERICAN ACADEMY

Since the limit of the ratio of Ar to any one, QQ', of the arcs cut
out of T by two circumferences of radii r and r -{■ \r respectively,
drawn around (? as a centre, is equal in absolute value to the sine of
the angle w^iich 0 Q makes with the external normal to T at Q, it is
easy to prove the second part of the theorem by integrating D^ U with
regard to 6 first, and then, after introducing proper limits, with regard
to r.

This theorem may be regarded as a useful special case of the
following

Theorem. — Let I =z f^ (x, y) and -q =: f^ {x, y) be two analytical

functions of a: and y such that the two families of curves fi (x, y) = c,

f^ix, y) = Z:, are orthogonal. Let Vhe any function of x and y which,

with its first space derivatives is finite, continuous, and single-valued

within a closed curve T, drawn iu the coordinate plane. Let hi and

Fig. 2.

^2 be the positive roots of the equations Iti^ = (D^ tf + {Dy ^)^,
h^ — (Z)^ r;)- + (Dyrj)'^. Then, if ^ has neither maximum nor mini-
mum values within 7] the surftice integral of /«i • //o • />,; I — ), taken

all over the area enclosed by T is equal to the line integral taken
around T'of FcosS, where 8 is the angle between the exterior normal
drawn to T at any point, and the curve of constant 77 drawn through the
point, and where the direction in which ^ increases is taken positive.
Similarly, if proper regard be had for signs,

f p'l ■ h A {T) ^« = r^sin 8 . ds.

If through any point, P, in the coiirdinate plane, two arcs .v, , So be

drawn along which ^ and ij are respectively constaut, d Sy = '' ,

2

OF ARTS AND SCIENCES. 23

ds^ = ~, and for the element of surface -— — ^ may be used. The

"l / y\ "1 ' "2

surface integral of hi-Ii.^.D^ i- J taken over the area enclosed by

7' is

o =JJh ■ K ■ A (;|-^) ds, . ds, =Jd r/Jn^ (J^ dc

Consider two curves 0 Q, 0' Q' along which r/ has respectively the
constant values r]o and r/o + A r/ ; and let ^ increase in the directions
OQ, O'Q'.

Let OQcutT 2n times at the points P', P", P'", ....,P^'-"\ where
the values oi' /u are h^', h.^", h.^", .... , /^j'""', respectively, and the cor-
responding values of V, V, V", V", ...., f^^-"\ The curved line
0 Q makes with the normals drawn to T at P', P", P'", etc., from
within outwards the angles 8', 8", l'", etc., and the two curves OQ,
O'Q', cut out of 7" the 2» arcs As', As", As'", ...., AP"K

where the integration is to be extended over all values of rj which
occur within T.

The angles 5', 5'" 5^^"""'^, or their negatives, are all obtuse and

their cosines are negative, but the angles 8", 8^'^', .... S^-"\ or their
negatives, are all acute and their cosines are positive, so that at every
point, P^"', where OQ cuts 7' we have

Limit r (-!)-• A s^-^ cos 8r-n_^

d n

and in the expression for O we may write ( — 1)'^ • cos S™ c?s™ for yj^..

"2

Hence,

where the sign of integration directs us to find a similar expression to
that in the brackets for every pair of consecutive curves of constant r/
which cut T, and to find the limit of the sum of the whole. This is
evidently equivalent to integrating T^cosS all around the curve T.

Jefferson Physical Laboratory,
Cambridge.

24 PBOCEEDINGS OP THE AMERICAN ACADEMY

lY.

THE QUANTITATIVE DETERMINATION OF ARSENIC, BY

THE BERZELIUS-MARSH PROCESS, ESPECIALLY

AS APPLIED TO THE ANALYSIS OF WALL

PAPERS AND FABRICS.

By Charles R. Sangeb.

Presented May 13, 1891.

The original method of Marsh* was published in 1836, and in the
following year Berzelius f proposed the modification which couples
his name with that of Marsh. It seems to have escaped notice, at
least I can find no mention of the fact, that Berzelius also suggested
the quantitative application of the method. He proposed to place
some pieces of copper in the reduction tube, and, after weighing
tube and copper, to heat and pass the arsenical gas through. The
copper would further the reduction of the arseuiuretted hydrogen,
and thus the arsenic might be collected and weighed. Wohler,! in
1861, proposes essentially the same method, but heats the tube in
two places, at the copper which he uses in the form of a spiral, two
inches long, and also just behind the copper, so that whatever gas
escapes decomposition by the first heating may be decomposed by
the copper. The first practical application of this method seems to
have been made by GautierJ in 1876, in the estimation of arsenic
contained m tissues. He omits the copper spiral, but heats the re-
duction tube by a layer of charcoal 20 to 25 cm. in length. To de-
termine the arsenic deposited, he weighs the tube, dissolves out the
mirror by nitric acid, and, after drying, weighs again. Croramydis,||
a year later, follows Gautier's method in a similar research.

Chittenden and Donaldson, H in 1881, investigated this method
with very satisfactory results, and suggested the improvements which

* Edin. Philos. Journ., XXI. 229.

t Berzelius, Jahresb., XVII. 191. J Mineral-analyse, p. 232.

§ Ann. d'llygiene publ. et de Me'd. legale, 1876, p. 136; also Bull, de la Soc.
Chim., [2.] XXIV. 250.

II Bull, de la Soc. Chim., [2] XXV. .348.

H American Chemical Journal, Vol. II. No. 4 ; Chem. News, XLIII. 21 ;
Moniteur Scient. de Paris, 1881, p. 227.

QUANTITATIVK DETERMINATION OF AKSENIC BY THE BERZELIDS-MARSH PROCESS

STANDAKD MJRRORS.

OP ARTS AND SCIENCES. 25

were necessary for its adoption as a trustworthy analytical method.
Their process has found its way into the text-books, and needs no ex-
planation here. I note its use by Hubbard* in 1882, and Prescottf
in 1886, and undoubtedly many others have employed it in toxicologi-
cal work. It seems remarkable, therefore, that the previous applica-
tion of the process to quantitative work should have escaped the
attention of Kiihn and Saeger,J as well as Polenske,§ to whom they
refer as the first to propose the method. Kiihn and Saeger's article,
published a few months ago, contains nothing new ; but as the paper
of Folenske is inaccessible to me, I cannot say what modification he
may have introduced.

I need not refer here to the numerous quantitative methods which
depend on the reduction by nascent hydrogen and absorption of the
arseniuretted hydrogen by argentic nitrate, the eventual determina-
tion of the arsenic being made from the silver solution in a variety of
ways. These methods form a class by themselves, and cannot be in-
cluded in the Berzelius-Marsh process.

All methods for the estimation of arsenic are open to a common
objection ; they do not allow the estimation of minimal, or even, with
accuracy, of small amounts. It happens so often that a small amount
of arsenic must remain unestimated, because unweighable, and only
an approximation to the real quantity can be made.

In the qualitative analysis of wall papers and fabrics by the Ber-
zelius-Marsh method, much confusion results from the careless man-
ner of reporting the amount of arsenic which makes its appearance in
the reduction tube. The reports, " trace," " small amount," "large
amount," are usually made without reference to any standard mirrors,
time of heating the reduction tube, or, in many cases, to the amount
of substance taken for analysis. On account of the want of a definite
quantitative method which could be easily applied to wall papers and
fabrics, there have been some propositions for a rough quantitative
determination, which should serve as a control as to whether the
substance contained more or less than a prescribed amount. The
Swedish law || of 1883, for instance, though not using the Berzelius-

— * -

* Pliysician and Surgeon, Ann Arbor, IV. 348 ; also, Contributions from the
Chem. Lab. Univ. Mich., Vol. I. Part I. p. 12.
t Chera. News, LIII. 79.
t Ber. d. deutsch. chem. Gesell., XXIII. 1798.
§ Arb. a. d. kais. Gesundheitsamt, Bd. V. Heft 2 (1889).

II Correspondence between the English and other Governments respecting
the Presence of Arsenic ... in Wall Papers and Textile Fabrics. Com"
mercial, No. 40 (1883).

26 PROCEEDINGS OP THE AMERICAN ACADEMY

Marsh method, prescribes that " 440 sq. cm. of the article ... by
reductiou with potassic cyauide and sodic carbonate, shall not produce
more than a partially opaque mirror in a glass tube of 1.5 to 2 mm.
inner diameter." Thoms * in 1883 proposes as a means of control
that the results from 100 sq. cm. of paper should be divided into
four grades : " strongly arsenical,"' " arsenical," " traces," and " free."
If, when the apparatus has been running ten minutes after the intro-
duction of the solution to be tested, a deposit is obtained no larger
than that corresponding to what is produced by 0.1 mgr. of arsenious
oxide under similar conditions, the paper may be considered to con-
tain a " trace," and need not be rejected.

A committee of the National Health Association of Great Britain,
consisting of Messrs. Bartlett, Heisch, and De Chaumout,! suggested,
in 1883, that a paper should be considered non-arsenical if, after being
treated by a modification of the Berzelius-Marsh method devised by
them, it failed to give a mirror in a tube of one eighth inch internal
diameter (about 3.3 mm.) sufficient to cut off at any point a black line
of a certain thickness (" thick rule, 8 to pica ") on a white ground.

All this is very crude, yet, without a means of easily estimating
the amount of arsenic present, it might answer until the exact de-
termination was called for.

The length of time necessary for any of the quantitative methods
precludes their use by analysts, especially when, as is generally the
case, the quantitative determination is not of especial importance.
If we attempt to apply the gravimetric Berzelius-Marsh method to
the analysis of wall paper, we are met, not only by the amount of
time necessary for the complete deposition of the arsenic mirror,
but by the large amount of paper that must be taken, or, if the
proportion of arsenic is very small, the unwieldy amount. Added
to this is the necessity for a delicate balance, and also the error in
weishinff small mirrors of arsenic.

A method is therefore desirable which will allow us to estimate
minimal amounts of arsenic, and, in such analyses as that of wall
paper, will give an approximate idea of the amount present without
requiring more time than that needed for the proper -conduct of the
ordinary Berzelius-Marsh method.

The process which is described in the following pages was sug-
gested by Professor H. B. Hill of Cambridge, about five years ago,

* Ref. Fres. Zeitschr., XXII. 474, from Ber. d. landw. chem. Vers. u. Samen-
controlstat. zu Riga, 1883.
t Brit. Med. Jour., 1883, p. 1218.

OP ARTS AND SCIENCES. 27

and has been in use in that hiboratory and others with excellent suc-
cess. The proof of the availability of the method was undertaken by
me, but, owing to numerous interruptions, the completion of the ana-
lytical work has been delayed until now.

The method consists, briefly, after getting the arsenic from a
measured amount of paper or fabric into solution, in the comparison
of the mirror obtained from an aliquot part of the solution with a
series of standard mirrors obtained from known amounts of arsenious
oxide. No method founded on exactly this principle has ever been
fully described, though Otto * gives cuts of mirrors obtained from
known amounts of arsenious oxide, with which some analysts may
have compared their mirrors. Selmi,t in 1880, states that he is
able to approximate to fractions of a milligram by comparing the
mirrors with those obtained from the following amounts : one twen-
tieth, one fiftieth, one hundredth, and one two-hundredth part of a
milligram. Thoms, as stated above, compares his " traces " with
a mirror obtained from one tenth of a milligram, and says that the
mirror can be kept any length of time as a comparison standard.
Blyth $ also suggests a comparison of mirrors.

I give in detail the method as I have used it in the analysis of wall
paper, making references to the analytical and experimental work
which follows.

The measurement of the paper is governed by (a) the quantity
of arsenic present, which may sometimes be judged by the color,
or by the rough test of the odor from the burning paper, and
(h) by the character of the paper ; i. e. whether a plain color, a
small or large figure. I have used 25 sq. cm., 50 sq. cm., and usually
100 sq. cm. As patterns for cutting, thin plates of glass § may be
used, on which are marked the dimensions. The advantage of the
glass is, that the figure of the paper may be seen while the paper is
being cut, and also, that, by washing or wiping the glass after each
cutting, any danger of contamination by adhering particles from a
previous arsenical paper may be avoided. I have used but one plate
for 25 sq. cm. (5x5); for 50 sq. cm., three (5 X 10, 4 X 12.5, and
2 X 25) ; and for 100 sq. cm., five (10 X 10, 5 X 20, 4 x 25, 8 X 12,5,
and 2 X 50) ; such a number allowing the variety in cutting that dif-
ferent papers necessitate.

* Graham-Otto-Michaelis, Lehrbuch, II. 2, 520.
t Gazz. Cliim. Ital., X. 435.

X Poisons, tlieir Effect and Detection, 1884, p. 634.
§ Dr. Charles Harrington.

28 PROCEEDINGS OP THE AMERICAN ACADEMY

The paper, cut into small pieces, is placed in a glazed porcelain
dish and moistened with 1 to 5 c. c, according to the amount of
paper taken, of strong sulphuric acid (sp. gr. 1.8) to which has been
added about one thirtieth of its volume of strong nitric acid (Appen-
dix, 2. e). The paper and acid are stirred with a thick glass rod
until the paper has absorbed the acid, and the dish is then placed on
a ring and heated by a low flame, the mixture being stirred continu-
ally, until the paper is thoroughly charred (App., 1). This may be
determined by the dry " crumbly " appearance, and by the amount
of fuming, it being necessary to heat until the nitric acid is expelled.
Usually, the first heating is enough, as, even if a trace of nitric acid
is left, it does no harm (App., 2./) ; yet, if one has reason to think
that considerable nitric acid is held back, it is best, after cooling, to
add a few cubic centimeters of water, and heat again until the fumes
of sulphuric acid appear. On cooling, the " char " is moistened with
a few drops of water, and then about 5 c. c. of water are added. The
mass is triturated with the thick rod until all lumps are thoroughly
broken up, heated to boiling to expel sulphur dioxide (App., 5. d) and
filtered hot (App., 3).

Filtration of the Extract. — In filtering, time is saved and greater
accuracy assured by using a filter pump, and for filtration I have
found most convenient a sideneck test tube of 25 to 30 c. c. capacity.
This is fitted with a rubber cork through which passes a small funnel,
the end reaching just below the side tubulus of the test tube. With
the tubulus is connected the pump, and in order to prevent accidental
contamination of the tubulus, (through which the extract is afterwards
poured into the Marsh apparatus,) I use a connecting glass tube,
which may be rinsed before and after each filtration. This tube is a
small U tube not over 25 cm. in total length, with two bulbs blown
in the lower part of the U. The filter paper is of small diameter,
not over 8 cm. and the lot should be tested for arsenic previous to
cutting the round filters. As a strengthening cone at the apex will
be found convenient for this, as well as for other filtrations, a square
piece of cheese cloth laid under the paper and folded with it.

After filtering, the char is washed with small quantities of hot
water until the filtrate and washings fill the test tube. The extract,
after cooling, is ready for weighing or measuring, and introduction into
the apparatus.

The Apparatus. — Two points will have been noticed in working
with the ordinary form of generating flask : the time required for dis-
placement of the air, and the impossibility of governing the evolution

OF ARTS AND SCIENCES. 29

of hydrogen during the analysis. The latter objection is partially
remedied by some such contrivance as that of Lehmann* or of Blond-
lotjt both of whom regulate the current bj' raising or lowering the
zinc by a glass rod working through the cork of the flask. Chitten-
den and Donaldson J regulate the evolution by the successive use of
acids of increasing strength. Both difficulties are, however, obviated
by the use of a constant hydrogen generator by which the air of the
flask may be swept out and the flow of the hydrogen controlled,
thus assuring the uniform rate of deposition of the arsenic, on which
the success of the process as a quantitative one largely depends.
This idea of a constant generator was apparently first proposed by
Verryken,§ and has been used also in 1888 by Wolff || in a modifi-
cation of Bloxam's electrolytic method. *[[ Any form of generator can
of course be used. To the delivery tube is attached a distributing
tube, which may be two-way or three-way, according to the number of
reduction flasks used. For two flasks an ordinary Y tube suffices,
each end of the Y being fitted with a thick rubber connecting tube
and a screw clamp, so that the supply of hydrogen may be shut off
entirely or controlled for each flask. The reduction flask is a wide-
mouthed bottle of about 75 c. c. capacity, fitted with a rubber stopper
pierced with three holes. Through one hole passes a right-angled
tube reaching to the bottom of the flask, the other end being con-
nected with the distribution tube of the generator. Through the
second hole passes the right-angled delivery tube, reaching just below
the rubber cork. The third hole serves for the introduction of acid
and extract. Through it passes to the bottom of the flask a tube
with the bore at the lower end somewhat reduced by melting.** In
the top of this tube is set a very small funnel.

To the delivery tube of the reduction flask is attached by a rubber
stopper a straight bulb drying tube filled with fused calcic chloride
(App., 5. a), and to the drying tube is connected by a short thick
rubber tube the reduction tube, which should be of the hardest and
best quality of Bohemian glass (App., 5. ^), and of as uniform bore as

* Pharm. Zeitschr. f. Russland, 1888, XXVII. 193.
t Memoires de la Soc. Roy. de Sci., Lett, et Arts de Nancy, 1845.
i Loc. cit.

§ Ref. by DragendorfE, Ermittelung von Giften, 1876, pp. 337 and 317, to Jour,
de Pharm. d'Anvers, 1872, pp. 193 and 241.
II Fresen. Zeitsclir., XXVII. 125.
TF Ibid.; also Biyth, loc. cit., p. 533.
** That the fluid introduced may not carry any air with it into the flask.

30 PROCEEDINGS OF THE AMERICAN ACADEMY

possible, about 7 mm. inside and 8 mm. outside diameter. It should
be drawn out before a small blast-lamp flame to as nearly as possible
definite bore, which ought to be from 1.5 mm. to 2 mm. at the place of
deposition of the arsenic mirror. The finely drawn tube is bent
slightly upward, and sealed at the end.

The Reagents (App., 4). — The zinc and sulphuric acid used in the
apparatus must be strictly free from arsenic, and one should not feel
satisfied of their purity unless a stream of hydrogen from the genera-
tor, led through the heated tube for several hours, fails to give the
slightest deposit. The granulated zinc used in the generator is best
of comparatively large size, while that used in the reduction flasks
should be quite fine-grained. The acid should have a concentration
of one part strong acid (1.82 sp. gr.) to eight parts of water, though
a more dilute acid can often be used.

The Course of Analysis. — In the reduction flask is placed a small
quantity, not over three grams, of zinc, and the apparatus is con-
nected together. Tightness must be assured, and can be tested for, if
the tip of the reduction tube is sealed, by adding through the small
funnel a few drops of acid. If these do not fall, the tip of the tube is
broken off, leaving an opening of not more than 1 mm. diameter, and
then about 20 c. c. of acid are added. The hydroiiea is now turned
on from the generator, aud, after expulsion of the air, lighted, and the
flame turned down to a height of 1 to 2 mm. The evolution of
hydrogen should be kept at this rate during the reduction of the
arsenical solution. Often it is not necessary to use the generator
during the reduction, as the evolution from the reduction flask is
sufficiently rapid. If it slackens, the generator may be used again,
and indeed it is generally necessary to use it toward the close of
the reduction. It may happen, on account of too strong acid or
increased action after introduction of the extract, that the evolution
is too rapid, and the flask becomes heated (App., 5. c). To obviate
this, the flask may be set in a vessel fitted with an exit tubulus,
and filled with cold water, which can be drawn off" and replaced
when necessary without disturbing the apparatus. An ordinary
crystallizing dish with a siphon would answer the purpose.

Shortly after (App., 5. e) lighting the hydrogen, the lamp is
placed under the heating place and the apparatus tested for absence
of arsenic for such length of time as the circumstances direct. In or-
dinary analyses of wall paper, I allow 15 to 20 minutes' free run
before adding the extract.

The lamp should give a large clearly defined flame (App., 5./),

OF ARTS AND SCIENCES. 31

and should heat the tube with its oxidizing flame only. An iron cone
may be used for increased draught, but not an iron or glass cylinder
for the greater radius of heat given by the latter tends to throw the
mirror farther along the tube, and to deposit it irregularly. A con-
venient rest for the reduction tube is made by soldering three stout
copper wires to an old binding screw or post, curving the ends to fit
the tube and branching them out, so that the tube lies flat in the
curved ends. The binding screw travels vertically on a brass rod
melted into a flat, heavy piece of lead, or screwed into tlie base of an
old Bunsen burner.

When the apparatus is found free from arsenic, the extract is added.
Previous to tliis it has been measured or weighed, preferably the
latter. The test tube having its weight marked on it, the weight
of the extract is quickly determined. Weighings can be made on a
balance sensitive to fifty milligrams, which is enough for all practical
purposes.

A few drops of the extract are at first added. If no mirror appears
in three or four minutes, one eighth to one quarter of the rest may be
added, and if in five minutes more there is no mirror, the whole of the
extract may be introduced. This cautious addition is necessary in
order not to obtain too large a mirror, else a difficulty in estimation
might arise, or a new determination might have to be made. A
twenty-five minute run is sufficient for the deposition of all the arsenic
when the size of the mirror formed in the first fifteen minutes is not
larger than that corresponding to 0.05 mgr. of arsenious oxide. If
the mirror forming is likely to be larger than this, it is better, after
weighing, to start another mirror with another portion of the extract,
than to wait for the complete deposition of a mirror which may be
too large for comparison with the standards.

The set of standard mirrors is made as follows. One gram of
arsenious oxide, purified by repeated sublimation, is dissolved with
the aid of a little sodic bicarbonate (free from arsenic), and, after
acidification with dilute sulphuric acid, is made up to a litre. Of
this standard solution (I.), containing 1 mgr. AS2O3 to 1 c. c, ten
c. c. are taken and made up to a litre, giving the standard solution
(II.) containing 0.01 mgr. to 1 c. c. Of this solution, 1 c. c, 2 c. c,
3 c. c, etc., are carefully measured from a burette and introduced into
the reduction flask of the apparatus, giving the mirrors corresponding
to the same number of hundredth-milligrams. It is necessary to
make two or more mirrors of the lower amounts, as, even with care-
ful drawing, the cross sections of the deposition tubes differ, so that

o2 PROCEEDINGS OP THE AMERICAN ACADEMY

the appearance of the mirror from the same amount of arsenic
varies, and the variation is more marked as the mirrors decrease in
size. Then, for greater convenience in interpolation, mirrors cor-
responding to the half-hundredths may be made, so that the set which
I use contains the following amounts. 0.005, 0.01, O.Ol.o, 0.02, 0.025,
0.03, 0.035, 0.01, 0.045, 0.05, and 0.06. Above 0.06 the difference
between the mirrors is so hard to estimate with accuracy, that it is of
no advantage to make any above 0.06. Yet in the set which I have
photographed, and which is shown in the plate, the following are added
for comparison : 0.07, 0.08, 0.09, 0.10.

When the set is not in use, it should be kept in the dark, and at
no time be exposed to direct sunlight. Although the tubes be sealed,
yet there is sufficient air inside to cause a rapid oxidation in direct
sunlight, especially with the lower amounts, and a film of arsenious
oxide results. By observing this precaution, the set can be kept a
long time without alteration.

The readings of the mirrors are far sharper by transmitted than
by retlected light, and comparison should always be made by the
former. For this purpose I use, at the suggestion of Professor Hill,
a small tin box painted black on the inside and outside, and similar
in shape and construction to the old-fjishioned stereoscopic camera.
The height of the box is 20 cm. ; width, 7.5 cm. ^ length at top,
10 cm. ; length at bottom, 15 cm. The bottom is open, while the top
is covered, and fitted with two eyeholes, 2 cm. in diameter and 4 cm.
apart. The set of standard mirrors, as shown in the plate, is
mounted, by means of sealing wax, in blackened wooden frames,
18 cm. long by about 7 cm. wide, and 4 to 5 mm. thick. The inner
dimensions of the frames are 12 cm. by 4.5 cm. The bottom of the
box carries a rabbet of tin, on which the frames may be slid, thus
bringing the mirrors under the eyehole.-^. A second rabbet above
the first allows the introduction of a frame carrying the mirror to
be compared, which may thus be brought between any two of the
standards and compared just as in nesslerization. The box is
mounted over a white paper or plate, in front of a good light.
The calculation of the amount of arsenic in the area 0^ paper taken
follows from the determination of the amount in the aliquot part of
the solution. From this can be found the number of milligrams
per square meter, which, when multiplied by the factor, 0.0128, gives
the number of grains per square yard.

OF ARTS AND SCIENCES. 33

Appendix. — Analytical and Experimental Work.

The experimental work in the investigation of the availability of
the method is given in detail in the following pages.

1. Necessity of Charring. — Considerable. difference of opinion ex-
ists as tx) the necessity for insuring the absence of organic matter in
the extract to be tested. Odling,* in 1859, in testing for arsenic
in tissues, found that he could obtain arsenic by Reinsch's test in the
presence of organic matter, but not by Marsh's. After getting rid
of the organic matter by distillation with hydrochloric acid, he had
no difficulty with the Marsh test. Blondlot f calls attention to the
necessity of completely destroying the organic matter. Chittenden
and Donaldson, t on the other hand, were able, in the presence of or-
ganic matter, to detect very small amounts of arsenic, and consider it
of no hindrance. My experience has been, that, while the organic
matter may not completely prevent the deposition of small amounts,
and is no hindrance in the detection of large amounts, yet the char-
acter of the mirror is so altered that comparison with the standards is
impossible. Besides, the organic matter causes an increased and
irregular flow of gas which tends (App., -">./) to carry undecomposed
arseniuretted hydrogen out of the tube. I therefore take the precau-
tion, by thoi-oughly charring, to insure the absence of organic matter
in the extract.

Thorns § digests 100 sq. cm. of the paper on the water bath with
(1-7) sulphuric acid, and adds the filtrate directly to the reduction
flask. Fleck || considers that digestion with a 25 per cent sulphuric
acid is sufficient to extract the arsenic completely, and ReichardtlT
concurs in this opinion. I cannot agree with this, and consider that
the chance of the ai'senic being held as arsenious sulphide is alone
enough to condemn the method (App., 2. c), not to speak of the
organic matter extracted.

2. Use of an Oxidizing Agent in Charring. — In case the paper con-
tains chlorides, it seemed likely that a portion of the arsenic would
be volatilized during the treatment with sulphuric acid. This was

* Guy's Hospital Reports, V. 367-374.
t Loc. cit. X Loc. cit. § Loc. cit.

II Rep. analyt. Chem., 1883, Heft 2.
1 Archiv d. Pharm., CCXXI. 271.

VOL. XXVI. (n. 8. XVIIl.) 3

34 PROCEEDINGS OF THE AMERICAN ACADEMY

found to be true to an extent sufficient to warrant the precaution of
adding a small quantity of nitric acid to the sulphuric acid, iu order
to prevent the formation of arsenious chloride.

a. The Presence of Chlorides in Wall Papers. — In order to deter-
mine to what extent chlorides exist in wall papers, several analyses
were made of papers taken at random. The method was as follows.
A measured amount of paper (400 sq. cm.) was thoroughly mois-
tened with a strong solution of potassic nitrate (free from chlorine)
on a broad porcelain plate, dried on the plate, and ignited over the
plate. The organic matter was practically destroyed. The residue
was transferred from the plate to a chlorine free filter, and washed
thoroughly with hot water. The filtrate, after acidification with
dilute sulphuric acid, was boiled to expel nitrous acid, and, after
cooling, titrated according to Volhard with approximately hundredth-
normal solutions. There is no danger of volatization of hydrochlo-
ric acid even on prolonged boiling, as Gooch and Mar* have
shown that a solution containing 0.12 per cent of hydrochloric acid
can be boiled down one half without appreciable loss. In titrating,
as well as in determining the strength of the standard solutions, it
was found that the dilution of the solution affected the end reac-
tion in that the color of the ferric sulpho-cyanate was obscured by
that of the argentic chloride, even if the latter were made to
" clump together." This difficulty was easily obviated by heating to
boiling after adding excess of argentic nitrate, filtering off the
chloride, and titrating back in the cooled filtrate. The end reaction
was then as sharp as possible.

Analysis of thirteen papers gave an average of 138 milligrams of
chlorine per square meter, or 1.38 mgr. in the usual amount (100
sq. cm.) taken for the determination of arsenic. The higliest amount
was 238 mgr., the lowest 56 mgr. In no case was a paper found
free from chlorides.

h. Effect of Chlorides. — A solution of sodic chloride free frorn ar-
senic was made up of such strength that 1 c. c. contained 1 mgr.
chlorine. Several trials were made of the effect of a known amount
of chloride on small amounts of arsenic. 100 sq. cm. S. and S. filter
paper (589) were used in each case. After addition of the arsenic
and sodic chloride solutions, the paper was dried in the evaporating
dish, and the arsenic determined exactly as described, but without
using nitric acid. Tiie following table shows the results obtained.

* Amer. Jour. Science, [3 ] XXXIX. 2U3.

OP ARTS AND SCIENCES.

35

Analyses 1-6 were made with duplicate readings which agreed
closely.

Mgr. AsjOg

C.c. NaCl, mgr. CI,

Mgr- As-jOj

Per Cent A 8^0.

taken.

taken.

recovered.

recovered.

1 . .

. 10.0

20.0

6.22

62.2

2 . .

. 10.0

20.0

4.96

49.6

3 . .

. 5.0

4.4

3.61

72.2

4 . .

. 0.5

2.0

0.44

88.0

5 . .

. 0.1

2.0

0.87

87.0

6 . . .

. 0.05

2.0

0.02

40.0

7 . .

. 0.01

2.0

0.005

50.0

Analyses 1 and 2 show that the presence of twice the theory of
chlorine causes a large loss, with an amount of arsenic comparatively
large. Analysis 3 shows a marked loss with less than the theory.
Analyses 4 to 7 were made with more chlorine than the average found
(1.38 mgr.), but less than the greatest amount (2 38 mgr.); the
amounts are respectively 4, 20, 40, and 200 times the theory, and the
conditions are more nearly those met with in practice than in analyses
1 to 3. It will readily be seen that the loss is sufficient to warrant the
precaution of using an oxidizing agent.

c. Loss by Retention of Arsenic in the Char as Arsenioiis Sulphide.
— The presence of the sulphide in papers is not very common. Yet
papers colored with ultramarine occur frequently, and the sulphu-
retted hydrogen set free from this by the sulphuric acid acts on the
arsenic compound, and thus a large part of the arsenic remains in the
char as the sulphide, insoluble in dilute sulphuric acid. The follow-
ing analyses of a paper are sufficient to show the danger of loss from
this .source. The paper had a light blue ground, and the odor of
sulphuretted hydrogen was apparent on addition of sulphuric acid.
100 sq. cm. charred with sulphuric acid and a few drops of nitric acid
gave aa amount of arsenic corresponding to 5.9 mgr. per square
meter. 100 sq. cm. charred with sulphuric acid alone gave 4.5 mgr.
per square meter. The char was then extracted with ammonia, and
the extract evaporated with sulphuric acid and a drop of nitric acid.
This yielded 0.01 mgr., corresponding to 1 mgr. per square meter
and bringing the total amount up to 5.5 mgr., which agrees with the
first analysis.

d. The Deposition of Arsenic from a Solution of Arsenic Acid. —
The question arises whether the conversion of the arsenious to arsenic
acid by use of an oxidizing agent would cause the arsenic to be held
back. This was quickly determined by comparison of the standard
mirrors with a set prepared from a solution of arsenic acid, of which

36 PROCEEDINGS OP THE AMERICAN ACADEMY

1 c. c. contained 0.01 mgr. AsgOs, as AsoOs- The mirrors agreed
sharply. This is at variance with the results of Headden and Sadler,*
who found, in using the method of Chittenden and Donaldson, that
it was necessary to subject the arsenic acid to preliminary reduction
in order to obtain all the arsenic present. The reason for this is,
probably, that a vei'y small quantity of arsenic acid, such as would
occur under the conditions of this method, is more quickly reduced
by nascent hydrogen, while the comparatively large quantity used by
Headden aud Sadler would take more time. They do not show that
prolonged treatment in the Marsh reduction flask would not have
eliminated this error.

e. The Use of Nitric Acid as an Oxidizer. — Nitric acid suggests
itself at once as the most convenient oxidizing agent. In this con-
nection may be mentioned the method proposed by Blyth,| which
consists in soaking the paper in potassic chlorate, drying, burning, and
extracting with water. Hager % substitutes sodic nitrate for the po-
tassic chlorate. Many destroy the organic matter by hydrochloric acid
and potassic chlorate, as in the case of tissues. Lyttkeus § uses sul-
phuric acid and potassic chlorate, and Lenz,]| commenting on this, con-
siders it the best means of treatment of the paper. The German
law IT of 1888 prescribes the treatment of fabrics with strong hy-
drochloric acid aud distillation with ferrous chloride. None of these
methods in the analysis of wall papers and fabrics have any advantage
over charring with the simple addition of nitric acid, and all require
more time.

That all the arsenic may be recovered when nitric acid is used is
shown by the following: 0.01 mgr. arsenious oxide was added to 100
sq. cm. of filter paper and a drop of strong nitric acid added before
charring. The mirror obtained corresponded to 0.01 mgr.

1 mgr. AS2O3 and 5 cc. sodic chloride solution (5 mgr. chlorine)
were added to 100 sq. cm. filter paper and treated with a mixture of
one part nitric to six parts sulphuric acid. Duplicate readings gave
1.095 mgr recovered, an error no greater than that wliich might
occur from the method.

10 mgr. As203,as AS2O5, were added to 100 sq. cm. filter paper, and

* Amor. Cliem. Journal, VII. 338; Ber. d. deutsch. Chem. Gesell, XIX. 116.

t Loc. cit , p. 632.

I Pliarm. Centrallialle, XIII. 145.

§ Ref. Frcs. Zeitsclir., XXII. 147, from Landw. Versuchsstat., XXVI. 305.

lIFres. Zeitsohr.. XXII. 147.

If Ibid., XXVII. 471.

M(fr. AsjOg
taken.

Mgr. A8.JO3
fouud.

Mgr. per Sq.
Meter taken.

Mgr. per Sq.
Meter found.

0.01

0.009

1.0

0.90

0.10

0.0995

10.0

9.95

1.00

0 977

100.0

9770

10.00

10.30

1000.0

1030.00

OP ARTS AND SCIENCES. 37

treated with sulphuric acid alone. Duplicate readings gave 9.94 mf^r.
recovered.

Finally, 100 sq. cm. of a wall paper containing no arsenic was,
after addition of varying amounts of AsgOa, charred with a nitrosul-
phuric acid containing one part nitric to thirty of sulphuric acid.
The following table shows the results of four analyses. The readings
on 2, 3, and 4 were in duplicate.

1 . .

2 . .

3 . .

4 . .

/. Effect of Free Nitric Acid in the Reduction Flask. — Rieckher*
does not consider the presence of free nitric acid to be detrimental, while
Fresenius t takes the opposite ground. Blondlot :j: thinks that free
nitric acid gives rise to the formation of a solid hydride of arsenic on
the zinc, thus causing the retention of part of the arsenic. Without
discussing the correctness of this statement, which seems to have been
quite universally accepted, we have only to consider the effect of a
trace of the acid on the deposition of the mirror, as the method of
charring would leave, at most, but a very small amount of free
nitric acid in the extract. 0.02 mgr. and 0.05 mgr. AsoO^ were added
to the reduction flasks in which were about 20 c. c. dilute sulphuric
acid, and immediately afterward a drop of strong nitric acid was
added to each. The dilution was considerable, but not so great as
might occur in practice. The mirrors were clearly defined, and cor-
responded sharply to 0.02 mgr. and 0.05 mgr. respectively.

3. Extraction of the " Char." — It is necessary that the char should
be thoroughly pulverized and extracted with hot water. Con.siderable
loss is likely to occur if these precautions are disregarded, as the fol-
lowing results show: —

To 100 sq. cm. of a wall paper free from arsenic, 1 mgr. AS2O3 was
added, and the char was extracted with 30 c. c. cold water. 0.77 mgr.
were recovered, = 77 per cent.

To the same amount of paper 5 mgr. were added, and the char ex-
tracted with 30 c. c. cold water. 3.82 mgr. were recovered, = 7G.4
per cent.

* Neues Jalirb. f. Pharm., XX. 3. f Comptes Rendus, LVII. 596.

t Free. Zeitschr., II. 3b9.

i38 PROCEEDINGS OP THE AMERICAN ACADEMY

To the same amount of paper 1 0 mgr. were added, and the char
extracted as before, but, in addition, the particles were well triturated.
9.19 mgr. were recovered, = 91.9 per cent.

That this loss was not due to error in estimation of the mirrors was
shown by repetition of the first two trials, mirrors being obtained from
separate portions of the same extract. The char was triturated in
each case : —

1 mgr. AS2O3 gave 0.9 mgr. and 0.9 mgr. Average, 90%.
• 5 " " " 2.92 " " 2.95 " " 58.00%.

That arsenic would be left in the char was shown as follows : 100
sq. cm. of paper as above, after addition of 5 mgr. AS2O3, were charred,
and the finely ground mass extracted first with cold water. 3.11
mgr. were recovered, — 62.2%. The residue was then extracted with
30 c. c. hot water. 1.63 mgr. were recovered, = 32.6%. A third
extraction with hot water gave a solution free from arsenic. Total
amount recovered, = 4.74 mgr. = 94.8%.

The quality of the paper has no effect on the loss : 5 mgr. AsoOs
were added to 100 sq. cm. filter paper and treated as before. Two
readings from the same cold extract gave 3.51 and 3.46 mgr. respect-
ively. Average, 69.9%. Finally, to show the practical completeness
of extraction with only 25 to 30 c. c. hot water, 100 sq. cm. of filter
paper were charred with different amounts of arsenious oxide.

1

2
3

10 mgr. of arsenious oxide per 100 sq. cm. of paper would corre-
spond to 1,000 mgr. per sq. meter, which is an unusual amount. Hence
any smaller amounts would be easily extracted. It should be borne
in mind that the error is necessarily great in such large amounts, for
the deposit which is compared with the standards is so small a propor-
tion of the total amount that a slight difference in reading is propor-
tionally increased.

4. Reagenfs. — Many chemists prefer to use hydrochloric acid in-
stead of sulphuric acid, on account of the quicker action of the former
on the zinc. Opinions vary considerably as to the error arising from
volatizatiou of zincic chloride, and consequent deposition at the heating

Mgr. AS2O3
takea.

Mgr,
Extract 1st.

. As„03

2a.

recovered.

3a.

Total.

Per Cent As^Oj
recovered.

5.0

499

0.08

0.0

507

101.4

5.0

4.98

0.07

0.0

5.05

1010

10.1

10 65

0.06

0.0

10.71

106 7

10.0

10.37

0.0

—

10.37

1U3.7

OF ARTS AND SCIENCES. 39

place, when hydrochloric acid is used. I.if\)ig,* very soon after the
publication of Marsh's method, called atteniiou to possible error from
this source, and several years later Wackeurodert coulirmed Liclii"',s
opinion. Beckurts t claims that there is no danger to be feared, but
does not show conclusively that this is the case. Brescius § recoof-
nizes the chance of error, and recommends passing the gas througli
sulphuric acid, if hydrochloric acid is used for generation. The com-
mittee of the British Health Association, above referred to, recom-
mends the use of hydrochloric acid, without comment on its possible
disadvantage.

While the volatilization of zincic chloride might not interfere with
the detection of arsenic in considerable quantity, yet it is of primary
importance in this method for the mirror to be of arsenic alone, and
hence capable of comparison with standards. We cannot, therefore,
run the risk of the small mirror being contaminated by any impu-
rity whatever. For this reason, if for no other, the use of hydro-
chloric acid is wholly unadvisable. Then the time gained in using
hydrochloric acid is not to be considered in this method, as the use
of a constant generator reduces the time of analysis so decidedly.

The same desire to hasten the evolution of hydrogen in the ordi-
nary Marsh process has led to the addition of stimulants to the action
in the shape of platinic chloride or cupric sulphate. Bernstein || has
shown that the use of platinic chloride is inadmissible, because arsenic
is thereby held back. At the same time, however, he finds no loss
when the zinc is platinized or silvered before being used. Headden
and Sadler 1[ agree with Bernstein in the case of platinic chloride, and
find that cupric sulphate also causes a loss. They also get low re-
sults by using a spiral of platinum wire in contact with the zinc.
Here, again, the use of the constant generator precludes the neces-
sity for increasing the sensitiveness of the zinc.

Mohr,** in 1837, called attention to the fact that the residual zinc
even after careful washinjr contained arsenic. This would seem to
substantiate the statement of Blondlot, quoted above, concerning the
solid hydride of arsenic. Fresenius tt ^ilso considers that the effect of
nitric acid is due to the formation of a hydride. When we consider
the case of the mere detection of arsenic by the Marsh process, where
all the arsenic is not necessarily reduced to arseniuretted hydrogen, it

* Ann. d. Chem. u. Fharm., XXIII. 217. || Inaug. Dissertation, Rostock, 1870.
t Archiv f. Pharm., LXX. 14. 1 Loc. cit.

t Ibid., CCXXII. 653. ** Ann. d. Pharm., XXIII. 217.

§ Dingl. Polyt. Jour., CLXXXVI. 226. tt Loc. cit.

40 PROCEEDINGS OP THE AMERICAN ACADEMY

is quite possible that part is left ia the ziuc, especially if the extract
be not free from nitric acid. But in this method, where all of a very
small amount of arsenic is to be reduced, and the action is pushed as
far as possible, the probability is that no arsenic is Jeftin the zinc. I
have often, after deposition of the mirror, pushed the action as far as
the complete solution of the zinc, and have never observed any in-
crease of the mirror at the close. It would seem hardly possible for
the solid hydride to remain in that state in a reduciug medium for
such a length of time (compare also 2./). A large amount of carbon
in the zinc is apt to cause too rapid evolution, but I cannot confirm
the statement of Headden and Sadler * that zinc containing carbon
causes a loss of arsenic, and that the zinc must in consequence be free
from carbon.

5. General Precautions. — a. Means of drying the Hydrogen. —
The use of sulphuric a.cid is not allowable. I have found the statement
of DragendorfE f to be true, that sulphuric acid absorbs arseniuretted
hydrogen. This is assumed byJanowskyJ to be due to decomposi-
tion into arsenic and hydrogen, but no proof is given. DragendorfFf
quotes the suggestion of Otto,§ that a small amount of fused potassic
hydroxide be placed before the fused calcic chloride, in order to absorb
any sulphuric acid which might be carried over, it being possible that
the acid might act on the calcic chloride, giving hydrochloric acid,
which might form arsenious chloride and thus cause loss of arsenic.
The potassic hydroxide would also hold back sulphuretted hydrogen.
It is well known that potassic hydroxide absorbs antimoniuretted hy-
drogen, and it has been recently shown by Kuhn and Saeger,|| as well
as by Headden and Sadler,* that arseniuretted hydrogen is also ab-
sorbed by it. There is, however, no need of its use as a precaution
against either of the contingencies mentioned above.

LyttkensIF considers sulphuric acid to be a better drying agent than
calcic chloride, and Lenz ** agrees with him, but neither shows that
there is no loss of arsenic attending ics use.

As to Headden and Sadler's* statement, unsupported by analysis of'
the calcic chloride used, that fused calcic chloride holds back arsenic
when moist, I have never met with any indication that there was
danger of loss from this source.

* Loc. cit. II Log. cit.

t Ermittehing von Giften, 2te Aufl., p. 336. 1 Loc. cit.

X Ber. d. deutsch. chem. Gesell., VI. 216. ** Loc. cit.
§ Ausmittelung der Gifte.

OP ARTS AND SCIENCES. 41

b. Impurities in the Glass of the Deposition Tube. — The errors le-
sulting from the presence of lead and arseuic in glass have been fre-
quently mentioned. The presence of lead in any hard glass fit for
use is scarcely probable. The formation of a mirror from either of
these sources would however be detected at the start, and the glass
rejected at once. 1 have not found any case of error attributable
to impurities in the glass.

c. Temperature of the Bediiction Flask. — Dragendorflf* quotes
Kolbet as having shown that sulphuretted hydrogen is always
formed by the action of sulphuric acid on zinc, if the reaction tem-
perature exceeds 30°, and recommends on this account that the flask
be cooled. The quotation is misleading. What Kolbe showed, and
Fordos and Gclis % called attention to the same point some time
before Kolbe, was, that if strong acid be added to the flask to in-
cf-ease the action, there was reduction of the acid at the temperature
mentioned. With acid diluted with two parts of water no reduction
took place. As one would hardly add strong acid to the flask, the
danger of formation of sulphuretted hydrogen from mere action of the
acid on the zinc is not to be feared. Yet it is necessary to keep
the flask cool in some such manner as suggested, if only to prevent
too violent action of the acid which may occur from accidental pres-
ence of organic matter.

d. Necessity of boiling the Char with Water. — The formation of a
" sulphur mirror " is often met with, due to the decomposition of sul-
phuretted hydrogen at the heating place. Then, too, Brunn § has
lately shown that sulphuretted hydrogen and arseniuretted hydrogen
form, when heated, hydrogen and arsenious sulphide. The necessity
of boiling- the char with water is therefore evident, in order that no
sulphur dioxide be left in the extract to be reduced by the nascent
hydrogen.

e. Beginning of Heating. — If the tube be heated as soon as the
hydrogen is lighted, the small amount of oxygen left in the flask
causes the formation of water in the deposition tube. This can be
avoided, if desired, by waiting a few minutes before setting the lamp
under the tube.

f Rapidity of Gas flow and Amount of Heat. — In the method of
Chittenden and Donaldson it is necessary to guard against too rapid
evolution of gas, and the heating surface must be very great in order

* Loc. cit., p. 336. t Ann. d. Chem. u. Pharm., OXIX. 174.

X Comptes Rendus, XIII. 437. § Ber. d. deutsch. cliera. Gesell., XXII. 3202.

42 PROCEEDINGS OP THE AMERICAN ACADEMY

that no arseniuretted hydrogen shall escape decomposition. In this
method, the amount of heating surface need not be greater than that
given by a good burner, and I have assured myself by direct experi-
ment tliat no arsenic escapes under the ordinary conditions vk^hen the
size of the mirror is not above O.OG mgr. It is necessary, however,
for the stream of gas to be slow and regular. Hence the disadvantage
of organic matter in the extract, causing an increased and irregular
flow of gas.

6. The following table contains some analyses, taken at random,
illustrative of the method.

Sq. Cm

Grams

Gr. Extract

Mirror obtained

Mgr. AS2O3

Mgr. ASoO;, found

taken.

Extract.

taken.

Mgr. As.Oj.

found.

per Sq. Meter.

1 ,

. . 100

33.17

33.17

0.030

0.030

3.0

2 ,

, . 100

32.27

6.00

0.017

0.094

9.4

6.81

0.015

0.083

8.3

3 ,

, . 100

31.84

5.66

0.045

0.253

25.3

1.90

0.015

0.251

25.1

4 ,

, . 25

29.69

2.64

0.016

0.179

71.9

2.65

0.017

0.191

76.2

5 .

. 111.15

33.66

0.78

0.035

1.510

136.0

0.46

0.020

1.460

132.0

6 .

. 25

30.67

1.01

0.040

1.214

485.6

2.08

0.080

1.180

472.0

1
1 •

. 12

28.83

0.48

0.030

1.802

1502 0

0.61

0.043

2.032

1693.0

7. Comparison of Results obtained hy the Berzelius- Marsh Process
with those obtained by other Methods of Analysis. — In order to test
the availability of the process, it became necessary to analyze, by
one of the general quantitative methods, some of the papers whicli
had been analyzed by the Berzelius-Marsh method. In the analyses
given below, the eventual determination was made by an approxi-
mately hundredth normal solution of iodine, and, when necessary,
titrating back with a sodic thiosulphate solution of corresponding
strength. I find a similar method to have been proposed some time
ago by Holthof.* Considerable dilliculty was met with at first in
finding a suitable method of getting the arsenic from the paper into
proper state for titration. A measured piece of paper (100-400
sq. cm.) was treated on a porcelain plate with a strong solution of

• Fresen. Zeitschr., XXIII. 378.

OP ARTS AND SCIENCES. 43

potassic nitrate, dried on the plate, burued over the plate, and the
residue washed ott* into an evaporating disli. The residue was then
treated with about 5 c. c. strong sulphuric acid aud evaporated until
sulphuric acid fumes appeared. To insure the complete expulsion of
nitric and nitrous acids, the mass was boiled down again, after addi-
tion of a little water. It was then taken up with hot water, hllerod,
and washed with hot water. The filtrate was kept at 60-70° for
half an hour, and sulphur dioxide passed through. After boiling the
reduced solution in the flask until sulphur dioxide was expelled, it was
made alkaline with sodic bicarbonate, cooled, and titrated. The results
obtained were far from satisfactory, but the analyses agreed closely in
three cases, which are given below. (Papers numbered 111, 395, and
42 in final table.)

Next, instead of ignition with potassic nitrate, the paper was
treated exactly as in the Berzelius-Marsh method, with nitrosidphuric
acid (1-30), taking care to get rid of all nitric and nitrous acids. The
extract was reduced at 60-70° by sulphur dioxide, the excess of the
latter driven off by boiling, and the cooled solution made alkaline with
sodic bicarbonate, and titrated. In one case the arsenic was precipi-
tated from the extract by sulphuretted hydrogen, and the arsenious
sulphide converted to arsenious acid, and titrated. The amount of
time required for this, however, off"set any advantage from it. Al-
though by care this method can be employed, yet the details require
much time and the chances for error are many. Duplicate analyses
agreed closely in two cases which are given below. (Papers num-
bered 194 and 155 in final table).

The method finally used was adapted from the well known process
of Schneider* and Fyfe.f A measured quantity of paper was cut
in*o small pieces and placed in a 500 c. c. distilling flask connected
with a cooler. Attached to the latter was a receiver with a second
tubulus carrying a long tube which served as an air cooler. About
100 c. c. of hydrochloric acid, diluted one half, were added to the
flask, and the mixture distilled, slowly, almost to dryness. It was
found by trial, that in nearly every case all the arsenic came over in
one distillation, and, if not, that a mere trace was left in the residue.
The distillate was transferred to a flask, potassic chlorate added, the
solution boiled down one half, transferred to an evaporating dish, and
evaporated to dryness, with the addition of a few drops of strong sul-
phuric acid. The residue was generally white. If dark, from pres-

* Jahrb. d. Chem., 1851, p. 630. t Jour. f. prakt. Chetn., L V. 103.

44

PROCEEDINGS OP THE AMERICAN ACADEMY

ence of volatile organic matter not destroyed by the potassic chlorate,
the addition of a few drops of strong nitric acid and evaporation ex-
pelled the organic matter. The residue was then vpashed into a flask
with about 50 c. c. water, and reduced and titrated as in the previous
cases.

The method was tested by the following analyses of filter paper to
which known amounts of arsenic were added, 200 sq. cm. of paper
beinw used in each case.

Mgr. As.Oa

C.

c. Iodine used.

Mgr. AsoOg

Per Cent AsjOs

taken.

1 c.c.

=0.9815 mgr. AS2O3.

found.

found.

la .
16 .

. . 5
. . 5

5.1

5.3

5.01 )
5.20)

102.1

2a .
26 .

. . 5
. . 5

5.0
5.2

4.91)
5.10 i

100.1

3a .

. . 25

25.1

24.64 )
24.54 )

98.4

36 .

. . 25

25.0

The residues from the distillations in the last analyses were, after
the addition of a few drops of nitric acid, charred with sulphuric acid
and " marshed," giving mirrors corresponding to 0.015 mgr. and 0.02
mgr. respectively, showing that a mere trace was left in the flasks.

The following papers were analyzed by this method.

Number of
Paper.

194 .

194 .

406 .

406 .

39 .

39 .

892 .

393 .

393 .

359 .

359 .

Sq. Cm. C.c. Iodine u.sed.

taken. 1 c.c. =0.!)815 Mgr.ASjO,.

. 200 17.20

. 200 17.30

. 400 3.10

. 400 3.00

. 400 2.80

. 400 270

. 380 1.80

. 400 0.38

. 400 0.36

. 400 130

. 400 1.10

Mgr. A.'SjOj found
per Sq. Meter.

844.00
849.00

76.07

73.61

68.71

66.25

46.50
9.33
8.83

31.90

27.00

Finally, the following table shows the comparison of the results
obtained by the volumetric method with those obtained on the same
papers by the Berzelius-Marsh method. The first two columns com-
pare the results in milligrams per square meter, and the second two
in grains per square yard. Each result is the mean of two, unless
specified.

OF ARTS AND SCIENCES.

45

'Jumher of

Mgr. ASjO;,

per Pq. Meter.

Grains per

Sq

Yard.

I'aper,

Berz. Marsh

Volumetric.

Berz. Marsh.

Volumetric.

393 . . .

8.8

9.1

0.11

0.12

859 .. .

24.0

29.5

0.31

0.38

392 ... .

4(5 8

46.5*

0.60

0.60*

39 . . .

64.9

67.5

0.83

0.88

406 .. .

72.8

74.9

0.94

0.96

Ill ... .

110.4

108.7

1.42

1.39

395t . . .

193.0

210.4

2.47

2.69

42 . . .

4788

421.8

6.13

5.40

194 .. .

842.0

832.7 1

10.78

10.66 1

155 . . .

1527.1

1478.5

19.55

18.93

determination,

Four determinations. } 395

was a piece of '

Turkey red '

Before the above described method was worked out, it was thousrht
that it would give merely an approximation of the amount of arsenic
in wall papers and fabrics, which would allow one, for instance, to pass
judgment on the articles from a sanitary standpoint. Not only, as
will be seen from the table, does the method give an approximation to
the actual amount when ordinarily conducted, but with care it can
be made to give results worthy of comparison with other quantita-
tive methods. The greatest error occurs naturally in the estimation
of large amounts, but in this case an approximation would answer
until a more exact determination was called for.

The process will, I think, be also found of great value in toxico-
logical work, not only as a rapid means of determining the quantity
of ar.senic present, and as a check on other methods, but also as the
only means of accurately determining the amount when the arsenic is
present in minute quantity. In such work the organic matter would
not generally be charred, but the arsenic would be extracted by the
method of distillation. I hope to investigate the extension of the
metliod to toxicological work.

The limit of arsenic that I have been able to detect with certainty by
the Berzelius-Marsh method is 0 001 mgr. As.jOg or 0.0007 mgr. As.
I shall not here take up the much discussed question of the delicacy
of this as compared with other methods, but I think it will be agreed
that no other method enables one to determine quantitatively such
small amounts.

In conclusion, I have to thank my assistant, Mr. Charles Walker,
very sincerely for his valuable services in most of the analytical work
of this paper.

United States Naval Academy, Annapolis, Md.,
February, 1891.

46 PROCEEDINGS OF THE AMERICAN ACADEMY

V.

CONTRIBUTIONS FROM THE CRYPTOGAMIC LABORATORY OF

HARVARD UNIVERSITY.

XV. — ON THE STRUCTURE AND DEVELOPMENT
OF CHOREOCOLAX POLYSIPHONI^, Reinsch.

By Herbert Maule Richards.

Presented by W. G. Farlow, May 12, 1891-

Heretofore comparatively little has been known concerning the
obscure parasitic alga, Choreocolax Pohjsiphonice, Reinsch. Little
has been written concerning it, and, so far as I know, only one other of
the various forms included in the original description under the same
genus has been the suliject of even a note. It was with a hope of
addinif something new to our knowledge of it that I undertook the ex-
amination of this plant. The observations resulted in the discovery of
several interesting facts, which, together with a general description of
the alga are embodied in this paper.

The literature concerning the genus Choreocolax, besides the origi-
nal description, consists of only a few scattered notes. The genus
was first described and figured by Reinsch in his " Contributiones ad
Alo-olofTJiim et F'unsoloQ-iara." * There he mentions several species
parasitic on various alga?, but described only from sterile specimen*.
Among them is Choreocolax Pohjsiphonice, growing on PoJysijyhonia
fastigiala, the only species that has been reported to have been found
on the American coast since Reinsch 's original descriptions. The next
notice of these parasites is by Farlow, in his New England Alga?,t
where in a foot-note he briefly mentions them. In a paper published
later. $ he makes mention of C. pohjsiphonicB^ descril)ing for the first
time the tetraspores of this plant. They develop from the terminal
cells of the plant, and may be either tripartite or cruciate ; usually the

* Pase 61, Plate XLIX.
t Pa fie 101.

X On some New or impcrfectl}' known Algae of tlie United States. Bull.
Torrey Bot. Club, Vol. XVI. No. 1, p. 6, Plate LXXXVII.

OF ARTS AND SCIENCES. 47

latter. Recently Reiiike and Schraltz found in one of the species for-
merly described as C. inirahilis cystocarpic fruit which enabled them
to ascertain its true affinities. They placed it among the Gelidiaceaj,
and gave to ft the new generic name of Ilarveyella.* Their reasons
for placing it in a new genus were twofold. They rightly thought
C mi'rabi/is to be widely different from G. Poh/ssiphonue. Besides this,
C Pohjsiplioniai w;\^ described before C. mirabiUs in Keinsch's original
account, so that the former is to be regarded as the type of the genus
Choreocolax ratlier than the latter. Besides these notes, Choreocolax
and Harveyella are scarcely more than mentioned by name in a few
other places. Schmitz in his arrangement of the genera of the Flori-
defe t places them both among the Gelidiaceae. Batters mentions
them also in his List of Berwick Algre, % and reports the collection of
tetrasporlc specimens of 0. PulyslphouicE, but has nothing new to add
regai'dlng them. I'hus it will be seen that the literature concerning
these interesting parasites is very scanty.

Choreocolax Polysiphonice grows, as has been already stated, upon
Polysiphonla fastigiata, a common alga along the Northern New Eng-
land coast. The parasite was sufficiently abundant at Nahant, Mass.,
to be collected in consideraljle quantities from the middle of Novem-
ber to the latter part of the following March. I found it also at
Newport, R. I., growing on the Polysiphonia, on the more exposed
points. C. Polysiphonice has been collected by Dr. W. A. Setchell
at various points along Long Island Sound, though its host is much
less common there than farther north. Mr, F, S. Collins has sent
me specimens he found at Mount Desert, on the Maine coast, during
the month of July, 1890. From all reports, however, it is found no-
where in such quantities as at Nahant. Batters § mentions C. Polysi-
phonice as having been collected on the British coast at Berwick Bay,
but adds that it is rare. The distribution of this alga, then, is fairly
wide, and It is probable that wherever its host is found it may also
be expected in greater or less quantities.

To the naked eye the fronds of Choreocolax appear as small whitish
brown dots of varial)le size, situated almost always in the dichoto-
mies of the Polysiphonia. In some specimens collected at Nahant,
almost every axil except the ones of the very youngest branches was

* Alpenflora der west. Ostspp. Deiitsch. Anth,, Kiel, 1880, p 28.
t Systemat. Uebersicht der bislier bekannten Gatt. der Florideen. Flora,
1889, Heft V p 489.
t Pages 1-26 and 142.
§ List of Berwick Algae, p. 142.

48 PROCEEDINGS OF THE AMERICAN ACADEMY

occupied by a frond of the parasite. Such cases as these, however,
are not usual ; ordinarily the host is not so completely covered with
the Choreocolax. That the parasite has a deleterious effect on the
Polysiphonia is very evident, for the fronds of the latter, on which
the Choreocolax is most plentiful, are always paler and less vigorous-
looking than the fronds not so affected. The growing tips of the host
plant, which usually give every appearance of active growth, are fewer
in number, small, and ofteu distorted. Often the terminal branches
can be seen to have shrivelled up and rotted away, probably from in-
sufficient nourishment. In fronds of Polysiphonia, all stages of this
decay may be seen, which varies in intensity according to the
amount of Choreocolax on the frond. Some exceptionally strong
plants seemed little affected, though considerably attacked by the
parasite.

On examining specimens of Choreocolax PolysiphonicB with a hand
lens, they are seen to be usually light-colored masses varying in shape
from a flattened hemisphere to almost a sphere. The surface of most
of them is smooth and the outline of the frond is regular (Fig. 1),
though some are divided unevenly into several lobes (Fig. 2). The
cause of this latter condition will be discussed subsequently. The size
of these masses that constitute the external portion of the Choreoco-
lax is very variable, ranging from small spots that can scarcely be
seen, even with a powerful hand lens, to bodies about 2 mm. in diam-
eter. The majority of them, however, are not over 1.5 mm. in diam-
eter. The small fronds are flattened, becoming more and more nearly
spherical as they increase in size and age. It is only the large fronds
that are lol)ed in the manner mentioned, the small ones are always
quite regular in slripe. The color of the fronds varies from the
translucent whitish color of the young ones, to the dirty reddish appear-
ance of the adult specimens. Occasionally the latter are dark brown-
ish red, though usually they are not very deeply colored, and may
sometimes be almost as white as the young fronds. The larger masses
of C. Poll/si phonicB are of a tough cartilaginous consistency, fii-mer
and more unyielding than the more gelatinous younger fronds. The
ap|)earaiice of C Polj/xiplionice is so uniquf, that, togetlier with its
habitat, the collector is easily informed of the identity of this rather
insignificant-looking alga.

Before describing the growth and development of the frond, it will
be best to explain the structure of the adult plant, wliich cannot be
well compared with the conditions presented by other algic. A sec-
tion through the frond shows it to be composed for the most part of

OF ARTS AND SCIENCES. 49

large, somewhat irrefrnlar cells, approaching a spherical or cylindrical
shape, which are lilled with coarsely grauular, almost colorless con-
tents. The cells are separated, except at narrow points of contact, hy
an almost structureless gelatinous intercellular substance (Fig. 8),
This gelatinous substance, which contains a large amount of water, is
found to a greater or less extent between all the cells of the frond, and
gives to it tlie consistency already mentioned. The large cells which
make up the interior of the frond are not at all regular in either size
or slm()e, some departing so far from the spherical as to become al-
most branched, by the excessive growth of some portion of the cell in
some other direction than that of the main axis. Tiiey are not ar-
ranged in filaments, or in any distinct order, but are joined in a loose
parenchymatous network. In the parts of the frond near its point of
attachment to the host plant, the cells are seen to be smaller and of a
different sha[)e than those of the rest of the plant. They are cylindri-
cal, with one axis considerably longer than the others, and are arranged
in filaments of greater or less length (Figs. 8, 14). The filaments
which may or may not branch, make their way beneath the peripheral
siphons of tiie Polysiphonia, encircling its axial row of cells. Usually
these filaments extend from the frond of Choreocolax fi'om which they
arise, through the length of three or four cells ; cases were observed,
however, where they had penetrated as many as ten cells from their
starting point. It is by means of these cells that the Choreocolax ob-
tains elaborated material from the Polysiphonia, on which it depends,
in a large measure at least, for its nourishment. The close connec-
tion which the filamentous cells of the parasite have with the cells of
the host may be easily demonstrated. A section shrunk in glycerine
and stained with Ilofraann's blue, enables one to see with the greatest
distinctness threads of protoplasm connecting the cells of the two
plants (Fig. 7). Material killed in osmic acid also shows this point
to advantage. At the same time it will be seen that the cells of the
Choreocolax attach themselves almost wholly to the cells of the central
siphon, although sometimes the walls of the peripheral siphon are
penetrated and the material afforded by them appropriated by the
parasite. C. PoJysiphonicB is then, as Reinsch first maintained, a true
vegetable parasite, which depends in the main for its nourishment on
the materials provided by its host, exerting upon the latter nothing
but a deleterious influence. These filaments were never seen to
connect with any of the external swellings except the one from
which they arose. Each swelling then represents a separate frond,
and there is no continuous growth of filaments which ramify through
VOL. XXVI. (n. s. xviii ) 4

50 PROCEEDINGS OF THE AMERICAN ACADEMY

the host, rising at places in external prominences for the purpose of
producing fruit. In the external portion of the frond, the proportion
of" the filamentous cells to the globose ones is very variable. Some-
times the former encroach far on the latter, while again the globose
cells may entirely exclude the filamentous ones from the external
frond. In the same way, the size of both kinds of cells varies a great
(leal ; in fact, in all the structures of the frond, even in rare cases in
the fruit, a great diversity in appearance may be noticed.

Besides the kinds of cells already described, the peripheral cells
of the frond present a very characteristic appearance. Nearer the
outside the cells are seen to be smaller, more nearly spherical in
shape, and more regularly arranged, than in the rest of the frond.
The gelatinous intercellular substance is also considerably diminished
in quantity in this region. Those cells which form the extreme
outer layer are still different in shape from any of the others. They
are somewhat elongated and pyriform, the smaller ends being directed
inwards (Figs. 5, 8). They constitute the growing part of the frond,
as will be described later, in discussing the development of the plant.
The contents of these outer cells is more granular than that of the
others, and the nuclei are more distinct ; in fact, they present all the
appearances characteristic of growing cells. Directly outside of
the pyriform cells there is a thick sheath of cellulose, which covers
and protects the whole frond (Figs. 8, 17). This outer skin of cel-
lulose is not formed by the fusion of the exterior walls of the pe-
ripheral cells ; they are only loosely connected with it, and may be
detached from it without injury. The cellulose sheath may be dis-
sected off in large pieces, when it is seen to be almost structureless,
except for the depressions left by the cells which had formerly been
attached to it, and for the irregular blotches of brownish red color-
ing matter in it. By this means the pigment to which the color of
C. PoJijsiphonice is due may be seen to be contained almost entirely
in this external covering. Sometimes the peripheral cells may be
also tinged with brown, while in one or two cases the whole tissue
of the frond partook of this color. The sheath is nothing more
than a thickening of the gelatinous intercellular substance on the out-
side of the frond. This gelatinous substance is itself but a modifica-
tion of a portion of the walls of the cells, and gives a cellulose test
with chloriodide of zinc as well as the sheath. As in the case of the
other cells of the frond, those near the periphery are subject to some
variation. In those fronds where the filamentous cells extend into
the external protuberance of the frond, the peripheral cells partake

OP ARTS AND SCIRNCES. 51

more or less of the same character. One extremely exaggerated case
was noticed, in which the cells around the circumference of the frond
were enormously elongated. This frond happened to be tetrasporic,
and the tetraspores were also greatly elongated and deformed. The
contents of the cells present little of interest. The outer cells are
filled with a very highly granular {)rotoplasm, in which the nucleus is
very conspicuous. The inner ones contain less of the granular proto-
plasm, and the nucleus is rather more indistinct. All the cells of the
plant are usually colorless, though sometimes they may be tinged with
purple, especially the filamentous ones in closest contact with the Poly-
siphonia. Of ordinary starch there is no trace to be found, iodine giv-
ing only a deep brown color to the whole of the contents of the cells.
Tlie walls of the cells are cellulose, and are not remarkable. One in-
teresting feature regarding the cells of C. Polysiphonioe is their great
variability in the amount of contained food material. This varies
from the condition found in the globose cells of the young plant, which
are gorged with protoplasm, to the decidedly contracted and starved
appearance presented by the cells of some of the adult plants. This
latter condition is particularly to be noticed in the tetrasporic fronds
which are almost ripe, the growth of the tetraspores having appar-
ently taken all the food material held in reserve by the plant. Figure 8
shows a tetrasporic frond where the cells are somewhat affected, and
is a good example of the condition of the average frond. In Figure 14
the cells are seen to be well gorged. The difference in appearance is
often so striking as to lead one to think at first sight that plants are
not of the same species.

The development of the frond I was able to follow with considerable
certainty, except in the youngest stages, of which fewer specimens were
found. As has been said before, the fronds of C Polysiphonice are
almost always found in the dichotomies of the host plant, and the
reason for this can be explained by the following circumstances. In
the axils of the branches of Pulysiphonia fastigiata there is often col-
lected more or less organic or inorganic material, and they are also
frequently occupied by some of the many epiphytic algse that grow
upon this plant. Besides the natural shelter afforded by the axils,
these growths enable the spores to become attached before they are
able to make their way through the tissues of their natural host. The
spore becomes buried in the organic matter collected in the axil, and
in this position begins to germinate. The earliest stage of a develop-
ing spore of Choreocolax that was found was one where five cells were
to be distinguished. The spore had apparently divided into four

52 PROCREDINGS OF THE AMERICAN ACADEMY

parts, and then after some growth one of these four cells had divided
au^ain into two, — a process which the other cells were probably about to
undergo. There is little to be ?aid regarding this ; what nourishment
the young frond reijun-ed to carry on ihe growth was probably taken
from the organic material in wliich it was buried. No giowth had
yet penetrated the cells of the Polysiphonia.

The next stage that was observed in the development of the
frond was rather more complicated. The young frond was here
composed of a considerable number of cells, which, however, pre-
sented as yet no very detinite arrangement (Fig. 3). The gelat-
moOs intercellular substance was present to some extent, and a
touo-her layer of it already covered the outside surface. The first
indications of the growth of the parasite into the Polysiphonia were
also seen here. lu the figure (Fig- 3) where this stage is shown, two
cells will be seen that have thrust themselves between the cells of the
host plant, and have grown some little distance inwards. Even as early
as this the young frond of Choreocolax must have obtained some nour-
ishment from the Polysiphonia, or it would not have given evidence of
60 much activity of growth. Other than this there is no differentiation
in the cells of the frond ; the characteristic ariangement of the ter-
minal layer that is developed in the adult frond has not yet made its
appearance. Having once forced tlieir way into the tissue of the
Polysiphonia, the cells of the Choreocolax grow more rajiidly, and
finally come to encircle cells of the host plant. New filaments push
their way in, and grow in both directions, between the central and pe-
ripheral siphons of the Polysiphonia, attaching themselves chiefly to
the former.

In the mean time the external portion of the frond has been in-
creasing in size. The cells which have pushed themselves into the
host plant have, besides fastening themselves to its cells, begun to
send out branches u^jwanls, which, by subsequent growth, are to form
a part of the external protuberance of the frond. As these cells in-
crease in number they press outwards, and, joining with the rapiilly
developing external portion {il ready .formed, displace the cells of the
host plant in the immediate neighborhood of this growth. Later, the
displaced cells of the Polysiphonia are entirely enveloped by the glow-
ing Choreocolax. No morbid growth is stimulated in them, however ;
they remain entirely passive, and are gradually absorbed by the para-
site, so that in adult specimens there is usually no trace of them left.

Comparatively early in the development of the frond, before the
internal growth of the vegetative filaments has pushed aside the cells

OP ARTS AND SCIENCES.

53

of the Polypi phonia to any great extent, it will be seen that the pe-
ripheral cells of the external portion of the Ciioreocolax frond present
an appearance different from those in the interior. They have become
arranged in a regular layer one cell deep over the entire surface of
the frond, covering the more or less promiscuous mass of cells be-
neath (Fig. 4). The internal cells divide and grow to some extent, but
it is fiom the outer, regular layer of cells that the larger part of the
exterior portion of the frond is to be developed. The growth inside
of the host plant also helps in the formation of this part of the frond,
but it is only for a short tune that it can be distinguished from the
growth of the peripheral cells just mentioned. As soon as the cells
which arise from the inner filaments make their way between the cells
of the Polysiphonia to the outside, they become arranged in this reg-
ular order and blend with the rest of the frond, becoming indistinouish-
able from it. Cases have been seen where, owing to irregular growth,
they did not unite ; and then, instead of one laige protuberance, there
were many smaller swellings closely bunched together. The surface
of the young frond, at first almost a plane, becomes rapidly convex by
the more active growth of the cells in the centre of the frond. Finally,
the hemispherical or almost spherical mass is formed in which the
fruit is later borne.

The ordinary method of growth of the frond in distinction to the
manner of development in its earlier stages is now to be considered.
It is essentially the same after the condition is reached where the
peripheral cells are arranged in a distinct layer. Before that time
the growth is irregular and unequal. Taking a single peripheral cell
and following its growth throughout, we find the method to be as fol-
lows. First, the cell is divided into two parts by the formation of a
transverse wall. The lowest half of the cell does not divide again, but
merely increases in size. The upper cell, on the other hand, is soou
divided in two by the foimation of a vertical wall, and these two cells
ultimately become four by the division of each into two, in a vertical
direction at right angles to the first vertical division (Fig. 6). The
four cells thus formed repeat the process of division first described,
and by this means the frond is enlarged in all three dimensions. The
number of cells into which the outer row of cells may divide vertically
is not necessarily four. There may be only three, or sometimes as
many as five cells so formed, but the important point is that they are
equally distributed, so that, besides extending the frond vertically, they
increase it almost equally in both of the other directions of space.
Other irregularities are also noticeable ; sometimes the transverse

54 PROCEEDINGS OF THE AMERICAN ACADEMY

division fails to take place in some of the cells, and leaves a conspicu-
ously long and ill-sliapen cell, which, however, continues its growth
like the others. At the time of the most rapid growth the formation
of the walls follows so quickly on one another that the newly furmed
cells do not reach their full size before they divide again. As a con-
sequence of this the outer portion of an actively growing frond is
made up of groups of small, closely compacted cells, each group
having originated from the division of a single terminal cell (Fig- o).
The cells of these groups gradually grow and assume the normal
appearance, the outer ones continuing to divide, though more slowly
than before, and the inner ones losing themselves in the inner mass of
the frond. The activity of the terminal cells almost enin-eiy ceases
as the frond approaches maturity, and in the adult frond there is no
sign of turtlier growth.

The tetrasporic fruit of Choreocolax was, as has been said at the
beginning of this paper, first mentioned by Farlow, who gives a brief
account of it. The tetrasporic fronds were not uncommon in the
material I collected at Nahant, and material was easily found from
which to study them. They were no more frequent at one time than
at anolhei", during the portion of the year in which I looked for
them. Externally, the tetrasporic plants present no characters by
which they may be invariably distinguished from sterile specimens.
The size of the frond bears very little relation to tlie presence of even
mature tetraspores, for it is not at all unusual to find in a very minute
frond not a millimeter m diameter tetraspores which are to all ap-
pearance perfectly developed. A vertical section of one of the
hemispherical swellings shows the tetraspores to be located on the
extreme periphery of the frond (Fig. 8). There is no definite limit
to the number of tetraspores to be found in a single specimen ; some-
times there are very few of them, while at other times there are so
many that they have quite crowded the terminal cells out of place.
All stages of growth of the tetraspores are present in one frond at
the same time, so that their development is not hard to trace. They
arise from the enlargement of certain of the terminal cells, but there is
no criterion by which it is possible to tell what ones will develop into
tetraspores. The first indication is a slight swelling of those cells
which are to form the spores (Fig. 9). They rapidly increase in size,
the contents of the transforming cells at the same time taking on a
more granular appearance than their unmodified neighbors (Fig. 10).
After the single cell has attained almost the size of the mature spore,
a transverse wall is formed across it (Fig. 11), and is soon followed

OF ARTS AND SCIENCKS. 55

by a vertical one, which thus divides the spore into four parts, pro-
ducing a very typical cruciate tetraspore (Fig. 12). Sometime-; in tlie
two-celled stage the longitudinal division of the distal cell precedes that
of the proximal one, giving the spore the appearance of being tripar-
tite. The longitudinal wall of the proximal cell is ultimately formed,
however, and then the spore presents the usual cruciate aspect. True
cases of tripartite spores are to be found, however, where the longi-
tudinal division of the lower cell has actually taken place in a direc-
tion at right angles to that of the upper one (Fig. 13). The contents
of the tetraspores do not differ very markedly from those of the other
cells, except that they are more highly granular. Fully adult spores
from fresh specimens are usually of a brownish color, and measure ou
the average 45.5 X 28 ,u. Some apparently mature ones were much
smaller, being only 25 X 18 fi. The curious case of distortion of the
tetraspores where they were so enormously elongated has already been
noted; they measured about SO fjc long by 15-20 /ut broad. The tetra-
spores make their way out by the breaking away of the outer cellulose
skin, which becomes very weak as the frond increases in age, and may
then be easily ruptured by slight pressure. Attempts were made to
germinate the tetraspores, but all proved unsuccessful. It may have
been that the conditions were unfavorable, or perhaps that the tetra-
spores rest some time before germinating

Besides the tetraspores, no one, so far as I am aware, has ever givea
an account of the reproductive organs of C/i9reocolax Pubjsiphonice..
When it was found that the non -sexual reproduction of C. Pohpi-
phonice was by means of tetraspores, this alga could be classed with
much more certainty among the Florideaj, and it was reasonable to
suppose that cystocarps might be found on further search. It was,
indeed, with this possibility in view that I was led to investigate
C. Polysiphoni(E. In all of the material collected during the fall
and early winter of 1890, nothing but the tetrasporic fruit was no-
ticed. Some specimens obtained at Nahant, on December 11th,
proved more interesting. In a few of the fronds, structures were
found which at once appeared could be nothing else than cystocarps.
These observations were corroborated later, and more cystocarpic
material was procured, which enabled me to make out definitely the
structure of the fruit, and to some degree also its development. It
was not until much later that the trichogynes were first seen, and as
it was not possible to trace out the com|)lete course of development
from the trichogyne to the ripe cystocarp, it will be best to begm
with a description of the latter.

56 PROCEEDINGS OP THE AMERICAN ACADEMY

The ripe cystocarpic fronds can usually be distinguished from the
others by the fact that they are more or less lobed, each lobe con-
taining a single cystocarp (Fig. 2). This is not a very reliable dis-
tinction, however, for when a frond contains only one cystocarp,
which not infrequently happens, its shape closely resembles that oF a
tetrasporic or sterile frond. On the other hand, the other fronds are
sometimes lobed, from abnormal conditions of growth, in manner not
unlike the cystocarpic specimens. Thus it will be seen that it is im-
possible to tell certainly, without microscopic examination, in what
state any particular frond may be. In the majority of the cysto-
carpic fronds there are several — from two to five — cystocarps pres-
ent, though a considerably larger one is, as has already been said,
often found in place of many.

Although it might seem to indicate, from the division of the frond
into lobes, that the cystocarps are in this instance external, closer
search shows that this cannot be considered to be the fact ; the growth
of so large a body as the cystocarp in so small a frond naturally neces-
sitates the condition found, and even as it is the lobes represent more
than merely the couceptacles themselves, for the ordinary tissues of
the frond go in part to make them up (Fig. 14). The cystocarps are
ovoid to almost spherical in shape, with the smaller end external.
They may be readily separated from the surrounding cells by a little
careful dissection, when they appear as small white dots, scarcely
visible to the naked ey%. From a vertical section of a cystocarp a
very good idea of its structure may be obtained (Fig. 14). The cells
surrounding the cavity in which the spores are borne are seen to be
more closely compacted than those of the rest of the frond, and of
a different shape. This closely compacted wall consists chiefly of
st'erile cells, with which on the inner surface the spore-bearing cells
are intricately associated. The conceptacular wall is always thickest
at the inner end of the cystocarp, gradually becoming thinner as it
approaches the outside, being represented in the region of the carpo-
stome by a single layer of cells. The carpostoine, which has always
been seen in these cystocarps, consists of a small circular opening
throush the cellulose covering of the frond. It is situated at the
small end of the cystocarp, where it approaches nearest the exterior
of the frond. The cells which compose the wall of the cystocarp,
when viewed in vertical section, are seen to be either spindle-shaped
or very thin and almost fiiliform (Fig. 14). This is due to the
collection of the protoplasm of the cell at the centre, leaving but
a small amount at the extremities. Other aspects show the cells

OP ARTS AND SCIENCES. 57

to be flat or tabular, with often a very irregular outline. They may
then, in a geiil-ral way, be said to be of a lenticular shape, although
they do not often approach the circular in form. They always
lie with their flattened faces presented to the interior surface of
the cystocarp, and consequently a vertical section, if it be not tan-
gential, always exhibits the cells in their fusiform appearance. The
amount of gelatinous intercellular substance between these cells is
much smaller than is found elsewhere, except perhaps in the growing
terminal part of the frond.

It is on cells very like the ones just described, {)erhaps somewhat
thicker in proportion to their other dimensions, that the carpospores
are borne. The cells from which the spores arise lie directly inside of
the conceptacular wall, and are, as has previously been said, closely in-
terwoven with it. They have no peculiarities in shape or structure
which distinguish them from the sterile cells forming the wall.
They do not, however, like the latler, always present their flat
faces inwards, but are more irregularly arranged. The spores are
borne from the ends or angles of the cells, or from protuberances
arising from their flat surface (Figs. 15, 16). The entire surface of
the cavity is lined with the spores, except in the immediate neigh-
borhood of the carpostome (Fig. 14). They are not arranged in chains,
but are borne singly. In shape they are irregularly ovoid or pyri-
form, with tap»ring and sometimes acute apices. A small basal cell
is always found between the spores and the spore-bearing cells proper.
From the basal cell there arises a sterile filament that always appears
to be present (Figs. 15, 16). This paraphysis is usually somewhat
longer than the spore, but as it arises from the side, and not the top, of
the basal cell, it extends out no farther. Its contents are almost hya-
line, in contrast with those of the spore. In the mature cystocarp the
spores are directed inwards, and somewhat upwards, towards the carpo-
stome, almost completely filling the cavity. The ripe spores are highly
granular, somewhat darker in color than the other tissues of the plant,
and have distinct nuclei. That the cystocarps examined were ripe,
there can be no doubt. Some specimens collected on December 23d
were placed uninjured in sea-water on a slide ; when they were exam-
ined, some fifteen hours afterwards, many spores were found to have
made their way out of the cystocarp, and to be lying loose in the sur-
rounding water. Attempts were made to germinate the carpospores
as well as the tetraspores, all of which failed, owing, no doubt, to the
same causes suggested in the case of the tetraspores.

Before leaving the subject of the cystocarps, it is necessary to de-

58 PROCEEDINGS OP THE AMERICAN ACADEMY

scribe a peculiar condition of the frond that was always found in cysto-
carpic plants. In the peripheral portion of such a frotid, instead of
finding the usual elongated pyriform cells, one sees a large number
of small spherical ones arranged in distinct chains (Figs. 14, 17).
The transition between the two conditions can be traced without much
difficulty, in fronds where the cystocarps have just commenced to
develop. It is seen that, when the main part of tlie growth of the
plant has taken place, the terminal cells, instead of dividing as fre-
quently vertically, divide more frequently transversely, forming short
chains of small cells, which afterwards increase considerably in size.
Frequently even in the adult fronds the chains of terminal cells are
seen to give place to the ones of the usual form at the base of the
frond (Fig. 14). There is nothing remarkable in the appearance of
the walls or the contents of these chains of cells that would lead one
to suppose that they have any special function. The condition of the
cells seems merely to be that which is very often seen in the cystocarps
of other alg^, where the outer wall of the conceptacle consists of a
great number of small cells in chains. In the case of C. Polysipho-
nice, the frond is so small in proportion to the cystocarp that the whole
of it becomes modified in this change.

It was not until late in the course of ray examination of C. Poly-
siphonice that I discovered the trichogyne and its accompanying
organs. The fronds containing them were searched for diligently, but
only a few plants were found that were in the right condition. It was
undoubtedly too late in the season when I first found the trichogynes
to expect them to be common, for then almost all of the cystocarpic
fronds were mature. In spite of this, however, sufficient material was
found to make out the structure of the undeveloped procarp, and to
some extent to follow its development. The trichogyne forms the
distal extremity of an irregular chain of cells, which are often con-
nected into a more distinct filament than is common with the interior
cells of C PoJysij)honice. It represents and is developed out of one of
a number of cells, which at first were ordinary terminal cells like the
others of the frond. The terminal cells arising from the same basal
cell as the developing trichogyne and trichophoric apparatus apparently
cease all growth after the latter begin to develop, and soon become
buried in the frond. Oiie or more of them often remain, as in Figures
18, 19, and 21, c. Tlie remaining cell which is to continue the growth
divides, the terminal cell developing into the trichogyne, the lower
ones forming the trichophoric apparatus. When fully developed, the
trichogyne is very long. The cell itself is about 2-2.5 ft iu diameter,

OP ARTS AND SCIENCES. 59

and often 115 yu, long. The diameter of the trichogyne is much
greater, however, by reason of a very thick apparently gelatinous
sheath of high refrangibility. The whole trichogyne, including sheuth,
measures 5-6 fx broad. At the apex the sheath becomes much thin-
ner, and consequently does not materially increase the length of the
trichogyne. The cell of the trichogyne is often irregular in diameter,
frequently exhibiting considerable swellings, which however are not
followed by similar swellings in the sheath (Fig. 19). At the base the
cell broadens out where it joins the trichophoric apparatus (Figs. 18,
19). The trichogyne usually pierces the cellulose covering without
bending or other distortion, and extends for a considerable distance be-
yond the frond. Not infrequently, however, the trichogynes instead
of immediately making their way through the sheath, become bent
when they come in contact with its lower surface. They often grow
for considerable time between the outer layer of terminal cells and the
under surface of the cellulose covering before they succeed in breaking
through it. Below the trichogyne are three trichophoric cells. They
are usually somewhat wedge-shaped cells, of variable size. The order
of their formation I do not know. The one next to the trichogyne is
generally smaller than the others, and is set at somewhat of an angle
to them (a, Figs. 18, 19). In other words, the axis of the procarp
curves here. The other two cells of the trichophore lie side by side,
and are usually of about the same size (a', a", Figs. 18, 19). The tri-
chophoric cells are of a decided brownish color, and their contents are
quite clear. The cell b (Figs. 18, 19), on which the lowest of the tri-
chophoric cells rest, presents very much the appearance of the cells of
the rest of the frond. The cells c (Fig. 18, 19, 21) I take to be un-
developed terminal cells, previously referred to in the development of
the trichogyne. The cell b, and apparently in some cases a number of
the cells beneath it, are probably the carpogenic cells of the procarp.
In all the cases seen it did not seem that the trichophore played any
part in the formation of the cystocarp. In Figure 20 the cells a, a', a",
probably correspond to the trichophoric cells indicated by the same
letters in the other figures. If this is the case, the fate of the tricho-
phoric cells can be accounted for. After fertilization they shrivel up
and finally disappear, without developing further. In the same fig-
ure tr indicates probable remnants of the trichogyne, while the cells
b', h" , and h'" also are the outcome of the division of the original
carpogenic cell b (Figs. 18, 19). Some cases were observed where the
growth of the cystocarp seemed to originate even farther down in the
tissue of the plant, but nothing definite was established concerning

60 PROCEEDINGS OP THE AMERICAN ACADEMY

this. After the stage represented in Figure 20, the development be-
comes hard to trace. In the next one figured (Fig. 21), the cells b, b', b",
etc., are probably to be compared with those indicated by the same let-
ters in the previous figure, while c is an undeveloped terminal cell
before noticed. The cells b', b", etc., have increased enormously in
size and in number as well. They have given rise to many smaller
cells the fate of which seems probably to be the formation of the
cystocarp proper. The course of growth becomes now even more
obscure. The cells in the neighborhood of the young cystocarp be-
come complicated in the growth of the wall, and effectually hide the
changes which at this time are affecting the cystocarp proper. It
seems probable from what was seen that the cells b, b', b", etc., start
up another growth, and, budding outwards, form, with the surrounding
cells whose growth has been already mentioned, the wall of the tabu-
lar cells found in the ripe fruit. Tiie small cells first formed from the
activity of the cells b, b', b", etc., are enclosed in this mass, and develop
into the spores and spore-bearing cells. It is to be regretted that the
material was so scanty for this work. A search will be made next fall
earlier in the season, to find if possible more trichogyne-bearing fronds,
and an attempt made to determine more definitely the development of
the fruit.

Up to the present time the relationship of C. Polysiphonice to the
rest of the Florideae has been very uncertain. Heretofore, it will be
remembered, nothing but tetraspores has been described. In his list
of the Florideae, Schmitz places it among the Gelidiacese with Bin-
derella in the sub-order Binderelleae. His reason for placing it in
the Gelidiaceae is presumably on account of its general likeness to
Harveyella mirahilis (Reiusch), Schmitz and Reinke, which in the
same list is placed in a separate sub-order, Harveyellese, next to the
sub-order to which Choreocolax is assigned. Others have followed
him in this arrangement, but no one, so far as I know, has placed
C. Polysiphonice in any other order of the Floridese. In view of what
has been described in this paper regarding the structure of the cystocarp,
this can scarcely be considered to be its true place. The cystocarp of
Harveyella is likened by Schmitz, in his note in Reinke's " Algen
Flora der westlichen Ostsee," to that of Caulocanthus, a resemblance
which would place Harveyella without doubt among the GelidiaceiB.
It certainly seems impossible to consider the cystocarp of Choreocolax
Polysiphonice as closely related to that found in the GelidiaceaB. It
lacks the most essential feature of similarity to the Gelidiaceoe in the
absence of the complicated axial placenta which characterizes that

OF ARTS AND SCIENCES. 61

order. The carpospores of Ohoreocolax Polysiphonice, it will be re-
membered, were found to be borne singly all over the inner surface
of the cavity, on cells or filaments projecting into its cavity. This
fundamental difference in structure certainly makes it impossible to
consider C. Polyslphonice one of the Gelidiaceae. The condition of
the cystocarp approaches far more nearly that found in the Chajtangi-
acese than in any other order. I examined cvstocarpic specimens of
Chcetangium ornaticm, in order to compare them with those of Clio-
ViOcolax Polysiphonice. Although the cystocarp of Chaetangium is
somewhat more complicated than that of Choreocolax, there is a great
resemblance between the two. The spores are borne in Chjetungium
on filaments projecting into the cavity of the cystocarp, much in the
same way as was observed in Choreocolax Polysiphonice. The fila-
ments in Chaetangium protrude farther into tl>e cystocarpic cavity
than in the other form, but that is not an essential difference. The
spores themselves resemble those of Choreocolax Polysiphonice in
shape, but are much smaller. Galaxaura was also examined, and an
even closer resemblatice seen. The cystocarp of Galaxaura is simple,
like that of Choreocolax Polysiphonice, and the spores are larger than
those of Chaetangium. The dissmiilarity of the fronds found in the
various genera of the Clia^tangiaceic from that of Choreocolax Poly-
siphonice can only be considered as a specific distinction, and not as a
valid objection against placing the plant in question in this order.
The fronds of the forms already included under the Chaetangiaceae
are so dissimilar that there cannot be said to be any typical frond in
this order.

Before closing, it seems well to say a few words regarding the
methods of work employed in investigating this rather unmanageable
alga. The immense amount of gelatinous matter in the frond was of
course a most ditlicult thing to preserve properly. Ordinary methods
of killing with corrosive sublimate and chromic and picric acids were
tried, and yielded partially satisfactory results. The material thus
killed was useful for maceration and dissection. All these reagents,
however, caused a great amount of shrinkage in the tissues of the
plant. Many other methods that seemed suitable were tried with a
hope of finding some reagent that would kill the cells and leave them
in a natural condition. Nothing was discovered, however, that served
this purpose ; in fact, it was found that just as soon as the cells of the
plant died, however cautiously they may have been killed, just so soon
did they shrink and contract into the grotesque shapes one finds them
in. JMore than this nothing could be found that would swell up the

62 PROCEEDINGS OP THE AMERICAN ACADEMY

contracted tissues to their lifelike appearance. Potassic hydrate
caused a general disintegration, and the various acids, unless used so
strong as to dissolve the whole mass, produced but little effect. For
this reason, the greater part of the work was done with sections cut
between pith, in sea-water, with a razor. Nothing else than sea-
water was allowed to touch them, and by this means sections were
obtamed that would keep alive three or four hours, after which time
they gradually contracted as they died. Almost all the drawings were
made from these fresh sections, and whatever work was done with
dried or alcoholic material was verified by means of them. Many
false appearances are given by the immense contraction which takes
place when the frond dies, and it was for this reason that these pre-
cautions were taken. In order to have a supply of fresh material con-
stantly on hand, I made excursions as often as possible during the
wmter to Nahant, where Choreocolax Polysiphoniai is fairly abundant.
The Choreocolax and the Polysiphonia on which it grew could with
care be kept for a considerable time, either in sea-water or moist in
a tin box. The latter way is perhaps the better, and if the box is
kept moderately cool, and has been well sterilized before putting
the material in it, the Choreocolax will keep alive from ten days to
two weeks.

In conclusion, I wish to thank all those who have kindly helped me
in my work. To Professor W. G. Farlow I am especially obliged,
and to Dr. W. A. Setchell I am also indebted for several valuable
suggestions.

Crtptogamic Laboratory, Harvard University,
March, 1891.

>^

2.

10.

14.

i/^^^Trn-..,.

13.

H^ n.i

■-tl.

m Mm

m

10.

15.

a

5

<~

H^

0

_■■)

n

/O'^j

VM\ ]U

w-

OF ARTS AND SCIENCES.

63

EXPLANATION OF FIGURES.

Fip. 1. Typical form of frond. X l\v about 3 diam.
" 2. Lobed cystocarpic frond. X by about o diam.

" 3. Very young frond just making its way into tbe Polysiphonia. X 180.
" 4. Somewbat older stage, where cells are nnorc regularly arranged. X 200.
'■ 5. Terminal cells of growing frond, from macerntion preparation. X 350.
" 0. Diagrams showing method of growtli of terminal cells.
" 7. From a preparation slirunk in glycerine to show connection of cells of

parasite with host. X 150.
" 8. Vertical section of tetrasporic frond. X 150.

" 9, 10, and 11. Three stages in the development of tetraspore. X 350.
" 12. Adult cruciate tetraspore. X 350.
" 13. Adult tripartite tetraspore. X 350.
" 14. Vertical section of cystocarpic frond. X 150.
" 15 and K!. Sliowing the two ways in which the carposporcs are borne.

X 350.
" 17. Portion of periphery of cystocarpic frond, showing chains of cells.

X 350.
" 18. Young procarp. X 600.

tr, trichogyne.

a, a,' a," trichophoric cells.

b, carpogenio cell.

c, undeveloped terminal cell.

" 19. Older procarp with fully developed trichogyne. X 600.

References as in Figure 18.
" 20. Procarp in which carpogenic cells have begun to develop. References

as in Figure 18, except b', b", b'", which are the cells newly formed

from tlie carpogenic cell b. X 600.
" 21. More advanced stage in development. References as before, b"", b^,
etc., are the result of further proliferation of cells b, b', etc. X 600.

Figures 1 and 2 were drawn free hand. The others were all drawn with the
camera, except Figure 6, which is merely a diagram. Figures 18 to 21 have been
reduced one third from the original drawings. All the drawings are from fresh
material except Figures 0 and 7, the latter of which was drawn from a section
shrunk in glycerine.

64

PROCEEDINGS OP THE AMERICAN ACADEMY

VI.

ON THE MATRICAL EQUATION 4>n = Clcf>.
By Henry Taber, Clark University.

Presented by Prof. W. B. Story, May 26, 1891.

For a given matrix, fi, the most general matrix <fi satisfying the
equation ^ fi = fl (/> may be found by the consideration of the canoni-
cal form of the matrix O. If the distinct latent roots of Q are gi,
an m-tuple latent root, ^2? an n-tuple latent root, etc., the canonical
form of fl is w ^ ixT^, where

o=C

Oi

^■2

in which all the constituents are zero except those in the square arrays
61, 62, etc., which correspond, respectively, to the latent roots (ji, g^i
etc., and are severally of order equal to the multiplicity of the latent
root to which they correspond; and if the characteristics of the latent
root gi are {in ; p, ^, r, — s, t), then

OP ARTS AND SCIENCES.

65

'1 —

/ * —

— .

,, ...rf\.—

.

,

»__

^

/>•

.'/lO .
0 <7i.

0 0 .

.0
.0

.6

1 0 .
0 1

0 0..

.0
0

'. 0

f
'A

.71 0 ..
0 </,..

.0

.0

1 0 .
0 1 .

.0
.0

.

0 0..

■9i

b b .

..0

r-

.91 0 .
0 ^1.

..0
.0

0 0 .

■•9i

■

s-

.<7i0 .
0 .91-

..0
.0

1 0 .
0 1 .

b b '

0
.0

.0

0 0 .

■■9i

^

.91 0 .

b 0 .

.0
.0

■9i

where the letters on the left and above denote the number of rows and
columns in the respective squares and rectangles below and to the
right. Similarly with respect to the square arrays 0^, etc., correspond-
ing to the other latent roots.* If, now, <^ is a matrix commutative
with Q, the most general expression for 0 is ^ = w i^ cj~ ^, where

c

'^l

V2

* This representation of a matrix, what I term its canonical or standard
form, was, I believe, first given explicitly by Weyr, Comptes Rendus, Vol. C. ;
it was, however, given substantially by Buchheim, earlier, in Proc. Lond. Math.
Soc, Vol. XVI.

VOL. XXVI. (n. 8. XVIIl.) 6

QG

PROCEEDINGS OF THE AMERICAN ACADEMY

in which all the constituents are zero except those in the squares along
the principal diagonal, and

Vi =

«U «12 •
021 ^22 •

•«1P
■02P

f>n ^12 • •

621 622 • •

•0-29

C'li Ci2 . •
C21 Coo . .

■Cir

etc.

etc.

etc.
etc.

Opi Up, .

..Opp

bpi bpo .

■ bpq

C,ii Cp2 •

.C^r

On 012 •
a.2i 022 .

hi hi •

box bn~2 . .

.b,r

a„ a,2 .

.a„

6,1 652 .

. V

«11 «12 •
CLry O22 ■

.Ajr

. .a.-r

Ori 0(2 •

■ -flrr

the a's, 5's, c"s, etc. being arbitrary ; the mode of filling up the re-
maining squares along the principal diagonal and the rectangles above
the principal diagonal is obvious. A similar expression obtains for
r}2, etc.

WoECESTEB, May 26, 1891.

OF ARTS AND SCIENCES. 67

VII.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF

HARVARD COLLEGE.

ON THE PRODUCTS OBTAINED BY THE ACTION OF

NITRIC ACID UPON BROMTRINITROPllE-

NYLMALONIC ESTER.

By C. LoRiNG Jackson and W. B. Bentley.

Presented May 13, 1891.

Our attention was first called to this subject by the appearance of
a vivid red color, when common strong nitric acid was used to set free
bromtrinitrophenylmalonic ester from its sodium salt, whereas no such
color was observed if hydrochloric or sulphuric acid was used in place of
nitric. Upon studying this action, we soon found that a new compound
was formed, which could also be obtained from the free bromtrinitrophe-
nylmalonic ester by treating it with hot nitric acid for a few minutes,*
or for some hours in the cold ; the free ester therefore acts with nitric
acid less readily than its salt, for with that the color appeared immediately.
The further study of this compound (which was obtained in colorless
prisms) proved that its formula was C6HBr(N02)3CN02(COOC2Hj2 ;
that is, that it had been formed from the bromtrinitrophenylmalo-
nic ester by replacing one of its atoms of hydrogen by the group
N02. We have also obtained the following similar derivatives : from
the bromdinitrophenylmalonic ester C6H2Br(N02)2CN02(COOC2H5)2,
melting point 111°, and from the trinitrophenylenedimalonic ester
CeH(N02)3CN02(COOC2H,)2CH(COOC2H5)2, also melting at 111°,
so that the reaction seems to be a general one. The position of the
group NO2 introduced by the action of nitric acid was determined by
the stuly of the acidity of these compounds. In the bromtrinitrophe-
nylmalonic ester C6HBr(N02)3CH(COOC2H5)2 there are only two
atoms of hydrogen, one on the benzol ring, the other on the side-
chain ; if the first of these were replaced by NO2, the effect, if any

* A preliminary account of this substance forms part of a paper by G. D.
Moore, and one of us (these Proceedings, XXIV. 265). The statements made
there are superseded by this paper.

68 PROCEEDINGS OP THE AMERICAN ACADEMY

would be to increase the acidity of the compound ; whereas if the
atom of hydrogen on tlie side-chain were replaced, the substance
would cease to show acid properties, as the NO2 would have taken
the place of the only atom of hydrogen which can be replaced by a
metal. Upon studying the behavior of the new compound with alka-
line reagents, we found that none of them affected it in aqueous solu-
tion, and that it was acted on only by sodic hydrate or ethylate in
alcoholic solution, thus showing a marked contrast to the behavior of
bromtrinitrophenylmalonic ester, which is acid enough to decompose
sodic carbonate in aqueous solution. As just stated, sodic hydrate or
ethylate does act upon our substance, but the red solution formed by
the sodic ethylate was proved to contain sodic nitrite, and there-
fore the formation of the red salt was preceded by the removal
of the group NO2, which proves that this radical has replaced the
hydrogen of the side-chain, as represented in the formula given
above. This result was confirmed by the study of the compound
CcH2Br(N02)2CN02(COOC2Hg)2, which proved even less suscepti-
ble to the action of alkalies than the corresponding trinitro com-
pound. The replacement of the hydrogen in the malonic ester
radical is not at all strange, since Franchimont and Klobbie * have
found that nitric acid converts malonic ester into nitromalonic ester.

To determine whether our substances were nitro compounds, that
is, with the radical NO2 attached to the side-chain by its nitrogen, or
nitrites, that is, with the radical attached by one of its atoms of oxy-
gen, we considered in the first place the further action of nitric acid upon
the compound C,;HBr(N02)3CN02(COOC2H5)2, which converted it
after warming for three hours into C6llBr(N02)3C01I( 0000,115)2,
that is, the bromtrinitrotartronic ester; the reaction seemed to run
smoothly, and the yield was 40 per cent of the theory. A similar
change was produced by heat; when exposed to a temperature of
124°-126°, the substance melted, turned blood- red, and gave off a
great deal of gas, with a striking increase in volume at the same
time.t A study of the gases showed that a part of the substance

« Ber. d. eh. G., XXIII., R. 62 from Rec. Trav. Chitn., VIII. 283. For a dis-
cussion of tiie efTict of different radicals upon tlie action of nitric acid on fat
substances, see Franchimont, Per. d. ch. G., XX., R. G89 from Rec. Trav. Cliim.,
VI. 224, and Ber. d ch. G., XXIII., R. 64 from Rec. Trav. Cliim., Vlll. 307.

t Tliese phenomena shown by our substance in melting, as well as many of
its other properties, are very similar to those observed by Gabriel in the cases
of the benzylidenplithaliddinitrite, Ber. d. ch. G., XVIII. 1255, and tlie etliin-
diphtlialiiidinitrite, Her. d. ch. G., XIX. 837, which confirms the view that our
substance is a nitrite.

OF ARTS AND SCIEN'CES. 69

had undergone complete decomposition, as they contained bromine ;
the residue, however, furnished a considerable amount of the substi-
tuted tartronic ester mentioned above. The action therefore has a
certain resemblance to the conversion of the nitrate of tartaric acid
into tartronic acid. These easy conversions of the compound con-
taining the group NO.2 into the corresponding tartronic ester can be
best explained on the theory that this group is attached to the side-
chain by oxygen, but cannot be considered as a strict proof of this
{loint. Accordingly we have reduced with tin and hydrochloric acid
the substance CeIl2Br(N02)2CNOo(COOC2ll5)2 (selected because it
can be obtained more easdy than the trinitro compound), and have
found that it yields amraouic chioride and a substance having tlie for-
mula C6H3NH,(CHOHCONH)HCl, that is, the chloride of amidoxy-
oxindol ; there can be no doubt, therefore, that the group NO2 is
attached to the side-chain by one of its atoms of oxygen, and the
substances are nitrites and not nitromalonic esters.

One of the most striking properties of the nitrite of bromtrinitro-
phenylmalonic ester is its slight stability. As has been already stated,
it decomposes rapidly, with change of color and evolution of gas, at
124°-12G°, but it is not necessary to heat to this temperature to bring
about the decomposition, as it also takes place slowly at 100°, and
partially even at 70°. Boiling with alcohol decomposes the substance
completely, and boiling water produces a similar but less complete
change. The nitrites of the corresponding dinitro compound and of
the trinitrophenylenedimalonic ester are much more stable.

As has been stated above, the nitrite of bromtrinitrophenylmalonic
ester, if warmed for three hours with strong nitric acid, or heated
alone to its decomposition point, yields a product in which the radical
of the nitrous acid has been replaced by hydroxyl, and which, is there-
fore bromtrinitrophenyltartronic ester ; this substance melts at 156°,
and possesses marked acid properties. It is not necessary, however,
to make it from the nitrite, as the bromtrinitrophenylmalonic ester is
converted directly into it by warming for three hours with nitric acid.
This therefore is a case of the direct oxidation of a tertiary liydrogen
to hydroxyl,* and, as we have succeeded in isolating the intermediate
product, our work throws some light on the mechanism of the reaction,
showing that it consists in the case of nitric acid of the formation of
a nitrite, followed by its saponification to the hydroxyl compound.

* Richard Meyer, Ann. Cliem , CCXIX. 234, CCXX 1 ; J. Bredt, Ber. d.
ch. G., XIV. 1780.

70 PROCEEDINGS OF THE AMERICAN ACADEMY

C6H(C6H5NH)(N02)3CNO,(COOC2H5)2, the nitrite of anilido-
trinitropheiiylauiloiiic ester, was obtained by the action of aniline on
the corresponding bromine compound ; it forms red rhombohedra,
which melt at 119°, decomposiug at one degree higher, and shows
marked acid properties, which must be due to the effect of the three
nitro groups upou the hydrogen attached to the nitrogen in the anilido
group, since, as has been already stated, the corresponding bromine
compound forms no salts. This conclusion was confirmed by the
study of the anilidotrinitrotoluol, which contains no other hydrogen
capable of being replaced by a basic radical, and yet formed the sodium
salt CGCH3H(C6H5NNa)(N02)3. As was to be expected, however,
this toluol compound was not so acid as the nitrite of anilidotrinitro-
phenylmalonic ester, which contains such a very negative radical in
place of the methyl.

The anilidotrinitrotartronic ester was also made from the corre-
sponding bromine compound, and was obtained in two modifications ;
one formed at higher temperatures appeared in orange-red prisms
melting at 143°, the other in rounded masses of yellow needles melt-
ing at about 122°. Both show the same percentage composition on
analysis, and one is easily converted into the other ; the yellow into
the red by allowing the alcoholic solution to crystallize at about 60°,
by heating the solid a little below its melting point, or by boiling it
with water ; the red into the yellow by solution in glacial acetic acid
and precipitation with water. Unfortunately we have not succeeded in
determining the molecular weights of these substances, as we have
not yet found a solvent that gives satisfactor}' results with Raoult's
method. The work on these substances will be continued in this
laboratory, especially with a view to determining whether the nitrogen
is the cause of the isomerism (if they are not polymeres), but we may
say now that this does not seem very probable to us on account of the
striking resemblance in properties * between our two substances and
the two forms of benzilorthocarbonic acid described by Graebe.f

The anilidotrinitrophcnyhartronic ester forms salts with one atom,
C6H(C6H5NH)(N02)3COM(COOC2H5)2, or with two atoms of uni-
valent basic radicals, CeH(CcH5NM)(NO,)3COM(COOCoH5)2. To
our great surprise, the disodic salt was formed even when the ester
was present in large excess. This tartronic ester is much more
strongly acid than the anilidotrinitrophenylmalonic ester (melting

* It is a curious fact that even the melting points are nearly the same.
t Ber. d. ch. G., X.XIH. 1344.

OP ARTS AND SCIENCES. 71

point 133°), a fact which confirms the formula assigned to it. We
may add, that we have failed in all our attempts to convert anilido-
trinitrophenylmalonic ester into its nitrite, or the corresponding tar-
tronic ester, by the action of nitric acid.

We have also tried without success to convert the bromdinitrophe-
nylmalonic ester into diiiitrophenylenedimalonic ester by the further
action of sodium malonic ester.

The remainder of the paper contains the experimental details of
the work.

Preparation of Tribromtrinitrobenzol.

The experience gained in making tribromtrinitrobenzol for the
work described in this paper has led us to introduce into the process
given in previous papers * from this Laboratory some improvements,
which are described in this section.

To the preparation of tribromaniline we have nothing to add, but
in the conversion of it into tribrombenzol we have found it best to
proceed as follows: — 50 gr. of dry tribromaniline were dissolved
with the aid of heat in 300 c. c. of alcohol containinsr 75 c. c. of
benzol to increase its solvent power, and 20 c. c. of common strong
sulphuric acid added to the hot solution from a pipette. If this
formed a precipitate, it was dissolved by longer heating, more of the
solvents being added if necessary. 20 gr. of finely powdered sodic
nitrite were then sifted into the hot liquid, as rapidly as the violence
of the action would permit ; after which the whole was heated until
the effervescence had ceased, and, after standing over night, filtered,
washed, and dried, when the product could be treated directly with
nitric acid to make the tribromdinitrobenzol in the manner already
described.! The yield of tribrombenzol was good, 42-44 gr. from the
50 gr. of tribromaniline. The filtrate and washings contained so little
of the organic substance that it was not worth while to work them up.

In the conversion of this body into tribromtrinitrol)enzol we have
found the cause of the much larger yield obtained by Dr. Moore X than by
Dr. Wing.§ This does not depend so much on the larger proportion of
fuming sulphuric acid used, as on the rapidity with which the boiling is
carried on. To get the best yield we found it necessary to raise the mix-
ture to the boiling point as quickly as possible, and to keep it boiling
violently during the whole of the three hours. Under these conditions
the yields varied from 8 to 10 gr. of tribromtrinitrobenzol from 20 gr.

* These Proceedings, XXIII. 139, XXIV. 258, 273.

t Ibid., XXIV. 274. X Ibid., XXIV. 268. § Ibid., XXIII. 140.

72 PROCEEDINGS OF THE AMERICAN ACADEMY

of tribromdiuitrol)enzol, that is, 36 to 45 per cent of the theoretical,
if both the nitric and fuming sulphuric acids were of the best quality.
To prove that this violent boiling was the cause of the large yields, in
one experiment the mixture was allowed to stand in the cold for a
week, and then kept barely at the boiling point for seven hours, when
it yielded only 18 per cent of the tribromtriuitrobenzol.

The Nitrite of Brointrinitrophenylmalonic J^sle?',
C6HBr(N02)3CN02(COOC2H5)2.

This substance was prepared as follows : — 3 gr. of bromtrinitro-
phenyhnalouic ester* were mixed with about 10 gr. of nitric acid,
specific gravity 1.38, and warmed in a dish for three minutes on the
water bath, when both the undissolved organic substance and the
acid became intense blood-red ; the mixture was then allowed to cool,
the acid liquid poured off, and the solid residue warmed once more for
two minutes with about the same volume of fresh acid. After this,
the acid was decanted off, and the solid crystalline product washed
with water till free from nitric acid, which changed it from a deep
blood-red to a pale reddish-white color. The puritication of this sub-
stance gave us much trouble at first on account of its slight stability,
since even the short warming necessary to dissolve it in alcohol was
sufficient to decom])ose it partially, while longer heating wTth alcohol
produced complete decomposition ; but at last we obtained satisfactory
results from the following method. The crude substance, after the
thorough washing with water mentioned above, was dissolved in warm
chloroform, in which it is freely soluble, but even with this solvent
care must be taken to warm the mixture for as short a time as pos-
sible; it was advisable, therefore, to achieve the solution rather by the
use of a larger qu ;ntity of eliloroform than by using a smaller volume
at a higher temperature. This chloroform solution was then diluted
with about its own volume of common alcohol, when the substance
graduRlly separted in well formed white prisms, and was obtained
pure after two of these crystallizations. It must be dritd in vacuo,
as it decomposes on the water bath ; in fact, even the temperature of
a steam radiator (about 70°) was sufiicient to bring about a partial
decomposition.

A small additional amount of this substance could be obtained from
the rei! nitric acid mother lifjuors formed in its preparation, either by
adding water or by evaporating them to dryness ; but tliis (luantity is

* Mellingut 104°. These Proceedings, XXIV. 258.

OF ARTS AND SCIENCES.

73

so small that it is not worth the amount of work necessary to free it
from the red viscous impurity with which it is mixed, especially as the
mother liquors can be used to more advantage for the preparation of
the bromtrinitrophenyltartronic ester made by the longer action of
hot nitric acid on broratrinitrophenylmalonic ester.*

The same substance was formed wlien the bromtrinitrophenyl-
maloiiic ester stood in the cold with nitric acid of specific gravity
1.38 for three days. At first there was no visible change, but after
two hours the mixture began to show a red color, which increased in
intensity to a deep blood-red. The product was puiified in the way
just described, but tliis method of making it is on the whole not so
good as that at a temperature of 100°.

The e.ise with which the substance breaks up under the influence of
heat rendered the combustion of it a matter of gre.it diffii;ulty, since
we encountered at one time an almost explosive evolution of gas, and
at another tiie formation of a partial vacuum in the tube. The latter
we are inclined to ascribe to the sudden absorption 1)y the plumbic
cbromate of the large quantity of bromine given off by the substance.
We finally succeeded in getting good results by spreudinj;- out the
weighed portion through the whole length of a long porcelain boat,
and then applying the heat so gradually that a layer of not more
than a few millimeters of ic melted at any one time. Care was also
taken that the temperature did not rise much above lla° until the
whole of the substance had been melted, after which the combustion
was finished in the usual way without trouble.

The analyses led to the fallowing results : —

I. 0.2157 gr. of the substance gave on combustion 0.2514 gr. of
carbonic dioxide and 0.0498 gr. of water.
II. 0.197"^ gr. of the substance gave 0.2310 gr. of carbonic dioxide
and 0.0428 gr. of water.

III. 0.1978 gr. of the substance gave 0.2324 gr. of carbonic dioxide

and 0.0430 gr. of water.

IV. 0.2014 gr. of the substance gave 20.6 c. c. of nitrogen at a tem-

perature of 24° and a pressure of 772.6 ram.
V. 0.1890 gr. of the substance gave 18.6 c. c. of nitrogen at a
temperature of 19° and a pressure of 752.8 mm.

* On one occasion the acid mother liquors yielded by spontaneous evapora-
tion hirge wliite prisms, wliicli melted in the crude state at about 99°, and we -e
apparently somewhat soluble in water; but, althoush we have tried in many
ways, we have not succeeded in obtaining this substance a second lime.

74 PROCEEDINGS OP THE AMERICAN ACADEMY

VI. 0.2101 gr. of the substance gave by the method of Carius
0.0813 gr, of argentic bromide.
VII. 0.2078 gr. of the substance gave 0.0792 gr. of argentic bromide.
VIII. 0.2054 gr. of the substance give 0.077G gr. of argentic bromide.

Found.
I. II. III. IV. v. VI. VII VIII.

Carbon 31.79 31.84 32.03

Hydrogen 2.56 2.40 2.42

Nitrogen 11. 60 11.20

Bromine 16.47 16.22 16.08

These analytical results agree best with the percentages required
by a substance formed from the bromtrinitrophenylmalonic ester by
replacing one of its atoms of hydrogen by the radical XO2, as is seen
from the following table.

Carbon
Hydrogen
Nitrogen
Bromine

Our reasons for the position assigned to the radical NO2 in this
formula, and for supposing that it is not a nitro group (— N = O2),
but a nitrite ( — O — N=:0), have been already stated in the intro-
duction to this paper. The yield is satisfactory on the whole ; 5 gr.
of bromtrinitrophenymalonic ester gave 3.26 gr. of the substance,*
mstead of the 5.5 gr. required by the calculation, that is, 59 per cent
of the theoretical yield.

Properties of the Nitrite of Bromtrinitrophenylmalonic Ester. — It
occurs usually in short thick well developed glistening white prisms,
with terminations consisting of two planes at both ends, which seem to
indicate that the crystals belong to the raonoclinic system. Less com-
monly the prisms are long and rather slender, with terminations similar
to those of the shorter form. Its behavior when heated is very charac-
teristic. If a tube containing some of it is dipped into an oil bath at
124°-126°, the substance turns red round the sides, then deeper, and
the action gradually runs through the mass, until after a second or two
the whole is melted forming a dark blood-red liquid, which occupies
mauy times the volume of the original substance, and contains bubbles

Calculated for

Found.

CcHBr(N02)30NO.>(C02C2H5)2.

Mean.

31.52

31.88

2.22

2.46

11.32

11.43

16.16

16.26

* In addition to this amount 0 7 gr. of bromtrinitroplienyltartronic ester
was obtained from the mother liquors.

OP ARTS AND SCIENCES. 76

of gas. The action is evidently a decomposition, and can be produced
at much lower temperatures, since if the tube containing the substance,
instead of being dipped into the heated bath as described above, is
gradually heated with the bath, the action takes place even below
120°. In fact, it can also be brought about by long-continued heating
in the steam bath, or partially even at 70°. It is evident, therefore,
that the compound has no definite melting or decomposition point, but
yet the temperature given at first (124°-126°) can be used as such in
purifying the substance, since it is essentially constant, if care is taken
always to heat the samples examined in the same way,

Tiiis decomposition seemed so interesting that we examined it more
carefully. F'or tliis purpose a considerable quantity of the nitrite of
bromtrinitrophenylmalonic ester was heated in a test tube inserted in
an air bath, and the gaseous products of the reaction drawn through a
solution of baric hydrate. At 103°-108° there was a quantity of red
vapor given off, and a precipitate of baric carbonate formed in the
tubes containing the baric hydrate solution. The red vapor was recog-
nized by the smell as bromine, and this was confirmed by the precipi-
tation of argentic bromide when argentic nitrate was added to the
acidified filtrate from the baric carbonate. On the other hand, we
could not detect a trace of nitrate or nitrite with ferrous sulphate
and sulphuric acid. After the substance had been heated to 103°—
108° for three hours, the tempei-ature was raised to 125° for two
hours and a half, but the additional loss at this higher temperature
was very small. We intended originally to determine quantitatively
the amounts of the various products, but abandoned this idea when
we found that the loss was not constant, three experiments giving
25 per cent, 30 per cent, and 21 per cent respectively. The appear-
ance of the bromine, too, shows that there has been a complete decom-
position of a part of the substance, and therefore the volatile products
are of so little interest that we did not care to spend the time neces-
sary for the identification of the others, which must have been formed
in addition to the bromine and carbonic dioxide. On the other hand,
we were much interested in the non-volatile product of the reaction
left in the test tube as a fused rather viscous mass of a brownish red
color, in which crystals were embedded. It was purified by washing
several times with alcohol, which removed much of the viscous por-
tion, then the residue was crystallized, first from dilute, and finally
from common alcohol, and when pure showed the melting point 156°
and the crystalline form of the bromtrinitrophenyltartronic ester
described later in this paper. The action of heat on our nitrite, there-

76 PROCEEDINGS OP THE AMERICAN ACADEMY

fore, is similar to the well known conversion of the so-called nitrotar-
taric acid into tartronic acid.

Tho nitrite of bromtriuitrophenylmalonic ester is almost insoluble
in cold water, perhaps a little more soluble in boiling water ; but if the
substance was boiled for some time with water, the crystals were con-
verted superficially into a daik red oily substance, and the water gave
a slight test for bromine; this decomposition seemed to be due, how-
ever, only to the heat (see above), nut to the presence of tlie water.
It is slightly soluble in cold, more freely in liot ethyl alcohol, but this
solution is easily decomposed by heat, since boiling for fifteen minutes
converts the substance completely into a brownish red viscous product
resembling half-dried varnish, from which nothing definite could be
isolated, and even a very short heating with alcohol is sufficient to
form some of this viscous body. Its solubility in methyl alcohol
resembles that in ethyl alcohol; it is freely soluble in chloroform, or
acetone; soluble in benzol, or glacial acetic acid; slightly soluble in
carbonic disulphide ; very slightly in ether, and essentially insoluble in
ligroine. The best solvent for it is the mixture of chloroform and
alcohol described above. Strong sulphuric acid seems to have no
action on it in the cold, or if heated, until the substance decomposes,
when it dissolves forming a reddish solution ; strong hydiochloric acid
has no action on it, either hot or cold, as long as the substance does
not decompose ; strong nitric acid has little or no action on it in the
cold, even when allowed to stand with it for some weeks, when heated
to 100° it converts it gradually into the bromtrinitrojjhenyltartronic
ester, as is described in detail later in this paper under the preparation
of that substance.

Ammoiiic hydrate acts on it little, if at all, even when the action
is assisted by the addition of alcohol ; sodic hydrate in aqueous so-
lution has no action, but if alcohol is added the crystals begin to dis-
solve slowly, imparting a red color to tlie liquid but only a partial
solution is effected m the cold. From this observation we inferred
that no salt was formed until the nitrite was decomposed, and to test
the accuracy of this inference we treated some of the nitrite with an
alcoholic solution of sodic etliylate, which gave at once a daik blood-
red coloration ; but even here ordy a part, and that not the larger part,
of the crystals of the nitrite was dissolved ; the red solution poured
off from the unaltered crystals gave a good test for a nitrite with
starch paste, potassic iodide, and dilute sulphuric acid, with ferrous
8ul[»hate and sidphuric acid, and by Liebermann's reaction, so that
there can be no doubt that sodic nitrite was formed in the experiment

OP ARTS AND SCIENCKS.

77

This proves the correctness of our inference that the substance melt-
inff at 124°-12G° cannot form salts ; and the blood-red salt observed
must have been derived from the decomposition product left after the
removal of the group NO2 from our substance. It may be added, that
acid or neutral sodic carbonate has no action on the nitrite in aqueous
solution, and very little, if any, in presence of alcohol.

Nitrite of Anilidotrinitrophenylnialonic Ester,
CoII(C«H5Nll)(N02)3CNO,(COOC2H5)2.

Aniline acts with great violence on the nitrite of bromtrinitrophenyl-
malonic ester. If the substances are mixed at ordinary temperatures,
the action is almost explosive, a good part of the mixture is thrown
out of the beaker, and the product seems to be principally carbon. If
the mixture is kept cool by immersing the beaker in water, the action
goes more mildly, but the product is still very black and impure. We
therefore carried on the reaction in ethereal solution with the best
results as follows : — 1 gr. of the nitrite of bromtrinitrophenylmalonic
ester was mixed with a small quantity of ether, and, disregarding the
fact that a portion of the crystals had not dissolved, a slight excess of
aniline was added ; the ether at once turned red, and the undissolved
crystals of the nitrite were taken up, while aniline bromide was de-
posited in their place. At the end of a few minutes the reaction was
complete, and, after washing out the aniline and aniline bromide with
water containing a little hydrochloric acid, the ether was allowed to
evaporate, when it left a vivid red mass, which was purified by dissolv-
ing it in hot chloroform avoiding long heating, and then adding alco-
hol until the crystals began to separate, as it had been found that this
Bubstance, like the corresponding bromine compound, was decomposed
by heating with alcohol. After the substance showed the constant
melting point 119° it was dried in vacuo for analysis.

The combustion of this substance was even more difficult than that
of the bromine compound, as it decomposed with almost explosive
violence at a temperature a few degrees above its melting point, and
did not begin to decompose at all at lower temperatures. We were
unable therefore to burn it in an open tube, but at last got satisfjictory
results by using a closed tube, mixing it with a long layer of plumbic
chromate and applying the heat very gradually. Its analyses led to
the foUowina: results : —

I. 0.1542 gr. of the substance gave on combustion 0,2538 gr. of car-
bonic dioxide and 0.0678 gr. of water.

78 PROCEEDINGS OF THE AMERICAN ACADEMY

II. 0.1494 gr, of the substance gave 18.9 c. c. of nitrogen at a tem-
perature of 24°. 3 and a pressure of 752.3 mm.

Calculated for Found.

C6H(C6H6NH)cno,)3Cno3(co2C2Hb)j. I n.

Carbon 44.97 44.88

Hydrogen 3.35 4.88

Kitrogen 13.80 14.02

In spite of the unsatisfactory number for the hydrogen, these results
prove that the substance has the formula which we have assigned to
it. The yield was good, 0.9 gr. of the nitrite of the bromtriiiitrophe-
nylmalonic ester giving 0.7 gr. of the anilido compound, instead of the
0.92 gr. required by the theory, that is, 76 per cent.

Properties. — The nitiite of the anilidotrinitrophenylraalonic ester
is a very beautiful substance, crystallizing in rhombohedra often two
millimeters long and one millimeter thick, with a very acute angle,
which is frequently, but not always, truncated by a basal plane. The
color of the crystals by reflected light is a rich full red, somewhat
darker than that of chromic anhydride, by transmitted light orange-
red. It melts at 119°, turning black and giving off a few bubbles of
gas, if the temperature has not risen above this point ; but at 120° it
decomposes with such violence that frequently a good part of the sub-
stance is thrown out of the capillary melting tube. It is essentially
insoluble in cold water, and its solubility does not seem to be uicreased
by heat; very slightly soluble in cold ethyl alcohol, more soluble in
hot, but still not freely, the hot solution seems to undergo partial de-
composition ; more soluble in methyl alcohol, whether cold or hot,
than in ethyl alcohol, but not freely even in this ; very freely soluble
in chloroform ; freely in benzol, or acetone ; soluble in glacial acetic
acid; slightly in ether, or carbonic disulphide; insoluble in ligroine.
The best solvent for it is a mixture of chloroform and alcohol. Strong
sulphuric acid seems not to act on it in the cold, but when heated
with it the substance decomposes, and then dissolves with a blackish
color ; strong nitric acid does not act on it in the cold, but when hot
dissolves it with a yellow color ; strong hydrochloric acid has no action
on it, whether cold or hot.

The nitrite of anilidotrinitrophenylmalonic ester showed marked
acid properties. Acid sodic carbonate in aqueous solution had no
action with it, but, if assisted with alcoliol, gave a red solution ; potas-
sic carbonate behaved in the same way ; sodic hydrate turned the
crystals black, forming at the same time a brownish solution, which
on addition of alcohol became deeper and blackish red, turning in

OP ARTS AND SCIENCES.

79

time to blackish brown ; ammonic hydrate in aqueous solution gave a
red color, but the action was not complete until alcohol was added,
when a very dark red solution was formed. The nitrite of auilidotri-
nitrophenylmalonic ester therefore shows much more acid properties
than the corresponding bromine compound, which is indifferent to all
these reagents except sodic hydrate in presence of alcohol, and this
fact puzzled us at first, until we decided that the hydrogen attached to
the nitrogen in the anilido radical (CcHsNH) might be rendered acid
by the presence of the three nitro groups. That this is the correct
explanation of the phenomenon we have proved by preparing and ana-
lyzing the sodium salt of anilidotrinitrotoluol, which contains no atom
of hydrogen that could be replaced by sodium except the one attached
to the nitrogen of the aniline radical. The description of this work
will be found in the following section.

An attempt to analyze the sodium salt of the nitrite of anilidotrini-
trophenylmalonic ester gave no satisfactory result, owing undoubtedly
to the partial formation of sodic nitrite from the organic nitrite by the
action of the sodic hydrate used in making the salt.

Sodhcni Salt of Anilidotrinitrotuluol, C6CH3H(C6H5NNa)(N02)8.

To prepare this salt, 1 gr, of anilidotrinitrotoluol * (melting point
151°) was dissolved in alcohol, and mixed with an alcoholic solution
of 0.09 gr. of sodic hydrate, that is, over 0.03 gr. less than ;he amount
required to convert the gram of substance into its salt ; a little ether
was then added, and the solution of the salt evaporated rapidly to
dryness in a narrow beaker sunk throughout its whole length in the
steam bath. During the evaporation there was an odor of phenyliso-
cyanide. The dry mass was washed thoroughly with benzol to re-
move the excess of anilidotrinitrotoluol, the residue dried at 100°, and
analyzed. As the salt explodes when heated with strong sulphuric
acid, it should be dissolved in water, treated with dilute sulphuric
acid, filtered, and the filtrate evaporated to dryness and converted
into neutral sodic sulphate.

0.7876 gr. of the salt gave 0.1662 gr, of sodic sulphate.

Sodi

lum

Calculated for OeCILiHlCjHsNNaXNOoJa. Found.

6.76 6.84

The salt prepared as described above forms a maroon-black powder,
which explodes gently when heated alone, or with strong sulphuric

* These Proceedings, XXIV. 255.

80 PB0Ci3EDINGS OP THE AMERICAN ACADEMY

acid, forming in the latter case a very loose spongy blackish mass.
It is soluble in alcohol with a yellowish brown color like that of ferric
chloride ; insoluble in ether, or benzol ; water decomposes the salt
almost completely, forming a yellow precipitate and a pale brotvn solu-
tion, which is strongly alkaline. The action of the anilidotrinitroto-
luol with alkaline reagents was also studied, and, as was to be expected,
no action was obtained with acid, or neutral carbonate of sodium, or
ammonic hydrate in aqueous solutions ; but if the action was assisted
by the presence of alcohol, each of these reagents gave a barely per-
ceptible reddish coloi'ation, so slight in the case of ammonic hydrate
that we were doubtful whether there had been any action ; sodic
hydrate in aqueous solution gave a slight yellow coloration, which, on
the addition of alcohol, was at once converted into a very dark blood-
red. The acid character of the anilidotrinitrotoluol therefore is much
less marked than that of the nitrite of anilidotrinitrophenylmalonic
ester, and as the position of the anilido group with regard to the nitre
groups is the same in both compounds, we must ascribe this to the
presence of the nitrite of the malonic ester radical CN02(COOC2H5)2,
with its large amount of oxygen in place of the indifferent methyl con-
tained in the toluol.

Bromtrinitrophenyltartronic Ester, C6HBr(N02)3COH(COOC2H5)2.

If the nitrite of bromtrinitrophenylmalonic ester, or the ester itself,
was heated for some time with nitric acid, it was converted into the
corresponding tartronic ester, which we found it convenient to prepare
as follows : — 1 gr. of bromtrinitrophenylmalonic ester was mixed
with about 10 c. c. of nitric acid of specific gravity 1.38, and heated on
the water bath for three hours in a dish covered with a watch-glass,
fresh nitric acid being added when necessary to replace that lost by
evaporation. At the end of this time the substance was found to be
completely dissolved in the small quantity of hot nitric acid left, but
upon cooling colorless crystals separated, which, after washing with
water, were essentially the pure substance, one recrystallization from
alcohol being all that was needed to make them melt at 156°, the con-
stant melting point of this compound. If, as was usually the case, the
bromtrinitrophenylmalonic ester was contaminated with acetylene-
tetracarbonic ester, a yellowish oil was also formed, which however
remained dissolved in the nitric acid mother liquor. The amount of
the substituted tartronic ester in this mother liquor was so small that
it did not pay for the working up.

The bromtrinitrophenyltartronic ester can also be made from the

OP ARTS AND SCIENCES.

81

nitrite of the corresponding malouic ester, but this method has no ad-
vantage over the preparation direct from the malonic ester described
above, except in the case of the nitric acid mother liquors from mak-
ing the nitrite, which yield on longer heating with nitric acid a small
amount of the tartronic ester, and this is worth saving on account of
the tedious and costly preparation of the mother substance.

The substance dried in vacuo was analyzed with the following
results : —

I. O0I95I gr. of the substance gave on combustion 0.2380 gr. of
carbonic dioxide and 0.0584 gr. of water.
II. 0.2049 gr. of the substance gave 16.2 c. c. of nitrogen at a tem-
perature of 17°. 5 and a pressure of 763.7 mm.

III. 0.2044 gr. of the substance gave according to the method of

Carius 0.0836 gr. of argentic bromide.

IV. 0.2272 gr. of the substance gave 0.0930 gr. of argentic bromide.

IV.

Carbon

Calculated for
C6HBr(N0o)3C0H(C02C,H5)j.

33.47

I.

33.28

Found.
II. Ill

Hydrogen

2.58

3.33

Nitrogen

9.02

9.20

Bromine

17.17

17.4

17.42

Our reasons for considering this a tartronic ester rather than a
phenol have been given already in the introduction to this paper. The
yield is satisfactory, considering the amount of loss to be expected
from such a method of preparation ; 1 gr. of bromtrinitrophenylmalonic
ester gave 0.4 gr. of the substance, or about 40 per cent of the theo-
retical yield.

Properties. — The bromtrinitrophenyltartronic ester crystallizes from
alcohol in long white prisms, terminated by planes at a very obtuse
angle to each other, so that the ends look almost but not quite square.
These crystals are often arranged in radiating groups, and in that case
are generally much more slender than when occurring in isolated prisms.
It is deposited from its solution in hot nitric acid in rather sharp slen-
der needles. It melts at 156°, and is nearly but not quite insoluble
in water, whether hot or cold ; soluble in cold alcohol, freely in hot ;
rather more soluble in methyl than in ethyl alcohol ; freely soluble
in benzol, glacial acetic acid, or acetone; soluble in ether, and some-
what more so in chloroform ; nearly insoluble in carbonic disulphide;
insoluble in ligroine. The best solvent for it is hot alcohol. Strong
sulphuric acid has no action upon it in the cold, but if hot, dissolves

VOL. XXVI. (N. 8. XVIII.) 6

82 PROCEEDINGS OP THE AMERICAN ACADEMY

it with difficulty with a reddish brown color ; strong hydrochloric acid
has no action on it hot or cold in open vessels, but decomposes it if
the two substances are heated toijether in a sealed tube ; nothiujj could
be obtained from the products, however, except a mass like a dry
varnish, which we did not succeed in bringing into a state fit for
analysis. An exactly similar result was obtained in an attempt to
saponify it with sulphuric acid of specific gravity 1.44. Strong nitric
acid has no action on it at first, whether cold or hot, but if heated with
it for some time seems to destroy it completely, and a similar result
was obtained upon boiling it with bromine and water, as in both
these cases we could not succeed in isolating any organic oxidation
product.

Acid sodic carbonate in aqueous solution had no action on brom-
trinitrophenyltartronic ester, but on the addition of alcohol a pale red
color appeared, which became more marked on standing ; potassic
carbonate also did not affect it in aqueous solution, but gave a strong
dark red solution on addition of alcohol ; aqueous sodic hydrate
turned the crystals brown, and dissolved some of them with a brown-
ish color ; on the addition of alcohol all the substance went into
solution with a dark red color ; ammonic hydrate gave a red solution,
but the action was not complete until alcohol was added, when the
solution became very dark red. The bromtrinitrophenyltartronic
ester therefore showed the strong acid character which we should ex-
pect, but no attempt was made to analyze its salts, as it was found
that the bromine atom was removed by alkaline solutions with great
ease, sodic bromide being formed, when it was treated with sodic
ethylate even in the cold.

A n ilidotrin itrophenyltnrtro n ic Ester,
C6H(CcH5NH)(NO,)3'cOH(COOaH5)2.

This substance was made by adding an excess of aniline to solid
bromtrinitrophenyltartronic ester ; the reaction ran smoothly with a
slight evolution of heat, and was complete after the mixture had been
allowed to stand for a few minutes, when the aniline bromide and ex-
cess of aniline were removed by washing with very dilute hydrochloric
acid, and the bright yellow residue purified by crystallization from alco-
hol. At first rounded masses of radiating needles looking like little
balls of fat of a bright yellow color were obtained, but as the crystal-
lization continued, orange-red prisms began to appear, and increased
in quantity until after several recrystallizations the greater part of the
substance had been converted into this form. This behavior during

OF ARTS AND SCIENCES.

83

crystallization suggested to us that the substance probably occurred in
two modifications, which was proved to be the case by the determina-
tion of the melting points of the two sorts of crystals, the yellow, after
thorough purification, melting at about 122°, the red at 143°. As the
substance passes from one form to the other with great ease, many ex-
periments were necessary in order to find methods for obtaining each
in a state of purity.

Red Modljication of Anilidotrinitrophenyltartronic Ester.

To obtain this form from the crude product of the reaction it was
crystallized several times from alcohol, until a mixture of the two forms
rich in the red had been obtained, and then a strong alcoholic solution
of the mixed crystals was allowed to evaporate slowly at temperatures
from 50° to 70°. The product thus obtained, which showed no sign
of the yellow form, was dried at 100°, and analyzed with the following
results : —

I. 0.1792 gr. of the substance gave on combustion 0.3114 gr. of car-
bonic dioxide and 0.0694 gr. of water.
II. 0.2016 gr. of the substance gave on combustion 0.3534 gr. of
carbonic dioxide and 0.0809 gr. of water.
III. 0.1938 gr. of the substance gave 19.9 c. c. of nitrogen at a tem-
perature of 22° and a pressure of 767.7 mm.

III.

11.76

The air-dried red crystals lost only 0.19 per cent when dried at 100°,
and therefore were free from water or alcohol of crystallization.

Properties. — The red form of anilidotrinitrophenyltartronic ester
crystallizes in well formed prisms, often half a centimeter long, with
at each end a pyramidal termination consisting of four planes. It has
an orange-red color like that of potassic dlchromate, and melts at
143° ; is essentially Insoluble in cold water, very slightly soluble in
boiling water, forming a pale yellowish solution ; soluble in cold al-
cohol, rather freely in hot. This solution if allowed to evaporate at
temperatures from 50° to 70° deposits crystals of the red modification,
as has been already stated ; but if allowed to evaporate at ordinary
temperatures, a mixture of red and yellow crystals is usually obtained,
although occasionally only red crystals are formed under these con-

Carbon

Calculated for
C6H{C6H5NH)(N0o),C0H(C02C2H5)2.

47.69

I.

47.40

Found.
II.

47.80

Hydrogen

3.77

4.30

4.46

Nitrogen

11.71

84 PROCEEDINGS OF THE AMERICAN ACADEMY

ditions. The red modification is more soluble in methyl than ethyl
alcohol ; very freely in acetone. Both these solutions deposit some
yellow crystals. Soluble in ether or chloroform ; slightly soluble in
cold benzol, freely in hot. The solution in benzol or chloroform de-
posits the substance in a viscous state. It is slightly soluble in cold
glacial acetic acid, more soluble in hot ; from this solution it is usually
precipitated by the addition of water in the yellow form, although
once or twice we thought we obtained the red modification ; very
slightly soluble in carbonic disulphide ; insoluble in ligroiue. Strong
sulphuric acid when cold has no action on it, but dissolves it with a
brownish yellow color when hot ; strong nitric acid dissolves it par-
tially when cold with a yellow color, and the action is increased by
heating ; cold hydrochloric acid has no action on it, but dissolves it
very slightly when hot.

Telloic Modification of Anilidotrinitrophenyltartronic Ester.

This substance was most conveniently obtained from the mixture of
the two forms, after it had been purified by several crystallizations
from alcohol, by dissolving it in warm glacial acetic acid, and, after
allowing the solution to stand for some hours, precipitating the solid
matter by the addition of water. The yellow powder thus obtained
was dried at ordinary temperatures over sulphuric acid, and analyzed
with the following results : —

I. 0.1879 gr. of the substance gave on combustion 0.3284 gr. of car-
bonic dioxide and 0.0726 gr. of water.
II. 0.1924 gr. of the substance gave 20.5 c. c. of nitrogen at a tem-
perature of 25° and a pressure of 769.7 mm.

Calculated for Found.

CeH(CuH5NU)tN0.)3C0U(C02C2H5)2. I II.

Carbon 47.69 47.66

Hydrogen 3.77 4.29

Nitrogen 11.71 12.05

The substance dried over sulphuric acid lost no weight when heated to
100°, showing that it contams neither water nor alcohol of crystal-
lization.

Properties. — The yellow form crystallizes in very fine needles
united into spherical groups like those of wavellite, looking frequently
like little balls of fat, and has a full yellow color like that of potassic
chromate. It is hard to determine the melting point with accuracy,
because this form is quickly changed into the red modification, at least

OF ARTS AND SCIENCES. 85

in part, when heated to a few degrees below that temperature ; it is
necessary therefore to heat the bath up to the meltiug point before im-
mersing the capillary tube containing the substance ; under these con-
ditions it melts, and then changes to the red form, and solidifies again,
after which it does not melt until about 140°. The melting point we
ascribe to this substance is 122°, but we are not sure that this is ac-
curate to one degree. This change from yellow to red can also be
brought about slowly by heating at 100°, by boiling the yellow form
with water for some time, or most conveniently by crystallization from
alcohol at temperatures from 50° to 70°. In its solubility in the differ-
ent solvents the yellow form does not differ much from the red, but
seems in general to be more soluble.

We have made many attempts to determine the molecular weights
of the two modifications by the method of Raoult, but have not as yet
succeeded in finding any solvent which gives satisfactory results.
Acetic acid, owing to its conversion of the red into the yellow form,
could at best give results only for the latter, but a number of deter-
minations showed us that it was of no use even for this purpose, as
results were obtained varying from 295 to 345 (the theoretical molecu-
lar weight is 478), therefore showing that some reaction (probably
the formation of an acetoxymalonic ester) had taken place between
the acetic acid and the ester. An experiment with phenol gave an
excellent result with the red form, 477 instead of 478 ; but with the
yellow only 307 was obtained, indicating a chemical action similar to
that observed with acetic acid, and we are not inclined to accept any
result with phenol, however excellent it may appear, as with a sub-
stance which is so easily affected by solution as this there can be no
certainty that it has remained in the same modification, unless it can
be recovered directly from the solution used, and this we have found
no means of doing in the case of phenol. Benzol gives such uncer-
tain results with substances containing a hydi'oxyl group,* that we did
not think it worth while to try it. We have hopes that naphthaline!
may give the desired result for the red form ; this will be tried in this
Laboratory, and a further attempt made to find a satisfactory solvent
for the yellow modification ; but as the departure of one of us from
Cambridge makes it necessary to postpone further work in this direc-
tion until next year, we have thought it best to publish at once our
present results, which have already reached a certain completeness.

* Ber. d. ch. G., XXI. 707. t Ber. d. ch. G, XXII. R. 128.

86 PROCEEDINGS OF THE AMERICAN ACADEMY

Saks of AniUdotrinitrotartronic Ester.

The anilidotrinitrotartronic ester has well marked acid properties.
An aqueous solution of acid sodic carbonate has essentially no action on
it, but if alcohol is added there is a slight but distinct action indicated
by the change of color ; potassic carbonate gives in aqueous solution a
slight red color, which becomes a strong red if alcohol is added ; sodic
hydrate in aqueous solution gives a very dark red, apparently convert-
ing the substance completely into its salt ; ammonic hydrate also gives
a strongf red color, which is intensified on the addition of alcohol.
Some of the alkaline salts were studied more carefully with the fol-
fowing results.

Monopotassium Salt, C6H(CoH5NH)(NO,)3COK(COOCoH5)o.

This salt was made by adding 20 c. c. of absolute alcohol to 0.5 gr.
of the ester, and then a large excess of pure potassic carbonate ; car-
bonic dioxide was given off, the solution became very dark red, and
after disestinir the substances for eight to ten minutes with the occa-
sional aid of a little heat, the dark brown solution was filtered from the
excess of potassic carbonate, evaporated to dryness in a beaker sunk
throughout its whole length in the steam bath, and analyzed with the
following results: —

I. 0.3458 gr. of the salt heated with strong sulphuric acid gave
0.0548 gr. of potassic sulphate.
II. 0.3748 gr. of the salt gave 0.0G9G gr. of potassic sulphate.

Calculated for Found.

C«H(C„H5NH)(N0,)3C0K(C0AH5)2- I- H.

Potassium 7.58 7.11 8.34

The absolute alcohol used in the second analysis had not been
freshly prepared, and the slight excess of potassium can be accounted
for by the assumption that it had absorbed a little water.

Properties. — The acid potassic anilidotrinitrophenyltartronic ester,
when prepared by the method described above, forms a brownish
black amorphous solid, freely soluble in water or alcohol, slightly sol-
uble in ether, and insoluble in benzol. The solutions have a dark
brown color.

Disodium Salt, CoH(C„Il5NNa)(NOo)3CONa(COOCoH5)2.

This salt seemed to be formed, to our great surprise, by the action
of an alcoholic solution of sodic hydrate on an excess of the anilido-
trinitrophenyltartronic ester. To prepare it, about 0.7 gr. of the ester

OP ARTS AND SCIENCES.

87

were mixed with a little alcohol, and somewhat less than the amount
of sodic hydrate (also in alcoholic solution) necessary to form a mono
sodium salt ; a little ether was then added to the dark red solution,
which was evaporated rapidly to dryness, the small beaker containing
it beiuff sunk throuajhout its leuijth in the steam bath. After this the
excess of unaltered ester was extracted with benzol, and the salt dried
at 100°, and analyzed with the following results: —

I. 0.2596 gr. of the salt gave after evaporation and ignition with

sulphuric acid 0.0672 gr. of sodic sulphate.
II, 0.2848 gr, of the salt gave 0.0698 gr. of sodic sulphate.

Calculated for
CoU(C„HiNNa)(NO,),CONa(C02C2HB)2.

Sodium 8.81

Found.

I.

8.39

II.

7.94

These results agree with the percentage corresponding to the for-
mula as nearly as can be expected, when it is remembered that the
salt was not crystallized, or purified in any other way ; but still we
cannot feel that they do more than make it highly probable that this
is the composition of the salt, as it is certainly strange that a neutral
salt should have been formed when such an excess of the acid sub-
stance was present, and we had no means of proving that the substance
analyzed was a pure salt rather than a mixture ; in fact, we observed
a slight smell of phenylisocyauide during its preparation, indicating a
deep decomposition of part of it, — only a very small part, however, if
we may judge from the smell.

Properties. — The salt prepared as described above formed a dark
reddish brown to black amorphous mass, soluble in water or alcohol ;
slightly soluble in ether ; insoluble in benzol. Its solutions have a
brownish red color.

The behavior of a solution of the ammonium salt of anilidotrinitro-
phenyltartronic ester, made by adding ammonic hydrate to an excess
of the ester, although it smelt strongly of ammonia, was tested with
various reagents and gave the following characteristic precipitates : —

With a zinc salt, orange-brown.

With a manganese salt, brown.

With a cadmium salt, reddish brown.

With a copper salt, yellowish brown.

With a lead, mercuric, or silver salt, reddish brown.

The fact that the ammonium salt gives no precipitate with salts of
magnesium, calcium, strontium, or barium, is also highly characteristic.

Although, as has been already stated, no definite results were ob-

88 PROCEEDINGS OF THE AMERICAN ACADEMY

taiued from our attempts to oxidize or saponify the biomtrinitrophe-
nyltartrouic ester, we thought that perhaps the anilido comjjound
might behave better, and accordingly the following experiments were
made, in the hope (unfortunately not realized) of decomposing the
anilidotrinitropheuyltartronic ester into substances which would con-
firm our inferences in regard to its nature. The substance was al-
lowed to stand in the cold with an aqueous solution of potassic
permanganate, but, although a considerable part of it disapj^eared,
no organic oxidation product could be detected. In the hope of sa-
ponifying the ester, we added to it an excess of sodic hydrate dissolved
in water, and allowed the reddish brown solution thus formed to stand
in a corked flask at ordinary temperatures for some weeks. During
this standing a strong odor of phenylisocyanide was developed, and,
if the amount of ester was small, the color changed to yellow ; if, on
the other hand, the quantity was large, it retained its dark brown color ;
at the end of the experiment, nothing was obtained except a brown
precipitate of most unpromising appearance, formed by adding an acid
to the solution, and the phenylisocyanide already mentioned as recog-
nized by its smell. As the formation of this substance would neces-
sitate a complete destruction of the benzol ring which carried the
nitro groups and malonic ester radical, we did not think it worth
while to repeat the experiment.

This formation of an isocyanide from the destruction of a benzol
ring containing nitro groups by means of an aqueous solution of sodic
hydrate recalls the work of Post and Hiibncr,* who found that ordi-
nary dinitrobenzol when boiled with a solution of sodic or potassic
hydrate was decomposed with formation of a cyanide quickly if the
solution was strong, slowly if it was dilute. They also found that
picric acid behaved in the same way, thus confirming the earlier ob-
servation of Wohler.f

A n ilidotrin itroph enjjJmalon ic Este r,
CoHlCcH.NH) (N0J3CII(C00C JI^)..

This substance was made to see whether it would be possible b.y
treatment with nitric acid to convert it directly into either its nitrite
or the anilidotartronic ester described above. It was found, however,
that standing for several days at ordinary temperatures with strong
nitric acid produced complete decomposition of part of the substance,
the only product insoluble in the acid being unaltered anilidotrinitro-

* T5cr. d. cli. G., V. 408 (1872). t Pogg. Ann., XIII. 488 (1828),

OP ARTS AND SCIENCES. 89

phenylmalonic ester ; and that upon heating it on the steam bath with
nitric acid lor three hours in the hope of making the anilidotrinitro-
phenyltartronic ester, the substance was destroyed completely, oxalic
acid being the only product which we could find. It seems therefore
that the less acid auilido compound cannot be converted into the ni-
trite or tartronic ester. But although these attempts to oxidize the
substance have failed, we add the description of the anilidotrinitro-
phenylmalonic ester, which has not been prepared heretofore.

It is easily made by adding an excess of aniline to solid bromtrini-
trophenylmalouic ester. The reaction ran smoothly in the cold with
slight evolution of heat, and the product was purified by washing
with very dilute hydrochloric acid to remove the excess of aniline
and the aniline bromide, and crystallization from alcohol until it
showed the constant melting point 133°, when after drying at 100°
it was analyzed with the following result : —

0.18G6 gr. of the substance gave 20.2 c. c. of nitrogen at a temperature
of 25°. 5 and a pressure of 764.3 mm.

Calculated for
0(.H(C6H5NH)cNO„)3CH(CO2C2H5)2. Found.

Nitrogen 12.13 12.12

Properties. — The anilidotrinitrophenylmalonic ester crystallizes in
long slender sharp needles arranged in radiating bunches. It has a
full yellow color, and melts at 133°. It is very slightly, if at all,
soluble in cold water, more soluble in hot, as shown by the faint yel-
low color of the solution ; freely soluble in hot alcohol, less so in cold ;
somewhat more soluble in methyl than in ethyl alcohol ; very freely-
soluble in chloroform; freely in benzol or acetone; soluble in ether,
or glacial acetic acid ; slightly in carbonic disulphide ; insoluble in
ligroine. Boiling alcohol is the best solvent for it. Strong sulphuric
acid or nitric acid dissolves it slightly in the cold with a yellow color;
the solubility is somewhat increased by heating ; strong hydrochloric
acid acts on it only very slightly, cold or hot. An aqueous solution
of acid sodic carbonate has no action upon it, but there is a barely per-
ceptible change of color if alcohol is added ; aqueous potassic carbon-
ate gives little if any action, but on the addition of alcohol a brown
solution is formed, the action however seems to be incomplete ; aque-
ous sodic hydrate turns the crystals dark and brown, the solution be-
coming yellow, the addition of alcohol seems to convert the substance
completely into the sodium salt ; aqueous ammonic hydrate gives a
barely perceptible yellow color, which is not increased by the addition
of a little alcohol, but a large amount gives a dark brown solution.

90 PROCEEDINGS OF THE AMERICAN ACADEMY

A solution of the sodium salt of anilidotrinitrophenylmalonic ester
made by adding a di'op of sodic hydrate to an excess of the ester mois-
tened with alcohol, and then diluting largely with water, gave charac-
teristic precipitates with the following reagents : —

With a calcitim salt, a heavy reddish brown precipitate.

With a strontium salt, a slighter precij^itate.

With a barium salt, a very faint precipitate.

A magnesium salt gives a reddish brown precipitate, as heavy as
that obtained with the calcium salt.

With a manganese salt, reddish brown flocks.

With a zinc salt, yellowish brown.

W^ith a cadmium salt, orange-yellow.

With a copper salt, yellowish brown.

With a lead salt, reddish brown.

With a silver salt, dark reddish brown.

Its most striking property is that the barium salt is the most, the
calcium salt the least, soluble of its salts with metals of the second
group, and in tiiis respect it resembles its mother substance, the brom-
trinitrophenylmalonic ester,* the corresponding acetacetic ester^f and
the orthonitrobenzoylmalouic ester of Bischoff.f

Nitrite of Trinitrophenylenedlmalonic Ester,
.C,H(NO,)3CH(COOC,H.)2CNO,(COOC2H5)2.

After we had studied the action of nitric acid on the bromtrinitro-
pheuylmalouic ester, it seemed of interest to determine whether the
dimalouic compound acted in the same way, and accordingly we pro-
ceeded as follows. A small quantity of the trinitrophenylenediraalonic
ester § (melting-point 123°) was covered with nitric acid of specific
gravity 1.38, and warmed in a dish on the steam bath for two min-
utes ; the solid turned dark yellow, and melted to a drop of oil. The
acid was then allowed to cool, poured off, and replaced by the same
quantity of fresh acid, when it was warmed as before for three min-
utes, making five m all. Upon standing, the oily product solidified to
a mass of crystals, which were washed thoroughly with water, and at
first recrystallized from alcohol ; but as m each crystallization a small
quantity of yellow oil was formed, we feared a partial decomposition,
and resorted to the method which had given excellent results in the
purification of the nitrite of bromtrinitrophenyhnalonic ester, that is,

» These Proceedings, XXIV. 261. t Ann. Cliem., CCLI. 362.

t Ibid., 278. § Tiiese Proceedings, XXIV. 268.

OP ARTS AND SCIENCES.

91

dissolving tlie substance with very little heat iu chloroform, and then
adding enough alcohol to start the separation of crystals ; in this way
a pure substance melting at 111° was obtained without diliiculty, which
was dried at about 70° and analyzed with the following results ; —

I. 0.19G0 gr. of the substance gave on combustion 0.3012 gr. of car-
bonic dioxide and 0,0714 gr. of water.
II. 0.1926 gr. of the substance gave 1G.7 c.c of nitrogen at a tem-
perature of 23° and a pressure of 771 mm.

Found
1. II.

41.91
4.05

Carbon
Hydrogen

Nitrogen

Calculated for
C8H(N02)3CH(C0,C2ll5),CN0s(C02C2Hr,)2.

41.81
3.83

9.76

9.93

The substance is therefore the mononitrite of trinitrophenylenedi-
malouic ester.

Properties. — The substance crystallizes in rather thick plates with
parallel sides terminated at each end by two planes at an acute angle to
each other, these acute angles being usually, but not always, truncated
by planes at right angles to the parallel sides. The crystals are often
much twinned and grouped into very irregular forms. It has a lemon-
yellow color, and melts without decomposition at 111°. It is very
slightly soluble in boiling water, giving a pale yellow solution, essen-
tially insoluble in cold ; slightly soluble iu cold alcoliol, freely in hot ;
more soluble in methyl than in ethyl alcohol, whether cold or hot ;
both these solvents seem to produce a slight decomposition of the sub-
stance when heated with it ; very freely soluble in chloroform ; freely
soluble in benzol, or acetone ; soluble in ether, or glacial acetic acid ;
slightly soluble in carbonic disulphide ; insoluble in ligroine. The
best solvent for it is the mixture of chloroform and alcohol used as de-
scribed above. Strong sulphuric acid has no action on it in the cold,
but, when heated with it, dissolves some of it with a pale yellow color
after it melts; strong hydrochloric acid has no apparent action on it
either hot or cold ; strong nitric acid has no action on it in the cold,
but dissolves it slightly when hot, and if heated with it for two hours
and a half destroys it completely, tlie only product which we isolated
being oxalic acid ; on one occasion, however, another product melt-
ing near 1 10° was obtained in small quantity, but we did not feel
sufficiently interested in this part of the subject to prepare enough of it
to determine whether it was the dinitrite or a tartronic ester. One
thing, however, these experiments have established, namely, that the

92 PROCEEDINGS OF THE AMERICAN ACADEMY

dimalonic compound is less stable toward nitric acid than the brom-
malouic compound, since the latter, after three hours' boiling with the
acid, had not gone further than the tartronic ester, while the former
usually underwent a total decomposition. On the other hand, the
nitrite of the dimalonic compound is not decomposed by melting, and
only slightly by heating with alcohol, and therefore in these two re-
spects is more stable than the nitrite of bromtrinitrophenylmalonic
ester.

The nitrite of trinitrophenylenedimalonic ester shows acid proper-
ties. With an aqueous solution of acid sodic carbonate it gives no
action, and very little when alcohol is added, unless in very large
quantity, when a yellowish solution is formed ; with potassic carbonate
in aqueous solution there is no action, but on the addition of alcohol a
very strong yellowish brown color ; aqueous sodic hydrate gives a red-
dish solution, on the addition of alcohol an orange flame-colored solution,
entirely different in color from that given vv'ith the potassic carbonate ;
ammonic hydrate imparts a strong yellow coloration to the solution,
intensified on the addition of alcohol. Upon comparing the acidity of
this substance with that of the trinitrophenylenedimalonic ester, it
seems as if this latter substance were somewhat more acid than the
nitrite, since it gives a slight coloration with aqueous potassic carbon-
ate, and seems to act more easily with acid sodic carbonate and alco-
hol, but the difference between the two in this respect is certainly very
slight, which we should not have expected, as it seemed probable that
the introduction of the nitrous acid radical (ONO) would have in-
creased the ease with which the hydrogen in the other malouic ester
radical was removed.

A solution of the sodium salt of the nitrite of trinitrophenylenedi-
malonic ester was made by adding one drop of sodic hydrate solution to
a large excess of the ester moistened with alcohol, and, after the action
had taken place, diluting with much water. Tlie solution thus ob-
tained was of the color of a solution of potassic dichromate, but much
less stronpjly colored than the solutions of the salts of any of the re-
lated substances. Its action with the various reagents was tried, and
the following characteristic precipitates observed ; —

With barium salt, rather heavy yellow flocks.

With morcurous or lead salts, heavy yellow flocks.

With silver or copper salts, yellow flocks.

Salts of calcium, sU'ontium, or magnesiimi gave only faint yellow pre-
cipitates, decidedly different from the heavy one produced by baric
chloride.

OF ARTS AND SCIENCES.

93

Nitrite of Bromdinitrophenylmalonic Ester,
CaIl2Br(N02)oCN02(COOC2lIii)o.

Although bromdinitrophenylmalonic ester does not change in color
when heated with nitric acid, as has been stated more than once in
previous papers* from this Laboratory, it really is affected in the same
way as the corresponding trinitro compound, being converted into its
nitrite, but without the formation of the i-ed secondary product which
made the reaction so striking in that case. The substance was pre-
pared as follows. A small quantity of the bromdinitrophenylmalonic
ester (melting point 75° -76°) was heated on the steam bath with nitric
acid of specific gravity 1.38 for five minutes, or longer (as the same
compound was obtained if the heating was continued three hours) ;
there was at first no sign of a reaction except a slight evolution of
nitrous fumes ; but, as the heating continued, the undissolved solid
melted, forming an oil drop, which after cooling solidified to a mass of
crystals, while at the same time the acid liquid deposited crystals look-
ing very much like those of the bromtrinitrophenyltartronic ester.
All of these crystals were purified by recrystallization from boiling
alcohol, until they showed the constant melting point 111°, when they
were dried at about 70°, and analyzed with the following results : —

I. 0.1973 gr, of the substance gave on combustion 0.2492 gr. of

carbonic dioxide and 0,0510 gr. of water.
11. 0.2084 gr. of the substance gave 17.4 c. c. of nitrogen at a tem-
perature of 22° and a pressure of 755.6 mm.

Calculated for Found

C6H,,Br(NO„)2CNOo(C02C2Hi;)2. I. II.

Carbon 34.66 34.44

Hydrogen 2.67 2.87

Nitrogen

9.33

9.41

The substance therefore is the nitrite of the bromdinitrophenyl-
malonic ester, and not the corresponding tartronic ester, as we had
expected from the long heating used in one of the methods of
preparation, and from the fact that no difficulty was encountered in
making the combustion of it.

Properties^ — The nitrite of bromdinitrophenylmalonic ester crys-
tallizes usually in thick rhombic crystal often one miUimeter m each
direction, which look somewhat like rhombohedra with a sharp acute
angle, but are seen to be twins by the lines of twinning and stria-

* These Proceedings, XXIV. 6, 257.

94 PROCEEDINGS OF THE AMERICAN ACADEMY

tions ; the obtuse angles on these crystals are often truncated. Less
commonly with the characteristic forms just described, long flattened
prisms occur, which are terminated by a single plane at a shfrp
acute angle, rarely by two. It is easy to see that the rhombic
crystals could be formed by the twinning of these prisms. The
crystals are very lustrous, of a white color, with a slight greenish
cast, and melt at 111° without decomposition. The substance is es-
sentially insoluble in cold water, very slightly soluble in hot; soluble
in cold alcohol, more so, but still far from freely, in hot. It shows no
signs of decomposition when boiled with alcohol. It is more soluble
in methyl than in ethyl alcohol, cold or hot ; very freely soluble in ben-
zol, chloroform, or acetone ; freely soluble in glacial acetic acid ; solu-
ble in ether or carbonic disulphide; nearly, if not quite, insoluble in
ligroine. Hot alcohol is the best solvent for it. Strong sulphuric
acid has no action on it in the cold, but when hot dissolves a little of
the substance after it has melted ; strong hydrochloric acid has no
action, hot or cold ; strong nitric acid has no action on it in the cold,
but when hot dissolves a little of it, which is deposited on cooling.
Neither acid sodic carbonate nor neutral potassic carbonate had any
action with it, even in presence of alcohol ; sodic hydrate in aqueous
solution was without action ; if alcohol was added, a portion of the
substance dissolved slowly with a yellow color, but most of the white
crystals were left unattacked ; ammonic hydrate in aqueous solution
had no action, with alcohol little or none. The substance behaves,
therefore, as we should expect, a salt being formed only by a reagent
like sodic hydrate strong enough to remove the NO2 group attached
to the side-chain

The nitrite of bromdinitrophenylmalonic ester is much more stable
than the corresponding trinitro compound, since it is not decomposed
by boiling with alcohol, or at its melting point, or by boiling with
nitric acid ; for, as has been already stated, it could be made by boil-
ing for three hours with nitric acid, whereas under these conditions the
trinitro compound was converted into the tartronic ester. Nor did
longer boiling of the dinitro compound with nitric acid produce the
tartronic ester, as even after seven hours it showed the melting
point of the unaltered substance 111°. As the trinitro tartronic
ester had been obtained also by heating the corresponding nitrite,
we tried the same experiment with the dinitro nitrite, and found that,
when heated a few degrees above its melting point, it turned rather
dark colored, and gave off bubbles of gas, in which bromine was rec-
ognized by the smell. The residue was oily, but after solution in

OF ARTS AND SCIENCES.

95

alcohol crystals were obtained which showed the characteristic rhombic
form of the uitrite of broradinitrophenylmalouic ester. This experi-
ment therefore seemed to show that the action of heat consisted only
in the complete decomposition of a portion of the nitrite of bromdini-
trophenylmalonic ester, and, as it did not promise to give the desired
tartronic ester, further work in this direction was abandoned.

Reduction of the Nitrite of Bromdinitrophenylmalonic Ester.

The conversion of the nitrite of broratrinitrophenylmalonic ester into
the corresponding tartronic ester by nitric acid, or by the action of
heat, indicates that the group NO2 in the side-chain is attached to
the molecule by oxygen instead of nitrogen, or, in other words, that
the substance is a nitrite, and not a nitro compound ; but, as we could
not consider this a conclusive proof of the oxygen attachment, we
have studied the reduction of a body of this class, selecting for this
purpose the dinitro compound, because it is more easily prepared than
the corresponding substance containing three nitro groups, and also
because the product obtained from it by reduction would probably be
more stable than one containing one more amido group.

Six grams of the nitrite of bromdinitrophenylmalonic ester divided
into three lots of two grams each were mixed with granulated tin,
strong hydrochloric acid, and a few drops of alcohol, and after adding
a piece of platinum to accelerate the reaction, were allowed to stand
on a steam radiator (50°-7U°) until all the organic matter had dis-
solved, and no further action was observed, which happened usually in
about an hour and a half. If quantities larger than two grams were
used a very dark colored solution was obtained. The solution poured
off from the excess of tin was freed from stannous and stannic chlorides
by means of sulphuretted hydrogen, when a residue was obtained by
evaporation of the filtrate, which gave off ammonia gas when treated
with an alkaline hydrate, and formed a precipitate of ammonic chlor-
platinate with chlorplatinic acid ; it evidently therefore contained am-
monic chloride in addition to the chloride of the organic base. The
washings of the suljihide of tin, which were worked up separately
from the filtrate, on the other hand, yielded crystals which showed only
a very slight amount of ammonic chloride by the same tests, and ac-
cordingly these crystals were dried at 100°, and analyzed with the
following results : —

I. 0.1980 gr. of the substance gave 25.25 c. c. of nitrogen at a tem-
perature of 23° and a pressure of 764.8 mm.

96 PROCEEDINGS OP THE AMERICAN ACADEMY

II. 0.1 954 gr. of the substance gave by the method of Carius 0.1480 gr.
of arijeutic chloride.

Calculated for

Found.

CeHsNHsClCCUOIlCONH).

I. 11.

Nitrogen

13.96

1447

Chlorine

17.70

18.7

These results are not all that we could wish ; but although they show
the presence of the small amount of ammouic chloride, which we had
detected by the qualitative tests, they leave no doubt in regard to the
composition of the organic substance, and therefore we have not
thought it worth while to spend the great amount of time which would
have been necessary to prepare a sample entirely free from ammonic
chloride ; especially as these analyses prove that the group NO2 is
attached to the side-chain by oxygen, since we obtained, by the reduc-
tion of the nitrite of bromdinitroj^lienylmalonic ester, ammonic chlo-
ride and the chloride of amidoxyoxindol instead of the chloride of
diamidoxindol, which would have been formed if the group NO2 had
been attached to the molecule by the nitrogen.

Properties of the Chloride of the Amidoxyoxindol,
C6H3NH3Cl(CHOHCONII).

This substance was obtained crystallized in rather large plates, usu-
ally in forms like a closed fan terminated by an obtuse angle, aud set
in rows one over the other with the obtuse angles parallel, or arranged
in branching arborescent forms like coral ; sometimes in thicker sharp
prisms. It had a dark yellow color as we observed it ; was soluble in
cold water, more so in hot ; slightly soluble in cold or hot alcohol,
and insoluble, or nearly so, in most of the other solvents. The strong
acids gave no striking reactions with it, except strong nitric acid,
which turned it orange-red ; but this seemed to be due to some nitrous
acid in the nitric, as on the addition of sodic nitrite the color was
much intensified. Sodic hydrate added to the aqueous solution gave
a few brown flocks, evidently from decomposition of a part of the base ;
ammonic hydrate gave a tolerably heavy flocculeut precipitate, at first
whitish, but turning brown on exposure to the air, but it showed no
signs of crystallization, and with the small amount of substance at our
disposal we have been unable to obtain the free base in any definite
form. The chloride mixed with alcohol and some strong hydrochloric
acid imparted after some time a dark red color to a piece of pine
wood.

OF ARTS AND SCIENCES.

97

Attempt to make Dinitrophenylenedlmalonic Ester.
The results just described having shown that the behavior of brom-
(liuitrophenylmalonic ester toward nitric acid was not so different from
that of the trinitro compound as we had at first supposed, it became
of interest to determine whether other apparent differences between
these two bodies were no better marked, and accordingly we tried
to make the dinitrophenylenedlmalonic ester by the method which
had given good results with the trinitro compound as follows : —
3 gr. of bromdinitrophenylmalonic ester were dissolved in ether, and,
after the addition of 1.5 gr. of malonic ester previously treated with
0.15 gr. of sodium in a large quantity of alcohol, the mixture was
boiled under a return-condenser for three hours ; at the end of this
time the solution had turned deep red, and a little solid had separated.
The product was treated with water and dilute sulphuric acid, the
ether separated, and the aqueous liquid tested for sodic bromide, which
it was found to contain in small quantity, but the crystals deposited
by the ether melted after one crystallization at 75° -7 6°, the melting
point of bromdinitrophenylmalonic ester. We infer from this experi-
ment, therefore, that, although a little dinitrophenylenedlmalonic ester
may have been formed, its quantity was exceedingly small, and that
the reaction, if it goes on at all, certainly takes place with much
more difficulty in the case of the dinitro than in that of the trinitro
compound.

VOL. XXVI. (n. s. xviii.)

98 PROCEEDINGS OF THE AMERICAN ACADEMY

VIII.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OP

HARVARD COLLEGE.

NOTE ON TRIBROMMONONITROBENZOL.
By C. Loring Jackson and W. B. Bentley.

Presented May 13, 1891.

The results of experiments on the behavior of tribromtrinitrobenzol
and tribromdiiiitrobenzol with various reagents have been described in
a number of papers from this Laboratory. For the sake of complete-
ness, we felt it necessary to include the tribrommononitrobenzol in our
investigation, and were the more inclined to do this as we hoped that
it might throw some light on the replacement of bromine by hydrogen,
so often observed in our work with the trinitro and dinitro compounds.
This hope has not, however, been fulfilled, as no such replacement of
bromine by hydrogen was observed, and the tribrommononitrobenzol
has proved to be so inert with various reagents, that it deserves only
the short investigation the results of which are given in this paper.

The tribrommononitrobenzol used in this work was prepared by
boiling symmetrical tribrombenzol (melting point 119°) for 'fifteen
minutes with a nitric acid made by adding to fuming nitric acid
(specific gravity 1.51) one quarter of its volume of common strong
nitric acid (specific gravity 1.38). The greater part of the tribrom-
mononitrobenzol separated as the mixture cooled, and the rest was
obtained by the addition of water to the acid mother liquor. This
process gave better results than when acetic acid was used to dilute
the fuming nitric acid according to the method given by V. von
Richter.*

Action of Tribrommononitrohenzol with Sodic Ethylate.

The tribrommononitrobenzol melting at 125° was acted on by hot
sodic ethylate giving a substance which we found it convenient to

* Ber. d. eh. G., VIII. 1426.

OP ARTS AND SCIENCES.

99

prepare as follows : — 5 gr. of tribrommononitrobenzol dissolved in
benzol were mixed with an alcoholic solution of the sodic ethylate
from 1 gr. of sodium, heated on the water bath just below boiling
for some time, and then allowed to stand at a temperature a little
above the ordinary for several hours. The red solution thus obtained
was filtered from a precipitate which had formed, (this gave a test
for bromide with argentic nitrate,) diluted with water, and acidified
with dilute sulphuric acid, after which the benzol was separated from
the aqueous liquid ; this was extracted with ether, and the combined
extracts from the organic solvents purified by crystallization from
alcohol until the constant melting point 91° was reached. Sometimes
a black tarry product was obtained from the reaction, instead of the
white crystals formed if the process had gone well. This tarry mass
was worked up most conveniently by distillation with steam, followed
"by crystallization of the steam distillate from alcohol. The pure
substance was dried at about 70°, and analyzed with the following
results : —

I. 0.1900 gr. of the substance gave by the method of Carius 0.2181
gr. of argentic bromide.
II. 0.3800 gr. of the substance gave 15.9 c. c. of nitrogen at a tem-
perature of 22° and a pressure of 756.5 mm.

Calculated for

CeHjBraNOjOCjHg.

I

Bromine

49.22

48.85

Nitrogen

4.31

Found.

II.

4.72

The substance is therefore dibromnitrophenetol, and was formed by
the replacement of one atom of bromine by the ethoxy radical.

Properties. — The dibromnitrophenetol crystallizes in bundles of
prisms, which are usually somewhat flattened and terminated by a
basal plane, also occasionally showing with this two other planes at an
obtuse angle to each other. It is white, but turns brown on exposure
to the air; melting point 91° ; essentially insoluble in cold water, per-
haps very slightly soluble in hot ; alcohol disolves it sparingly when
cold, freely when hot ; more soluble in methyl than in ethyl alcohol ;
very freely soluble in benzol, chloroform, acetone, or carbonic disul-
phide ; freely soluble in ether ; slightly soluble in cold glacial acetic
acid, soluble in hot ; very slightly soluble in ligroine. The best solvent
for it is hot alcohol. It distils with steam. Strong sulphuric acid
has no action in the cold, but when hot gradually forms a black solu-
tion ; cold nitric acid has no action on it, but when hot dissolves it,

100 PROCEEDINGS OF THE AMERICAN ACADEMY

depositing crystals on cooling, which seem by their melting point to be
the unaltered substance ; strong hydrochloric acid seems to have no
action, whether hot or cold ; sodic hydrate seems not to act upon it.

Action of other Reagents on TribrommononitrohenzoL

Sodic methylate behaved in much the same way as the sodic ethyl-
ate, except that it was necessary in this case to heat for a longer
time, and even after this a certain amount of tribrommononitrobenzol
was apt to be left unaltered. The product after crystallization from
alcohol showed the constant melting point 104°, but its analyses gave
percentages of bromine differing from those corresponding to dibrom-
nitroanisol by about one per cent. The cause of this is probably the
same as that of the similar want of agreement between the calculated
and observed percentages in the case of the bromdinitroresorcine
dimethylether ; * but we have not thought the substance of sufficient
importance to try to obtain better analytical results, which, to judge
by our work on the bromdinitroresorcine dimethylether, would be a
matter of great difficulty.

From these experiments it appears that the tribrommononitrobenzol
acts with sodic alcoholates less easily than the tribromdinitrobenzol,
which is attacked by them in the cold forming the ethers of bromdi-
nitroresorcine, whereas it is necessary to heat the mononitro com-
pound to bring about any action. When the tribromdinitrobenzol is
heated with an alcoholate, the third atom of bromine is replaced by
hydrogen and an ether of the dinitroresorcine is formed, but we found
no trace of any action of this sort with the mononitro compound. It
is to be noted, too, that only one atom of bromine is removed from
the mononitro, but two from the diuitro compound.

The inertness of tribrommononitrobenzol as compared with the cor-
responding dinitro compound is made even more evident by the study
of the action of aniline upon it. This substance converts tribromdi-
nitrobenzol into trianilidodinitrobenzol, when heated with it for a short
time, whereas the corresponding trinitro compound is formed in the
cold ; from the mononitro substance we were unable to obtain any
anilido compound, even after heating for some time to the boiling
point of aniline ; a little aniline bromide was formed, it is true, but
almost all of the tribrommononitrobenzol was recovered unaltered,
as recognized by its melting point and very characteristic crystalline
form.

* These Proceedings, XXV. 175.

OF ARTS AND SCIENCES. 101

No better result was obtained with sodium malonic ester, which
acts on the dinitro and trinitro compounds in tiie cold, but even after
boiling for some time with the tribrommononitrobenzol yielded only a
very small quantity of sodic bromide, while almost all of the tribrom-
mononitrobenzol was recovered unaltered.

102 PEOCEEDINGS OF THE AMERICAN ACADEMY

IX.

ON A KEPHIR-LIKE YEAST FOUND IN THE
UNITED STATES.

By Charles L. Mix.

Presented by Professor W. Q. Farlow, May 26, 1S91.

A SPECIES of 3'east which causes alcoholic fermeutation of n)ilk is
well known in Europe, the attention of the leading scientists having
been called to it by Edouard Kern, in an article published during
November, 1881, and entitled, " Ueber ein neues Milch-ferment aus
dem Kaukasus." In order to give an intelligible description of a
similar ferment which exists in the United States, a summary of
Kern's paper becomes a necessity.

Kern's milk-ferment is found in the region of the Caucasus Moun-
ta'ns, and so far as is known in no other j)lace. It is called by the
Caucasian peasants "kephir," "kiphir," "kiaphir," or "kefir." The
country being a mountainous one, agriculture is impossible, so that
milk and flesh are the food of the peasants. However, they do not
drink their milk fresh, but ferment it, adding to it what are known
as " kephir-graiiis " in the proportion of one volume of the grains to
six or seven volumes of milk. The whole is then exposed to the air
for twenty-four hours at an ordinary temperature, and shaken fre-
quently. The '' ferment-milk " thus formed is poured from the grains
and mixed with twice its volume of fresh milk, which it ferments in
turn, eliminating a large amount of carbonic acid gas, and forming
from ?, % to 1% of alcohol. When kephir is made successfully, it
is a thick fluid without any very large coagulated dumps, and with a
pleasantly acid taste ; by longer fermentation it becomes a frothing,
foaming, strongly acid drink, like the koumiss of the Steppes.

According to Kern, this ferment is used not onlv as a driidv, but
also as a curative for various diseases, with great success, various
gastric and ])ulmonary complaints, it is said, being cuied by it. Its
reputation, Kern continues, has extended beyond tlie narrow limits
of the mountainous region where it originated, and has already
reached many cities of the Caucasian district.

OF ARTS AND SCIENCES. 103

Examined with the eye, the kephir-grains when fresh are found to
consist of white, conipact, elastic masses, enveloped by a slime, and
with a spherical or elliptical contour, varying from 1 mm. to 5 cm.
in diameter. The very small grains have a smooth spherical exterior,
while the larger ones are provided with outgrowths and furrows, look-
ing more or less like a very small cauliliower. When the grains are
dried they assume a yellowish brown color, and shrink a good deal
by the loss of water. When examined with the microscope there
are found in each grain, whatever its form or size, two different
structures, yeast cells and Bacteria, The latter form the mass of the
grain in which the yeast cells are embedded.

The yeast cells occur in pairs or rows of cells of all shapes and
sizes. Most of them are elliptical or spherical, the former varying
from 3.2 /x to 9.6 /x, by 3.2 jx to 6.4 ju, the spherical ones varying from
3.2 /A to 6.4 /x in diameter. Each yeast cell has a plainly visible mem-
brane with a double contour, brought out by stains. Within the cell
is a vacuole, at the poles of which are often found small fat globules
in no definite number, but which increase in number as the cell is
dried, the protoplasm at the same time becoming granular, the vacu-
oles diminishing in size and ultimately disappearing. The yeast cells
increase by budding.

Kern discusses the question of the possibility that the yeast cells
may be the spores of some Mucor, as 3f. racemosus for example, since
these are known to cause alcoholic fermentation ; but since cultures
continued for weeks failed to show him a trace of mycelium, he con-
cludes that there can be no doubt that these are true yeast cells.

The origin of the kephir grains was unknown to Kern. He could
find no wild form of yeast from which they might have been culti-
vated ; nor could he gather any information as to their source from
the peasants. They are said to grow in little clumps or granules on
peculiar bushes found on the mountains just beneath the snow line.

Kern could not induce the kephir yeast to form spores. He ex-
plains the matter by saying that these yeast cells have for an infinity
of generations grown in milk only, and have increased only by bud-
ding. Hence, when they are exposed to conditions favorable for
spore formation, they are unable to form spores. He declares the
kephir yeast to be ordinary Sacchnromyces cerevisice, Meyen, saying
that he cannot agree with Dr. Max Reess in classifying yeasts accord-
ing to their form and size. The form and size of the cells vary too
much ; besides, the variations are not constant, being conditioned
partly by age, partly by the nature of the nourishing medium, and
partly by the temperature.

10 i PROCEEDINGS OF THE AMERICAN ACADExMY

The other portion of the kephlr-graiu is made up of Bacteria em-
bedded iu a zoogloea mass which is firm and ehistic, comprising the
bulk of the grain. The individual cells are short, cylindrical, and
rod-shaped, 3.2 /x to 8 /j. long and 0.8 yu. broad, with homogeneous pro-
toplasm. These cells increase by the regular splitting process charac-
teristic of Schizomycetes.

The Bacteria ia the zoogloea are motionless ; but in addition to
these, when the kephir grain is placed in a nutrient solution, there are
to be seen moving cells, exactly like the motionless ones in both
form and size. By allowing these moving cells to dry upon a slide,
then staining with Extract Campech. and removing the excess of the
stain, Kern was able to demonstrate a very thin thread-like wavy
cilium on but one of the ends of each cell.

Exposed to unfavorable conditions the Bacteria cells grow out into
Leptothrix threads, varying from 10 yx to 40 /a in length, which are
merely the necessary consequences of successive cell-division in which
the products do not separate from each other. At various intervals
in the length of such a thread agglomerations of protoplasm occur.
At first there is hardly an indication of the splitting of such a mass,
there being merely tiny incisions on either side ; but these become
larger and larger, until finally a single protoplasmic mass has given
rise to two spores, separated by a regular ceK wall. Thus it happens
that in a Leptothrix thread each cell has tw( spores situated one at
either end. Kern mentions still another kind of spore formation, seen
in the individual cells, which differs markedly frum that just described.
In these cells spore formation begins with the appearance of a small
bright point at each end of the cell. The points enlarge more and
more, assume a well defined contour, and uliimately become true
spores. The form is always round, the diameter never exceeding
that of the mother cell before they are freed, but reaching 1 yu, after
liberation.

Kern names his kephir Bacterium Dispora Caiicasica, uov. gen. et
nov. sp., with the following distinguishing characteristics: —

(1.) The vegetative cells are in the form of short cylindrical rods,
3.2 /A to 8 /A long, and 0.8 /j. broad.

(2.) In the zoiigloca condition the cells form white elastic clumps
of considerable size.

(3.) The moving vegetative cells liave on one end a thin, thread-
like, wavy cilium.

(4.) The spores are round ; when in the cells their diameter never
exceeds that of the mother cell ; when free, they may reach 1 yu, in
diameter.

OP ARTS AND SCIENCES. 106

(5.) The round spores are alwajs two in number, one at each end
of the cell.

From this subject Kern passes to his last topic, the power of re-
sistauce of the kephir-grains when subjected to external influences.
Drying does not seem to deprive them of life. They contract a
great deal, become dirty brown and hard as stone, so that they
have even been called " little stones " or " pebbles " by the inhab-
itants of the Caucasus district. In this dried state they are kept for
long periods of time, yet under suitable conditions they are always
ready to cause fermentation again. Kern himself kept some for two
months in his room. They were thoroughly desiccated, yet when
placed in milk agahi they became gradually white, and in a few days
could not be distinguished from fresh specimens. Under the micro-
scope the dried clumps show that the yeast cells suffer most, very
many being dead ; the Bacteria seem to suffer very little, since they
form spores.

Having thus summarized Kern's paper rather fully because it bears
directly upon my subject, I am in position to describe an American
milk-ferment which I hope to show is almost, if not quite, identical with
the European kephir. The material which I studied consisted of two
sets of specimens placed in my hands by Professor Farlow of Har-
vard University, to whom they had been sent by Dr. George Thurber,
of Passaic, N. J., and Mr. J. Dearness, of London, Ontario. In both
cases the specimens were in the form of rather small granules, very
few being above a centimeter in diameter, of a dirty brown color, and
presenting on their surfaces numerous lobes and fissures, thus remind-
ing one of rather dirty gum-arabic. The material from Dr. Thurber
was received in 1 888, and at that time had already lain in a dried con-
dition in his herbarium for several years. The specimens from Mr.
Dearness, undistinguishable to the naked eye from those of Dr. Thur-
ber, were received in January, 1891, under the name of "California
bees' beer," with the note that " housekeepers through this country
(Ontario) keep a self-sealing jar of this Saccharomycete half filled or
more with sweetened water. The fermented product is drawn and
drunk for a tonic."

The material from New Jersey and that from Ontario were prac-
tically identical in gross and microscopic characters, the Ontario grains
being as a rule somewhat smaller, and the following description ap-
plies to both of them. In my experiments on the action in fermenta-
tion I used principally the New Jersey material, which, in spite of the
long time it had been dried, revived when placed in a nutritive fluid.

106 PROCEEDINGS OF THE AMERICAN ACADEMY

I experimented with the Ontario material so far as to make sure that,
like the New Jersey form, it caused a fermentation of saccharose ; but
in studying the fermentation of other sugars I used only the New
Jersey form.

When soaked for a time in water, the grains become whitish, very
firm and compact, and quite elastic. Examination under the micro-
scope shows them to consist of two elements, a small proportion of
yeast cells embedded in zoJgloea masses of rod-shaped Bacteria.

Although in the dried specimens the ytast cells seem entirely dead,
yet when placed in a nutrient solution they begin to grow vigorously.
They vary in size and shape, from elliptical to spheiical, the aver-
age diameter of tlie latter being 4.2 fx, and the former varying from
10.5 ft to 6.0 /A by 6 /x to 4 /i,. On careful examination, each yeast cell
is found to have a plainly marked double contour, within which is an
almost homogeneous protoplasm containing a small vacuole. Culti-
vation of tlie yeast cells in water increases the size of the vacuoles,
and causes the formation of small fat globules at the poles ; and culti-
vation in strong solutions of saccharose produces two or even three
vacuoles in each cell, together with numerous fat globules. When
such cells are mounted in a mixture of acetic acid and glycerine, the
vacuoles disappear, and the protoplasm becomes finely granular.

The yeast cells increase by budding, growing best in solutions of
dextrose and in milk, both of which they ferment ; and it is in these
substances that the best colonies are to be found. In pure water, the
yeast cells for a short time increase slowly in numbers by budding,
but no colonies are met with since the daughter cell separates from
the mother cell as soon as it is formed. In cane-sugar or saccharose
solutions, which the yeast is unable to ferment, the cells increase very
rapidly in numbers, but it is hard to find a colony of more than
three cells ; whereas in milk and in solution of dextrose, colonies num-
bering at least from ten to fifceen cells are very common.

It was impo>sible to induce spore formation ; and indeed the very
fact that the yeast cells gave rise to new cells by the simple process of
budding after they had been dried for several months seems to war-
rant the conclusion that there is no spore formation. That they are
yeast cells, and not spores of Mucor racemosus or any otluT Mucor, is
shown by the fact that not a particle of mycelium was found during
the three months in which the yeast was under observation.

Kern decided his yeast to be a form of Saccharonvjces cet-evisits,
Meyen. In the case of the North American kephir, the species evi-
dently is not S. cerevisicE, however much it resembles that species in

OF ARTS AND SCIENCES.

107

general appearance, for it cannot invert cane-sugar as ordinary beer
yeast sliould do. Although I cultivated it in saccharose solutions of
all strengths, it never caused a trace of fermentation. As soon,
however, as T placed the yeast in a grape-sugar solution, i. e. a solu-
tion of dextrose, fermentation ensued. Unfortunately, Kern did not
try the effect of his yeast upon saccharose, and we are therefore
unable to compare his yeast with that found in American kei)hir iu
this important point. But the absence of information can hardly be
urged as an evidence that the two forms are not the same. Again,
ordinary beer yeast forms spores, while the kephir yeast does not,
thus affording another reason for reganliog them as distinct species.

Beyerinck has described the yeast which occurs iu the Caucasian
kephir grain in the " Ceutralblatt fiir Bakteriologie," Vol. VI. page
44, naming it Saccharotuyces kefyr with the following distinguishing
characteristics : —

(i.) The cells are of various sizes and shapes, from spherical to
elliptical, the former measuring from 3.2 /a to 6.4 /x in diameter, and
the latter varying from 3.2 /x- 9.6 /a in the major axis to 3.2 /a — 6.4 /i
in the minor.

(2.) The yeast is associated with a rod-shaped Bacterium in a
granular mass.

(3.) The yeast is not able to ferment saccharose or cane-sugar.

(4.) It is able to ferment lactose or milk-sugar.

(5.) It has no known spore formation. Since the North American
yeast agrees with all these characteristics, while it differs in an im-
portant point from S. cerevisice, it will be sufficient for the present
purpose if I apply the name S. kefyr to our American fosra without
attempting to discuss at lengtli disputed points in synonymy.

Let us turn now to the Bacteria. The cells are short cylindrical
rods with homogeneous protoplasm, varying fiom 8.5 fx. to 4.5 yu, long
by 0.8 /x broad ; precisely agreeing with Kern's measurements. The
cells increase by splitting perpendicularly to the long axis, the result-
ing cells being somftimes joined together, thus producing Leptothrix-
like threads of all lengths, even to I20yu., and sometimes completely
separated. Many oi the isolated cells possess the power of motion,
but after repeated efforts I was unable to demonstrate the presence
of cilia.

It is not such an easy matter to induce these cells to form spores as
Kern implies that it was in the case of his B;cteria. The best method
is to place a clump of the yeast in a watch-crystal with a little water,
covering the whole with another crystal. In twenty-four hours

108 PROCEEDINGS OF THE AMERICAN ACADEMY

Leptothrix threads, which seem to precede spore formation, begin to
form, and within thirty-six to forty-eight hours the spores api'ear. It
will be remembered that Kern gives two distinct methods of spore
formation, — one occurring in isolated cells, and the other in the
Leptothrix threads. It is no wonder, therefore, that neither method
has received general credence. My investigations on the North
American form have led to results diametrically opposed to those of
Kern. First, I found but one method of spore formation ; secondly, I
found this method occurring only in the Leptothrix threads, although
I sometimes found isolated cells bent or curled in such a manner
that spore formation was well simulated. Spore formation in the
Leptothrix threads takes place as follows. At each end of every cell
of the thread a small bright dot appears, which becomes brighter,
larger, and much more highly refractive than the rest of the cell, until
finally it assumes a well defined spore wall and develops into a mature
spore. Each cell has therefore two spores, one at each end, and each
originating independently of the other. In no case did I see two
spoi'es formed, as Kern states, by the division of a single agglomerated
mass of protoplasm into two portions.

There are two or three other important points in which the Ameri-
can alcoholic milk-ferment closely resembles the Caucasian kephir.
So far as I know, no one has ever tried the experiment of making
the North American yeast cause the alcoholic fermentation of milk.
Struck with its gross and microscopic resemblances to kephir, I was
induced to try the experiment, and to my pleasure I obtained
alcoholic fermentation, the evolution of carbonic acid gas being
suflftcient to force the cork from the flask. I easily obtained a large
precipitate of CaCOs from lime-water by the usual test for carbonic
acid gas. The presence of alcohol was proved by the iodiform test.
Since ethyl alcohol, Co H5OH, cannot be detected in the presence of
lactic acid, CsHgOg, (for lactic acid forms iodiform as easily as ethyl
alcohol,) I neutralized with Na^COs after filtering off the fluid por-
tion of the milk, distilling finally the neutralized filtrate. Thus every
trace of lactic acid was removed. With KOH and iodine the dis-
tillate gave iodiform, thus proving the presence of alcohol. This fact
alone, viz. that this yeast causes alcoholic fermentation of milk, is
sufficient to establish a near relation to the Caucasian kephir. More-
over the fermented milk agrees closely with the description of the
kephir drink. The milk does not sour in the ordinary sense, for it
does not coagulate in large masses ; still it is acid, contains some
carbonic acid gas and alcohol, and is by no means unpleasant to the
taste.

OP ARTS AND SCIENCES. 109

Another way in which this American milk-ferment resembles
kephir is, that it causes alcoholic fermentation of dextrose. De Bary
is authority for the statement, that the "• kephir yeast, like its con-
stituent the Saccharomycete, working by itself, gives rise to alcobolic
fermentation in a nutrient solution of grape-sugar, though of a less
active kind than that caused by beer-yeast." * The specimen which
I had gave a good alcoholic fermentation with dextrose solutions, but
caused no fermentation with saccharose. It seems, therefore, to have
the power of fermenting only two of our natural sugars, — dextrose
and milk-sugar.

When we consider the remarkable similarity of these American
grains with the kephir granules in color, shape, and general appear-
ance; the great similarity between the j^east cells and bacteria of
each in appearance, habits, mode of growth, form, and size; the fact
that both of these yeasts cause alcoholic fermentation of milk ; the
fact that the drink formed by the American kephir closely resembles
the description of kephir ; the minor resemblance between the two,
that of fermenting dextrose solutions, and that of its great capacity
for resisting external influences, — we are justified in concluding the
American milk-ferment to be a very near relative of the European
kephir, if it be not indeed identical with it.

One point remains, viz. How can this yeast cause alcoholic fermen-
tation of milk-sugar ? This question, which did not present itself to
Kern, De Bary has tried to explain in his ' Lectures on Bacteria.'
Speaking of kephir, he says, " The changes in the milk which produce
the drink here described are brought about by the combined activity
of at least three ferment-organisms." There is the yeast cell, the
Bacillus of the kephir-grain, and the Bacterium of lactic fermentation.
He goes on to say that " the acidification is caused by the conversion
of a portion of the milk-sugar into lactic acid by the bacterium of that
acid. The alcoholic fermentation, that is, the formation of alcohol
and of a large part at least of the carbonic acid, is indebted for its
material to another portion of the milk-sugar, and for its existence to
the fermenting power of the Sprouting Fungus (yeast). , . . But al-
coholic fermentation is produced in milk-sugar as such neither by Sac-
charomycetes, with which we are acquainted, nor, as experiment has
shown, by those of which we are speaking. To make this fermenta-
tion possible the sugar must first be inverted, split into fermentable
kinds of sugar." De Bary continues: "According to Nageli, the for-

* Lectures on Bacteria, De Bary, translated by Garnsey and Balfour, p. 96.

110 PROCEEDINGS OF THE AMERICAN ACADEMY

mation of an enzyme which inverts milk-sugar is a general phenome-
non in Bacteria, and Hueppe has shown that it is probable in the case
of his Bacillus of lactic acid in particular." De Bary then concludes: >
" The inversion required in this case to enable the Saccharoraycete to
set up alcoholic fermentation is the work therefore of the Bacillus of
lactic acid, or of the Bacterium of the Zoogloea, or of both."

But De Bary has since revoked this explanation. A. Levy of
Hagenau discovered that kephir may be made without any kephir
grains "simply by shaking the milk with sufficient violence while it is
turning sour. A trial convinced me," says De Bary, " of the correct-
ness of this statement. The kephir obtained by shaking was not per-
ceptibly different in taste or other qualities from the kephir of the
grains, and the determination of alcohol, kindly made for me by Pro-
fessor Schmiedeberg gave 1 per cent in some specimens of the former
kind and 0.4 per cent in one of the latter; sour milk not shaken con-
tained no trace of alcohol or only a doubtful one. Our former expla-
nation, therefore, must be abandoned, and there is no other ready at
present to take its place."

Beyerinck has also proposed a theory to account for alcoholic fer-
mentation of milk. There are at present, he sa,ys, four yeasts which
are known to cause such fermentation: (1) that of Duclaux,* (2) that
of Adametz, called Saccharomyces lactis,^ (3) that of the kephir
called by Beyerinck Saccharomyces hefyr, and (4) Saccharomyces Ty-
rocola.% As a matter of fact there is one other which he overlooked,
Saccharomyces galacticola, described by Pirotta,§ of which I intend to
speak later. Beyerinck supposes that these yeasts secrete an enzyme
which he names lactase, since it inverts lactose or milk-sugar, and
which he declares to be in every way analogous with invertine. The
inverted milk-sugar is next acted upon by the yeast, carbonic acid gas
eliminated, and alcohol formed. Hence, if his supposition be true,
Saccharomyces kefyr should ferment sweet milk by first producing
its enzyme and then by acting upon the inverted product ; but it is
universally agreed by all who have written upon kephir that the lactic
acid fermentation must precede the alcoholic, or else the latter will
not take place. Beyerinck's theory therefore fails, in that it pays no
heed to the Bacteria of lactic fermentation.

* Ann. d. I'lnst. Pasteur, 1887, I. 573. See Ibid.. 1889, III. 201.

t Centralblatt. f. Bakt. u. Parasit., V. 116.

\ Ibid , VI. 44.

§ Pirotta et Rib. Studii eul Latte. Pavia, 1879.

OP ARTS AND SCIENCES.

Ill

I have experimented with our North American ferment and find
the facts to be these: (1.) It causes alcoholic fermentation of milk-
sugar or lactose, Ci2H220n. (2.) It causes fermentation of dextrose,
CsHioOg. (3.) It does not cause fermentation of saccharose or cane-
sugar, which has the same empirical formula as lactose. In addition to
these three facts it is also known, as Ilueppe has shown, (1) that the
Bacillus of lactic fermentation causes to some extent the inversion of
milk-sugar; (2) that lactic acid, according to Hammarsten, by stand-
ing with milk-sugar inverts it to dextrose and galactose just as does
any mineral acid ; (3) that the Bacillus of lactic fermentation acts
further on the galactose, CeHjaOe, convening it into two molecules of
lactic acid, CsHgOs.

From these data it seems evident that alcoholic fermentation of
milk takes place in the followiug manner. The Bacillus acidilactis
begins the process by forming some lactic acid, which in turn, assisted
by the Bacillus itself, inverts the milk-sugar to galactose and dextrose.
The galactose is further acted upon by Bacillus acidi-lactis and con-
verted into lactic acid ; the dextrose is acted upon by the yeast, and
converted into alcohol and carbonic acid gas. In the kephir drink,
therefore, we should find plenty of lactic acid, a little milk-sugar, not
inverted, the amount depending upon the duration of fermentation,
some alcohol, and carbonic acid gas, — precisely what is found.

One vital objection may, however, still be urged against this theory.
If it be true, as I have said, that the Bacillus acidi-lactis to some
extent, and the lactic acid to a greater extent, cause the inversion of
milk-sugar, then should not ordinary beer yeast, SaccJiaromyces cere-
visice, Meyen, cause alcoholic fermentation in sour milk, since the
milk-sugar, according to the theory, must be here inverted to fermenta-
ble dextrose and to galactose ? It should cause such fermentation ;
and if it does, the theory is confirmed. Upon experiment, I found that
ordinary beer yeast when added to sour milk or to milk on the point
of souring did cause fermentation, much carbonic acid gas being elimi-
nated and some alcohol formed ; sweet milk, however, did not ferment
with beer yeast.

It will be observed that I have given no function to the Bacteria of
the kephir granules. The very fact that they remain almost wholly
in the Zoogloea masses during fermentation, comparatively few going
out into the milk, seems to indicate that they have little to do with
this alcoholic fermentation ; and this is made still more probable by
the additional fact that, though absent in the fermentation of sour milk
by beer yeast, still fermentation ensues.

112

PROCEEDINGS OP THE AMERICAN ACADEMY

To this theory De Bary has objected that kephir can be made
simply by shaking milk which is on the point of souring, such kephir
being called Pseudo- or Schijttelkepliir. He refers to a paper by A.
Levy,* of Hagenau, in which Levy claimed that ordinary sour milk
shaken in a flask with eight or ten parts of cold boiled milk at about
10° R. gave carbonic acid gas, lactic acid, alcohol, and peptone. But
Levy says, by shaking, the air is introduced, and the fermentation and
peptonization are probably brought about by micro-organisms, which
are very numerous in milk. Franz Kogelmanu f has also published
a method for obtaining kephir easily and cheaply. Take one volume
of ordinary buttermilk, shake with two of fresh, and there is obtained
a fluid " identical with kephir," containing carbonic dioxide, alcohol,
lactic acid, casein, etc. Notwithstanding these claims, there is some
doubt about the identity of Schiittelkephir and Kogelmann's kephir
with the true sort, as Rudeck's % table shows.

Milk.
One Litre.

Kogelniann's
Kephir.

Pseudo
Kephir.

True Kephir,
36 hours.

Casein

48.00

35.00

38.00

36.50

Butter

38.00

11.00

16.00

18.00

Lactose

41.00

9.00

13.00

18.00

Lactic acid

14.50

11.00

6.00

Alcohol

....

....

Trace

5.00

Albumen

1.80

ii

1.50

Peptonized albumen . .

0.90

ii

2.00

Lactosyntonid ....

0.40
t

Trace

0.80

Peptone

• • • •

0.48

Salts and Water . . .

871.20

929.20

922.00

911.72

Totals

1000.00

1000.00

1000.00

1000.00 •

Despite this table, however, it is not improbable that alcoholic fer-
mentation often does actually take place in Kogelmanu's and Levy's
methods, as the following paragraph may show.

* Die wahre Natur des Kefirs. Deuts. Med. Ztg., 1886, p. 783.

t Ueber Milchwein (Kefir). Ibid See also Pharm. Cent. Halle, XXVII. 42.

X Pharm. Ztg. Berl., XXXIII. 420.

OF ARTS AND SCIKNCKS.

113

Levy aud Kooelmann were by no means the first to experiment on
this j-ubject. Blondlot,* as early as 1872, found tliat he could obtain
alcoholic fermentation in milk simply by shaking it. Pirotta investi-
gated this fermented fluid, and found a yeast, wliich he named Sur-
charomyces galacticola, identical with Succharomyccs ccrevisice in
appearance, size, spore formation, and in the fact that both ferment
sour milk. It is not improbable, therefore, that this yeast may be
nothing more nor less than Saceharomyces cerevisice itself, and that
ordinary beer yeast is one of the micro-organisms which sprung up in
Levy's kephir, of which he unfortunately omitted to make a microscop-
ical examination. Hence De Bary's objection that sour milk, simply
shaken, veill give alcoholic fermentation, loses its significance.

Throughout Germany and Russia kephir has become a very cele-
brated drink, simply because a considerable portion of the albuminoids
of the milk are peptonized. For persons of weak digestion, for chil-
dren, and for dyspeptics generally, it is an excellent diet, since it re-
lieves the stomach of much of its work. Hence the fame of kephir
has spread far and wide, and a kejihir factory has been started at
Hamburg. The following table, taken from J. Biel's " Ueber die
Eiweisstoffe des Kefirs," shows this peptonization very neatly.

In 100 parts of kephir were obtained : —

Kepliir fcriiiiMitt'd
One Day.

Kephir fermented
Two Days.

Kephir fermented
Three Days.

Lactic acid

0.510

0.5625

0.6525

Lactose

3.750

3.2200

3.0940

Casein

3.340

2.8725

2.9975

Albumen

0.115

0.0300

0.0000

Acid albumen

0.095

0.1075

O.'J.-jOO

Peptonized albumen . . .

0.190

0.2815

0.4085

Peptone

0.035

0.0460

0.0815

From the table it is evident that the casein and albumen decrease
during fermentation, while the peptone, peptonized albumen, and acid
albumen increase. This is shown still better by another table.

* Comptcs Rciidus, LXXIV. 5-34.

VOL. XXVI (n. S. XVIII.) 8

114

PROCEEDINGS OF THE AMERICAN ACADEMY

In 100 parts of albuminoids were obtained : —

Kephir fermented
One Day.

Kcphir fermented
Two Days.

Kejiliir ftrmented
Three Days.

Casein .... . .

Albumen

Acid albumen

Peptonized albumen . . .
Peptone

88.47
3.05
2.52
5.03
0.93

86.07
0.90
3.22
8.43
1.38

80.20
0.00
6.69

10.93
2.18

An analysis of the milk fermented by the American yeast shows
the presence of peptone, in some quantity, whereas sour milk fermented
by beer yeast gave only a trace, thus agreeing with Rudeck's analysis
for Kogelmann's kephir. From these analyses there is but one in-
ference,— the peptonizing power must lie, not in the Bacillus acidi-
lactis, which is common to all these true and false kephirs, but in the
yeast which Beyerinck has named Saccharomyces l^efyr, and which
exists in the United States.

In conclusion, I would return my thanks to Prof. W, G. Farlow
and Prof. H. B. Hill, for advice given during the progress of my
work.

OF ARTS AND SCIENCES. 115

X.

DAMPENING OF ELECTRICAL OSCILLATIONS ON

IRON WIRES.

By John Trowbuidge.

Presented May 27, 1891.

It has generally been assumed by those who have studied the subject
of very rapid oscillutious of electricity, such as occur in Leyden jar
discharges, that the magnetic character of the conductor has very
little influence upon the character of the discharge. Thus, in a note
to an article on electrical waves, W. Feddersen states that electrical
oscilhitions may suffer a slight weakening on iron ; but this diminu-
tion is very slight : —

" Beim Eisen konnte in Folge der Magnetisirungen eine Abweichung
hervortreten ; in dess zeigt der Versuch, dass dieselbe keiuenfalls be-
deutend ist, iibrigens in dem Sinne erfolgen miisste, als wenn die
Elektiicitiit beim Eisen ein grossere Hinderniss fande, wie bei den
ubrigen Metallen."*

In Dr. Lodge's treatise on Modern Views of Electricity (ed. 1889),
we find the following : —

" But in the case of the discharge of a Leyden jar iron is of no
advantage. The current oscillates so quickly that any iron intro-
duced into its circuit, however subdivided into thin wires it may be,
is protected from magnetism by inverse currents induced in its outer
skin, and accordingly does not get magnetized ; and so far from in-
creasing the inductance of the discharge circuit, it positively diminishes
it by the reaction effect of these induced currents ; it acts, in fact,
much as a mass of copper might be expected to do." (p. 3Go.)

Fleming writes as follows : —

" With respect to the apparent superiority of iron it would natu-
rally be supposed that, since the magnetic permeability of iron bestows
upon it greater inductance, it would form a less suitable conductor for
discharging with great suddenness of electric energy. Owing to the
fact that the current only penetrates just into the skin of the conductor,

* Annalen der Physik und Chemie, No. 108, 1859, p. 499.

116 PROCEEDINGS OF THE AMERICAN ACADEMY

there is but little of the mass of the iron magnetized. Even if these
instantaneous discharges are capable of magnetizing iron, . . . the
electromotive impulses or sudden rushes of electricity do not mag-
netize the iron, and hence do not find in it any greater self-inductive
opposition than they would find in a non-magnetic but otherwise simi-
lar conductor. Dr. Lodge's further researches seem to show that
there is a real advantasre in using iron for lightniiiij conductors over
copper, and that its greater specific resistance and higher fusing point
enables an iron rod or tape to get rid safely of an amount of electric
energy stored up in the dielective which would not be the case if it
were copper." *

Fleming describes in full Dr. Lodge's experiments to prove the
non-magnetizability of iron by sudden discharges : —

" In the experiments on alternative path, as described by Dr.
Lodge, the main result is very briefly summed up by saying that,
when a sudden discharge had to pass through a conductor, it was
found that iron and copper acted about equally well, and indeed iron
sometimes exhibited a little superiority, and that the thickness of the
conductor and its ordinary conductivity mattered very little indeed.
... In the case of enormously rapid oscillations the value of the
impulsive impedance varies in simple proportion to the frequency of
the oscillations, and depends on the form and size of the circuit, but
not at all on its specific resistance, magnetic permeability, or diame-
ter. . . . For discharges of a million per second and ujiwards, such as
occur in jar discharges and perhaps in lightning, the impedance of all
reasonably conducting circuits is the same, and independent of con-
ductivity and permeability, and hardly affected by enormous changes
in diameter." f

Turning now to the observations of Hertz, we find it stated that the
material, the resistance, and the diameter of the wire of the microm-
eter circuit employed by him, have very little influence on the result.
The rate of propagation of an electrical disturbance along a con-
ductor depends mainly on its capacity and coelHcient of self-induction,
and only to a small extent on its resistance. Hertz concludes that,
owing to the great rapidity of the alternations, the magnetism of the
iron is unable to follow them, and therefore has no effect on the self-
induction. When a portion of the micrometer circuit employed by
Hertz was surrounded by an iron tube, or replaced by an iron wire,
no perceptible effect was obtained, and thus the result was apparently

* Fleming, Inrluction of Electric Currents, p. 398. t Ibid., p. 411.

OF ARTS AND SCIENCES. 117

confirmed that the magnetism of the iron is unable to follow such
rapid oscillations, and therefore exerts no appreciable efiect. The
velocity of propagation in a wire has a definite value independent of
its dimensions and material. Even iron wires offer no exception to
this, showing that the magnetic susceptibility of iron does not ])lay
any l^urt in the case of such rapid motions.*

Altlioiigh the iin[)ulsive impedance is apparently not affected by the
magnetic character of the wire, experiments lead me to believe that
discharges of the quick period of a Leyden jar are affected very appre-
ciably by tlie magnetic nature of iron, steel, and nickel conductors.
This effect, is so great that it dampens the electrical oscillations, and
makes it difficult to determine whether the time of oscillation is also
affected by the permeability of the conductor.

The apparatus em[)loyed was similar to that described in the in-
vestigation of electrical oscillations with an air condenser.! Certain
important modifications, however, were made. The plane mirror which
was used in the former research was replaced by a concave mirror of
ten feet focus and three and a half inches in radius. This mirror
was mounted upon the end of the armature shaft of a one-half horse-
power electric motor.

The discharging apparatus consisted of a sharp cutting tool, insu-
lated, and mounted on the edge of the rotating disk bearing the mirror.
It was metMllically connected with a grooved ring of brass miounted
upon the shaft and insulated from it by hard rubber. Around this
was wound a copper wire, one end of which was connected with the
discharging wire, and the other drawn taught by a rubber band. The
electrical discharge was thrown on to the circuit by thrusting forward
a lever which brought a solid hinged frame containing a strip of soft
type-metal into contact with the rapidly revolving steel cutting tool.
An electrical contact was thus insured by the tool cutting a groove in
the strip of type-metal. In order to avoid a spark at the contact, the
type-metal was thickly covered with a wax of peculiar composition.
The only spark that occurred, therefore, was the one the oscillations

* " Ersetzen wir den bisherigen Kupferdraht durch einen dickeren oder
dunneren Kupferdraht oder durch einen Draht aus anderem Metall, so belialten
die Knotenpunkte ilu'e Lager bei. Die Fortplanzungsgeschwindigkeit in alien
solchen Driihten ist daher gleich, und wir sind berechtigt, von derselben als
einer bestimmten Geschwindigkeit zu redcn. Auoh Eisendrahte machcn keine
Ausnahme von der allgemeinen Kegel, die Magnetisirbarkeit des Eisens kommt
also bei so schmalen Bewcgungen nicht in Betracht." — Ann. der Physik und
Cheniie, No. 34, 1888, p. 558.

t Tiiese Proceedings, Vol. XXV. p. 109.

118 PROCEEDINGS OF THE AMERICAN ACADEMY

of which I desired to study. At each trial the type-metal was moved
so as to expose a new cutting surface. The type-metal was insu-
lated from the rest of the apparatus, but connected with the outer
coating of the Leyden jar ; first both terminals of the Holtz machine
were thrown off, and immediately after the cutting tool, ploughing its
way through the type-metal, placed the' outer coating of the Leyden
jar in circuit with one of the two parallel wires leading to the ter-
minals of the spark. The other wire was permanently in connection
with the inner coating of the jar.

Beside the short lead wires above described, the discharging circuit
consisted of two parallel wires 30 cm. apart and 510 cm. long. These
were the only portions of the apparatus changed during the experi-
ment, and they were replaced by wires of different material and of
different size. The other conditions — length of spark, lead wires,
and the copper cross wire connecting the outer end of the long parallel
wires — remained undisturbed throughout the experiment.

The Leyden jar was charged each time as nearly as possible to the
same potential, judging by the number of turns given the Holtz ma-
chine. It is unfortunate that no more accurate means of measuring it
were at hand, although the different negatives showed but slight varia-
tion. The capacity of the jar to alternations of this period was 5060
electrostatic units.

I describe the discharging portion of the apparatus minutely, for
the success of an investigation of this na'ture depends upon the sup-
pression of all sparks save that which one wishes to observe ; and the
method surely and completely accomplished this. The photograph of
the spark could thus be made to fall very accurately on the sensitive
plate. When one considers that the image of the spark was flying
through the air on a circle of a radius of ten feet with a velocity of a
mile a second, it will be seen that an extremely small deviation in the
point of contact between the cutting tool and the type-metal would
have thrown the image entirely off the sensitive plate, A singular
phenomenon was noticed in this connection. When a comparatively
low potential was used, such as that afforded by the air condenser used
in our previous investigation, the cutting tool ploughed two or three
millimeters along the surface of the type-metal before a spark passed
at the point in the circuit where it was de>ired. With higher poten-
tials this phenomenon was also observed, but the extent of cutting
was diminished.

It is possible that that the insulating wax may have melted under
the sudden blow of the cutting tool, and, flowmg around it, prevented

OP ARTS AND SCIENCES. 119

instant contact. This seems to us improbable, for a deep and clear-
cut groove was made in the soft type-metal. Great attention was
paid to the solid structure of this contact apparatus. It was entirely
separate from the support of the revolving parts, and was perfectly
steady.

The other end of the armature shaft was lengthened into a cylin-
drical chronograph, similar to that described in the article already
cited, and its performance left nothing to be desired. A small
Kuhmkorf coil, excited by two storage cells, and interrupted by a
seconds pendulum, gave a record of the speed of the mirror. The
stylus which drew the spiral turns on the barrel of the chronograph
was drawn along the barrel by means of a small heavily loaded car-
riasfe, which, ou beins released at the moment the lever arm threw
the type-metal in contact with the cutting tool, descended an inclined
plane of adjustable height.

A small Topler Holtz machine charged a large Leyden jar, and
it was found to work admirably in all states of the weather. The
apparatus which I have thus described was the result of the expe-
rience of the previous year, and worked for months without failure ;
and the taking of photographs of the oscillatory discharge by it be-
came a mere matter of routine.

The following cases were tried : —

(1.) When the long parallel wires were of copper (diameter
.087 cm.), the number of double oscillations visible on the negatives
averaged quite uniformly 9 or 9.5.

(2.) When the wires were of German silver (diameter .061 cm.),
three oscillations were visible.

(3.) But when an annealed iron wire (diameter .087 cm.) was
substituted, only the first return oscillation was distinctly visible, with
occasionally a trace of the first duplicate discharge.

(4.) On substituting fine copper wire (diameter .027 cm.), five
complete oscillations were quite uniformly visible.

(5.) Fine German silver wire (.029 cm.), nickel wire (.019 cm.),*
soft iron (.027 cm.), and piano steel wire (.027 cm.), gave but faintly
the first return discharge after the pilot spark.

The pilot sparks were in all cases strong.

The single return discharge through the iron wire did not admit of
measurement sufficiently accurate to furnish any basis for calculation
of its self-induction. The time did not apparently differ, if at all, by

* Obtained by the kindness of Joseph Wharton, Esq., of Philadelphia.

120 PROCEEDINGS OF THE AMERICAN ACADEMY

more than fourteen or fifteen per cent. Some general reasoning
based upon the number of oscillations may be of interest. It must
be acknowledged, however, that' this reasoning is open to criticism,
although it affords the most plausible explanation. The phenomenon
itself is not a doubtful one.

The time of a double oscillation for the large-sized copper wire was
.0000020 sec; for the small copper wire, .0000021 sec. The others
as far as could be determined did not differ much from these values,
and for this purpose either is sufficiently accurate. Denote by R' the
ohmic resistance of the parallel wires to alternating currents of this
periodicity ; by R, the resistance to steady currents.

p^ — = 3,000,000 (practically).

Taking the cases up in order :
(1.) Large copper wire,

R = 0.285 X 10^

and substituting in Lord Rayleigh's formula, R' = y/^plixR,

i?' = 0.66 X 10^

(2.) Large German silver wire,

i? = 9.2 X 10^
and substituting in the series

^ ^ V + 12 ~^r ~ i8o E* +•••;'

R> = 9.2 X 10^

(3.) Large iron wire,

R =2.5 X 10^
and if there is a true time lag, as often stated, such as to prevent
action of the magnetic property of the iron, and if on this assumption
we make /a = 1,

R' = 2.78 X 10^

(4.) Fine copper.

R =3.S X 10^
R' =3.5 X 10^

(5.) Again, as before, call yu, = 1 in iron, nickel, and steel. The
lenjith of these circuits was 7.41 meters, the remaindi^r of the 10.20
meters — 2.79 meters — being of copper wire of R' =: 0.94.

Tlie value of R' in the separate cases, including in each the resist-
ance 0.94 of the copper portion, was as follows : —

OP ARTS AND SCIEN'CES.

121

Soft iron
Piano steel
Nickel . .
German silver

15.0 X 109

20.7 X 10^

30.G X 109

23.0 X 109

The ratio of the strengths of successive discharges during the oscil-

rT

lation is given by the function e^^, where r is the ohmic resistance,
T the time of a double oscillation, and L the self-induction. The
ratio of one discharge to the nth one after it is e'^2/;. If we as-
sume— and it is a large assumption, but one which perhaps the result
will in some measure justify — that the ratio of the strength of the first
to the strength of the last visible discharge is more or less a constant,

T
we may make use of the above data. Denote ^^ by A, and call

the unknown resistance of tlie short connecting lead wires and of the
spark X. Then will r =. R' -\- x, and ?i will be the number of com-
plete oscillations visible.

Take cases (1) and (2), large copper and large German silver
wires : —

n^ {R\ + X)^ «2 (i?'2 + x) ;

9.5 (0.66 + x) =z 3 (9.2 + x) ;

X ■= 3.4 ohms.

Taking cases (1) and (4) similarly,

«i (E\ + x)=n, (R', -f x) ;
9.5 (0.66 + x) = 5 (3.5 -f- x) ;

X = 2.6 ohms.

Experiments with other copper wires having 7?' equal to 3.4 and
1.27 gave 0 and 8 for the values of n respectively, or

X = 2.4 ohms.

The resistance (R') of the lead wires forming part of x was 0,8
ohm, leaving as a possible value for tiie resistance of the spark
about 2 ohms.

If, taking this value of x, we calculate the value of R' necessary to
damp out the oscilhition in one complete double discharge iu the case
of the large iron wire, we shall have

9.5 (0.66 X 3) =. 1 (R' + 3);

R' = 30 ohms.

122 PROCEEDINGS OP THE AMERICAN ACADEMY

But neglecting the magnetic property of the iron, its calculated re-
sistance to alternating currents of this periodicity was R' = 2.78
ohms. This is obviously inadequate, and would point to the conclu-
sion that the oscillation is not, as sometimes stated, too rapid to admit
of the magnetic action of the iron.

If we substitute this value H' = 30 in the equation

we have for the resulting value of the magnetic permeability jx = 230.
This lies between the limits yu, = 103 and /x = 1110, found by taking
the number of oscillations one and a half and one half respectively for
the case of the iron wire.

It should be noticed that this estimate of fx necessitates assuming
that T and L remain the same within broad limits. Measurements
of the single oscillation on the negatives show that this is near enough
the case. Part of the more rapid decay of the oscillation in the iron
may be well ascribed to the dissipation of energy by hysteresis.
While we cannot place much reliance upon an estimate of its value
in such a case, — its percentage effect probably increasing rapidly with
the decay of the spark, — it is not difficult to show that its influence
may be very great.

There still remains the fact, not generally recognized, that, in Ley-
den jar discharges through iron wires, the magnetic property of the
iron has time very materially to modify the character of the spark.

We give an example of the measurement of the half-oscillation
which was the only one visible on the photograph of the discharge
over iron wires, all the others having been dampened or extinguished
by the iron, in compari>on with the measurement of the similar half-
oscillation on copper wires of the same diameter as the iron wires.
The number of oscillations on the copper wires was eight.

The total duration of the discharge on iron wires was only three
millionths of a second, while that on similar copper wire was three
hundred-thousandths of a second. A steel wire gave the same results
as the annealed iron wires.

Comparative Leiigths of First Half-oscillation in Millimeters.

Fine Iron Wire. Fine Copper Wire.

.23 .19

.21 .20

.19 .20

.21 .19

OF ARTS AND SCIENCES. 123

Large Iron Wire. Large Copper Wire.

.20 17

.20 18

.19 .20

I9 .18

I wish to express my deep obligations to my assistant, Mr. W. C.
Sabine, for bis valuable suggestions and for his skill in the mechanical
details of this investigation.

Conclusions.

1. The magnetic permeability of iron wires exercises an important
influence upon the decay of electrical oscillations of high frequency.
This influence is so great that the oscillations may be reduced to a
half-oscillation on a circuit of suitable self-induction and capacity for
producing them.

2. It is probable that the time of oscillation on iron wires may be
changed. Since we have been able to obtain only a half-oscillation
on iron wires, we have not been able to state this law definitely,

3. Currents of high frequency, such as are produced in Leyden
jar discharges, therefore magnetize the iron.

Jefferson Physical Laboratory,
Cambridge.

124 PROCEEDINGS OP THE AIMERICAN ACADEMY

XI.

CONTRIBUTIONS TO AMERICAN BOTANY.
By SiiKKXo Watson.

Presented April 8, 1891.

1. Descriptions of some nevj North American Species, chiefly
of the United States, with a Revision of the American Species
of the Genus Erythronium.

Arabis Macounii. Biennial, branched from the base, slender, pu-
bescent below with mostly stellate spreading hairs, glabrous above or
but sparingly puberulent, a foot high : leaves small and narrow, -A inch
long or less, the lower very rarely fevv-toolhed, the cauline sagittate
at base : flowers very small, pale rose-color, 2 lines long: pods very
narrow, 1 to \^ inches long and about -j line broad, glabrous, slightly
curved, mostly divaricate on very slender pedicels 2 to 4 lines lung,
acute, the stigma nearlj'^ sessile: seeds (immature) appi-oximately
1-rowed, apparently wingless. — At Revelstoke, British Colinnbia;
collected by Prof. John Macoun, May, 1890. Near A. hirsiita.

Erysimum arenicola. Caudex much branched and densely tufted,
the branches slender; flowering stems about 6 inches high: leaves
narrowly oblanceolate, sparingly toothed, acute or acutish, attenuate
to a slender base, about an inch long, sparsely appressed-pubescent :
pedicels slender, spreading and 2 or 3 lines long in fruit : calyx 4
lines long: pods ascending, slender, 1^ to 2 inches long and less than
a line broad, compressed and thin-valved, usually attenuate above to
a slender style tipped by the depressed lobed stigma : seeds narrowly
oblong, a line long; the cotyledons very obliquely incumbent. — In
volcanic sand on the Olympic Mountains, Wasliingt'Mi, at 5.000 feet
altitude; collected by Mr. C. V. Piper, Septoinber, 1S<>0 (n. 91()).

SiLENE Macounii. Stems very slender, from a slender branching
rootstock, a foot high, minutely puberulent, glandular above: leaves
linear-oblanceolate, 3 inches long or less : flowers few, on pedicels ^
to 1 inch long; calyx inflated, oblong-campanulate, 4 or ;"> lines long,
with short obtuse teeth ; petals little ex.^erted (G lines long), with a
broadly auricled glabrous claw and large thin quadrate and nearly

OP ARTS AND SCIENCES. 126

entire appendages, the flabelliform bifid blade with a linear tooth on
each side : cap?ule equalling the calyx, oblong-ovate, on a carjiophure
lA- lines long. — Snnunit of the Rocky i\I()Uiitains, Biitish Columbia ;
collected by Prof. J. Macoun, August, 18'J0.

MiMULUS (Eu.MiMULus) FiLiCAULis. A dwarf annual with very
slender and thread-like lax stems, 1 to 4 inches high, simjile or nearly
so, sparsely glandular-pubescent : leaves thin and nerveless, entire, ob-
lanceolate or oblong or the lowest obovate, obtuse, narrowed to a very
short petiole: flowers on long pedicels, the narrowly obKnig or tur-
binate calyx 3 lines long, acutely and unequally toothed ; corolla
funnelform with a nearly equally lobed limb, 7 to 9 lines long, bright
rose-color in various shades, with more or less of purple and yellow in
the throat and tube. — Collected by J. W. Congdon on Snow Creek.
Mariposa County, California, in June, 1800. Near 31. Pahneri, from
which it differs in its much more slender and less branching habit, the
leaves more narrowed at base, the calyx-teeth acuter, and the corolla
different in shape and coloring.

Cladothkix cryptantha. Apparently annual, canescent through-
out with a fine dense much branched pubescence, slender, repeatedly
branched somewhat di- or trichotomously : leaves alternate, or sub-
opposite at the forks, ovate to obovate, 4 to 6 lines long or less, on
slender petioles : flowers in close clusters of 2 to 5, involucrate and
more or less enclosed by two or more sessile floral leaves which are
united below into a somewhat indurated cup ; bracts and bractlets
minute; sepals thin, ^ line long: ntricle included, thin and hyaline,
obtuse, the 2-lobed stigma nearly sessile. — Collected by Dr. C. C.
Parry at Colton, California, in 1881 (n. 274), and by Mr. C. R.
Orcutt in November, 1890 (n. 218G), at Causo Creek in San Diego
County.

Eriogonum (Ganysma) MINUTIFLORU5I Of the E. ptcsiUum group,
very slender, G inches high or less, diffusely branching, glabrous, ex-
cepting the small ovate rosulate leaves which are densely white-tomen-
tose on both sides, becoming less tomentose above; bracts minute;
peduncles filiform, divaricately spreading, 3 to 8 lines long; involucres
very small (t line long), broadly turhinate-campanulate, purplish:
perianth yellow, minutely puberulent, very small. — Found by Mr.
C. R. Orcutt in the desert region of San Diego County, California,
April, 1890. Resembling E. subreniformr, but the leaves not reni-
form nor cordate, and the smaller flowers more pubescent.

EniOGONrn deserticola. Apparently an annual of the same
group (base and foliage unknown), tall, several times dichotomously

126 PROCEEDINGS OF THE AMERICAN ACADEMY

branched, wliite-tomentose becoming mostly glabrous and yellowish
green; bracts all small and deltoid i involucres shortly pedicellate or
subsessile toward the end of the branches, erect or spreading, tur-
binate-campanulate, a line long : perianth villous, the elliptical segments
yellow with greenish or reddish midveins, 1 to 1^ lines long. — In the
southwestern part of the Colorado desert, San Diego County, Cali-
fornia; C. R, Orcutt, November, 1890 (n. 2189).

P:RYTHR0N1UM, Linn. It is within the limits of the United
States that this genus reaches its fullest development. On this conti-
nent it is found scarcely beyond our own boundaries, and in the Old
World it shows a far narrower range of variation than here. Much
unavoidable uncertainty has long existed respecting the species of
western North America. Having taken advantage of such opportuni-
ties as presented themselves for studying these various forms, I now
propose, though with some hesitation, the following revision of the
genus. For material and for field-notes upon the ditficult western
species, tlianks are due especially to Mr. Carl Purdy, G. R. Kleeber^er,
and Volney Rattan, of California, Mrs. P. G. Barrett, Thomas Howell,
and W. C. Cusick, of Oregon, L. F. Henderson and W. N. Suksdorf,
of Washington, and Prof. John Macoun of the Canadian Geological
Survey.

The eastern and western species are conveniently separated, as will
be seen, upon good distinctive characters. The Old World species,
considered as a unit, is most nearly allied to the eastern group in its
always solitary flowers, the want of a gibbous crest upon the petals,
and the shape of the capsule, while in its mode of underground prop-
agation it more resembles the western species.

The characters that must be relied upon for the distinction of species
are rarely constant. The thinly coated corms produce new ones, either
as in the eastern species at base within the old coats or at the ex-
tremit}'^ of long oflHshoots, or as in nearly all the western species along
a rhizome, sometimes in near succession for several years, sometimes
at intervals of an inch or less. The habit of spreading by offshoots,
where it occurs at all, appears to depend much upon the season or lo-
cality, and is usually attended with a diminished production of flowers
and seeds. The form of the leaves is only exceptionally of any value.
In all the species the leaves in the cauline pair are unequal, one being
as a rule narrower and more acuminate than the other. The mottling
varies greatly in degree in the same species, or may even be wholly
wanting, and like the minute dotting, which is generally present, it
very often disappe irs in dried specimens. Only in J^. propullans do

OP AUTS AND SCIENCES. 127

the petioles form a closed slieath about the peduncle, and only in
E. Hartwegi are they alternate instead of opposite.

Tl.e auricles or appendages at the base of the inner petals are uni-
form and constant, so far as known, in each species where they occur,
though always greatly obscured in other than fresh flowers. Of the
eastern species, E. Americanmn is the only one with such special or-
gans, having a rather thickened auricle upon each side of the petal,
somewhat as in E. dens-canis. The western species, with the excep-
tion of E. HoweUii, have the inner petals appendaged with a trans-
verse crest of four more or less saccate gibbosities, the two inner the
more prominent, the outer forming lateral auricles, so that the crests
of the three petals when appressed to the ovary close completely over
the basal cavity of the perianth. These crests ditfer in some degree
in the different species, but usually not in such a wny as to make a de-
scription of the differences easy ; nor have they all been examined in
the fresh flower.

The stamens show little that is specifically cliaracteristic. They are
in two unequal series, with more or less dilated filaments, the yellow,
white, or occasionally purple anthers varying in length under different
conditions, a moistened anther becoming very much longer than the
same when dry. The relative lengths of the style and stamens vary
with the age of the flower. The coherence or divergence of the stig-
mas appears to be in general a good sectional character. In all cases,
however, the stigmas are at first coherent, and where separation occurs
it may be more or less delayed after anthesis and more or less com-
plete. Even in species with persistently coherent stigmas it is proba-
ble that separation occasionally occurs. The form of the capsule, while
differing in the two grdiips, is essentially uniform iu each. In the
western species it varies much in length, in some species more acute
than in others, in some pioportionately wider. No marked differences
have been observed in tlie seeds.

* Eastern species. Corni i^mnll (G to 9 lines lon^), oblong-ovate, often prop-
agatinp: by lengtliened offshoots, but also producing new cornis more or less
frequently at the base of the old: scapes low, 1 flowered : inner petals not
crested: capsule obovate (mostly 5 to 9 lines long).

1- Offshoots produced from the base of the corm.

1. E. Americanum, Ker. Leaves mottled: flowers yellow, often
tinged without with purple and finely dotted within ; segments 10 to
20 lines long, the iimer auricled near the base : style scarcely lobed
at the summit. — Bot. Mag. t. 1113; Bigelow, Med. Bot. t. 58; Bar-
ton, Fl. N. Amer. t. 33 ; Gray, Struct, and Syst. Bot., fig. 1247-1251 ;

128 PROCEEDINGS OF THE AMERICAN ACADEMY

Meehan, Nat. Flowers, 1st Ser., 1, t 17. E. Jlavum, Smith; Eaf.
Med. Fl. fig. ob. The ^'E. Carolinianum, Walt." of Poiret, Roemer
& Schiiltes, etc., was based upon Walter's '■^ Anonymos, Erythronio
a^ni's?" which must have been Ucidaria inrfoliata. Damp open
woodlands and banks ; Nova Scotia to Ontario and Minnesota, and
south to Florida and Arkansas.

2. E. 'ALBiDUM, Nutt. Leaves mottled : flowers white with more
or less of a bluish or purplish tinge, yellow within near the base, not
dotted, the segments strongly recurved, not at all auricled : stigmas
short (1 to H lines long), becoming recurved. — Similar localities;
eastern New York to Ontario and Minnesota, and south to Pennsyl-
vania, Virginia, Tennessee, and central Texas. The var. coloratum,
Sterns (Torr. Bull. 15. Ill), is the more deeply colored form.

3. E. mesochoreTjJI, Knerr. Resembling the last, but the leaves
narrower (| to 1 inch wide) and not mottled ; segments of the peri-
anth not recurved ; capsule larger (6 to 15 lines long). — Grassy prai-
ries or wooded slopes, from western Iowa to central and eastern Kansas.
First noted, as a variety of E. albidum, by Mr. R. Buigess (Bot. Gaz.
2. 115) and Mr. M. H. Panton (same, 2. 123) ; perhaps well separated
from that species by Prof. E. B. Knerr (Midland College Monthly,
2. 5).

-1- H- Offslioots produced from tlie sheathed portion of tlic scape.

4. E. PROPDLT.ANS, Gray. Leaves small (2 to 4 inches long)
above the close sheath, from within the base of which the offshoot
springs, slightly mottled: flowers rose-color with yellow base, small
(^ inch long), the inner segments not grooved nor auricled : stigmas
united ; capsule unknown. — Am. Nat. 5. 228, fig. 74. Southern
Ontario {fide IMacoun) ; southern Minnesota.

* * Western sjiecies. Corms usually elon<i:ated, rarely if at nil propacfating
by offshoots (except in n.O), the new conns borne upon a short rhizome : scapes
often tall, l-several-flovvered : inner petals auricled and transversely crested at
liase (except in n. 14) with four prominent gibbosities : capsule oblong, attenuate
below.

■t- Stigmas at length distinct and recurved.

++ Leaves not mottled (or rarely '') : flowers bright yellow.

5. E. GiJANniFLORUM, Pursh. Scape 1- (rarely 2-3-) flowered,
becoming 1 or 2 feet high : perianth strongly recurved, H to 2 inches
long: capsule 1 to 2 inches long, rounded or slightly retuse at the
summit. — Ijindl. Bot. Reg. t. 1786. E. yrandiflorum^ var. niinuy.
llnok., jNIorren, etc. ]\rouiitains of northern Idaho, AVashington, and
British Columbia near the boundary.

OP ARTS AND SCIENCES. 129

Var. PARVirLORUM. Scape usually low ; fluvvors smaller, the se^-
meuts 12 to 15 lines long. — E. Natlalliunum, Kej,a'l, Gartenil. t. G'J5,
not R. & S. ; E. grundlflorum, Murray, Gard. Cliron. 1874, fig. 173.
lu the mountains from Colorado and northern Utah to British Amer-
ica, in the Blue Mountains of Oregon, and in the Cascade Mountains
of Washington and British Columbia ; the more common form.

The var. Murrayi of Morreu (Belg. Hort. 26. 105, t. G) is a doubt-
ful cultivated form similar to this, but said to have mottled leaves. In
a single specimen collected by IVIr. Henderson above the timber-line
on the Olympic Mountains, Washington, the leaves are evidently
mottled.

++ ++ Leaves more or less mottled. Pacific Coast species.
= Corms producing slender offshoots from the base.

6. E. Hartwegi, Watson. Corms small (6 to 8 lines lono-') :
leaves (rarely 3) often alternate: flowers 1 to 3, pale yellow with an
orange base, the segments 1 to 2 inches long by 3 to 6 lines broad or
more, spreading or scarcely recurved : capsule unknown. — Proc. Am.
Acad. 14. 271. In the Sierra Nevada, California, from Mariposa to
Plumas Counties. This species, or a similar one, is also reported
from near Healdsburg, Sonoma County (^R. H. Thompson).

= = Corms (1 to 2 inches long) produced in succession upon a usually short

rhizome.

7. E. REVOLUTUM, Smith. Leaves attenuate to a usually narrow
petiole : scape often over a foot high, 1-2-flowered : perianth " white"
to pale yellow, yellow at base, the segments narrowly lanceolate (3 or
4 lines broad) : capsule abruptly acutish at the apex, 12 to 15 lines
long. — Eees' Cyc. E. grandijlorurn, var. Smithii, Hook. E. gran-
dijiorinn, var. revolutum, Baker. Vancouver Island to the lower
Columbia valley. Described by Smith as having purplish flowers, and
an original specimen in Herb. Kew bears the note " fl. rubr. purp." by
Sir W. J. Hooker, but it is rarely that the petals assume a pinkish
tinge in drying, and the ground for the statement is unknown. No
purple-flowered species is now found on Vancouver's Island, where
Menzies's specimens were collected.

Var. BoLANDERi. Usually low, 1-3- (rarely 4-) flowered ; perianth
white with yellowish centre, becoming rose-purple ; appendages very
prominent. — In the redwoods of Colusa, Mendocino, and Trinity
Counties, California.

8. E. GiGANTEUM, Lindl. Leaves narrowed to a usually short
and broadly margined petiole : scape often tall, 1-6-flowered or more :
perianth cream-color (often described as white) with a light yellow or

VOL. XXVI. (n. s. xviii.) 9

130 PROCEEDINGS OP THE AMERICAN ACADEMY

orange base and sometimes a transverse darker or brownish band, the
segments more broadly lanceolate (1 to 1^^ inches long by 4 to 7 Hues
broad) : capsule oblong-obovate (7 to 9 lines long), very obtuse or re-
tuse at the summit. — Hook. Bot. Mag. t. 5714 {E. granclijioriun^
Fl. de Serres, t. 2117). E. grandijlorum, var. cdhijiorum, Hook.
E. giganteum, var. alhijlorum, Gard. Chron. 1888^ fig. 74. From
the lower Columbia valley southward to Mendocino and Sonoma
Counties, California : March to June.

9. E. MONTANUM. Like the last, but the leaves (not mottled?)
more or less abruptly contracted and rounded at base; scape 1-2-
flowered ; perianth white vlth an orange base, often drying pinkish. —
On the high mountains of Oregon and Washington (Mt. Hood, Mt.
Adams, etc. ; Mrs. P. G. Barrett, Hoicell, and Snksdorf) ; in flower
from July to September.

I- -1- Style short-clavate, undivided ; scape a foot high or less ; leaves mottled ;

corms as in the last group.

++ Inner petals appendaged.

10. E. ciTRiNUM, Watson. Corms unknown : scape 3-flowered :
segments of the perianth broadly lanceolate, an inch long, strongly
recurved, light yellow with an orange base, the tips becoming pinkish ;
crest rather thin : capsule unknown. — Proc. Am. Acad. 22. 480. Deer
Creek Mountains, southern Oregon {T. Howell).

11. E. Hendersoni, Watson. Scape 1 -3-flowered : perianth 1 to
1^ inches long, strongly revolute, pale purple with a very dark purple
and yellowish base, the lobes of the crest subglobose-inflated : capsule
an inch long, very obtuse. — Proc. Am. Acad. 22. 479 ; Gard. Chron.
1888S fig. 86 ; Garden and Forest, 1, fig. 50; Bot. Mag. t. 7U17. In
the mountains of southern Oregon.

12. E. ruupuRASCENS, Watson. Corm 1 to 2 inches long: leaves
undulate : scape 1-8-flowered : perianth 9 to 12 lines long, spreading,
light yellow tinged with purple, deep orange at base: capsule obtuse
or retuse, 1 to 1^ inches long. — Proc. Am. Acad. 12. 277. In the
Sierra Nevada, from Placer to Plumas Counties, California. Inflo-
rescence more crowded than in any other species.

** +^ Inner petals not appendaged.

13. E. HovvELLii, Watson. Scape 1-3-flowered . perianth re-
curved, pale yellow with deep orange base, becoming pinkish : capsule
unknown. — Proc. Am. Acad. 22. 480. Josephine County, southern
Oregon {T. Howell).

OF ARTS AND SCIENCES.

131

ZoSTERA Okkg.vxa. Stem slender: leaves H to 2 lines broad,
rather faintly o-uerved : spathes 2 or 3 inches long, very obtnse, wIlIi
a short scarious crest but without a foliar appendage at the sumuiit ;
spadix short-acuminate : utricle with a long straight beak ; seed (im-
mature) narrowly oblong, with IG longitudinal striie. — Collected by
Elihu Hall in 1871, probably near the moutli of the Columbia River,
Oregon ; in Herb. Gray, without number, and not mentioned in the
I)ublished list of his distributed Oregon collection. Readily distin-
guished by the uuappendaged spathe. The spadix examined was
about 15-lruited; the anthers had all fallen.

Zostera Pacifica. Stem stout ; cauline leaves 4 to G lines
broad, 5-y-nerved, those on the branches 2 lines broad and 3-nerved :
spathe nearly 3 inches long, with a long foliar appendage ; spadix
acutish, many-flowered : utricle with a long straight beak ; seed
broadly elliptical, compressed, 1^ lines long, not evidently striate but
appearing under the microscope very finely transversely striolate. —
About Puget Sound (Rev. R. D. Nevius) ; Monterey {Dr. G. L. An-
derson) ; Santa Barbara {Mrs. R. F. Bingham). A specimen received
from Mr. Nevius is eight feet long, with some of the leaves four feet
in length. The anthers are mostly arranged in vertical pairs, alter-
nating with solitary ovaries. The seed is remarkably distinct from
that of Z. marina, which is shorter, oblong, terete, and conspicuously
20-striate. The species has been described by Mr. Morong (Torr.
Bull. 13. 160) as Z. marina, var. (?) latifoUa.

2. Descriptions of nciu Mexican Species, collected cliicfiy Vy
Mr. C. G. Prinrjk in 1889 and 1890.

Ranunculus vagans. Near R. hydrocharoides and R. stoloniftr,
low, spreading by elongated stolons, glabrous : leaves narrowly lanceo-
late or the lowermost ovate-lanceolate, entire or with a few often slen-
der teeth toward the apex: petals 8 to 10, oblong-obovate, about 2|-
lines long and nearly twice the length of the sepals, with a prominent
gland above the narrow claw: carpels smooth, in a dense globose
head 1\ lines in diameter. — Found iu shallow ponds near Flor de
Maria in the State of Mexico, August, 1890 (n. 3177).

Nasturtium bracteatum. Annual, erect, unbranohed, glabrous
or slightly and finely pubescent (G inches high or less) : leaves sessile,
auricled at base, narrowly oblong, pinnately toothed, 6 to 9 lines long:
raceme many-fiowered and becoming elongated ; fruiting pedicels di-
varicately spreading, 3 lines long, the lower solitary in the axils of the

132 PROCEEDINGS OF THE AMERICAN ACADEMY

upper leaves : flowers apparently white, a line long : pod elliptical, l^-
to 2 lines long, beaked with a short style; stigma small. — N. paliistre,
Benth, PI. Hartw. 9, Described from a few small but fruiting speci-
mens in Herb. Gray, collected by Ilartweg (n. 39) at Aguas Cali-
entes in the Mexican State of the same name. The lower axillary
pedicels are an unusual character.

Sisymbrium multikacemosum. Finely stellate-pubescent through-
out, the lax and slender stems procumbent, branching, 2 feet long or
more : leaves narrowly lanceolate, attenuate to both ends, not auricled
at base, sinuately serrulate or sometimes serrate, 1 to 2 inches long :
racemes in most of the axils, on very short leafy peduncles or nearly
sessile, 1 to 3 inches long in fruit: flowers very small, white: pods
divaricately spreading on pedicels about a line long, pubescent, narrow
and subcylindrical, about 3 lines long, beaked by a short slender style :
seeds in one row, 8 to 10 in each cell. — At Las Canoas, San Luis
Potosi ; October, 1890 (n. 3522). A species of strongly marked habit.

PoLYGALA SUBALATA. Annual, the several stems erect from an
ascending base, branching above, narrowly wing-angled, leafy, gla-
brous or slightly puberulent, G inches high or less : leaves mostly ver-
ticillate, oblanceolate, very acute, 4 to 6 lines long, scabrous on the
margin, the lowermost small, obovate to spalulate and obtusish, the
upper becoming linear ; spikes sessile, dense, acuminate, becoming
elongated and looser: flowers small, very shortly pedicellate; petals
white with a broad green or purplish mid vein : capsule very broadly
elliptical, equalling the petals (a line long) : seeds oblong, sparsely cov-
ered with a very fine appressed silky pubescence, the linear appendages
of the hilum as long as the seed. — In low grounds at Flor de Maria,
State of Mexico, September, 1890 (n. 3240). Resembling P. aiha,
especially its var. siispecta, from which it is most positively distin-
guished by the much less pubescent seed nearly equally broad at both
ends.

Talinum Coaiiuilense. Stems very short and leafy, from a slen-
der branching rootstock or rhizome bearing oblonij or ovate tubers :
leaves rhombic-obovate, cuneate at base and nearly sessile, inflated-mar-
gined, ^- inch long or less : flowers solitary on short pedicels (^- inch
long) ; sepals round-ovate, acute, strongly concave, 2 to 2^ lines long;
petals round-obovate, 3 lines long : stamens numerous with very slen-
der filaments and short anthers: capsule ovate. — On limestone hills
at Carneros Pass, Coahuila ; INIay, 1890 (n. SOOG). Much resem-
bling T. hrevifoUum and T. breriroit/e, especially the latter, differing
iu the tuberous roots and broader leaves and sepals.

OF ARTS AND SCIENCES.

133

SiDA Alamosaxa. I'erenniul (?), herbaceous, erect, slender and
brauchiug, finely «ilaiidiilar-|)ubescent : stipules filil'orin, 2 to 4 lines
long ; leaves on slender petioles, ovate to lanceolate, acuminate, more
or less cordate at base, lallier acutely serrate, 2^ inches long or less,
somewhat hairy especially on the nerves (not glandular-pubescent),
the hairs simple or forked : peduncles axillary, slender, mostly an inch
long or more: corolla "orange" (apparently white), 3 lines long, ex-
ceeding the acuminate-toothed calyx : carpels 5, little over a line long,
ovate, glabrous except the summit, thin-walled, terminating in a con-
tracted truncate dehiscent cavity above the seed. — Collected at Ala-
mos, Sonora, by Dr. Edward Palmer (u. G83j in September, 1800.
Closely resembling S. uhnifolia in habit, but more glandular-pubes-
cent, with longer filiform stipules, more acutely serrate leaves, and
especially distinguished by the unusual character of the carpels.

Ayenia Berlandieri, Watson, Proc. Am. Acad. 21. 419, in part.
Mr. Pringle has collected flowering specimens which accord in foliage
with the original specimens of Berlaudier, but with different flowers
from those of the plants of Dr. Palmer's Cliihuahuan collection upon
which my description was largely based. It becomes necessary to
separate the two forms and to redescribe this species. — Plant 3 to 6
feet high, the herbaceous branches more or less stronofly angled and
sulcate : leaves ovate to ovate-lanceolate, acute, rounded or slightly
cordate at base, densely pubescent beneath with a soft stellate tomen-
tum, greener above, serrulate, 2 to 4 inches long: peduncles 1-2-flow-
ered, fascicled in the axils or in a naked terminal inflorescence: calyx
and petals dark purple, the former 2^ lines long, the blade of the
petal parted into two broad quadrate aud truncate lobes : anthers 3-
celled : ovary and capsule rather long-muricate. — Berlandier ; Prin-
gle (n. 3309), in low lands at Las Palmas, San Luis Potosi.

Ayenia Jaliscana. Differing from A. Berlandieri in its more
terete branches and thinner, less pubescent, and more coarsely toothed
leaves; flowers smaller and paler, the sepals li lines long, and the
lobes of the petals oblong and acutish ; anthers 2-celled and capsule
more shortly miiricate. — Southwestern Chihuahua, Dr. E. Palmer
(n. I'J and 83), and apparently also at Guaymas (n. 2-13 of his 1887
collection).

BuNCHOsiA Prtnglf.i. a shrub or small tree (12 to 20 feet high) ;
branchlets and inflorescence appressed-hairy : leaves thin, oblong, ob-
tusely short-acuminate, acutish at base, eglandular, glabrous above,
sparsely hairy beneath, 3 or 4 inches long by 1 or 2 broad, on petioles
3 lines long : racemes mostly solitary in the axils and simple, shorter

134 PROCEEDINGS OP THE AMERICAN ACADEMY

than the leaves ; pedicels glanduliferous, stout and 2 or 3 lines long
in fruit: ovary pubescent, globose, beaked by the subpersistent style;
fruit 2-3-lobed, the seeds 4 lines in diameter. — In Tamasopo Canon,
San Luis Potosi ; July, 1890 (n. 3540). Near B. lanceolata of Turcz-
aninovv.

Sargentia (?) Pringlei. a shrub, 10 to 15 feet high: leaves
coriaceous, pinnately 1-foliolate or sometimes palmately 3-foliolate, the
lower surface (with the petioles and branchlets) pubescent with very
short spreading hairs, glabrate above; leaflets oblong-oblanceolate,
acutish or obtuse, cuncate at base, entire, puncticulate, 1 to 2\ inches
long : sterile flowers small, in a very short terminal few-flowered pani-
cle, with very short triangular sepals and 4 (or 5) narrow oblanceo-
late valvate petals, concave and spreading : disk inconspicuous or
none, the stout filaments (usually 4) inserted at the base of the small
abortive ovary ; anther-cells distinct to above the middle, acute at
base, rounded above : fruit on short stout pedicels solitary at the ends
of short branchlets, depressed-globose (6 to 9 lines broad), mostly 2-3-
celled and 2-3-seeded, the oblong-obovate seeds G lines long. — The
fruit is essentially that of S. Grc(jfjii, though less lobed, but the flowers
differ so widely in being usually tetramerous, with valvate petals, no
disk, and dissimilar anthers, that the correctness of the reference to
Sargentia is very doubtful. The fertile flowers are as yet, however,
unknown. In S. Greggii the anthers are cordate with an acute apex.
Found in the mountains of San Luis Potosi at San Jose Pass, in June,
1890, in fruit, and in July in flower (n. 3220), It appears to have
been also collected by Berlandier in 1828 at INIonterey, without flowers
or fruit, the specimen in the Gray Herbarium without num])cr, under
the name of Choisya iernata.

Xanthoxylum Pringlei. A tree, becoming 40 feet liigh, gla-
brous : leaflets 3 to 5 pairs, lanceolate, acuminate, rounded at base,
entire, coriaceous, epunctate or with a {ew large glandular dots near the
apex, 1^ to 2.} inches long on petiolules 1 to 2 lines long : flowers di-
oecious, very numerous in large terminal corymbs ; pistillate flowers
with a very short 3-lobed calyx and three elliptical petals: carpels
solitary ; fruit globose, tubcrculate, 1^ lines in diameter. — In Tama-
sopo Caiion, San Luis Potosi, in June, in flower ; October, in fruit
(n. 3102).

NEOPRINGLEA. As the genus JJm'ea of Liebmann is long ante-
dated by the Llavea of Lagasca, a genus of ferns that is generally
considered as well founded, it is necessary to substitute another name
for that of Liebmann. I take pleasure in dedicating the genus to one

OF ARTS AND SCIENCES. 135

most wortlij'. whose botanical collections in Mexico have been un-
excelled in character and who has added very greatly to our knowl-
edge of the Mexican flora.* The genus has hitherto been placed
jirovisionaliy in the Cclastracece. Its affinities are rather with Alva-
rudoa in the Saplndacece.

Neotringlea integrikolia. {Llavea inter/rifolia, Ilemsl ) Com-
plete specimens of this hitherto imperfectly known dioecious shrub
have been collected by Mr. Priugle on San Jose Pass, San Luis Po-
tosi ; in flower, July (n. 3137), and in fruit, October (n. 3248). The
staminate flowers are in short crowded axillary racemes, tetramerous,
the four greenish petals orbicular, pubescent, not exceeding the sepals :
stamens 12, in threes opposite to the petals, in the sinuses and inter-
vals between the pubescent lobes of the disk ; pistillate flowers with-
out petals or disk, the 3-winged ovary 3-celled at base, with solitary
anatropous ovules on the axis : capsule 1-celled, 1-seeded ; seed com-
pressed-obovate, the embryo straight in rather thin albumen, with flat
cotyledons and rather slender inferior radicle.

Desmodium subspicatum. Stems erect from a narrow fusiform
root, somewhat woody at base, glabrous or slightly puberulent, 18
inches high : leaves nearly sessile, the single leaflet linear-lanceolate, ob-
tusish at base, acuminate and ver}^ acute, 1 to 3|- inches long, glabrous
or nearly so : racemes pubescent, terminal, simple and spike-like,
the small greenish flowers solitary or in pairs, on jjedicels a line long
or less, the lanceolate bract about equalling the calyx : legumes erect,
pubescent, 9 lines long or less, 3-5-joiuted, nearly equally indented
on both sutures and the joints elliptical. — On grassy hillsides at Las
Canoas, San Luis Potosi ; July. 1890 (n. 3211). Kear D. amjustifo-
liiim, DC, as that species is described.

Desmodium amaxs. Tall, erect, the stem, petioles and inflores-
cence clothed with shoi-t hooked hairs : stipules attenuate from a
broad base, villous ; petioles about \ inch long ; leaflets rather thick
and reticulately veined, narrowly oblong, acute, rounded at base, gla-
brous above, villous beneath, 14^ to 2 inches long by 6 lines broad ;
stipels subulate, attenuate: racemes in a naked terminal panicle;
bracts ovate, acuminate, 3 lines long ; pedicels becoming 3 or 4 lines
long : calyx-teeth exceeding the tube, unequal, the lower longer and
narrower ; petals purple, 3 lines long : legume 5-7-jointed, with the
two sutures nearly equally sinuate, the suborbicular joints \-\ lines

* The Privgleoph]itiim dedicated to Mr. Prinsle by Dr. Gray is identical with
Berginia, Harvey, as was first suggested by Mr. Brandegee.

136 PROCEEDINGS OF THE AMERICAN ACADEMY

broad, densely uncinulate-liispid. — Ou hillsides at Las Cauoas, San
Luis Potosi ; October, 1890 (n. 3291).

CoLOGANiA Jaliscana. This name is to be substituted for C.
Pringlei upon page 147 of the last volume of the Proceedings, on
account of a previous species so named in Mr. Pringle's collection of
1887.

Begonia (Weilbachia) Pringlei. Rhizome slender, covered
with brown ovate acuminate glabrous stii^ules : petioles tomentose, 2
to 4 inches long ; leaves obliquely rhombic-ovate, acute, cordate at base,
coarsely and subsinuately toothed, sparsely pubescent above, somewhat
tomentose and densely papillose beneath, 2 or 3 inches long : pe-
duncle 4 to 8 inches high, red, slightly i)ubescent, 4-6-flowered at the
summit : flowers apetalous, the two sepals round-reniform or orbicu-
lar, 4 to 6 lines bread : fruit with two broad truncate-rounded wings,
the third smaller. — Ou cold ledges in Tamasopo Canon, San Luis
Obispo ; October, 1890 (n. 3514).

Ertngium Mexicanum. a span high, spreading, twice branched :
basal leaves with a nodose petiole sheathing at base, the blade linear
or linear-lanceolate, sparsely serrate, usually with a pair of similar but
shorter lobes at base ; cauline leaves nearly sessile, similar : heads
pedunculate, ovate or oblong-ovate, 4 to G lines long, terminated by
one or sometimes two or three slender foliaceous appendages, entire
or occasionally cLft above, and subtended by about ten longer linear
acute bracts, whitish above and usually with a slender tooth on each
side ; bractlets none or minute : fruit covered with numerous white
scales, the calyx-lobes dark blue. — In wet meadows at Del Rio, State
of Mexico; August, 1890 (n. 3229).

Arracacia Mariana. Stems decumbent from a branching cau-
dex, slender, simple or sparsely branched, a foot long, glabrous : leaves
near the base, thin, pinnate, the leaflets 4 or 5 pairs, an inch long or
often much less, acute, sharply serrate, the lower usually with a pair
of linear lobes at base : umbels on long terminal peduncles ; bracts of
tiie involucre and involucels narrowly linear; rays (6 to 10) in fruit 4
or 5 lines long : flowers yellow : fruit nearly sessile, ovate, acutish, 3
or 4 lines long, strongly and obtusely ribbed ; ventral sinus of the seed
completely closed. — On hillsides at Flor de Maria, State of Mexico ;
July, 1890 (n. 3480).

Arracacia multifida. Tall and stout, glabrous : leaves ample
(a foot long or more), sessile upon a short sheathing base, ternate and
several times pinnate, the ultimate segments very narrowly linear,
entire, acute, an inch long or usually less : umbels on stout peduncles,

OF ARTS AND SCIENCES.

137

without iuvolucreor iuvolucels ; rays numerous, 1 to 1^ inches luii" in
fruit ; pedicels 2 or 3 lines long : fruit 3 or 4 lines long, oblonji-, the
carpels narrowed at each end and beaked by the stout erect stylophove
and style, acutely angled by the thin prominent ribs ; vittse quite varia-
ble, 1 to 3 in the broad intervals and usually 2 on the narrow com-
missure ; the deep sulcus very narrow. — On hills at liio Hondo, State
of Mexico; August, 1890 (u. 3G20).

CiiOMELiA Pringlei. A Small tree (15 feet high); branchlets pu-
bescent with short spreading hairs : leaves oblong-ovate, rounded at
base, short-acuminate, finely pubescent beneath, slightly scabrous
above, 1^ to 3 inches long, on petioles 1 or 2 lines long: peduncles
slender, G to 12 lines long; flowers few, sessile or nearly so, subcapi-
tate or in a very small loose cyme ; calyx pubescent, equally toothed,
scarcely a line long ; corolla reddish, 3 or 4 lines long, rather broadly
tubular and but slightly dilated above, the lobes appressed-pubes-
cent, a line long : fruit compressed-oblong, sjuirsely pubescent, 3
lines long. — In Tamasopo Caiion, San Luis Potosi; August, 1890
(n. 3209).

Crusea megalocarpa. {Spermacoce megalocarpa, Gray, Proc.
Am. Acad. 21. 381.) Collected by Mr. Pringle in the barranca near
Guadalajara; September, 1889 (u. 2968). This species was referred
by Dr. Gray to the section Borreria of Spermacoce, but the charta-
ceous cocci separate from a persistent septum which is cleft nearly to
the middle and crowned by the persistent linear-lobed calyx-limb. It
therefore accords perfectly with Crusea as that genus was character-
ized in De CandoUe's Prodromus and by Bentham & Hooker in the
Genera Plantarum. The " calycis limbi lobi persistentes " of the lat-
ter is criticised by Dr. Gray, and rightlj^ if it were to be understood as
meaning that the calyx-limb is persistent upon the cocci. It always
separates from the cocci, and may either remain persistent upon the
persistent axis, as is the statement of De Candolle, or break away
from this also, as in C. svhulata and some other species.

Edpatorium Madrense. Branches woody, slender and lax, the
branchlets finely pubescent, leafy : leaves opposite, small (an inch
long), very shortly pedicellate, ovate-lanceolate, acute, rounded or
subcordate at base, 3-nerved, serrate except toward the apex, roughish
above, tomentose beneath : heads few (3 to 6), terminal and axillary
on short peduncles, many-flowered, narrow below, 4 lines high ; invo-
lucre 3 lines long, shorter than the disk, its unequal scales in several
series, nerved, subtomentose, acute, the outermost narrowly ovate, the
innermost linear : achenes slightly hispid toward the base on the acute

138 PROCEEDINGS OP THE AMERICAN ACADEMY

angles ; pappus in a single series, slightly scabrous. — In the Sierra
Madre near Monterey; June, 1888 (n. 2201).

EuPATORiUM (?) Chapalense. Woofly, the branches terete, gla-
brous : leaves opposite, ovate-lanceolate, acute, rounded at base but
decurrent upon the slender petiole, 3-5-nerved above the base, serrate,
puberulent above on the veins and villous-tomentose on the nerves
beneath, 1|- to 2^ inches long, the petioles about an inch long: heads
on slender bracteate puberulent peduncles in a dense terminal corymb,
20-flowered or more, narrovp at base ; involucral scales mostly nearly
equal, 4 lines long, thin, linear, acute, pubescent; receptacle depressed-
conical, smooth and naked : achenes (immature) apparently obtusely
pentagonal, papillose; pappus-bristles in more than one row, distinctly
barbellate, unequal, the outer very short. — In the mountains near
Lake Chapala ; December, 1889 (n. 2974). Near E. multisei-ratum,
Schultz Bip., in habit and involucre, but the pluriseriate unequal pap-
l)us is abnormal. The unequal pappus suggests Piptolhrix, but the
setce are much more numerous and not caducous, while the heads,
flowers, and pappus are twice longer than in either of the species of
that genus.

OLIGONEMA, a new genus of the homochromous Asteroidece.
Heads many-flowered, radiate and heterogamous, the ray- and disk-flow-
ers all fertile. Involucre broadly campanulate, of many nearly equal
herbaceous scales in several series, the inner series becoming some-
what coriaceous with herbaceous tips. Receptacle nearly flat, naked.
Ligules linear, yellow, slightly 3-toothed; disk-corollastubidar,tlie throat
scarcely dilated and about equalling the oblong acute lobes. Anthers
obtuse at base. Style-branches tipped with attenuate conical appen-
dages. Achenes of the ray compressed-triangular, those of the disk
obcompressed, elliptical, glabrous, rather thin-margined, truncate, the
thickened margin of the summit bearing 1 to 4 very delicate lax
scabrellate deciduous hairs. — A stout annual aquatic or marsh herb,
with alternate linear entire leaves, the submerged pinnately parted ;
iieads rather large, terminal on the branches. A genus most nearly
related to Grindelia^ strongly marked by its habit and the characters
of the achenes and pappus.

O. HETEuorriYLLA. Mostly glabrous, somewhat glandular-pubes-
cent above, the stem 2 or 3 feet high, fistulous, sending out delicate
rootlets from the submerged nodes : leaves sessile, clasping, acumi-
nate, 2 or 3 inches long, the submerged pinnate with narrowly linear
entire or toothed segments : heads half an inch broad, the orange lig-
ules G to 8 lines long ; involucral scales lanceolate, acuminate : achenes

OF ARTS AND SCIENCES.

139

about 1^ lines long, somewhat longer than the delicate pappus. — In
shallow water at Del Rio, State of Mexico ; August, 181)0 (u. 3236).

Acii^ETOGEHON LiNEARiFOLius. Biennial or perennial (?), branch-
ing from the base, the stems ascending or decumbent, a foot high or
less, somewhat strigose-pubescent: leaves numerous, dark green, nar-
rowly linear, entire or rarely with one or two narrow lateral lobes,
mostly 1 to 1^ inches long, acute, narrowed to the base: heads hemi-
spherical (4 or 5 lines in diameter), with hemispherical receptacle and
linear acuminate involucral bracts: rays very numerous in several
rows: acheues (immature) compressed (?), sparsely hispidulous ; pap-
pus coroniform, dentate and laciniately denticulate, nearly as long as
the proper tube of the corolla. — BlutTs and plains near Flor de Maria,
State of Mexico ; Sept., 1890 (n. 3242). The distinctions between the
genera Aphanostephus and Acltcetogeron are very vaguely defined,
the latter genus appearing to rest upon the compressed Erigeron-like
acheues.

PsiLACTis TENUIS. Erect, sleudcr (2 feet high), branching above,
the slender spreading branches simple or few-flowered, rough-hispid
with spreading hairs : cauliue leaves oblong-oblanceolate, acute, nar-
rowed to a short winged petiole, sharply serrate, 1 to 1^ inches long,
those on the branches gradually smaller and narrower: heads solitary
and terminal, small (2 lines high) ; involucral scales thin and scarcely
herbaceous, narrowly linear, acuminate : ligules purplish, 3 lines long:
acheues sparsely pubescent ; pappus of disk-achencs of numerous
(about 30) barbellate setae. — In the Sierra Madre near Monterey;
June, 1888 (n. 2238).

Aster Carnerosanus. Stems a foot high or less, from slender
rootstocks, slender, purplish, pubescent: leaves oblanceolate, 1^ inches
long or less, mostly entire, glabrous or nearly so, shortly ciliate, the
bracteal oblong, 2 or 3 lines long : heads middle-sized (5 lines high),
solitary on the short branches, the numerous foliaceous-tipped scales
somewhat spreading, oblanceolate, acute or the innermost subacumi-
nate : rays pale purple. — At Carneros Pass in the mountains of Coa-
huila ; Sept., 1889 (n. 2859). Rather closely resembling A. surculosics
of the Alleghanies, but with smaller heads.

Melampodium glabrum. Nearly glabrous (slightly scabrous
above, especially on the peduncles), decumbent, branching : leaves
oblong- to linear-lanceolate, scarcely narrowed to the broad clasping
base, acutish, sparingly toothed, 1| inches long or less: heads small
(2 to 2^ lines high), on slender peduncles, the 5 or 6 ovate scales
obtuse or acutish ; ligules oblong, shorter than the scales, yellow :

140 PROCEEDINGS OF THE AMERICAN ACADEMY

achenes triangular-obovate, over a line long, transversely rugose. —
Valley near Irapuata, Guanajuato ; Sept., 1889 (u. 2821). Near
M. montanum.

Melampodium (Unxia) bibracteatuji. Annual, glabrous, a foot
high; leaves oblong to oblauceolate, acute, sparingly toothed, |- to H
inches long: heads sessile; outer involucre of two foliaceous broadly
ovate opposite bracts, 3 lines long, the inner saccate, closely enclosing
and conformable to the achenes, thin and somewhat reticulated, un-
a^jpendaged : disk-flowers 5, sterile; ray-flowers 5, the yellowish ligule
very short (^ line long or less) : achenes obliquely obovate and com-
pressed, smooth and coriaceous, a line long; pappus none. — In fields
at Del Rio, State of Mexico; August, 1890 (n. 3230).

TiTHONiA MACROPiiYLLA. Leaves thin, scabrous above and on the
veins beneath, the cauline large (about a foot long) with a broadly
ovate blade cuneate from a shallow sinus and decurrent into the winged
petiole, deeply 3-lobed, the lobes narrowly acuminate and serrate ;
uppermost leaves much smaller, ovate, not lobed : peduncles and heads
sparingly pubescent; outer involucral scales foliaceous, narrowly lan-
ceolate, acuminate, an inch long or more, the inner somewhat shorter
and broader: rays orange, l^- inches long: achenes appressed-silky,
4 lines long ; pappus of two awns and a crown of six broad and very
obtuse denticulate scales half as long. — Barranca, near Guadalajara ;
Sept., 1889 (n. 2798).

ViGUiERA LEPTOCAULis. Annual; stem erect, slender, 3 or 4
feet high, sparingly branched and somewhat strigose above, glabrous
below : leaves opposite, narrowly lanceolate, abruptly cuneate to a
short petiole, attenuate upward from near the base, sparsely serru-
late, very scabrous above, 4 inches long : heads solitary on the few
branches; involucral scales linear-lanceolate, long-acuminate: ray-
flowers sterile ; ligules orange, an inch long : achenes glabrous, com-
pressed, bearing a single caducous attenuate chaffy palea dilated and
lacerate at base. — In the Sierra Madre near Monterey: July, 1888
(n. 2247).

OroPAPPUS ACUMiNATDS. Tall, scabrous: leaves opposite on slen-
der petioles (3 to 6 lines long), ovate- to oblongdanceolate, narrowly
long-acuminate, rounded or subcordate at base, distantly glandular-
serrulate, very scabrous above and scabrous-hirsute on the veins be-
neath, 3 to G inches long by | to H broad: heads in terminal and
axillary pedunculate corymbs, broad and many-flowered, radiate, 3
lines high ; involucral scales in several series, short, narrow, somewhat
squarrosely tipped: ligules small (1^ lines long) : achenes of the ray

OF ARTS AND SCIENCES.

141

3-wingcd, the two outer wings narrow, the inner broader, produced
above and adnate to tlie awn of the ])appus ; disk-achenes compressed,
simihirly winged on one edge, scarcely at all on the other, the pappus
reduced to the two very unequal processes at the angles : appendages
of the style-branches acuminate. — Barranca near Guadalajara ; Oc-
tober, 1889 (n. 2999.)

SqiLANTiiES BoTTKKii. Annual, branching, about a foot high, spar-
ingly pubescent or glabrate : leaves ovate to oblong-ovate, cuneate at
base and shortly petiolate, acute or shortly acuminate, 1 or 2 inches
long : peduncles slender, mostly equalling or somewhat exceeding the
leaves, nearly glabrous : heads ovate-conical, 3 to 5 lines long, dis-
coid, the glabrous narrowly oblanceolate involucral scales 2 or 3 lines
long: achenes ciliate, the sides nearly glabrous, striate; pappus of
two slender bristles. — Near Guadalajara; November, 1889 (n. 2946);
also collected in Orizaba by Botteri (n. 825 in Herb. Gray).

Salmea Palmeri, Stem erect, rather stout, 3 to 5 feet high or
more, branching, glabrous or slightly pubescent above : leaves subcori-
aceous, opposite, very shortly petiolate, ovate to ovate-lanceolate, acute
or acuminate, rounded at base, distantly glandular-serrate or serrulate,
strongly veined and slightly pubescent, scabrous on the margin, 2 to 4
inches long : heads 2 to 5 lines high, i'ew to many, closely corymbose ;
scales acute, finely pubescent above, as also the chaff of the recepta-
cle : corolla-tube as long as the throat, the limb deeply cleft : achenes
compressed or more or less acutely 3-4-angled, short-ciliate on the
angles, the sides nearly glabrous ; pappus of two nearly equal atten-
uate bristle-like palea3, or of three or four more unequal ones. —
Near Guadalajara; n. 528 Palmer and n. 2345 Pringle (1889), both
distributed through some inadvertence as S. grandiceps, Cass., to
which they have little resemblance; also Pringle's n. 2155 and 2170
of 1889.

Dahlia dissecta. Stems slender, lax, ascending from a decum-
bent or procumbent woody base, 1| to 2 feet high, glabrous through-
out : leaves deltoid in outline, 3 to 6 inches long, twice or sometimes
thrice pinnate, the lanceolate to linear segments mostly laciniately
lobed : heads long-pedunculate ; outer involucral scales suborbicular,
3 lines long, the inner narrowly oblong, 6 to 8 lines long ; rays pur-
ple or "mauve-colored," 1 to 1| inches long : achenes linear-oblanceo-
late, truncate, I lines long, nerved and with a strong central ridge on
both sides, somewhat minutely pubescent on the inner surface, — On
limestone ledges at San Jose Pass, San Luis Potosi ; July, 1890
(n. 3167).

142 PROCEEDINGS OP THE AMERICAN ACADEMY

Dahlia pubescens. Stem erect from a cluster of small tubers,
1^ to 3 feet liigb, leafy, more or less pubescent witb mostly sbort stiff
pointed bairs : leaves pinnate, 3 to 4 incbes long, tbe 5 to 7 leaflets (an
incb long) ovate to lanceolate, coarsely and acutely tootbed: heads erect
on loug peduncles, coarsely pubescent at base, tbe outer scales ovate-
lanceolate, acutisb, 4 or 5 lines long, the inner lanceolate, 9 lines long
in fruit; rays purple, 12 to 15 lines long: acbenes nearly as iu tbe
last, but usually shorter (3 lines long) and broader. — On limestone
bluffs at Flor de Maria, State of Mexico; August, 1890 (n. 3164).

BiDENS DAHLioiDES. Low and slender, glabrous, bearing tubers
alonw tbe slender branches of a divided rootstock : lower or lowermost
leaves entire, lanceolate or oblanceolate, acute, attenuate to the peti-
ole; the upi)er (3 inches long) pinnately 3-5-lobed, tbe terminal seg-
ment broadly ovate to lanceolate, less deeply 3-lobed, the lateral
ovate-lanceolate, entire : heads solitary upon elongated terminal pe-
duncles, the purplish ray 1^ inches long : scales of the two involucres
similar and nearly equal, many-nerved, 6 lines long : acbenes oblan-
ceolate, 5 lines long, flattened, with iw^o short stout retrorsely bearded
awns at the angles. — Dahlia-like in habit, but with the awned acbenes
and short-appendaged style-branches of Bidens. On hillsides at Flor
de Maria, State of Mesico ; Sept., 1890 (n. 3168).

Bahia Schaffneui. Annual, branching from the base, more or
less decumbent or procumbent and strigulose-puberulent: leaves sub-
ternately decompound, the segments linear and usually short : heads
about 3 lines high, the involucre of 8 to 10 broadly oblanceolate ob-
tuse or acutisb bracts : ligules deep yellow, short : acbenes very nar-
rowly obpyramidal, somewhat hispid on the angles, especially near
the base, 1^ lines long or more; pappus of five obovate imbricate
scales, about |- line long, nearly equalling the slender corolla-tube. —
Sandy plains near San Luis Potosi ; May, 1889 (n. 3028), Also col-
lected by Dr. Schaffner (n. 327, in part) in the same locality, and by
Parry & Palmer (n. 494), referred to Ji. anthemoidcs. It is the same,
moreover, as a plant cultivated from Mexican seeds in the Jardin des
Plantes iu 1838 and 1839, of which there are s])ecimens in Herb.
Gray. B. anihemoides, Gray, with which the species has been con-
fused, is represented in Herb. Gray by specimens collected at Toluca
by Berlandier, and by 327 Schafl'ner (in part), probably from the
valley of Mexico ; it is also 3143 Pringle, collected near the city of
Mexico. As described and figured in HBK. Nov. Gen. & Spec, it is
characterized by white ligules and by a much shorter and broader and
more pubescent achene.

OF ARTS AND SCIENCES.

143

Senecio Jaliscana. Perennial, tall and erect, rather stout, white-
tomentose throughout; leaves ovate to oblong-ovate, thickish, subcor-
date or truncate at base, acute, sinuately and acutely few-toothed and
glandular-denticulate, densely tomentose beneath, subglabrate above,
3 to 4 inches long or less ; heads discoid, many-fiowercd, in rather
dense compound corymbs terminating short peduncle-like branches
which are bracteate only above ; involucre short (2 lines long), of 8
or 10 imbricated scales, calyculate at base: achenes glabrous; pappus
white, the somewhat rigid setje scabrous and often thickened at the
tips. — With the foliage and habit nearly of S. sinuatus, but the in-
volucre and pappus of S. barha-Johannis. In the Chapala Mountains
near Guadalajara; December, 1889 (n. 2931).

Cacalia (Conophoka) poculifera. Nearly glabrous, puberu-
lent above, 3 or 4 feet high, branching and naked above: leaves few,
the lower long-petiolate, centrally peltate with depressed centre, 6 to S
inches broad, deeply 7-9-lobed with narrow sinuses, the broad lobes
sinuately toothed ; upper leaves subreniform, lobed and decurrent into
a short very broadly auriculate-clasping petiole, or cordate-ovate, ses-
sile and coarsely and acutely toothed : heads 5-flowered on short pedi-
cels in rather dense terminal corymbs ; involucre of 5 glabrous scales,
2 lines long, slightly calyculate; receptacle flat, naked : corolla-lobes
as long as the tube : achenes glabrous or nearly so ; pappus fuscous.
— Near Guadalajara, Jalisco ; July, 1889 (n. 2879).

Cnicus velatus. Stem slender, erect from a tuberous-fascicled
root, branched above : radical leaves linear-oblanceolate, 6 to 12 inches
long by |- to 1 broad, long-petiolate, glabrous or slightly floccose-pubes-
cent above, white-tomentose beneath, entire with short scattered spines
upon the margin or sparsely sinuate-lobed, the lobes acute and spar-
ingly spinulose ; cauline leaves sessile and decurrent, linear-lanceolate,
the shoi't lobes and auricles slightly more strongly spinulose : heads
solitary on the elongated branches, campanulate, 9 to 12 lines high,
the lanceolate outer scales tipped with a short appressed spine and
covered by a veil of floccose hairs : flowers Y>^\e pui'ple : anther-tips
attenuate. — In low meadows, Flor de Maria, State of Mexico; Au-
gust, 1890 (n. 3228).

Cnicus (Echinais) linearifolius. Stem erect, very leafy and
wing-angled, simple : leaves linear, the radical 12 to 1/5 inches long by
an iTich broad or less, petiolate, pinnately many-lobed to the middle,
the broad lobes and sinuses undulately margined and spinulose, rough-
ish above, white-tomentose beneath ; the cauline similar but smaller,
sessile and decurrent, acuminate, erect : heads small (9 lines high).

144 PROCEEDINGS OF THE AMERICAN ACADEMY

broad, nearly sessile in a terminal cluster ; outer scales shortly spi-
nose-tipped, lanceolate, scariously dilated above, the margin entire or
somewhat lacerate: corolla purple: anther-tips acuminate. — In low
meadows near the city of Mexico; August, 18'J0 (n. 3145).

Perezia collina. Stout and tall, glabrous or the inflorescence
slightly puberulent : leaves thick and rigid, broadly oblanceolate, acute,
narrowed to the sessile auriculate base, irregularly toothed, 6 to 7 inches
long or less, the upper ones narrower : heads in rather close panicles
terminal on the branches, 8-flowered, the narrow acuminate scales
somewhat tomentose, not glandular-puberulent, the longer 5 lines long :
achenes {2^ lines long) glandular-puberulent and hispidulous. — In
foliage and habit very closely resembling P. rigida (which is collected
near the same locality), differing in the narrower acuminate involucral
scales, rather fewer-flowered heads, and longer hispidulous achenes.
The five nearly equal long-linear lobes of the corolla are coherent into
the two lips only at the tips, or are at length entirely distinct. Hills
near Guadalajara; December, 1888 (n. 2123).

Sttrax Jaliscana. Leaves round-ovate to oblong-obovate, acute,
at base obtuse or somewhat cuneate, vehite-tomentose and reticulately
veined beneath, green and becoming sparsely pubescent above, 2^ to
4 inches long, on petioles 2 or 3 lines long : peduncles axillary and
1-flowered, or terminal and 2-5-flowered : calyx very shortly toothed ;
corolla 6 to 8 lines long, the pubescent filaments adnate to the short
tube: fruit depressed-globose, valvately dehiscent, usually 3-seeded, 5
or 6 lines broad. — In the Sierra de San Esteban and on rocky hills
near Guadalajara; May and November, 1890 (n. 3486 and 2978).

ScHULTEsiA Mkxicana. Glabrous ; stems stout, erect, 2 or 3
inches high, sparingly branched above with short erect branches, 5-9-
flowered : leaves oblong-ovate or -lanceolate, sessile and clasping, 3 to G
lines long: flowers shortly pedicellate; calyx strongly winged, nerve-
less excepting a stout nerve at the base of each wing, 5 lines long,
not cleft to the middle, the teeth long-acuminate ; wings strongly
cross-veined; corolla yellowish, becoming purplish, 7 to 9 lines long:
filaments not appendaged ; anthers oblong, sagittate. — Damp places
on the plains near Guadalajara; October, 1889 (n. 2598). Distributed
as a new species of Microcala. .

Ehretia Mexicana. a shrub, with the young branches tubercu-
late and somewhat hispid : leaves lanceolate, short-acuminate, suh-
cnneate at base, serrate, minutely appressed-strigulose above, pubescent
and reticulately veined beneath, 1 or 2 inches long on a i)ubescent
])etiole 2 to 4 lines long r flowers small, in dense compound terminal

OF ARTS AND SCIENLE.S.

145

pubescent corymbs ; calyx deeply cleft, -I to nearly 1 line long ; co-
rolla white, neaily 2 lines long: fruit unlvnowu. — At the base of the
mountains near Lake Chapala; May, 1890 (n. 3085),

BoKRHAAViA OCTANDKA. Stems slender, dichotoraously and di-
varicately branched, glabrous or {)ul)erulent above: leaves broadly
ovate, acutish or abruptly short-acuminate, rounded at base, sparsely
pubescent and shortly ciliate, |- to 1| inches long- umbels terminal,
few - many-flowered, the flowers neaily sessile ; perianth tubular to
funnelform with a short slightly dilated limb, green with a tinge of
red, 2 lines long: stamens 8, exserted , fruit oblong, 4 lines long by
1^ broad, glabrous with a i'cw scattered tubercles.- — Much resembling
B. scandens in habit and foliage. On river-banks near Guadalajara ;
October, 1889 (n. 2958).

Ahistolochia (Gymnolobus) NANA. Stems procumbent, from
a slender subterranean branching rhizome, slender and flexuous, 3 to
6 inches long, leafy, nearly glabi'ous : leaves from renlform-cordate to
deltoid- cord ate, very obtuse or acutish and with broad rounded basal
lobes, 3 to 8 lines long, on short petioles : flowers solitary in the axils,
nearly sessile; ovary pubescent, narrow, 2 lines long; perianth dark
brown, narrowly tubular and nearly straight, \^ to 2 J inches long,
the elongated narrow blade exceeding the tube, the scarcely dilated
base of the tube closed by a glabrous diaphragm with a circular cen-
tral orifice : anthers 5 : capsule depressed-globoso, 6 lines broad. —
Collected by Prof. A. Duges of Guanajuato in 1883 at Guadalcazar
in the State of San Luis Potosi, and by Mr. Pringle iu August, 1890
(n. 3630), on dry limestone hills at San Jose Pass in the same State.

Piper (Enckea) Jaliscanum. Shrubby, 8 feet high, glabrous :
leaves oblong-ovate to round-ovate, acute or short-acuminate, abruptly
short-cuneate at base, 5-7-nerved, \\ to 3 inches long, on slender peti-
oles 3 to 7 lines long, not punctate, becoming thiekish, rather rigid and
glaucous : spikes slender, on peduncles nearly equalling the petioles, 9
to \ii lines long, becoming 24- inches long in fruit, densely flowered:
flowers 6-androus : fruit sessile, oblong, obtusely quadrangular, a line
long. — Canons near Guadalajara, in dense raoist shade, Dr. E.
Palmer, June, 1886 (n. 122), in flower, and Mr. C. G. Pringle iu
December, 1888 (n. 2153), in young fruit.

Peperomia Jaltscana. Herbaceous, the short stem from a small
tuberous root, glabrous: leaves 2 to 4, one radical, the rest cauline and
alternate, suborbicular, cordate at base, very obtuse or rounded at the
summit, thin, 2 to 5 inches broad, on petioles G to 12 lines long or
more: spikes 2 to 4, axillary and terminal, pedunculate, slender and

VOL. XXVI. (n S. XVIII.) 10

146 riiOCEEDINGS OF THE AMERICAN ACADEMY

elongated (4 iuches long or less) : flowers scattered, sunk in pits in the
fleshy rhachis : bract very minute and fleshy : stamens two, scarcely
exserted : ovary oblique-ovate, the stigma apical and sessile. — On
rich shaded banks in the barranca near Guadalajara ; September,
1889 (n. 2953).

Euphorbia (Ctttarospermum) digitata. Near E. dioscore-
oides, probably annual and 2 feet Iiigh, erect, with rather numerous
very slender ascending branches, ghibrous : raraeal leaves about equal-
ling the very slender petioles, ovate-lanceolate, acute, rounded at base,
6 to 9 lines long, gradually diminishing upward, basally peltate, entire,
ciliate ; peduncles solitary in the axils (rarely in pairs), usually ex-
ceeding the leaves, binodose and glandular-brai'teate (bracts very
rarely filiform), bearing otdy a single terminal involucre: involu-
cres turbinate-campanulate, | line long; appendages of the roundish
glands regularly 4— 6-parted into narrowly linear segments, or these
sometimes more or less united : capsule long-exserted • seeds | line
long, ovate, pitted, the pits with a central cavity and their margins
rather obscurely tuberculate. — On limestone hills near Las Pahnas,
San Luis Potosi ; October, 1890 (n. 3525). Distinguished from
E. dioscoreoides by its habit, less pubescence, solitary glandular-brac-
teate peduncles and more divided appendages. The seeds are also
smaller, scarcely more than half as large, similarly pitted, but the
margins of the pits less distinctly tuberculate.

Euphorbia (Cyttarospkrmum) subpeltata. Perennial, the
stems somewhat woody from a thickened or subtuberous root, erect,
glabrous, with numerous ascending or divaricate branches : leaves
alternate, on very slender petioles (3 to 8 lines long), semi-orbicu-
lar, short-cuneate at base and attached to the petiole slightly within
the margin, entire, glabrous, 3 to 8 lines broad : involucVes in short
slender axillary racemas, with elongated filiform bracts, carapanulate,
J line long ; lobes minute, lacerate; glands very small, subreniforra,
the purple or purplish appendages palmately divided into 3 or 4 linear-
subulate obtuse lobes as long as the involucral tube: capsule glabrous,
the subgloi)ose greenish seeds (| line long) marked with broad shal-
low pits and somewhat tuberculate. — On limestone ledges in Tama-
sopo Canon, San Lnis Potosi ; August and September, 1890 (n. 3272).
Nearly allied to E. dioscoreoides.

Euphorbia (Tithymalus) misella. Annual, erect, branching
alternately below, dichotomously above, low and slender (2 or 3 inches
high), slightly pubescent : lower leaves alternate, tht^ upper opposite,
petiolate, round-obovate, entire, 1 or 2 lines long : involucres solitary

OP ARTS AND SCIENCES.

147

in the forks, pedunculate, scarcely \ line long, the lobes fimbriate, the
glands (3 or 4) broadly stipitate, minute, rounded, entire: capsule
smooth, J line long: seed ovate, smooth or very obscurely indented,
ecarunculate. — On wet grassy borders of prairie ponds, Flor de Ma-
ria, State of Mexico; October, IH'JO (n. ooOo). 2s'ot nearly related
to any other of our species of the section.

Phvllanthus Pkinglei. a small tree (15 feet high), with
smooth gray bark on the numerous branches, and the slender herba-
ceous branchlets sulcate-angled : leaves distichous, thin, round-ovate to
orbicular or round-obovate, acutish or usually obtuse or retuse at the
summit, as also at base, 6 to 12 lines long or le^s, on petioles about a
line long ; stipules short, obtuse and scarious : pistillate flowers soli-
tary (or only 2 or 3) in the axils, on very slender pedicels 2 to 5 lines
long ; calyx G-parted, the oblong segments nearly equal ; disk cupu-
late; styles bifid, spreading: staminate flowers and fruit unknown. —
On limestone ledges at Las Palmas, San Luis Potosi ; June, 1890
(d. 3532). The material is insuflricient for a full description, but it
seems quite unlike any known species that is likely to be found in
Mexico.

Croton (Eucroton) calvescens. Shrubby, herbaceous above,
the young branches and leaves densely covered with a white or gray-
ish stellate tomentum, soon glabrate and more or less scabrous with a
rigid substellate puberulence : stipules obsolete ; leaves ovate to ovate-
lanceolate, acuminate, rounded and biglandular at base, serrulate, 2 or
3 inches long on petioles 3 to 12 lines long: racemes terminal, sessile,
3 to 9 lines long, dense, pistillate at base ; pedicels a line long: sta-
mens 9 to 12; calyx-lobes of pistillate flowers deltoid, obtuse, not be-
coming reflexed : ovary densely stellate-pubescent and hispid; styles
once divided; capsule becoming nearly glabrous, ellipsoidal: seed
smooth and shining, 2| lines long. — Collected by Dr. E. Palmer in
1886 (n. 706) near Chapala, Jalisco, and by Mr. Pringle in Novem-
ber, 1890, on hillsides near Patzcuaro in Michoacan (n. 3346). Near
forms of C. flavus.

Croton (Eutropia) el^agnoides. A shrub or small tree, 10
to 15 feet high: leaves 3-5-nerved at base, eglandular, ovate to lan-
ceolate, acutish to acuminate, green above and roughish with a slight
scurfy puberulence, white beneath with a dense compact lepidote coat-
ing (as also the inflorescence and fruit), | to 2 inches long, short-petio-
late : racemes becoming 4 to G inches long, pistillate below ; staminate
flowers nearly 3 lines broad, with narrowl}^ lanceolate acutish pubes-
cent petals a;id about 15 stamens ; pistillate flowers scattered, the

148 PROCEEDINGS OF THE AMERICAN ACADEMY

sepals oblong or oblong-obovate, acutish, 2 lines long: styles thrice
dichotomous ; capsule depressed, 2^} lines broad : seeds triangular-
ovate, minutely and irregularly jjitted. — At Las Palmas, Sau Luis
Potosi ; June, 1890 (n. 3080).

Manihot Pringlei. Apparently herbaceous, glabrous : leaves
long-petioled, o-parted to the base, the divisions 3 or 4 inches long,
narrowly lanceolate or oblong-lanceolate, very acutely short-acumi-
nate, usually obtusely lobed by a more or less broad and deep sinus
on each side, glaucous beneath ; stipules caducous, small and subulate ;
bracts of the long-pedunculate corymbose raceme foliaceous, narrowly
lanceolate and aciuiiinate, denticulate, G to 12 lines long; pedicels
erect, usually bearing a bracilet or two: perianth of the staminate
flowers glabrous, campanulate, G to 0 lines long, cleft nearly to the
middle, the lobes valvate ; stamens 10; disk large, 54obed ; pistilhite
flowers narrower, the calyx 5-parted ; disk conspicuous, entire : IVuiting
peduncles 2 inches long or more, erect; capsule glabrous, 8 lines long.
— On limestone hills at Las Canons, San Luis Potosi; July, 1890
(n. 3558). Somewhat resembling M. Cartliaginensis.

AcALYPHA DissiTiFLORA. Perennial, herbaceous, slender, a foot
high or more, dioecious, the fertile plant branching above, the stami-
nate simple above the base, pubescent : leaves thin, ovate, 3-5-nerved,
acute or shortly acuminate, rounded at base, serrate, somewhat ap-
pressed-hairy, 1 to 1^ inches long on slender petioles 3 to 6 lines long:
spikes axillary, very slender, pedunculate, 1 or 2 inches long, the stami-
nate very rarely with a pistilhite flower at base, the pistillate with
flowers much scattered ; bracts 1 -flowered, scarcely a line high, acutely
5-7-toothed, shorter than the pubescent capsule : styles short, pecti-
nately divided. — On limestone ledges in Tamasopo Caiion, San Luis
Potosi; July, 1890 (n. 3083). A strongly marked species, in ]Muel-
ler's arrangement falling near A. clUptica.

AcALYPiiA MULTISPICATA. Perennial, herbaceous, the numerous
stems simple, about a foot high, pubescent with recurved woolly hairs;
difccious: leaves subsessile, 3-5-nerved at base, the lower ovate or
obovate and obtuse or acute, the upper lanceolate and acute, serrate,
strigose-pubescent, 1 or 2 inches long: spikes pedunculate in nearly
all the axils, the staminate slender, dense, 4 to 1 indi long, the pis-
tillate short and mostly few- (1-10-) flowered; bracts 1-flowcred,
reniform, acutely 7-11-toothed, becoming 2 or 3 lines long: styles
pectinately divided; capsule pubescent. — On hillsides near Guada-
lajara; July, 1889 (n. 2903). In the same group with the last
species.

OP ARTS AND SCIENCES. 140

AcALYPHA FLAVKSCKNS. A slirub 5 to 10 feet high, the young
branches ami petioles somewhat pubescent: leaves thin, S-o-nerved,
ovate to ovate-lanceolate, rounded at base, acuminate, serrate, very mi-
nutely puncticulate and finely rough-puberulent, 3 to 4 inches long on
petioles 1 to I,]- inches, on the short fruiting branchlets smaller and
nearly sessile ; stipules rigidly setaceous from a broad base : si)ikes
sessile, 1 to 1^ inches long, the staminate axillary, dense, the pistillate
terminal and rather loose ; bracts small, 1-flowered, thin and loose,
broadly renifbrm, many-nerved, 7-tootlied, the teeth attenuate above:
styles sparingly pinuatifiil ; ovary densely pnbescent. — Jn Tamasopo
Canon, San Luis Potosi ; June, 1890 (n. 3073). Kear A. carpinifolia
as grouped by Mueller.

AcALYPHA (LiNOSTACHYs) LONGiPES. Suffrutescent, the young
herbaceous branches sparsely pubescent : leaves thin and glabrous or
slightly hispid on the nerves, 3-5-nerved at base, oblong-lanceolate to
lanceolate, acuminate, subcordate at the narrowed base, serrulate, 1 or
2 or sometimes 4 inches long, on pedicels 1 to 9 lines long; stipules
attenuate-subulate : staminate spikes axillary, sessile, 3 or 4 inches
long ; pistillate racemes pedunculate, axillary, very slender, 2 to 5
inches long, the pedicels solitary or in paiis and unequal, the longer
1 to 12 lines long; bracts minute: ovary densely muricate. — On
limestone ledges in Tamasopo Canon, San Luis Obispo; June, 1890
(d. 3082).

Sebastiania Pringlei. a glabrous shrub with slender branches,
dioecious or the sterile ameuts with sometimes (?) a pistillate flower at
base : leaves rather thin, on short slender petioles, from elliptical and
obtuse to lanceolate and acute or short-acuminate, rounded or sub-
cuneate at base, eglandular, obsoletely crenate-serrate, 9 to 18 lines
long : spikes terminal, nearly sessile ; staminate bracts" very short,
broad and abruptly apiculate, 2-flowered ; flowers nearly sessile, dian-
drous ; calyx of 1 to 3 minute distinct linear acuminate sepals : distil-
late spike 2-flowered, the upper flower usually retarded in development
or abortive; bract thicker, biglandular : capsule glabrous, 4 lines long,
chartaceous, dehiscing dorsally and ventrally ; seed not seen. — In
rocky gulches at San Jose Pass, San Luis Potosi ; July, 1890 (n. 3136,
distributed as Gymnanthes Pringhi). An evidently closely related
species, but with much larger leaves, recently collected by Dr. Palmer
near Alamos in Sonora, has nearly globose seeds with a very minute
caruncle. The presence of this caruncle has determined the reference
of the present species to the genus Sehrtstinnia, with which in other
respects the characteis accord very satisfactorily. Some specimens

150 PROCEEDINGS OF THE AMERICAN ACADEMY

that were received from Prof. A. Duges, as collected by Prof. Jose
Ramirez on the banks of the Alamos River in Sonora, closely resem-
ble Mr. Pringle's specimens excepting that the spikes are all bisexual,
the staminate bracts 4-5 flowered, and the stamens 2 or 3. The
few loose seeds which accompanied these specimens show, however,
no caruncle, though otherwise like those of Dr. Palmer. It is prob-
able that this is a third species of the same genus, and that too much
weight has been given to the presence of a caruncle as a generic
character. Tlie fruit of the Sonora plant is said by Ramiiez to be
that in which the Garpocapsa saltdans is found ; and this is certainly
true of Dr. Palmer's species. Doubtless the "jimiping beans" are
the product of more than one of these nearly allit'd shrubs. The
fruit of Sehastiania bilocularis is found to be attacked by a similar
insect, though of a diflferent genus, which has been named by Mr.
C. V. Riley Grapholitha SebastianicB. The capsules which Dr. Palmer
collected, like those of the other collections, have the cocci dehiscing
nearly to the base, and the rather thin valves become more or less
contorted. In fruit occupied by the Carpocapsa, such as I have seen,
the cocci remain closed, but the walls are chartaceous and complete
dehiscence is readily effected.

Ficus (Urostigma) .Jaliscana. Young branches, buds and peti-
oles pubescent : leaves coriaceous, round-cordate with broad more or
less overlapping basal lobes, acute, 3-5-nerved at base, 3 to 4^ inches
long, on petioles nearly as long, soon glabrate above and smooth though
minutely puncticulate, more puberulent beneath, especially on the
nerves : fruit in pairs, on stout pedicels 2 or 3 lines long, globose,
densely tomeutose, 4 or 5 lines broad, subtended by a broad somewhat
3-lobed involucre ; fertile flowers pedicellate, with unequal sepals, one
cucullate, the others shorter, broad and concave, the style rather short
and stigma subcapitate ; abortive pistillate flowers similar but smaller
and sessile, the style elongated and stigma bifid ; staminate flowers not
found : orifice of the receptacle closed within by several rows of rigid
closely imbricated broad bracts. — On cliffs near Guadalajara; Decem-
ber, 1889 (n. 2932).

Ficus (Uuostioma) Pringi-rt. Young branches and petioles
densely pubescent with spreading hairs : leaves ovate, 3-nerved and
slightly cordate at base, obtuse or acutish, 2 to 4 inches long by
1.^ to 3 broad, on stout petioles 2 to 4 linos long, very rough above
with fine hispidulous reticulations and prominently puncticulate, pu-
be.scent and strongly reticulated beneath: fruit sessile in pairs in the
axils, involucrate with two opposite orbicular silky-pubescent bracts,

OF ARTS AND SCIENCES. 151

globose, finely pubescent, 4 or 5 lines broad, the orifice somewhat um-
bonate and closed by numerous rows of imbricated rigid bracts;
flowers shortly pedicellate, the pistillate with short style and nearly
equal concave petals, the starainate with two strongly cucullate sepals
and a broad obtuse nearly sessile anther. — In the barranca near Gua-
dalajara ; December, 1889 (n. 2928).

Ficus (Phakmacosyce) Guadalajauana. Young branches
sparsely pubescent: leaves coriaceous, pinnately veined, oval, acutish
at each end, 2 to 4^ inches long by 1 to 2| broad, on pubescent petioles
3 to 8 lines long, very scabrous above, reticulately veined beneath, and
rather soft-pubescent especially on the prominent veins: fruit solitary,
on peduncles 4 lines long, globose, very shortly stipitate and with a
very narrow undulate involucre, 6 to 9 lines in diameter ; the bracts
within the orifice linear and strictly inflexed, rufous ; staminate and
gall-producing flowers on rather slender bracteate pedicels, the fertile
nearly sessile; sepals of the staminate flower 4, broadly elliptical, the
2 nearly sessile anthei's ovate-elliptical, obtuse ; sepals of the pistillate
flower linear, acuminate; bracts and sepals rufous. — In the barranca
near Guadalajara; October, 1889 (n. 2947). The galls were found
occupied by a black winged insect, — the only instance in which I have
detected the gall-fly in any of our species, though doubtless often pres-
ent. Mr. Riley informs me that he finds in this same fruit gall-
insects of three different genera.

Ficus (Pharmacosyce) radulina. A tree with rather stout
finely pubescent brancblets : leaves thin-coriaceous, oblong-lanceolate,
acute or short-acuminate, 3-5-nerved and acutish at base, very mi-
nutely rougliish-punctate above becoming smooth, glabrous beneath, 3
to 6 inches long by 1^ to 2^ broad, on petioles 8 to 16 lines long : fruit
slightly pubescent becoming glabrous, obovate-globose, 10 lines broad,
iuvoluciate with 3 short-deltoid deciduous bracts, solitary, on peduncles
2 or 3 lines long; orifice somewhat prominently margined, closed by
numerous intruded narrowly linear bracts: staminate flowers p( dicel-
late, with deeply 4-5-cleft perianth, the lobes lanceolate, acute ; sta-
mens 2 or sometimes l,the anthers elliptical, obtuse; pistillate flowers
sessile or pedicellate, 4-5-parted, the sepals narrowly linear. — Col-
lected by Dr. Edward Palmer at Hacienda San Miguel near Hatopilas
in southwestern Chihuahua, in 1885 ('' L."), and again in March,
1890 (n. 367), at Alamos in Sonora. The species much resembles
F. radula and F. anthelmintica. In the fruit examined an appar-
ently perfect flower was occasionally found, perhaps however only
pseudo-hermaphrodite, as in the few East Indian species of which

152 PROCEEDINGS OP THE AMERICAN ACADEMY

Dr. King forms his section Palceomorphe, based upon this characteristic
mark.

Ficus FASCiCULATA, Watson, Proc. Am. Acad. 24. 78. This spe-
cies was described from specimens in quite young fruit, found in
cultivation at Guavmas, but said to be native in the same region. A
very similar, if n(jt the same, species has been recently found by Mr.
Pringle (n. 3554) in Tamaso[)o Canon in the mountains of San Luis
Potosi, and this appears to be identical with what was collected by
Ervendberg (n. 332) near Tantoyuca in Huasteca, and by Botteri still
farther southward in Orizaba. The leaves, however, vary considera-
bly in size, and from obtuse to quite sharply acuminate, and the only
specimen seen by Mr. Pringle was a small erect shrub very different
in habit from those at Guaymas as described bj^ Dr. Palmer. The
orifice of the small thin fruit is in the Guaymas specimens much im-
pressed, while in the others it is conspicuously prominent, which may
be due to the stage of growth. All may possibly be referrible to
F. sapida, Miquel, of Costa Rica and Panama, as forms of one poly-
morphous species.

PiLEA GLABRA. Low and hcrbaccous, glabrous : leaves thin,
showing on the upper side numerous linear cystoliths, entire, lanceo-
late or broadest near the middle and narrowed each way, acutely
acuminate, rounded at the very base, 3-nerved, the nerves continued
to the apex, 2 to 4 inches long by 8 to 16 lines broad, on petioles an
inch long: panicles pedunculate, very loose and slender, exceeding
the petioles, solitary or in pairs in the axils, androgynous. — In
Tamasopo Canon, San Luis Potosi; August, 1890 (n. 3550).

Myriocarpa brachystachys. Young branches, petioles and
lower surface of the leaves densely tomentose : leaves ovate, rounded
at base, short-acuminate, acutely serrate, finely bullate, ^nearly gla-
brous above, 3 or 4 inches long, on petioles ^ inch long : pistillate
inflorescence sparingly branched, nearly sessile, the longer spikes 3
to 5 inches long, very densely flowered : sepals lanceolate, sparingly
<:iliate, a third as long as the sparsely hispid ovary. — In the bar-
ranca near Guadalajara; IMay, 1888 (n. 3024).

JuGLANS Mexicana. Foliage as in ./. nigra, but with the pubes-
cence nearly of,/, cinerea: fruit large, subcompressed-globose, 2^
inches high by 2 inches broad ; nut H inches broad, very obtusely
rugose, obtuse or slightly apiculate. — On hills at San Josi^- Pass, San
Luis Potosi; Octol)er, 1890 (n. 3322).

Michostylis (Dienia) tenuis. Stem slender fiom a small tu-
berous base, 4 to 6 inches high, with a single narrowly ovate acutish

OF ARTS AND SCIliNCES.

153

basal sheathiug leaf 1^ inches long : flowers greenish ochroleucous, in
an open raceme 2 inches long ; pedicels very slender, 1 to 3 lines
long ; bracts very snuUl : sepals and petals linear-lanceolate, acumi-
nate, 2 lines long, the lip a little shorter, attenuate from a broader
base. — In low meadows, Flor de Maria, State of Mexico; July,
1890 (n. 3186).

Sfiranthks Pringlei. Root of fascicled fusiform tubers an
inch long ; stem slender, 4 to 8 inches high, puberulent, the scattered
sheathing bracts (5 or G) tliiu. acute or acuminate, ^ to 1 inch long:
spike loosely few-flowered, 1 or 2 inches long, the narrowly lanceolate
bracts equalling the ovaries : flowers white, the lanceolate se[)als
3 lines long; lip a little longer, dilated above into a reniform un-
dulately margined blade; column short, its crest short and obtuse;
beak of the anther oblong, acutish : capsule oblong-ovate, 3 lines long.
— Moist plains near Guadalajara; June, 1889 (n. 2877). Radical
leaves unknown.

Spikanthes (Stenorhynchus) Jaliscana. Radical leaves un-
known ; stem from a fascicle of long tuberous roots, leafless, a foot
high, glandular-pubescent above, partially covered with acute or short-
acuminate bracts an inch long: flowers in a rather loose slender spike,
subtended by linear-lanceolate acuminate bracts 4 to 6 lines long;
sepals and petals red, 9 lines long, narrow above the prominent gib-
bosity, acuminate, the lip much narrowed above the dilated and
auricled basal portion: capsule ^ inch long. — Plains near Guadala-
jara; June, 1889 (n. 2874). Related to S. speciosa.

Bletia Palmeri. Stem from a tuberous-thickened base upon a
slender rootstock, slender, 1 to 1^ feet high, G-12-flowered : leaves
shorter than the stem, broadly linear, 6 to 10 inches long by 3 to 6 lines
broad ; bracts srnall : perianth purplish, G to 9 lines long; sepals and
petals nearly equal, oblong, acutish ; lip 6 lines long and nearly as
broad, with broad rounded lateral lobes, a rhomboidal middle lobe, and
seven very prominent contiguous laminae extending from the base to
the apex: capsule an inch long, on a pedicel 3 or 4 lines long. —
Collected at Rio Blanco, Jalisco, in August, I88fi, by Dr. E. Palmer
(n. 336), and in the barranca near Guadalajara in May, 1889, by
Mr. Priugle (n. 3023).

GovENiA elliptica. Basal sheaths very broadly dilated, the
lono-er 6 inches long ; leaves lanceolate above the enclosed petiole,
acuminate, 8 inches long by 3 broad, nearly equalling the loose but
many-flowered spike ; floral bracts lanceolate, acuminate, shorter than
the slender ovaries: sepals and petals brown, 6 lines long, bilabiately

154 PROCEEDINGS OF THE AMERICAN ACADEMY

divergent, the lower sepals falcate and rather narrow, the upper ob-
long ; lip yellow, elliptical with cuneate base, obtuse or emargiiiate,
3 lines long. — Cool rich cailons in the mountains near Monterey,
San Luis Potosi; June, 1890 (n. 2797).

Arethusa granuiflora. Flowering stem leafless from a tuber-
ous base (6 to 8 lines in diameter), 6 to 8 inches high, with 2 to 4
very short closely sheathing bracts ; foliar stem contiguous, sheathed
below and bearing two long- acuminate narrow leaves (G to 12 lines
broad) exceeding the scape : flower solitary, large, the unguiculate
sepals oblong-lanceolate, acuminate, falcate, 15 lines long; lip 2
inches long or more and \^ broad, purple, 3-lobed, erosely denticu-
late ; column shorter than the sepals : ovary slender, 8 lines long. —
Banks of cailons near Guadalajara; October, 1889 (n. 2997). The
condition of the flowers prevented an examination of the androecium,
but there seems no reason to doubt the correctness of the generic
reference.

PoGONiA (Triphora.) Mexicana. Stem 2 to 4 inches high from
a small tuber, sheathed at base and bearing 4 or 5 sessile leaves, the
lower round-ovate, acute, about 6 lines long, the uppermost lanceo-
late: flowers 2 or 3, pedicellate, soon recurved; perianth 5 lines long,
the greenish sepals and white petals linear, acute ; lip purplish above,
with three green median nerves, cuneate to a short broad claw,
3-lobed, the middle lobe subdeltoid, undulate-margined : column 3^
lines long. — In Tamasopo Caiion, San Luis Potosi; August, 1890
(n. 3557). Resembling P. pendula, and probably the same as Parry
& Palmer's plant from the same region, mentioned by Mr. Hemsley
as in Herb. Kew.

Habenaria filipeua. Low, from a small oblong-ovate tuber,
glabrous, 4 to 8 inches high: leaves erect, longer than the internodes,
oblong-ovate or the lowest ovate, acute, mostly sheathing at base, 1 to
1^ inches long; floral bracts foliaceous, acuminate, nearly equalling
the flowers : spike loosely rather few-flowered, 2 or 3 inches long :
sepals acute, the lower oblong-ovate and subfalcate, the upper sub-
orbicular, 2 lines long ; petals bifid, the upper lobe oblong, falcate,
nearly oqiialliiig the upper sepal, the lower filiform, 3 lines long or
more ; lip o-lobed, the lobes all filiform, the lateral 4 lines long, ex-
ceeding the somewhat broader middle one ; spur 7 or 8 lines long,
acute, a little enlarged toward the end. — On moist grassy slopes,
Flor de Maria, State of Mexico; July, 1890 (n. 3187). Tliis was
also previously collected by Mr. Pringle in Chihuahua in 1887 (n.
1375''), distributed as //. Gnadalnjarana, var (?). It differs from that

OF ARTS AND SCIENCES. 155

species in the fewer flowers, the longer lower lobes of the petals, the
longer and narrower lobes of the lip, and the longer acuminate spur.

Hechtia pedickllata. Leaves long-attenuate from the base,
about 2 feet long and 1^ inches broad at base, white-scurfy beneath,
less so above, spinose on the margin : flowering stem flexuous, 2 or 3
feet high, covered by numerous thin lanceolate attenuate bracts serru-
late on the margin : panicle glabrous, long and narrow, the numerous
spreading branches about 2 inches long, mostly simple : flowers nu-
merous, scattered, racemose; sepals and petals small, deltoid, persist-
ent, the latter 1|- lines long: capsules oblong-ovate, 4 or 5 lines long
on pedicels 2 or 3 lines long. — On ledges in the barranca near Guada-
lajara ; October, 1889 (n. 2970).

TiLLANDSiA (Anoplophytcm) Pringlei. Basal leaves (15 to 20
or more) abruptly convolute-linear from a dilated base (1^ inches long
by an inch broad), densely lepidote with appressed centrally punctate
scales, 6 to 8 inches long, recurved, the cauline shorter and soon re-
duced to sheathing oblong acuminate or acute lepidote bracts 12 to 6
lines long : flowering stem nearly 2 feet high, very sparingly branched ;
spikes 4 or 5 inches long, 6-8-flowered, glabrous, slightly flexuous ;
calyx little exceeding the acutish narrow appressed bract and about
equalling the internodes, 7 to 9 lines long ; petals long-exserted. — At
Las Palmas, San Luis Potosi ; June, 1890 (n. 3530).

TiLLANDSiA (Platystachys) cylindkica. Basal leaves unknown ;
peduncle stout, a foot long or more, covered with numerous imbricated
erect scurfy-canescent leaves with dilated base 1^ inches broad and at-
tenuate upward into an elongated convolute-linear termination a foot
long : inflorescence scurfy-canescent, cylindrical, 8 to 10 inches long
and over 2 inches broad, of numerous (20 to 30) distichous sessile
ovate-lanceolate 8-10-flowered spikes which are 2 or 2^ inches long
and an inch broad ; bracts tinged with red on the margins, those on the
rhachis more or less attenuate and the lower often much exceeding the
spike, those of the spike acute, carinate, 1 to 1|- inches long : calyx more
or less scurfy, an inch long; petals greenish, 1| inches long, convolute
below into a tube as long as the calyx ; stamens and style exceeding
the petals. — From Guanajuato, Mexico ; Prof. Alfred Duges. Near
T. Dugesii.

SiSYRiNCHiUM PLATYPHYLLUM. Perennial, glabrous ; stem stout,
nearly 2 feet high, branching above, broadly winged : leaves ensiform,
acute, 6 or 8 inches long by 6 to 9 lines broad, smooth on the margin,
the uppermost lanceolate; floral bracts equal, broad, acute, 1| inches
long ; pedicels slightly exserted : ovary small, subpuberulent : peri-

156 PROCEEDINGS OP THE AMERICAN ACADEMY

anth yellow, the segments oblong, obtuse, 9 lines long by 3 broad,
faintly nerved except toward the base: free portion of filament nearly
2 lines long ; style cleft nearly to the base. — In the barranca near
Guadalajara; July, 1889 (n. 287G). Resembling S. Arizomcum, but
more glabrous, the broader leaves less acuminate, and the style more
deeply cleft; fruit unknown.

Agave (Litt^a?) Hartmani. Acaulescent; leaves very numer-
ous in a dense rosette (becoming a foot in diameter), broadly linear
above the short dilated base, the blade 3 inches long by about |- inch
broad, very thick, convex on the back, marked on both sides by broad
intersecting gray lines (as in A. ^lif era), ending abruptly in a stout
brown spine 2 or 3 lines long, the margins at base acute and finelj'
serrulate, obtuse above and bearing on each side about half a dozen
very slender recurved ash-gray threads : flowers and fruit unknown. —
Collected by Mr. C. V. Hartraan, botanist of the Leraholtz expedition,
and in cultivation at the Cambridge Botanic Garden.

Agave (Manfreda) brunnea. Acaulescent: leaves rather few,
fleshy, recurved, channelled above, 4 inches long or less by 6 to 10 lines
broad in the middle, the margin armed with scattered broad hooked
spines a line long: flowering stem 2 feet high, with lanceolate narrowly
attenuate bracts shorter than the nodes : flowers few (about 6), ses-
sile ; ovary 6 lines long; perianth 15 to 18 lines long, the narrow tube
nearly twice longer than the dark brown narrowly oblong spreading
lobes : stamens and style much exserted : capsule broadly ellipsoidal,
about 9 lines long. — On the battle-field of Buena Vista, Tamaulipas ;
July, 1888 (n. 2218).

EciiEANDiA NODOSA. Roots coarsely fibrous ; the base of the plant
surrounded by the more or less fibrous remains of numerous dead
leaves : leaves broadly linear, 18 inches long by 12 to 1 ") lines broad,
attenuate to each end, very shortly ciliolate : stem 3 feet high, panicu-
lately branched, the branches spreading and very slender with numer-
ous nodes (mostly approximate, 3 to 6 lines apart, sometimes more
distant) ; pedicels very slender, 3 to 6 lines long, jointed a little below
the middle: perianth pale yellow, 5 lines long: capsule broadly ellip-
tical, 2 or 3 lines long. — In the barranca near Guadalajara; Nov.,
1888 (n. 2151). This genus must include several species, though it is
difficult to identify the forms that have been described as distinct, and
which have of late all been embraced under 7i'. ternijlora. A form
more nearly resembling typical E. ternl flora is Mr. Pringle's n. 3183
of 1890. This has long and more or less tuberous-thickened "oots,
and linear-lanceolate leaves attenuate from near the base to (^ nar-

OF ARTS AND SCIENCES.

167

rowly acuminate apex (4 to 8 inches long by 9 to 18 lines broad),
more distinctly ciliate : branches less slender, ascending, with distant
nodes ; pedicels jointed toward the base ; perianth yellow becoming
pale, o or 6 lines long, and capsule 3 or 4 lines long. Both of these
species must be distinct from the common form with very elongated
and narrow leaves.

Dasylirion inkrme. Tree-like, 20 to 30 feet high and 2 to 4 feet
thick at base, with few branches : leaves 2 to 4 feet long by 6 lines
broad, long-attenuate, thin and scarcely at all carinate, the margin
unarmed, very minutely serrulate : inflorescence paniculate, a foot
broad or more: fruit ti'iquetrous, broadly winged, 5 or 6 lines long. —
On limestone ledges at Las Palmas, Saa Luis Potosi ; June, 1890
(n. 3108). A very remarkable species.

Tradescantia Pringlei. Stems slender, decumbent and rooting
at base, glabrous or with a pubescent line on one side : leaves ovate,
short-acuminate or only acute, 1 to 1| inches long, minutely puberu-
lent, the short petiole and loose sheath villous-ciliate : peduncles axil-
lary and terminal, a little shorter than the leaf, bearing a small head
of nearly sessile flowers : bractlets and sepals glabrous or a little
ciliate, the latter 2.', lines long; petals purple: stamens 6, unequal,
the longer with filiform filaments bearded in the middle (or one
naked) and broadly oblong anthers, the shorter filaments very densely
bearded in the middle with green gland-tipped hairs, the anthers
orbicular ; anther-cells contiguous. — In the Sierra Madre near Mon-
terey; July, 1888 (n. 2226).

Cham^dorea Pringlei. Acaulescent or nearly so : leaves erect,
pinnate, nearly 3 feet high ; leaflets 12 to 15 on each side of the triangu-
lar rhachis, linear-lanceolate, acuminate, 6 or 8 inches long by 3 or 4
lines broad: inflorescence dioecious, the staminate spadix arising from
the base and about equalling the leaves, simjJe ; peduncle covered
with sheathinw bracts; flowering rhachis about 8 inches long: flowers
scattered ; calyx 3-parted, the segments orbicular ; corolla 3-parted,
valvate ; abortive ovary columnar, greenish : pistillate flowers and
fruit unknown. — In Tamasopo Canon, San Luis Potosi; June, 1890
(n. 3527). The most northern locdity known for any species of the
genus. While the absence of pistillate flowers renders the section
to which the species belongs uncertain, it nevertheless appears to be
clearly distinct from the few other known species that have a simple
spadix.

Eriocaulon Jaliscanum. Dwarf: leaves subulate, acuminate, 5
to 8 lines long by nearly a line broad, little exceeding the bifid

1*)8 PROCEEDINGS OP THE AMERICAN ACADEMY

sheaths : scapes very slender, ^ to 3 inches high ; heads globose to
oblong-ovate, ^ to H lines high ; involucral bracts scarious, white or
nearly so, oblong to obovate, obtuse, the inner white becoming some-
what fuliginous, acutish : flowers trimerous, mostly pistillate, a few at
the apex staminate ; petals linear-spatulate, slightly ciliate above :
seeds very minutely papillose. — Wet places near Guanajuato ; No-
vember, 1889 (n. 2936).

3. Upon a wild Species of Zea from 3Iexico.

Prof. "W. H. Brewer, in a communication to Dr. Sturtevant, to be
found in the paper of the latter upon '' Indian Corn " in the Report of
the New York State Agricultural Society for 1878, gives a statement
which Roezl, the well known German collector, made to him in 1869
to this effect: that "he found in the State of Guerero a Zea which
he thinks specificially distinct, and he thinks undescribed ; the ears
very small, in two rows truly distichous ; the ear (but not each grain
separately) covered with a husk, the grain precisely like some varieties
of maize, only smaller and harder." Specimens of a Zea which is
in all probability the same that Roezl referred to were received by me
in 1888 from Prof. A. Dug^s of Guanajuato under the designation of
Mais de Coyote. It was reported to him as growing wild at Moro
Leon, to the south of the State of Guanajuato, and as not at all re-
sembling ordinary varieties of maize. The specimens sent were two
very slender stalks about four feet high, with a small terminal stami-
nate inflorescence but no trace of fertile spikes. These were probably
very depauperate stalks, that had been selected for easy carriage.
Accompanying them was a united cluster of about half a dozen small
ears enveloped in their husks, each about two inches long and bearing
a few rows of small white pointed kernels.

Some of the peculiarities of this remarkable corn were noted at
the time, but nothing more was done until last year, when an attempt
was made to grow it at the Botanic Garden, Cambridge, with quite
unexpected results. The corn was planted early under glass, and as
soon as danger from frosts was over the plants were transferred to a
warm sunny location, where they .soon began to grow vigorously and
to send out numerous ofTshoots from the base. These "suckers" grew
as rapidl}' as the main stalk, so that the plants, wiiich had fortunately
been placed some feet apart, had the appearance of two " hills," one
of the two having nine and the other twelve stalks ascending from a
common base. The tallest were over ten feet in height, with a diame-
ter of nearly two inche.'', and they would have become yet taller had

OF ARTS AND SCIENCES. 159

the season permitted. Their foliage and pubescence were in every
way as in ordinary field corn, the staminate tassels M'ith conspicuously
longer and more drooping racemes, the habit of growth wholly unusual.
In our ordinary form the erect culm is always apparently single, bear-
ing solitary axillary ears which are terminal upon a usually short leafy-
bracted peduncle. This peduncle is in fact a lateral branch, bearing a
terminal pistillate spike corresponding to the staminate panicle on the
main stem. In this Mexican corn, on the other hand, the better de-
veloped stalks were evidently branched from several of the axils, the
branches often becoming three or four feet in length, very leafy, and
having at least a rudimentary ear in the axil of every leaf. Several,
sometimes half a dozen, perfect ears were formed upon each branch.
The terminal ear was always androgynous, staminate at the summit.
On the shorter stems the branches were reduced to a more or less
crowded axillary cluster of §ars similar to the one received from Prof.
Duges. The last year's season was a long one, and there was no heavy
frost in Cambridge until near the end of October. The corn however
was at that time still very green, and the stalks were finally cut and
stored under shelter in the hope that the ears would ripen in the
stack; but upon very few did any of the kernels mature.

The natural supposition was that we had here at last the original
wild state of our cultivated maize. A careful comparison of the two,
as thorough as the material at hand of the cultivated forms w^ould per-
mit, has led me first to doubt the probability of this, and now to con-
sider the form in question a distinct species. The differences upon
which this conclusion is based are in the habit of growth, the arrange-
ment of the staminate spikelets and the nervation of their glumes, the
form of the glumes of the pistillate flowers, and the ready disarticula-
tion of the rijjened ear.

It appears from descriptions, figures, and such specimens as I have
seen of cultivated maize, that the staminate spikelets are in pairs at
the joints of the rhachis, and their empty glumes 7-9-nerved. In the
Mexican plant there are usually three and sometimes four together,
one of them short-pedicelled, the others more nearly sessile. The
empty glumes are 3-5-nerved and bicarinate, the flowering ones more
narrow than in Z. Mays. The pistillate spikelets are in pairs at the
joints of the rhachis, the intei nodes of which are more or less strongly
margined and cupulate, and finally become hard and shining. The
glumes are very broad, strongly concave and enfolding each other,
much more so than in the flowers of Z. Mays that I have examined,
and than they are represented in the figures of Nees and Doell. The

160 PEOCEEDINGS OF THE AMERICAN ACADEMY

lower glume becomes very hard and rigid, excepting its margin, and
fiimly embraces the lower part of the kernel.

The ears upon the plants raised in the Botanic Garden were very
variable, from scarcely two inches to four inches in length and three
fourths of an inch broad, tapering slightly to an acuti^h apex, and
with the kernels in four, eight, ten, and sometimes twelve, but most
frequently in ten rows. A con)parison of these shows clearly the struc-
ture of the ear. When there are only four rows, the ear is flattened
and distichous, and the opposite pairs of rows are evidently the result
of the pairs of spikelets regularly alternating upon the opposite sides
of an extremely short-jointed and very flexuous rhachis. In the eight-
rowed ear the rhachis is four-sided instead of two-sided, and in the
ten-rowed it becomes five-sided. This latter case corresponds to the
arrangement in the terminal raceme of the staminate inflorescence,
where the spikelets are usually in five rairks. In the eight-rowed ear
each joint bears two opposite pairs of spikelets, alternating with those
of the joints immediately above and below, and in the twelve-rowed
there are three pairs to each joint, alternating in the same way. The
kernels are somewhat imbricated in the rows, and usually alternat-
ing, owing to one of the spikelets in each pair being slightly pedicel-
late. They are small, ovate, somewhat flattened dorsally and pointed,
the lower part constricted by the closely embracing glume. In struc-
ture they are hard and corneous, with a central starchy layer extend-
ing from the base nearly to the apex. The ripened ear breaks readily
at any point, so that the eight-rowed ear, for example, may be separated
into its several joints, each bearing two opposite pairs of kernels. I
would therefore characterize the new species as follows: —

Z. CANINA. Culms several from the same root, ascending, branched :
staminate racemes often elongated and drooping ; spikelets 2 to 4 (usu-
ally 3) at each node, one or more short-pedicelled ; empty glumes 3-
5-nerved, bicarinate : pistillate spikes sessile in the axils and terminal,
the terminal staminate at the apex; ears small, 4-12-rowed, dividing
more or less readily at the joints ; kernels small (3 to 4 lines long),
white, hard and smooth, ovate, acutish, constricted at base.

The location from which the specimens were obt.iincd for Prof.
Duges is stated more definitely to be IMoro Leon (otherwise Congre-
gacion), near Uriangato, about four Mexican leagues nortli of T^ake
Cuitzco. It is therefore near the boundary line betweiMi the States
of (juanajuato and INIichoacan. The natives of the district are said
by I'rof. Duges to believe the mn'/'s (J" cnyofe to be the source of the
cultivated varieties of maize, notwithstanding the recognized dill'er-

OF ARTS AND SCIENCES.

161

ences between them. The kinds of corn in cultivation about Gua-
najuato are described by Prof. Duges as the ma'is arribeno, with hard
sind shining kernels ; t7iais commun, softer and less shining, either yel-
low, black, or red ; ma'is molonco, resembling the last, but with smaller
kernels, growing in the " bad lands," and the difference probably due
to the poverty of the soil ; and a popcorn, mais rosero. All of these
are affirmed to be very unlike the ma'is de coyote, which appears to
be known only in a wild state.

As shown by Dr. Sturtevant in the lleport of the N. Y. Agricul-
tural Experiment Station for 188G, the varieties of corn cultivated
generally by the Indians of Mexico all come into the group desig-
nated by him as "soft corns" ("Zea* amyfacea "), in which the broad
flattened kernels are composed wholly of starchy matter in addition
to the embryo. A peculiar kind was found by Dr. Palmer at the
Indian village of San Padro, Guadalajara, of which Dr. Sturtevant
forms the distinct group of " starchy sweet corns " (" Zea^ amylea-
saccharata "), inteimediate between the sweet and soft corns, having
a wrinkled exterior and the summit of the kernel corneous. It is
evident that so far as the grain is concerned these have little resem-
blance to the ma'is de coyote.

4. Notes ujyon a Collection of Plants from the Island of
Ascension.

During the visit of the U. S. Eclipse Expedition of 1S89 to the
island of Ascension, Mr. E. J. Loomis of the Nautical Almanac Bu-
reau, Washington, made a collection of such plants as chanced to at-
tract his attention. Though very small, it adds some species to the
previously known flora of the locality, a full account of which is given
by Mr. W. B. Hemsley in the botanical report of the voyage of the
" Challenger." The vegetation of the island, as there stated, is re-
stricted almost wholly to " Green Mountain," an elevation of nearly
3,000 feet altitude, which condenses the moisture of the southeast
trades, and is consequently subject to frequent rains and fogs. At
about three hundred feet below the summit the peak is encircled by
a nearly level road, known as " Elliott's Pass," which generally fol-
lows the very irregular contour of the mountain, but is occasionally
carried through a spur by means of a tunnel. When these tunnels
are short, they are lined throughout, roof and sides, with the most
delicate and beautiful ferns, as are also the longer tunnels so far as
daylight extends. It was on this part of the mountain that much
of the collection was made.

VOL. XXVI. (n. S. XVIII.) 11

162 PROCEEDINGS OP THE AMERICAN ACADEMY

The entire plia^nogamous flora native to the island, as given by Mr.
Hemsley, includes but twelve species, and ten of these are marked
by him as only doubtfully indigenous. The two other species, Hed-
yotis Adscensionis and Euphorbia origanoides, are endemic. The
cryptogamous flora is somewhat more extensive, the same authority
giving a list of fourteen ferns (including a Lycopodium), thirty-five
mosses and hepatics, and a dozen lichens.

The following is a list of the species of every kind collected by
Mr. Loomis, of which those in italics are without doubt introduced
plants. The species of ferns described as new have been submitted
to Prof. Eaton and to Mr. J. G. Baker. The mosses were determined
by Prof. C. R. Barnes, the single hepatic by Prof. L. M. Underwood,
and the lichens by Prof. Farlow. It is to be regretted that the oppor-
tunity could not have been improved to make a larger and more
complete collection of the plants of the island, for it is probable that
a thorough search would increase considerably the number of indige-
nous species now known.

Argemone Mexicana, L. Abundant.

Senebiera didyma, Pers.

Oxalis corniculata, L.

RuBUS NANUS. Stems very short (1 or 2 inches high), from a
woody caudex, erect, armed with numerous short recurved prickles,
about 3-flowered at the summit: stipules narrow, acuminate, entire or
sparsely toothed ; petioles short, pubescent and prickly ; leaves mostly
simple and rounded, snbcordate at base, unequally and subincisely ser-
rate, somewhat .'3-lobed or sometimes ternate, sparsely villous above,
prickly on the veins beneath: calyx small (about 4 lines broad), the
base prickly : carpels rather numerous, on a depressed receptacle, gla-
brous ; styles short. — Described from a single specimen found near
the entrance of a tunnel in Elliott's Pass. It appears to be a very
distinct and peculiar species of a genus not otherwise represented in
the flora of the islands of the South Atlantic.

liidcns pilosa, L.

Lactuca Scan'oh, L. (?) Foliage only.

Euphorbia origanoidks, L.

Commellna nudijlora, L.

CvpiiRus UMBELLATUS, Beuth. Foliagc only.

Lycopodium cernuum, L.

Pteuis incisa, Thunb.

Pteris flabellata, Thunb., var. Ascensionis, Hook. & Bak.

OF ARTS AND SCIENCES. 1G3

AsPLENiUM AsCENSlONiS. Stipes slender, tuftetl, an inch Ion" or
less, naked, dull brownish green ; frond 3 to G inches long, often root-
ing at the apex, 6 to 10 lines broad, the rhachis channelled above
and narrowly winged ; pinna? about 20 (15 to 25) on each side,
very irregular in shape (nearly as in A. fragile), in the middle of the
frond somewhat quadrilateral and nearly twice longer than broad,
often more or less deeply lobed at base on the upper side, obtuse,
irregularly and obtusely few-tooihed or the lower margin entire, the
lowermost usually round-deltoid, nearly equilateral and often lobed
both sides, the uppermost becoming cuneate-obovate : sori linear-ob-
long, oblique, 1 to 6 on each side. — This adds another species to the
closely allied -4. vmc/e group, intermediate between A. fragile and
A. vagans. It is probably the same as the A. dentatum reported by
Bory as collected on the island by D'Urville. Mr. Baker informs me
that he also finds a specimen in Herb. Kew, among their A. fragile,
which was given by " Don " (probably George, the younger, and col-
lected by him on his trip to Sierra Leone) to Lindley in 1831.

Nephrodium molle, Desv. Specimens very variable and mostly
much reduced.

Nephrodium (?) viscidum. More or less densely glandular-fibril-
lose throughout ; stipes from a slender creeping rhizome, brown, about
an inch long, bearing a lanceolate acuminate pinnate frond 3 inches
long ; pinnae (except the upper ones) cleft nearly or quite to the mid-
vein, the thin broadly oblong segments (1 to 2 lines long) obtuse, more
or less crenate-serrate, glandular-villous beneath. — A single speci-
men, not in fruit, and the genus consequently indeterminate. Though
strongly marked by its glandular-fibrillose character, it is not recog-
nized by either Mr. Baker or Prof. Eaton. The former thinks it
a starved Nephrodium, but "it cannot be N. Ascensionis" the only
species excepting the last that is known from the island.

Gyjinogramme Ascensionis, Hook.

Camptlopcs introflexus, Brid.

Ehacopilum gracile, Mitt.

Dicranella ?

Lejeunia pterota, Taylor.

Theloschistes chrtsophthalma, Norm., var. flavicans, Wallr.

Parhelia saxatilis, Ach. ?

164 PROCEEDINGS OF THE AMERICAN ACADEMY

XIL

contribution from the gray herbarium of
'harvard university.

DESCRIPTIONS OF NEW PLANTS, CHIEFLY GAMOPET-

ALJS, COLLECTED IN MEXICO BY C. G. PRINGLE

IN 1889 AND 1890.*

By B. L. Robinson.

Presented by Sereno Watson, April 8, 1891.

Xylosma Pringlei. Monoecious, glabrous, armed at the nodes
with simple acicular spines 3 to 5 lines in length : leaves small, fasci-
cled in the axils of the spines, coriaceous, elliptical, obtuse, attenuate
to a subsessile base, serrulate, veiny, shining above, pale beneath, 9 to
16 lines long by half as broad ; the edges more or less revolute and the
teeth incurved : flowers fasciculately grouped, 3 to G together, with no
common peduncle ; pedicels slender, 3 or 4 lines long : sepals 4, ovate,
acutish, scarcely exceeding half a line in length, glabrous or puberulent
on tlie outer surface, ciliate, minutely pubescent within : stamens 12 to
18, twice as long as the sepals: style 2-3-cleft, and the ovate ovary in-
completely 2-3-celled ; fruit not seen. — In the Sierra Madre near
Monterey, August, 1889 (n, 2784). This is perhaps the same as Dr.
Palmer's n. 10G2 (collected between San Luis Potosi and Tampico),
which differs principally in its ovate leaves and verrucose stem.

Desmodium Jaliscanum, AVatson, var. (?) ohtusum. Stem 5 to
10 feet high : leaves elliptic, oblong, or even slightly obovate, very
obtuse, apiculate : calyx-teeth ovate, obtusish, the upper one refuse ;
pods appressed-pubescent, very numerous in dense simple or branched
racemes. — Rocky slopes, Tamasopo Canon, San Luis Potosi ; Octo-
ber, 1890 (n. 3290).

PiMPiNELLA IMexicana. Glabrous : root more or less thickened ;
stem 4 feet high, with scanty foliage, panioulately branched above :
the radical and lower cauline leaves very long-petioled, ternate ; leaf-

* Of tlie plants here described the last three only were sent from California
and Wasliington by other collectors.

OP ARTS AND SCIENCES. 165

lets deeply 2-3-parted, the segments ovate or lanceolate, acuminate,
more or less narrowed at base, 1 to 1^ inches long, 10 lines or less in
breadth, green above, pale beneath, serrate, the margins of the teeth
being slightly thickened and cartilaginous ; upper leaves scattered in
the diifusely branched inflorescence, pinnately or ternately parted with
long linear toothed or serrate segments ; the highest leaves reduced
to Hliform bracts ; involucres and involucels none : umbels numerous,
with few (2 to 6) often unequal rays 6 to 9 lines in length ; umbellets
about 12-flowered, only 2 or 3 flowers in each being fertile: corolla
white (?) or in the sterile flowers not infrequently purple : fruit a line
long, the ribs not very prominent ; stylopodiura depressed, the margin
crenute ; oil-tubes numerous ; inner surface of the seed plane or very
shallowly concave. — Hills of Patzcuaro, Michoacan ; November,
1890 (n. 3331).

EuPATORiUM EspiNOSARUM, Gray, var. subintegrifolium.
Leaves ovate, acuminate, subentire, very glutinous on both sides :
scales of the involucre a little longer and more acute than in the
typical form, being in these respects more as in var. amhigimm, Gray.
— Shaded ledges of lime-rock, San Jose Pass, San Luis Potosi ; Oc-
tober, 1890 (n. 3311).

Gymnolomia decumbens. Stems several from a ligneous base,
decumbent, simple or branched from near the root, smoothish below,
roughened above with very short appressed hairs : lower leaves oppo-
site, elliptical, acute, narrowed to a sessile base, subentire, thickish,
rough-pubescent, green on both sides, \^ inches long by half as broad ;
the upper similar but narrower, alternate, sparse : heads terminal, soli-
tary or 2 or 3 together; involucre ^ inch in diameter; outer scales
ovate, acuminate, very rough, the inner larger, smoother, obtusish ;
chaff oblanceolate, acuminate; rays 12 to 15, narrow, over half an
inch in length, yellow ; achenes (immature) smooth, angled, and with
no trace of teeth or awns. — Rocky hills, Tultenango, State of Mex-
ico; September, 1890 (n. 3263). This species resembles G. rudis,
Gray, but differs in its decumbent habit, smaller and more entire
leaves, smaller heads, ovate not oblong involucral scales, and longer
narrower rays. From G. multijiora, Nutt., it differs in the ligneous
clearly perennial base, as well as in the involucre, etc.

Otopappus ALTERNiFOLius. Stem 3 to 6 feet high, striate, smooth-
ish or slightly tomentulose : leaves alternate, ovate-lanceolate, acumi-
nate at both ends, subsessile, with a short roughish pubescence above,
white-tomentose beneath: heads in an open corymb, about 15, half an
inch or more in diameter ; scales of the involucre in several rows, nar-

166 PROCEEDINGS OP THE AMERICAN ACADEMY

rowly oblong or lanceolate, some rounded at the apex, others acute .
rays 12 to 25, little exceeding the disk; pappus of two very unequal
awns, the inner more than half as long as the achene, the outer much
shorter ; achenes smooth, with a single rib on each face, edges acute,
the outer usually wingless, the inner with a narrow wing, which ex-
tends up upon the inner awn and there broadens into an auricular
appendage. — Limestone liills, San Jose Pass, San Luis Potosi ; Oc-
tober, 1890 (n. 3310). From its peculiar pappus with wing-appen-
dages this plant must be referred to the genus Otopappus as extended
by Herasley (Biol. Cent. Am. Bot. 2. 191), and placed near 0. epale-
aceus, Hemsl. The whole habit of the plant with its alternate leaves
is that of a Verbesina, and it seems thus to connect the two genera,
especially since there are several species of Verbesina, as V. Hum-
boldtii, Spreng., and V. perymenoides, Schultz Bip., in which the
achenes are more strongly winged on one side than on the other.

Senecio Guadalajarensis. Herbaceous; stem erect, leafy, sul-
cate-striate, smooth : leaves elongated lanceolate, 6 to 8 inches long,
9 to 15 lines broad, acute, narrowed at base to a very short naked
petiole, pinnately veined, callous-denticulate and minutely ciliate, cori-
aceous, glabrous on both sides except the veins beneath, the upper
surface green, covered with lighter-colored warty blotches (pathologi-
cal ?), the lower surface very glaucous : corymb ample, with rather
numerous medium-sized heads; scales of the involucre about 10,
strongly carinate, acutish, 4 lines long : rays about 5, rounded at the
apex ; teeth of the disk-corollas exceeding the short campanidate
throat and half the length of the slender tube proper: achenes ribbed,
puberulent. — Rich slopes of barranca near Guadalajara, Jalisco ; Sep-
tember, 1890 (n. 3280).

Laurentia ovatifolia. Annual, 6 to 12 inches high, puberu-
lent ; stem slender, flexuous, angled, subsimple or branched above :
leaves petiolate, ovate, acute, abrupt or subcordate at base, finely and
regularly serrate, about an inch long, two thirds as broad, the lowest
smaller, rounder, obtuse, the upper lanceolate ; petioles 3 to 5 lines
long; pedicels filiform, cui-ving upward, 9 to 12 lines long, two to four
times longer than the linear-filiform bracts : flowers small, racemose ;
calyx-tube hardly any, the lobes linear, acute, not quite equalling
the corolla-tube ; corolla 2 lines long, the nearly white tube split
half-way to the base behind; limb blue, with very dissimilar lobes,
the three lower obovate, somewhat united to form a spreading lip, the
two upper erect, contracted to narrow points : stamens inserted near
the middle of the tube, the two smaller anthers peuicillate : capsule

OF ARTS AND SCIENCES.

167

ovute-oblong or elliiJtic in outline, papery, almost entirely superior,
2-valved at the apex. — Shaded bauks near Guadalajara; September,
1S89 (u. 2985) ; also near Cardeuas, Sau Luis Potosi, 1890 (u. 3302).
This plant, which was distributed under the name of Laurentia Michoa-
caiia, var. ovatifolia, appears ou further examination to be distinct from
n. 3337, of which it was at first regarded as a variety. It has the
habit of a Lobelia, but the corolla is split behind only part way to the
base. While in the limb of the corolla and the insertion of" the sta-
mens it agrees rather closely with Palmerella, the corolla-tube is short,
as in Laurentia. The plant thus furnishes additional evidence that
Sclionland is right in uniting the two genera.

Mr. Pringle's n. 3337, the plant which was distributed as Laurentia
Michoacana, Robinson, is without doubt closely related to the species
just described, but is distinguished by its sessile leaves of considerably
different form and toothing. It appears to be identical with Parry &
Palmer's n. 557, which Hemsley refers to Lobelia micrantha, HBK.,
and with Mr. J. Donnell Smith's n. 22, the Lobelia Turckhelmii of
Vatke (unedited?). The corolla in all these plants, however, is cleft
only from one third to half-way to the base, a character which should
certainly exclude them ft'om Lobelia, as at present defined by Ben-
tham & Hooker, Baillon, Schonlaud, and others. While the speci-
mens in question agree in most points with Laurentia ramosissima,
Benth. &; Plook. {Lobelia ramosissima^ Mart. & Gal.), they are dis-
tinguished by their much smaller flowers. Unfortunately, however,
Martens and Galeotti have characterized, under the name of Lobelia
farvijiura, a plant which also appears from their rather unsatisfac-
tory description to differ from their Lobelia ramosissima chiefly in the
same respect, namely the shorter corolla. This being the case, it seems
best that all three of the numbers just cited siiould be for the present
doubtfully referred to Lobelia parvijiora, JMart. & Gal., since it is
highly probable that the type of this species will be found ou investi-
gation to be a Laurentia, just as the Lobelia ■*'amosissima of the same
authors has already proved.

Lobelia novella. Stem a foot high, branching, with a few rather
large leaves near the base, almost naked above : leaves petiolate,
broadly ovate, acute or obtusish, rounded or subcordate at base, un-
dulate and finely raucronate-denticulate, hispid on both sides, scarcely
})aler beneath, 2^ to 3 inches long, 2 inches in breadth ; upper leaves
few, much smaller, sharply and doubly denticulate ; racemes long, se-
cund ; bractlets minute, scarcely exceeding a line in length ; pedicels 6 to
9 lines long : calyx-tube turbinate, in fruit equalling or surpassing the

168 PROCEEDINGS OF THE AMERICAN ACADEMY

narrowly Imear obtuse lobes ; corolla blue and white, the lobes of the
lower lip obovate-spatulate, longer than the very narrow upper lobes ;
capsule half-inferior. — Cool ledges and gravelly banks, Tamasopo
Canon, San Luis Potosi ; October, 1890 (n. 3288). Eesembles L.
Sartorii, Vatke (ex char.), but is distinguished by the elongated ra-
cemes, and by the size, shape, and indentation of the leaves.

Nemacladus oppositifolius. Stems a foot high, much branched
from near the ground ; branches becoming erect from a decumbent base,
usually simple, each bearing about six pairs of subopposite leaves,
and then continued as a lone; naked racemose inflorescence: leaves
petiolate, ovate, acuminate, rounded at base but inconspicuously decur-
rent upon the petioles and stem, sharply dentate, smooth, 9 to 15 lines
long ; petioles 2 or 3 lines in length ; bracts minute, awl-shaped, scarcely
a line in length, the bractlets subtending the calyx similar ; pedicels
not exceeding 2 lines long : calyx-tube short, rounded at base, the
segments equal, awl-shaped, acute, a line long ; corolla-tube not equal-
ling the calyx-lobes, the lower lip consisting of two ovate spreading
segments, much longer than the three upper lobes: stamineal column
ascending against the upper lip of the corolla and then curved forward ;
anthers stellately divaricate around the stigma, not at all appendaged :
mature capsule nearly equalling the calyx-lobes, two-valved a*^ the apex.
— Dry calcareous bluffs near Cardenas, San Luis Potosi ; October,
1890 (n. 3300). This plant, which is identical with Dr. Coulter's
n. 30, differs greatly in habit from the known species of Nemacladus.
It possesses, however, just the characters which distinguish this genus
in such a marked manner ; namely, the stellately divaricate anthers,
and the division of the corolla into a three-lobed upper lip and a two-
lobed lower lip. The subopposite leaves appear to be anomalous
among the Loheliacece.

Symplocos Pringlei. a small tree, 20 to 30 feet high ; branch-
lets and petioles covered with a very short rufous tomentum : leaves
elliptical, varying from obtusish to abruptly acuminate, cuneate at base,
sharply and regularly serrulate, subcoriaceous, green and glabrous
above, slightly paler beneath and pubescent especially along the mid-
rib, 2.^ to 4 inches long by 1^ broad ; petioles 4 to 6 lines long : flowers
usually aggregated by twos and threes, subsessile upon a scarcely dis-
tinguishable common peduncle, occasionally solitary ; calyx silvery-
sericeous ; lobes ovate, rounded or more or less pointed at the apex;
petals 5, broadly spatulate, coherent for a third of their length, either
smooth or very minutely pubescent on the outer surface : stamens ob-
scurely 4— 5-delphous, united high up. adnate to the corolla for half its

OF ARTS AND SCIENCES. 169

len«'tli : lower part of the style as well as the apex of the ovary strongly
hirsute ; ovary 3-celled ; fruit oblong, dark-colored, with slight puberu-
lence, tipped with the hardened calyx-teeth. — Hills of Patzcuaro,
Michoacau ; November, 1890 (n. 334a). Resembles in many respects
S. prionophyUa, Hemsl. (ex char.), but has leaves not at all ovate,
and the petals nearly smooth and united with each other and with the
stamens for about half their length. In the species just mentioned
they are described as nearly or quite free.

GoxoLOBus PARViFLOKUS, Gray, var. brevicoroxatus. Corolla
erect, instead of being reflexed as in the type ; the outer crown of five
short rather thick truncate scales, which are not at all produced into
lobes at the edges; the inner crown of five horns as in the typical
form. — Laredo, Texas; July, 1889 (n. 3029).

BuDDLEiA Chapalaxa. Suffruticose ; branches with a smooth yel-
lowish bark ; branchlets woolly : leaves ovate or ovate-oblong, 1^ to
2^ inches long, an inch broad, appearing more or less rhombic in out-
line from being contracted at base into a broadly winged petiole, au-
riculate-perfoliate, acute, rather coarsely and bluntly toothed, woolly
on both sides when young, becoming nearly smooth at maturity and re-
taining only a slight mealiness ; floral leaves much smaller, lanceolate,
scarcely perfoliate : flowers in globular heads ; the heads half an inch
in diameter and borne in pairs on axillary peduncles about 4 lines
long; corolla 1\ lines in diameter, short, not or scarcely exceeding the
woolly calyx. — On dry cliff's in the mountains near Luke Chapala,
Jalisco; December, 1889 (n. 2972).

CoRDiA Pringlei. A shrub 10 to 15 feet high, belonging to
the section Dasycephal(E ; branchlets, peduncles and petioles stellate-
tomentose : leaves ovate or ovate-oblong, more or less narrowed to
an obtuse apex, often one-sided, all but the smallest abrupt or even
subcordate at base, crenate-dentate with irregular teeth, green and
finely stellate-pubescent above, lighter and tomentose beneath, 3 inches
long, 1 or 2 inches broad ; petioles 6 lines in length : heads globular,
an inch in diameter on subterminal peduncles 1^ inches long ; calyx
narrowly campanulate, the tube white-pubescent, and the filiform teeth
stellately ferruginous-pubescent, curled irregularly, mostly reflexed ;
corolla funnel-form, white, 5 lines broad, with obtuse lobes: stamens
inserted near the summit of the pubescent throat. — Las Palmas, San
Luis Potosi ; June, 1890 (n. 3091). Differs from C. omhigua, Cham.
&, Schl., in its larger heads and the abrupt or subcordate bases of the
leaves, as well as in the absence of the long spreading hairs on the
petioles and peduncles. Mr. Pringle's plant much resembles a speci-

170 PROCEEDINGS OF THE AMERICAN ACADEMY

men from Nicaragua, collected by C. Wright (N. Pacif. Expl. Exped.
1853-56). The latter differs, however, in its larger flowers aud more
finely toothed leaves, which are more densely aud roughly pubescent
upon the upper surface.

Heliotkoi'IUM Phinglei. a spreading annual, branching from
the base, 2 to 5 inches high, silky-villous : leaves elliptic-lanceolate,
acute at both ends, 4 to 6 lines long, 2 lines broad; petioles 1 or 2
lines long : flowers scattered along the leafy branches from near the
base ; calyx-lobes lance-linear, acuminate ; corolla-tube 1 to 1^ lines
long, the limb abruptly expanding, 2^ lines broad, with five acutish
primary lubes, with which five short teeth alternate : fruit separable
at maturity into four nutlets, minutely reticulate-roughened on the
back but not hispid. — Santa Eulalia Mountains, Chihuahua. October,
1886 (n. 1160) ; also in thin soil of rocky hills, Las Canoas, San Luis
Potosi, July, 1890 (n. 3207). This species resembles in habit //
jjhyllostachyum, Torr., but has flowers more than twice as large. It
is distinguished from H. Umbatum, Benth., under which name n. 1160
was distributed, by its broader leaves, very scattered flowers, more
acute lobes of calyx and corolla, rougher nutlets, and its more decid-
edly annual character. The inflorescence in H. Uinbatum is distinctly
terminal and compact, and the root often clearly perennial.

Omphalodks ACUMINATA. Stem weak, procumbent, dift'usely
branched, nearly smooth, 18 inches long or more, leafy throughout :
leaves deltoid, attenuate-acuminate, entire, 2 inches long, 15 lines in
breadth, cordate with rounded lobes and broad sinus, membranaceous,^
nearly smooth above, puberuleut with very short hairs beneath ; the
upper leaves smaller, lanceolate, subcordate or rounded at base ; peti-
oles 1 to 1^ inches long: peduncles lateral, scattered, very slender,
flexuous-sprcading, 9 to 12 lines in lengtli : calyx-lobes oblong, acute;
corolla 4 or 5 lines broad, rose-color, with short tube and rounded
spreading segments ; scales very prominent, nearly globular, minutely
papillose : nutlets orbicular, only one in each flower ripening, 1^} lines
in diameter at maturity, flat, horizontal over the gynobase, the wing
reflexed, about 30-tootlied. — Li the Sierra Madre. near Monterey ;
June, 1888 (n. 2220). Resembles 0. cardioplii/Ua^ Gray, but is
smoother throuirhout and has much longer stems of more strairuling
habit, the leaves twice as large, of more delicate texture, and much
more attenuate at the apex. In that species, moreover, two to four
nutlets mature and the wing-borders have but half as many teeth.

IroAiai;AL Leonensis. Root tuberous, an inch thick ; stem twining,
striate-angulate, minutely roughened : leaves deeply palmatQly 5-lobed,

OF ARTS AND SCIENCES.

171

cordate or subcordate, 3 to <) inches broad, punctate above and with a
soft deciduous tomentum beneath, the lobes lanceolate, rounded at the
apex and inconspicuously niucronate, or in young leaves sometimes
acutii?h, the two outer lobes of the larger leaves each appendaged at
base with a large blunt auricle ; petioles an inch long or more, rough-
ened especially near the base by numerous small tuberculate glands :
peduncles 2 inches long, l-flowered, puberulent ; bractlets minute or
obsolete ; pedicels half an inch long: sepals ovate, with rounded apex,
4 lines long, canescent with very short ai)pressed silvery hairs ; co-
rolla nearly 3 inches long, twice the length of the stamens, purple. —
Calcareous ledges near Monterey ; June, 1889 (n. 2840). Allied to
Ji digitata, L., but differing in its l-flowered peduncles and glandular
petioles, as well as in the shape and pubescence of the leaves. The
flowers are also larger and the throat relatively narrower.

Bassovia Mexicana. A shrub, 5 to 15 feet high; branches un-
equally dichotomous, slightly striate-angled and minutely verrucose :
leaves membranaceous, geminate, ovate, acuminate, the larger 1^ to 2
inches long, | to 1 inch broad, the smaller half as large, all abrupt at
base or slightly decurrent on one side upon the slender petiole, green
on both sides, slightly scabrous above, sparsely pubescent on the veins
beneath : inflorescence umbelliform, sessile, axillary ; pedicels 2 to 6,
l-flowered, 4 to 5 lines long, striate, slightly thickened upward : calyx
spreading, shallow cup-shaped, strictly truncate, persistent; corolla 3
to 3^ lines in length, divided to the middle or even more deeply ; tube
short, slender; limb campanulate, of 4 or 5 lanceolate acuminate seg-
ments, puberulent upon the outside; stamens exserted, the filaments
at maturity exceeding the anthers ; throat of the corolla and base of
the filaments bearded : fruit the size of a pea, red (turning black?),
many-seeded ; seeds a little over half a line in diameter, compressed,
roughened with pits. — Tamasopo Cafion, San Luis Potosi ; June,
1890 (n. 3071). Resembles some species of the genus Brachistus
in its truncate calyx, but has the deeply cleft valvate corolla of Bas-
sovia, and shows close affinity to B. macroplnjlla, Benth. & Hook.
(Witheringia macropl/ylla, auct.), from which it differs principally in
its ligneous stem, much smaller leaves, fewer-flowered fascicles, and
more persistent calyx. The distinctions between Bassovia and Bra-
chisfus are not at all satisfactory, and considering the intermediate char-
acter of this new species it seems best to refer it to the older genus.

AVithania(?) melanocystis. a shrub, 5 to 8 feet high, minutely
pubescent on the leaves and younger parts; branches gray, striate :
leaves single or geminate, lanceolate to ovate-lanceolate, acute, entire

172 PROCEEDINGS OF THE AMERICAN ACADEMY

15 lices long by 6 wide, narrowed at the base into a petiole 3 to 4
lines iu length : flowers 2 to 4 at each node, usually but one of each
group fruiting ; common peduncle none ; pedicels slender, 3 lines long :
calyx-teeth 5, at authesls ovate, one third as long as the tube ; fruiting
calyx much enlarged, bladder-like, neither angled nor conspicuously
veined, becoming black, 9 lines in diameter, with throat contracted
and teetli relatively short ; corolla campanulate, 5 lines long, with a
more or less spreading limb of five ovate acutish segments equalling
the tube, pale yellow with conspicuous dark spots in the puberulent
throat; filaments inserted in the middle of the short tube, half the
length of tlie oblong anthers : berry globular, red, much smaller than
the enclosing calyx ; seeds roughened, compressed, yellow. — Tama-
sopo Caiion, Sau Luis Potosi ; June to October, 1890 (n. 32S5). A
plant of puzzling affinities, possessing the habit of a Phi/salts, but hav-
ing a woody stem, clustered flowers, calyx without conspicuous angles
or veins, and anthers much exceeding the filaments. Some of the
characters resemble those of Athencea ; in that genus, however, the
fruiting calyx is much more deeply divided.

Herpkstis auriculata. Roots fibrous; stem erect, smooth, sharply
4-angled, much branched above, a foot high ; branches mostly simple:
leaves lanceolate, acute, serrulate, sessile with an auriculate-clasping
base, punctate, 8 to 10 lines long, 3 lines broad ; floral leaves gradu-
ally reduced to linear bracts 2 to 3 lines long : flowers small, axillary,
mostly in pairs, opposite at each node; pedicels (iliform, 3 lines long;
divisions of the calyx acuminate, a line long, becoming l-^- lines long
in fruit, very unequal in breadth, tlu-ee of them ovate, the inner two
rather narrowly lanceolate, none of them at all oblong ; corolla pur-
plish, 3 lines long, slightly bearded within near the summit of the
tube ; rudimentary stamen present, very small, subulate : capsule de-
pressed-ovoid, suhglobose, not exceeding the calyx, 2-valved. — AVet
places near Guadalajara; November, 1889 (n. 2937).

Gerardia punctata. Stem puberulent, nuich branched: leaves
small, numerous, opposite or appearing fascicled, linear, entire, acute,
smooth, 7 to 10 lines long, a line or less in breadtli ; bracts similar:
flowers rather numerous, opposite in long subspicate racemes ; pedicels
1 to 2 lines long: calyx smooth, 10-nerved, often punctate with small
purple dots, the five teeth lanceolate, acuminate, exceeding the tube,
tlie intermediate nerves terminating at each sinus in a minute glandu-
lar prominence ; corolla purple, tubular-campanulate, an incli long,
the lobes a fourth as long as the tube; stamens conspicuously didyna-
mous, the upper pair being much shorter and having considerably

OP ARTS AND SCIENCES. 173

smaller anthers than the lower ; filaments glabrous ; anthers hirsute,
mucronate at the base of each cell : style piiberuleut ; capsule ovate,
acuminate, equalling the calyx-lobes; seeds oblong, light brown, the
external seed-coat loose, cellular, and iridescent. — Hills of Zacatccas ;
October, 1888 (n. 2183).

Castilleia m acrostic. ma. Perennial, 8 inches to a foot hi'di,
nearly smooth or more or less hispid with weak spreading hairs :
leaves small, rather numerous especially at the base, linear, acute, en-
tire, subamplexicaul, 1-3-nerved, 9 to 15 lines long, seldom exceed-
ing a line in breadth ; floral leaves broader and shorter, lanceolate or
narrowly ovate, angled or sparingly toothed toward the apex, often
yellow-tipped : flowers closely sessile in a short spike, slender, 8 to 12
lines long, surpassing the bracts; calyx light yellow, equally cleft be-
fore and behind, the lobes emarginate or two-toothed ; corolla a third
longer than the calyx, yellow, the lower teeth linear, acuminate, erect:
stigmas conspicuously exserted, thickish, sometimes more than a line
long, recurved. — Grassy slopes, Flor de Maria, Mexico ; July, 1890
(n. 3194). This appears to be the same as Mr. Pringle's n. 1545
from Chihuahua, which was distributed under the provisional and
unedited name of C. h'ihospermoides, HBK., var. (?) jlava, Watson.
Specimens of this number differ from 3194 chiefly in being more his-
pid and in having the stigmas less prominently exserted, which differ-
ences may be in part due to the less advanced stage of development.
Both plants differ from C. lithospermoides in the size and color of the
corolla, in the deeply cleft conspicuously exserted stigma, and also in
the shorter relatively thicker and more coarsely reticulated seeds.

RuELLiA BouRG^i, Hemsley. The mature cajjsules of this species,
which were wanting in the type specimens and could not therefore be
described, are excellently shown in Mr. Pringle's u. 2951 (barranca
near Guadalajara, September, 1889). They are rigid, lanceolate,
acutish, cuspidate, glabrous, 2 to 2\ inches long and ^ inch broad, the
valves with a conspicuous median furrow; hooks ascending, bifid;
seeds orbicular, nearly 3 lines in diameter.

JusTiciA Pringlei. Herbaceous, 2 feet high, hirsute with spread-
ing white hairs : leaves ovate, acuminate, entire, petiolate, l-g- to 2^
inches long, half as wide ; the lowest smaller, orbicular, obtuse : flowers
small, distant, borne in a spreading panicle ; lower bracts foliar, ovate,
the upper much smaller, lanceolate ; bractlets minute, awl-shaped, half
the length of the calyx ; calyx-teeth 4, narrowly linear, attenuate,
2^ to 3 lines long ; corolla 4 to 5 lines long, pale rose-color, the limb
nearly equalling the tube, the upper lip narrowed to a retuse summit,

174 PROCEEDINGS OF THE AMERICAN ACADEMY

the lower lip rounded : anther-cells nearly equal in size, the upper not
ciliate, minutely mucionate at the base, the lower with a conspicuous
spur: capsule 5 lines long, soft-pubesceut, 4-seeded; seeds compressed,
very hairy especially at the edge. — On hills near Guadalajara; Oc-
tober, 1889 (n. 2967). This species seems anomalous among the
New "World members of the genus. It resembles more nearly the
species of the § Rhaphidospora of the Old World, which have a diffuse
inflorescence and echinate-hispid seeds. It differs from § Sarotheca
chiefly in lacking the ciliation of the anthers.

CiTiiAREXYLU.At Berlandieri. Ten to fifteen feet high; branches
gray, striate ; branchlets pubescent, striate, acutely 4-angled but soon
becoming round ; leaves ovate or rhombic, varying in the same indi-
vidual from acute to refuse at the apex, shortly cuneate at base, mem-
branaceous, puberulent above, soft-pubescent beneath, 1 to If inches
long, 9 lines broad, entire, or some with a few blunt teeth near the
apex ; inflorescence termhial, subspicate, more or less flexuous, 1 to 2
inches long, becoming more rigid in fruit; pedicels half a line long,
with minute filiform bractlets of the same length: calyx turbinate, 2
lines long, striate, pubescent, the five subequal teeth erect, very short
and blunt; corolla-tube just equalling the calyx, pubescent within, the
spreading limb of five rounded lobes, puberulent on both sides : fertile
stamens four, filaments hairy; the fifth stamen present as an an-
therless rudiment: drupes dark brown, as large as peas, crowded in
the spikes. — Rocky hills. Las Canoas, San Luis Potosi ; July, 1890
(n. 3222). Apparently identical with n. 3182 of Berlandier, coUectecf
near INIatamoras, May, 1832. This specimen of Berlandier's was
doubtfully referred by Dr. Gray to C villosum, Jacq., and has formed
the sole evidence that this species occurs in Mexico. In the light of
]\Ir. Pringle's better material, however, it is clear that the Mexican
plant is quite distinct from the species just mentioned.

Scutellaria hispidula. Perennial with ligneous base; stems
numerous, mostly sim[)le, hispidulous, a span high, often dark purple as
well as the floral leaves and calyx : leaves small, ovate or ovate-elliptic,
subglahrous or very minutely hispid but not at all cinereous-pubes-
cent, 4 to 5 lines long (about half the length of the internodes), 2 to
3 lines broad, all except the very lowest abrupt at the base and sessile :
flowers numerous, of medium size, with short pedicels ; calyx strongly
accrescent, hispid with short white hairs ; corolla blue, white in the
throat, soft-pubescent on the outside, 5 to 7 linos long, the upper lip
but slightly cucuUate ; upper stamens often exserted. — Meadows, Flor
de Maria, ^lexico ; August, 1890 (n. 3233). A plant with the habit

OF ARTS AND SCIENCES. 175

of -S. Dnomnondii, Bentli., and S. Wrif/ldil, Gray, but distinguished
from the former by its ligneous base and clearly perennial nature,
and from the latter by the form of the leaves, character of the pu-
bescence, smaller blue flowers, and less cucullate ujiper lip. Whilt!
it may ultimately prove to be a variety of ^S. WrigJitii, such a disposi-
tion of it in the absence of connecting forms and with the difference of
geographic occurrence would at present be unwarranted.

Mi.MULUS CoNGDONii. Very small, at flowering subacaulescent ;
stems glandular-pubescent, becoming in iVuit 1 to 4 inches higii, much
branched near tlie base ; branches simple : leaves ovate or lanceolate,
obtuse, entire, dark purple beneath, \ an inch long, narrowed at base
to ciliate petioles of nearly equal, length ; peduncles in fruit 1-^ to
2 lines long, often reflexed : flowers very small; calyx-tube very
slender, prismatic, glandular-puberulent, moderately gibbous at base,
becoming strongly so in fruit, ending obliquely in short teeth, the
upper tooth the longest; corolla rose-purple, the slender tube 4 to 6
lines in length, with little or no distinctly enlarged throat, the sub-
regular abruptly spreading limb H to 2 lines in diameter; stamens
strongly didynamous, the upper pair much shorter and occasionally
with abortive anthers: style puberulent above; capsule cartilaginous,
very gibbous, laterally compressed, narrowed from a moderately broad
base, acute, deeply furrowed on the sides ; seeds minute, acute at each
end. — Collected by Mr. J. W. Congdon, in Mariposa County, Cali-
fornia, at Zimmerman's Ranch, in March, 1887, in April, 1888
(flowers and fruit), and in iMay, 1888 (fruit); also at Stockton Creek,
March, 1889 (flowers) ; and at Agua Fria (fruit). The diminutive
size and nearly acaulescent character of flowering specimens of this
plant made it at first appear probable that it represented merely a
dwarfed, early-spring form of one of the larger-flowered species. The
constant characters of Mr. Congdon's specimens, however, collected as
they were at different dates and localities, and representing verj' dif-
ferent stages of development, prove it a normal form and a distinct
species. While the vegetative habit is much like that oi 31. Kelloygii,
Curran, it is distinguished from that species by its much shorter co-
rolla-tube and smaller limb, as well as by its acute and not at all
oblong capsules. From M. pnlcheUns, Greene, it differs in its smaller
rose-purple corolla without the yellow lip, in its much shorter calyx-
teeth, and in other ways. In its short corolla-tube and very gibbous
capsule it resembles 31. latifoUus, Gray, but differs in its smaller size,
in its habit of branching from the base, (the stem of 3f. latifoUvs
although branching al)ove is simple below,) in its very slender calyx-

H I

176 PROCEEDINGS OF THE AMERICAN ACADEMY

tube, longer and usually reflexed peduncles, and relatively slender and
more acute capsules. Like M. latifoliiis, this species is somewhat in-
termediate between §§ (Eiioe and Eunanus.

MiMULUs GRAciLiPES. A delicate annual, 3 to 5 inches high ; stem
usually simple, glaudular-puberulent under a lens : leaves about four
pairs, very small, seldom exceeding 4 lines in length, sessile, lanceolate,
obtusish or scarcely acute, entire or minutely denticulate, the round-
ish cotyledonary(?) leaves persisting at base: peduncles springing by
pairs from each node, an inch long, filiform, spreading horizontally
and gracefully curved upward : calyx oblong-campanulate, a little
over 2 lines in length, a line in diameter, the equal teeth slightly
spreading, obtuse, strongly ciliate ; corolla rose-purple with yellow
markings in the throat, the tube not quite twice the length of the
calyx ; limb spreading, 3 or 4 lines in diameter, strongly bilabiate, the
upper lip much shorter than the ample 3-lobed lower one : capsule
symmetrical, of delicate texture, elliptical in outline, 2 lines in length,
included in the scarcely inflated calyx ; seeds obtuse, apiculate at each
end. — Collected by Mr. J, W. Congdon, at Mormon Bar, Mariposa
County, April, 1889, Near M. ruhellus, Gray, and M. dejlexiis, "Wat-
son, but differing from the former in its simple stem, much larger co-
rolla, and longer peduncles ; from the latter in the color of its corolla;
and from both these and various other related species in its ciliate
calyx.

Aster Engelmanni, Gray, var. (?) paucicapitatus. Slender;
stem weak and flexuous : leaves numerous, oblong-elliptic, obtusi?h,
apiculate, puberulent above and with scanty traces of tomeutum be-
neath, 10 to 15 lines long, a fourth as broad : heads solitary, terminal,
or sometimes 3 to 5 ; scales of the involucre more equal and less closely
imbricated than in the typical form or other varieties; rays white or
pink. — Collected by Mr, C, V, Piper iu the 01ym{)ic Mountains,
Washington, September, 1890. This is a very puzzling form and
perhaps a distinct species. If, however, Dr, Gray was right in class-
ing A. ledophylliis, Gray, and A. Engehnanni, var. glaucescens, as va-
rieties of the same species, Mr. Piper's plants, which are in some
regards intermediate between these varieties, should probably also be
included iu ji. Engelmanni.

In closing, the writer wishes to express his warmest thanks to Dr.
Serene Watson for his kind and ready aid in many points of dillicult
classification and perplexiug uomeucluture.

OF ARTS AND SCIKNCE3. 177

XIII.

CONCERNING THE LIFE-HISTORY OF SACCORHIZA
DERMATODEA, (De la Pyl.) J. Ag.

By William Albert Setchell.

Presented by W. G. Farlow, June 10, 189L

We owe our first knowledge of this interesting species to Baron de la
Pylaie, who discovered it iu 1816, on his first visit to Newfoundland
and the neighboring islands of St. Pierre and Miquelon. He deposited
specimens of this plant with the Museum of Paris in 1817, and gave
to it the name of Laminaire en forme de cuir, or Laminaria der-
matodea. In his second voyage to Newfoundland, in 1819, De la
Pylaie extended his observation on this species. In 1824 he pub-
lished the name given to the herbarium specimens, as well as a few
remarks on the affinities of the species, in the " Annales des Sciences
Naturelles." But it was not until 1829, however, that he gave any
detailed account of its form and structure. In that year he published
the first part of the " Flore de Terre Neuve," a work which was to
include descriptions of all the plants collected by him in the "New
World." This part, however, was the only one ever published, and
that lacks the plates intended to accompany it.

The description both of the external form and the internal struc-
ture published in the " Flore de Terre Neuve " is very full. De la
Pylaie considered that it replaced on the shores of Newfoundland the
Laminaria bulbosa of Europe, whose near relative he rightly judged
it to be. The name was suggested by the thickness and leathery con-
sistency of the adult plant.

The plant thus discovered and described has since been found in
several other places, and a number of references have been made to
it. A list of as many as have come under the notice of the writer
is appended to this paper.

Distribution. — The range of SaccorMza dermatodea seems to be
circumpolar. It is found at Vancouver's Island, on the western coast
of North America,* and along the shore of Southwestern Kam-

* Harvey, Prnc. Linn. Soc, Vol. VI. p. 106, 18G2.
VOL. XXVI. (n. 8. xvai.) 12

178 PROCEEDINGS OF THE AMERICAN ACADEMY

tschatka.* Kjellman also credits it to the North Pacific. f It is,
indeed, unknown either in the American Arctic Sea t or in the Si-
berian Sea ; but the former has uever been explored to any extent,
and although the latter has been more thoroughly searched, the phy-
sical conditions over the greater portion of it are unfavorable to the
growth of any algce. § The species reappears in the Murman Sea,||
extends into tlie White Sea, If and thence along the northern coasts of
Norway.** In the Greenland Sea it occurs at Spitzbergen,tt and at
" Beeren Eiland."t$ It is found in Baffin's Bay on the western coast
of Greenland, §§ and extends down along the eastern coast of North
America nearly to Cape Cod.||||

From the scanty references to the occurrence of Saccorhiza der-
matodea in the North Pacific, one might infer that it is very scarce
there. The locality in the Arctic Ocean where it is most abundant is,
according to Kjellman, the northern coast of Norway ; but even there,
he says, " it does not occur in any such numbers as are comparable in
any way with other Lmninariacece ." On page 9G of his "New Eng-
land Algai," Farlow has called attention to the fact that this species
is of much rarer occurrence on the New England coast than almost
any other species of the same family ; and that, while rare at its
southern limit in the region of Boston, Mass., it becomes more abun-
dant on the coast of INIaine. My own experience in collecting on
the New England coast fully confirms this. It has never been
found growing south of Nahant, Mass., where it is ordinarily rather
rare. Farther north at Portland and Mount Desert Island, Maine,
it is more plentiful, while at Eastport, Maine, it is said to be fairly
abundant.

Habitat. — Eaton HIT records that a specimen was brought up by the
dredge through twenty-five fathoms of water off Campobello Island,

* Riipr., Tange Och., pp. 352 and 409, 1847.

t Arct. Alg., p. 38, 1883.

\ Cf. Kjellman, he. cit., pp. 4 and 5, for the divisions of the Arctic Sea.

§ Kjellm., he. cit., p. 26.

II Kjellm., Alg. Murm., p. 36, 1877.

1 L. Bcvrll, Post. & Kupr., 111. Alg., p. ii, 1840 (>/<• Areschoug).
** Aresdi., Obs. Phyc., Part III. p. 11, 1875. Kjellm., Arct. Alg., p. 225.
tt J. Ag., Spets. Alg. Till., p. 31, 18G8. Kjellm., Arct. Alg., p. 225.
Jt Kjellm., Arct. Alg., p. 225.
§§ J Ag., Griinl. Lam. och Fuc, p. 11, 1872.

lill De la Pyl., Ann. Sci. Nat., Ser. I.Tom. IV. p. 179, 1824 ; Fl. Terre Neuve,
p. 48, 1829. Fallow, N. E. Alg., p. 95, 1881.
IfH Trans. Conn. Acad., Vol. II. p. 343, 1873.

OF AUTS AND SCIENCES.

179

near Eastport. Kjellman * says that " in the Arctic Sea proper the
present species occurs in company with other Laminnriacece^ and is
usually met with here at a depth of 2-10 fathoms on rocky or stony
bottom"; but ''on the coast of Norway," he says, "it does not be-
long to the proper formation of Lnminariacece, but descends deeper
than this, even to a depth of twenty fathoms. But it is most common
here," he adds, " in shallow, rather exposed bays on gravelly bottom
in 4—5 fathoms." Farlow^ states f that at Eastport it is found in deep
pools, but that elsewhere it is an inhabitant of deep water. I have
found it at Peak's Island, growing in shallow pools between tide
marks, and down to eight or ten feet below low-water mark. I
did not search for it at a greater depth. At Marblehead, Mass., it
occurred in a very shallow pool at extreme low-water mark. At
Nahant, in May and June, 1889, it grew in abundance just below
low-water mark, and in pools between tides. But a number of speci-
mens were found in two rather deep pools at the uppermost tide
limit.

Season. — De la Pylaie mentions } that numerous very young
specimens were cast ashore about the end of October; and Kjellman
states § that young individuals were common during the winter on the
north coast of Spitzbergen, but that on the "coast of Norway younger
and older specimens are of rather the same frequency during the sum-
mer months, in July and August." The season on the coast of New
England seems to correspond Vfith that of the coast of Norway. I
have seen young specimens of Saccorhiza dermatodea only in late
spring and summer. The first young specimens I found were growing
at Nahant, just below low-water mark, June 23, 1888, and were only
two or three in number. On April 24, 1889, on a visit to Nahant, I
found the pools full of young Laminariece of small size, among which
young specimens of the present species were tolerably abundant.
These were of different lengths, and a number of specimens were col-
lected. A careful search was made for the very earliest stages, but
none were found. Some very small plants were obtained, but none of
the very earliest. It may be present in large numbers in a given
locality at one season, and almost entirely wanting during the next.
In June, 1888, on the northeastern side of Peak's Island, hundreds
of beautiful large plants were growing just a few feet below low-
water mark ; but in June, 1889, not a single plant was to be found in

* Arct. Alg., p. 225.
t N. E. Alg., p. 96.

\ Fl. Terre Neuve, p. 49.
§ Loc. cit., p. 225.

180 PROCEEDINGS OF THE AMERICAN ACADEMY

the same locality, even after a very careful search. lu June, 1888,
only a very few plants were to be found in the tide-pools at Nahant,
and those occurring in the very lowest; but in May and June, 1889,
all the tide-pools were full of young plants, even the very upper-
most. In the mass of kelps thrown ashore after storms are to be
found comparatively few individuals of this species. They are usually
more abundantly cast up in February and March than in the other
months, and are for the most part old plants, which are often much
mutilated.

Material. — The material on which the present paper is based was
collected, for the greater part, by the writer on the shores of Maine
and Massachusetts. During the year 1889 the tide-pools at Nahant,
Mass., furnished an excellent place for observing this species in almost
all of its stages. The first collections were made in the last week of
April, and the locality was visited from time to time imtil June of the
following year. Collections were also made at various times at Marble-
head, Mass., and at Peak's Island, near Portland, Maine. From these
various collections specimens of all ages were preserved in alcohol of
about fifty per cent strength, and numerous notes and observations
made from them while living. Consequently a number of interesting
facts concerning the development have been revealed, and certain rela-
tionships more firmly established. In the following paper the mor-
phology of the different periods in the life-history will be given first,
and then the histological details of these same periods.

Morphology.

Literature. — The first description of the plant is the one by De la
Pylaie in the " Flore de Terre Neuve." He describes in full the
morphology of the adult plant and that of several stages of develop-
ment. Pie emphasizes one point about the origin of the permanent
holdfast, which seems to have been entirely overlooked by later
writers. Areschong* was the next writer to add much to our stock
of information. He gives a number of very valuable details concern-
ing the size and form of some plants of different ages, and is the
first to announce the discovery of this species in Northern Norway
and to distribute specimens of it.f On the label of the specimen
distributed he remaiks, "cum speciminibus e Lapponica Rossica
nomine L. Bcerii, Post. & Rupr. acceptis melius quara cum specimini-

* Obs. Phyc, Tart. III. p. 11.

t Aig. Scand. Exss., No. 213, under tlie name of Laminaria lorea.

OP ARTS AND SCIENCES. 181

bus e Terra Nova convenit nostra plaiita." The first part of the
sentence is very vahiable, as it establishes the identity of the L. Bcerii
of Postels and Ruprecht. As to the second part, however it may
have been with Areschoug's American specimens, the plant distrib-
uted agrees so completely with New England specimens that there
can be no doubt that the Norwegian and New P]ngland foi ms belong
to the same species. Farlow* gives a very careful account oi; the
New England plant, and Kjellman t adds greatly to our knowledge
of the Arctic forms of the species.

An interesting paper closely connected with this subject, and one
to which I am indebted for many valuable suggestions, is one by Mr.
P. A. Barber, entitled, " On the Structure and Development of the
Bulb in Laminaria bulbosa.'^ t This paper has made possible a fairly
satisfactory comparison between the earlier stages of Laminaria bul-
bosa, Lamour., and of L. dermatodea, De la Pyl., and helped to make
the discussion of the relationship existing between the two species
much more satisfactory.

Periods. — For convenience of description, the life-history of Sac-
corhiza dermatodea may be divided into four periods, each of which is
characterized by some important changes in development.

First Period. — This period naturally deals with the development
from the spore. The material is wanting for any satisfactory treat-
ment, and as the plants are at first of microscopic size, these earlier
stages will be discussed under the portion devoted to the histology.

Second Period. — 1. Earliest Stages found. — The smallest speci-
mens available are two or three which measure 5 to 6 mm. in length.
They were found growing in the same clusters with larger ones of the
same species, and were detected only by very careful search. Smaller
ones were sought for through each bunch collected, but none were
found. A hand lens is necessary in examining these specimens. The
point of attachment has been lost in every case. The stipe is dis-
tinct, and is about 1 mm. long and 0,14 mm. wide at the middle. It
seems to be nearly cylindrical in shape, and expands gradually into
the blade at the upper portion. The blade, which has thus a wedge-
shaped base, is generally ovate or obovate, the broadest portion being
in most cases about two thirds of the way from the base. The tip of
the blade either tapers to a more or less acute point, or is blunt and
truncated from having lost the terminal portion. Below, the stipe

* N. E. Alg., p. 95. t Arct. Alg., p. 223.

t Annals of Botany, Vol. III. No. IX. p. 41, February, 1889.

182 PROCEEDINGS OF THE AMERICAN ACADEMY

appears to be prolonged into the blade one third to one half the way
up, as a sort of broad midrib. The midrib-like portion is rather
wide at the base, and tapers gradually to a point. The edges are
not straight lines, but are somewhat notched or toothed. As seen
with the lens, the apex of this portion is not acute, but is somewhat
blunt. (Cf. Figs. 1 and 2.)

The midrib is very striking in fresh specimens as it is of a deeper
color than the rest of the blade. When examined with a low power
(150 diameters), it is seen to be thicker and of more complex structure
than the rest of the blade, which appears on focusing to be a simple
membrane composed of a single layer of cells. The further details
of this structure properly come under the sections devoted to the
histology.

On some of these young specimens are to be found small clusters of
hairs arising from the surface not only of the more complex portion,
but also in some cases from the membranaceous outer part of the blade.
These are the cryptostomata or " Fasergriibchen." By the presence of
these organs the young plants of Saccorhiza dermatodea may be dis-
tinguished from the young plants of any of the other species of Lam-
inariecB of the New England coast, except from those of species of
Alaria. From these they may be distinguishetl, after a little experi-
ence, by the outline and the color. Saccorhiza dermatodea in these
early stages is usually elliptical to obovate, while the species of Alaria
uniformly have a narrowly linear frond. The color of the young
Saccorhiza is light brown. They become somewhat greenish in dry-
ing at times, but generally take on a more yellow tint. The young
AlaricE are of a much lighter brown, and become decidedly greenish
on drying. The midrib-like portion in Alaria, too, is more linear in
shape, and narrower, more like a real midrib. As soon as the Alaria
attain any considerable size the true midrib makes its appearance, and
distinguishes them at once from the young plants of any of the other
species.

2. Later Stages. — As the specimens become larger a number of
changes may be noticed, which are well shown in a specimen about
o cm. high. This specimen possesses a distinct organ for attachment,
consisting of the slightly dilated basal portion of the stipe and a
sort of small cushion by which this is attached to the substratum.
The cushion has a spongy appearance and a dirty brown color, and
differs decidedly from the stipe both in color and consistency. The
stipe itself is about 3 mm. high and about 0.5 mm. in diameter. It is
cylindrical in shape, and expands gradually into the lanceolate or

OF ARTS AND SCIENCES. 183

oblanceolate blade, which is about 6 mm. wide. The terminal portion
of the blade has disappeared through erosion, but at the tip on either
side are seen two narrow strips (cf. Fig. 3, E) of the membrane of
the original blade still jiresent. The rest of tlie blade is more com-
plex in its structure. The cryptostomata are much more numerous
and much more conspicuous. They are much larger towards tlie tip
of the blade, and become smaller and smaller as they approach the
base, until at the place where the blade passes over into the stipe
they disappear altogether. Hence it is reasonable to assume that this
transition place, as we may call it, the " Uebergangstelle " of German
writers, is a region of active growth.

Third Period. — 1. Development of the Rhizogen. — Very soon after
the primitive or simple membranaceous blade has thus been very nearly
replaced by a blade of more complex structure, and the stipe has
elongated to a somewhat greater extent, we come to the beginning of
a set of changes in the lower portion of the stipe, which are of great
importance in establishing the relationships between this species and
others of the same group. The holdfast or organ of attachment iu
this stage differs from that of the preceding merely in its somewhat
greater size. In some of the specimens, however, the holdfast is a
vertically elongated body of narrowly conical shape. This variation
seems to be the result of peculiar conditions in the substratum to which
the plant is attached. The stipe iu the specimen chosen as a good
illustration of this stage (cf. Fig. 4) is about 8 mm. long and 0.5 mm. in
diameter. It is for the greater part of its length cylindrical, but at
about 2 mm. from the holdfast it appears somewhat swollen (cf. Fig.
4, B). This swelling, which is regularly ellipsoidal in general outline,
is in the specimen chosen about 2 mm. in height and 1 mm. in diam-
eter. Even before this swelling becomes evident, that portion of the
stipe where it is to appear seems denser, and is therefore of a slightly
different color from the neighboring portions, although at the time it
is hard to detect that the surface is at all curved outward. The swell-
ing or node is very readily seen, both in fresh and in alcoholic
specimens, and is very nearly as distinct in specimens pressed for
preservation in the herbarium.

This swelling is described by De la Pylaie in the " Flore de Terre
Neuve " (pp. 49 and 50), and was observed by him in plants thrown
ashore about the end of October. His description of a young plant
is so excellent that I reproduce it here. " Tout le vegetal n'avait
alors que 2 a 3 centimetres de hauteur, et sa fronde, d'une delicatesse
infinie, n'avait en largeur que 3 millimetres au plus. Cette fronde ua

184 PROCEEDINGS OF THE AMERICAN ACADEMY

peu rcservee en pointe a son sommet puis eiisuite lineaire sur la
plus graude partie de sa longeur, s'amiucit successivement jusqu'a se
terminer par uu petit stipe presque capillaire, oil Ton distingue '
neanmoins, au dessus de son point radical, I'espece de petit noeud
(Rhizogene), que j'ai vu plus tard s'accroitre et donner naissance aux
priucipales raciues qui fixent cette Algue aux corps solides." It is
very evident from this description that De la Pylaie recognized this
swelling just above the {jrimitive discoid attachment as well as its
function, and that he gave to it the name of " Rhizogene." In regard
to this organ and the term applied to it, he explains in a foot-note as
follows : " Organe exclusif a cette hydrophyte et au L. bulbosa ; il
n'existe que chez ces especes dans tout le regne vegetal ! n'ayant pas .
encore ete dcsigne par un nom particulier, je propose celui-ci dont je
me servirai pour eviter periphrases." This name, Anglicized to rhizo-
gen, will be adopted in this paper to designate this organ. The blade
in specimens in this stage varies from 3.5 cm. to 6.5 cm., and is per-
haps at times even longer. It is narrowly cuneate at the base,
and still bears some traces of the primitive blade at the apex (of.
Fig. 4, E).

2. Development of the First Hapteres. — For a number of stages
now the changes which take place in the rhizogen are of particular
interest. It increases, chiefly in diameter, until it projects out around
the stipe in the form of a circular ridge, as shown in Figure 5 at B.
This specimen is 7.3 cm. long. The holdfast is a disk-shaped body
about 2 mm. in diameter. It will be best hereafter to distinguish this
first organ for attachment to the substratum as the primitive disk or
primitive holdfast. The rhizogen is situated about 2 mm. above this,
and forms a conspicuous projecting circular ridge, about 3 mm. in
diameter. Its upper surface is convex and rounded, while the lower
surface is concave to a slight extent. Its projecting edge is coarsely
crenate into five small lobes. The stipe in this specimen is 7 mm,
long and 1.5 mm. in diameter. The base of the lamina is narrowly
cuneate, as in the precedhig period, but the blade itself is more lance-
olate, occurring very linearly lanceolate in some specimens. A trace
of the primitive blade (cf. Fig. 5, E) still lingers in some specimens.
The cryptostomata are very abundant in this stage.

As the plants increase in size the rhizogen continues to grow. It
increases in diameter proportionally with the stipe. The lobes, scarcely
noticeable at the beginning, grow out into finger-like processes (cf.
Fig. 6, F), which extend out at first nearly horizontally, and then
more and more obliquely downwards. Figure 6 represents a specimen

OF AHTS AND SCIKNCES. 185

in this stajre ; its leuszth is 13 cm. The rhizo<fen is 5 mm. from the
primitive cli.sk, wliich is slightly larger than in the last specimen
described, but not otherwise different. The rhizogen itself is about
4 mm. in diameter. It has several lobes of some length, which are
already started downward. The blade presents no particular differ-
ences in form, but in specimens of this stage it is very rare indeed to
find any trace of the primitive blade remaining.

In somewhat older specimens the finger-like projections have reached
the substratum, have flattened their tips against it, and have become
attached to it. There are usually five projections, and they form a
whorl of attaching organs, which constitute the second holdfast. They
are called bapteres by most recent writers,* after the term used by
AYarming f to designate very similar organs in the Poduslemacece.
The hapteres form a regular whorl (cf. Figs. 7 and 9), and their
points of attachment form a circle about the point of attachment of
the primitive disk (cf. Fig. 8). The development of the hapteres
marks the end of the first stage of what is to be the permanent
holdfast of the plant, and confirms De la Pylaie's statement quoted
above.

3. Development of the Second Hapteres. — In the next series of
forms, the second and last stage of the permanent holdfast is formed
in a manner essentially the same as that in which the first stage was
developed. On the rhizogen above the primary whorl a new whorl of
protuberances appears. These protuberances are in most cases regu-
larly alternate with the hapteres of the first whorl, but may at times
be so placed that one or more of this set are situated above similarly
placed hapteres of the first whorl. While the new hapteres are
developing, the older hapteres increase in length, and on this account
come to occupy a more upright position (cf. Fig. 7 with Fig. 9).
The new hapteres, therefore, have the opportunity of growing down
obliquely, and of attaching themselves to the substratum, after the
manner of those of the first set. Their points of attachment form a
second but rather irregidar circle outside the one made by the ends
of the primary hapteres. We now have three sets of attaching
organs : first, the primitive disk ; second, the primary whorl of hap-
teres, all of which are usually present at this stage ; and third, the
secondary whorl of hapteres. It often happens that some of the hap-
teres of the second set remain as mere protuberances after the rest of

* Foslie, Die Lam. Norwegens, pp. 6 and 7, 1884. Barber, Antials of Botany,
Vol. 111. p. 48.

t Bot. Zeit., Bd. XLI. pp. 193 and 197, 1883.

186 PROCEEDINGS OF THE AMERICAN ACADEMY

the members of the whorl have reached the substratum and have
become attached. This generally occurs at a place where two of
the hapteres of the first set remain close together. In such a case
the secondary hapteres may never develop at all, and their place
may be permanently supplied by members of the first whorl. But, as
a general rule, all the hapteres of this second set reach their full
development, and constitute the important part of the permanent hold-
fast. The smallest specimen examined which shows the protuberances
that are to grow out into the secondary hapteres is 16 cm. long, while
one that has reached a leugth of 30.5 cm. has most of the second set
attached, but many of them only slightly. In this specimen the primi-
tive disk has entirely disappeared, and there are five hapteres of the
first set, which is apparently the original number. Of the second set
only two are firmly attached ; three are just touching the substratum ;
and there are two more just showing as small protuberances from
the disk of the rhizogen. The stipe of this specimen is 5 cm. long
and 3 mm. in diameter at the base. Below, it is cylindrical, but in
the upper portion it is noticeably flattened. Even in a specimen of
this size the stipe expands gradually into the blade, which in what
seems to be the most common form in New England is broader in pro-
portion to the length than in previous stages. This blade is elliptical-
lanceolate, 25.5 cm. long and 4.8 cm. wide in the broadest part, which
is about half the length from the base. Tlie cryptostomata are very
abundant and very conspicuous.

All of the previously mentioned forms were obtained from the pools
at Nahant in the gatherings made on April 24 and 25, 1889. The two
or three largest and most developed specimens came from below low-
water mark, while the younger specimens came for the most part from
the uppermost pools. On May 12, some specimens collected from the
higher pools were about as large as the largest specimens of the first
collection.

On May 25, a few specimens were collected at Peak's Island,
Maine, the measurements of some of which are given in the following
table : —

Length of Stipe.

Length of Blade.

Width of Blade

0.

cm.

cm.

cm.

1 . . .

3.2

15.9

1.6

2 . . .

5.1

26.9

1.4

3 . . .

. 4.45

26.7

1.6

4 . . .

10.8

55.9

1.9

5 . . .

10.2

35.6

1.9

6 . . .

10.2

99.1

3.8

OP ARTS AND SCIENCES.

187

Jfo.

Length of Stipe.

Length of Uludo.

Width of Blade

cm.

cm.

cm.

7 .

. . . 11.4

85.1

3.2

8 .

... 9.6

85.1

6.1

9 .

... 9.6

66.5

8.5

10 .

. . . 15.2

85.1

3.2

11 .

. . . 12.7

76.2

3.3

As can be seen from the above table, which represent specimens
taken at random, all the plants of this collection are remarkably long
ami narrow. The blade of No. 1 was broken off short, as if by
some accident. The tips of the rest were eroded in the customary
fashion. Nos. 1 to 9 were found growing in a .shallow pool just
above low-water mark, while 10 and 11 were found cast ashore on a
small beach not far away. All of them were of a remarkably light
yellow-brown color, and very pellucid for specimens so long as some
of them were. About thirty specimens of much the same character
were collected at this locality on the same date. All sorts of transi-
tions between this narrow form and the more ordinary form are
found. I have seen plants of this narrow form at Nahant and Marble-
head also, but they were few or single, and this is the only time that
I ever found so many together.

Fourth Period. — 1. Transitional Forms. — Some specimens were
collected at Nahant, June 12, 1889, which show some very striking
peculiarities. The specimens, between twenty and thirty of which
were carefully examined at the time, were all about the same size,
and apparently of about the same age. They grew not only just
below low-water mark, but in the highest pools as well. They were
all in the neighborhood of 55 cm, long. The holdfast was well de-
veloped, and the hapteres were stout and of a dark brown color. The
stipe was decidedly flattened above, and from 6 to 10 cm. long. At the
top where it passes over into the blade there is a noticeable ridge
made by the abrupt transition from the thick stipe to the much thinner
blade. The stipe has become a very dark brown, and is opaque, hav-
ing lost entirely the translucent appearance characteristic of the organs
of the younger plant. But the most noticeable and interesting change
is one which is to be seen in the blade itself. In the upper part of
this the color is still decidedly yellowish, and the cryptostomata are
abundant and conspicuous ; but toward the ba.se there is a noticeable
transverse constriction (cf. Fig. 11, A), and below this the blade is of
a dark brown, agreeing more nearly with the stipe in color, and almost
if not entirely destitute of cryptostomata. Tlie cryptostomata grow
smaller and smaller as they approach the constriction and disappear at

188 PROCEEDINGS OF THE AMERICAN ACADEMY

a small distance above it. Below the constriction not a cryptostoma
is to be seen. The outline in the figure just referred to above was
traced from the plant itself while fresh, and is to be compared with
Figure 10, in which the permanent oar-shaped base of the adult
plant is represented.

As the blade grows older and the part below the constriction in-
creases in size, the constriction gradually disappears. In a specimen
of this same collection, 73 cm. long, the last traces of the constriction
are to be seen at about 10 cm. above the base of the blade. The base
of the blade now becomes more blunt, and begins to take on the shape
of the base of the blade of an oar, which is very characteristic of the
adult plant. The distinction between stipe and blade becomes for
the first time sharply marked.

This process resembles in almost all respects the process of renewing
the blade so well known now in the European species of Laminai'ia*
and seen by the writer in most of the species of Laminaria and
Agarum of New England. But in this process the old blade, which
is thick and is fruiting or has fruited, is cast off, and is succeeded by a
new blade of more delicate consistency. In Saccorldza dermatodea, on
the contrary, as shown more fully below, the new blade is thicker and
denser, and destined to bear the reproductive organs.

It is about this time in its life-history, also, that the plant grows
most actively and reaches its maximum of size. The largest speci-
mens I have ever found were growing at Peak's Island, June 28, 1888.
The largest specimen of my collection of that date measures 112 cm.
when dried. The stipe is 1G.5 cm. long, and the blade reaches a width
of 12 cm. Hundreds of plants of this size, and even larger, were
growing in that locality at that time.

Farlow says in his New England Algae, f that the stipe some-
times reaches a length of two feet, and the blade a length of six feet.
The figures are probably taken from Eastport si:)eciraens, where the
species is said to grow more luxuriantly than it does farther south. A
specimen in the herbarium of Mr. F. S. Collins of Maiden, Mass.,
collected at Eastport in August, was in this same stage of growth.
The stipe measured GG cm. The blade was short and fragmentary,
but measured 21.4 cm. in width.

In these large specimens the blade is very thin and papery on dry-
ing. The color is still a rather yellowish brown, and the cryptostomata

* Cf. Le Jolis, Examen I., p. 549, 1856; Foslie, Die Lam. Norwegens, p. 23;
and earlier writers.
t Page 95.

OF ARTS AND SCIENCES.

189

on the upper part are more numerous and more conspicuous than at
any other stage. They are larger in these fronds than at any time
previous, as will be noted below. In the basal third or half of the
blade, however, the cryptostomata are entirely wanting. The color
of this portion is darker, and it is thicker and more leather}'.

The stipe is more flattened at the top, and the holdfast is larger and
firmer. Occasionally one is found in which one or more of the hap-
teres is branched. The branching, however, is only a simple forking
of the haptere, and may perhaps be regarded as indicating two hap-
teres, fused together above but free below. This brandling, as far
as my experience goes, is decidedly rare.

At about this period, and in some cases also somewhat earlier,
the blade splits vertically into two or more divisions. Two segments
seems to be the rule in the narrower blades, but in the wider ones there
are generally three. I have seen a number of specimens where there
were four, and in the herbarium, of Mr. Collins I have examined one
in which there were five. The splits are continued almost to the very
base of the frond, and the segments are more or less strap-shaped.
The old fronds split very readily in this vertical direction, and one
often finds them cast ashore in the springtime split down to the hold-
fast itself. In Mr. Collins's herbarium there is a specimen which had
been split down thus ftir when young, and each half had proceeded to
continue growing, and there are two more or less perfect plants in that
case from the same holdfast. It reminds one of the famous bifurcate
specimens of Laminaria digitata, of which a summary is given by
Le Jolis.*

2. Adult. — As the season progresses, the plant ceases to grow so
actively as at first, and the tip wears away faster than it is renewed.
Consequently, the plant becomes shorter as the summer advances, but
it also grows thicker and denser. I have unfortunately no specimens
collected early in July, but through the kindness of Dr. G. H. Parker
of Harvard University I have some excellent specimens collected late
in July. These are shorter than the earlier specimens, are of a dark
brown color, and very leathery. They show no traces of ciypto-
stomata. In Se[)tember I collected very similar specimens at Nahant
with zoosporangia just forming. In October, specimens with well
formed zoosporangia were not uncommon at the same place. Fruit-
ing specimens of this species may be found from October on until
May of the next year.

* Examen I., p. 540.

190 PROCEEDINGS OF THE AMERICAN ACADExMY

Kjellraann says * that the proper season for the formation of zo-
ospores in this species is in September or the beginning of October,
on the Norwegian coast. At Spitzbergen he found specimens with
zoosporangia in July and August, and at Nova Zembla in July.
There is apparently in the more northern regions an earlier ripening
of these bodies.

The fruit occupies about the lower half of the blade in the well de-
veloped specimens. The collection of sori, or cloud-like patches of
fruit, covers the surface on both sides at the very base of the frond,
but separates above into irregular-shaped blotches of considerable size.
When the water has evaporated from the surface of the frond, these
Sori have a peculiarly soft and velvety appearance, and are of a some-
what litihter brown color than when wet. When dried, the mature
plants are of a decidedly black color, but tlie sori are even blacker
than the frond, and may thus be detected readily even in herbarium
specimens.

De la Pylaie f describes the fruit as entirely covering both surfaces
of the frond, and as occurring toward the end of autumn. AreschougJ
says : " Fructificatio . . . non format maculas extensas 1. fascias, nt
in Laminariis, sed maculas minutissimas fere punctiformes, nunc sepa-
ratas, nunc approximatas, quo fit, ut cum siccata est, lamina farina
nigrescente conspurcata videatur." The specimens from the New
England coast do not agree with this description, as may be seen from
what has been said above. §

Specimens are cast ashore in the greatest numbers from November
until May. I have generally found them most abundant in February.
A specimen thrown ashore at Lynn, Mass., February 9, 1890, was in
very perfect condition, and so will serve as a good illustration of these
very old plants. The lower part of this specimen is represented of
the natural size in Figure 10. It measured 70.5 cm. in length, and
the holdfast was about 2 cm. in diameter. The primitive disk was still
present, although mucli shrivelled and wrinkled. Of the first set of
hapteres three were present in a good state of preservation on one
side, and the places where the other two had been were to be seen on
the other. There were nine hapteres present in the outside set. All
the hapteres were unbranched. The stipe was 12.1 cm. long. It was
cylindrical at the base and flattened at the top. The blade had the

* Arct. Algr., p. 225. t Fl. Terre Neuve, p. 49.

t Obs. Phvc, Part. III. p. 12.

§ Cf. also kjelhnan, Spetsb. Tliall., II., p. 14, 1877.

OF ARTS AND SCIENCES.

lOl

oar-shaped base characteristic of these old specimens, and expanded
gradually until it reached the truncated apex. It was 58.4 cm. long,
and 12.7 cm. wide at the top.

A curious thing to be observed in tlie specimens cast ashore is that
the great majority of them lack the holdfast, although the attachment
by this organ does not seem to be a very firm one ; for if one catches
hold of a growing plant, even by the tip of the blade, and gives it a
strong and sudden pull, it almost always comes away from the sub-
stratum uninjured, olfering in this respect a strong contrast to the
ordinary species of the Laminar iece. But the majority of the old
specimens cast ashore in the spring look as if they had been twisted
off at the lower end of the stipe, and occasionally a bunch of fionds
will be found cast up, all twisted togdther in a most complicated
manner.

This species appears to be an aniuial. De la Pylaie * and Ares-
choug t were also of this opinion, and all the data that are at hand
seem to support it. On the coast of Massachusetts I have never seen
an adult plant after May (when old, decaying, fructified specimens
were cast ashore) until September. The plants growing in the pools
become very battered and torn, and covered with parasites in De-
cember and January ; and the adult plants found in the springtime are
always very old and about to decay. There is, besides, no indication
in the structure that the plant is either biennial or perennial.

Summary. — From what has been said above, it will appear that
each of the four periods into which the life-history has been divided
for convenience has its characteristic features.

The lack of suitable material, and the small size of the germinating
plants, prevent any satisfactory discussion here of the changes taking
place in the first period and the early portion of the second. Toward
the end of the second period, however, the young plant is provided
with a distinct disk for attachment, a short cylindrical stipe, and a
blade which for the most part is more than one layer of cells thick.

The third period has for its characteristic feature the development
of the permanent holdfast. A small swelling, the rhizogen, appears
a short distance above the primitive disk, increases in size, and pro-
duces two whorls of short fibres, the hapteres, one above the other.
The hapteres attach themselves to the substratum by their tips.

The fourth period extends to the death of the plant. In all the
preceding stages the base of the frond has been narrowly cuneate, and

* Fl. Terre Neuve, p. 49.

t Obs. Phyc, Part. III. p. 12.

192 PROCEEDINGS OF THE AMERICAN ACADEMY

it has been difficult to determine either where the stipe ended or where
the blade began ; but at the beginning of this period occurs the pro-
cess compared above to tlie process of the renewing of the blade in
other Laminarieae, after which the boundary between stipe and blade
is well marked, and the blade itself becomes thicker and firmer, and
ceases to possess cryptostomata. When this set of changes has fairly
ended, the zoosporangia are produced, and the plant, having served its
purposes, dies, and gives place to a new generation.

Histology.

Having thus considered with some detail the various morphological
changes which Saccorhiza dermntodea undergoes in the course of its
life-history, we are in a position to turn to the histological changes
which take place during the same period.

Literature. — The histology of this species has been seldom referred
to. De la Pylaie's account* does not express to us any very definite
idea. Areschoug t has described the structure of the cryptostomata
and of the organs of fructification. KjellmannJ has spoken of the
structure of the stipe, and especially mentioned the " very long tubular
cells, with very thick walls," found in this organ in this species, and
lacking in the Pliyllaria lorea found by him at Spitzbergen and Nova
Zembla. Barber, as mentioned above, has described the histological
changes occurring in connection with the development of the bulb in
Saccorhiza bulbosa, De la PyL, and has thus furnished a valuable
means for comparing these two species.

The account of the histology will be arranged according to the
periods of development sketched out above.

First Period. — The first period begins with the zoospore and its
germination to form a filament of a single row of cells, and extends to
the formation of a membrane of a single layer of cells in thickness.
No plants of Saccorhiza dermatodea belonging to this period have
been seen, but evidences of the existence of these early stages are
present in the youngest plants known.

As far as I know, Thuret § is the only one who has described or
figured either the zoospores or germinating plants of any species of
the Laminarieee. He figures the entire development of Saccorhiza
bulbosa II for this period, as well as the early stages of Laminaria

* Fl. Terre Nenve, pp. 50 to 52. t Obs. Pliyc, Part. III. p. 12.

X Arct. Alg., pp. 224 and 225.

§ Ann. Sci. Nat., Ser. 3, Tom. XIV. PI. 29 and 30, 1850.

II Loc. cit., PI. 30, Figs. 5 to 10.

0? ARTS AND SCIENCES.

193

saccharina.* Kiitzingf figures and describes young plants of Lami-
itaria saccharina consisting of a simple membrane. I have seen a
young plant, probably belonging to Laminaria PhylUlis, (Stackh.)
J. Ag., which consisted of a simple filament.

It seems fairly certain, then, that the germinating zoospore of the
LaminariecB gives rise first to a simple filament, and later to a simple
membrane ; and it seems very probable also that the earliest stages
of Saccorlnza dermatodea will be found to correspond very closely
to those known for Saccorhiza bidbosa.

Second Period. — The preceding period is doubtless very short, and
changes soon begin which convert the simple membrane into a mem-
branaceous frond of more complex structure. The changes take place
in the early portion of what has been marked off as the second period
of the development. The youngest specimens m my possession show
these changes already well advanced, and have been partially de-
scribed above in the account of the morphology. They are of very
delicate consistency, and are therefore difficult to manipulate. Fur-
thermore, there were only two or three of the very small ones, and
on these accounts many of the details of the histology here given had
to be derived from somewhat older plants, belonging however to the
same period.

When one of the youngest plants is examined under a low power
(150 diameters), it is seen that the stipe and the midrib-like portion
of the blade have already arrived at some degree of complexity of
structure. The outline of this midrib portion, as mentioned above, is
irregular (cf. Figs. 1 and 2). In Figure 1 it is seen to have several
tooth-like projections from the main portion. One of these (labelled A)
is situated at the tip. Somewhat below, at A', there is another one,
and at A" still another. These are only more prominent than the
others at A'", and so on, but do not differ in kind. At the very
tip the blade is eroded, and consequently we find here, as has been
described for other LaminariecB, that the growth is not apical, but
from below, and that the apex of the blade is the oldest portion. As
a still further evidence of this, it can be seen that the apical portion of
the blade is of simpler structure, and becomes more complex down-
ward. This is exactly what we should expect from what we know
from other members of the group. Consequently, a set of transverse
sections of one of the plants, arranged in a series from above down-

* Ann. Sci. Nat, Ser. 3, Tom. XIV. PI. 30, Figs. 1-4.
t Phyc. Gen., p. 345, Taf. 24, I, Fig. 5, 1843.
VOL. XXVI. (n. s. xviii.) 13

194 PROCEEDINGS OF THE AMERICAN ACADEMY

wards, will throw a considerable liglit on the changes of structure
which it has undergone in its previous development. In such a series,
projecting portions, such as A', etc., are of significance m showing
that the growth in complexity from the simple membrane did not
originate from one point and extend out regularly, but originated at
several points irregularly situated. A comparison between Figures 1
and 2 will also show this.

To get these sections some series were made by hand with the
razor, and more complete ones were cut on a microtome from speci-
mens which had been embedded in paraffine. Some of the latter
were from specimens which had been treated with corrosive sub-
limate. They were stained with Kleinenberg's haematoxylin before
being embedded, and mounted in balsam after they were cut. Some
of the series gave very good results, others were very much shrivelled.

The upper portion of the blade, or the primitive blade as it is con-
venient to term it, is but a single layer of cells thick, as appears in
sections as well as by focusing. As seen on the flat surface of the
blade the cells are nearly square or slightly rectangular, but in some
cases elongated in the direction of the long axis of the blade. In
transverse section they appear nearly as thick as broad, but they vary
much in this respect. The cells contain a number of large and con-
spicuous lenticular chromatophores, usually arranged about the inner
walls. The protoplasm in these cells is in the form of a thin film
lining the wall. The nucleus is nofvery conspicuous, and is hard to de-
monstrate. These cells possess all the characteristics of age, and have
probably ceased to take any active part in the growth of the plant.

On the one-layered portion were found in some of the specimens
the small clusters of hairs already alluded to. Attempts to get a g^od
section through such a bunch were fruitless. As far as could be seen
by focusing, the hairs protruded only on one surface, and were gener-
ally five or six in number. AVhen viewed from the other side from
that on which they were situated, they were seen to arise from basal
cells formed from the divisions from one (or at times from two) of the
ordinary cells of this portion of the blade. The basal cells of the hairs
are almost isodiametric, are densely filled with protoplasm and chro-
matophores, and possess a single large nucleus. Farther up, the cells
become elongated in the direction of the long axis of the filament ; the
protoplasm becomes filled with large vacuoles; and the nucleus be-
comes central, and connected with the peripheral protoplasm by fine
threads. At the distal end the cells are four to six times as long as
broad, and are destitute of contents. These hairs accordingly have

OF ARTS AND SCIENCES.

195

the region of active growth at the base, as do most of the hairs iu the
Pha:ospure(e.

In the series of sections the first evidence of an increase in com-
plexity is shown in Figure 12. In this transverse section it will be
seen that two cells, at A and A, have divided by walls parallel to the
surface of the frond, and that these two cells are not situated side
by side, but are separated by one cell. Sections from other plants
represent the initial cells somewhat differently. In one case there was
a single one, and in another two were side by side. A section from
the same series as the one represented in Figure 12, and about 30 /x
below it, is represented in Figure 13. In this the process has gone
further, and the central portion of the blade is two-layered, while the
edges are only one-layered. A longitudinal section of a young blade
showed that in that particular specimen the two-layered portion was
five cells in height. Soon we come to a section such as that repre-
sented in Figure 14, where there is a third cell of different shape
and appearance in the centre, and the two layers are separated from
each other in the middle to make room for it. This cell lacks the
chromatophores, is round in cross-section, and has a lining of dense
protoplasm, in which are embedded numerous small nuclei, which take
on a deep stain with ha^raatoxylin.

In the same figure will be noticed two groups of hairs, one project-
ing from each surface. E)ach group seems to have arisen in this case
from the divisions of a single cell. • In the several-layered blade they
are frequently opposite in this fashion.

Figure 15 is drawn much less magnified, and the cells are all repre-
sented on one side to show the proportions existing between the many-
layered and the one-layered portions of the blade. It will be seen
here that there are three of the central cells in this section. In
longitudinal section these central cells appear as elongated tubes, the
ends of which are not readily found. They will be spoken of as
tubes, or tubular cells, for the present.

Figure 16 shows a section still lower in the original series. The
many-layered portion is wider, and the cells appear irregularly placed ;
but the arrangement is seen better in Figures 17 and 18, still lower
down. The tubular cells are readily seen at a and A. The many-
layered blade, about half of which is represented in the figures, has
increased chiefly \\\ width. At B in Figure 17 is seen a group of four
cells, situated apart in the one-layered portion, which recalls the state
of things represented in Figure 12, and looks as if the cells were the
initial cells of a new many-layered portion. Such is the case in fact.

196 PROCEEDINGS 0? THE AMERICAN ACADEMY

This is the origin of the tip of sucli a portion as that labelled A" in
Figure 1. At B in Figure 18, about 30 /x below the section figured
in the preceding, this group of four cells has increased by dividing to
form six. The divisions are repeated, and this portion grows vpider
until the intervenijig space is traversed and it unites with the main
part.

This fact is significant in indicating that there is no one cell, nor
even one region, whence all growth proceeds, but several cells and
regions, and these somewhat irregularly placed.

As we go down, the outer cells, one by one and at first somewhat
irregularly, divide into two by a wall parallel with the surface of the
frond, until at length we have a layer of smaller cells on each side, a
layer of large cells adjoining each of these, and the tubes in the mid-
dle. The outside layers on each side may be called the limiting layers
(L in Figs. 19 and 20) ; the next layers, of larger and more irregular
cells on each side, may be spoken of as the cortical layers (C in
Figs. 19 and 20); and the tubes alone represent what later demands
attention as the medulla.

The lower portion of the blade does not increase in complexity for
some distance, but widens until the whole width of the blade is occu-
pied by the many-layered portion. But in the region where the blade
narrows and passes gradually over into the stipe, in what has been
spoken of above as the transition place, a new set of structures arises,
which are intimately concerned with the thickening of the blade and
stipe, and the increase in the complexity of the tissues. The fact
that this set of changes originates in this portion shows that the prin-
cipal meristematic region is situated here in this species, as it is in
others of the same group.

At the transition place in such specimens the limiting layer is dis-
tinct, and consists of rather small cells of nearly equal dimensions in
the three principal directions. The cortex has two, or even three,
layers of two more or less distinct kinds of cells. Those of the outer
layer are like the cells of the simple cortical layer. The inner cells
appear in transverse section much like those of the outer layer, except
that they are more rounded and lie more separated both from one an-
other and from the outer layer, and in the middle layer among the
tubes. In longitudinal sections, however, they are seen to be much
more elongated than the outer, some of them being five to six times
as long as broad. From them, too. are given off short hyphal branches,
which extend out laterally from the cells, and protrude into the mi'ldie
portion. It was impossible to see the mode of origin of these hyphal

OF ARTS AND SCIENCES.

197

branches in the one or two specimens of this period which showed
them, and so a more complete account of them is left until the next
period is discussed.

Tiie stipe, towards the end of this period, hecomes a rather solid
structure, composed of the following tissues. Outside is a limitiii*.'
layer similar to that described for the blade, but vvhicli is in a state of
more active growth. The cells composing it are somewhat elongated
radially, and divide actively by tangential walls. By divisions in the
two directions perpendicular to this, it also keeps pace with the growth
of the stipe in circumference and length. Within this are several
layers of cortical cells, and the centre is occupied by a mass of
elongated tubes. No liyphal cells are present except at the transi-
tion place.

All except one of the younger specimens had lost the point of at-
tachment. This individual is 6 mm. high. The base is decidedly
bulbous, and is 18 mm. wide, while the stipe above is about 0.1 mm.
wide. The bulbous base is about 0.13 mm. high. At the bottom it
is attached by a number of hair-like bodies, which it seems proper to
term rhizoids. These are comparatively numerous (probably between
50 and GO), and arise as prolongations from the cells of the limiting
layer of the basal portion of the enlargement. The rhizoids them-
selves are irregularly cylindrical, and at this stage seem to be for the
most part unicellular. Toward the distal end the shape is more ir-
regular, the wall thicker and wrinkled, and at this portion too they
are not infrequently and somewhat irregularly branched. The branches
are very short and the tips of the rhizoids are somwhat swollen. This
specimen was growing among a mass of other young plants and fine
filamentous forms of other species, and this allowed the rhizoids to be
seen to greater advantage.

The above description is to a certain extent fragmentary, but much
of that is due to the character of the available material. That previ-
ous to this period there exist stages in which the frond consists of a
simple membrane of a single layer of cells seems to be evidenced by
the fact that, in the youngest specimen seen, the blade is for the greater
part in that condition. It is supported, too, by what we know of the
early stages of some nearly related species. What the stages are by
which the development proceeds from this simple frond to such a stage
as that represented by the youngest specimen described is largely a
matter of conjecture. That the complexity comes about by a series
of definite and gradual changes seems probable. It is supported,
moreover, by what is desi-ribed above. For in the blade we find at

198 PROCEEDINGS OF THE AMERICAN ACADEMY

the top, which is the oldest portion, the simplest structure of a single
layer of cells. It then becomes two-layered, three-layered, four-
layered, five-layered, and finally still more complex by the growth of
the hyphal branches.

That the stipe develops somewhat differently is evidenced, not only
by its different structure, but also by its different shape. In the ab-
sence of a series of earlier stages to indicate the way in which the
complexity came about, we have almost nothing to help us ; for in a
specimen only 6 mm. long the stipe is 12 or 13 cells in diameter. We
get a slight clue, however, from a young specimen of Laminaria
Phyllitis collected at Nahaut. In this the stipe consists of a central
vertical row of cells and a cylinder of cells about it only one layer
thick, so that a transverse section is only three cells across. This
seems to indicate that changes to produce the stipe have followed
something of the same course that has been followed in the blade,
but have been modified in such a way as to produce a cylindrical
rather than a flattened structure.

Kiitzing * figures some details of the development of Laminaria
saccharina during this second period. Reinket also mentions the
partially one-layered stage of the same species, and gives a rough
sketch of it.

Third Period. — The most important feature of the third period, as
marked out above, is the development of the rhizogen and the two
whorls of hapteres. At the same time, the blade and stipe increase
greatly both in size and complexity. The early stages are almost
exactly like the later stages of the previous period, except for the
development of the swelling which marks the appearance of the
rhizogen. It will be best, therefore, to describe first the changes in
structure which accompany this.

A longitudinal section through a very young rhizogen shows that
the central portion consists wholly of elongated cells arranged in ver-
tical rows. There are no long tubular cells present in tins portion.
The changes which result in the swelling begin at the centre and pro-
ceed outwards. The very central cells increase greatly in diameter,
and are followed in this by the outer rows. They tend also to in-
crease in length, but, being hemmed in on all sides, soon have the
appearance of being cramped. This process, thus taking place in a
certain definite region of the stipe, causes the sides to bulge outwards,

* Phyc. Gen., p 345, Taf 24, I, Figs. 1 to 6,

t Pringsh. Jahrb.,Bd. X. p. 375, Taf. XXVII. Fig. 18, 1876.

OF ARTS AND SCIENCES. 199

and the swelling called the rhizogen results. It is produced by the
great increase in the size of the cells, not by any large increase in the
number of cells, as Barber* has already described for the same organ
in Saccorhiza bulbosa.

Not only do the inner cells of the rhizogen enlarge, but the walls
thicken and separate from one another, except at certain points on
each face, wiiere they still remain united. The space between the
separated walls is filled for some time with a mucilaginous modifica-
tion of the outer layers of the cell wall (cellulose) ; but after a while
it is impossible to detect anything of this kind, and the substance
probably becomes very watery indeed. However that may be, the
spaces between the walls finally become filled with small cells in lines
and rows, the origin of which is not readily determined. They ap-
pear to bud out from the larger cells in the form of small hyphae,
which, however, become so crowded that they seem to form a compact
tissue almost at once, and to lose all trace of their hyphal structure.
(Cf. Fig. 21,^..)

The limiting layer has continued growing, and at certain places
along the equator of the rhizogen, in company with the underlying
cortical cells, it takes on a much greater activity than it does else-
where. The increase at these portions is in the radial direction, and
is brought about both by the tangential division of the cells of the
limiting layer and by the increase of the cortical cells in the radial
diameter. As a result several protuberances are formed in the zone
of the equator destined to push out into hapteres. The cortical cells
increase to from six to eight times their original radial diameter, and,
new cortical cells being supplied by the activity of the limiting layer
at the distal end of each protuberance, the haptere begins to elongate.
The limiting layer is more active at the summit of a haptere than
along its sides ; in fact, very few new cells are formed in the latter
region at first. Later, however, when it has finished its growth in
length, these conditions are reversed, and its growth is then in thick-
ness. When the haptere has reached the substratum in its outward
and downward growth, the cells of the limiting layer along the surface
in contact with the substratum grow out into rhizoids for attachment
like those mentioned in the case of the primitive disk.

The second set of hapteres arises after the first set is fully formed.
They develop in exactly the fashion that the members of the first
set did.

* Annals of Botany, Vol III p. 55.

200 PROCEEDINGS OF THE AMERICAN ACADEMY

This brief sketch shows how the rhizogen with its two sets of hap-
teres develops. They do not increase iu complexity to any extent
during this period. But while this has been taking place there have
been important changes in the other organs of the plant.

If we start at the tip of the blade of a specimen which has the first
set of hapteres well developed, we shall find it to possess a well defined
limiting layer on each surface, and within this, on both sides, a cortical
layer of a single series of cells, and in the centre the medulla, repre-
sented by a few scattered tubes. As we approach the base, we find
the limiting layers more and more sharply defined. Their cells are
smaller and more elongated radially, and give evidence of greater
activity. Within these the cortical layers are more developed, and
are to be distinguished mto an inner and an outer layer. The outer
layer consists of smaller pai'enchymatous cells, while the inner con-
tains laiger elongated cells. We have, too, at this point a distinct
medulla, m which the long tubular cells form a conspicuous element.
Surrounding them are elongated tube-like cells of an appearance in-
termediate between that of the original tubes and the elongated inner
cells of the inner cortex. Their walls are thin, while the walls of the
original tubes are now noticeably thickened, and they are in general
destitute of the dense contents of the latter. They have attached to
them and entwining about them numbers of hyplial filaments, which
usually have a lateral or oblique direction of growth.

If one examines the mnermost rows of cells of the inner cortex in
transverse, or better in longitudinal section, he will see whence these
hyphffi arise. As the cells of the inner cortex elongate, the middle
layer of the common wall becomes swollen, and the cells become
finally separated from one another, except at a few points on each
face. As the layer continues to swell and to become indistinct, the
connections at these points become drawn out into short tubes ; and
since the process continues, these may become quite long at times.
The tubes or hyphie may be known as connecting hypha3 (Verbiud-
ungshyphen). In the other LaminariecB, as far as investigated, they
are said to arise only at points on the walls where there are pits, and
this seems to be true of the specimens examined ; at any rate, iu all
places where the point of origin could be plainly seen, this was the
case. In the region where the cells begin to separate, the cell walls
proper are thin, and the pits are not conspicuous. But if the section
be stained with a dilute solution of Boehmer's ha^matoxylin the
appearances shown in Figure 23 are brought out. In this figure it
is seen that at the place where the two cell walls remain in contact

OF ARTS AND SCIENCES. 201

/
there is a thickening which takes on a deeper stain than the com-
paratively thin walls do. In Figure 23, also, one of these thicken-
inofs shown in surface view is seen to be in the form of a rinji. In
cells which have separated more, and have a connecting hypha, there
appears to be just such a thickening at the point in the hypha which
represents the place of contact in the originally adjacent cells.

On the longitudinal walls of these Inner cells of the inner cortex
may also be seen small rounded protuberances which grow out into
hyphoe. These hyph«, which in contrast to the connecting hyphae
may be called free liypha3, grow horizontally inwards, and twine
among the separated elements of the medulla. They sometimes arise
from pits, and probably always do, although this was not to be plainly
seen in all cases. The hyphae when just starting do not seem to have
a thickened ring at the base, but nearly all of the longer ones do.
There seems to be no fundamental distinction between free liyphiB and
connecting hyphte. A connecting hypha may be separated from one
of the two cells, either at the middle or at one end, and, thus becoming
a free hypha, continue its growth. Where, too, the free hyphae origi-
nate as small protuberances, it is probable that at these points had
been situated pits, but that the two cells had nevertheless entirely
separated from each other.

The reason why the cells of the medulla and inner cortex become
separated from one another is not at all apparent. It is suggested
that these cortical cells, growing rapidly in length, become longer than
the corresponding region of more external cells, and so, becoming
somewhat bent, have a tendency to separate except at places where
they are most firmly attached, viz. at the pits. This would seem to
go a step farther in the medulla than it does in the inner cortex. In
the former the cells separate almost entirely from one another, except
as they are united together by the connecting hypha3, and, the ends
being free to slide up or down, the elements can elongate to a much
greater extent than they could when they were members of definite
vertical rows.

We find, then, at the end of this period, four sorts of cells in a
medulla of this stage, viz.: first, the original tubes; second, the tubes
arising from the elongation of the cells separated off from the inner
cortex; third, the connecting hyphoa ; and fourth, the free hyphre.
The further changes in the medulla introduce no new kind of ele-
ments, but are brought about by the addition of new elements of the
last three kinds, and the changes which these undergo.

The development of the cryptostomata may be readily investigated

202 PROCEEDINGS OF THE AMERICAN ACADEMY

in the thicker blades at this stage of growth. A series of sections at
and a little above the transition place will give all of the early stages.
The first indication that a cryptostoma is forming is the appearance of
a shallow saucer-shaped depression. This will be seen to be due to
the fact that the cells of the limiting layer in the region of this de-
pression cease to divide as actively in a tangential direction as they
have been doing, or as actively as their neighbors are doing. Further-
more, the cells of the outer cortex immediately below this point do
not increase in size either so rapidly or to so great an extent as their
neighbors do. So very soon this region is left behind by the more
active growth of the rest of the blade, and forms a slight depression.

Very soon also the cells of the limiting layer at the bottom of this
depression begin to grow out into filaments, and a stage of this is
represented in P'igure 22. In this drawing the filaments or hairs are
only three or four cells in length, and still possess the original rounded
terminal cell. The growth is at the base, ;is may be seen from the
fierure, and as was described in the cases of the clusters of hairs found
on the primitive blade. The oldest cryptostomata of specimens of this
stage are decidedly bowl-shaped, but they lack entirely the prominent
overhangmg margin to be found m older specimens.

Fourth Period. — Transitional Forms. — Having developed the
permanent holdfast, the plant enters the final period of changes.
Early in this, its active vegetative life ends, and it begins to mature
and to ripen. At the very beginning, it starts to throw off the old
blade, which has been, by means of its thinness and abundant crypto-
stomata, so well adapted for obtaining nourishment and for assimilat-
ino- it, and proceeds to develop a frond especially adapted for bearing
the reproductive bodies.

Through an oversight, specimens of the stage showing the renewal
of the blade were not preserved otherwise than as pressed for the
herbarium, and so I am unable to give an account of the stage where
the transition takes place. But the change is brought about by
a thickeninc of the walls and a modification of the contents of the
cells, rather than by any change in the kinds of elements or in their
distril)ution.

As mentioned above, the growth of the blade is very rapid at about
this time, and at the boundary of these two periods the plant attains
its maximum length. The most striking thing, perhaps, about one of
these large plants, such as were collected at Peak's Island and at
Naliant in the latter part of June, is the abundance and size of the
cryptostomata. It is in these plants that they reach their final com-

OF ARTS AND SCIENCES.

203

plexity, for in the new blade just forming they are entirely absent.
Tiiey are saucer- or shallow bowl-shaped depressions, about 0.3 mm.
in diameter, 0.15 mm. deep, with a large number (60 to 70) of hairs
projecting from the bottom. The margin of the saucer has a sort
of rim projecting in over the edge of the depression, such as Ares-
(•lioug* describes for the Norwegian plant.

Adult Forms. — Specimens collected by Dr. Parker at Nahant, on
.July 26, 1889, are shorter than the last, have no cryptostomata, are
of a darker color, and more leathery in consistency. Sections through
the various organs of some of these specimens show the arrangement
of tissues much as described in the last ; but show also that the walls
of all the elements of these tissues have undergone extensive thicken-
ings, and that the elements themselves are somewhat larger. It will
be well to glance, then, at the charges which have thus taken place in
the tissues of a plant more matured, but not yet fruited. Such plants
are common at Nahant in the middle and latter part of September.

a. Limiting Layer. — In these plants the limiting layer is still
present, but is covered with a thick secretion resembling a cuticula.
The latter may be separated readily from the section, and then has a
pitted appearance where it has been parted from the rounded ends of
the cells. The cells of the limiting layer, both in stipe and in blade,
are usually somewhat elongated radially (cf. Fig. 24, L). Tiie very
outside walls and the cuticula give a yellow reaction with chloriodide
of zinc solution, but the inner walls turn a pale bluish. At the base
of most of the cells small cells may be seen in cross-section (cf. Fig.
24, a) which have recently been formed by tangential division. The
cells of the limiting layer always have dense contents and numerous
brown chromatophores, so that the section needs to be particularly thin
to show the cell structure without the use of clearing reagents.

b. Cortical Layers. — The next layers differ in stipe and in blade
in the details. In the blade the cortical layer is not at all distinctly
divided into inner and outer, as seen in cross-section. In a longitudinal
section, however, the innermost cells are elongated and have hjqjhal out-
growths, while the outer ones are shorter and more irregularly placed.
In the stipe the outer cortex appears distinct in a transverse section,
from its cells being somewhat elongated radially and arranged in
radial rows, although it is not as plain in this species as it is in those
of the genus Laminaria. The cells of the inner cortex of the stipe
appear more nearly hexagonal in transverse section, and are not

* Obs. Phyc, Part. III. p. 12.

204 PKOCEEDINGS OP THE AMERICAN ACADEMY

arranged in radial rows. In a longitudinal section of the stipe the
cells of the outer cortex are short and more or less irregularly placed,
while the cells of the inner cortex are much elongated and in vertical
rows. In the blade there are only a couple of layers of cells in each
cortex, but in the stipe there are several layers of cells in each, and
the cells of the layers are well supplied with pits.

The pits among the Laminariaeece are very conspicuous structures,
and deserve esi)ecial study. They are mentioned and described by a
number of writers on the subject, but no satisfactory account of their
development has yet been published. They appear quite early in the
second period of growth, but are not readily seen until reagents are
applied. When the section, however, is treated with chloriodide of
zinc solution, or dilute Boehmer's or Kleinenberg's hasmatoxylin,
they appear as circular uncolored places, and are large in proportion
to the size of the cells on the walls of which they occur. They are
to be found for the most part on the walls of the cells of the inner
cortex and medulla, but are present sparingly on the walls of the
innermost layers of cells of the outer cortex. In the inner cortex of
older specimens the cell walls are very much thickened. There is an
appearance like the middle lamella of wood cells, and the walls them-
selves are plainly stratified. The pit appears to be situated in the
stratified portion, and to consist of two opposite depressions in the
adjacent walls of neighboring cells, wliich are separated at this point
only by a very thin membrane. When treated with chloriodide of
zinc or iodine followed by sulphuric acid, the inner lininjis of the cells
turn rapidly blue, and after a time there is a faint blue coloration
through the whole thickness between two cells. The whole thickness,
too, stains deeply with hcematoxylin.

c. Medulla. — The tubes (or long tubular cells) are a conspicuous
feature in sections of the adult plant, both of blade and of stipe. As
we have seen, they arise very early in the history of the plant. It is
somewhat obscure as to just how the first tube arises. Probably a
middle cell is cut off between the two layers, and has a chance to
elongate before another one is formed ; for, as we have seen, the
tubes in the young plant were few and scattered. Later, however, new
tubes are formed from the inner cells of the inner cortex, and come to
lie with the old ones in the medulla. The walls of these tubular cells
begin to thicken before those of the other cells, and are of consider-
able thickness even in the preceding period. These cells form a laj'er
adjoining those of the inner cortex on either side in the blade, and in
the stipe this layer appears in cross-section as an ellipse of occasion-

OP ARTS AND SCIENCES.

205

ally as many as four to six cells in thickness. The very central mass,
both of the medulla of the stipe and of the medulla of the blade, con-
sists of densely interwoven hyplia3, which have made their way in
from the surrounding cells, and have crowded the tubes out to the
periphery.

In the very old stipes, and to a somewhat less degree in the blades,
the tubes form an important factor in rendering them as tough and
leathery as they are. When such fi'onds are macerated in potash,
the limiting and cortical layers may be removed, and a firm mass
of fibrous tissue left, which is in the form of a fiat plate in the blade,
while that from the stipe is a cylinder hollow or spongy in the
middle. This layer is unbroken at the transition place, which shows
that in the adult plants this region is no longer the seat of active
growth.

It is easy to isolate individual fibres from macerated specimens to
get a better idea of their shape and size. They have, without ex-
ception, a longitudinal course, very nearly vertical, and lie closely
crowded with overlapping ends. They are for the most part simple,
but some are slightly branched at one end.

In a macerated stipe 9.3 cm. long, the longest fibre (or thickened
tubular cell) was 7.5 cm. long, and was terminated only at one end,
the other being broken off short. A whole one obtained from the
same stipe was 7.3 cm. long, and one of more average length was
6.2 cm. long. A ver}' small one measured 3.2 cm. The fibres from
the blade are much shorter, averaging from 3 to 4 cm. in length.

In Figure 25, a, b, c, are represented the tips of various fibres as
seen in optical section. They are of various shapes, and frequently
have spiny projections which are probably traces of the connecting
hyphae with which in their earlier stages they were united with
adjacent cells. In June plants, they have a thickened and stratified
wall and a wide lumen. The contents at this time also line the
wall, and are dense and of a yellow color. But in the specimens
figured, which are from an October plant, the lumen is almost
obliterated. The walls are composed of very distinct layers. The
treatment with potash and a prolonged soaking in water are respon-
sible for some of this, but not for the greater portion. The walls
when treated with chloriodide of zinc solution finally become a pale
watery blue.

These fibres, both as regards shape and arrangement, remind one
of the fibres in the bands of brown tissue so common in tlie stems and
stipes of ferns which are called sclerencfiyma fibres. There seems to

206 PROCEEDINGS OP THE AMERICAN ACADEMY

be no reason why these structures should not be called by the same
name.

Sclerenchyma fibres occur also in Saccorhiza bulbosa, as I have
learned from the examination of specimens in Professor Farlow's
herbarium, but they do not seem to be so regularly parallel as in
S. dermatodea. They have never been studied by any one, so far as I
know. Even the references to them are few. Janczewski* mentions
"cellules tres-longues a membrane epaissie simulant des fibres libe-
rieuues," in " Haligenia,''^ referring of course to Saccorhiza bidbosa.
Kjellman, as I have mentioned above, both describes and figures them
for his Phyllaria dermatodea. He says that they are also present
(i. e. " the long tubular cells ") in Phyllaria lorea, " but their walls are
always thin, not differing in thickness from the walls of the adjoin-
ing parenchymatous cells." t

d. Splitting. — The s[)litting of the frond, although fairly regular
and uniform, yet seems to be of the nature of a wound. In some
specimens this splitting does not occur at all, but the majority of
plants split early. The blade widens, and, being exposed to the buf-
fetings of the waves, naturally tends to split. On account of the struc-
ture of the fibrous masses of tissue, it splits readily in a longitudinal
direction rather than transversely. If the splitting happens when
the plant is still young and actively growing, the limiting and cortical
layers continue to increase, while the medulla, having no growth of
its own, is left behind. Consequently, there is soon a depression or
channel formed on the wounded edge. Under favorable conditions
the sloping sides of this channel continue to grow until they meet and
fuse, and then sections show no trace of the wound. In old speci-
mens, however, these outer layers are not growing, and so, when the
split occurs in such a frond, the edges exposed simply thicken and
harden, and a section always gives evidence of the wound. Almost
all the stages between these two conditions are met with.

e. Fruit. — The last organs to be developed are the reproductive
organs. This occurs at Nahant and Peak's Island in the autumn and
winter. Specimens just beginning to develop sporangia were found
at Nahant in the last part of September, and fruiting specimens are
to be found from that time on until May.

The reproductive organs are borne on the surface of the blaJe in
large patches known as " sori." They are of two kinds, zoosporangia

* Mem. Soc. Nation. Sci. Nat. Cherbourg, Tom. XIX. p. 98 (p. 2 of reprint),
1875.

+ Arct. Alg., pp. 224, 225.

OF ARTS AND SCIENCES. 207

and paraphyses. The September specimens show many details of the
development. The cells of the limiting layer elongate perpendicularly
to the surface of the frond until they are ten to twelve times as long
as broad. These elongated cells are slightly swollen at the tips, and
have granular brown contents and constitute the paraphyses. At ma-
turity, each is cut off from the cell whence it arose by a tangential
wall. The cuticula has been carried up and still covers the tips of
the paraphyses. After they are fully formed, the zoosporangia begin
to appear as small and broad swellings at the bases of the para[)hyses.
They generally arise from the basal cell of a paraphysis; at least, in
all cases where the point of origin could be distinctly determined this
was the case.

In November specimens, the zoosporangia are well developed. The
cuticula has disappeared from over the tops of the paraphyses, and
the sori in consequence have taken on their characteristic velvety
appearance. The paraphyses themselves have a slightly thickened
tip. The zoosporangia are in the shape of obovoid sacs, about two
thirds the length of the paraphyses. The contents are of a light
yellow brown and are already divided up into spheres, the future
zoospores. The tip of the zoosporangium is noticeably thickened.

Farlow* has called attention to the fact that the paraphyses are
destitute of the hyaline appendage of the paraphyses of the species of
Laminaria. Areschoug t has given measurements of both zoospo-
rangia and paraphyses. Kjellman | figures a zoosporangium and a
paraphysis ; but these drawings do not correspond exactly to the zoo-
sporangia and paraphyses of my specimens. Barber § has figured
and described the zoosporangia and paraphyses of Saccorhiza bulbosa,
which seem to agree very well with those of our plant.

General Summary.

I. The development of Saccorhiza dermatodea agrees in its general
features with what is known of the development of other Laminariece.

II. The permanent holdfast is developed from a peculiar organ,
the rhizogen, which produces two successive whorls of hapteres, the
first organ of attachment or primitive disk being only a temporary
structure.

* N. E. Alg., p. 96.

+ Obs. Phyc, Part. III. p 12.

t Spetsb. Thall., II., Taf. I. Fig. 8, 9, 1877.

§ Annals of Botany, Vol. III. p. 58, PI. VI. Fig. 22.

208 PROCEEDINGS OF THE AMERICAN ACADEMY

III. The cryptostomata are present on tlie one-layered portion of
the blade as clusters of hairs upon the fiat surface ; in the more com-
plex portions, they occupy saucer- or bowl-shaped depressions ; and
these finally are provided with a prominent projecting margin, as
shown above.

IV. The medulla is largely made up of hyphse developed in the
same way as in the other LamniariecE.

V. Peculiar sclerenchyma fibres are developed in the medulla both
of the stipe and of the blade.

VI. The paraphyses are destitute of the conspicuous terminal ap-
pendage present in almost all the other Laminariece.

VII. There is present a series of stages in which the maximum of
growth is attained, and this series of stages is more or less sharply
marked off from the adult stages by a process very similar to that
known as " the renewing of the blade " in other Laminariece.

VIII. The adult frond is destitute of cryptostomata.

Relation to Phyllaria lorea, Kjellm. — Kjellraan has made a careful
study of the old Laminaria lorea, Bory,* and has decided that it is a
distinct species, but high Arctic in its distribution. He has given his
reasons for considering it distinct from the Laminaria dermatodea,
De la Pyl., in the " Arctic Algne." Stromfelt f follows Kjellman in
keeping it distinct. I have searched for the species most carefully
among my collections from the New England coast, and have studied
Kjellman's descriptions in order that I might be able to recognize
the plant. The following comparison between my experience with
New England Saccorhizce and Kjellman's with Spitzbergeu and Nor-
wegian specimens may throw some light on the subject.

Beginning with the young plants, Kjellman says that there is a
marked difference between the two species. Young plants of S. der-
matodea have a distinct stipe, a dark brown color, are '' only little
translucent, and provided with very few short-haired cryptostomata."
The other species {Ph. lorea) has only a "very short stipe, passing
without definite limit into a narrow, sometimes almost filiform, linear,
or more usually lanceolate lamina. Their lamina is thin, very light
brown, almost yellowish brown, perfectly pellucid with numerous long-
haired cryptostomata." X

The great majority of the young Saccorhizce to be found along the

* In J Aff., Spec. Alp., Vol. I. p. 130, 1848.

t Algvcg. Isl., pp. 42 and 76, 1886. } Arct. Alg., p. 224.

OP ARTS AND SCIENCES.

209

New England coast answer perfectly to this description of Phi/Uaria
lorea. Only a very few were found which would at all agree with the
description of Ph. derinatodea, and these few always seemed to he
the product of unfavorable circumstances which had dwarfed them.
The shapes too of the young specimens figured in the "Arctic Alg;c"*
agree very well with numerous specimens collected at Nahant. Kjell-
mau further says, that " older specimens of the two species are easily
distinguished from each other by several good characteristics. In Ph.
lorea the stipe collapses in drying, and becomes flat, thin, almost
membranaceous, and brittle ; even in vevy large specimens, it has the
same color as the lamina, being pellucid like this, lu older specimens
of Ph. dermaiodea the stipe is far more solidly built, dark brown,
0[)!ique, flat upwards, but almost terete downwards. "f Of my New
England specimens, very large ones, in fact many of the very largest,
answer almost exactly to this description of Phyllaria lorea. When
dried, both stipe and blade become very thin and membranaceous, and
extremely brittle. The stipe is light yellow-brown still in many cases,
and the cryptostomata are very numerous and long-haired. In such
specimens, too, the tubular cells are thin-walled, and with a very wide
lumen. It is not until very late that these plants have cryptostomata
with a prominent overhanging margin. They have all the characters,
so far as I can see, that distinguish Ph. lorea from Ph. dermatodea.
But these same specimens, as observed in tide pools at Nahant, soon
begin to lose these characteristics. They become thicker and more
solid both in stipe and in blade; they grow darker; they become
shorter ; the distinction between stipe and blade becomes more
marked ; — in fact, they come to have all the marks which serve to
distinguish the Phrjllaria dermatodea of Kjellman. Furthermore, all
the plants of Phyllaria lorea taken by Kjellman were young. Only
one was taken which had zoosporangia in course of development.!

Professor Kjellman has very kindly sent me specimens both of his
Phyllaria lorea and his Ph. dermatodea. The specimens o^ Ph. lorea
agree exactly with ray young specimens, which undoubtedly belong to
Sitccorhiza dirmatodea, while the specimens of Ph. dermatodea corre-
spond in all characters except size to older specimens of the same.
Most of my browner specimens are larger, but, as the species is very
variable in size, this does not seem to me to be of any great importance.
I have little hesitation in believing that the Laminaria lorea, Bory,

* Taf. XXIV. Figs. 1, 2.
X Loc. ciL, p. 227.

VOL. XXVI. (n. S. XVIII )

t A ret. Alg., p. 224.

14

210 PROCEEDINGS OP THE AMERICAN ACADEMY

and the Phyllaria lorea, Kjellm., are younger, and perhaps in certain
cases more delicate, stages of Saccorhiza dermatodea.

In his recent paper on the Alg:e of East Finmark (p. 74), Foslie
describes some young stages of Saccorhiza dermatodea which seem to
agree very closely with those occurring on the New England coast,
and his remarks on these tend to confirm the views expressed above.

Forms of Saccorhiza dermatodea. — Kjellman * describes two forms
of his Phyllaria dermatodea, one of which, f. typica, has abundant
cryptostomata ; and under this he makes a reference to Farlow's
plant described in the New England Algaj. But, as I have shown,
the adult plant has no cryptostomata, which character would tend
to refer it to the other form, arctica, which has no cryptostomata
in the adult condition. But f. arctica has few in the young con-
dition, according to the description. Here, however, 1 may refer to
what I have shown above, that Kjellman's description of the young
Ph. dermatodea does not agree entirely with mine. As the Norwegian
plant distributed by Areschoug agrees well with our New England
specimens, it certainly seems to me that these two varieties or forms
of Kjellman represent merely different stages in the life history of one
plant.

In the paper on the Algae of East Finmark alluded to above, Foslie
describes two new forms of Saccorhiza dermatodea (p. 74). These,
as nearly as can be judged from the description, exist also among
New England specimens ; but there is no constant difference, and
they seem to me to be merely states of the same plant, differing in
shape and color on account of difference in age, or with varying con-
ditions of light, exposure, etc. ¥. lanceolata, as far as measurements
and description indicate, is mostly a younger and more delicate form
than f. oUonga. On the New England coast the matured fruiting
plants are remarkably uniform in size, shape, and color for a species
of the Laminar iacecE, and so I can see no reason for separating our
plant at least into different varieties or forms.

Relation to Saccorhiza bulbosa, De la Pyl. — The resemblance
between S. bulbosa and S. dermatodea is pointed out by De la Pylaie
in his first paper. f The rhizogen and the production of the two
whorls of hapteres are the same in both species. After that the rhizo-
gen of *S'. dermatodea ceases to form hapteres, while that of S. bulbosa
continues to do so indefinitely, and produces that characteristic organ,
" the bulb." The development of the tissues in this organ is, so far as

* Arct. Alg., p. 223. t Ann. Sci. Nat., Ser. 1, Tom. IV. p. 179.

OF ARTS AND SCIENCES.

211

I can ju<l;j:e, epseniially the same in the two specie3. S. bulhosa, too,
is the only other alga, as far as is known, that possesses the peculiar
sclerenehyma fibres. Furthermore, S, bulbosn has paraphyses of the
same peculiar kind with S. dermatodea. Both are annual plants,
while the majority of the Laminariece are perennials. These two
species seem to constitute a distinct and natural genus, in which are
also placed several smaller European species of more or less doubtful
autonomy. Saccorkiza, proposed by De la Pylaie m 1829,* has the
right of priority. H<dt(je)iin, proposed by Decaisne in 1842, f included
»S^. bulbosa only. It was extended in 1856 by Le Jolis t to include
.S*. dermatodea. Le Jolis divided Haligenia into two sections : Sacco-
rkiza, to include S. bulbosa, etc., and Phyllaria, to include S. der-
matodea and others. Phyllaria has since been used by Gobi § and
Kjellman || as an independent genus. Kutzing made S. bulbosa the
type of the genus Phycocastamim in 1843.1[ There seems to be no
need of separating the two species, and consequently they should both
be included under the older name of SaccorJdza.

Relation to the other Laminariece. — The relation of these two
species to the rest of the species of the Laminariece is at once complex
and interesting. The first organ that suggests itself as important in
this discussion is the holdfast. Chorda has in all probability the
simplest holdfast of any member of the group. The holdfast in this
genus appears to be merely the primitive disk expanded. Nothing
like hapteres is produced. Laminaria solidungida seems to repro-
duce the holdfast of Chorda, but on a larger scale. Next comes
Saccorhiza dermatodea, in which the primitive disk does not produce
hapteres, and where only two successive whorls of hapteres are pro-
duced from the rhizogen. Saccorhiza bulbosa differs not only in
producing a greater number of whorls of hapteres from the rhizogen,
and the consequent enlargement of that organ, but also in producing
hapteres from the primitive disk as well.**

In the rest of the Laminariece the holdfast is more complex. In
most cases the hapteres are numerous and branched, and it seems a
difficult task to bring them into homology with these simpler forms.
But they may be separated into two classes : Jirst, those with hapteres

* Fl. Terre Neuve, p. 23.
t Ann. Sci. Nat., Ser. 2, Tom. XVII. p. 330.
t Examen I., p. 591.
§ Alg-Fl. Weiss. Meeres, p. 75, 1878.

II Arct. Alg., p. 223. 1 Phyc. Gen., p. 346.

** Cf. Barber, Annals of Botany, Vol III. p. 46, PI. V Figs. 1, 2, 3, and 6.

212 PROCEEDINGS OP THE AMERICAN ACADEMY

in fairly distinct whorls ; and second, those in which no arrangement
into whorls can be traced.

Laminaria digitata, as limited by the Swedish algologists,* is a
good example of the first of these. The primary disk here produces
a few hapteres. The second set of hapteres is produced in a whorl,
slightly above the region of the primitive disk. As all the hapteres
of this whorl start out at about the same time, there is a bulirinfr of
the stipe in that region, giving it for a very limited period somewhat
the appearance of a rhizogen ; but it does not long continue^ and at
no time is there as definite a structure as the rhizogen of the Sacco-
rhizce. The hapteres pus^h out very early, and a second whorl soon
appears above the first, another above that, and so on. Although
there is no real rhizogen present here, yet this whorled arrangement
and the consequent bulging appear so very similar as to suggest that
the two methods are very closely homologous.

Of the second class, the common Alaria esculenta, as now limited, is
a good example. The primitive disk here produces hapteres, and the
next hHpteres above generally come out one by one, and apparently
irregularly, although at times they seem to form a more or less
regular whorl or two.

As fiir as the holdfiist is concerned, then, Chorda is the simplest
form. Between Chorda and Saccorhiza bidhosa we might expect to
find a simple form with hapteres arising from the primitive disk, but
we do not. The intermediate form is Saccorhiza dermatodea, with a
simple primitive disk and a rliizogen of limited growth. In regard
to the holdfast, too, Saccorhiza dermatodea seems directly related to
-S. bullosa, on the one hand, and the digitate Lamina) ictp, on the
other.

In structure, Chorda is perhaps the simplest : it has no trumpet
hypha?, and few ordinary hypha?. The diaphragms, however, show
a complication not present in other genera. Saccorhiza is of more
complicated structure than Chorda. It lacks the trumpet hyph;e so
characteristic of the other Laminariece, but possesses the peculiar
sclerenchyma fibres in their stead.

Tlie paraphyaes of Saccorhiza are more like those of Chorda than
those of the other Laminariecc, being simpler in their structure.

Saccorhiza dermatodea, then, seems to be of simpler form and struc-
ture than any of the rest of the group except Chorda. It is Arctic
in distribution, and circumpolar. This simplicity of structure and

* Cf. Foslic, Die Lam. Norwegens, p. GO.

OF ARTS AND SCIKNCES.

213

extended distribution seem to indicate that it is a primitive form, and
more nearly related to the primitive ancestor of the Laminariece than
any other known form except Chorda. Chorda, however, needs to
be studied more carefully in order to determine its relationships more
clearly.

Literature.

Below are given as many of the references to Saccorhiza dermalodea
as have come to the knowledge of the author : —

Laminaria dermalodea., De la Pyl., Ann. Sci. Nat., Ser. 1, Tom.
IV. pp. 179, 180, PI. IX. Fig. G, 1824.

Laminaria dermalodea, De la Pyl-, Fl. Terre Neuve, pp. 48 to 52,
1829.

Laminaria Bcerii, Post. & Rupr., 111. Alg., p. ii, 1840 (nomen).*

Laminaria dermalodea, Rupr., Tange Och., pp. 352 and 409, 1847.t

Laminaria dermalodea, J. Ag., Spec. Alg., Vol. I. p. 131 (exclus.
syn. L. Bongardiana)., 1848. J

Laminaria lorea, Bory, in J. Ag., Spec. Alg., Vol. I. p. 130, 1848.

Phyllitis dermalodea, Kiitz., Spec. Alg., p. 567, 1849.

Laminaria dermalodea, Harvey, Ner. Am. Bor., pp. 91 and 92 (ex
parte), 1852.§

Haligenia (Phgllaria) dermalodea, Le Jolis, Comptes Rend us,
Tom. XL. p. 473, 1855.

* Areschoug referred this name as a synonym under his Laminaria lorea, No.
213 of liis "Algae Scandinavicse Exsiccatae," by virtue of a specimen received
from Iluprecht. He makes a furtlier reference to it in the "Observationes
Phycologicae, Particula Tertia" (p. 12). Gobi places it under his Phijllaria
dermalodea in "Die Algen Flora des Weissen Meeres " (p. 75), giving Rosta-
finski as his authority for doing so. Kjellman refers to Gobi as his authoritj'
for quoting this name as a synonym under the Phyllaria dermalodea of his
"Algae of the Arctic Sea" (pp. 223 and 225).

t Le Jolis in Examen II. puts under Haligenia (Phyllaria) dermalodea, "nee
Rupr.," referring probably to this reference.

t Agardh quotes the Laminaria Bongardiana, Post. & Rupr., as a synonym of
Laminaria dermalodea, De la P.vf, in the " Species Algarum," and has thus given
grounds for the assertion (cf. Le Jolis, Examen I., p. 583) that the L. dermalodea
of the " Species " was not the true L. dermalodea, De la Pyl. Agardh, however,
in "De Laminarieis" (p. 83) explains how it was that lie came to quote L. Bon-
gardiana as a synonym of L. dermalodea, but asserts that nevertheless his
reference really includes the true L. dermalodea, De la Pyl.

§ Harvey does not quote L. Bongardiana, Post. & Rupr., as a synonym; but
his description of the old plants (for he says that he had seen only young
specimens) was evidently taken from Agardh, and refers to some form of the
true digitate Laminarioe.

214 PROCEEDINGS OP THE AMERICAN ACADEMY

Haligenia {PUyllaria) dermalodea, Le Jolis, Examen II., p. 312,
1855.*

Haligenia (Saccorkiza) Bcerii, Le Jolis, Examen II., p. 312, 1855.

Haligenia (^Phyllaria) dermatodea, Le Jolis, Exameu I,, p. 591, 1856.

Laminaria Bcerii^ Nyl. & Sael., Herb. Mus. Fenn., p. 73, 1859.

Laminaria deimatodea, Kemp, Caiiad. Nat., Vol. V. p. 34, 18G0.

Laminaria dermutodea, Harv., Proc. Liun. Soc, Vol. VI. pp. 158,
160, 166, and 167, I862.t

Saccorhiza dermatodea, J. Ag., De Lam., pp. 31 to 33, 1867.

Laminaria dermatodea, J. Ag., Spetsb. Alg., p. 8, 1868.

Saccorhiza dermatodea, J. Ag.. Spetsb. Alg. Till., p. 31, 1868.

Laminaria dermatodea, Lawson, Canad. Nat., Vol. V. p. 100, 1870.

Saccorhiza dermatodea, J. Ag., Gronl. Alg., p. 110, 1871 {^fide
Kjellman).

Laminaria dermatodea, J. Ag., Gronl. Lam. Och. Fuc, p. 11, 1872.

Laminaria dermatodea, D. C. Eaton, Traus. Conn. Acad., Vol. II.
p. 343, 1873.

Laminaria dermatodea, Jordan, Canad. Nat., Vol. VIII. p. 402,
1874.

Saccorhiza dermatodea, Aresch., Obs. Pliyc, Part. III. pp. 11 and 12,
1875.

Laminaria dermatodea, Farlow, Proc. Am. Acad., Vol. II. p. 355,
1875.

Laminaria dermatodea, Farlow, Report U. S. Fish Comm. for 1875,
p. 707, 1876.

Saccorhiza dermatodea, Kjellm., Spetsb. Thall., II., pp. 14, 15,
Taf. I. Figs. 8 and 9, 1877.

Saccorhiza dermatodea, Kjellm., Alg. Murm., pp. 36 and 65, 1877.

Phyllaria dermatodea, Gobi, Alg. Fl. Weiss. Meeres, p. 75, 1878.

Laminaria dermatodea, Robinson, Flora of Essex Co., p. 163, 1880.

Saccorhiza dermatodea, Farlow, N. E. Alga?, pp. 95 and 96, 1881.

Saccorhiza dermatodea, Collins, Torr. Bull., Vol. IX. p. 71, 1882.

* Examen I. was presented to the Imperial Academy at Breslau in 1854, but
was not piiblislicd until 18-30. Examen II. was publisiied in the Memoirs of
the Imperial Academy at Clierbourg in 1855. In tiie mean time a summary
had appeared in the Comptes Rendus (Tom. XL. p. 470), and reviews in the
Annals and Magazine of Natural History (2d Ser., Vol. XV. p. ,118, April,
1855), and in the Ilegcnsburg Flora, as is mentioned in a note appended to
Examen II.

t J. G. Agardh say8, on page 18 of " De Laminarieis," that L. dermatodea,
Harv., "in scedula," is his Laminaria Jtssilis, and adds, "sec. specim. distrib."

OF ARTS AND SCIENCES. 215

Phyllaria dermatodea, Kjellm., N. Isliufv. Algfl., pp. 278 to 282,
Tuf. 20j Figs. 1 to 4, 18«3.

Phyllana lorea, Kjellm., N. Ishafv. Algfl., pp. 282 and 283, Taf. 24,
Figs. 1-3, Taf. 25, Figs. 5, G, 1883.

Phyllana dermatodea, Kjellm., Arct. Alg., pp. 223 to 225, Taf. 25,
Figs. 1 to 4, 1883.

Phyllana lorea, Kjellm., Arct. Alg., pp. 22G and 227, Taf. 24, Figs.
1 to 3, Taf. 25, Figs. 5, 6, 1883.

Phyllaria dermatodea, Wille, Bidr. Alg. Phys. Anat., pp. 27, 28,
and 81, 1885.

Laminar ia dermatodea, Ue Toui et David, Catal. Herb. Zanardiui,
p. 115, 1885.

Phyllaria dermatodea, Wille, Bot. Centralbl., Bd. XXVII. p. 246,
1886.

Phyllaria lorea, Strijmfelt, Algveg. Isl., pp. 42 and 76, 1886.

Saccorhiza dermatodea, Hay, Free. Roy. Soc. Canada, Vol. V.
Sec. IV. p. 172, 1888.

Saccorhiza dermatodea, Ilariot, Journ. de Bot., p. 182, 1889.

Phyllaria dermatodea, Foslie, East Fiuraark Algae, pp. 74 to 76, 1890.

Exsiccatce.

Laminaria lorea, Aresch., Alg. Scand. Exss., No. 213.
Saccorhiza dermatodea, Farl., Eaton and Anders., Alg. Am. Bor.
Exss., No. 120, 1879.

I desire to express my thanks to Prof. W. G. Farlow of Harvard
University, in whose laboratory and under wiiose direction the present
paper was prepared, for assistance at every point; also to Dr. G. H.
Parker of Harvard University, and to Mr. F. S. Collins of Maiden,
Mass., for valuable aid in the way of materials.

Cryptogamic Laboratory, Harvard University,
June 1, 1891.

216 PROCEEDINGS OF THE AMERICAN ACADEMY

EXPLANATION OF THE FIGURES.

Platk I.

Fig. 1. Diagram of a very young plant ratlicr narrower than the typical form.
Tlie midrib-like portion is shaded. X 10.

Fig. 2. Diagram of a very young plant of the typical form. The midrib-like
portion is shaded as in Fig. 1. X 10.

Fig. 3. Sketch of a young plant from Nahant, showing no trace of a rhizogen.
A = the primitive disk ; C = the stipe ; D — the more complex blade ; E =
the primitive blade. X 1.75.

Fig. 4. Sketch of a young plant showing the rhizogen at B. Other letters as
in Fig. 3 X 1.

Fig. 5. Sketch of a young plant in wliich the hapteres are just appearing.
X 1.

Fig. 6. Sketch of a plant somewhat more advanced. Xl

Fig. 7 Sketch of the base of a young plant in which the hapteres (F) are well
developed, but in which the primitive disk (A) still persists. X 1-

Fig. 8. Diagram showing the point of attachment of the primitive disk ( 1) and
of each of the first set of hapteres (2). X 1.

Fig. 9. Sketch of the base of a young plant showing the primitive disk (A)
still remaining, the first set of hapteres attached, and the second set of hap-
teres just appearing. X 1.

Fig. 10. Sketch of the base of a small adult plant collected at Nahant in
February, 1890, showing the permanent holdfast, the flattened stipe, and the
" oar-shaped " base of the blade. X 1-

Fig. 11. Tracing of the outline of the base of a young plant having the appear-
ance of " renewing the blade." At A is the constriction between the young
and the adult portions. X 1

Fig. 12. Diagram of a portion of a transverse section of the blade of a speci-
men 1.2 cm. long, showing at A and A the beginnings of the more complex
portion. X 500.

Fig. 13. About 80 n below the section represented in Fig. 12. X 500.

Fig. 14. Diagram of a portion of a transverse section still lower down on the
same plant, showing the three-layered blade in the centre and the one-layered
blade on the sides. In the very centre is sliown the transverse section of a
young sclcrenchynia fibre, and at one side are the bunches of hairs belonging
to young cryptostomata. X 500.

Fig. 15. Diagram of a little more than half of a transverse section below that
represented in Fig. 14, showing the proportion existing between the three-
layered blade of the centre and the one-layered blade at the sides. X 385.

Fig. 16. Diagram of a portion of a transverse section below that represented
in Fig. 15. X 000.

Fig. 17. Diagram of a portion of a transverse section showing at B the origin
of a second more complex portion of tlie blade. At A is the transverse sec-
tion of a sclerenchyma fibre. X 385.

Fig. 18. About 15 IX below Fig. 17. a, a, represent transverse sections of
young sclerenchyma fibres. X 385.

1 )

r J

i'i:y:i:tn3X:

■-1

rf

^

III •!%

"f

23

25 c

25 b W

«t

J
lf(

Z5a

' <i :^.-.-V:'y^' -:»^u- '^

OF ARTS AND SCIENCES. 217

Fig. 19. Diagram of a portion oi a transverse section still lower down. L, L =

limiting layers ; C, C := conical layers. X 385.
Fig. 19 a. Diagram of a small portion of a similar section showing two young

sclerenchyraa fibres. X o85.
Fig. 20. Diagram of a portion of a longitudinal section in the same portion of

the blade as that represented m Fig. 19, with tlie same lettering. The

shaded baud m the middle represents a young sclerenchvma fibre. X 3b5.

Plate II.

Fig. 21. Diagram of a portion of the centre of a transverse section through a
rhizogen which has produced one set of hapteres. a, sclerenchyma cells;
b, hyphal cells. X 500.

Fig. 22. Diagram of a section through the centre of a young cryptostoma in
the transition place of a frond about 20 cm. long. X 500.

Fig. 23. Diagram representing portions of two elongated cells of the inner
cortex of a young specimen, showing the thickenings around the pits.
X 500.

Fig. 24. Diagram of a small portion of the margin of a transverse section of
the stipe of a specimen collected at Nahant in September, 1889 L = the
limiting layer ; a = a couple of cells which have recently divided off from
the peripheral cell. X 500.

Fig. 25, a, b, c. Diagrams of tips of sclerenchyma fibres seen in median longi-
tudinal optical section From a specimen collected at Nahant m December,
1889. X 385.

218 PROCEEDINGS OP THE AMERICAN ACADEMY

XIV.

ON SOME SIMPLE CASES OF ELECTRIC FLOW
IN FLAT CIRCULAR PLATES.

By B. O. Peirce.

Presented May 13, 1891.

1. It is well known * that, if in a thin conducting plate of indefinite
extent there are two sources and a sink, each of strength numerically
equal to m, situated respectively at points A, B, C, which lie in order
upon a straight line, one of the lines of flow consists in part of a cir-
cumference of radius \/0A ■ CB drawn around C as a centre; so
that the flow inside the circumference would be unchanged if the part
of the plate outside it were cut away. In other words, if a circum-
ference be drawn in a thin conducting plane plate of indefinite extent,
the " image " in this circumference of a source, of strength m, situated
at a point P in the plane, is made up of a sink, of strength m, at the
centre of the circle, and a source of the same strength at Q, the inverse
point of P with respect to the circumference.

If a sink be regarded as a negative source, it follows from this that
if inside a circumference drawn in a thin plane conducting plate of

indefinite extent there are sources at the points Ai, A^, As, A^,

of strengths algebraically equal to mi, m^, m^, .... m^, respectively,
and sources of strengths algebraically equal to — Wj, — m.2, — m^,

— wjj, at the corresponding inverse points, then, if m^ -\- m.,

+ m3 + — -\- m^ = 0, there is no flow of electricity across the
circumference.

2. If, at a fixed point P in a thin plane plate there is a sink of
strength numerically equal to m, and at another point P' in the plate
an equal source, and if P' be made to approach P as a limit and the
product m ■ PP' be kept always equal to a given constant, /a, we have

* Kirclihoff, Popg. Ann. 1845, p. 497 ; W. R. Smith, Proc. Ed. Roy. Soc,
1869-70; Foster and Lodge, Phil. Mag., 1875; Minchln's " Uniplanar Kinemat-
ics," p. 213; etc., etc.

OF ARTS AND SCIENCES.

219

as a limit a "doublet"* of strength fj., the axis of which is PX, the
limiting position of the straight line drawn from P to P'. We shall
find it convenient to represent sources and sinks respectively by black
and unshaded circles, and doublets by circles half black and half un-
shaded. The black portion of a doublet circle indicates the directions

in which there is a flow awai/ from the point where the doublet is
situated ; the unshaded portion indicates the directions from which
there is a flow towards this point. The axis of a doublet bisects both
the black and unshaded portions of the doublet circle. (See Fig. 1.)
If P be used as origin and PX a.s axis of abscissas, the velocity poten-
tial function due to the doublet is <^ = — ^

„ „ and the flow func-
If X -{- yi ^ z, these are respectively the real

tion IS i/^ = „ , .,

part and the real factor of the imaginary part of the function ^

z

Fig. 2.

The equipotential lines and the lines of flow are circles (see Fig. 2)
touching the axes of ^^ and x respectively at the origin.

In uniplanar motion, the velocity at a point 31 due to a doublet of

strength ^u, at a point P is numerically equal to ^ ^ , and is directed

* Basset, Treatise on Hydrodj'namics, Art. 47 ; Mascart and Joubert, Trea-
tise on Electricity and Magnetism, Art. 151; Neumann, Untersiichungen iiber
das Logarithmische und Newton'sclie Potential Chap. TV.- etc.; etc.

220

PROCEEDINGS OP THE AMERICAN ACADEMY

along a straight line MN such that PM bisects the angle between
MN a,\\di the axis of the doublet. In all that follows, the motion will
be assumed to be uniplanar.

3. If, in the subjoined figure, OL ■ ON = OL' • ON' = 0^, aud
if there are sources, of strength m, at L' and N', and equal sinks at

Fig. 3.

L and iV^, there will be no flow across the circumference drawn around
0 with 0^ as radius. The value at P, the co-ordinates of which are
X and y, of the flow function due to these sources and sinks is, if
OL = a, LL' = 8, 0A = r,

i/r „ = m ; tan~^ ^

tan

-1 y

^y

= mJtan ^

( {x—af—h{x—a)-fif

X — a

■tan"

+ tan

-1

y . -tan-^

y

X

a + S

X — -

r^i

a

{ax—i^y-\-8x(ax—r')-\-a{a-\-?))y

'>)A'

If now 8 be made to approach zero as a limit, and m to increase in
such a way that m • 8 is always equal to a constant, /i, the circumfer-
ence just mentioned continues to be a stream line and the flow function
due to the combination of a doublet of strength /x at X with axis
coincident with the radius drawn through Z, and the image of this
doublet in the circumference is

Limit if/p= fx) ^

y

h^)'

(1)

+ y

It is evident from this result (see section 2) that the image of a
doublet of strength /x so situated at a point L — wliich is at a distance a
from the centre of a circumference drawn with radius r on a tliin plane
indefinitely extended plate — that its axis coincides with the radius

OF ARTS AND SCIENCES.

221

drawn through L, is a doublet of strength* /a— and axis making an

angle of 180° with that of the first, at the inverse point of L with
respect to the circumference.

It is to be noticed that the flow function given by (i) is the real
factor of the imaginary part of the function

t^ = — /x

z — a

a'

2 — _

a

4. By differentiating (1) partially with respect to y and — cc,
respectively, we get the velocity components at the point (x, y) cor-
responding to the case which we have been considering. They are

u = D^\p = ix

-Ji^-^y-f

r- ("-!^T

r

\{x-af-\-y''Y a

' ih^f

+f

(2)

v = — D^^^ = 2ixy.

X — a

r^ {ax — 7^)

[{x-ay + y'Y

"1(^-9"

y-

(3)

If we make a approach zero in order to obtain the flow in a thin
circular plate due to a doublet at the centre, tp as given by (1) grows
larger without limit, but u and v approach the definite limits

Mn

( a;2 — ?/2 1 )

/-]- r- )

\ [_x' +

Vq =

r]-

2 ixxy

(4)

(5)

and these expressions solve the problem.

This case may be treated in another way, however. In Figure 4 let
OL =0L' = 8 and 03f. OL = OL' .OM' = OA' =■. r^; then, if
there are sources of strength m a,t L and 31 and equal sinks at L' and
jW, one of the lines of flow due to this combination will consist in part
of the circumference of radius r drawn about 0 as centre.

* This does not as^ree with the statement marie in the first edition of
Basset's Treatise on Hydrodynamics, p. 56, where tiiere seems to be a typo-
graphical error.

222

PROCEEDINGS OF THE AMERICAN ACADEMY

Fig. 4.
The value at the point (x, y) of the flow function is
x^/ z= m,\ tan~ ^ — ^— ^ + tan~ ^ — ^ — - — tan~^ — ^ — ^ — tan~^ ^

X — 8

= m\ tan-^ _— — J-— + tan"!

a: + 8

7M *

8

8M^' + y')

If, now, 8 be made to approach zero as a limit, and m to increase in
such a way that 2 m 8 is equal to the constant ^,

^o^Limit^^-^(-^+/;-/^); (6)

an expression which is equal to zero at every point of the circumfer-
ence, x^ -\- ■if- ^ r^.

The velocity components Mq, Vq, obtained by differentiating (6), are
identical with those given by equations (4) and (5).

Fig. 5.

5. In Figure SJet OA = OA^ ^a, 0B= OB' = b, and OA OB
= OA' OB' = OC'^ = r\ and let there be at A and B sources of
strength m, and at A' and B' sinks of equal strength m ; then one
of the lines of flow due to this combination of sources and sinks in an

OP ARTS AND SCIENCES.

223

indefinitely extended plate will consist in part of the circumference of
which 00 is a radius. This shall be called the circumference 0.

The value, at P, the co-ordinates of which are x and y, of the flow
function is

^p= —m\ta.n~^^ + tan ^-l -^ — tan~^ ^

( a; — a cos ^ x — Oisosd x — a

^-m ^tan-^ ^^ ~ ^^ ^^ - a sin 6) —y (x ~ a cos 0)
I {x — a) [x — a cos t>) + y {y — a sin 0)

— tan

-1 y >

x — b)

j^^- 1 {x - h) (y -bsine)-y(x-b cos ^))

■ sin ej\'

(x — b) {x — b cos 0) -\- y (y — b sin

If now we let 6 approach zero as a limit and m increase in such a
way that the product of m and the straight line AA' shall always be
equal to the constant fi, we have

.2

ij/Q = Limit iff =: -{- fi-

X — a

+

X — —
a

0 = 0 ^{x-ar + f ' a^ L-'lX + f

the real part of the function

,(7)

w

1

z — a a'

At every point of 0, \f/o has the value

2 — —

a

t.
a

The velocity components at P corresponding to i/^o are
M = — 2/xy

X — a

[_{x - af + y'Y "^ a

r

X

a

v = -\- fl-

(x — ay — y^

[{X - af + ff

+ -,

(^-^T-^1
(^-^T-'

a'

(^-9'-^T

(8)

(9)

If a is made to approach zero as a limit, these expressions approach
respectively the limits

Wo =

_ — 2 [xxy

PTTP

/ a;2 - ^2 1 X

(10)

(11)

224 PROCEEDINGS OP THE AMERICAN ACADEMY

which agree with the expressions (4) and (5) obtained by using
different co-ordinate axes.

We infer from (7) that the image of a doublet of strength jx so
situated at a point A — which is at a distance a from the centre of a
circumference drawn with radius r on a thin plane indefinitely ex-
tended plate — that its axis is perpendicular to the radius vector

o

drawn through A, is a doublet, of strength fx — and axis parallel to

that of the first doublet, at the inverse point of A with respect to the
circumference.

If we make a equal to r, (7), (8), and (9) give us the flow function
and the velocity components inside a circular plate when there is a
doublet at some point of the circumference with its axis coincident with
the tangent to the plate at the given point. The forms of these expres-
sions might have been inferred from the results given in section 2.

6. It is to be noticed that if in Figure 6 O'B' and OB are parallel
lines, and 00', AA', and £B' perpendicular to them, and if there are

f-0' Q-A^ B^

G-0 ^A B-#

Fig. 6.

sources of strength m at 0', A, and B, and sinks of equal strength m
at 0, A', and B\ a circumference drawn around 0 as centre with
radius ^/OA . OB will form part of a line of flow due to the combina-
tion of the sink at 0 and the sources at A and B, but not one of the
lines of flow due to the combination of three doublets obtained by
making O'B' approach OB and increasing m at the same time so that
m . 00' shall always be equal to the constant /x.

7. In Figure 7 let OA.OB= OA' . OB' =0C^= r\ and OA = a,
and let the curve AA' make with OM the angle 8. If, then, there are
sources of strength m at A' and B', and sinks of strength ;» at A and B,
one of the lines of flow will consist in part of the circumference (C)
drawn with radius r about 0 as centre. If if/p is the value at P of the
flow function due to this combination, the limit appronchod by ij/p as
A' is m ide to approach A along the curve A'A, and Ji' to approach
^ along the corresponding reciprocal curve B' B, will be the value at
P of the flow function due to a doublet at A, the axis of which makes
the angle 8 with the radius drawn through A of the circumference C,
and to the image of the doublet in C.

OP ARTS AND SCIENCES.

225

Fig

(. -1 y—KA , _i y , * _i y—LB' , _, y
— «i fan i if _ — tjin i _£ — -f-tan ^ , — tan — ^-

5^^ = m tan~

X

-a-^^

X — a

x-OB-BL

x-OB

= m •] tan ^

y.AK-KA {x-a)

(x — ay + f — 1/. KA' — AK {x — a)
_i y.BL-LB' (x-OB)

+ tan

(x - OB)'^ ^f--y. LB' - BL{x- OB)

-0B)\

If ^' be made to approach A as a limit, we have

Limit

AA

; Limit — sin 6 — Limit
OA" AA BB'

Limit — cos 8 — Limit

AA' BB'

and, if m be made to change so that m . AA' shall always be equal to

the constant fi, Limit (m . BB') will be equal to the constant, — . fi.

a"

Limit i^p = i/'o = )(^ Limit
AA' = 0 AA'=0

tan

_i y.AK-KA (x-a)

(x—a)'^+y^—y . KA—AK(x—a)

+ —„ Limit
BB' = 0

a'

( tan-^

< (x-'

AA'

y.BL-LB' (x-

I

(x-OB)'+y'-y

BB'

B' (x-OB) \

TB'-BLix-OB) >

VOL. XXVI. (y. s. XVIII.)

15

226 PROCEEDINGS OF THE AMERICAN ACADEMY

AK KA' , .

y ^ I ^ d I

1/^0 = /^. Limit J ' JJi A A'

AA' = 0

{x — a)2 + r — y • ^-^' — -4^ (a; — a)

BL LB' .

a'

J3B> = 0

{x -OBy + f-y.LB'- BL {x -OB)

\ X . ., y.cos8+(x — j sinS)
J y . cos 8 — (x — a) sin 8 r^ \ a I ' / 1 )\

It follows from (1 2) that the image of a doublet of strength /a so situ-
ated at a point A — which is at a distance a from the centre, 0, of a
circumference drawn with radius r on a thin, plane, indefinitely ex-
tended plate — that its axis makes the angle 8 with the radius drawn

through ^, is a doublet of strength - — and axis making an angle of

180° — 8 with the line OA, at the inverse point of A with respect to
the circumference.

8. It was long ago noticed * that the functions

log«, -'

112 6

^5' ~3 ' "i' •

each of which is the derivative with respect to z of the one wliicli
precedes it, yield a series of pairs of conjugate functions which repie-
sent in order the velocity potential functions and the flow functions
due to a source at the origin, to a doublet at the origin, to a quadru-
plet t at the origin, to an octuplet at the origin, and so on.

I will use the word motor to denote in general a source, a sink,
a doublet, a quadruplet, or any other combination of sources or sinks
at a single point. The strengths of two motors of the same kind
shall be in the ratio of /> to q if, when they are similarly placed at
the same point successively, they give rise to velocity potential func-
tions and flow functions which have values in the ratio oi p to q at

* See, for instance, Klein's Lectures on tlie Potential, or his paper, " Ueber
lliemann's Tlieorie der Alpebraisclien P^inctionen unrl ilirer Integrale."

t A quadruplet is formed of two equal and opposite doublets in the same
manner that a doublet is formed out of a source and an equal sink. An octu-
plet is formed in a similar way of two equal and opposite quadruplets, and so on.

OF AKTfc AND SCIENCES. 227

every point. The unit motor of any kind and the direction of its
axis, if it has any, may be chosen at pleasure.

Let there be in a plane any distribution, A, of n motors, either all of
the same kind or of different kinds, of strengths respectively equal to
till, ni^, Wg, m„.

Let there be also a system, B, of n motors, each respectively equal
but opposite to a corresponding motor of A, and so situated in the
plane that every motor of A could be superposed upon its equivalent
in B by moving A parallel to itself through a distance 8 in a direc-
tion making an angle a with the axis of x.

Let u — f(x, y) und ^" = x (^' V) ^^ ^^e values at the point (x, y)
of the velocity components due to the system A^ and let (^ (a-, y) and
i/f {x, y) be the corresponding values of the velocity potential function
and the flow function.

It is evident that the velocity components due to B have values at
(ic, y) equal but opposite to those at the point (x — 8 cos a, y — 8 sin a)
of the corresponding velocity components due to A. We may, there-
fore, take for the values at any point (x, y) of the velocity potential
and flow functions due to A and B existing together, the expressions

(^ (x, «/) — <^ (x — S cos a, ^ — 8 sin a) ; (13)

\p{x, y) — ij/ {x — 8 cos a, y — 8 sin a). (14)

If now Ji be made to approach A, by decreasing 8 and keeping a
constant, and if the strength of each of ^'s motors be made to grow so
as to keep the product of itself and 8 constant, the expressions just
given approach as limits the velocity potential function and the flow
function which A would yield if every one of its motors were doubled
up with an equal but opposite motor approaching it from a direction
which makes an angle a with the axis of x.

If mi 8 = yu-i, mgS = |Lt2? ^38 = )U,3, etc., where the ya's are con-
stant, the limits of the expressions (13) and (14) are the values
obtained by changing every m into its corresponding /x in

Limit {^^^±±^^\ and Limit (^^^ + ^y^'
8 = 0^ ^ / 8=^0^ S

that is, in cos a . D^. ^ -\- m\ a . Dy (ft, (15)

and cos a . Z)^. xp -\- sma . Dy if/. (1 6)

If IV =zf(z) := (f) (x. y) + i . if/ (x, y), the function of z which corre-
sponds to the new velocity potential function and flow function may
be found by changing every m into its corresponding fj. in

— (cos a -{- i . sin a) . B^ ic. (17)

228 PROCEEDINGS OF THE AMERICAN ACADEMY

9. The equation "- — = c represents a family of closed curves

each of which is made up of 2 n equal mutually distinct loops sym-
metrically situated about the origin ; each curve passes n times through
the origin and each loop is tangent there to two straight lines which

include an angle of _ . The equation — = c represents the same

n r"

IT

family of curves turned around the orisfin through an angle — .

Tlie motor which gives rise to the flow function ~ may be

r

denoted by the symbol shown in the upper part of Figure 8, in which

the black portions show the directions in which fluid flows out from

the point at which the motor is situated, and the open portions the

directions from which fluid flows towards this point. The motor

, • , , , _ , . A sin n ^ , ,

which corresponds to the now lunction — may be represented

similarly by a circle drawn about the origin, and properly divided
into n shaded and n unshaded sectors.

Let the motor A^ which corresponds to the flow function \pJ^ = ,

r"

be doubled up with an equal opposite motor which approaches it from
a direction making an angle a with the axis of x. The flow function
i{/^, due to the resulting motor £, is given, apart from a constant factor,
by the equation

ips = Dx\pj^. cos a + Z>j, i/r^ . sin a

= A xlf^ . cos {6 -a)- ^^ . sin (6 - a)

n . sin \{n + 1) ^ — «]

(18)

Referred to a new initial line drawn in a direction making an angle

T with the old, this is

w . pin (n + 1 ) ^

^B = ^UT ^ ; (19)

and it is evident that the orientation of i?'s axis, but not ^'s charac-
ter, depends upon a.

For instance, the flow function due to a unit doublet at the origin
with axis coincident with the axis of a: is

a:' + /

OP ARTS AND SCIENCES.' 229

If this motor be doubled up with its negative approachins; from tlie
positive purt of tlie axis of x, the resultaut quadruplet (see Fiy. 8)
corresponds to the flow function

(x- + y^Y

e «

Fig. 8. Fig. 9.

If, on the other hand, it be doubled up with an equal opposite
doublet (see Fig. 9) approaching it from the positive part of the axis
of y, the resultant quadruplet corresponds to the flow function

•2\2

One of these quadruplets is evidently equivalent to the other turned
through an angle of 45°.

10. Besides the simple quadruplets obtained by combining two
equal and opposite doublets just as a doublet is formed from a source
and a sink, other combinations of two doublets sometimes appear
when one attempts to find the image of a simple motor in the cir
cumference of a circle in its plane.

In one common case, a fixed doublet D^ of strength fx. is ap^
preached by another doublet D^, the axis of which remains always
parallel but opposite in direction to that of D^. The path of D.^ is a
straight line, but its strength is equal to fxf{S), where B is the dis-
tance between the two doublets at any time, and f (S) is a finite,
continuous, and single-valued function such that ^(0) =: 1. Th^
product of fx and 8 is kept always equal to a constant, A.

The flow function due to a doublet of strength fx at the origin

with axis coincident with the axis of x is -J~^— , and the flow

x^-\-y^^

function due to a reversed doublet of strength (xf (8) at the point
{x — 8 . cos a, y — 8 . sin a)

— fx . (y — 8 sin a) ../'(8)

IS

(a; — o . cos ay -\- {y — 8 . sin a)^

The flow function due to the two doublets existing together is

A ( y _ (;/ — 8 . sin a) /*(8) )

8 Ix^ -\- y'^ (x — 8 . cos a)'^ + (y — & . sin a)' \

230 PROCEEDINGS OF THE AMERICAN ACADEMY

and the limit of this as 8 approaches zero is

X ^ (a:-2 - y-^) sin g - 2 xy cos a — y (x- + f)/' (0)^ . ,^^^.

{x' + fr ' ^

and this is equivalent to a simple quadruplet superposed upon a
doublet of strength — A ./' (0) at the origin.

In general, let fx . vf/ (x, y) be the flow function due to a fixed
motor of strength \x.^ and of any kind ; and let this motor be approached
(keeping /x 8 constantly equal to A) by an opposite motor of strength
\x . /'(8) from a direction making an angle a with the axis of x. The
value at the point (x, y) of the flow function due to this second motor
i«* the same as the value at the point (.r — o . cos a, 5/ — S . sin a) of
the flow function due to the first motor, so that the flow function due
to the two motors existing together is

K [lA (^' y) — /(S) • "A (^ — ^ cos a.,y — Z sin a)],
o

and the limit approached by this as 8 approaches zero is

\\P^^\, . cos a + Z)^ ./. . sin a — i/. {x, y) ./(O)]. (21)

11. Let there be two equal opposite motors of the same kind at
the points A and J5 in a plane equally distant from an origin, (9, and
let the axis of the motor at A make with the straight line OA the
same angle that the axis of the motor at B would make with OB
if its flow function were the negative of what it really is. If now
OB l)e rotated about 0 so as to approach OA as a limit, and if the
product of the strength of one of the motors by its distance from the
other be kept equal to a constant, A, we get as a limit a kind of motor
which sometimes appears in practical problems.

Let II . R {r, 6) and /* . (^ (r, 0) be the values at the point P, the
co-ordinates of which are r and ^, of the velocity components due to
the motor at A^ taken respectively parallel and perpendicular to the
radius vector, OP. If the angle A OB (see Fig. 10) be a, the values
of the velocity components at P due to the motor at Ji are equal but
opposite in sign to values at the point P', the co-ordinates of which
are r and 0 — a, of the velocity components due to the motor at A.
Hence the motors at A and B together cause at P the velocity
components

(x.R(r,6) -i^.R (r, e - a),

,x . 0 (r, ^) - M . 0 (•'•, 6 — a).

OF ARTS AND SCIENCES.

231

Fig. 10.

The limits of these expressions as a is made to approach zero, while
fx . a . OA = A, are

" and

A®

OA OA

so that the flow function due to the newly found motor is obtained by
changing /x to A in the product of the reciprocal of OA and the
derivative with respect to 6 of the flow function due to the original
motor at A alone ; that is, in

where

OA

fx li = '- and fi ® =: — JJr if/.

(22)

A single doublet on the axis of x, at a distance a from the origin,
with its axis perpendicular to the axis of x, yields, when combined
with another similar motor in the way described in this section, the
flow function

(23)

^^Xy(x' + f-a')

a [{X - ay + ff

This is equal to zero on the axis of x and on the circumference of
the circle drawn with the origin as centre so as to pass through the
motor. It gives, therefore, a case of flow in a semicircular disk.

The flow function just found is the real factor of the imaginary
part of the function

A« __^.(1, « )

w = —

(24)

a(z — oif a i z — a ' (z — af \

The motor which corresponds to this flow function may be looked
upon, therefore, as formed by superposing a doublet upon a quad-
ruplet.

The image of such a motor (3f) as this in a circumference drawn
in its own plane with 0 as centre is a motor of the same kind as M,

232

PROCEEDINGS OF THE AMERICAN ACADEMY

^2
but of strength — times as great as 3/'s strength, situated at the
ar

inverse point with respect to the circumference of the point where
JJ/lies.

The flow function in this case is easily found, by a method analogous
to that used in Section b, to be

,2

X ^ f x^ + y'^ — a^ 7^

a"

x^ + y^ -

a'

a \[{x- a') + y^Y <^'' \L_t\2., f^

(25)

The combination in the same manner of two pairs of motors of the
kind just mentioned yields still another case of flow inside a circular
disk due to a certain motor at a point Q, the co-ordinates of which are
a and 0, and another motor at the inverse point of Q with reference
to the circumference. The flow function in this case is

a-

7? + y^ — c?

[(X - af + /]^

A

a;2 _j. ^2 _

{x (x — a)^ + y^ (x — 4 a)}

+

a'

and this process might be carried on indefinitely.

12. The image of a simple quadruplet in the circumference of a
circle drawn in its own plane is not another simple quadruplet, but
a more complex motor.

On a straight line which passes through the centre of a circum-
ference of radius r drawn in a plane, and which is taken for axis of x,
let there be two equal and opposite doublets of strength fi with axis
coincident with the line, at points dis,tant respectively a and o + Aa
from the centre of the circumference. Let there be also on this line
the two images of these doublets with respect to the circumference.

The value at the point (x, y) of the flow function due to these
four doublets is

^y\^x-af+y^

{x — a — Acf)'^ -I- y

+

^''^^"'Ai^-v^l+A '^[(^-:7T+^^

OP ARTS AND SCIENCES. 233

and if Aa be made to approach zero, and fj. to increase so that fi . Aa
always has the constant value A, this function approaches as its limit

If the axis of a simple quadruplet be a diameter drawn through the
centre of the shaded portions of the symbol which represents the
motor, the expression just found gives, with proper algebraic sign,
the flow function due to a simple quadruplet in a circular disk with
axis parallel or perpendicular to the radius drawn through the point
at which the motor is situated.

The flow function, which vanishes at every point on the circum-
ference, is the real factor of the imaginary part of the function

w = — X -{ + — .

('-^)

,.4

2 r^

{z-df a* /^~_^V «' _^'r (28)

\ a J a

The image of the simple quadruplet in question is made up, therefore,
of a simple quadruplet and a doublet existing together at the inverse
point, with respect to the circumference, of the point where the
original quadruplet is situated.

13. If the axes of the doublets mentioned in the last section had
been perpendicular to the axis of x instead of coincident with it, the
flow function due to the four doublets would have been

1,
i

X — a X — a — Aa

(x — ay + f (x—a — Aay + y'

2

^2

X — — o

r^ a _ ''

">-F)

+ y''

X —

7-2

a +

Aa

Ir-

r"

y ^ ,,

The limit of this, as Aa aproaches zero and n Aa is kept equal to
the constant X, is

234 PROCEEDINGS OF THE AMERICAN ACADEMY

^^ ( f — (x — ay r*

([(x-af + ff «4- / _^

h^J^^

This has the value — — at all points of the circumference, and gives

a^

us the flow function inside a circular disk in which there is a simple

(jiiadruplet with axis making an angle of 45° with the radius drawn

tlirough it.

J t is to be noticed that (29) is the real part of the function

which appears in equation (28).

14. The flow function due to a simple quadruplet at the centre of
a circular disk of radius r may be found by putting a equal to zero
in (27). It is

^ = -2Axy j, ,! ,,-\\^

1 2^

and the corresponding w is — + - .

In general, consider the function

1 , s" r2« + p2" (cos 2 w ^ + I . sin 2 n ^)

_n „a n

^m ^2n _ pu (^gQg nO -\- i . sin n 0)
r^" . cos nO -\- p-"' (cos 2 n6 . cos ?i ^ + sin 2 ?2 ^ . sin n 9

+

r2» . p"
i {p^" (sin 2n 6 . cos w ^ — cos 2 w ^ . sin w 6) — r^" sin w 6]

^ . cos nO -\- —^—z smn 6. (61)

OF ARTS AND SCIENCES. 235

The function

is constant along the circumference drawn about the origin with
radius r, and is evidently the flow function due to a simple
2" + ^ -plot at the centre of a circular disk.

15. The rational algebraic function 0, , , where the numerator

■^ (z)

and denominator are algebraic polynomials involving only integral
powers of 2, can be put by simple division into the form of a rational

algebraic polynomial in z plus the proper fraction, -LA^ , where the

F(z)

numerator is of degree at least one lower than the denominator.
This proper fraction can be represented in the usual way as the sum
of a series of fractions the numeratois of which are in general com-
plex constants and the denominators of the form [z — {a -\- b i)]".

The real part of -^ >-^ is therefore the velocity potential function

of a collection of motors such that each has for its own velocity po-
tential function the real part of a function of the form

a -{- (3 t ^ fi (cos 8 + ^ sin 8) .oo\

(z — {a+b i)y (z — (a + b i)y

Fig. 11.

Let us take as the axis of a simple multiplet a line which, drawn as
a diameter through the symbol which represents the motor, bisects
the black portion nearest the axis of x. If this line goes through
two black portions, the sense in which the direction of the axis is
taken is of no consequence. If it goes through one black sector and
one unshaded sector of the symbol, the direction shall be from the
unshaded part to the black part. The simple motor, for instance,
which has for its flow function the real factor of the imaginary part of

— , has its axis coincident with the axis of x.

z"

236 PROCEEDINGS OF THE AMERICAN ACADEMY

The values at the point (x, y) of the velocity compoueiits due to
any motor at the point (a, h) must be equal respectively to the two
velocity components at the point {x — a, y — b) due to the same
motor at the origin with axis parallel to its old direction.

If ^ {x, y), ip (x, y) ,a.ve the velocity potential and flow functions
due to the motor at the origin, cb (x — a, y — b), ip {x — a, y — Z>),
yield velocity components for the case of the same motor at (a, b)
which satisfy the condition stated above.

If f{z) = (p (x, y) + i if/ (x, y), however,

f(z— a- bi) = cfi(x — a, y — b) + i . xfj {x — a, y — b),

and, as a particular case, — — and ~ correspond to two

^ [z— (a + b i}]" z"

equal and similarly placed motors, one at the point (a, b), the other
at the origin.

Let us now compare the motors the flow functions of which are
the real factors of the imaginary parts of

i and 'i+li

z" z"

respectively ; that is, of

ces K d -, fi CIS 8 . ces k 6

where z = p (cos $ + i . sinO), a -\- ^i = p (cos 8 -\- i . sin 8), and
where (cos 8 -\- i sin 8) and (cos 6 — i . sin 6) are written cis 8 and
ces 0 respectively.

The two flow functions are

ft . sm

sin K ^ J

and

'(-!)

p" p"

and the values at the point (r, 6) of the velocity components respect-
ively parallel and perpendicular to the radius vector drawn from the

origm are

K . cos K 0 J K . sin K ^
— — and -j-^—

pK + l pK+l

in the case of the first motor, and

OF ARTS AND SCIENCES. 237

fi K . COS kIO — - j ^ K- . sin K (6 — - 1
- ^TTl - and ^;^+i

in the case of the second motor. The second motor gives at everv
point velocity components which the first motor gives at a point
equally distant from the origin, but differing in vectorial angle from

the first point by - . The second motor is then equivalent to the first

turned through the angle - .

If, then, the simple motor which corresponds to the function ~ were
transferred parallel to itself to the point (a, h) and then rotated
counter-clockwise through the angle -, it would correspond to the

function ^ -, , .^-, . It is to be noticed that a rotation of

the motor corresponding to — through the angle — - would by
reason of symmetry give the same motor, and multii^lying — by i

would turn the symbol representing the motor through half the angle
which corresponds to a black sector.* The rational algebraic proper

f («)
fraction -p.~. corresponds then, in general, to a distribution of vari-

. . f(z)

ously oriented simple multiplets, and the function A z + -rr ^ \ > which

frequently appears in two dimensional problems in magnetism, to
a distribution of such multiplets in a uniform field. t

16. Ihe real parts of the functions , — , are the ve-

locity potential functions which correspond to two simple multiplets
of the Kth order, L and J/, of strengths respectively equal to X and /x.
Both L and M are situated at the origin. X's axis coincides with the
axis of X, but Jf s axis makes with the axis of x the angle a.
If we superpose L upon M, we get

_ — {X -\- fi . cos K a -\- i . (i . sin *c a)

= -'Li^^, (34)

* This explains the identity of equations (28) and (30).

t See, for instance. Maxwell's Treatise on Electricity and Magnetism, Vol. II.
Figs. XV. and XVI.

FHOCEEDINGS OF THE AMERICAN ACADEMY

where

cr^ = A^ + fji' -\- 2 \ fjL COS K a, (35)

. _i ( u sin K a ) /oo\

T = sm N — ^ >■ . {6b)

'- V -^^ + M^ + 2 \fi cos Ka^

This is a motor of the same kind as L and 31. Its strength is a, and
its axis makes with the axis of x the angle - . In the case of doublets,

K

that is, when k = I, this result may be put into a simple form. If
two doublets L and M exist together at a point 0, and if the direc-
tions of the two straight lines OA, OB, show the directions of the
axes of L and M respectively, and the lengths of OA and OB the
strengths of L and 31 on some convenient scale, then the direction of
the axis of the result of L and M will be given by the directions and
the strengths of the resultant bv the leno-th of the diagonal of the
parallelogram of which OA and OB are adjacent sides. Doublets
then can be compounded and resolved by compounding and resolving
their axes like forces or velocities.

17. Let there be any motor J/ at a distance a from the centre of
a circumference of radius r drawn in a plane, and let the radius
drawn through 31 be taken as axis of x. Let n . x {x, y-, a) be the
flow function due to M, and its image in the circumference, existing
together. If another motor B, of the same kind and strength as A
but of opposite sign, exist at a point on the axis of x at a distance
a + Aa from the centre, the flow function due to B and its image
together is

- /* • X (^» y' « + -^ «)•

The flow function due to A, B, and their two images existing
together is, therefore, — /^t . A„ ^ (x, y, a).

If now Aa be made to approach zero and ^ to increase so that
H . Aa is equal to the constant X, we get as the flow functions due to
the resulting new motor at A and its image in the circumference

.// = - X . z>„x (^, y^ «)• (3")

F (2, n), of which i// is the real factor of the imaginary part, is con-
nected with/(2, a), of which ^ is the real factor of the imaginary part
by the equation

F{z,a)=-\ D„f{z,a). (38)

OF ARTS AND SCIENCES. 239

If, for insUiuce, we apply (38) to

■( z — a a- r-

w = ^^ II

\

which yieltls (5), we get

1 r^ r^ — 2 ^2

±X-{{z-af a' (^_>_

a

which, after reduction, gives (30).

By meaus of (38) and other similar equations, a great number of
cases of flow inside circular disks may be found.

Jefferson Physical Laboratoev,

Cambridge, Mass., August 1«, 1891.

240

PROCEEDINGS OF THE AMERICAN ACADEMY

XV.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF

HARVARD COLLEGE.

A REVISION OF THE ATOMIC WEIGHT OF COPPER.

FOURTH PAPER.

By Theodore William Richards.

Presented June 10, 1891.

Table of Contents.

PAGS

Introduction 240

Balance and Weiglits 242

I. Analysis of Cupric Sulphate . . 244

Materials 245

Description : Three Series . . 249

Cause of the Deficiency . . . 260

Atomic Weight of Sulphur . 268

Atomic Weight of Copper . . 270

II.
III.

IV.
V.

Sj'nthesis of Cupric Sulphate
Analysis of Cupric Oxide .
Tests for Impurities . .

PAGE

273
27tj
278

Determination of Occluded

Gases 281

Upon Typical Copper . . . 291

Summary of Results .... 293

Introduction.

It will be remembered that recent investigation upon the atomic
weight of copper has pointed toward the existence of a value for that
constant very different from the one indicated by earlier determina
tions. During the winter of 1886-87 a number of experiments * were
made in this Laboratory upon the replacement of silver from solutions
of argentic nitrate by means of metallic copper; and during the follow-
ing winter these were supported by similar series t involving material
from different sources. The consistent verdict of all these experi-
ments indicated a considerably higher atomic weight than that which
had previously been accepted ; but although the new method seemed
to be in every way satisfactory, the presumption must nevertheless re-
main in favor of the older value until the presentation of more evi-

* These Proceedings, XXII. 342.

t These Proceedings, XXIII. 177 ; Fres. Zeitschr., XXVIII. 392.

OF AIITS AND SCIENCES.

241

dence against it. Accordingly, another method of determination was
devised, and the two distinct but concordant series of results obtained
by this method fully confirmed the higher value. The aspect of the
case was now wholly changed, — it presented two series of coinci-
dences, of which one must necessarily be due to chance alone ; and
in such a case certainty can be reached only through a complete
and intelligent reconciliation of the conflicting evidence. A careful
examination of the possible sources of inaccuracy involved in all
recent analyses failed to reveal the probability of a constant error
greater than two or three units in the second decimal place. There-
fore a complete examination of the older work seemed imperative,
and the object of the present paper is to describe such a revision.

A brief review of earlier determinations is given in the first paper
of the present series, but for the sake of convenient reference a com-
plete summary is repeated below. On account of the present uncer-
tainty with regard to the ratio of oxygen to hydrogen, the standard
to which the figures in the following table are referred is the arbitrary
one, 0= 16; and this standard has been adopted throughout the
present paper.

Atomic Weight of Copper.

0= 16.000.
Berzelius, 1828, from CuO

(Pogg. Ann., VIII. 182.)

Erdmaun and Marchand, 1844, from CuO

(J. pr. Chem., XXXI. 391.)

Dumas, 1859, from CuO and CugS (?)

(Ann. de China, et de Phys., [3J, LV. 129, 198.)

Millon and Comaille, 1863, from CuO

(Compt. Rend., LVn. 147; Fres. Zeitschr., II. 474.)

Hampe, 1874, from CuO
" " from CUSO4

(Fres. Zeitschr., XIH. 352 ; XVI. 458 )

Baubigny, 1883, from CUSO4

(Compt. Rend., XCVII. 854, 906.)

Shaw, 1886, through electrolytic equivalent

" " " " " corrected

(Phil. Mag., [5], XXIII. 138.)

Richards, 1887, 1888, by relation to Silver

(These Proceedings, XXII. 342; XXIII. 177.)

" 1890, from Cupric Bromide (two series)

(These Proceedings, XXV. 195)
VOL. XXVI. (n. s. xviii.) 16

Cu =

63.29

63.47

63.50

63.12

63.34
63.32

63.47

63.48
63.51

63.60

63.61

242 PROCEEDINGS OF THE AMERICAN ACADEMY

The low value obtained by Hampe has been until recently univer-
sally accepted, because of the care employed in his analytical work,
and the agreement between his results. Hence the present discussion
will be confined almost entirely to the two substances which formed
the bases of his operations.

The following values for the atomic weights of various elements
entering into the present investigation will be assumed throughout
the paper.

Oxygen = 16.000 Hydrogen = 1.008

Barium = 137.10 Silver =^ 107.930

Bromine — 79.955 Sodium == 23.053

Carbon = 12.002 Sul^jhur = 32.060

Balance and Weights.

The balance used in the work to be described was made with especial
care by Henry Troemner, of Philadelphia, and was procured particu-
larly for the present research. The beam and pans are composed of
aluminium, and all the remaining metal-work is plated with gold. By
raising the centre of gravity, the pointer may be made to swing with
great constancy as much as eight divisions of the scale for one tenth
of a milligram ; but since this degree of sensibility is far beyond the
range of accuracy obtainable even in atomic weight investigation, the
balance was usually adjusted so that the pointer moved ten divisions
for a milligram. A reasonable increase in load altered this relation
but slightly. The balance was kept in a small closet lined with cur-
tains, and was protected as much as possible from changes in tempera-
ture. The air of the case was dried by means of potassic hydroxide
and sulphuric acid, and large dishes of the former substance were kept
in the closet outside.

The standard weights have already been described.* They were
carefully rubbed with chamois skin and again compared before the
present work. The slight corrections, which differed scarcely at all
from those previously found, were applied to each weighing.

A double rider attachment upon the balance made the method of
weighing by substitution accurate and easy ; and accordingly this
method was invariably adopted, except in the first series of experi-
ments upon cupric suli)hate. In general, the vessel to be weighed was
carefully tared with an adjusted set of common gilded weights, and the

* These Proceedings. XXV. 196.

OP ARTS AND SCIENCES. 243

vibrations of the pointer were noted. After the removal of the vessel
standard weights were added to the left-hand pan until the same con-
dition of equilibrium was reached. The vessel was then once more
substituted for these weights, and any slight change of centre point
was of course manifest at once. In the rare cases when such slight
change appeared, the weights and vessel were alternately substituted
for each other until constancy was reached. In the case of hygro-
scopic substances the already ascertained weights were first placed
upon each scale pan, then the vessel was quickly removed from the
desiccator and substituted for the standard weights, and, finally, the
latter were again put in place.

All desiccators were allowed to remain in the closet with the bal-
ance three or four hours before the weighing, and objects were often
weighed on successive days, to furnish assurance of constancy. In
every case the barometer and thermometer were read, and any correc-
tion due to change of relative buoyancy amounting to more than the
fiftieth of a milligram was applied to the result. With large vessels
an invariable slight loss of weight, amounting sometimes to as much
as one twentieth of a milligram, was noticed after the object had re-
mained for some time upon the balance pan. The loss may have been
due to the replacement of the perfectly dry air from the desiccator by
the less completely dried air of the balance case. Possible error from
this source was avoided by the employment of uniform conditions
suited to the particular substance in hand.

All weighings were of course reduced to the vacuum standard by
calculation from the specific gravities of the substances and weights
involved.* Through the great kindness of Professor Mendenhall of
Washington, two of the Laboratory's ten-gram weights (one of brass
and one of platinum) have been compared as carefully as possible
with the standards of the Washington Bureau of Weights and Meas-
ures. Five comparisons of the Sartorius ten-gram weight with these
gave the following results for its value in vacuum : —

1891. Grams.

February 4. By comparison with the brass weight, 10.00025

March 4. " " " platinum weight, 10.00025

March 14. " " " " " 10.00026

June 12. « « " " " 10.00026

June 12. " « " " " 10.00024

Average, 10.00025

* These Proceedinp^s, XXV. 196. Specific gravity of brass = 8.3.

244 PROCEEDINGS OP THE AMERICAN ACADEMY

The third aud fifth of these comparisons were made with the bal-
ance in the condition of greatest sensibility, while the others were
made exactly in the usual manner. Jt is not pretended, however,
that all the weighings which follow have probable errors so small as
these. The results are nevertheless an excellent gauge of efficiency
of the balance, especially since the observations were made at widely
varying temperatures and pressures. The correction involved, apply-
ing in the same proportion to all weighings, has been omitted from the
following figures; but any weight given below may of course readily
be reduced to the Washington standard by multiplication with the
factor 1.000025.

I. The Analysis of Cupric Sulphate.

The only published analyses of cupric sulphate which have had for
their object the determination of the atomic weight of copper are those
previously mentioned. Hampe has very clearly described two ex-
tremely concordant electrolytic determinations of the copper contained
in the so-called anhydrous salt. The greatest care had been taken to
separate minute traces of metallic impurities from the preparation,
which had been dried at 250° Centigrade ; but apparently it was not
realized that for the purpose in hand the retention of a trace of water
by the salt was a much more serious possible cause of error than
any other. Baubigny's analyses are less fully described, so that it is
less easy to judge of their value. A different method, the conversion
of cupric sulphate into the oxide by heat, was adopted for these ex-
periments. The result was very different from that obtained by
Hampe, but was more nearly accurate, because complicated with a less
serious combination of constant errors.

The entire reconciliation of all these results, not only with each
other, but also with the more recent analyses, involved the complete
analysis and synthesis of cupric sulphate. A somewhat detailed ac-
count of the various operations is given below. Many of the less
important points are necessarily omitted on account of the already
too great accumulation of material, and many relations between the
figures can be worked out by those whom they may interest. It is to
be hoped that the effort to spare the overburdened literature of to-day
any unnecessary additions may not have interfered with the clearness
of the description. The first three series of experiments upon the
analysis of cupric sulphate constituted a study of the effect of pro-
gressive refinement, and the first of the three perhaps hardly deserves
a place in a paper upon atomic weights. Unless otherwise stated, the
(lata are given in full.

OF ARTS AND SCIENCES. 245

Materials used in the Analysis.

Capric Sulphate. — The substance used in the first series of experi-
ments was partly prepared from very pure copper remaining from the
ciipric bromide research, and partly from other sources. It was per-
fectly neutral and reasonably pure. For the second and third .series
four hundred grams of the so-called '' chemically pure " cupric sulphate
of commerce were dissolved in a very large amount of water, and
the solution was twice successively treated with a very small amount
of potassic hydroxide in dilute solution ; the mixture being occasion-
ally shaken and allowed to stand each time for a week before decant-
ing the clear supernatant liquid. One quarter of a gram of amnionic
bromide was added to this liquid ; and after standing two weeks the
solution was carefully filtered, evaporated to very small bulk in a por-
celain dish, and again filtered from the deposited basic salt. The
finely divided crystals obtained from this solution by agitation and
cooling with ice were drained with a reverse filter,* redissolved in
hot water, and repeatedly recrystallized in platinum vessels. The
first three mother liquors were rejected, but the later ones were all
combined. After three more recrystallizations the substance con-
tained in these mother liquors was dissolved in cold water and allowed
to crystallize by slow evaporation in pure air. The prejiaration thus
made was used in the second series of analyses.

The crystals obtained by eight successive crystallizations in platinum
vessels were dissolved in cold water, and the salt was slowly recrystal-
ized over sulphuric acid in a vacuum. The substance thus obtained
was dissolved in water which had been distilled in a platinum retort,
and after standing six days the perfectly clear solution was again
brought to crystallization in a vacuum. These crystals were dried in
a platinum dish over partially dehydrated cupric sulphate, and were
used for the third series of experiments. The neutrality of each of
these preparations was determined as nearly as possible by means
of methyl orange, after the manner described in these Proceedings,
Volume XXV., page 201.

Sodic Carbonate. — As sodic carbonate formed one of the chief
bases for the determination of the sulphuric acid, its manufacture in
a pure state was a matter of great importance. From the " chemicallv
pure" material of commerce one may easily procure by five recrystal-
lizations a material which gives no coloration with ammonic sulphy-

* These Proceedings, XII. 124.

246 PROCEEDINGS OF THE AMERICAN ACADEMY

drate, nor, after neutralization, with potassic sulphocyanide.* Sodic
sulphate and chloride are even more readily separated than traces of
iron.

Three separate preparations of sodic carbonate were used in the
course of the work. For the first series of comparatively crude ex-
periments the solution of ordinary " chemically pure" material, which
almost invariably contains a perceptible amount of finely divided solid
matter, was after filtration twice recrystallized in a platinum dish. It
was subsequently found advantageous to prevent the clogging of the
filter by two or three preliminary recrystallizations, in the course
of which most of the solid matter finds its way into the mother
liquor.

The second sample of sodic carbonate was recrystallized six times
after filtration. The vessels, rod, and reverse filter were all of plat-
inum, and water was used which had been distilled in a platinum
retort.

Besides all these precautions, several additional ones were taken
in the preparation of the purest sample of sodic carbonate designed
for the third series. The water used had been distilled four times :
first alone, next over potassic permanganate, then over acid potassic
sulphate, and finally in a platinum still without the addition of foreign
matter. Every precaution was taken to exclude dust and acid fumes,
and the sodic carbonate was recrystallized ten times after filtration.
The salt prepared in this way gives an absolutely odorless warm con-
centrated solution. The second and third samples gave essentially
identical results. Each of the three [)reparations was pure snow-
white, whether fused or unfused ; each gave a perfectly clear and
colorless solution with water ; and not even the first save the least
test for hydrochloric or sulphuric acid.

On the other hand, as Stas has already indicated, the most elaborate
precautions are unable to free sodic carbonate wholly from traces of
silica and a basic oxide, which is probably alumina. The former
impurity was determined in the usual manner. The alumina and
silica were determined together in new portions of the salt by exact
neutralization witii hydrochloric or sulphuric acids, ignition of the
evaporated product, and weighing of the insoluble residue. Phenol
phthalein was used to determine the neutral point, and hence the

* Stas appears once or twice to liave had more ilitticulty in accomplishing
this result. At other times his experience seems to have coincided with tliat
described above. Compare Aronstein's translation (1867), pp. 112, 270, with
p. 275.

OF ARTS AND SCIENCES. 247

insoluble bases present were necessarily precipitated. It is needless
to state that the operations were conducted in platinum vessels. Ne-
glecting two determinations which were vitiated by known impurity,
the weight of total residue obtained varied from 0.15 to 0.41 milli-
gram, while the weight of sodic carbonate taken varied from 1.02 to
3.06 grams. The average of twenty determinations showed about
thirteen parts of impurity to exist in one hundred thousand parts
of the salt. Silica constituted about forty per cent of this impurity.
The third specimen did not ditit^r essentially from the second on these
averages. Owing to the slight solubility of silica in sodic chloride
solutions,* the observed amount of this impurity may be slightly too
low, but for the present purpose the correction is unimportant. The
filter papers used in this part of the investigation left upon ignition an
ash of 0.00004 gram.

Since the silica and alumina are probably present as sodic silicate
and aluminate in the original carbonate, and may be considered as
replacing carbon dioxide molecule for molecule, it is evident that the
correction to be applied will not equal the whole weight of the residue.
Upou this assumption the correction when phenol phthalein is used as
an indicator amounts to about half the weight of the impurity, or
about 0.007% of the weight of the sodic carbonate. Toward methyl
orange aluminic hydroxide is alkaline ; hence in this case the errors
due to the two impurities tend to counterbalance each other, and the
correction may be omitted. The presence of a small amount of
alumina thus explains the slight difference observed later between the
results obtained with these two indicators.

The crystallized sodic carbonate was dried over pure boiled sulphuric
acid in a vacuum, and subsequently ignited to a dull red heat in a
double crucible over a Berzelius spirit lamp. The latter apparatus
was employed in order to avoid the possible introduction of impurity
from illuminating gas. It was found that the salt could be ignited to
perfectly constant weight at any temperature between dull redness
and its fusing point. The light powder contracts and " sinters to-
gether " at a low red heat, and then remains essentially unchanged
in weight until it melts. The unfused salt was not observed to gain
perceptibly in weight upon an hour's exposure to the air of the bal-
ance case ; upon the other hand, the fused salt, which of course was
never used as a basis of determination, was markedly hygroscopic.

* Stas, Aronste'm's translation, p. 279. In tliis place the correction is
applied to sodic chloride, but not to the nitrate formed from it.

248 PROCEEDINGS OF THE AMERICAN ACADEMY

Acids. — The chemically pure hydrochloric and nitric acids were
each distilled three times successively in a platinum still. Large
quantities evaporated in open dishes left unweighable and scarcely
visible residues. The insignificant amount of the residues was a proof
of the purity of the air, as well as of the acids. Neither acid gave a
test for iron, and the nitric acid contained no chlorine even at the
beginning of the operations.

Since the platinum still was very large, it was found more con-
venient to distil sulphuric acid from a small hard glass retort heated
by a ring burner, immediately before it was needed. After three
such distillations, a preparation was obtained which left no weighable
residue upon the evaporation of any quantity used in the work.

All sulphuric acid used for drying was boiled with ammonic
sulphate.

In order to keep the air of the laboratory quite pure, during the
greater part of the time neither volatile acids nor ammonia were
allowed in the room devoted to the investigation, and all available
precautions were taken against dust.

Indicators. — As is well known, the salts of most of the heavier
metals are acid toward phenol phthalein and neutral toward methyl
orange. The most important consequence of this relation has already
been pointed out. When the former indicator was used, the slightly
acidified solution was of course first freed from carbonic anhydride
by long continued heating upon the steam bath, and the end point was
determined with pure caustic alkali. It is a noteworthy fact, that
methyl orange is useless in very concentrated solutions of sodic sul-
phate, the color change becoming apparent only upon dilution. This
indicator is naturally less serviceable in the presence of a strong color
like that of a copper salt. It is less sensitive with cupric sulphate
than with cupric bromide. In such cases as these, colorimetric com-
parison can alone afford accurate results ; but even here the effect of
a personal equation must be more or less perceptible.

Since the amount of either indicator used in any one case was not
more than the thirtieth of a milligram, it could not have seriously
influenced subsequent operations with the solution.

Water. — The distilled water taken hot from the tin-lined condenser
around a steam-drying oven contained no ammonia discoverable by
Nessler's reagent, and in the first crude experiments this water was
used without further treatment. After having been once more dis-

OF ARTS AND SCIENCES. 249

tilled in a platinnra still, it appeared to be very pure. The first por-
tions of the distilhite wei-e rejected, although no ammonia was found
in them. The water left absolutely no trace of residue upon evapora-
tion, even after long standing. Such doubly distilled water was used
in the second series of experiments.

For the third series all the water used, even for minor operations,
was distilled four times : first alone, then over alkaline potassic per-
manganate, next over acid potassic sulphate, and finally once more
alone in the platinum still. Since this water was not different in
any of its properties from the second preparation, the last refinements
were probably unnecessary ; and in later work water which had been
distilled only twice was used.

The value of the present work has been largely increased by the
unlimited supply of platinum ware placed at the disposal of the writer
through the kindness of Professor Cooke. Because of the large size
of some of the retorts and bottles, it was impossible to free these vessels
from iron after the method of Stas. Accordingly they were digested
alternately with strong hydrochloric and nitric acids until pure acid
which had remained in them for days gave no test for iron after
evaporation upon tiie steam bath.

Smaller vessels were freed from iron in the usual manner, by treat-
ment with the vapors of aramonic chloride at a red heat, as well as
with fused acid potassic sulphate. They were usually protected dur-
ing ignition by an outer crucible, from which they were separated by
a coil of platinum wire. Crucibles weighing twenty grams rarely
showed a variation of more than one twentieth of a milligram between
the weights taken before and after any ordinary operations, except-
ing when ignited with cupric oxide under conditions which will be
explained later.

Description of Analyses.

First Series.

It is evident that the most probable constant error in Hampe's
work lay in the great hygroscopic power of cupric sulphate dried at
250°. A possible means of detecting such error existed in the use of
the crystallized salt as a starting point. This salt was therefore
coarsely powdered, dried to constant weight over a mixture of crys-
tallized and partially dehydrated cupric sulphate, and analyzed by
electrolysis in the usual fashion. Although the crystalline powder
must have contained occluded water, the quantity of copper found in

250

PROCEEDINGS OP THE AMERICAN ACADEMY

several samples from different sources was much larger than that
demanded by the old atomic weight, and closely approached the
theoretical quantity based upon the new value.

In the third experiment the weight of water lost by heating the
salt to constant weight in the manner described by Hampe * was also
determined, and this amount proved to be nearly three tenths of a per
cent less than it should have been according to any hypothesis. Since
larger crystals of cupric sulphate exposed to air above the desiccating
mixture already mentioned lost none of their lustre, and since the salt
under these conditions comes to perfectly constant weight, it seemed
highly probable that the apparent deficiency of crystal water was due
to occlusion in the so-called anhydrous salt, and not to any previous
efflorescence of the crystals.

The sulphuric acid produced during the electrolysis f was preserved
in each case, and approximately determined by means of the first prep-
aration of sodic carbonate ; also afterwards by the evaporation of the
solution of sodic sulphate. Since the manipulation was more or less
imperfect, and the materials and water not quite pure, it is thought
unnecessary to include these somewhat bulky data in full here, but
the evidence which they furnished was conclusive upon one point.
The percentage of acid was not equal to the complement of the other
percentages. In other words, over one tenth of a per cent of material
liad not been determined.

Found.

Theory
Cu =: 63.60.

Theorv
Cu - 63.33.

Percentage of Water . . .
" Copper . .
" S()4 . . .

35.958
25.455
38.46±

86.0695
25.4665
38 461

36.109
25.385
38.506

Total

09.87

100.000

100 000

It is seen that we are dealing, not with infinitesimals, but with grave
error evident to the crudest analysis. The reason for the deficiency
has already been suggested, but the proof is yet wanting.

* Loc. cit.

t In tliis ronnection it may he noted tliat Hart and Croasdale liavo quite
independently used tiie sulphuric acid tims formed as a basis of alicalinietric
analysis. (Cliem. News, LXIII 93, 1891.) The idea appears to be a good one

OP ARTS AND SCIENCKS.

251

The ratio between the copper and the sulphuric acid, or rather
between the copper aiul sodic carbonate on the one hand and sodic
sulphate on the other, affords a new twofold basis for the calculation of
the atomic weight of copper. This basis is entirely independent of the
always uncertain weight of the cupric sulphate and the objectionable
method of calculation from difference. The results correspond to the
new atomic weight, but the complete discussion of this phase of the
subject will be reserved until more definite data have been given.
It is interesting to note that these comparatively crude experiments
differ only by one or two units in the second decimal place of the
result from the far more carefully executed work yet to be described.

Analysis of Cupric Sulphate. — First Series: Data and Results.

Welgltts reduced to Vdcuum Slandard.

No. of
Experiment.

Weight Cupric
Sulphate
(cry St, ).

Weight CuSOi
(250^) found.

Weight Copper
found.

Per Cent Water

lost at 250^.

Per Cent
Copper.

1
2
3

2.8815
2.7152
3.4639

2.2184

0.7337
0.6911

0.8817

• • • •

• • • •

35.958

25.462
25.452
25.454

Avenige . . .

35.958

25.455

Second Series.

Although from the earlier experiments it appeared that the defi-
ciency in the sum of the analytical results might be explained by the
hypothesis of the existence of water in the so-called anhydrous cupric
sulphate, it was possible that at least a part of this deficiency might be
due to experimental error. The next step of the research was therefore
to start again upon a more refined basis. It was important, too, that
some proof of the above mentioned hypothesis more definite tlian an
indirect inference should be obtained. The cleare.st idea of the various
operations may probably be acquired through a detailed statement of one
of the experiments, and accordingly such a statement is given below.

Experiment 4. — The cupric sulphate,* having been cautiously
powdered in an agate mortar and dried to constant weight over the
desiccating mixture already described, was gradually heated to 255°
in a platinum crucible with a very closely fitting lid.

* See page 245 of this paper.

252 PROCEEDINGS OF THE AMERICAN ACADEMY

Grams.

Weight of crucible + CUSO4 . 5 HgO = 23.36188

" " alone = 20.30300

" cupric sulphate in air := 3.05888

Correction to vacuum (Sp. gr. = 2.274) = .00118

Corrected weight CUSO4 . 5 HgO, iu vacuum = 3.06006

Weight crucible + CUSO4 dried for three hours at 255° = 22.26243
" " dried for two hours more = 22.26233

The crucible and contents gained in weight upon the balance pan
at a rate of only one tenth of a milligram in four minutes, and hence
the last weighing is without doubt sufficiently accurate. When the
weight had thus become constant, twenty drops of strong pure sulphu-
ric acid were added to the white powder, and the whole was kept at a
temperature of 365° for three and one half hours in a suitable air bath
composed of a nest of porcelain crucibles. An air thermometer was
used in measuring the temperature.

Grams.

Weight of crucible and contents = 22.25905
The same, after reheating for four hours with six drops

more of 11.^0^ = 22.25900

Weight in air of cupric sulphate dried at 365° = 1.95600

Weight in vacuum of cupric sulphate (Sp. gr.* = 3.61) = 1.95637

Loss of weight between 255° and 365° = .00333

The crucible and contents gained in weight even less rapidly than
before.

This experiment shows that the so-called anhydrous cupric sulphate
of Hampe loses about seventeen one-hundredths of a per cent of its
weight when subjected to the temperature of the boiling point of mer-
cury ; a loss nearly equal to the deficiency which we are seekinj; to
explain. The residual salt may be slowly dissolved in water without
the need of very great caution. It gives a perfectly clear solution,
which is neutral to methyl orange and does not deposit basic salt even
upon indefinite standing. These facts alone show that the decrease
in weight was not due to loss of acid, but a much more definite proof
of this point is given later.

For electrolysis, the clear solution of the 1.95637 grams of cupric
sulphate was transferred to a large platinum crucible. This crucible
had been previously coated inside with pure copper, washed with

* Tliis is a mean of the results of Hampe and Karsten.

OP ARTS AND SCIENCES.

253

water and alcohol, and dried at 103°. The crucible was supported
upon a platinum triangle in the interior of a large breaker, and was
covered by a watch glass into 'which was fused the positive electrode.
The wires conducting the current into the beaker were of platinum.
The whole apparatus was covered with a large watch-glass, which
effectually excluded impurity. Several times before the conclusion
of the electrolysis the drops adhering to the inner watch-glass were
washed back into the crucible by means of apparatus especially
adapted for the purpose. After the maintenance for 120 hours
of a weak current,* the electrolysis was interrupted, and every trace
of the acid iu the crucible was washed as rapidly as was consistent
with scrupulous care into a large platinum bottle. The crucible was
then washed with pure alcohol, dried at 103°, and finally weighed.

Crucible + Cu before electrolysis (Bar. = 767 ; t° = 20°)
+ additional Cu after (Bar. = 750 ; t° = 20°)

Apparent weight of copper

Correction for change in temperature and pressure
" to vacuum

True weight of copper in vacuum

Percentage of copper in CUSO4 . 5 HgO

" CUSO4 dried at 255°
" " " " dried at 250° (Hampe)

" " « " dried at 365°

Grams.

= 36.12120
= 36.90000
= 1).77880
= + .00007
= — ■00001
= .77886

Per Cent.

= 25.452

= 39.744

= 39.725

= 39.811

Further comment here is unnecessary.

The earlier experiments had already shown approximately the
equivalent ratio of sodic carbonate to copper. Accordingly a very
little less than the amount of sodic carbonate corresponding to the
data above was carefully weighed out from the second sample of this
material, and the salt was ignited at a dull red heat until constant
in weight. f

Weight crucible -f- NaoCOg

" " alone

((

sodic carbonate in air

" " in vacuum (Sp. gr.

Grams.

= 21.60095

= 20.30260

= 1.29835

2.466) = 1.29880

* Tliese Proceedings, XXV. 202.

t See description of sodic carbonate, page 247 of tliis paper.

254 PROCEEDINGS OF THE AMERICAN ACADEMY

The sodic carbonate was dissolved in water and poured into the
very dilute solution of sulphuric acid which had been formed by the
electrolysis. The lid of the crucible, as well as that vessel itself,
was very thoroughly washed with water. On account of the gieat
dilution of the liquid, carbonic anhydride was not at once set free,
but upon warming for six or eight hours upon the steam bath it was
easily expelled. During this operation the platinum flask was of
course suitably covered, a Gooch crucible being found most con-
venient for this purpose.

When gas evolution had long ceased, the solution was wholly trans-
ferred to a large platinum dish, and was evaporated on the steam bath
to a volume of about fifty cubic centimeters. The very slight excess
of acid was then titrated by means of sodic hydroxide, using metiiyl
orange as an indicator, and was found to require 0.09 cubic centimeter
of a decinormal solution for neutralization. The probable error of
the end point was not greater than one drop, or one ten-thousandth
part of the sodic carbonate present.

Grams.

Sodic carbonate weighed out = 1.2988

" " added in titration = 0.0005

" " required (in vacuum) = 1.2993

Cupric sulphate taken (CuSOi . 5 H2O) = 3.06006

Copper found in " « = .77886

Per cent of SO4 in CuSOi . 5 H2O = 38.439
Atomic weight of copper :

NaaCOs : Cu = 106.108 : 63.606

In transferring the sodic sulphate from the evaporating dish to the
capacious platinum crucible in which it was to be weighed, the pre-
caution of washing with water which liad been thoroughly boiled was
absolutely essential to prevent the possibility of subsequent mechan-
ical loss during evaporation. At best, the concentration of a solution
in an open crucible over the steam bath is an unsatisfactory operation,
and accordingly for the third series a new method was devised. The
water left the sodic sulphate very slowly ; but in the end it evaporated
so completely that after covering there was a scarcely audible decrepi-
tatiim on heating the crucible to redness. The pure white sodic
sulphate was finally fused at a bright red heat. During this last opera-
tion the salt occasionally became tinged with a slightly yellowish hue,
due probably to a trace of iron from the platinum vessels ; but the
purest specimens remained wholly colorless. The salt lost only one

OF ARTS AND SCIENCES.

255

or two tenths of a milligram duriug the fusion. A drop of very dihite
permanganate solution was not decolorized by the solution of tiie
residual salt, showing that no reduction had taken place. A very small
amount of insoluble residue, consisting of the apparently unavoid-
able impurity in the sodic carbonate together with any iron or copper
which might be present, was determined in each case ; and the amount
was subtracted from the weight of sodic sulphate. Naturally, besides
this, it is necessary to subtract the amount of salt corresponding to the
acid added in titrating back and forth when determininsf the end
point of the acidimetric reaction. This end point had been taken as
the average of a number of readings. Below are the data of the
experiment under discussion.

Weight of crucible + NagSOi (etc.)

" " alone

Uncorrected weight of NagSO^
Subtract weight of salt added in titration

" " insoluble residue

Corrected weight of Na2S04 in air
Correction to vacuum (Sp. gr. = 2.631)
Weight Na2S04 in vacuum

Grams.

= 18.67040
= 16.92594

1.74446

.00362

.00034

1.74050

+.00063

= 1.74113

Per cent of SO4 in CUSO4 . 5 HgO = 38.445

« " " " found above = 38.439

Atomic weight of copper (if Na2S04 = 142.166, from table) :

Weight NaaSOi : weight Cu = 142.166 : 63.595.

No copper was found in the solution of the sodic sulphate, and a
minute trace only in the insoluble residue.

The object of the second experiment of the series was to determine
whether the method of weighing cupric sulphate in a tightly covered
crucible was an accurate one. The determination was in every respect
modelled after the mode of procedure adopted by Harape, a tightly
stoppered weighing-bottle being used to contain the salt. The result
sufficiently confirmed the previous experiment. For later experiments
a crucible was accordingly used, since the evaporation of sulphuric
acid at high temperatures is much more feasible in such an apparatus.
A mishap prevented the accurate determination of the acid.

For the third experiment, a much larger amount of material was

256 PROCEEDINGS OP THE AMERICAN ACADEMY

used, the method being essentially that of the first experiment. The
large crucible of one hundred and fifty cubic centimeters' capacity
used for the electrolysis was not previously coated with copper. The
end point of the acidimetric determination was taken with both phenol
phthalein and methyl orange, the color change with the latter indicator
being rather unsatisfactory in the presence of so large an amount of
dissolved substance. The amount of insoluble residue found in the
sodic sulphate was only 0.0003 gram; it contained no trace of copper.
The results showed that little was to be gained by the use of larger
quantities of substance, since the unavoidable errors of quantitative
work were multiplied nearly in proportion to the quantities of mate-
rial. Mechanical errors are at the present day inessential compared
with the constant ones which complicate so many analyses. It is
obvious that a far more accurate result could be obtained from four
tenths of a gram of really anhydrous cupric sulphate, than from four
hundred grams of a substance still retaining nearly two tenths of a per
cent of water.

Adding in the small amount of water which is lost by cupric sul-
phate between 255° and 365", the percentage composition of the
crystallized salt, as indicated by the second series of analyses, gives
a much more satisfactory total result than before : —

Water lost at 255° = 35.960

Additional water lost at 365° = 0.108

Copper =: 25.450

Sulphuric acid radical = 38.436

Total = 99.954

Although greatly improved, the analysis still leaves much to be
desired. Even in materials prepared with the described precautions,
traces of impurity were manifest. In the last experiment, where a
large amount of material was used, a few minute spots, which might
have been arsenic, appeared on the clear surface of the electrolytic
copper, and traces of a brownish coloration were observed upon the
positive pole. Both of these impurities were so infinitesimal in
amount that they could not reasonal)ly have been expected seriously
to affect the final result; but the thought that they might be responsi-
ble for a part of the remaining deficiency prompted the execution of a
still more elaborate series of experiments. The data and results of
the second series are appended : —

OF ARTS AND SCIENCES.

257

Analysis of Cdpric Sulphate. — Second Series: Data.
Weiffhts reduced to Vacuum Standard.

No. of
Experi-
ment.

CuSOiSUiO

taken.

CUSO4

(260)
taken.

CUSO4

i360°)
found.

Metallic
Copper
found.

Sodic Carbonate
required.

Sodic

Sulphate

found.

Methyl
Orange.

Phenol
Phthalein.

4

5
6

grams.

3.0G006
2.81840
7.50490

grams.
1.9597

1.8048

4.8064

grams.

1.95(337

• • • •

4.79826

grams.

0.77886

0.71740
1.90978

grams.
1.2993

3.1859

grams.
3.1865

grams.

1.7411
4.2679

Second Series: Results.

No. of
E.^peri-
nient.

Water lost

at 260°.

Additional
loss at
360°.

Total Per

Cent of

Water (360°).

Per Cent
of Copper.

Per Cent of

SO4 by NajCOa

(average).

Per Cent of

SO^by

NajSOi.

4

5
6

35.950
35.964
35.957

0.109

• • • >

0.108

36.068

36.065

25.452
25.454
25.446

38.439
38.435

38.444
38 424

Averages

35.960

0.108

36.067

25.450

38.437

38.434

Third Series.

It is only by gradual approach that very accurate quantitative work
may be realized, and the earlier series were absolutely necessary as a
training and preparation for the present one. Tliis series had for its
object, not only an investigation of the effect of increased refinement,
but also a definite proof that the amount of acid present in cupric
sulphate was unaffected by the operations involved in the dehydration
of the salt. To this end the first sample was not dehydrated at all, the
second was heated only to 250°, and the third was exposed to the full
heat of 365°, after the addition of sulphuric acid. Since the percent-
age of acid found in these samples did not vary beyond a reasonable
limit of error, the proof is satisfactory.

The additional precautions taken in the preparation of the materials
have already been mentioned under appropriate heads. Besides these,

VOL. XXVI. (n. S. XVIII.) 17

258 PROCEEDINGS OP THE AMERICAN ACADEMY

many refinements of manipulation were employed, too numerous to be
fully described. For example, pbospliorous pentoxide was used as a
drying agent in the desiccator, and immediately after the introduction
of a hot crucible the air was exhausted with a good air-pump and
readmitted through a series of drying tubes. This method of course
absolutely prevents the absorption of moisture by the contents of the
crucible, while cooling.

Again, in the ninth experiment, the solution of the perfectly neutral
sodic sulphate was transferred to a small flask, heated to 103°, and
evaporated by means of a current of pure dry air. When dry, the
sulphate was gently ignited and weighed in the flask, and subsequently
fused in a platinum crucible. The flask lost 0.00010 gram during the
operations, and the sodic sulphate lost 0.0003 gram upon fusion.
With the exception of the small flask, all the apparatus was of plati-
num. The sodic sulphate formed in the second experiment, and the
sodic carbonate required in the third, were not determined.

The last experiment of the series was not of the same grade of
refinement with the others. It is only included here because the
same preparation of cupric sulphate was used in its execution.

The determination of the sulphuric acid by still another method was
the object of this experiment. Baric sulphate was precipitated from a
boiling solution of cupric sulphate strongly acidified with hydrochloric
acid. After weighing as usual, the perfectly white precipitate was
fused with pure sodic carbonate, and the resulting cake thoroughly
lixiviated with boiling water and dilute sodic carbonate solution. In
the filtrate, the chlorine, which had been originally present as baric
chloride occluded in the sulphate, was determined in the usual manner.
Traces of baric sulphate and argentic chloride must have been dis-
solved in the wash water, but these errors tend to counterbalance one
another, and hence no correction was made for them. Finally, the
weight of baric chloride, calculated from the amount of chlorine found,
was subtracted from the weight of baric sulphate ; and from the cor-
rected weight thus obtained the percentage of sulphuric acid in cupric
sulphate was calculated. The astounding agreement of this experi-
ment with the others may be nothing more than accident. In any
case, the uncertainty of the atomic weight of barium, and the very
unsatisfactory nature of baric sulphate, combine to make a repeti-
tion of the experiment of little value for the present purpose ; but the
method seems to be of value, and will form a subject of future inves-
tigation in this Laboratory. It is well known that baric sulphate has

OF AETS AND SCIENCES.

259

great power of occluding many salts, but few experimenters seem to
have realized that the occlusion of most metallic sulphates tends to
decrease the amount of precipitate obtained. This fact was recognized
by Professor Jannasch and the writer* in 1889, and would influ-
ence the controversy between Ostwaldf and Kruss.J The occlu-
sion of baric chloride of course increases the amount of precipitate
when sulphuric acid is to be determined, and diminishes it when
barium is to be determined. With care this last error may be reduced
to a very small amount, but it is doubtful if it has ever been wholly
avoided. A trace only of copper was found in the precipitate ob-
tained in the experiment described above, showing that cupric chloride
is not occluded to any essential extent.

In spite of the fact that the important analyses of the third series
were far more carefully performed than those of the second, the result
was no more satisfactory than before. It was therefore apparent that
a point had been reached beyond which further refinement was unavail-
ing, and that the reason of the discrepancy must be sought, not in
accidental or variable impurity, but rather in some property inherent
in the purest cupric sulphate.

Analysis of Cupric Sulphate. — Third Series: Data.
Weights reduced to Vacuum Standard.

No. of
Experi-
ment.

OusOi.snjO

taken.

CUSO4

found.

Metallic Cu
found.

Na,C0, found

by Metli3'l

Orange.

NaXOs found
by Phenol
Phthaleia.

Sodic

Sulphate

found.

7
8
9

grams.

2.88307

3.62913
5.81352

grams.

at 260=
2.82373

at 370=±
3.71680

grams.
0.73380

0.92344

1.47926

grams.

1.2242
1.5407

grams.

1.22435

1.54080

....

grams.

1.68994
3.30658

CUSO4 . 5 H.O
taken.

Baric Sul-
phate found
(uncor.).

Argentic

Chloride

found.

Baric

Chloride
calculated.

Baric
Sulphate
corrected.

10

grams.

3.1902

grams.
29967

grams.

0.0284

grams.
0.0206

gram.s.
2.9761

* J. fiir prakt. Chemie, XXXIX. 321 (1889).
t Lehrbuch der Allgem. Chem., I. 53.
1 Annalen, CCLXII. 40.

260

PROCEEDINGS OF THE AMERICAN ACADEMY

Third Series : Results.

No. of Ex-
periment.

Water lost
at 260°.

Water lost
at 370°±.

Copper
found.

SO4 found from
NajCOs (average).

SO4 found
from NajSOi.

7
8
9

35.970

• • • •

36.067

25.452
25.446
25.445

38.443
38.435

38.433

38.431

10

From Baric
Sulphate.

38.434

Average

35.970

36.067

25.448

38.439

38.433

Percentage Composition of Cupric Sulphate.

Second Series.

Third Series.

Theoretical.

Cu = 63.60.

Cu = 63 33.

Water

Copper

SO4

36.068
25.450
38.436

36.067
25.448
38.436

36.0695
25.4665

38.464

36.109
25.385
38.506

99.954

99.951

100.000

100.000

The Cause of the Deficiency.

It was by no means easy to trace the cause of the disappearance of
less than one two-thousandth part of the material to its true source.

The notable agreement between the percentages of water given in
the first three columns of figures immediately above of course sug-
gested the possibility of a loss of copper and sulphuric acid during the
electrolysis ; and it became an important matter to test the point.

Accordingly weighed portions of a dilute solution of pure sulphuric
acid were estimated with weighed amounts of sodic carbonate. Other
portions of the same solution, weighed alternately with the previous
ones, were then evaporated over the water bath with small amounts

OF ARTS AND SCIENCES.

2C1

of cupric nitrate and successive additions of water. The cupric sul-
phate formed by this treatment was wholly decomposed by electroly-
sis, and the resulting sulphuric acid was determined exactly in the
usual manner. The comparison of the acid solution before and after
electrolysis is capable of showing at once whether any sulphuric acid
was mechanically or chemically lost during the process. If any nitric
acid had been held by the sulphuric acid in the second case, it could
not have escaped decomposition during the electrolysis.
Following are the results : —

Strength of H.2SO4 before Electrolysis.
Weights reduced to Vacuum.

Number of
Experiment.

Solution taken.

NajCOa required.

Na.,C03 for
10.000 grams Solution.

11

12

grams.
5.5932

11.4175

grams.
1.0175

2.0768

grams.
1.81917

1.81896

Average . . .

1.81906

Strength of H2SO4 after Electrolysis.
Weights reduced to Vacuum.

Number of
Experiment.

Solution taken.

NajCOs required.

Na.,C03 for
10.000 grams Solution.

13
14

grams.
16.8214

5.5612

grams.
3.0592

1.0115

grams.
1.81865

1.81886

Average . . .

1.81876

Difference between averages

0.0003.

Known mishaps tended to make the first experiments of each of
the two groups respectively too high and too low. It is seen that the
second experiments gave nearly identical results. Considering the
increased transference and manipulation involved in the second group,

262

PROCEEDINGS OF THE AMERICAN ACADEMY

the coaclasion seems to be warranted that little if any sulphuric acid
is lost during electrolysis.

In order to prove like relations with regard to copper a similar
method was adopted. Weighed amounts of electrolytic copper were
dissolved in pure nitric acid in a flask provided with bulb tubes for
the condensation of spray. The cupric nitrate was then evaporated
with an excess of sulphuric acid upon the water bath, and the cupric
sulphate was electrolyzed as usual.

Electrolysis op Copper.

Number of
Experimeat.

Copper taken.

Copper found.

Loss of Metal.

15
16

grams.
1.24156

1.00613

grams.

1.24139
1.00600

grams.
0.00017

0.00013

Here again, the manipulation was so much more elaborate than in
a simple electrolysis that it is difficult to decide where the slight loss
took place. In another experiment (No. 17) both copper and sul-
phuric acid were weighed before and afterward, but the experimental
operations were so doubly involved that the determination was of very
doubtful value. The result was not materially different from the
others, however. The significant portions of the data are given in
anotlier place.*

In the course of these experiments it was found that a compact and
brilliant copper film may be safely washed by decantation if sulphuric
acid alone is present. Indeed, during most of the preceding deter-
minations this method was adopted to insure complete collection of
the acid, and in only one case (P^xperiment 7) was as much as one
twentieth of a milligram of copper found in the filtrate. In this case
that amount was added to the larger quantity, and the sum is given in
the table. Potassic ferrocyanide was the reagent used in the colori-
metric tests.

Assuming the deficiency of copper and acid as found above to
represent a real loss during electrolysis, the composition of cupric
sulphate would be : —

* Page 268.

OP ARTS AND SCIENCES.

263

Water .
Copper
SO4 .

Uncorrected.

36.007
25.449
38.435

99.951

Corrected.

36.067
25.452
38.442

99.961

That is to say, an uudoubtedly excessive value for this correction
would account for only one fifth of the deficiency for which explana-
tion was sought. The correction is at best very doubtful, and it is
not applied in the final calculation of the atomic weight. Its appli-
cation would make no essential difference in the final result, but
would render the individual variations much less marked.

The only probable interpretation of the deficiency now lay in the
assumption that cupric sulphate still held a volatile impurity at 360°.
Such an amount of any non-volatile foreign substance would have
increased the weight of sodic sulphate found at the conclusion of the
analysis by about eight one-hundredths of one per cent, an amount
entirely too large to escape detection. Besides, such a source of error
had already been shown to be unlikely.

After it had been experimentally proved that cupric sulphate had
no tendency to hold ammonic sulphate at 370°, the hypothesis of the
occlusion of a small amount of water became the last resource. At
first sight it seems improbable that any material could hold appreciable
quantities of water in the presence of sulphuric acid when so strongly
heated. But it must be remembered that at this temperature the acid
is dissociated, and water is actually present in the vapor. It was
hoped that sulphur trioxide might be more efficient as a dehydrating
agent, but a single experiment (No. 18) showed that cupric sulphate
possessed more affinity for water even at 300° than did the sulphur
trioxide. 2.3787 grams (in vacuum) of cupric sulphate lost 0.8554
gram in weight on heating to about 300° in a glass tube under a
current of dry air. Upon continuing the application of heat in air
charged with sulphuric anhydride from Nordhausen acid, no essential
change in weight was observed. The apparatus was somewhat com-
plicated in order to avoid rubber connections, but a description of it is
superfluous. The total loss was 35.962 per cent of the weight taken.

264 PROCEEDINGS OP THE AMERICAN ACADEMY

Experiments showed that between 360° and 400° cupric sulphate
does not essentially lose in weight. For example (No. 19 or 23),
1.28563 grams of the substance dried at the former temperature
weighed 1.28558 grams after heating for ten hours at 400°. At the
temperature of dull redness the salt slowly decomposes.

If the loss of weight during this decomposition could be compared
with the deficiency of sulphuric acid in the residue, it is evident that
an indirect means would be at hand for the detection of a possible
simultaneous loss of water. Such a circuitous road seemed to be the
only one open to the present search.

In the series of experiments tabulated below, pure cupric sulphate
was heated in the usual manner to 250°, and then after the addition of
sulphuric acid to 400°, until the salt became constant in weight. The
heat was then increased to dull redness, and after cooling the loss of
weight was determined. Upon solution of the residue in water a small
amount of basic salt was naturally deposited. Since the filtrate was
apparently quite normal, it is evident that the amount of sulphuric
acid necessary exactly to dissolve this precipitate must have been
equivalent to the anhydride driven off.

After standing a considerable time, the basic salt was carefully
filtered off and a measured amount of twentieth normal acid was used
for its solution. The crucible used for the ignition was also washed
with a measured quantity of acid. The clear solutions were all com-
bined, and the excess of acid was determined by sodic hydroxide and
methyl orange. Since the loss of weight upon ignition was noticeably
more than the amount of anhydride corresponding to the quantity of
sulphuric acid used to dissolve the basic salt, something beside sulphur
trioxide must have been expelled by the heat.

The last column of the table gives the difference between the loss
of weight of the cupric sulphate and the amount of sulphuric anhy-
dride required to dissolve the basic salt, expressed in percentage of
the original crystallized compound. This difference probably repre-
sents a small amount of water held even at 400°. The results showed
a very noticeable variation, and at first sight appeared somewhat un-
satisfactory. The first two experiments agreed well with each other,
and were apparently very trustworthy. In the third, on the other
hind, the amount of basic salt was so large as greatly to interfere
with the accuracy of the result. In Experiment 23 the lossr of weiglit
upon ignition was so very large, amounting to about three per cent of
the anhydrous sulphate, that the experiment was rejected. After
such a failure, it was natural that the next sample should not have

OP ARTS AND SCIENCES.

265

been heated enough. This result is hence probably too low. Tlie
last experiment was more carefully regulated, and is more trustworthy.
In this case the cupric sulphate was heated for an hour at very dull
redness.

Action of Heat upon Cdpric Sulpuate.

Weights reduced to Vacuum.

IS"

d

^5

1 — 1

o

S a

O

bCOQ

Weight after
Ignition at
dull Redness.

Loss between
400° and
Redness.

n'i, normal H.^S04
required for Basic Salt.

Per Cent of
CuS04.5H.,0
unaccounted
for.

Uncor-
rected.

Corrected.*

20

grams.

grams.
1.5822

grams.
1.57945

grams.
1.57325

milli-
grams.
6.2

cubic
ceutim.
2.65

cubic
centim.

2.65*

0.044

21

2.71828

1.7413

1.73886

1.73130

7.06

2.95

2.85*

0.050

22

4.7325

—

3.0258

2.9928

33.0

16.86

[15.86*?]

[0.030?]

23

2.01084

1.2880

1.2856

1.2504

35.2

—

—

—

24

7.211

—

4.6075

4.6055

2.0

0.33

0.32*

0.019

25

3.979

—

2.5443

2.5403

4.0

0.93

0.90*

0.055

Total

Average
ige omitti

0.040

Avera

ng22

0.042

The filtrates from the preciftitates of basic salt did not deposit any
further solid upon long standing. That from the last determination
was rendered very distinctly acid to methyl orange by the addition of
a milligram of sulphuric acid, in spite of the difficulty in detecting the
color-change in the presence of the blue cupric sulphate. These tests
indicate that the filtrate was wholly normal.

In order to test still more definitely the accuracy of the method,
the following mode of procedure was devised. To neutral solutions
of cupric sulphate — prepared either from the purest crystals or by
long standing after the neutralization of the trace of acid in ordinary
" chemically pure " material — were added small measured amounts
of a standard sodic hydroxide solution. After a time, the precipitate
was filtered oft' and dissolved in standard sulphuric acid, exactly as
if it had been obtained by the expulsion of the acid through heat.

* See page 266, Experiments 26-37.

266

PEOCEEDINGS OF THE AMERICAN ACADEMY

Test of Mkthod.
Precipitation and Solution of Basic Citpric Sulphate.

Number of
Experi-
ment.

Concentration
of Solution.

Time between

Precipitation and

Filtration.

Twentieth

normal XaOH

used.

Twentieth

normal H,jS04

used.'

Factor
c. c NaOH

e C.H2SO4

26

27
28

Strong.
Strong.
Medium.

10 minutes.
90 minutes.
15 hours.

cub centim.

3.00
3.00
3.00

cub. centim.
2.85

2.94

3.14

1.05

1.02

.95

29
30
31
32
33

Dilute.

<(

tc
<(
«

1 hour.

90 minutes.
<i <(

(( it

6.00
6.00
3.00
3.00
3.00

6.24
6 2T
3.12
3.14
3.03

.90
.95
.96
.95
.99

Average of Experiments 28-33

0.96

It is hence evident that from dilute solutions the basic salt is wholly
precipitated in a comparatively short time. Even under the worst
conditions no deficiency approaching that noticed in Experiment 25
was observed. Another series (Experiments 34-37), made in very
dilute solutions without filtering, gave values for the factor which
respectively equalled 0.97, 0.98, 0,95, and 0,98. The average is
0.97, not seriously different from that found above. The first value
of the factor was used in correcting the results of Experiments 20-25.
Tlie correction involved is extremely slight except in Experiment 22,
where the application of the full value of the factor to a very large
quantity is of doubtful advisability. Hence it is better to omit this
experiment from the final average.

One contingency alone is not provided for by these test experiments
that is to say, the formation of a cuprous compound and its solution in
the cupric salt. The formation of a cuprous salt by gentle ignition of
cupric sulphate in the air is far from probable ; but for fear that such
a reaction had taken place, the solution of the cupric sulphate was
usually exposed to the air from twelve to twenty-four hours before
filtration. The fact that exposure to the air for a long time after

OP ARTS AND SCIENCES.

267

filtration caused no further formation of precipitate is a satisfactory
{)roof that no cuprous salt remained iu sokitiou, even if it had been
formed in the first place.

The outcome of all these more or less indirect experiments points
to the conclusion that cupric sulphate retains about 0.12 per cent of
its water of crystallization, even when heated to 400° C. This last
trace of water is given olf only at a temperature at which the cupric
sulphate itself begins to decompose. It must be admitted that the
evidence upon the point is not absolutely certain, but there is no doubt
of its great probability. Upon such a matter as this it is dilhcult to
see how more definite results could have been procured.

Total Percentage Composition of CuSOi . 5 H2O.

Per Cent.

Water lost at 360° — 400° = 36.067

" " between 400° and redness = 0.042

Copper, average = 25.449

SO4 = 38.436

Total = 99.994

If the amounts of copper and sulphuric acid supposed to be lost
during the electrolysis are added, the total becomes 100.004.

Either of these total results is now accurate within the limits of
error of ordinary analysis. But it will be seen that the total percent-
age of crystal water indicated by the analyses is much too large as
compared with the sulphuric acid, which is our only certain standard
of reference. This excess of water must have been occluded in the
original crystals, which were not very finely powdered. Accordingly
two specimens of the purest cupric sulphate were much more finely pul-
verized, and their loss of weight at 400° was then found to be 36.057
(average of Experiments 21 and 23) instead of 36.067 per cent. Mr.
H. M. Richards kindly measured many diameters in the two powders,
and found that the particles of the coarser powder approximated one
tenth of a millimeter, while those of the finer were less than one
hundredth. Since more than half an hour's continuous powdering, at
the very slow rate adopted to prevent warmth from friction, had been
necessary to reduce a small amount of material to the finer state, it
was concluded to abandon the attempt at obtaining cupric sulphate
free from occluded water.

It must be remembered that the salt had been very slowly crystal-
lized by the evaporation of the solution, and of course had had every

268

PROCEEDINGS OF THE AMERICAN ACADEMY

opportunity to imprison traces of the mother liquor. Crystallization
from hot solutions could not be adopted because of the risk of the
formation of basic salt.* But even if the true amount of the water
of crystallization could have been determined, it would have been of
little present use for the determination of the atomic weight of copper,
because of the uncertainty in its molecular weight. It is hoped that
an investigation of the quantitative relations of water in other com-
pounds may be in progress during the coming year at this Laboratory.

The Atomic Weight of Sulphur.

It is seen that the ratio between the weights of sodic carbonate and
sodic sulphate found in each analysis of cupric sulphate forms a valu-
able check upon the purity of the first substance. Conversely, if we
assume the purity of each of the substances and the molecular weight
of one of them, we obtain data for calculating the molecular weight of
the other. Since sulphur is probably the least definitely determined
of the elements involved, the figures may be used with advantage for
the calculation of its atomic weight.

Molecular Weight of Sodic Sulphate.

Weujhts rcdncid to Vacinim.

No. of Ex-
periment.

Source of
Acid.

Na.,C03
taken.

Indicator.

Na^SOi
fouud.

Molecular Weight

Na,S04 if Nii-.CO ^

= IOC 108. ■

4

Electrolysis

grams.
1.29U;30

M. 0.

grams.
1.74113

142.189

6

((

3.18620

Av.

4.26790

142.131

11

Distillation

1.01750

M.O.

1.36330

142.169

12

<(

2.07080

M.O.

2.78200

142.160

7

Electrolysis

1.22427

Av.

1.63094

142.133

17

Distillation

1.77953

Phth.

2.-38465

142.189

17

Electrolysis

2.04412

Tilth.

2.73920

142.189

38

Distillation

3.06140

Tilth.

4.10220

142.180

Averas

e Molecular V

l^eight of Na.SO. . . .

142.169

Avuraff
reliat

e omitting Ex
)le than tlie ot

periments 4,0, and 7, wli
licrs

ch are less

142.181

* Hampe, loc. cit.

OF ARTS AND SCIENCES. 209

The first column of this table indicates the ori<;inal number of the
experiment. The data of Experiments 17 and 38 are given for
the first time. The second column shows the source of the acid. The
third column gives the weights of sodic carbonate taken ; while in the
fourth are recorded the indicators used in determining: these weishts,
the abbreviations standing for methyl orange, phenol phthalein, or the
average of the two. The last two columns contain respectively the
weight in grams and the molecular weight of the sodic sulphate.
Since half of the results were obtained with methyl orange, it is un-
necessary to correct the result for the trace of impurity in the sodic
carbonate.

Subtracting from the averages the quantity 'Na.fii = 110.106 we
obtain two values for the atomic weight of sulphur : —

From the first average, S = 32.063

" second average, S = 32.075

Usually accepted value, S = 32,06

A large alteration in the assumed atomic weight of sodium would of
course make very little difference in the result, while an alteration
of xxiVo ^^ '^^^ value in the atomic weight of carbon changes the
atomic weight of sulphur by only about the twentieth of one per cent.
Although it is certain that the results are hardly capable of deciding
the present uncertainty in the atomic weight of sulphur, the outcome
of the comparison is nevertheless an interesting check upon the work
previously described.

The check just mentioned proves that the sodic carbonate contained
the normal amount of carbon dioxide, but does not prove the absence
of a very small amount of neutral impurity from the salt. Such
impurity would have but a slight effect upon the ratio of the weights
of the two compounds. It is true that this effect would be to di-
minish instead of to increase the observed atomic weight of sulphur;
but it was thought advantageous to attempt the comparison of the
sodic carbonate with some salt having a very different equivalent
weight, in order to obtain more light upon this point. Accordingly
the material to be investigated was converted with suitable precautions
into sodic bromide. In Experiment 39, 1.2198 grams (in vacuum)
of sodic carbonate yielded 2.3685 grams (in vacuum) of the bromide ;
or 2 NaBr : NaoCOg = 206.022 : 106.103. Upon its face the result
was very satisfactory, but the great hygroscopic power of sodic bromide
and various other complications afforded a wide possibility of error;
hence the experiment was not repeated. To be of value, any such

270 PROCEEDINGS OP THE AMERICAN ACADEMY

determination must be made the subject of an especial investigation ;
and for this, with so remote a purpose as the present one, time was
wanting.

TJie Atomic Weight of Copper.

The calculation of the atomic weight of copper from the amount of
metal contained in cupric sulphate which has been dried at 250-270°
gives the value 63.35. This result sufficiently approaches Hampe's
(63.32) to show the identity of the materials used. But it has been
proved that such cupric sulphate still contains a comparatively large
amount of water ; and this fact renders useless the careful analyses
made in 1874. The effect of the coi-rection is considered in a follow-
ing paragraph.

Besides this method of calculation " from difference," the data
which have been given afford twelve other ratios for the computation
of the atomic weight of copper. Six of these ratios are useless be-
cause of the known error in the crystal water ; namely, 5 H.,0 : Cu ; —
SO4 . 5 H2O : Cu ; — 5 H^O : CuSO^ ; — NaoCOa : CuSO^ . 0 HoO ; —
NasSOi : CUSO4 . 5 H2O ; — and BaSOi = CuSOi . 5 H2O. Three of
these last ratios give values which are much too low, and the other
three give values which are much too high, as would naturally be
expected. The mean is curiously near to the true value, but need not
be further discussed. Neglecting these results because of the known
constant error which vitiates them, there still exist seven ratios which
do not involve the uncertain amount of the water of crystallization.

I. From calculation based upon the results tabulated on page 267
anhydi'ous cupric sulphate is seen to contain certainly over 39.807
per cent of copper, and very probably as much as 39.832 per cent.
The first number gives 63.53 as the atomic weight of copper, and this
value must be regarded as the lowest possible limit. The second
much more probable figure, which takes into account the water held
by cupric sulphate at 400°, gives the jDroportion :

(CUSO4 - Cu) : Cu = (100.000 - 39.832) : 39.832 = 96.06 : 63.593.

II. From Experiments 4, 6, 7, and 8 it is found that 7.2501 grams
of sodic carbonate correspond to 4.34583 grams of copper when
methyl orange is used as an indicator ; and from Experiments 6, 7,
and 8 it is found that 5.95165 grams of sodic carbonate correspond
to 3.56697 grams of copper when phenol phtlialcin is used as an indi-
cator. It has been shown that the impurities of the sodic carbonate

OF ARTS AND SCIENCES.

271

probably rendered the latter quantity of the salt about 0.007 per cent
too high; hence the true weight should have been o. 95123. The
correction is so small that it has not heretofore been applied.
From these values we obtain the proportions :

NaoCOg : Cu = 7.25010 : 4.34583
= 5.95123 : 3.56697
Average

106.108 : 63.003

106.108 : 63.597

63.600

The highest individual result was 63.604, the lowest 63.592, when
entirely uncorrected.

III. From Experiments 4, 6, 7, and 9 it is apparent that 10.95552
grams of sodic sulphate are equivalent to 4.90165 grams of electrolytic
copper. Hence :

NaaSOi : Cu = 10.9555 : 4.90165 = 142.166 : 63.607.

Highest value 63.614

Lowest " 63.595

The last determination (Experiment 9) is by far the most trust-
worthy. The value deduced from it is 63.600.

IV. From Experiments 4 and 6 it is evident that 4.4855 grams of
sodic carbonate (average) are equivalent to 6.75463 grams of cupric
sulphate which has been dried at 360°. But it is reasonably certain
that this cupric sulphate still contained sixty-six hundred-thousandths
of its weight of water. Making the appropriate correction we have :

NaaCOa : CuSO, = 4.4855 : (G.7546 - .0044 = 6.7502)

= 106.108 : 159.681

Highest value
Lowest "

159,703
159.673

Subtracting SO4 = 96.060, the average value Cu = 63.621 is
obtained.

V. In the same way, from Experiments 4, 6, and 9, 9,31558 grams
of sodic sulphate are seen to correspond with 10.4715 — 0.0069 =
10.4646 grams of really anhydrous cupric sulphate. Hence:

Na2S04 : CUSO4 = 9.3156 : 10.4646 = 142.166 : 159.701

Highest = 159.73

Lowest = 159.65

From the average, Cu = 63.641. This result has a very large
probable error, in the chemical as well as in the mathematical sense.

272

PROCEEDINGS OF THE AMERICAN ACADEMY

VI. and VII. Only one experiment was made with baric sulphate,
but the result is appended for the sake of completeness. From No.
10 it is seen that 100.000 parts of crystallized cupric sulphate are
equivalent to 93.289 parts of baric sulphate. But the same amount of
the same specimen of the salt had been shown to contain 25.448 parts
of copper and 36.109 parts of water. (See pages 260 and 267.)
From these data : —

VI. BaS04 : Cu = 93.289 : 25.448 = 233.16 : 63.603
VII. BaS04 : CuSOi = 93.289 : 63.891 = 233.16 : 159.685

From the last value Cu = 63.625. The atomic weight of barium
probably lies between 137.0 and 137.2; the mean value is assumed
above. The corresponding variations of the atomic weight of copper
would be from 63.57 to 63.63 with the first ratio, and from 63.54 to
63.69 with the second.

The results of the seven ratios dependent upon the analysis of
cupric sulphate are collected in the following table.

The Atomic Weight op Copper.

Ratio.

Lowest.

Highest.

E.xperimental
Mean.

I.

II.

III.

IV.

V.

VI.

VII.

(CuS04-Cu): Cu
NaoCOg : Cu
Na2S04 : Cu
Na^COg : CUSO4
Nn.^SOi : CUSO4
BaS04 : Cu
BaS04 : CUSO4

[63.53]
63.592
63.59
63.61
63.59
[63.57]
[63.54]

[63.62]
63.604
63.61
63.64
63.67
[63.63]
[63.69]

63.593
63.600
63.607
63.621
«3.641
63.003
63.625

Avera
Avera

ore

63 612
63.605

ge, omitting V. and VII. for obvious reasons . .

The correction of the experimental mean, in each case, for the
amount of copper and sulphuric acid lost during electrolysis, dimin-
ishes the variations noticeable above to a remarkable degree. But
since the validity of this correction is doubtful, and since its applica-

OF ARTS AND SCIENCI-:S, 273

tion scarcely influences the two most important results (II. and III.),
the figures have been allowed to stand uncorrected. 'Ihe final aver-
age would have remained essentially the same. The hi'diest and
lowest valut-s given in this table are in several cases much more
seriously in error than the actual experimental results. The reasons
for this widening of the limits will be sufllciently understood throu<di
a careful perusal of the matter which immediately precedes the table.

Clianges in the atomic weights of the elements used as standards of
reference of course cause slight changes in the corresponding atomic
weights of cop{)er. Sulphur enters into a larger number of the ratios
than any other element, excepting of course oxygen and copper; but
a reasonable change in its assumed atomic weight, while slightly
affecting individual results, has no effect upon either of the averages.
Carbon and sodium are both determined with a hisih deirree of accu-
racy, and oxygen is our standard of reference. A fuller discussion of
these results will be found in the concluding summary.

II. Synthesis of Cupric Sulphate.

Parallel with the experiments which have just been described was
commenced a series of syntheses of cupric sulphate. It was hoped
that the results might furnish a valuable confirmation to the conclu-
sions based upon the analytical work, but the outcome of the series
was very disappointing.

The method of experiment was as follows. A weighed amount of
pure electrolytic copper,* which had been ignited in hydrogen, was
dissolved in the purest nitric acid in a platinum flask provided with
bulb tubes for the condensation of spray. A slight excess of pure
sulphuric acid was then added, and the whole was evaporated to small
bulk in a platinum dish. After transferring to a small crucible en-
closed in a larg3r one, the remaining solution was evaporated to dry-
ness on the steam bath ; and the resi<lue was gradually raised to a
temperature of 400°. After some time the crucible was quickly
covered, placed in a phosphoric oxide desiccator, and allowed to cool
in a vacuum. Air having been admitted through suitable drying
tubes, the crucible was quickly weighed. The product of the experi-
ment soon reached constant weight upon reheating, but it was in-
variably found to contain a very j)erceptible amount of imprisoned
sulphuric acid, which rendered its solution strongly acid to methyl
orange. It is evident from the preceding section that this impurity is

* For mode of preparation, see these Proceedings, XXV. I'JO, 206.

VOL. XXVI. (n. S. XVIII.) 18

274

PROCEEDINGS OF THE AMERICAN ACADEMY

not, retained when pure sulphuric acid is added to powdered cupric
sulphate which has already been nearly dehydrated, but only when the
salt is crystallized from strongly acid solutions. This fact rendered
the present experiments useless for their original purpose. The re-
sults of the two determinations show how insidious constant error
may be, if not guarded against by check experiments.

Synthesis of Cupric Sulphate fkom Metallic Copper.
Weights reduced to Vacuum.*

Number of
Experiment.

Metallic Copper taken.

Cupric Sulphate found.

Per Cent of Copper
in Cupric Sulphate.

40
41

grams.
0.67720

1.00613

grams.
1.7021

2.5292

39.786
39 781

Average

From Analysis of Cupric Sulphate dried at 400° . .

39.784
39.807

In order to show that the large differenre between these two aver-
ages was due solely to the occlusion of sulphuric acid, the following
experiment (No. 42) was made. 3.1227 grams (in vacuum) of the
purest crystallized cupric sulphate were dried at 400°, the residue
weighing 1.9'J63 grams. The nearly anhydrous salt was then care-
fully dissolved in water, and after the addition of small quantities of
nitric and sulphuric acids it was evaporated and ignited at a tempera-
ture even somewhat higher than before. A gain of 0.0013 gram was
apparent, very nearly corresponding to the difference noted above.
The acids, tested immediately afterwards, left no weighable residue
upon evaporation.

While it was apparent that these experiments were useless for their
original purpose, it was hoped that their comparison with similar
syntheses from cupric oxide might furnish valuable indirect evidence
with regard to the quantitative relations of the latter imj)ortant sub-
stance. The hope was not in vaiu.

The cupric oxide used as the basis of the three following experi-
ments was prepared essentially in the manner described in these Pro-
ceedings, Volume XXV., page 199. The three experiments represent

* Specific gravitj' of Cu = 8.95; of CUSO4 (Hampc and Karsten) - 3.61.

OF ARTS AND SCIENCES.

275

as many different preparations, in making which various different
precautions were observed ; but in a paper ab-eady far too prolix the
omission of these minor points will be well pardoned. The cupric
oxide was ignited to constant weight at a dull red heat in a double
crucible, and cooled in a vacuum as usual. The outside crucible
was of platinum in the second and third experiments, and in these
cases the inner crucible was observed to have gained noticeably in
weight during the ignition. The last weight was adopted, for ob-
vious reasons. In the first and last cases the solution of the oxide
was conducted in the crucible, without transferring, while in the second
experiment the cupric oxide was dissolved in the platinum flask with
bulb tubes. In each case both nitric and sulphuric acids were added,
that the conditions might be similar to those of the preceding syn-
theses. The cupric sulphate used by Baubigny in his analyses was
prepared in a somewhat similar way, and hence in this respect his
result is comparable with these.

Synthesis of Cupric Sulphate from Cupric Oxide.
Weights reduced to Vacuum.*

Number of
Experiment.

43
44
4.5

Cupric Oxide taken.

grams.

1.0084
2.7202
1.0144

Cupric Sulphate found.

grams.

2.0235
5.4770
2.0350

Average

Theory if Cii = 63.6

" =0.3.34

Baubigny's result

Per Cent Cupric Oxide
in (Jupric Sulphate.

49.835
49.830
49.848

49.838

49.856
49.774

49.815

The comparatively close agreement between the average and the
theory is solely due to an elimination of opposite errors. This fact
will be more evident in the following section.

In order to show that no material was mechanically lost during the
syntheses, the last sample of cupric sulphate from each of the two
series was electrolyzed. The result of one of these electrolyses has

* Specific gravity of CuO = 6.3 ; of CUSO4 = 3.61.

276 PROCEEDINGS OF THE AMERICAN ACADEMY

already been given in Experiment 16; the other yielded 0.8096
gram of metal from 2.0350 grains of the sulphate, or 39.78 per cent
of copper.

It is evident that impure cupric sulphate prepared in this manner
contains about 39.784 per cent of copper and about 49.838 per cent of
ordinary cupric oxide. That is to say, 39.784 parts of copper corre-
spond to 10.054 parts of the remainder of cupric oxide. Assuming
this remainder to consist solely of oxygen, the atomic weight of copper
would be 63.312. But upon comparing the quantity 10.054 with the
quantity of sulphuric anhydride found by difference, 50.162, it is
evident that a grave error exists in the former figure. This error
becomes only more apparent when allowance is made for the occluded
acid.

These inferences are based upon data of somewhat uncertain accu-
racy, it is true ; but the error is nevertheless so large as to be appar-
ent even to cruder analysis than this. The full explanation of the
results, as well as of the variation exhibited by Baubigny's analysis,
must be deferred to the next section. The doubts raised by tlie data
under discussion formed a useful introduction to the study of cupric
oxide.

III. The Analysis of Cupric Oxide.

Deprived of the support of the results from cupric sulphate, Hampe's
oxide determinations possess little more weight than those of any
other experimenter. It will be remembered that the values of the
atomic weight of copper deduced from this soui-ce have varied from
63.1 to 63.5, Such a fluctuation alone, without the confirmatory
evidence which has just been given, is sufficient to cause the suspicion
of an undiscovered error in cupric oxide.

With the hope of detecting the possible error, a number of analyses
of the substance were made under varvinjj conditions and with differ-
ent samples of material. No difficulty was found in obtaining results
varying as widely as those cited above. Part of the cause of this
variation was traced to differences in the preparation, and part to
differences in the temperature and tension of the surrounding air
employed in ignition.* Cupric oxide ignited to constant weight at
a very dull red heat lost a very perceptible amount of material upon

* Compare Bailoy and Hopkins, J. Ch. Son. Trans., 1890, p. 269. Also Schut-
zenberger, quotdl in Am J. Sc, [3], XXVI. 05.

OF ARTS AND SCIENCES.

277

heating to the highest temperatures whicli hiiril glass would bear.
On the other hand, copper reduced by hydrogen at the lowest possible
temperatures is well known to lose in weight on heating to bright
redness. Some of P^rdmaim and Marchand's experiments* show
that the exhaustion of the tube at the time of weighing makes very
little if any variation in the weight of the materials ; but from others,!
as well as from the fiftieth experiment below, it would appear that
exhaustion at the time of ignition introduces a somewhat more serious
correction. In this last case the apparatus was necessarily rather
complicated, to admit of the ignition and weighing of the copper and
its oxide in a Sprengel vacuum, but the full description of the con-
trivance would demand more space than it is worth. In the four pre-
ceding experiments Hampe's method was carefully followed.

The mode of preparation of the cupric oxide was essentially that
recommended by Ilampe, and described in these Proceedings, Volume
XXV., page 199. In some cases the basic nitrate was not washed
with water before ignition, and in other cases it was thus washed. In
general, none but platinum vessels were used. Various slight unim-
portant modifications were introduced, which need not find a place here.
The essential conditions to be borne in mind are the invariable use of
the nitrate as the source of the oxide, and the variable temperatures
employed in the ignition.

Analysis of Cl'pric Oxide.

Weights reduced to Vacuum Standard.

No. of
Exp't.

CuO

dull rednes.s

550"±

CnO

750" ?

Cu
found

400"i:.

Cu
found

550''±.

Cu
found
750''?

Atomic Weight of

Copper
0 : Cu^l6:a;.

46
47
48
49
50

grains.
2.08491

1.11936

1.00300

1.91703

2.62410

grams.

1.11851
1.062.53
1.91656
Ignited

grams.

[0.8490]
in vacuum

grams.
1.6G405

0.89355
2.0956

grams.

0.84831
1.5298

63.26
63.29 to 63.53
63.23 to [63.62 j
63.20 to 63.29

63.44

mpnprnl Avprnp-e

63.37]

* Loc. cit. (see p. 241).

t J. prakt. Chem , XXVI. 461.

278

PROCEEDINGS OF THE AMERICAN ACADEMY

The average of results so heterogenous in nature can naturally have
no important meaning, but it is interesting to note the approximation
to the old value of the atomic weight. The high value given in Experi-
ment 48 is manifestly only a compensation of errors. From Experi-
ments 48 and 49, the only two in which both materials were ignited
at a bright red heat, 2.97909 grams of cupric oxide yielded 2.37811
grams of metallic copper. Computiug from these data the atomic
weight of copper, we obtain the value 63.313, which is comparable with
the result 63.312 obtained indirectly through the synthesis of the sul-
phate as well as with the number 63.346 fixed upon by Hampe. The
absolute identity of the first two figures must be attributed to chance,
since the agreement of the individual results was not perfect.

These results again pointed to the existence of a volatile impurity
in cupric oxide, but no proof was afforded that the impurity was
wholly driven off at the temperature of fairly bright redness. The
determination of this point, as well as of the nature of the occluded
material, became a matter of great importance.

Tests for Impurities-

The first hypothesis suggested was the possible imprisonment of a
small amount of water.* To subject this hvpothesis to proof, pure
cupric oxide might be dried at a red heat, and reduced by means of
carbon monoxide, when of course any occluded water would be set free
and mijjht be weighed.

The cupric oxide used in the execution of this plan was some of
that which remained from the oxygen research. It hud been prepared
in the usual manner. The carbon monoxide was made from oxalic
and sulphuric acids, and after a preliminary purification with caustic
potash the gas was collected over water in a glass gasholder. From
this receptacle it was passed through very large amounts of potassic
hydroxide, over calcic chloride and red hot platinum sponge, and
finally through a flask containing sulphuric acid and two tubes con-
taining phosphorus pentoxide, before being allowed to come in contact
with the material to be reduced.

In Experiment 53, — which may be taken as a type, — fifteen
grams of the cupric oxide contained in a hard glass tube were dried
at a red heat in a stream of pure air for forty-five minutes. A small
weighed phosphoric oxide tube was then connected at the exit, and
the current of air was continued for over an hour.

* See MuUer-Erzbacli, Jahresbericht, 1885, p. 74.

OF ARTS AND SCIENCES.

"Weight of tube before connection
» after "

279

Grams.

= 55.1920
= 55.1922

The temperature changed from 15°. 7 to 16°.8 between the two
weighings, involving a correction of — .OOUl to the latter weight. The
tube therefore gained only the tenth of a milligram, showing the cupric
oxide to have reached a constant hygroscopic condition.

During the ensuing reduction with carbon monoxide, the drying
tube was again connected with the apparatus. Before the last weigh-
ing of the tube, all carbon dioxide was expelled by pure air. The
tube weighed 55.2177 grams, showing a gain of 0.0255 gram, or
0.017 per cent of the weight of the cupric oxide.

Two similar experiments led to like results.

Ueduction of Clpric Oxide by Carbon Monoxide.

No. of
Experiment.

CuO taken.

Water formed.

Volume of
Carbon Monoxide.

Weight of H_.0 for
lU grams CuO.

51
52
53

grams.

0.90
20.00
15 00

grams.
.0105

.0347

.0255

litres.
5.5

gram.s.
.0152

.0173

.0170

Nothing in these results proved that the water was not formed from
the oxidation of hydrogen or hydrocarbons possibly contained in the
carbon monoxide, and indeed further examination showed that it
originated from this source. The metallic copper from the last ex-
periment was immediately oxidized by a stream of pure dry air, and
once more reduced with carbon monoxide. Four litres of gas required
to reduce the partially reoxidized copper yielded 19.3 milligrams of
water. Since this water could not possibly have come from the
oxidized copper, it must have been solely obtained from impurities in
the carbon monoxide. The quantity closely agreed with the amount,
18.3 milligrams, which one would have expected to find from the
volume of gas used, upon the basis of previous results. Considerii.g
the fact that the volumes of gas were not very accurately measured,
the difference is not greater than the experimental error. Hence
it may be safely concluded that the carbon monoxide contained about
four one-hundredths of a per cent of hydrogen by weight, but that the
cupric oxide retained no appreciable amount of icater at a red heat.

280 PROCEEDINGS OF THE AMERICAN ACADEMY

Experiments 54 and 55. — The idea that some cupric nitrate might
remain uudecomposed in cupric oxide is evidently not a new one, for
almost every experimenter upon the subject tested for nitric acid in
the vpater formed during the reduction. It seemed possible, however,
that the nitric acid might be reduced as well as the cupric oxide, and
hence that the test might not be a sufficient criterion. Accordingly,
five grams of cupric oxide were dissolved in very pure dilute sulphuric
aiud, and several successive portions of water were distilled off from
the mixture. After evaporation to small bulk in a platinum dish, the
distillates were found to be neutral to meihyl orange and acid to
phenol phthaleio. When the amount of standard alkali necessary
to neutralize this acidity had been determined, hydrochloric acid was
added and the sulphuric acid present was estimated as baric sulphate.
Since the weight of baric sulphate (3.7 mg.) was nearly equivalent
to the amount of alkali used (0.39 c. c. of a decinormal solution),
since no nitric acid was found in the filtrate by the most sensitive
tests, and since a very perceptible amount of copper was found there,
it may be reasonably concluded that cupric sulphate had been carried
over mechanically in fine drops. A second experiment yielded like
results. Hence, so far as this test was concerned, no nitric acid was
to be found in ciipric oxide.

Experiment 56. — The tendency of various oxides to hold carbonic
acid even at high temperatures is well known. Thudichum * has
observed that, in order wholly to free cupric oxide from this impurity,
it is necessary to ignite it in a vacuum. In order to test this point
so far as it concerns the present investigation, the following experi-
ment was made. After a small amount of nitric acid had been added
to twelve grams of cupric oxide, the substance was gradually brought
to redness and maintained at this temperature for ninety minutes.
The material was then rapidly transferred to a flask provided with a
stoppered funnel and two bent tubes. One of the latter was connected
with two test tubes containing a clear solution of baric hydroxide, and
the other permitted a gentle current of air to be blown through the
entire apparatus. Before either baric hydroxide or cupric oxide were
added, all the tubes and flasks were of course freed from carbon
dioxide. Finally, dilute boiled hydrochloric acid was run in at the
funnel tube, the cupric oxide was dissolved, and the solution was
gradually brought to boiling. A slight bluish precipitate of cupric
hydroxide appeared after some time in the first test tube, while the

* Chem. Soc. Journ., 1876, [2], p. 364.

OF ARTS AND SCIENCES. 281

second remained perfectly clear. The excess of baric hydroxide iu
the first tube was quickly neutralized with weak hydrochloric acid and
phenol phthalein, and the precipitate was filtered off and washed.
The absence of barium from this precipitate proved the absence of
carbon dioxide from the cupric oxide.

In order to test the adequacy of the method, 3.4 milligrams of so-
dic carbonate were added to the cupric chloride through the funnel
tube. This quantity yielded only one tenth the amount of carbon
dioxide necessary to account for the difference between the atomic
weights of co|)per in question, but it nevertheless produced a heavy
white precipitate in both newly filled baric hydroxide tubes. The
precipitate when tested showed large quantities of barium ; hence the
method was quite competent to decide that the absorption of carbonic
acid by cupric oxide was not the error for which search was being
made.

Experiment 57. — Although it seemed very improbable that cupric
oxide could contain even traces of a higher oxide after ignition at a red
heat, proof of the point was not at hand. Accordingly, four grams of
the substance were boiled with pure hydrochloric acid in an apparatus
somewhat resembling the last one, in which all the joints were of
sealed or ground glass. The vapors were driven through a reversed
air cooler into bulbs containing a strong cooled solution of potassic
iodide. The very small amount of iodine set free after some time was
determined by means of sodic thiosulphate. Tiiis amount corre-
sponded to 0.05 milligram of oxygen but even this was undoubtedly
due to copper in uncondensed spray. Eight tenths of a milligram of
potassic dichroraate, subsequently added to the cupric chloride, at once
set free more than double this amount of halogen in the bulbs. The
method was hence shown to be adequate for the purpose. The results
vrove the absence of an essential amount of any higher oxide or oxi-
dized nitrogen from ordinary cupric oxide.

The Determination of Occluded Gases.

After so much negative evidence, the only remaining hypothesis
which could account for the irregularity of the atomic weight of
copper deduced from cupric oxide is that of the occlusion of gases by
the substance. This idea is by no means new. As long ago as 1842
Erdmann and Marchand* showed that at least a small amount of air

* J. f. prakt. Clieni., XXVI. 4G1.

282 PROCEEDINGS OF THE AMERICAN ACADEMY

was condensed by the substance. Frankland and Armstrong* found in
1868 that cupric oxide prepared from the nitrate contains both carbonic
acid and nitrogen. Hilditch f states that oxygen gas is occluded by
cupric oxide at a red heat, and estimates the effect of this impurity upon
the atomic weight of oxygen, without, however, adducing any experi-
mental data. More recently Morley t has found that cupric oxide
slowly gives off a gas in vacuum. It is a noticeable fact that cupric
oxide is almost as serviceable as platitium in preventing bumping
during ebullition.

The simplest and surest method for the liberation of an occluded
gas is completely to dissolve the cupric oxide in a pure acid. It is by
no means certain that indefinite heating in a vacuum would accomplish
the end in view. Several pieces of apparatus more or less suitable
for the j^resent purpose were devised in succession, and by their means
it was proved that some varieties of cupric oxide occlude important
quantities of aeriform material. Nearly fifty determinations were
made.

According to the first method a tube of the shape shown in Fig-
ure 1 was filled at the end with freshly ignited cupric oxide, and in the
middle with pure boiled sulphuric acid. When the whole had been

¥lG. 1.

completely exhausted by means of a Sprengel pump, the liquid was
allowed to come in contact with the solid, and the gas evidently set
free was pumped out and measured. But the action was of course
only a very superficial one, and no amount of long standing or violent
shaking could accomplish a more complete change. Three experi-
ments indicated that at least a small amount of gas, which appeared to
be chiiifly nitrogen, was set free. Since it was difficult to decide how
much cupric oxide had been combined, the method was abandoned.

* Cliem. Soc. Journ., XXI. 89,03 (1868).

t Chem. News, XLIX. 37 (1884).

t Am. J. Sci., XLI. 231, March, 1891.

OF ARTS AND SCIENCES. 283

Subsequently, more dilute boiled sulphuric acid was used ; and the
apparatus was sealed by fusion after exhausting, but before mix-
ing the acid and base. The mixture was digested upon the water
bath until the action was nearly or (juite completed. It was found
that this com2)k'tiou was greatly facilitated by violent shaking at the
moment when the acid was poured upon the cupric oxide, for other-
wise the materials inevitably formed a hard cake which dissolved very
slowly. In one or two cases a small amount of unchanged cupric
oxide was weighed, and subtracted from the original amount.

The apparent evolution of gas was at first very violent; but this
violence might have been partly due to ebullition caused by the heat
of the chemical action. The tubes were opened under boiled water or
mercury, and the gas was analyzed by means of liempel's apparatus,
somewhat modified for accurate use with very small quantities of mate-
rial. The portion of the gas which was not absorbed by caustic potash,
but was absorbed by pyrogallol, is tabulated below as oxygen. The
remainder refused to support combustion, suffered no diminution of
volume upon the addition of oxygen, and appeared perfectly inert ;
hence it was undoubtedly nitrogen. In order to test the solution of
pyrogallol, analyses of air were made from time to time, with satis-
factory results. The volumes were of course corrected for tempera-
ture and pressure, as well as for the tension of aqueous vapor, before
the weights of gas were calculated. The conditions varied so little
from the mean, 24° C. and 760 mm., that a statement of these data
seems inessential.

The occlusion of an inert gas by cupric oxide was so peculiar a
phenomenon that many experiments were needed to carry conviction.
Several blank experiments were tried, in order to test the freedom of
the acid from dissolved air. Tlie fact that only a very small amount
of gas was evolved in the second experiment is also valuable evidence
upon this point.

The material used in Experiments 58, 63, and 64 was four years
old, while that used in Experiments 59, 65, 66, and 67 had been
recently prepared. Both of these samples were made from cupric
oxynitrate, whereas the one used in Experiment 61 was precipitated
by caustic alkali. The tubes of Nos. 64 and 67 were opened under
mercury. In No. 64 the cupric oxide was cooled for forty-eight
hours after igniting, while in No. 66 the oxide was still very hot
when introduced into the tube.

284

PROCEEDINGS OF THE AMERICAN ACADEMY

Occluded Gas ix Clpric Oxide. — First Series.

Xo. of
Erperi-
menc.

Weight

ofCupric

Oxide.

Vessel em-
ploved for
Ignition.

Tempera-
ture rf
Ignition.

Vol. of tjas
evolred un-
corrected.

Vol ofO,

for 1 srm.

ofCuO.

Vol of S.

forlgrm.

ofCuO.

Per Cent of
GasinCuO
by Weight.

JT:ini.=.

cub. cent.

cub. cent.

cub. cent.

iS

4.3.5

PorL-elain

Dull red

2.70

0.03

0.59

0.075

59

3.12

{ Double xble
( Platinum

Bright red

0.30

0.01

0.09

0.012

60

Blank

—

—

0.10

—

—

—

61

1.20

"^

Not ignited

S.oofCOs

Trace

Trace

62

Blank

—

—

0.10

—

—

—

63

1.20

I orcelain

Dull red

0.73

0.03

0.58

0.073

64

2.20

tc

it

1.34

o.as

0.58

0.073

65

2.00

<f

u

0.8±

Trace

0.40

0 05±

•

3.20

(Double xble
1 Platinum

ei

1.70

0.05

0.48

0.064

67

2.10

Porcelain

«

1.13

0.00

o..5;3

0.065

The series was discontinued because a new method had been de-
vised which was much less tedious and much more direct in its exe-
cution. According to the new method, the cupric oxide to be analyzed
was introduced into a tube shaped in the manner shown in Figure 2.

Fig. 2.

After the bulbs had been nearly filled with thoroughly boiled warm
water, the open mouth of the tube was connected with an air pump to
remove any mechanically adhering air. The water was usually allowed
to boil for several minutes under the reduced pressure. The air was
then readmitted, and the tube was entirely filled with pure boiled
water, being placed perpendicularly for the purpose. When the tube
had been returned to its usual position, acid was added through a fine

OF ARTS AND SCIENCES. 285

funnel tube in such a manner that it at once attacked the oxide.
A short piece of glass rod in the large bulb together with constant
agitation prevented the formation of cakes.

In the most carefully performed experiments the sulphuric acid as
well as the water was boiled, but repeated trials showed this precau-
tion to be unnecessary. In one blank experiment even ordinary dis-
tilled water and sulphuric acid evolved only about one thousandth of
a cubic centimeter of gas, while boiled water, which was invariably
used in tbe determinations, gave no trace even of the finest bubbles
with unboiled sulphuric acid.

Tlie gas evolved on the solution of the cupric oxide was collected
chiefly in the largest bulb. The amount collected in the smaller bulbs
decreased in proportion with the increase of distance from the first,
and only a very small amount was lost through the open mouth. The
gas was measured in the right hand graduated tube, or drawn off for
analysis through a rubber connector attached at this end.

The apparatus was afterwards modified to avoid the slight loss
through the open mouth. It then consisted of a small stoppered flask
with two concentric funnel tubes, and a delivery tube closed with a
rubber connector and a pinch cock. When the whole apparatus had
been filled with boiled water through the larger funnel tube, the
smaller one was filled with pure sulpiiuric acid and quickly placed
inside the larger one. Upon inclining the flask, it is evident that the
loss of minute bubbles must have been reduced to a minimum, while
the small annular openin": between the funnel tubes was quite suffi-
cient to allow the escape of the displaced water. The gas was readilv
drawn otf for analysis through the delivery tube, which just penetrated
the smooth rubber stopper.

Many varieties of cupric oxide were examined. Material obtained
from the carbonate, or by precipitation from the sulphate, evolved little
or no gas, thus affording an admirable proof of the purity of the acid
and water. Check experiments with such material were accordingly
made from time to time. Other circumstances being equal, the purer
the oxide, the more gas was evolved. Solid impurities, such as alka-
line salts, seemed partially to prevent the occlusion.

The volumes of gas are given without any correction, the temper-
ature and pressure remaining so nearly constant that the observations
are not recorded. The following table is intended only as a basis of
rouah comparison. The gas was frequently analyzed, with results
similar to those tabulated above. The results confirm those of the
preceding series.

286

PROCEEDINGS OF THE AMERICAN ACADEMY

Occluded Gas in Cupric Oxide. — Second Series.

Xo. of
Exp't.

Nature of
Cupric Oxide.

Weight

of
CuO.

Heat and Time
of Ignition.

Volume

of Gas

evolved.

Per Cent of
Ga.s in CuO
by Weight.

Remarks
on Method.

68
69

Blank exp't.

grams.

—

cubic

centim.

0.00

0.00

—

H.,S04+H20

70
71

Precipitated
tt

1.00
1.60

Dull red

010
0.20

0.07
0.10

0.011
0.013

Bulbs.
tt

72
73

From Carbonate

1..50
2.00

Dull red
( Dull red ex- )
( posed to air )

0.006
0.006

Bulbs.

74

75
76

(^rurle, of
Commerce,
contained

Cuprous
Oxide.

3.00
8.00
7.00

Not ignited
Dull red
Not ignited

0.70
2.00
2.00

0.026
0.028
0.032

Bulbs.

Flask.
( Another
I sample.

77

From Carbonate

1.50

Dull red

0.05

0.003

Flask.

78
79
80
81

82

83

Six Experi- "
ments witii
CuO prepared
by action of

CuO on

Cu(N03).3&

subsequent

ignition of

basic nitrate

1.70
3.20
1.20
1.70
2.00

2.00

Very low
Dull red, 3 hrs.
Dull red, 7 hrs.
Bright red
Dull red

Dull red

0.34
1.03

0.25(?)
0.09
0.26

0.73

0.023
0.036
0.023(?)
0.006
0.015

0041

Flask.
<<

"

j Pump not
j used.
( Pump not
( used.

84
85
86
87
88
89
90
91
92
93
94
95

After Hampe's

method

(( ti

(1 <(

(( C(

It t<
a ti
tt (t
tt tt

New preparation

<( <(

« ((
tt tt

0.89
1.10
1.82
0.39
0.50
0.60
0.61
0.44
2.00
2.98
0.98
0.50

310°; 2 hours.

600° ; 2 hours.

700°; ihour.

600° ; 1 hour.

600° ; 4 hours.
<( ((

>■ >(

700° ; f hour.
600° ; 2 hours.

See below.
900°; Ihour.

0.17
0..30
1..35
0..30
0.33
041
0.36
0.29
1.37
2.12
0.75
0.17

0.050
0.035
0.084
0.087
0.077
0.079
0.070
0.079
0.080
0.082
0.088
0.039

Bulbs.

Flask.
Bulbs.

t(
i(

Used IICl.

Bulbs.
See below.

Flask.
See below.

96

97

( From Carbon- )
} ate prepared >
( from Nitrate )

0.65
0.50

600° ; 1 hour.
290°.

0.09
0.10

0.016
0.023

98
99

( From Carbon- )
} ate of Com- >
( merce )

1.00
0.75

600° ; 1 hour.
It tt

0.10

0.03

0.011

0.004

100

Wire form

1.00

—

O.Oli

0.001 ±

( Contained
) Cu.,0.

OF ARTS AND SCIENCES. 287

The liiglier temperatures recorded above are merely approximate
guesses. GOO^ signifies a dull red heat, 700° a medium red heat, and
900° a very bright red heat. In Experiment 92 the oxide was
heated at the highest temperature obtainable by a Berzelius spirit
lamp. In No. 94 the material was heated for an hour and a half in
the Berzelius lamp, and then for an hour in a glass tube under a
stream of oxygen. The oxide was transferred while warm to the
bulb apparatus. It evolved upon solution 0.088 per cent of gas, one
eighth of which was oxygen. In the ninetieth experiment hydro-
chloiic acid was used for the solution of the oxide.

The two series of determinations, which were perhaps more numer-
ous than necessary, showed the following important facts: —

First, that cupric oxide prepared by the ignition of the oxynitrate
after Hampe's method contained between four and five times its
volume of occluded gas. (Nos. 58, 63-67, 84-94.)

Second, that ignition of the oxide at very bright redness was ca-
pable of expelling a portion of this gas. The temperature required
was nevertheless considerably above that which Hampe apparently
employed. (Nos. 59, 81, and 95.)

Tiiird, that beyond a certain limit the time of ignition makes no
important difference. (Nos. 86-91.)

Fourth, that the gas was not absorbed from the air on cooling.
(Nos. 64, 65, 66, and others.)

Fifth, that specimens of cupric oxide prepared in different ways
contained very different amounts of gas. It is a curious fact that
material which has been heated only to 300° contained less gaseous
nitrogen than that which had been ignited at a higher temperature.
In this case the nitrogen may still have existed in the combined form.
(Nos. 78, 79, 84, 85.)

An explanation for the observed phenomena must necessarily be
hypothetical, and will not be attempted at present.

In order to find the direct connection between the loss of weight
noticed upon heating cupric oxide and the quantity of gas retained by
the substance, the following experiment was made.

Experiment 101. — Cupric oxide which had been heated to constant
weight at dull redness was ignited at bright redness in a double
platinum crucible.

288 PROCEEDINGS OF THE AMERICAN ACADEMY

Grama.

"Weight of platinum crucible = 18.0889

" crucible + CuO, 2^ hours at dull redness = 19.6o8o2

« " " 4 hours more " = 19.6584G

After ignition at very bright redness (20 minutes) = 19.6574

" " (50 m. more) =19.6571

Total loss of weight = A)014

One gram of this specimen was shown to contain 0.00081 gram of
<iccluded gas before ignition, and 0.00039 gram afterward (Experi-
ments 86 to 95). In the quantity taken above this difference corre-
sponds to a loss of weight of nearly 0.0007 grams, or about half of
that observed. The remainder of the loss may have been due to a
partial reduction of the oxide. (Compare page 275.)

Experiment 102. — In order to discover if heating cupric oxide in
a vacuum produced any important evolution of nitrogen, six grams of
cupric oxide were dried in a glass tube with the greatest care, exactly
after the manner of Hampe, and weighed in the air. One end of the
tube was then sealed, the other was attached to a Sprengel pump, and
the whole was heated to about 500° C. for half an hour. Upon seal-
ing and weighing; the loss of weight was found to amount to 4.4 milli-
grams. The amount of air displaced was found by opening the tube
under boiled water, which wholly filled it. The weight of water,
reduced to 4°, was 3.48 grams, while the temperature and pressure
of the air at the time of weighing the exhausted tube were respectively
24° C. and 754 mm. Hence the weight of displaced air must have
equalled 4.1 milligrams, or only 0.0003 gram less than the loss of
weight actually observed. Therefore the cupric oxide could not have
lost more than 0.005% of its weight under the treatment described.*
The moist cupric oxide yielded approximately the usual volume of
nitrogen upon solution in acids ; but no very accurate measurement
was made.

It has been said that these experiments were intended onlv for
rough comparison. Since it appeared possible to apply a correspond-
ing correction to the atomic weight of copper deduced from tin- oxide,
it became importatit to make a more accurate series of determinations.
The cupric oxide used in these final experiments represented tln-ee
different specimens, similar to or identical with those used in Nos.
43, 45, 48, and 49.

* In IvNixrimont 50 the temperature of ignition was somewhat liiglier.

OF ARTS AND SCIENCES.

289

Occluded Gas in Cdpric Oxide. — Third Series.

Number

of
Experi-
ment.

CuO
taken.

Gas evolved.
[Apparatus.]

Temperature
ot Gas.

Atmos-

|)lieric

Pressure.

Per Cent of
Nitrogen in

CuO
by Weight.

Per Cent of
Oxygen in

CuO
by Weight.

103
104
105

grams.
1.84

1.80

1.30

cubic centimeters.

1.38 (bulbs)
1.55 (flask)
1.10 (bulbs)

Centigrade.
25°

22°

24°

millimeters.
757

7G3

760

0.080
0.094
0.090

0.003
0.004
0.004

Average

0.088

0.004

The cupric oxide of the last determination was ignited for three
hours in a stream of air, and for twenty minutes in a stream of oxygen,
at a medium red heat. It was impossible to determine whether the
material had reached constant weight, for the hard glass tube in which
the ignition was conducted was very much altered by the heat.

As might have been expected, the very pure cupric oxide employed
in these experiments contained a slightly greater amount of occluded
gas than the less carefully prepared material of the earlier experiments.
Since the latest samples of material were precisely similar to those
used in the quantitative analysis of cupric oxide, it is fitting that the
latest correction should be applied to those quantitative results. The
2.97909 grams of cupric oxide used in Experiments 48 and 49 (page
277) must have contained 0.00276 gram of occluded gas. That is to
say, 2.97633 grams of pure cupric oxide correspond to 2.37811 grams
of metallic copper. The difference between these weights now really
represents the weight of combined oxygen present. The atomic
weight of copper upon this basis is easily found to be 63.605.

Correcting the indirect results from the synthesis of cupric sulphate
in the same way (see page 276) the number 63.603 is obtained.
Averaging this result with that immediately above, the value 63.604
is found as a mean of five determinations including two wholly distinct
methods.

Since it is impossible to learn exactly the amount of gas occluded
by Hampe's cupric oxide, the correction of his results will not be
attempted. The residue left after the ignition of cupric sulphate can
hardly contain any nitrogen ; and this supposition more than accounts
for the difference between the results of Baubigny's analysis and the
present syntheses of cupric sulphate (see page 275).

VOL. XXVI. (N. 8. XVIII.) 19

290 PROCEEDINGS OF THE AMERICAN ACADEMY

Upon referring to the other work upon the atomic weight of copper,
it is apparent that nearly all of the discrepancies have been explained.
A portion of Millou and Comaille's work consisted of weighing the
water formed by the reduction of cupric oxide ; the reason of the low
result of this part of their work is not even now evident, but at the
present day a discussion of the possible causes of its error would be
uhprofitable. Shaw's recent result, which formerly ranked among the
highest of the observed values, now appears among the lowest.

Evidently the occlusion of nitrogen by cupric oxide must have a
very serious effect upon experiments in which oxygen is determined
by the loss of weight of material prepared by the ignition of the basic
nitrate ; notably in the determination of the atomic weight of the
element after the method of Dumas. When the oxide is obtained
by the ignition of the metal in oxygen, the correction is of course in-
applicable. Unfortunately, most of the experimenters upon the subject
have omitted to state the source of their cupric oxide. The prepara-
tion used in the later experiments of Erdmann and Marchand * must
have contained nitrogen ; and it will be remembered that the second
series of his results gave a much higher result than the first.

In view of the correction, it is remarkable that the resultant atomic
weight of oxygen in this case was not even higher than 16 ; for the cor-
rection is about five times as large as the amount necessary to account
for the diflTerence between 15.87 and that figure. From a difference
either in the mode of preparation or in the temperature of ignition, the
oxide used by Erdmann and Marchand must have cont-ained much
less nitrogen than the amount found in Experiments 103 to 105; or
else some opposite error must have partially counterbalanced this one.
Considering the remoteness of all the experiments, a present discussion
of the results is unnecessary ; but the conclusions are at least sufficient
to throw a serious doubt upon the applicability of cupric oxide for a
quantitative source of oxygen, as well as to support the modern low
atomic weight of that element. The presence of nitrogen could not
seriously influence the results of Keiser or those recently obtained
here ; and it is evident that the material used by Noyes could have
contained no occluded gas.

The use of cupric oxide prepared from the nitrate must also intro-
duce a serious error into organic nitrogen determinations after the
method of Dumas, as Frankland and Armstrong have already pointed

* J. fiir prakt. Chemie, XXVI. 461.

OF ARTS AND SCIENCES. 291

out. The usual " wire form " of cupric oxide is not subject to this
cause of inaccuracy, however.

In the uear future other oxides will be investigated at this laboratory
with regard to their possible occlusion of gases.

IV. Upon Typical Copper.

It has been noticed that most of the calculations described in the
present paper have been referred to electrolytic copper. Whether
such copper represents the typical element or not is a most impor-
tant question, but is one which unfortunately cannot be conclusively
answered here.

Many authorities have discussed the subject,* and both sulphur and
occluded gases have been found in metal prepared in this way. Hampe
has shown that the occasional presence of sulphur is probably due to
minute drops of solution imprisoned between the electrode and the
deposit. When pure copper was to be prepared in the present inves-
tigation, the metal was invariably detached from the dish in order to
prevent such imprisonment. In the determination of copper already
in solution, the current was always made as weak as possible in order
to avoid the chance of error. Besides, it will be remembered that
there was a slight loss rather than a gain daring electrolysis.

The occlusion of gases is at worst very slight. In order to reduce
the error from this cause to a minimum, the crucible which formed the
negative electrode was usually coated beforehand with copper, that the
initial and the final surface might be subject to the same error. To
determine the amount of volatile material absorbed by electrolytic
copper, very thin strips were ignited in a stream of hydrogen, with the

following results :

Gram.

107

0.80565

gram

lost

0.00005

108

0.07724

it

(f

0.00004

109

1.24158

«

((

0.00004

110

0.80712

u

((

0.00001

111

1.00616

u

u

0.00003

Average 1

gram

lost

0.00003

This correction is too small to be applied, even supposing that the
method of precipitation just mentioned did not render the application

* Soret, Compt. Rend., CVIT. 733, CVIII. 1298; Hampe, loc. cit.; also
Macintosh, Classen, Foote, Smith, and many others.

292 PROCEEDINGS OF THE AMERICAN ACADEMY

unnecessary. The amount of hydrogen absorbed even by copper
reduced from the oxide is exceedingly small. Erdmann and Marchand
found that one gram of copper absorbed three one-hundredths of a
milligram of hydrogen, and Dumas's results were essentially the same.*
Hampe was unable to find a trace of the gas in the copper remaining
from his experiments. While this correction may have a sensible
effect upon the old atomic weight of oxygen, it cannot seriously eiFect
the atomic weight of copper.

No method for the preparation of pure copper has been suggested
which is not open to possible objections. The present standard of
reference has the merit of simplicity. Moreover, the results of the
analysis of cupric sulphate, unless concealing some unknown error,
show that electrolytic copper cannot be very different from the typical
element as it exists in combination, otherwise the summation of the
results would not so nearly equal one hundred per cent. Again, cupric
oxide gives essentially identical results, whether analyzed by reduction
or by electrolysis. It must be remembered that, while some of the
copper used in the present series of researches came from Lake Su-
perior, the greater part was prepared from " chemically pure " Ger-
man cupric sulphate. Some unknown constant impurity may have
vitiated all the preparations, but the present evidence seems to show
that pure electrolytic copper is as definite a substance as most of
the other " elementary substances " to which our atomic weights are
referred.! A more elaborate comparison of copper from different
sources and different modes of preparation would neverthless be one
of considerable interest.

* Ann. Chira. Phys., [3], VIII. 189, 205. See also Thudichum and Hake,
Jahresber. 1876, p. 966; and Johnson, Jahresber. 1878, p. 286.

t Griinvvald and Brauner have independently come to the conclusion that
copper is a compound. If this is the case, the compound must be a very defi-
nite one. See Chem. Soc. Abstracts, 1890, p. 434.

OF ARTS AND SCIENCES.

293

V. Summary op Results, 1887 to 1891.
The Atomic Weight of Copper.

0 = 16.000.

Ratio.

Result.

I.

2 Ag : Cu

Eleven Determinations

63.601

11.

2 AgBr : Cu[Bro;

Three

a

63.G09

III.

2 Ag : Cu[Br2]

Three

(C

63.605

IV.

[CUSO4 - Cu] : Cu

Three

a

63.593

V.

Na^COg : Cu

Seven

(I

63.600

VI.

Na^SOi : Cu

Four

ti

63.607

VII.

Na^COs : [CuSOi - SO,]

Two

«

63.621

VIII.

Na^SO, : [CUSO4 - SOJ

Three

((

63.641

IX.

BaSOi : Cu

One

((

63.603

X.

Baso, : [CUSO4 - so;

One

«

63.625

XL

]CuO — Cu] corrected : Cu

Five

u

63.604

General Average of the eleven Series

63.610

" " of fortj

'^-three Determinations

63.606

Of these results five — namely, I., II,, III., V., and VI. — are
incomparably more trustworthy than the others. Their computation
involves only very accurately determined elements, and they are least
affected by a given change in the molecular weights of the standards
of reference. Their experimental errors are far smaller than those
of any of the others, and they involve the simplest and most direct
processes, and the minimum number of necessary corrections. They
are in no case computed by the objectionable method of difference.
Finally, each one of the five is wholly independent of the weight of
any copper compound whatsoever. These reasons constitute a suffi-
cient ground for the separation of such results from the others.

Selected Series.

Ratio.

Salt
decomposed.

If 0 = 16.01
Cu =

I.

2 Ag : Cu

AgNOs

63.601

II.

2 AgBr : Cu

CuBra

63.609

III.

2 Ag : Cu

CuBr2

63.605

V.

NasCOs : Cu

CUSO4

63.600

VI.

Na2S04 : Cu

CUSO4

63.607

Final average

63.604

294 PROCEEDINGS OF THE AMERICAN ACADEMY

If the oxygen is taken as 15.96, copper becomes 63.44. With
oxygen 15.87, the value sinks to 63.09.

The mathematical discussion of these results is omitted, because
it gives an exaggerated notion of the accuracy of the final average.
The real probable error of this average is much more dependent upon
the chemical purity of the electrolytic copper than upon the mere
mechanical or experimental error with which alone the theory of
least squares is competent to deal.

It is apparent that any one of the methods, thoroughly investigated,
would have long ago afforded a fairly accurate knowledge of the atomic
weight. Too much cannot be said against the multiplication of incom-
plete or carelessly obtained data ; for such data carry with them not
only uncertainty and confusion in the present, but also additional
labor for a reviser in the future.

In the present investigation every reaction was assumed to involve
some constant error, and every substance was assumed to contain some
constant impurity, until a proof of the contrary was obtained. The
research consisted, in fact, of a succession of mutual checks. The
attempt was made to adopt precautions of a consistent order of refine-
ment ; and the still remaining causes of possible error have been care-
fully pointed out in the description. The not inconsidei'able outlay of
time, thought, and labor has been more than repaid by a conviction
of the definiteness of the combining proportions which four years ago
would have seemed to the writer impossible.

OP ARTS AND SCIENCES. ^95

XVI.

CONTRIBUTION FROM THE CHEMICAL LABORATORY OF
CLARK UNIVERSITY, WORCESTER, MASS.

THE ACTION OF ACETOACETIC ETHER ON QUINONES :
SYNTHESIS OF BENZOFURFURAN DERIVATIVES.

By M. Ikuta.

Presented by J. V. Nef, October 14, 1891.

When chloranil is treated with sodium acetoacetic ether, one or two
of its chlorine atoms can be replaced by acetoacetic ether groups. 1 he
study of the mono- and the di-substitution products obtamed has led
to an interesting synthesis of a number of furfuran derivatives.

Trichlorquinoneacetoacetic Ether, CeClaOaCCsHoOs).
10 crams of acetoacetic ether, dissolved in about 100 c.cm. of ben-
zine, are converted by means of sodium in the form of fine wire into
the sodium salt. To the emulsion thus obtained is slowly added 18
grams (1 molecule) of finely powdered chloranil with continual shak-
ing The reaction sets in immediately, and the solution becomes col-
ored blue, dark green, and finally brown. In order to complete the
reaction the mass is either warmed on a water-bath or allowed to stand
for several hours. The part insoluble in benzine, which is removed
by filtration and repeatedly washed with fresh benzine, is a very
bulky resinous mass, not practicable to work up further ; it consists
of sodium chloride, the sodium salts of the mono- and the di-substi-
tuted acetoacetic ether derivative of chloranil, besides much resinous

matter *

The benzine from the clear orange-colored filtrate is then distilled

* From this portion, by acidifying and extracting with benzine and crystal-
lizing from alcohol, a very slight amount of greenish yellow prisms melting at
183° soluble in alkalies with blue coloration, and also dark brown needles
insoluble in alkalies and melting above 270°, were obtained, but in too small
quantity to admit of further study.

296 PROCEEDINGS OF THE AMERICAN ACADEMY

off, and the last traces of it removed by prolonged heating in an open
dish on a water bath. A dark orange-colored oil is left, which is a
mixture of acetoacetic ether, and the mono- and the di-substituted
acetoacetic ether derivatives of chloranil. On adding a small amount
of absolute alcohol, and stirring, the oil gradually solidifies to a crys-
talline mass. Recrystallized from alcohol, large orange rhombohedra
of trichlorquinoneacetoacetic ether are obtained, as well as yellow
rhombic crystals of dichlorquinonediacetoacetic ether, which two sub-
stances can readily be separated mechanically from one another; the
latter is also somewhat less soluble in alcohol.

The red mono-substitution product was recrystallized from alcohol
until it showed the constant melting point 94°, and then dried at 70°
to 75° and analyzed.

0.3014 gram substance gave 0.4737 gram COg and 0.0738 gram HgO.
0.2524 " " " 0.3158 " AgCl (Carius).

Theory for CtsHaClsOg.

Found.

c

42.42

42.86

H

2.69

2.72

CI

31.37

30.95

Trichlorquinoneacetoacetic ether is easily soluble in benzine, warm
alcohol, and chloroform, moderately so in acetic acid, and difficultly
soluble in ether. It dissolves in dilute alkalies and ammonia with blue
color, changing quickly to orange-yellow on standing. An alcoholic
solution of the substance gives a blood-red coloration with ferric chlo-
ride. With aniline it forms a colorless addition product melting at
146°, whereas fuming nitric acid changes it into a yellow very unsta-
ble substance. Reducing agents readily convert it into the corre-
sponding hydroquinone.

Trichlorhydroquinoneacetoacetic Ether, C6Cl3(OH)2(CeH903).

The reduction of the quinone can be effected quantitatively by dis-
solving it in about 6 parts of alcohol, and adding alternately zinc dust
and dilute sulphuric acid in small quantity until the solution becomes
colorless. On adding water the hydroquinone separates out as an oil,
which soon solidifies. Recrystallized from benzine, it is obtained in
colorless rhombohedra, melting at 132°.

0.2204 gram substance dried at 105° gave 0.3444 gram COg and
0.0707 gram H2O.

OF ARTS AND SCIENCES.

297

Theory for CiaHiiClsOs

Found.

c

42.20

42.61

H

3.22

3.62

The hydroquinone is difficultly soluble in most organic solvents in
ihe cold. It dissolves in alkalies with pale yellow color, changing
quickly to dark green and yellow. An alcoholic solution gives with
ferric chloride a blood-red coloration (vide the quinone).

The most noteworthy property of this substance is the ease with
which it loses water going over into a benzofurfuran derivative.

Trichlor p-oxyhenzofurfuran a-methyl ^-carhoxylic Ether,
CI O

CI

HO

CH,

(a)

COaC.Hs (i8)

CI

One part of trichlorhydroquinoneacetoacetic ether is dissolved in 10
parts of acetic acid, and, after adding a few drops of concentrated sul-
phuric acid, the solution is boiled for about an hour, or until on add-
ing water to a portion a white precipitate is obtained which is not
affected by ferric chloride. On adding a small amount of water to
the hot acetic acid solution, the furfuran derivative crystallizes out in
colorless transparent needles melting at 138°. The substance was
recrystallized once from acetic acid, dried at 110°, and analyzed.

0.2530 gram substance gave 0.4131 gram CO2 and 0.0672 gram H2O.
0.2005 " " " 0.2638 " AgCl (Carius).

Theory for Ci^HgClA-

C 44.51

H 2.71

CI 32.92

Found.

44.53

2.95

32.54

The substance is easily soluble in chloroform, less so in alcohol,
ether, and cold acetic acid ; it gives no coloration in alcoholic solution
with ferric chloride, and dissolves in alkalies forming a colorless solu-
tion. On warming with concentrated sulphuric acid, a dark reddish
blue coloration is obtained, which is very characteristic of the sub-
stance, as well as of all benzofurfuran derivatives. On heating with
alcoholic potash, it is saponified to the corresponding monobasic acid.

298

PROCEEDINGS OF THE AMERICAN ACADEMY

The furfuran derivative is unquestionably formed according to the
following equation : —

CI

CI

o

CI
HO

-OH HO-C-CH3 CI
C-CO2C0H5 HO

C-CH3

IC-CO2C2H5

+ H/i

CI

CI

TricMorhydroquinoneacetoacetic Ether.

Benzofurfuraa Derivative.

That acetoacetic ether is not a ketone, but a phenol-like body (^-oxy-
crotonic ether), has been made highly probable by the work ot" Nef,
soon to api^ear in Liebig's Annalen. In order, therefore, to explain
the above reaction, the assumption of a pseudo or mobile modification
of trichlorhydroquinoneacetoacetic ether is unnecessary ; and the ease
with which this anhydride formation takes place depends entirely oii
the position of the atoms in space.*

* In some cases only the anli3'dricle can be obtained (see Ann. Chem. Lie-
big). In this connection I would lilce to point out, that diacetylsuccinic ether,

CH3 CH3

I I

COH HOC

HgC.O.C-C-

-C-C0.,C,H5,

according to Knorr, splits off on heating either water (200°), forming the fur-
furan derivative,

H3C— C C— CH3

II II

H5C2O2C-C C-C O2 C^ H5,

(Ber. d. chem. Ges., XVII. 2863), or alcohol (170°), forming isocarbopyrotritaric
etiier (Ber. d. cliein. Ges., XXK. 159). Knorr thinks it probable that the latter
substance is a ketopentamethylene derivative, just as Fittig (Ber. d. chem. Ges.,
XVIII. 3410) regards pyrotritaric ether as a ketopentamethylene derivative.
It is however far more probable that the splitting off of alcohol from diacetjl-
succinic ether takes place as follows :

H,C-C-OH C.,H,0-CO

H,C,0.,C-C-

*5'-'2'-'2

Diacetylsuccinic Etber.

-C=COH-CHo

H3C-C

H5C..O0C-C-

/0\

CO

I

-C=rCOH-CH,

+C,H50H

Isocarbopyrotritaric Ether.

OF ARTS AND SCIENCES.

299

Benzofurfurau derivatives were' first obtained by Fittig and Ebert*
from cumariue derivatives, and later Rossingf obtained cumarone
from o-aldehydopheuoxyacetic acid.

A general method for obtaining such compounds has been published
by IIantsch,t which consists in heating sodium salts of phenoles with
a-chloracetoacetic etlier, e. g. :

H

H O

H

H

ONa CI-C-CO2C2H5 H

+
H HO-C-CH3 jj

'C-CO2C2H5

-f NaCl

H HO-C-CH,

H H

Sodium Phenolate. a-Chloracctoacetic Ether. Intermediate Product.

H

O

H
H

H

C-CO2C2H5 (a)

+ H2O

C-CHg ifi)

Benzofurfuran Derivative.

While the above method leads to a-methyl /3-carboxylic derivatives,
that of Hantsch leads to /3-methyl a-carboxylic derivatives of benzo-
furfuran.

i. e. that under certain conditions of temperature tlie OH group of an acetyl
radical is nearer in space to tlie OC2H5 group of a COOC2H5 group, whereas at
a higher temperature (200°) the two OH groups of the acetyl radicals are nearer
to one another in space, so that either alcohol or water is split off according to
the temperature. Knorr also puts forward the above formula for isocarbopyro-
tritaric ether, but favors the ketopentamethylene formula because of tlie be-
havior of the ether towards alkalies (Ber. d. chem. Ges., XXII. 167). On the
other hand, all the oilier reactions of isocarbopyrotritaric ether are best ex-
plained by the above lactone formula. It is therefore highly probable that a
similar relationship exists in the very puzzling pyrotritaric acid series as has
been shown by Feist (Liebig, Ann. Chem., CCL(VII. 253) to exist in the case of
dehydracetic acid. — J. U. Nef.

* Ann. Chem. (Liebig), CCXVI. 170.

t Ber. d. chem. Ges., XVII. 3000.

X Ber. d. chem. Ges., XIX. 1292, 2930, XX. 1332.

300 PROCEEDINGS OP THE AMERICAN ACADEMY

Trichlor p-oxyhenzofurfuran a-methyl ^-carhoxylic Acid,

CI O

CI

HO

CH3
CO.,H.

CI

The ethyl ether just described is readily saponified by heating with
twice the calculated amount of alcoholic potash on a water bath.
After driving off the alcohol, and taking up with water, addition of
dilute acetic acid precipitates the free acid in the form of colorless
needles. It is purified by crystallization from glacial acetic acid, from
which solvent it separates out into colorless transparent needles melt-
ing at 258° and subliming without change. The crystals become
opaque on drying above 100°.

0.2736 gram substance dried at 130° gave 0.4090 gram CO2 and 0.0439

gram H2O.
0.2130 gram substance dried at 130° gave 0.3076 gram AgCl (Carius).

Theory for CjoHgOlsOi. Found.

C 40.61 40.76

H 1.70 1.78

CI 36.04 35.72

In its behavior towards organic solvents, concentrated sulphuric
acid, and alkalies, the free acid is like that of the ethyl ether.

p-DlcJilorquinonediacetoacetic Ether, C6Cl202(C6H903)2.

The method of preparing this substance is exactly the same as that
described above for the mono-substitution product, except that twice
as much sodium acetoacetic ether is taken. The crude product ob-
tained is purified by crystallization from alcohol until the melting point
of the substance reaches 127-128°. The alcoholic mother liquors
always contain some mono-substitution product as well as greenish
yellow prisms melting at 183° (mentioned in the foot-note, page 295).
The substance is thus obtained in pure yellow rhorabohedra or rhom-
bic plates.

0.2492 gram substance dried at 110° gave 0.4515 gram CO2 and

0.0944 gram HgO.
0.3003 gram substance dried at 110° gave 0.1947 gram AgCI (Carius).

OP ARTS A^D SCIENCES.

301

Theory for CigHjsOgCl,.

Found.

c

49.88

49.41

H

4.16

4.21

CI

16.39

16.04

The compound is easily soluble in benzine and chloroform, less so
in alcohol, acetic acid, and ether. It dissolves in dilute alkalies, mild
or caustic, forming deep bluish violet solutions, which soon change to
brown. Sodic nitrite reacts on an alcoholic solution in the cold, form-
ing a yellow substance which crystallizes in prisms, melts at 184°, and
contains both chlorine and nitrogen. Fuming nitric acid converts the
quinone into a very unstable yellow substance. An alcoholic solution
is colored deep blood-red by addition of ferric chloride. One of the
noteworthy properties is the ease with which the quinone forms addi-
tion products, and these are treated specially further on. A great
many experiments were made in order to prove that the two aceto-
acetic ether groups are in the para position, but without success. This
is, however, extremely probable, since in all cases yet known where
two chlorine atoms in chloranil have been replaced the substitution
takes place in para position (compare also Stieglitz, where two ma-
lonic ether groups are introduced in the para position*).

p-Dichlorhydroquinonediacetoacetic Ether ^ C6C1o(OH)2(CgH903)2.

The above quinone is reduced to the hydroquinone in exactly the
same way as described under trichlorhydroquinoneacetoacetic ether.
The solid obtained by adding water to the alcoholic solution is purified
by crystallization from benzine, from which solvent it separates out in
colorless transparent rhombohedra, melting at 154°.

0.2114 gram substance dried at 130° gave 0.3816 gram CO2 and
0.0881 gram H^O.

Theory for CigHooCljOg.

Found.

c

49.65

49.23

H

4.59

4.62

In its behavior towards ferric chloride, alkalies, and solvents, the
hydroquinone is quite similar to trichlorhydroquinoneacetoacetic ether.
The substance treated with dehydrating agents loses two molecules of
water with ease, going over into a benzodifurfuran derivative.

* American Chem. Journal, XIII. 38.

302

PROCEEDINGS OF THE AMERICAN ACADEMY

p-Dlchlorbenzo-p-difurfuran a-dlmethyl ^-dicarhoxylic Ether,

CI O

HjCoOoC

H,C

CH,

CO2C2H5.

O

CI

The above hydroquinone, when boiled for about an hour in acetic
acid solution containing a few drops of concentrated sulphuric acid, is
converted quantitatively into a benzodifurfuran derivative, which, since
it is insoluble in alkalies, can easily be separated from the original
substance. After precipitating by addition of water and treating with
dilute sodlc hydrate, the substance is crystallized from acetic acid
(95%) and obtained in long transparent prisms melting at 175°.

0.2006 gram substance dried at 120° gave 0.3965 gram CO2 and

0.0748 gram H2O.
0.2647 gram substance dried at 120° gave 0.1886 gram AgCl (Carius).

Theory for CisHieCl^OB.

Found.

c

54.13

53.95

H

4.01

4.16

CI

17.79

17.63

The fiirfuran compound is unquestionably formed according to the
following equation : —

CI

CO,C,H,-C

CH,-C-OH HO-

OH HO-C-CH,

-C CO2C2HS

CI

p-Diohlorhydroquinonediacetoacetic Ether.

CI o

C00COH5

CH3

CH,

CO2C2H5

+ 2H2O

O CI

^-Dichlorbenzo-^-difurfuran o-dimethyl j3-dicarboxyIic Ether.

OP ARTS AND SCIENCES.

303

The substance dissolved in concentrated sulphuric acid and warmed,
gives a dark bluish green coloration which is very characteristic. It
is difficultly soluble in most organic solvents except chloroform. By
passing chlorine into a chloroform solution of the substance it forms
a pale yellow oily addition product, which, on treating with zinc dust
and acetic acid, is converted back to the original substance. Alcoholic
potash saponifies it to the corresponding dibasic acid, —

p-Dichlorhenzo-p-difurfuran a-dimethyl ^-dicarboxylic Acid,

CI O

CO,H

CO9H.

The ethyl ether described above is easily saponified by heating with
twice the calculated amount of alcoholic potash. After driving off the
alcohol, taking up with water, and acidifying with dilute acetic acid,
the free acid separates out as a fine white powder, which, examined
under the microscope, consists of crystalline needles. It sublimes at
a higher temperature without melting, and is insoluble in organic sol-
vents and water. It was therefore dissolved in potassic carbonate,
precipitated by dilute acetic acid, and, after thorough washing with
hot water, dried at 130° and analyzed.

0.2087 gram substance gave 0.3732 gram CO., and 0.0482 gram H2O.
0.2461 "

u

" 0.2013 "

AgCl

(Carius).

Theory for CiiHgCUOs.

Found.

c

48.95

48.76

H

2.33

2.56

CI

20.69

20.24

Concentrated sulphuric acid gives a bluish green coloration similar
to that of the ether.

The Addition Products of p-Dichlorquinonediacetoacetic Ether.

As above mentioned, this substance is characterized by the ease with
which it forms addition products. Treated in chloroform solution
with one molecule of bromine, a complete absorption of bromine takes
place quickly without the slightest evolution of hydrobromic acid, and
a colorless odorless substance separates out in heavy scales, which melt

304

PROCEEDINGS OP THE AMERICAN ACADEMY

at 216°, The analysis of the product, made at different times and
washed with chloroform, gave no constant figures. The halogen de-
terminations gave results varying from 1% to 5% too high for that
calculated for a dibroinide, CigHisCLOs, Br^, and this points to the
presence of some tetrabromide. That the bromine has been added to
the double bonds present in the quinone group and not to the aceto-
acetic ether residues follows: 1st, because the substance sublimes
without decomposition ; 2d, the corresponding dichlorhydroquinone-
diacetoacetic ether is acted upon only very slowly by bromine, and
then with evolution of hydrobromic acid and formation of pungent
derivatives affecting the eyes ; 3d, the above addition product treated
with zinc dust and acetic acid, or in alcoholic solution with zinc dust
and dilute sulphuric acid, is converted chiefly into jo-dichlorbenzo-/)-di-
furfuran a-dimethyl ^-dicarboxylic ether.

The study of the product formed with bromine, though incomplete,
is thus sufficient to show that it is an addition product, and its study,
as well as that of other addition products, will be continued.

Noteworthy is the ease with which ^-dichlorquinonediacetoacetic
ether adds water, and is then converted into a new quinonefurfuran
derivative.

Quinone p-d/furfuran a-dimethyl ^-dicarhoxylic Ether Dihydrochloride,

CO,C,H,

CO.C.Hg.

;)-Dichlorquinonediacetoacetic ether is dissolved in 3 parts of acetic
acid (95%) and warmed with a kvf drops of dilute sulphuric acid
(1 : 5) on a water bath until the orange-colored solution becomes al-
most colorless. On diluting with water a pale yellow substance sepa-
rates out in oily drops, which soon solidify to a white crystalline solid.
The yield is almost quantitative, and after crystallizing from a mixture
of chloroform and ether the substance was analyzed with the following
result.

0.2010 gram substance dried in vacuum gave 0.3G41 gram CO2 and

0.074 G gram H2O.
0.2 11 2 gram substance dried in vacuum gave 0.1369 gr. AgCl (Carius).

OP ARTS AND SCIENCES.

305

Theory for C,8Hi8Cl;j08.

Found.

c

49.88

49.40

H

4.15

4.12

CI

1G.39

16.03

The substance melts at 171° to a pale yellow liquid, and sublimes
without decomposition. It crystallizes in star-shaped leaflets or short
prisms, and is easily soluble in organic solvents, but insoluble in water
or mild alkalies. Warmed with concentrated sulphuric acid it gives
a deep bluish green coloration, which is characteristic for furfuran
derivatives.

Treated with twice the calculated amount of alcoholic potash, it is
saponified to a dibasic acid, ChHioCLOs, which crystallizes from hot
water in colorless needles. The acid is best purified by dissolving in
soda and addition of dilute acids, when it separates out very slowly on
standing. It decomposes with evolution of carbon dioxide on heat-
ing to 220°. An analysis gave results which agree closely with the
formula C14H10CI2O8, but show that the acid was not quite pure.
(Found 43.53% C, 2.96% H, 20.50% CI. Theory, 44.56% C,
2.65 % H, 18.95 % CI.)

The formation of the above ester, CigHigClaOs, isomeric with p-di-
chlorquinonediacetoacetic ether is probably as follows. At first, an
addition of two molecules of water takes place, just as has been ob-
served in the case of many olefine derivatives:

O

CH3-C OH CI
CO2C2H5C

-C-CO.,CoH

2"5

+ 2HoO

CI HO C-CH,

O

p-Dichlorquinonediacetoacetic Ether.

CH3-C OH
COAH5-C

VOL. XXTI. (X. S. XVIII.

Cl\

HO/
H\

O

NH

O

Addition Product.
20

-C-CO,C,R

2'-'2i»6

/OH HOC-CH3
\C1

306

PROCEEDINGS OF THE AMERICAN ACADEMY

The addition product formed then loses two molecules of water, going
over into the furfuran derivative,

CH

CO2C2II5

CO.Calls.

Benzofarfuran Derivatives from Quinone and Acetoacetic Ether.

Von Pechraann * describes a condensation product obtained from
quinone and acetoacetic ether by means of alcoholic chloride of zinc.
He gives it the formula CisHigOe, and, as he states, it is formed
according to the following equation :

C6H4O2 + 2C6H10O3 = CiellicOo + H2O + CJloO.

From the product CisHigOe, he obtained further, by saponification
with alcoholic potash, a dibasic acid, C14H12OC.

Since v. Pechmann says that he does not intend to study the product
obtained any further, I have taken up the investigation of these sub-
stances, especially since the behavior of the compound CieHigOg is
very similar to the benzofurfuran derivatives obtained from chloranil,
and it therefore seemed probable that it possessed a similar consti-
tution. The formulae given by him for the above products are de-
duced from the mean result of a number of analyses. It is difficult,
for instance, to see how from a neutral substance, CigHieOe, by sa-
ponification, a dibasic acid, ChRi^Oq, can be formed, since the former
substance is not likely to contain methoxy groups. The study of
these substances soon showed that the empyrieal formulae given to the
condensation product as well as the dibasic ncid were incorrect.

The reaction between quinone and acetoacetic ether really takes
place according to the following equation :

I. CCII4O2 + 2 CcIIioOa = CisTIisOe + 2 H,0 + 2 II ;

the two hydrogen atoms are used up in converting a molecule of qui-
none into hydroquinone :

IT. C6H4O2 -f 2 n = CJI4 (011)2.

* Ber. d. chem. Ges., XXI. 3005.

OP ARTS AND SCIENCES.

307

The condensation product thus has the formula CisHigOc, instead of
CigIIicOg ; by saponification with alcoholic potash two molecules of
alcohol are split off, as follows :

CisHisOe + 2 H.O ^ Ci JlioOg + 2 CJIfiO.

The above mentioned dibasic acid has therefore the formula
C14H10OG, instead of CuIIioOc, and indeed the previous analyses
aofree better with this formula.

Theory for C„H,o08.

For ChHisOg-

Von Pechmann'a Analysis.

c

61.25

60.80

60.80

H

3.65

4.30

3.80

Benzo-p-difurfaron a-dimethyl ^-dicarhoxylic Ether,

H O

CO2C2H6

H,C

co^aHg.

When quinone (one molecule) is heated in alcoholic solution with
two molecules of acetoacetic ether in the presence of zinc chloride,
the condensation product CigHigOe is obtained, and as regards prop-
erties, etc., I can confirm in all detail the given statements. The
alcoholic mother liquors contain, however, hydroquinone as well as a
new substance, C12H12O4, described below, which products have been
overlooked. The analysis of the condensation product crystallized
from acetic acid, gave figures agreeing well with the formula CigHigOe.
For the combustion it was carefully mixed with powdered chromate
of lead, and the results obtained prove that it contains about 2%
more carbon than found by v. Pechmann. The benzofurfuran de-
rivatives all burn with great difficulty, so that most of the analyses
given in this paper were carried out with lead chromate.

0.1942 gram substance dried at 130° gave 0.4625 gram COg and

0.0983 gram HaO.
0.2027 gram substance dried at 130° gave 0.4840 gram CO2 and

0.1033 gram Hp.

c

Theory for

CigHigOe.

65.45

Found.
I. II.

64.96 65.12

Mean of former
Analyses.

63.00

Theory for

63.10

H

5.45

5.62 5.66

5.30

5.20

308

PKOCEEDINGS OF THE AMERICAN ACADEMY

I give here, for convenience' sake, the properties of the substance.*
" Coloi'less needles, melting at 184°. Insoluble in water or alkalies,
and soluble in chloroform and boiling acetic acid. The solution in
concentrated sulphuric acid turns deep blue on warming. Bromine
reacts, forming a crystalline derivative. Phenylhydrazin, benzoyl
chloride, and hydriodic acid have no action on the substance."

The proof that the compound CisHigOe has the constitution above
given is as follows : chlorine converts it into a disubstitution prod-
uct C18H16CI2O6, which is identical with ^-dichlorbenzo-jo-difurfuran
a-dimethyl /3-dicarboxylic ether,

O CI

CO2C2H5,

CH,

HsC.,0,C

already described and obtained from chloranil.

The substance is dissolved in acetic acid, and chlorine is passed into
the solution for several hours in the sunlight, until a portion treated
with zinc dust gives on addition of water a crystalline precipitate melt-
ing from 155-160°, when the rest of the solution is treated in the
same manner. By repeated crystallization from acetic acid (95%),
the crude product was finally converted into a substance which melts
constant at 175°, and crystallizes in long transparent needles iden-
tical in every respect with the p-dichlorbenzodifurfuran derivative
CisHieCloOe described above. This was also confirmed by a halogen
determination.

0.2138 gram substance dried at 120° gave 0.1499 gram AgCl (Carius).
CI

Theory for CigHioCUOg.

17.79

Found.

17.35

The substitution of the two hydrogen atoms of CigHigOo is un-
doubtedly preceded by the formation of an addition product, just as
CisHjcCloOe gives, when treated with chlorine, an oily addition product
as before mentioned, but the cldorine in these addition products is
easily taken out by zinc dust and acetic acid.

These results prove conclusively the constitution of the condensa-
tion product CigHifjOfi; the constitution of the dibasic acid Ci4HioOe

* Ber. d. chem. Ges., loc. cit.

OF ARTS AND SCIENCES.

309

obtained from it by means of alcoholic potash necessarily follows ; it
is a benzo-j9-difurfuran a-dimethyl jS-dicarboxylic acid,

O

IlaC

HOoC

H

H

CO.,H

CH„

O

p-Oxyhenzofurfuran a-dimethyl ^-carboxylic Ether,
H O

H;

no

CH«

CO2C2H5.

H

The above alcoholic chloride of zinc mother liquors, from which the
product CisHigOs has been removed, always contains hydroquinone
and a furfuran derivative, C12H12O4, besides acetoacetic ether. The
hydroquinone was isolated^ after driving off the alcohol, by taking up
water and extracting the aqueous solution with ether. The condensa-
tion product C12H12O4 is obtained by allowing the alcoholic mother
liquors to evaporate spontaneously, when it separates out in pearly
scales. It is formed from quinone and acetoacetic ether according to
the following equation :

CeH,02 + CeHioOs = C12H12O4 + H2O.

The best yield (50% of quinone taken) is therefore obtained by
heating quinone (1 molecule) with a molecule of acetoacetic ether in
the presence of alcoholic zinc chloride. After filtering off the insoluble
product CigHigOc, which is invariably formed, the filtrate is allowed
to evaporate in the air, and the pearly scales which separate out are
brought on porous plates and washed with water. Recrystallized
twice from petrolic ether (bpt. 70-80°), the substance is obtained in
colorless scales or in flat needles, melting at 137°.

0.1889 gram substance dried at 110° gave 0.4510 gram CO2 and
0.0997 gram HoO.

310 PROCEEDINGS OF THE AMERICAN ACADEMY

Theory for CjjUiA-

Found.

c

65.45

65.12

H

5.45

5.86

HC-CO-CH2

H^C-CO-CHBr

II 1

and

(II.) 1 1

HC-CO-CHBr

BrHC-CO-CHa,

The benzofurfuran derivative is insoluble in water, but, since it con-
tains a phenol group, soluble in alkalies. It is easily soluble in most
organic solvents except petrolic ether. Warmed with concentrated
sulphuric acid, it gives a reddish purple color.

The question now arises how the formation of the condensation pro-
ducts CigHigOe and C12H12O4 can be explained. Sarauw * has shown
that, when quinone is treated with hydrobromic acid, hydroquinoue
and mono- and di-bromhydroquinone are found.

Since Nef f has unquestionably proved the presence of two double
bonds in quinone,

CH-CO-CH

II II

CH-CO-CH,

it is highly probable that the action of the hydrobromic acid is to form
addition products,t

(I)

and that these are then converted into monobrom- and jo-dibrom-
hydroquinone respectively, two of the hydrogen atoms in the addi-
tion product II. being removed by quinone, which goes over into
hydroquinone.

It is becoming more and more apparent that all the reactions shown
by quinone which have hitherto not been understood are due to the
ease with which such addition takes place ; the activity of the double
bonds in quinone has been further shown above by the addition of
bromine and of water to jo-dichlorquinonediacetoacetic ether. It is
therefore very probable that acetoacetic ether reacts on quinone, just
as hydrobromic acid, forming at first addition products :

/H /^COH-CHg

,j . HC-CO-C C-COaCoHg

(I.) 2 . 5

HC-CO-C Ha

* Ann. Cliem. (Liebig), CCIX. 99.

t Anier. Chem. Journal, XII. 48.3, XIII. 422.

t Amer. Chem. Journal, XIII. 427.

OF ARTS AND SCIENCES.

311

and

(II.)

H3C-HOC,,
H5C2O2C-C-C-CO-CH2

h/ H

/
H2C-CO-C C-CO2C2H5

^ COH-CH3

and that these are then converted into the hydroquinoues,

H

H

HO

OH HO-C-CH

-C-CO2C2H5

H

HoC-C-OH HO

H

and

HsCsOqC-C-

-C-CO2C2H5

OH HO-C-CH3

H

(two of the hydrogen atoms in the addition product (II.) being taken
up by quinone, forming hydroquinoue).

The hydroquinones just as the corresponding hydroquinones obtained
above from chloranil, then lose water under the influence of the dehy-
drating agent present, and go over into the benzofurfuran derivatives,

H

HO

H

0

/\

CHs
CO2C2H5

H

and

O H

H«C

H5C2O2C

C02C2Hg

'CH«

H 0

312 PROCEEDINGS OP THE AMERICAN ACADEMY

The above explanation becomes the more probable, because the ad-
dition of acetoacetic ether in the form of the sodium salt to define
derivatives has already been accomplished in many cases by Michael,*
Auwers,t and others.

The work of which an account is given in this paper has been car-
ried out under the direction of Prof. J. U. Nef, to whom I wish here
to express my sincere thanks.

* J. prakt. Chem. (2), XXXV. 349, 449; Michael and Freer, Ibid., XLIII. 90.
t Ber. d. chem. Ges., XXIV. 307.

Worcester, Mass., September, 1891.

OP ARTS AND SCIENCES. 313

XVII.

NOTE ON THE VARIATION OF MOLECULAR

PRESSURE.

By Carl Barus.

Presented December 9, 1891.

1. The Method. — If a quautity of heat, 8 Q, be imparted to a sub-
stance, the result is usually expressed as 8 f^ + 8 U', where hU '\%
the internal, and 8 C/^' the external work done. For liquids at at-
mospheric pressure, 8 W is negligible.

Regarding 8 U from an experimental point of view, two salient
considerations present themselves. If I measure latent heats (A),
8 U includes both the dissociation energy and the expansion energy
necessary to effect fusion, as follows : —

a. The dissociation energy per gram necessary to change the
solid molecule into the liquid molecule, isothermally, without change
of volume,

h. The energy per gram for the isothermal expansion of the
molecule.

c. The energy per gram for the isothermal expansion of the con-
figuration of molecules which constitutes the given substance from
solid to liquid.

If I measure specific heats isopiestically, 8 U includes the purely
thermal energy necessary to effect changes of the body's temperature
and the expansion energy, viz. : —

d. The purely kinetic energy per gram per degree, by which the
solid is appreciably changed as to temperature. This is a temperakire
function only.

e. The energy per gram for the corresponding isothermal expansion
of the molecule.

/. The energy per gram for the corresponding isothermal expansion
of the configuration of molecules ; with similar quantities, o?', e\ f,
for the solid state.

The permissibility of thus separating 8 U into a thermal (tem-
perature) and a volume and thermal component is given, if the iso-

314 PROCEEDINGS OF THE AMERICAN ACADEMY

metrics are straight lines.* For liquid thymol this is very nearly
true. § 10.

2. In relation to a, I shall postulate : (1) A liquid or a gas
solidifies whenever, in consequence of changes of the physical or
chemical environment (temperature, stress external or internal, etc.),
the cohesive valency of the constituent atoms is sufficiently increased
to admit of the permanent construction of a tridimensional chain, the
type structure and the molecular weight of which (solid) are each a
multiple of the given structure and molecuhir weight. (2) A gas
liquefies whenever in consequence of changes of environment the cohe-
sive valency of the constituent atoms is increased, without affording
means for the permanent construction of the chain in question.

Thus, it is supposed that the same molecular nucleus is present in
the three states of aggregation, and that the aggregations are produced
by relatively weak affinities. Since, therefore, it is my purpose to
bring these forces to bear, at constant external pressure, when the mole-
cules are at successively different distances apart, it will facilitate
reasoning to assume at the outset that the energy necessary to effect
the change a, § 1, is constant as to temperature, at least within the
interval 0° to 50°, under experiment. §§ 8, 9, 12.

3. In relation to b and c, § 1, conjointly, I will proceed as fol-
lows. Anywhere within the given body let a small sphere be described ;
and conceive the expansion work between solid and liquid done
within this sphere by the applied heat to be represented by the iso-
thermal expansion against a pressure, p, acting on the surface of the
sphere. Then p may be called the molecular pressure and the expan-
sion work will be,

w

^Jpdv (1)

4. Now, if a substance can be found which under the same pres-
sure may be kept either in the solid or in the liquid state through a
considerable range of temperature, then if the latent heats within this
range regarded as functions of temperature, be co-ordinated with the
corresponding volume changes solid-liquid, relations for the energies
b and c, § 1 , are deducible.

Again, if with the knowledge of b and c in hand, specific heat
regarded as functions of temperature, be co-ordinated with the corre-
sponding thermal expansions, the characters of the energies e and f,

* Fitzgerald, Proc. Roy. Soc, Vol. XLII. p. 50, 1887.

OF ARTS AND SCIENCES.

315

for the solid and the liquid state, respectively, are more nearly ripe
for discussion.

This, in brief, is my method. What it thus far has yielded I will
now indicate.

5. Tliermal Expansion. — In virtue of the occurrence of volume
lag,* or hysteresis, substances of the kind needed in § 4 are forth-
coming. I infer, therefore, that as regards isopiestic thermal expan-
sion the liquid and the solid, as well as the gas, obey certain ideal
laws, which may be conceived to hold good even though temperature
be decreased below condensation or solidifying points, indefinitely.

I have worked with the beautifully crystalline solid thymol, which
melts slightly below 50° C, and when freshly distilled can be cooled
even below 0° C.

TABLE I. — Expansion of Thymol, v =:

vo

1-kd

Substance.

Vo X 105.

k X 10'.

Tempera-
ture
Interval.

No. of
Obser-
vations.

Bemarks.

Solid

96293
96369
96306
96579
96565
96565
96584

2300 + 90
2456

24S6 + 26
2300 + 80
2200 + 60
2260 + 6 0
2100 + 90

o o

0-32
0-84
0-33
0-34
0-84
0-41
0-44

8

10

10

6

o

7

10

j Pyknometer measure-
( ments.

) Expansions of povv-
> dered thymol under
) water.t

Expansions of a single
lump under mercury |

Liquid

101128
101128

7600 + 26
7500 + 6 0

0-46
0-53

12
5

S Pyknometer measure-
1 ments.

Pyknometer measure-
ments, liquid boiled
in vacuo. Air also
expelled by vacuum
solidification.

* Am. Journal, Vol. XLII. p. 125 et seq., 1891. Cf. Vol. XXXVIIL p. 408,
1889 ; Vol. XXXIX. pp. 490-494, 1890.

t Evaporation in the first set not allowed for ; hence the high Vq — .96369.

} The lump unavoidably contains vacuities and fissures within. Powdered
thymol cannot be put under mercury, free from air.

316

PROCEEDINGS OP THE AMERICAN ACADEMY

Measurements of thermal expansion made with a variety of methods
gave me the data of Table I. They have been conveniently put in
the form of an equation (Mendeleeff), r = t'o /l —kd, where v is
the specific volume of thymol at 6° C, and k is all but constant as to
temperature.

The discrepancies here observed are easily explained, being in part
due to unavoidable insulficiencies of the methods, and in part to the
special properties of thymol. The numbers in Italics correspond to
the conditions under which the calorimetry must be done, and are
therefore selected.

It will be seen that at zero the difference of specific volumes solid-
li<]uid is about .0485 c.c, whereas at 50° the same difference has in-
creased to .07G0 c.c, an increase of more than 50%. This cannot be
regarded as a mere differential. Indeed, the relations are such that
the curves if prolonged intersect at about — 100% after which the
volume of the solid exceeds tliat of the li(|uid. In other work * I
referred to the same point as the transitional temperature, and inferred
it from the tendency of the volume lag to vanish when temperature is
decreased below a certain value. The new evidence in favor of a
continuous jiassage of the normal type of fusion into the ice type is
noteworthy. Solidification contraction decreases with temperature at
a rate very much more rapid than tlie coetRcients of expansion.

G. Latent Heat. — A general survey of my best results for latent
heat (X) are given in the following table, for the intervals of temper-
ature specified.

TABLE II. — Latent Heat of Thymol varying with Temperature.

Temperature
Interval.

\

Temperature
IntervaL

\

Temperature
Interval.

A.

o o

3.2-5.1

23.0

25.9-30.7

26.3

35.6-37°7

22.6

4.2-6.1

23.4

25.8-30.4

26.9

35.1-39.0

23.4

2.5-4.7

24.8

25.8-30.4

26.9

36.2-39.2

22.7

25.7-30.3

26.9

36.0-39.6
36.4-39.9

26 0
23.7

Mean.

Mean.

Mean.

Mean.

Mean.

Mean.

4?.%

23.7

28°.l

26.8

37°.5

23.7

* Am. Journal, Vol. XLII. p. 145, 1891.

OP ARTS AND SCIENCES.

317

The difficulty in obtaining such results is excessive ; for, apart from
the serious complications of the method itself, (thus it takes hours
before complete solidification sets in at the higher temperatures, and
minute stages must be observed in order to allow for radiation,) the
observer is dealing with au (under-cooled) substance, which at the mere
suggestion of careless handling begins to freeze prematurely, and
which in the liquid state tends to become impure by absorption of air
or water vapor. These decompositions are gradual, yet they cannot
be disregarded, because they lower the melting point and thus induce
early partial fusions and late complete solidifications. The table also
shows the difficulty in working at other than room temperature.

7. Specific Heat. — For the reasons just stated, I shall have to give
my data for the specific heat of thymol provisionally ; for thougli I
made such experiments with great care and in considerable number, I
was not at tlie time fully conversant with the variety of precautions to
be taken to keep the substance pure.

TABLK III. — Specific IIkat of Sor.in and of Liquid Thymol.

Substance.

Temperature
latcTVul.

Specillc
Ueat.

Substance.

Temperature
Interval.

Specific
Ueat.

Solid

o o
20-32

22-43

.388
.430

Liquid

o o
25-54

.508

Solid
«

«

22-32
22-4 ;
22-48
24-49

.374
.429
.643
.969

Liquid
«

22-38
23-51

.506
.515

Solid after 1
distillation j

26-41

.504

1

Liquid dis- 1
tilled J

26-42

.509

These data, obtained with diflferent charges, arc much below the
sensitiveness of the method. The values for solid thymol, moreover,
show the unreasonably large influence of temperature, accounted for
in § 6.

8. Molecular Pressure. — Taking the data of Table II. at their face
value, the only justifiable conclusion to be derived is that A is con-
stant for the interval 0" to 40°. AVithin the same interval the differ-

318 PROCEEDINGS OF THE AMERICAN ACADEMY

ences of specific volume increase from .048 c.c. to .071 c.c, nearly 50%.
Hence, if either of these quantities be more than a vanishing increment,
their difference must also be.

From a different point of view : if the dissociation energy, q, is
large as compared with the expansion energy, then the thermal con-
stancy of X is of little consequence here. I make the supposition,
therefore, that q is not large relative to fpdv, a point which I will
endeavor to test in § 12.

Returning to equation (1), § 3, it therefore follows experimentally
that

XV, pV
pdv = q + I pdv (2)
"'^

where q^ and q are the dissociation energies, p the molecular pressure,
Vq and Vq the specific volumes solid and liquid respectively at zero
Centigrade, and where v and V have the same meaning at any given
temperature between zero and the melting point. If by § 2, ^'o = q,
then equation (2) may be reduced to

I p dv — I p dv = 0

(3)

an equation in which the molecular pressure is expressed in terms
of the thermal expansion of the liquid and the solid within the same
thermal limits, and by which any reasonable law of internal pressure
may be preliminarily tested.

Now suppose these integrations be successively taken at zero, and
all succeeding stages up to the melting point : then will the distance
apart of the limits of either term vary in any ratio relative to the
distance apart of the limits of the other. An equation

p(V-i()=c (4)

will therefore at least partially satisfy (3). In how far it may do so
throughout the whole interval 0° to 50° may be gauged by determin-
ing the constancy of x throughout this interval. This is done in
the next table, where ?', V, and >t are given for successive tempera-
tures d.

The constancy of x is thus marked, and quite within the range
of experimental errors ; and hence the equation (4) expresses the
law of force as well as any other function fitted to equation (3).

OF ARTS AND SCIENCES.

319

TABLE IV. — Values of k.

e

V

V

K

Mean k

o

0
30

.96300

.97035

1.01128)
1.03508 )

.94140

.94145

10
40

.96540

.97289

1.01905 )
1.04334 S

.94148

20

50

.96785
.97550

1.02698 )
1.05178 )

.94148

Since X is the minimum volume to which the h'quid may be reduced
when pressure increases indefinitely, it is interesting to note that
this volume is (1.05 — .94) /1. 05, or 10% below the liquid volume at
the melting point. Amagat* found that even at 3,000 atm. the vol-
ume of water at 17°. 6 was not decreased more than 10%. Again,
X = .9415 is very near the volume of solid and liquid thymol at the
transitional temperature, this volume being .9420. Thus through-
out this tentative work a certain degree of consistency is apparent,
always remembering that the approximations q^ ^ q and p^ ^ p is
not vouched for.

9. Equation (4) may be regarded as a point of departure from
which the further construction of the equation may be attempted.
Thus it is next in place to endeavor to ascertain how temperature
may be said to lurk in the quasi constant, p ( V — x). In a sub-
stance like thymol, which boils at 233°, the interval 4°-40° is too
small to bring out thermal variations appreciable by the above
method. Hence the deduction from (2),if^ {V — x) = g) {T),
though easily integrable, is not as yet available. I have therefore
sought to throw some light on the thermal variations of p, and on
the relative importance of q — qo by other methods.

In the first place I will note the possibility of getting rid of con-
siderations relative to the thermal variation of the dissociation energy,
q, as follows.

Suppose the specific heat at constant volume is the same for the

* Amagat, Compt. Rend., Tom. CUT. p. 429, 1886. According to Riicker
(Nature, Vol. XLI. p. 362, 1890) converging lines of evidence show that liquids
cannot be compressed more than .2 or .3 of their normal bulk.

320 PROCEEDINGS OF THE AMERICAN ACADEMY

liquid and for the solid, kept at the same volume and temperature.
Suppose also that the solid and the liquid isopiestics for a given
pressure do actually intersect at the transitional temperature, § 5.
If, therefore, the latter be taken as a point of departure, the total
energy communicated to the liquid up to the temperature t and
volume V, when a is the transitional volume, will be

q + J PdV+F(£),

where the three terms represent dissociation, expansion, and purely
thermal energy, respectively. The corresponding total energy im-
parted to the solid up to the temperature t will be

0+ r pdv+f(t).

a being the common volume at the transitional temperature. The
difference between these quantities is the latent heat, X, at t. Hence
the increase of latent heat from t^ to t will be, (since X is constant, and
F (t) — f (t) is assumed to be constant,)

f Pdv + f pdv = 0 (3')

an equation which differs from equation (3) in so far as P and p de-
note the internal pressures for the liquid and the solid states respect-
ively, under conditions in which external pressure is pronouncedly
variable. The equation (3') would still apply if the specific heats
at constant volume differ only by a constant appreciably within the
limits of observation (0°-50°), but it is not available for practical
comparisons.

10. 27ie Isometrics. — Some time ago I showed* that within a range
of 1,000 atmospheres of external pressure, at least, and within rea-
sonable limits for the thermal stability of organic bodies, the iso-
metrics of liquids, and particularly of thymol, f are very nearly
straight. Thus the extension of equation (4) would be

ip+p') (V->i)=cT,

* Phil. Mag., (5,) Vol. XXX. p. 348, 1890.
t Ibid., p. 358, and Plate XI.

OF ARTS AND SCIENCES.

321

where /)' is the external pressure and T is temperature. In my
experiments, only a single value of volume is introduced, and hence
the work is not further available. Referring to the iutriusic equation
of Ramsay, Young, and Fitzgerald (loc. cit., p. 51),

p'^RTl{v-v,)-iiiv\

so that here the internal pressure has the form p = fi/v\

11. Isothermals. — The question ceases to retain so simple an
aspect when the isothermals of liquids are considered. Of the two
equations which I discussed in my work * on the compressibility of
liquids, viz. :

dp'

d

1 + ap'

and

dV
d^

fi

(1 + vp')

IV2

• (6)

where p' is the external pressure, and d, a, jW, v are constants, the
latter is at once compatible with the results of § 9. For equation (3)
leads to

dV _ _ ^(T)

dp p^

(7)

and hence in equation (4), llv=p is the internal pressure active
under the isothermal circumstances of compression, and

The results actually found for thymol are given in Table V., 2" being
the absolute temperature of the isothermals, and pressures being in
atmospheres.

TABLE V. — Compressibility of Thymol,

dV
dp

nlv^

(l/r + /)2

T

IJ. X lOe

V X io»

Vv

M/r2

Remarks.

o

Atm.

301

66

300

3370

750

Initial pressure, p'q — 20 atm.

338

73

330

3050

680

Interval of observation, 400 atm.

373

97

435

2300

510

-

lj.lv - const. = .222.

458

162

730

1370

304

Melting point, 49°.5 C.

583

481

2160

460

102

Boiling point, 233° C.

* Am. Journal, (3,) Vol. XXXIX. pp. 497 and 506, 1890.

VOL. XXVI. (N. S. XVIII.) 21

322 PROCEEDINGS OF THE AMERICAN ACADEMY

Thus it is seen that thej internal pressures (Ifv) decrease from a
value of nearly 4,000 atm. at zero Centigrade, indefinitely. The same
is true of jtt'r', but in neither case are the observations sharp enough
to indicate the nature of this variation. Indeed, to co-ordinate all the
results, I used the smoothing process fi/v = const. Hence I shall have
to sketch the mere trend of the data here involved, by grouping the
values fi/v"^ along some mean straight line like .29 (560 — 7^ r^ g) (7^),
wherein preference is given to low temperatures. Thus the volume
equation takes the form

_ 2^9(0^^

When p :=^ — p', the liquid will boil, and consequently, since v — x
remains finite, 7*= 560. This number stands not unreasonably for
the absolute boiling point of thymol at the external pressure p',
which in the last table is 20 atmospheres.

12. The question now arises how the result (8) compares with the
calorimetric equation

q + \ pdv = X,

where X = 25 is at the outset considered constant as to temperature.

Availing myself of equation (8), expressing pressures in degrees
per square centimeter instead of in atmospheres, and remembering
that Joule's equivalent is 4.2 X 10'', I find

/

^^ 2 9 F — X

pdv = ^ X (5Q0- T)X In ^ ... (9)

Table IV. shows that throughout the interval 0° to 50°

hi ( V— x) / (v — x) = 1.18 nearly.

Hence the values of the dissociation energy q are in gram calories,

0° pdv

20°
40°

Thus the dissociation energy is small as compared with the expansion
energy ; but q increases with temperature which is unreasonable. To
explain this discrepancy it is necessary to revert to Table II., supposing

23.4

q^ 1.6

21.8

3.2

20.1

4.9

OF ARTS AND SCIENCES.

323

the constancy of k no longer tenable. Thus it appears that the march
of q is within the errors of I, and hence it has no meaning. Cf. § 9.
It follows, in general, therefore, that the results obtained iu measure-
ments of the compressibility of thymol are not inconsistent with the
calorimetric data ; and it has thus been brought out, inasmuch as the
dissociation energy is not large relative to the expansion energy, that
the premises of § 8 are sustained.

18. Energy and Volume. — I have finally to touch upon the data
obtained for specific heat, and their bearing on the purely thermal
energy of thymol. It is interesting in the first place to compare the
heat absorbed per unit of volume increase in the solid state, the liquid
state, and during the change of state from solid to liquid. This has
been done in the following table.

TABLE VL — Volume Relations of Thermal Capacity.

Liquid Interval 22^ to 38='.

Fusion.

Solid Interval 22° to 32^.

Mean spec, heat, .506

\ Mean volume in-

i crease .01307 cc.

[coefficient .000817

AQ
Ratio -- 620
A V

Latent heat, 25.4

C Volume increase,
\ .0589 at 19°.3
( .0655 at 31°.8

S 4-0 1
\ 390 S

Mean spec, heat, .381

f Mean volume in-

, crease . . .002-'^0

L Coefficient . .00025
1520

Thus it appears that the energy expended per unit of volume in-
crement is in marked degree greater for the solid than for the liquid
state, and is greater in both of these states than during the inter-
mediate fusion.

14. Sufficient data are now in hand for the computation of the
specific heat at constant volume of thymol. Applying the well known
thermodynamic relations (cf. Clausius, Chap. VIII., § 5) at 29°,
Table V. shows (d Vejdp) = 66/10^- in terms of dynes per square
centimeter and atmospheric pressure nearly. Under the same con-
ditions {d Vpidd) = 801 /1 0« from Table I. Hence the specific heat
at constant volume C„ is, from Table III.,

0, = .506 - ((29 + 273)/42 x 10'') ((801)2/66) = .436,

or about .86 of the specific heat at constant pressure C^ = 506.

324 PROCEEDINGS OF THE AMERICAN ACADEMY

The value C^^ A4 is larger than the solid specific heats at constant
pressure for the same interval (Table III.), and hence a fortiori larger
than the corresponding solid specific heats at constant volume, which
are as yet indeterminable. § 9.

With this result, a direct computation based on equation (9) above
may be compared. Integrating betvreen 22° and 38° (Table VI.), —
i. e. betveeen the specific volumes 1.02860 and 1.04167, — the expan-
sion energy is found to be 2.56 to 2.42 g. cal., according as the mo-
lecular force at the beginning or the end of the interval is inserted.
The total energy imparted within this interval is .506 X 16- Hence
the purely thermal energy per degree between 22° and 38° has the
mean value of .35, or about two thirds the specific heat at constant
pressure.

Compared with the preceding datum, .44, this result is considerably
too small ; i. e. the expansion energy applied is about twice as large
as its true value. I might point this out as an indication of the diffi-
culty encountered in deducing internal pressure from the curvature
of the isothermals (§ 11) ; but the true nature of the discrepancy is
an error in principle, since the change of internal pressure with tem-
perature must be radically different when the body is kept at constant
external pressure in one case, and at constant volume in the other.
Hence I do not regard the present result as conflicting with § 12.

15. Melting Point and Pressure. — The data of Tables I. and H.
suflfice for a preliminary computation of this value. Using the nota-
tion of Clausius (Warmetheorie, Chap. VII,, § 2), Table I. shows that
at T = 50° (melting point, nearly), a — r = .0763. From Table II.
the mean value (0°-50°) of r' is 24.7, the minimum value 23, and
the maximum 27. Hence, if pressures be expressed in atmospheres
instead of dynes per sq. cm., the mean relation of melting point to
pressure is d T/ dp = .024 ; or, more clearly,

.025 >dTldp > .022,

where the upper limit is probably more nearly correct. This value
lies well within the margin of data (.020 to .036) thus far in hand,
and found with wax, paraffine, spermaceti, and naphthalene. From
a perusal of my work on the continuity of solid and liquid,* it appears
clearly, however, that a more accurate value of o? Tdp is obtainable
from direct experiment, from which a value for latent heat may
then be computed by inverting the thermodynamic ecjuation.

* Am. Journal, Vol. XLII. p. 144, 1891.

OF ARTS AND SCIENCES. 325

16. In the above paragraphs I have endeavored to exhibit the
stage of progress attained iu the work on molecular pressure, at the
time when it had to be abandoned ; for the calorimetric work can
only be satisfactorily done in midwinter, Cf. §§ 5, 6, 7. It will be
seen, I think, that the subject, as a whole, is ripe for sharper tests,
and the work will therefore be resumed at an early opportunity.

Physical Laboratory, U. S. G. S.,
Washington, D. C.

PROCEEDINGS.

Eight hundred and thirty-third Meeting.

May 27, 1890. — Annual Meeting.

The Vice-President in the chair.

The Corresponding Secretary read an obituary notice of
the late Rowland G. Hazard.

The following gentlemen were elected members of the
Academy : —

Henry Newell Martin, of Baltimore, to be an Associate
Fellow in Class II., Section 3, in place of the late John
C. Dal ton.

Sir Henry Enfield Roscoe, of London, to be a Foreign
Honorary Member in Class II., Section 1, in place of the
late James P. Joule.

On the motion of the Corresponding Secretary it was

Voted, To meet, on adjournment, on the second Wednes-
day in June, at half-past seven o'clock.

The Treasurer's report was read and accepted.

The report of the Librarian was read and accepted.

On motion of the Corresponding Secretary it was

Voted, To appoint a committee to consider a proposed
change in the statutes as they affect annual assessments.

The Treasurer, the Recording Secretary, and Dr. H. W.
Williams were appointed.

Voted, To appropriate the sum of twenty-five hundred dol-
lars ($2500) for the expenses of publication for the ensuing
year.

Voted, That an appropriation of twelve hundred dollars
(81200) be made for the purchase and binding of books for
the ensuing year.

328 PROCEEDINGS OF THE AMERICAN ACADEMY

The annual election resulted in the choice of the following
officers : —

Joseph Loveetng, President.
Andrew P. Peabody, Vice-President.
JosiAH P. Cooke, Corresponding Secretary.
William Watson, Recording Secretary.
Eliot C. Clarke, Treasurer.
Henry W. Haynes, Librarian.

Council.
Amos E. Dolbear,
Arthur Searle, \ of Class I.

William E. Story,

William G. Farlow,

Samuel H. Scudder, \ of Class II.

David W. Cheever,

William Everett,

Martin Brimmer, \ of Class III.

Edward J. Lowell,

Rumford Committee.

Wolcott Gibbs, Joseph Lovering,

John Trowbridge, George B Clark,

JosiAH P. Cooke, Erasmus D. Leavitt,

Benjamin O. Peirce.

Member of the Committee of Finance.
Thomas T. Bouvj^.

The following papers were presented by title : —

On Electrical Oscillations or Waves. By John Trowbridge
and W. C. Sabine.

On the Spectroscopic Study of Electrical Waves. By John
Trowbridge.

On the Action of Nitric Acid upon Bromtrinitrophenylma-
lonic Ester. By C. Loring Jackson and W, B. Bentley.

On the Action of Sodic Alcoholates on Nitrotribrorabenzols.
By C. Loring Jackson and W. H. Warren.

OF ARTS AND SCIENCES. 329

Eight hundred and thirty-fourth Meeting.

June 11, 1890. — Adjourned Annual Meeting.

The Peesident in the chair.

The report of the Council was read and accepted.

The following report was read : —

The Rumford Committee present the following brief report
for the year ending with the annual meeting of 1890.

1. The Committee have had under consideration several
applications for grants from the income of the Rumford
Fund, to assist in meeting the expense of original investiga-
tions relating to light or heat. They have acted favorably
upon only one of these applications, and their report upon,
it was approved by the Academy on October 5, 1889.

2. The Committee have also given much time to the dis-
cussion of the relative claims of various suitable candidates
for the Rumford Medal ; but they are not yet prepared to
make any recommendation upon that subject.

Respectfully submitted,

Joseph Lovering, Chairman.
June 11, 1890.

The report of the committee appointed to consider a pro-
posed change in the statutes as they affect annual asssess-
ments was read and accepted.

On the motion of the Corresponding Secretary it was

Voted, That a committee, consisting of the Recording
Secretary and Messrs. H. W. Williams and H. G. Denny,
be appointed for the better ordering of the meetings of the
Academy.

Remarks upon the subject were made by Judge Nathaniel
Holmes, Dr. Williams, and the Recording Secretary.

The President appointed the following standing commit-
tees : —

Committee of Publication.

JosiAH P. Cooke, William G. Faelow,

John C. Ropes.

830 PROCEEDINGS OF THE AMERICAN ACADEMY

Committee on the Library.

Henry P. Bowditch, Amos E. Dolbear,
Edward J. Lowell.

Auditing Committee.
Henry G. Denny, Thomas T. Bouv6.

The following papers were presented by title : —

On the Candle Power of Incandescent Lamps, as related
to Current, Voltage, and Energy consumed. By Louis A.
Ferguson and David A. Center.

Notes on Zonaria variegata. By Herbert M. Richards.

The Recording Secretary called the attention of the Acad-
emy to the proposed monument to be erected at Genne-
villiers, near Paris, to the memory of the celebrated sanita-
rian, Alfred Durand-Claye.

Professor Dolbear made a communication on Vortex Rings,
and remarks on this subject were made bj^ Professor Story.

Eight hundred and thirty-fifth Meeting.

October 8, 1890. — Stated Meeting.

The President in the chair.

The Corresponding Secretary read a letter from the Presi-
dent of the Academy of Natural Sciences at Catania, an-
nouncing the death of Orazio Silvestri, Professor in the
University of Catania ; also, a letter from the President of
the Geographical Society of Berne,' announcing the approach-
ing meeting of the International Geographical Congress of
1891.

Sir William Bowman, Bart., of London, was elected a
Foreign Honorary Member in Class II., Section 4, in place
of the late Franz Cornells Donders.

On the motion of the Treasurer it was

Voted., That Section 2 of Chapter I. of the statutes be
amended by substituting for it the following : —

OF ARTS AND SCIENCES. 331

" 2. Fellows, resident in the State of Massachusetts, only,
may vote at the meetings of the Academy. Each Resident
Fellow shall pay an admission fee of ten dollars and, such
annual assessment, not exceeding ten dollars as shall be
voted by the Academy at each annual meeting."

Voted, That the assessment for the current year be five
dollars.

The following papers were presented by title : —

On the Composition of Certain Petroleum Oils and Refin-
ing Residues. By Charles F. Mabery.

The Analysis of Cupric Bromide, and the Atomic Weight
of Copper. By Theodore W. Richards.

The Effect of Electrical Oscillations on the Molecules of
Iron. By John Trowbridge.

A biographical notice of James Prescott Joule was read by
Professor Dolbear.

Eight hundred and thirty-sixth Meeting.

November 12, 1890. — Monthly Meeting.

The President in the chair.

Mr. George B. Clark made a communication on telescope
lenses.

The following papers were presented by title : —

On the Influence of the Strength of the Magnet in a
Magneto Telephone Receiver. By Charles R. Cross and
Harry E. Hayes.

On Electrical Condensers. By Charles Nutt.

Eight hundred and thirty-seventh Meeting.

December 18, 1890. — Monthly Meeting.

The President in the chair.

The Corresponding Secretary read a letter from the Inter-
national Congress of Ornithology, inviting the Academy to
send delegates to its approaching meeting at Budapest in
May, 1891 ; also, a letter from the American Chemical Soci-

332 PROCEEDINGS OF THE AMERICAN ACADEMY

ety, giving notice of its next meeting, and inviting all chem-
ists to be present.

A paper by Dr. Oliver W. Huntington, entitled " A New
Meteoric Iron from Stutsman County, South Dakota," was
presented by title.

A notice of the late Maria Mitchell, written by her brother,
Henry Mitchell, was read by the Corresponding Secretary.

Eight hundred and thirty-eighth Meeting.

January 14, 1891. — Stated Meeting.

The President in the chair.

The Corresponding Secretary read a communication from
the Massachusetts Historical Society, inviting the Academy
to send two delegates to its Centennial Anniversary on Jan-
uary 24, 1891 ; and it was

Voted, To accept this invitation, and to appoint the Presi-
dent and one Fellow of the Academy, to be selected by him,
as delegates.

The following gentlemen were elected members of the
Academy : —

Arthur Messinger Comey, of Somerville, to be a Resident
Fellow in Class I., Section 3.

Charles Edward Munroe, of Cambridge, to be a Resident
Fellow in Class I., Section 3.

John Ulric Nef, of Worcester, to be a Resident Fellow in
Class I., Section 3.

Theodore William Richards, of Cambridge, to be a Resi-
dent Fellow in Class I., Section 3.

Charles Robert Sanger, of Cambridge, to be a Resident
Fellow in Class I., Section 3.

The President, as chairman of the Rumford Committee,
reported the recommendation of the Committee, that one
hundred dollars be appropriated from the income of the
Rumford Fund to assist Dr. E. H. Hall in his study of the
periodic changes in the walls of a steam-boiler.

Also, that two hundred dollars be appropriated from the

OF ARTS AND SCIENCES. 333

same income to assist Professor B. O. Peirce in his investi-
sration on the conduction of heat in the interior of sohd
bodies. It was accordingly

Voted, That the Treasurer be authorized to pay these ap-
propriations from the income of the Rumford Fund.

Voted, That the sum of two hundred dollars be appropri-
ated to defray the expenses of the social meetings of the
Academy for the present year.

Professor Amos E. Dolbear read a paper on Electro-mag-
netic Waves.

Professor W. G. Farlow presented, by title, a paper by
Roland Thaxter, entitled, " Supplementary Notes on North
American Laboulbeniacese."

I^ight hundred and thirty-ninth Meeting.

February 11, 1891. ^Monthly Meeting.

The President in the chair.

The Corresponding Secretary read the following letters :
from the National Society of Horticulture of France, invit-
ing the Academy to send delegates to the Seventh Horti-
cultural Congress, to be held at Paris in 1891 ; from the
Secretary of the Society of Natural Sciences at Milan, an-
nouncing the death of Professor Antonio Stoppani ; from
Professor Charles E. Munroe, acknowledging his election as
a Fellow of the Academy ; and from Mr. William C. Collar,
resigning his Fellowship.

Professor Mark presented, by title, " A Preliminary Notice
on Budding in Bryozoa."

Eight hundred and fortieth Meeting.

March 11, 1891. — Stated Meeting.

The President in the chair.

In the absence of the Recording Secretary, Professor
Haynes was appointed Secretary pro tern.

334 PROCEEDINGS OF THE AMERICAN ACADEMY

The Corresponding Secretary announced the decease of
members since the 29th of May, 1890, as follows : —

Henry Jacob Bigelow, Charles Otis Boutelle, and William
Prescott Dexter, Resident Fellows ; George Bancroft, John
Charles Fremont, and Christian Henry Frederick Peters,
Associate Fellows.

The following correspondence was presented : a letter
from the President of the Royal Academy of Sciences of
Belgium, announcing the death of its Permanent Secretary,
Jean Baptiste Joseph Liagre ; and a circular from the Com-
mittee of Organization of the Fifth International Congress of
Geologists, inviting the members of the Academy to attend
the meeting of the Congress in Washington in August next.

Professor Henry P. Bowditch brought to the attention of
the Academy a circular, signed by numerous men of science,
soliciting subscriptions for a marble bust and medal in
commemoration of the seventieth birthday of Hermann von
Helmholtz.

Professor Farlow presented, by title, a paper by W. Albert
Setchell, entitled, " Preliminary Notes on the Species of
Doassansia, Cornu."

Professor Dolbear gave an account of his researches to
discover evidence of momentum from electrical action.

Dr. William Everett spoke at length upon the deficiencies
of elmentary text-books in secondary schools.

Eight hundred and forty-first Meeting:.

April 8, 1891. — Monthly Meeting.

The Academy met at the Harvard Medical School, Boston.

The President in the chair.

The following papers were presented : —

Demonstration of Electrical, Optical, and Mechanical
Apparatus for Instruction and Research. By Henry P.
Bowditch.

On the Means of teaching Anatomy, illustrated by new
Models and Specimens. By Thomas Dwight.

OF ARTS AND SCIENCES. 335

The following papers were presented by title : —
Contributions to American Botany: I. Additions to the
Flora of Western North America. II. Descriptions of New
Species of Plants collected in Mexico, chiefly by Mr. C. G.
Pringle, in 1889 and 1890. By Sereno Watson.

Contribution from the Herbarium of Harvard University :
Descriptions of New Plants, chiefly Mexican Gatnopetalce,
collected by Mr. 0. G. Pringle in 1889 and 1890. By B. L.
Robinson.

Eight hundred and forty-second Meeting*

May 13, 1891. — Monthly Meeting.

The President in the chair.

The following papers were presented by title : —

On some Theorems which connect together certain Line
and Surface Integrals. By B. O. Peirce.

On some Cases of Electrical Flow in Thin Conductors.
By B. O. Peirce.

The Quantitative Determination of Arsenic by the Marsh-
Berzelius Method, especially as applied to the Analysis of
Wall Papers and Fabrics. By Charles R. Sanger.

On the Products obtained by the Action of Nitric Acid
upon Bromtrinitrophenylmalonic Ester. By C. Loring Jack-
son and W. B. Bentley.

Note on Tribrommononitrobenzol. By C. Loring Jackson
and W. B. Bentley.

On the Structure and Development of Choreocolax Poly-
siphonisB, Reinsch. By Herbert M. Richards.

Professor Edwin H. Hall gave an exposition of his thermo-
electric method for the study of cylinder condensation in
steam-engines. The apparatus, provided for by an appropri-
ation from the income of the Rumford Fund, was exhibited.

i

AMERICAN ACADEMY OF ARTS AND SCIENCES.

Report op the Council. — Presented May 26, 1891.

BIOGRAPHICAL NOTICES.

Henry Jacob Bigelow By Oliver Wendell Holmes.

Charles Otis Boutelle Edward Goodfellow.

Alfred Hosmer Henry W. Williams.

George Bancroft A. McFarland Davis.

Julius Erasmus Hilgard O. H. Tittmann.

Christian Heinrich Friedrich Pet kiss Arthur Searle.

Charles John Maximowicz Sereno Watson.

Karl Wilhelm von Naegeli .... William G. Farlow.

Eduard Schonfeld Arthur Searle.

Notices of William Prescott Dexter, John Charles Fremont, John Le Conte,
and Joseph Leidy, must be deferred until the next volume.

EEPORT OF THE COUNCIL.

Since the last annual meeting, May 27, 1890, the Acad-
emy has lost by death twelve members ; — viz. four Resident
Fellows, Henry Jacob Bigelow, Charles Otis Boutelle, William
Prescott Dexter, and Alfred Hosmer ; six Associate Fellows,
George Bancroft, John Charles Fremont, Julius Erasmus
Hilgard, John Le Conte, Joseph Leidy, and Christian Heiu-
rich Friedrich Peters ; and three Foreign Honorary Members,
Charles John Maximowicz, Karl Wilhelm von Naegeli, and
Eduard Schonfeld.

RESIDENT FELLOWS.

HENRY JACOB BIGELOW.

Henry Jacob Bigelow was born in Boston, March 11, 1818,
and died in Newton, Mass., October 30, 1890. He was the oldest of
five children of Jacob and Mary (Scollay) Bigelow. His father was
distinofuished in various branches of science and literature ; he was a
former President of the Massachusetts Medical Society, and a Presi-
dent of tliis Academy ; a man of great ability, a leading practitioner in
Boston during his long life, and especially memorable as the founder
of Mount Auburn, the earliest of our garden cemeteries. His son
inherited many of his father's qualities. After attending Mr. Thayer's
school, which he entered in 1826, he joined the Latin School, then
under the charge of Mr. Leverett. When Mr. Leverett left the
Latin School and established one of his own, he followed his instruc-
tor, having: amonor |,is schoolmates William M. Evarts and William
W. Greenough. He entered Harvard College in 1833, graduating in
1837. "If he does not become a distinguished man," Dr. James
Jackson is reported to have said of him, " it will be because Boston
is not a large enough field for his ability."

Mr. Henry Lee writes an interesting account of the early years he
and Henry Bigelow passed together, from the age of three until Mr.
Lee left to go to college, a year before his companion. He describes
his young friend as a slender boy, lithe and active, a good gymnast
and dancer, and full of contrivances and ideas of all sorts. He had a

340 HENRY JACOB BIGELOW.

rather remarkable facility for mechanical work, — took early to shoot-
ing, a taste which lasted to the later years of liis life ; he was also
foud of bird's-nesting, with the usual knowledge, or rather more, of
birds and their haunts and habits ; like his father, he had a taste for
botany, which came agaiu very strongly in his later years. He was
a fair thouoh not remarkable scholar, through school and colleofe.

He early showed his independence of character. There was a
rebellion while he was in college, and anxious parents went out to
look after their sons, — among them Dr. Jacob Bigelow, who remon-
strated with Henry. The latter reminded him that there was a re-
bellion in his own day. '• Yes," said his father, " but I have seen the
folly of it." " Well, I want to see the folly of it too," was Henry's
(characteristic) answer.

He graduated with respectable rank in 1837. After leaving college
he had threatening symptoms of pulmonary disease, for which he went
to Havana ; but he was able to continue the study of medicine which
he had already commenced, in the prosecution of which he went to
Europe, passing his time chiefly in Paris, visiting London, more es-
pecially to hear the lectures of Sir James Paget. He took his med-
ical degree at Harvard University in 1841, and entered upon practice
in Boston. He had determined to devote himself to surgery, and soon
found himself in active business.

In connection with Dr. Henry Bryant, he established a kind of
surgical dispensary, which was the subject of no little comment and
some harmless satire from unknown rivals, which amused him and his
friends as much as it did any of the medical community.

He soon became known as an enterprising and aspiring practi-
tioner, who was mapping out his own path, deterred by no fear of
rivals, and not afraid of his critics.

First on the list of Dr. Bigelow's publi>hed writings stands a
"Manual of Orthopedic Surgery," being a Boylston Prize Disser-
tation for the year 1844.

The Boylston Prize Fund was provided by the generosity of Ward
Nicholas Boylston, a Boston merchant. Its two annual prizes of
fifty dollars each invited the competition of the younger members of
the medical profession, and the gaining of them was a favorable intro-
duction of the young practitioner to the medical world and the gen-
eral public. The question, or one of the questions, for the year
1844 was the following : " In what cases, and to what extent, is the
division of muscles, tendons, or other parts, proper for the relief of
deformity or lameness ? " Dr. Bigelow did not confine himself

HENRY JACOli BIGELOW. 341

strictly witliiii the limits of the question, but extended his labor until
it took the form of tlie Manual above mentioned, au octavo volume
of more than two hundred pages. This was a systematic and lucid
treatise, far beyond the ordinary standard of the annual dissertations
in scope and completeness.

For more than thirty years — from 1849 to 1882 — he was Pro-
fessor of Surgery in the Medical School of Hai'vard University. In
1847 he was appointed one of the surgeons to the Massachusetts Gen-
eral Hospital. During all this active period of his life, he published
many important papers, bearing more especially upon, but not con-
fined to, surgical practice. The following list is furnished by Dr.
R. H. Fitz, at the close of his tribute to Dr. Bigelow at the me-
morial meeting of the Society for Medical Improvement.

A List of some of the more importatit of Dr. Bif/elow's Contributions

to Medical Literature.

Manual of Orthopedic Surgery. Boylston Prize Dissertation. 1845.
Fragments of Medical Science and Art. An Address delivered before the

BoyLston Medical Society. 1818.
Insensibility during Surgical Operations produced by Inhalation. Boston

Medical and Surgical Journal. 1816.
On a New Physical Sign, a Clicking in the Throat. Boston Medical and

Surgical Journal. 1817.
Anaesthetic Agents, their Mode of Exhibition and Physiological Effects.

Transactions of American Medical Association. 1818.
Ether and Chloroform: a Compendium of their History, Surgical Use,

Dangers, and Discovery. 1848.
On the Employment of a New Agent in the Treatment of Stricture of the

Urethra. Boston Medical and Surgical Journal. 1849.
An Introductory Lecture. 1849.
Dr. Hallow's Case of Crowbar Injury to the Head. Philadelphia Medical

Journal. 1850.
Notes from Clinical Lectures on Surgery. 1851.
Science and Success. A Valedictory Address. 1859.
Surgical Cases and Comments. Boston Medical and Surgical Journal.

1804.
Rhigolene, a Petroleum Naphtha for producing Anaesthesia by Freezing.

Boston Medical and Snrgical Journal. 1866.
New and Successful Operation for Ununited Fractures, with Cases. Bos-
ton Medical and Surgical Journal. 1867.
Nitrous Oxide Gas for Snrgical Purposes in 1848. Boston Medical and

Surgical Journal. 1868.
The Mechanism of Dislocation and Fracture of the Hip. Boston. 1869.

842 HENRY JACOB BIGELOW.

Medical Education in America. Address before the Massachusetts Medi-
cal Society. 1871.

Death by Chloroform and alleged Death by Ether. Boston Medical and
Surgical Journal. 1872.

Alleged Death from Ether. Letter to the Editor of the British Medical
Journal. Boston Medical and Surgical Journal. 1873.

Turbinated Corpora Cavernosa. Boston Medical and Surgical Journal.
1875.

The True Neck of the Femur: its Structure and Pathology. Boston Med-
ical and Surgical Journal. 1875.

A History of the Discovery of Modern Ansesthesia: a Century of Ameri-
can Medicine. Philadelphia. 1876.

New Methods and Treatment of Extrophy of the Bladder and Erectile
Tumors. Boston INledical and Surgical Journal. 1876.

Lithotrity by a Single Operation. American Journal of Medical Sciences.
1878. Boston Medical and Surgical Journal. 1878.

Rapid Lithotrity with Evacuation. 1878.

Litholapaxy. New York Medical Record. 1879.

Litholapaxy. Boston Medical and Surgical Journal. 1879.

Litholapaxy. Letter to the London Lancet. Boston Medical and Sur-
gical Journal. 1879.

Litholapaxy : an Improved Evacuator. Boston Medical and Surgical
Journal. 1880.

The Code of Ethics adopted by the Massachusetts ^ledical Society. A
Minority Report. Boston Medical and Surgical Journal. 1880.

Remarks on Modern Lithotrity. Lancet. 1881.

Radical Cure of Umbilical Hernia. Boston Medical and Surgical Journal.
1882.

A Case of Disease of the Liver. 1882.

Lithotrity with Evacuation. 1882.

A Simplified Evacuator for Litholapaxy. Boston IMedical and Surgical
Journal. 1883.

A Radical Cure for Umbilical Hernia. Boston Medical and Surgical
Journal. 1889.

Fees in Hospitals. Boston Medical and Surgical Journal. 1889.

An Old Portrait of a Surgeon. Boston Medical and Surgical Journal.
1889.

The second publication on the list is entitled " Fragments of Medi-
cal Science and Art." Under this head is printed " An Address
delivered in 1846,"

The great aim' of this essay is to show the importance of the
imagination in science. The " Numerical Method " of Louis was at
that time looked up to, by the more ardent disciples of that admirable
observer and teacher, as the master-key which was to unlock all the

HENRY JACOB BIGELOW, 343

secrets of disease and its remedies. Observe all the facts in a case,
in a hundred or a thousand cases ; tabulate them, add, subtract, mul-
tiply, divide them, and the laws of pathology and therapeutics will
come out in your sums and quotients as inevitably as a clerk's balance
at the end of his account-book. Dr. Elisha Bartlett's " Philosophy
of Medical Science," published in 1844, presented the Numerical
Method in a form which might be thought to exclude the imaginative
element, and reduce the man of science to a mere statistician.

Dr. Bigelow's essay was a vindication of the true office and the
importance of hypothesis. To illustrate his argument, he appealed to
the history of great discoverers and inventors, of Copernicus, of Kep-
ler, of Newton. "I am aware," he says, "that this position, namely,
that hypothesis is essential to the discovery of scientific truth, is not
recognized by many philosophers, especially in medical science of the
present day. Bacon himself, feeling that unfounded theory, gratui-
tous assertion, had been a stumbling-block to all preceding science,
was led to attaching too exclusive value to facts. ' We must not
imagine or invent,' he says, ' but discover the acts and properties of
nature.'"

In the face of Bacon's proposition, in the presence of the chimpl-
ons of the statistical school of observers. Dr. Bigelow maintained
effectively and convincingly the true office of that higher faculty,
which, instead of counting columns of figures, sees, in virtue of its
special gift of insight, the hidden relations between a few facts re-
mote from one another to all appearance, but which, connected by
an hypothesis, are often verified by large observation, and become a
part of accepted knowledge or true science.

It was not so much the originality of the thesis maintained by Dr.
Bigelow as the reasonable and forcible method by which he expounded
and illustrated it, and the peculiar fitness of his choice of a subject at
that particular time. He knew when to strike, as well as how to strike.
One of the most distinguished of our Boston practitioners said to me
that he almost regretted Dr. Bigelow's having given so much time to
special practical points, instead of applying himself to the larger prob-
lems of medical philosophy. I would not go so far as that, remem-
bering how much he accomplished in the improvement of mechanical
surgery, and the amount of human suffering which his inventive ge-
nius has relieved ; but, after reading this essay, one may be pardoned
for regretting that so good a thinker and reasoner was willing to allow
his skilful handiwork to usurp so large a po^-tion of his time and
labor.

344 HENRY JACOB BIGELOW.

Had Dr. Bi'tfelow left no other record, the association of his name
with the great inventive discovery of artificial anaesthesia would
preserve his memory to the latest period of civilizatiou. On the
evening of November 2, 1846, he called at my house in Charles
Street with a j^aper which he proposed reading at the meeting of
tiie American Academy of Arts and Sciences, to be htld the next
day, and which he wished me to hear. He began by telling me
of the successful use of the inhalation of a gas or vapor which pro-
duced insensibility, during which a capital operation had been per-
formed at the Massachusetts General Hospital. He was in a state of
excitement as he spoke of the great discovery that the gravest oper-
ations could be performed without the patient's knowing anything
about it until it was all over. In a fortniglit, the news of this won-
deiful discover}-, he said, will be all over Europe. He then proceeded
to read to me the paper he had prepared, — the first formal presenta-
tion of the subject to the scientific world. The following is the
official report, copied from the records of the Academy: —

"November 3d, 1846.

" Dr. Henry J. Bigelow read a paper giving some account of the new
method of inhalation employed by Dr. Morton of this city to produce in-
sensibility to pain during the performance of operations by the dentist
and the surgeon."

No person took hold of Dr. Morton's discovery with such far-see-
ing, almost prophetic appreciation as the young surgeon who had been
but a few years in practice, and who threw all the energy and ardor
of his early manhood into his advocacy of the new and startling in-
novation which was destined to change the whole aspect of surgery.
It was not merely by his sagacious foresight that he recognized the
importance of this epoch-making novelty, but throughout its subse-
quent history, until its universal acceptance, he was the foremost
champion of the claims of artificial aniBsthesia. After the use of
chloroform was introduced Dr. Bigelow remained faithful to the origi-
nal anassthetic agent, and was always ready to battle in the cause of
ether as against chloroform, which, though more convenient, and in
many cases useful, is a more dangerous agent than the other. His
writings on this subject extend through a period of thirty years, from
1846 to 1876.

In the year 1850 Dr. Bigelow published a remarkable article on a
case which may be considered on the whole as the most extraordinary
in the anuals of surgical injury. This was the famous "crowbar

HENRY JACOB BIGELOW. 345

case," the account of which seemed to many incredihle, and its mech-
anism beyond exphmatiou. The story was briefly this. A man
was ramming down a charge of powder in a hole drilled in a rock,
when the charge exploded, and the tamping iron — a short rouud bar
— was driven up through the side of his face, out at the top of his
head, breaking upward through the top of his skull as if it had
been pie-crust, .^hooting up into the air, and falling at some distauce.
Dr. Bigelow accepted the story as true, and undertook to show how
the bai" could have found its way up and out through the bones of the
face and skull, traversing the brain, and cutting one of the optic nerves
on its way. He prepared a skull to illustrate the course taken by the
implement. The subject of this extraordinary accident lived many
years, but an opportunity was found to inspect the injured parts
after death, and Dr. Bigelow's explanation of the accident was fully

confirmed.

In the midst of his scientific researches Dr. Bigelow never forgot
the practical aim and end of the healing art. Pie cared quite as much
for '' common sense" in a medical man as he did for scientific acquire-
ments ; indeed, he rather undervalued pure science as compared with
practical skill. His lectures are eminently practical, and most of his
scientific researches tend to some important curative purpose. No
man knew better than he what were the needs, and what should be
the training, of the young practitioner who would make his way m
the world ; and his Lecture on " Science and Success " gives some
of the best results of his wise experience.

In 1.S69 he published his essay, " The Mechanism of Dislocation
and Fracture of the Hip." This subject had been long and diligently
studied by the great surgeons of tlie past, more especially by Sir
Astley Cooper. Dr. Bigelow threw new light upon the whole mat-
ter. I have requested Dr. Richard M. Hodges, who knew the his-
tory of Dr. Bigelow's researches more intimately than any other of
his pupils and assistants, to make a brief statement of the leading
points of his doctrine and practice in dislocations of the hip. The
following is his answer to my request : —

" Hip Dislocations. — A\iho\xgh Winslow and Weitbrecht had de-
scribed the two fasciculi of the ilio-femoral ligament, or ligament of
Bertin, Dr. Bigelow first drew attention to the great strength of the
anterior part of the capsule of the hipjoint, and defined with preci-
sion the two bands of the abovenamed ligament, diverging like the
branches of an inverted Y.

" Dr. Bisrelow showed that, so long as it remained unbroken in one

346 HENRY JACOB BIGELOW.

or both of its branches, the Y ligament dominated all the dislocations
of the hip joint with established features, and that it was the chief
obstacle to reduction ; the muscles playing only a subordinate and
occasional part in giving position to the limb, or in hindering the
reduction.

" Dr. Bigelow classified dislocations of the hip into regular and
irregular.

"The regular dislocations, seven in number (four of them being
new varieties), are those in which, one or both branches of the Y liga-
ment being unbroken, the head of the femur is thereby held near the
acetabulum, and their signs are constant.

" The irregular dislocations are those in which the Y ligament is
wholly ruptured, and they therefore offer no constant signs. The
head of the femur, being loosed from the acetabulum, is free to go
anywhere.

" In the regular dislocations, manipulation of the Y ligament will
alone effect reduction.

*' The principle of this manipulation is flexion, which is efficient
because it relaxes the Y ligament.

" The Y ligament being flexed, and therefore relaxed, the head of
the femur is drawn or forced into the desired direction by ' traction,'
which disengages it from behind the acetabulum and directs it toward
the socket, — or by ' rotation,' which winds the Y ligament around
the neck of the bone and so shortens it, thus compelling the head of
the femur, as it sweeps around the acetabulum, also to approach the
socket, into which it can be easily lifted.

" Dr. Bigelow converted random, ill devised, and fruitless move-
ments into accurately conceived, instructed, and well directed manip-
ulation,"

Growing out of his investigations of this subject was his original
study of the anatomical neck of the femur. By a series of parallel
sections through the head and neck of the bone, he demonstrated the
column or lamina of condensed bone in the midst of the cancellated
tissue forming a line of support rendered necessary by the obliquity
of the neck of the bone.

In 1878 Dr. Bigelow published his essay, " Lithotrity by a Single
Operation," of which Dr. Hodges speaks as follows : —

^^Rapid Lithotrily with Evacuatio7i at a slnrjh sitting ; or Lithola-
faxy. — The normal urethra having been shown to admit instruments
of greater size than surgeons had previously supposed possible, Dr.
Bigelow constructed a lithotrite, improved in many of its details,

HENRY JACOB BIGELOW. 347

(especially by devices which prevent the blades from clogging or be-
coming impacted with crushed material,) of a size much larger than
had before been used. This permitted the attack of calculi exceeding
in dimensions the limits previously thought allowable by crushing
alone, i. e. without evacuation.

" Dr. Bigelow also constructed thin silver tubes, easy to be intro-
duced, notwithstanding their large size (27-31 Charriere), through
which evacuation of the crushed stone was made practicable by means
of an elastic exhausting bulb of sufficient suction power to draw out
the fragments previously comminuted to a size enabling them to
enter and pass through the tube, — pulverization being no longer
essential.

" Dr. Bigelow established the fact that with these instruments a
sitting — two minutes having been, up to that time, assigned by Sir
Henry Thompson as the proper average duration — could be pro-
longed, with the aid of anaesthesia, one to two hours, harmlessly for
the patient and without detriment to the bladder. ' Lithotrity with
a single sitting ' has been shown to have a mortality less than that
of ' Lithotrity with many sittings,' and it has entirely superseded
the latter.

" The operation of Litholapaxy, at first supposed applicable only to
adults, has been within the last few years extended in its use to chil-
dren from two years of age upwards, with great success. They have
never been supposed to come within the scope of old-fashioned litho-
trity. This practice, adopted originally in India (Lihore), has latterly
been introduced in England and America.

" Dr. Bigelow's invention may justly be said to have acquired a
world-wide reputation."

I add a few words to this description by Dr. Hodges. He was led
to think that a principal source of failure in that operation was the irri-
tating effect of the fragments of stone allowed to remain in the bladder,
which left it inflamed and sensitive, not in condition to be the subject
of a second or third operation. If the bladder could be completely
cleared at one sitting, this danger could be avoided. To effect this
object, he designed new instruments, or modified such as were in use,
so as to make them serve his purpose. He spared no pains in per-
fecting his apparatus. It is not to be supposed that his surgical inno-
vations were at once accepted without question or opposition. The
end of it all was, that his principal rival in the treatment of calculus,
Sir Henry Thompson, became a convert to Dr. Bigelow's mode of
dealing with stone in the bladder, and that this new m'^thod of opera-

348 HENRY JACOB BIGELOW.

tion is generally recognized as one of the great improvements of
modern surgery. I myself bad the opportunity of observing some
of his experiments, and well remember the patient and persevering
labor they involved. I recollect, more especially, the pains he took
in getting plaster casts of the bladder and the urethra, and I learn
from others that he bestowed the same care upon the instruments he
contrived or adapted for the rapid removal of a calculus, by the me-
thod to which he gave the name of Litholapaxy.

Among Dr. Bigelow's other professional labors, I may mention his
suggestion of a new refrigerant for producing local anfcsthesia. This
was brought forward in an article published in the Boston Medical and
Surgical Journal, in 1866, under the title, " Rhigolene, a Petroleum
iVaphtha for producing Anaesthesia by Freezing." ^

A new anatomical observation was published by Dr. Bigelow in the
same journal, in the year 1875, "• Turbinated Corpora Cavernosa."
The anatomical expert will recognize at once the analogy hinted at in
this designation. The suddenness with which the air passage through
the nostrils will become obstructed, and the equal suddenness with
which it will be cleared, without the removal of any secretion, might
well suggest the idea that some kind of erectile tissue was concerned
in this familiar phenomenon. Dr. Bigelow examined the mucous
membrane, and detected a spongy tissue with large cells, capable of
being rapidly filled with l)lood and as rapidly emptied, — a structure
resembling that of the corpora cavernosa, as the name he gave it im-
})lies. This is one of the ver}'^ few additions to human descriptive
anatomy which have been made in this country.

Dr. Bigelow was not a collector of books, nor a great reader. He
opened a book as he would open a jackknife, to use it for some
special purpose, which having accomplished, he shut it up and had
done with it. I may be allowed to quote my own words, as they
stand in the report of the memorial meeting held shortly after his
decease by the Boston Society for Medical Improvement : —

" He read men and women as great scholars read books. Pie took
life at first hand, and not filtered tlirough alphabets. He was not
ashamed of his want of erudition, and would ask questions on matters
with which he was unacquainted with the simplicity of a child. l>ut
he would get what he wanted out of a book as dexterously, as neatly,
as quickly, as a rodent will get the meat of a nut out of its shell. In
the address before spoken of, on the use of imagination in science, he
handled his rapidly acquired knowledge of the great authors he cited
so like an adept in book lore that one might have thought he was

HENRY JACOB BIGELOW. 349

born iQ an alcove and cradled on a book-shelf. He got what he
wanted out of his authority, and the next day the volume he had evis-
cerated would be kicking about his floor, in the midst of the pam-
phlets, instruments, and all sorts of learned litter, which half covered
his carpet. This power of finding what he wanted in the midst of
rubbish he did not want, was hereditary. I remember Dr. James
Jackson's saying to me, that, if there was a grain of wheat in a bushel
of chaff. Dr. Jacob Bigelow would lind it quicker than any man he
ever knew."

Though Dr. Bigelow was not as much given to general reading as
many less occupied professional men, it is not to be supposed that his
active mind could fail to find subjects enough to interest it when not
absorbed in some important investigation. He had many tastes and
fancies which furnished him abundance of pleasant work, and called
forth all his enthusiasm, each special pursuit in its turn. To this one
object, whatever it might be, he gave himself enthusiastically for the
time. When he had mastered all its details, when he had got at
all its secrets, he left it for some new and inviting subject.

At one time he undertook the keeping and raising of fancy pig-
eons. For this purpose he established a columbarium at the top of his
house in Chauncy Place, and showed his fantails and pouters, and
other curious varieties, with great satisfaction, until he had learned
their ways and become familiarly acquainted with their various graces
and accomplishments. At another time his visitor would be startled
by a most unceremonious address from a mino bird, to which he had
taught certain phrases which could not fail to arrest the attention of
his visitor. Another of his pets was a little bird which used to run
up his sleeve in the most uncanny way. Between these two familiar
spirits, he might well have been hanged as a wizard in the days of
witchcraft. At another time he amused himself with the study of
the ways of ants, placing the sand for their dwelling between two
plates of glass, so that their operations could be observed. Again,
he found his recreation in the royal handicraft of the locksmith, and
studied the intricate contrivances of Bramah and Hobbs as he has
studied the arrangements of the hip joint. With this fondness for
animal life it is not strange that he held in great aversion the too
frequent abuse of vivisection. I have often heard him express himself
very strongly on this subject. I think his longest and strongest fancy
was for paintings. He did not care to refer to the fact that he was
color-blind, so far as the difference between red and green was con-
cerned. When he was a boy he could not distinguish between the

350 HENRY JACOB BIGELOW.

color of cherries and that of the leaves of the tree. Still, he had a
passion for a picture, and spoke with enthusiasm of the color of some
that pleased hiiu. A bright patch on an old canvas attracted him
in a moment ; he would wet his finger and rub off the dust as eagerly
as a gold-hunter explores a pebble with shining yellow particles
scattered through it. He bought a good many pictures, and it was
generally for their color, rather than for any other excellences, so
fur as my observation has gone. Another of his hobbies, if I may
call them so, was the study of agates. He made a large collection of
them, and examined some points of their internal formation with great
interest.

Dr. Bigelow was not in the habit of speaking of his health, but
he suffered at various times from symptoms of different kinds. The
earlier pulmonary symptoms, which have been referred to, do not
appear to have troubled him after the period of early manhood.

A few years before his death he was thrown from a vehicle, and
received a blow on the head, which was followed by what seemed to
be an inflammation of some of the membranes of the brain, leading to
what he thought and what proved to be some thickening of the dura
mater. His fatal illness seemed to be entirely disconnected with the
injury referred to. Occasional passages of gall stones, inflammation of
the bile ducts extending to the liver, and producing abscesses, with
other marks of internal inflammation, inability to take food without
extreme suffeiing, ended in gradual faiUire of bodily strength, the
mind remaining bright and clear to very near the close of life. It
was noted, in examination of the brain, that its convolutions presented
an unusual complexity, suggesting a greater amount of vesicular
matter than is common.

Dr. Bigelow wrote upon vai'ious important subjects of a more gen-
eral nature. In 1871 he delivered an address upon medical education
in America, before the Massachusetts Medical Society. In 1880 he
wrote a minority rejiort upon the code of ethics adopted by the Medi-
cal Society ; and in 1889, an article upon fees in hospitals, in which he
took strong ground against certain practices alleged to have grown
up in some of these institutions. The last paper on the list of his
works is entitled, "An Old Portrait of a Surgeon." A painting was
presented many years ago to the Society for Medical Improvement,
supposed to be a portrait of the great surgeon, Ambroise Fare. The
truth of this supposition had been questioned, and remained undecided
for thirty or forty years, when Dr. Bigelow thou<;]it it was time to
settle it authoritatively. For this purpose he instituted the most search-

CHARLES OTIS BOUTKLLE. 351

ing inquiry ; had photographs taken of numerous portraits bearing on
the (juestion ; carried on a correspondence with experts in Euroi)e ;
and linally established beyond doubt tlie fact that the portrait was
not of Ambroise Pare, but of another practitioner of a certain reputa-
tion, but by no means so great a name as the illustrious surgeon's of
whom it had been thought to be a likeness.

Dr. Bigelow was, unquestionably, a man of true genius. Sagacity
in divining the truth ; the power of continuous, patient, and searching
investigation ; inexorable determination to have the truth, if nature
could be forced to yield it, characterized his powerful intelligence.
The record of his printed publications is not a very long one, but it is
weighty with original thought and practical discovery. He inherited
a distinguished name, and his labors have rendered it memorable «nd
illustrious, — one of the brightest in the annals of American surgery, —
not to claim for it a still higher place in the history of the healing art.

Dr. Bigelow was married in 1847 to Susan, daughter of the Hon.
William Sturgis. She died on June 9, 1853. One son, William
Sturgis Bigelow, survives his parents.

CHARLES OTIS BOUTELLE.

Charles Otis Boutelle was born in Lexington, Massachusetts,
August 4, 1813. His grandfather was an officer who served hon-
orably throughout the Kevolutionary War. His father, a skilful
physician and a man of brave and earnest temperament, was a sur-
geon in the Navy during the war of 1812. His mother, a daughter
of General Nathaniel Goodwin, of Plymouth, who served also dur-
ing that war, was a woman loved and revered by all who knew her.
She lived to nearly the age of one hundred, and her son never
ceased to mourn her loss.

With such ancestry, many features of Mr. Boutelle's character
can be traced to their source. Having while yet at an early age
lost his father, he w^as educated by his uncle, the Peverend Ezra
Shaw Goodwin, of Sandwich, Massachusetts, and received from
him a thorough training in both the classics and mathematics.
It soon became necessary for him to earn his own living; so he
taught school, studied surveying, and one day, having heard that
a friend who owned a Avork on that subject was willing to lend it
to him, he walked twenty miles to get it. His skill in practical
surveying soon became known, and a place was given to him on the
survey of his native State by its director, Simeon Borden.

352 CHARLES OTIS BOUTELLE.

Having served creditably as Mr. Borden's chief assistant, he was
appointed by Alexander Dallas Bache, Superintendent of the U. S.
Coast Survey, to a position upon that work, in January, 1844. His
service was at first in the office, but his active temperament and
robust physique demanded less sedentary occupation, and his spe-
cial capabilities for the field were quickly recognized by his distin-
guished chief. His advancement was rapid. In 1846 he was made
an assistant in the Survey, and from that time forward gained
steadily in standing on the work, being intrusted with the charge
of important operations, which he conducted with his accustomed
energy, and with the professional skill and fertility of resource
always at his command.

For some years he carried on the reconnoissance for the primary
triangulation upon the coast of Maine. He made the reconnois-
sance and selection of sites for three primary base-lines, and had
personal charge of the measurement of a primary base-line (the
Atlanta base) in Georgia. This measurement was three times
repeated as a test of accuracy, the line being measured twice in
winter and once in summer, with an accordance of results so
close that the greatest divergence did not exceed a millionth part
of the whole length of nearly six miles. He conducted the pri-
mary triangulation which was carried from the Atlanta base north-
ward and northwestward along the Blue Ridge, to connect with the
primary triangulation which was advancing southward and south-
westward from the Kent Island base, and had charge of the sur-
veys upon the coasts of South Carolina and Georgia.

During this period the bent of his mind was shown by the im-
provements he introduced into the methods and processes of the
work; among these may be mentioned the form of preliminary base
apparatus described in the Coast Survey Report for 1855; his form
of tripod and scaffold observing signal, 1855; his experiments with
lights for geodetic night signals, carried on for several years, and
brought to a successful termination in 1880 by the adoption of the
magnesium lights and the student-lamp reflectors.

In 1884 the charge of the Coast and Geodetic Survey office was
assigned to him, and after his relief from that duty he was placed
in immediate supervision of geodetic operations in the States which
had organized their own geological and topographical surveys.

For a number of years he was a member of the board of commis-
sioners for tlie improvement of the harbor of Norfolk.

On February 16, 1884, soon after taking up his residence in

CHARLES OTIS BOUTELLE. 353

this city, he was elected a menibor of the Philosophical Society
of Washington.

No notice of Mr. Boutelle's life would be complete that should
omit reference to the important services which he rendered to his
country at a critical period of its history. In common with the
great majority of his brother officers assigned to duty with the
military and naval forces, he participated in the hardships and
dangers of the civil war. Soon after the outbreak of hostilities
he was assigned to the command of the steamer Vixen and schooner
Arago, as hydrographic officer of the South Atlantic Squadron,
serving under Admirals Dupont and Dahlgren, and Commodore
Lanman, U. S. N. This duty lasted throughout the, war, and it
devolved upon him the responsibility for the safetj'' of naviga-
tion of the squadron along its entire cruising ground. With
what patriotic devotion and professional ability this service was
rendered, the records of the civil war amply attest.

Admiral Dupont, in his report to the Navy Department of the
capture of Port Royal, refers to the fact that all aids to navigation
had been removed by order of the Confederate authorities, and ac-
knowledges the able assistance of Captain Boutelle in sounding
out and buoying the channel, and thus enabling the squadron to
advance to the attack.

General W. T. Sherman, U. S. Army, commanding the land
expeditionary force, concludes a report, dated November 8, 1861,
as follows: ''It is my duty to report the valuable services of Mr,
Boutelle, assistant in the Coast Survey. . . . His services are in-
valuable to the army as well as to the navy, and I earnestly recom-
mend that important notice be taken of this very able and scientific
officer by the War Department."

Personally, Captain Boutelle (as he was known to his friends
after the civil war) was a man of varied reading and a most reten-
tive memory, genial and witty in conversation, of uniform kindness
of heart, and of a generous and hospitable nature, always assuming
that others were guided by motives as unselfish as his own.

He combated manfully the advances of age and the inroads of
disease, and it was not until the approach of his seventy-eighth
year that, yielding to the solicitations of his family and friends, he
sought relief from active duty. He died on the 22d of June, 1890,
at the home of his son. Dr. Boutelle, in Hampton, Virginia.

VOL. XXVI. (n. 8. xviii.) 23

354 ALFRED HOSMER.

ALFRED HOSMER.

Alfred Hosmer, of Watertown, Massachusetts, a Fellow of the
Academy from 1879, was born at Newton, September 11, 1821, but
removed with his mother to Walpole, New Hampshire, when nine
years of age. Notwithstanding the limited opportunities for educa-
tion afforded by a small country town, he qualified himself for ad-
mission to Harvard College, where he graduated with honor in
1853, and received his degree of M. D. in 1856. He spent nearly
a year of study in Europe before beginning the practice of his
profession at Watertown, where he soon gained reputation as a
worthy successor -of his uncle, Dr. Hiram Hosmer, a physician
and surgeon of great rejKite. From this time until his last illness
he enjoyed in an exceptionally high degree the confidence of the
community in his own and the neighboring towns, and won the re-
spect of his confreres by his untiring energy, accurate observation,
sound judgment, fertility of resources, and unswerving fidelity to
every duty.

Soon after his establishment at Watertown he married Helen
Augusta, daughter of Josiah Stickney, Esq., and she, a daughter,
and a son survive him.

Dr. Hosmer's published papers were marked by originality and
independence, by alertness in discerning the presence of unusual
features in disease, and readiness and good judgment in the adap-
tation of suitable means to the existing circumstances.

Dr. Hosmer was President of the Middlesex South District
Medical Society; President of the Boston Obstetrical Society; and
for many years a Councillor of the Massachusetts Medical Society,
and was its President in 1882 and 1883.

Exceptionally brilliant as was Dr. Hosmer's professional career,
this by no means included all of his public service. His vigorous
activity and faithfulness were conspicuous iu the affairs of citizen-
ship, as well as in the duties of his chosen vocation. Among the
offices of honor and trust to which he was for many years re-elected
b}' his fellow-townsmen, he was a member of the School Commit-
tee, Trustee of the Public Library, Trustee and President of the
Savings Bank, and President of the Watertown Historical Society.
He was for many years surgeon of the United States Arsenal at
Watertown.

Through the influence of the Massachusetts Medical Society, the
Legislature of the State, during the administration of Governor

GEORGE BANCROFT. 355

Rice, abolished the corrupt and inefficient coroners' system, and
substituted therefor a carefully selected corps of skilled medical
examiners, appointed by the Governor for a long term, and, hap-
pily, continued in office in most instances by reappointment;
so that the State, the judicial authorities, and the people retain
the advantage of experience as well as of ability in these chosen
experts. No event in our time has contributed so largely to the
honor of medicine and to the safety of the community as the crea-
tion of this body of skilled and trained officials, which has worked
so intelligently and faithfully for the promotion of justice in the
protection of the innocent and the disclosure of guilt. Great
credit was due the Governor for his just appreciation of the im-
portance of the proposed innovation, and for his judicious selection
of the appointees, on which its success would so largely dej^end.
Dr. Hosmer was appointed Medical Examiner for the District of
Middlesex County, in which he resided ; and upon the organiza-
tion of the Massachusetts Medical Legal Society, composed of the
seventy or more Examiners of the Districts of the Commonwealth,
he was elected its first President, and by his zeal and executive
ability greatly promoted the triumphant issue of this eventful
experiment. Thanks to the high character and talents of Dr.
Hosmer and his associate examiners, Legal Medicine has been re-
created in Massachusetts, and the new system has been adopted
in other States, as a long desired efficient auxiliary of public
justice.

On the 29th of December, 1888, in returning from professional
visits. Dr. Hosmer had an attack of cerebral hemorrhage. From
this he largely recovered, although he continued to have some dis-
ability to find the desired words when conversing; but on May 14,
1891, after a few days' illness, he suddenly expired from rupture
of the thoracic aorta, which, as well as the entire arterial system,
was found at the autopsy to be extensively degenerated.

ASSOCIATE FELLOWS.

GEORGE BANCROFT.

George Bancroft was born at Worcester, October 3, 1800.
He was the son of Aaron and Lucretia (Chandler) Bancroft. His
father was the leading clergyman of the Unitarian denomination

356 GEORGE BANCROFT.

in Central Massachusetts; was honored with the degree of D.D. by
his Alma Mater, Harvard College; was the author of a standard
Life of Washington which has been republished in England and
has also been recently reprinted in this country in popular form;
was Vice-President of the American Antiquarian Society for many
years; was President of the American Unitarian Association from
its organization in 1825 to 1836; and was a Fellow of the American
Academy of Arts and Sciences. His mother was a daughter of
Judge John Chandler, and was a recognized leader among the wo-
men of the little town in which her lot was cast. Her time during
married life was engrossed in the care of a large family of children,
who were reared in a manner suited to their social position, with
no other income ordinarily at command than her husband's salary
as a country clergyman, eked out by such increase as could be
obtained from teaching the children of parishioners. The lesson
of George Bancroft's boyhood was that he must rely for his future
support upon himself alone.

A sister of Mr. Bancroft married John Davis, the greater part of
whose life was spent in public service, as Governor of Massachu-
setts, E,epresentative in Congress, or United States Senator. A
nephew, J. C. Bancroft Davis, succeeded Mr. Bancroft at the Ger-
man Embassy. The present Chief Justice of the United States
finds a common ancestor with Mr. Bancoft, one generation back
of the historian's father.

It was Dr. Bancroft's earnest desire that his son should adopt
his own profession ; and, in preparation for that career, George was
sent to Harvard College, where he graduated in 1817, second in
his class. The opportunity was then offered him to complete his
studies in Europe at the expense of the College. Of this he gladly
availed himself, and the next five years of his life were spent in
study and travel in Europe. He studied at Gottingen and Berlin,
taking the degree of Ph.D. at Gottingen before he was twenty
years of age.

This period of his life was rich in friendships of distinguished
men belonging to a generation now passed away. It was fraught
with benefits to him in his future career. The familiarity with
the German language which he then gained was of incalculable ser-
vice to him, both in his diplomatic and in his literary labors. The
philosophic cast of his studies somewhat influenced his methods of
thought, and found sufficient expression in his writings for his
critics to charge him with mj^sticism. Yet their tendency was to

GEORGE BANCROFT. 357

raise his strong, nervous personality above the plane of prejudice,
and to help him to measure events upon a just standard. The
estimate in which he was then held may he judged by the language
in which Humboldt introduced him to Pictet. Bancroft was de-
scribed by Humboldt as a "young American, who has made an
excellent study of philosophy and philosophic history in Germany."
His reception at this time by prominent men in the world of Euro-
pean letters doubtless inspired in him the confidence which was
required when, a few j^ears later, he concluded to undertake what
was to prove the work of his life.

The privilege of study in Europe had been afforded him by
Harvard College, with a view to his thus preparing for a professor-
ship in that College when there should be a vacancy. So far as
his immediate prospects were concerned, he had not on his return
to the United States, in 1822, ultimately abandoned his intention
of adopting the profession of the Christian ministry; and he did
in fact preach in his father's pulpit, and from time to time else-
where, for a brief period. He was appointed Tutor of Greek at
Harvard College in 1822, and filled that office for a year.

From 1823 to 1830, his time was devoted to teaching at Round
Hill School, Northampton. In 1827 he married Sarah H. Dwight.
She died in 1837. By this marriage he had two sons and one
daughter. The sons survive their father. In 1838 he married
Mrs. Elizabeth (Davis) Bliss, who died in 1886. She bore him
one daughter ,who died while yet a child.

It was while connected with Round Hill School that he launched
his first literary venture, a volume of poems, published in 1823.
His pen during this period of his life was also busily at work upon
translations from the German, and on articles for the prominent
reviews of the day. An oration delivered on the 4th of July, 1826,
and published the same year, in which he announced the democratic
nature of his political views, and a pamphlet on " The Bank of the
United States," published anonymously in 1831, indicate the in-
terest he took in the political questions of the day. In 1830 an
opening was afforded him through which he might have entered
political life if he had cared to do so. He was elected to the legis-
lature, but declined to take the seat. Next year he was tendered
the nomination to the Senate, but, notwithstanding the fact that
his election would have been certain, he declined the honor.

In 1834 he published the first volume of his History. *' I have
formed the design," he says in his Preface, " of writing a History

358 GEORGE BANCROFT.

of the United States from the discovery of the American Continent
to the present time." It will be seen that he did not attempt to
carry out the plan as thus promulgated. History requires the aid
of intervening time as a guaranty that the writer shall escape the
prejudices and passions provoked by current events. It was im-
possible that Mr. Bancroft should treat with uniform impartiality
of men and affairs belonging strictly to the past, and of events
which had been shaped by the generation then controlling the
destiny of his country. His History, a% published, concludes
with the establishment of a constitutional government in the
United States, and to its production he devoted the remainder of
his life, publishing from time to time the several volumes of the
series, as intervals of leisure in his i^olitical and diplomatic labors
permitted.

In 1835 he removed to Springfield, and next year ran for Con-
gress in that district on the Democratic ticket, but was defeated.
From 1838 to 1841 he was Collector of the Port of Boston. In
1844 he was Democratic candidate for Governor of Massachusetts,
but failed of election. In 1845 he was Secretary of the Navy under
James K. Polk, and while he held that position he founded the Na-
val School at Annapolis. The need of a training school for naval
officers similar to that at West Point for the Army had long been
felt. The usual sectional jealousies and political prejudices which
impede useful legislation at Washington were to be apprehended if
a direct appeal for the establishment of such an institution were
made. An opportunity offered to secure the post at Fort Severn,
Annapolis, without legislation. Existing laws permitted the
stationing of certain officers and instructors at this post. It was
also possible to detach midshipmen from vessels as they arrived in
this country and order them to Annapolis. All of this was accom-
plished. Orders for the establishment of the school were issued,
and a scheme was promulgated which, in 1846, was published at
Washington, under title of "Plan and Regulations of Naval School
at Annapolis." The school was formally opened October 10, 1845.
On the 1st of January, 1846, its membership was composed of
forty-nine midshipmen and seven acting midshipmen. The Navy
thus became indebted to Secretary Bancroft for a training school.

While Secretary of the Navy, he issued the order to take posses-
sion of California in the event of a war between the United States
and Mexico, and while acting as Secretary of War, during the
temporary absence of Marcy, Bancroft issued the order to Greneral
Taylor to march into Texas. ^

GEORGE BANCROFT. 359

The stirring events in which Bancroft as a member of Polk's
Cabinet participated were of momentous consequence to liis coun-
try. Every believer in what was then termed the ''manifest
destiny " of the country was compelled to admit that it was true
statesmanship to seize the opportunity for gaining the magnificent
territory offered for our control by the republic of Texas. Every
antislavery man felt that Texas would not have been annexed,
except that it would add another slaveholding State. Whether to
sacrifice an opportunity which might not occur again, or to take
the chances of a future adjustment of the powers of the two sections,
was the question which antislavery men were forced to solve. To
most Northerners the overwhelming antislavery sentiment in the
North made the solution of the problem easy; but to an antislavery
Democrat holding an office of power, the position was a difficult
one. Mr. Bancroft was a Democrat, an ardent lover of his coun-
try, and an antislavery man. Bev. Dr. Hale is authority for the
statement that, while his confirmation as Secretary of the Navy was
before the Senate, he was approached by a Senator on the subject
of slavery, and in reply to questions as to his views he said that
he was an antislavery man, and if he were to go tlirough the
Senate he must go erect and not on his knees. The annexation
of Texas carried with it a possibility of war with Mexico. The
general opinion at that time was that the days of Mexico's occu-
pation of Upper California were numbered, and it was feared that
England would take possession whenever opportunity offered. That
it would be in the interest of civilization if the United States
should by any chance secure this territory, there was no doubt
in any person's mind. It was under these circumstances that
Bancroft, to quote the language of Von Holtz, "never wearied
of impressing this one precept upon Sloat, Stockton, and Biddle.
See to it that as soon as practicable Upper California at least be in
our hands, in order that we may retain it if peace is concluded upon
the basis of uti possidetis.^ ^ The same author measures tlie situ-
ation in the United States at that time in the following language:
"After its territory had once been extended thus far to the West,
it w^as a proper, nay, an inevitable thought, that its banner must
overshadow the entire continent, in its whole extent from ocean to
ocean."

In 1846, Bancroft was transferred from the Cabinet to the Court
of St. James. During the three years that he remained in London
as Minister Plenipotentiary, he negotiated a postal treaty between

360 GEORGE BANCROFT.

Great Britain and the United States. He took great interest in
the debates in Parliament on the Navigation Laws, and exerted his
influence to secure a moditication of their rigor.

When Bancroft went to London, he found that his History, three
volumes of which had then been published, had made him a famous
man. He was cordially welcomed in society, and was admitted to
the friendship of the prominent literary men and statesmen of the
day. '^1 met him everywhere," says Robert C. Winthrop, ''and
witnessed the high estimation in which he was held by literary
men like Rogers, and Hal lam, and Alison, and Milman, and Lord
Mahon, and by statesmen like Peel, Palmerston, and Russell."

Under such circumstances the opportunity to consult original
documents bearing upon the topics of his History were exceptional,
and he availed himself of it to amass an amount of material such
as no man working in the same field had before that time had at
his command. While Minister to England, he received from the
University of Oxford the degree of Doctor of Civil Law.

In 1849 he returned to the United States, and took up his abode
in New York, where he devoted himself to historical labors. His
life was one of great regularity. Each day had its aj)propriate
hours set apart for labor, for exercise and recreation, and for social
duties. He faithfully followed the allotted programme, thus secur-
ing the best results for mind and body. The fourth volume of the
history — the first of the Revolution — was published in 1852, the
fifth in 1853, the sixth in 1854, the seventh in 1858, the eighth
in 1860, and the ninth in 1866. Besides these, he published in
1855 a volume of ''Literary and Historical Miscellanies." In
February, 1866, he delivered before Congress an oration in memory
of Abraham Lincoln.

In Ma^', 1867, he was appointed Minister to Prussia. He re-
mained at Berlin until 1874, representing our government at this
Court during its successive changes from the kingdom of Prussia
to the North German Confederation, and finally to the German
Empire. While Minister to the North German Confederation, he
concluded naturalization treaties which included in their scope
Prussia, Baden, Bavaria, and Wurtemberg. These treaties were
the first to recognize the right of a citizen to change his alle-
giance, and could only have been secured by a man of exceptional
influence. While at Berlin, he presented the American case in
the arbitration between Great Britain and the United States
concerning the title to the island of San Juan. This act was

GEORGE BANCROFT. 3G1

performed by him under a separate appointment. The decision
of the Emperor sustained tlie title of the United States, and finally
disposed of the vexed question. In 18G8, he received from the
University of Bonn the degree of Doctor of Laws, and in 1870, on
the fiftieth anniversary of his Doctorate of Philosophy at Got-
tingen, that University conferred upon him an honorary I'h. D.
The circumstances connected with this event attracted much atten-
tion. Congratulations were showered upon the head of Mr. Ban-
croft in person, by letter, and by telegraph, from crowned heads,
from learned societies, from generals, and from men of letters.
Bismarck sent his from the field. Americans when they read the
story of these events realized that rare honors had been heaped
upon the head of their Minister.

In 1874 he was at his own request recalled. On his return
to this country he divided his life between Washington and
Newport, making the former place his winter, and the latter his
summer home. Almost to the last day of his life, he continued
his methodical habits, parcelling out his labors and his pleasures,
allotting specific periods of the day to each, and rigidly adhering
to his plan. Those who were present at the opening session of
the American Historical Association in Washington, in 1886, will
remember an incident which will illustrate the value he attached
to punctuality. At the appointed hour he was ready to open the
meeting, but owing to circumstances over which he had no control
was unable to do so. As soon as he was able, he called the audience
\o order, saying, "I pray you all to bear witness that I was here,
prepared to open this meeting, at precisely ten o'clock."

Volume X. of his History, which practically completed the work,
was published in 1874. The Centenary Edition, revised and con-
densed into five volumes, was published in 1876. An edition was
published in London in 1862, and a German edition in Leipzig in
1875. Perhaps these foreign editions did not have his personal
supervision. Volumes XI. and XII. were separately published
under the title of "History of the Formation of the Constitution
of the United States," in 1882. A revised edition, which included
the history of the Constitution, and which was termed by him
the ''final revision," was published in six volumes in 1883-85.
The last volume which he published was a "Life of Martin Van
Buren." The manuscript for this sketch had been prepared many
years before. The last work actually given by him to the public
Avas a trenchant criticism of the Supreme Court in the Legal Tender

362 GEORGE BANCROFT.

Case. This was published in 1886, under the title of ''A Plea
for the Constitution of the United States of America, wounded
in the House of its Guardians."

Allibone gives a list of his miscellaneous publications. The
Annual Report of the American Historical Association for 1889
gives a list of his historical works, including papers communicated
to historical societies. Still another list of his publications will
be found in Appleton's Cyclopaedia of American Biography.

The social position which Mr. Bancroft held when lie returned
from Germany to this country was enviable. His friendships
comprehended the great men of two hemispheres for half a century.
Learned societies at home and abroad had elected him to active and
honorary membership. He bore honorary degrees from American,
English, and German universities. A partial list of these socie-
ties and degrees occupies nearly half a column in the Quinquen-
nial Catalogue of Harvard University. The Senate of the United
States extended to him the unprecedented honor of free access to
the iloor of their chamber. His society was eagerly sought both
at Washington and at Newport, and it required all the restraints
of his methodical habits to preserve strength for the work still
before him. Towards the end of his life the anniversaries of
his birthday Avere made much of by friends. Flowers, messages,
and congratulations were showered upon him. October 3, 1887,
Browning cabled him as follows : —

" Bancroft, the message-bearing wire
Which flashes my all-hail to-day
Moves slower than the heart's desire
That what hand pens tongue's self might say."

He took great pleasure in the cultivation of roses; and the *' George
Bancroft " in the catalogues of rosarians bears witness to the
recognition of this taste on the part of horticulturists. He was
fond, especially during the latter i)art of his life, of riding; and
visitors to Washington or Newport felt that they had missed one
of the sights of the place if they had failed to see the slight,
erect form of the historian, crowned with his snow-white hair and
beard, as he took his daily exercise on horseback. He was grateful
for the opportunities afforded him at Exeter, where he went to
school, and founded there a scholai"shi]i. At Harvard he founded
a fellowship which he named, after his old teacher, the "John
Thornton Kirkland Fellowship." In honor of his parents, he

GEORGE BANCROFT. 363

placed in possession of the city of Worcester a fund for the main-
tenance of a scholarship which he called the "Aaron and Lucretia
Bancroft Scholarship."

He died, January 17, 1891, at his home in Washington. His
friends were not unprepared for the event, as he had been percepti-
bly failing for a long time. The Emperor of Germany caused
flowers to be laid upon the casket which contained his remains.
The funeral services were held at Washington; but the body of the
historian was interred in the same cemetery in Worcester which
holds the remains of his father and mother.

The position of Bancroft's History as the standard history of the
United States has left for the critics to discuss only the question
how long the work will be able to maintain this position. Alli-
bone has grouped in his columns the opinion of a number of com-
petent writers. A few extracts from these opinions expressed
during the progress of the work, a brief analysis of some of the
objections which have been made to the History itself, and a few
quotations from later writers, in which they state their estimate of
the work, will help to determine this question. Heeren, under
whom Bancroft studied and who was his personal friend, reviewing
the first three volumes, says, "We know few modern historic
works in which the author has reached so high an elevation at
once as an historical inquirer and an historical writer." Edward
Everett, reviewing the first volume, says, "As far as it goes, it
does such justice to its noble subject as to supersede the necessity
of any future work of the same kind." Prescott, reviewing the
third volume, says, "His Colonial History establishes his title to
a place among the great historical writers of the age." Dr. Gris-
wold in his " Prose Writers of America," treating of Volumes I.,
II., and III., thinks that "he becomes insensibly the advocate of
the cause of freedom, which invalidates his testimony." The
Edinburgh Review, on the other hand, says, "The real liberal-
ity, the general fairness, the labor and conscientious research it
evinces, deserves, and we are assured will receive, his [the English
reader's] warmest approbation." The Westminster Review pre-
dicts, "with confidence, that his work will be reckoned among the
genuine masterpieces of historical genius." Lecky, in his "Eng-
land in the Eighteenth Century," accuses him of violent par-
tisanship, and charges that it greatly impairs his "very learned
History." The foregoing will illustrate the reception of the
History by literary men during its progress. It is difficult to

364 GEORGE BANCROFT.

conceive of tributes more gratifying to an author. If there had
been no voice in England to raise the charge of partisanshij)
against this ardent American while engaged in depicting the
preliminary struggles of the Colonies with the mother country,
it would have been because he had failed to accomplish what he
had undertaken. If the English people, as a whole, had not been
able to appreciate Bancroft's labor and conscientious research, his
fairness of purpose, and the real liberality beneath his sharp, inci-
sive criticism, it could only have been because they had become
less tolerant than we know them to be.

In the composition of the first three volumes, which are devoted
to Colonial History, Bancroft relied upon resources such as are
at command of ordinary writers. ''I have been most liberally
aided," he says in his Preface, **by the directors of our chief
public libraries; especially the library at Cambridge, on American
history the richest in the world, has been opened to me as freely
as if it had been my own." The period covered topics concerning
which an expression of opinion did not necessarily raise a contro-
versy. Believers in the historic value of the Sagas did not feel it
their duty to attack one who accepted the visits of the Norsemen
to this coast as a natural probability, because he thought the
Sagas themselves mythological in form and obscure in meaning.
The exercise of discriminating judgment as to the voyages of the
period, while it might arouse criticism as to the accuracy of the
adoption of this or that narrative, did not provoke acrimonious
discussion. The wide difference between the reception of the ear-
lier volumes, and of the series which bore upon Revolutionary
topics, calls attention to the scheme of the History, and empha-
sizes the manner in which that scheme was developed. The whole
History is divided into three parts: the history of the Colonies;
the history of the Revolution, which in turn is separated by the
Declaration of Independence into two subdivisions; and the his-
tory of the formation of the Constitution. The publication of the
portion dealing with the Revolution stirred up a series of con-
troversies. Bancroft was compelled to express himself frankly
concerning men who had living descendants. Family pride was
aroused, and pamphlets were issued in defence of ancestors whose
reputations were supposed to have been injured by the strictures of
the historian. Mr. Winsor, in a note in the eighth volume of the
''Narrative and Critical History of America," gives an account of
the literature of this description bearing upon the most important

GEORGE BANCROFT. 365

of these pamphlet battles. The controversies there recapitulated
were based upon language used by Bancroft concerning Colonel
Timothy Pickering, General Greene, General Schuyler, General
Sullivan, and Joseph Reed. In the last case it was shown that
Bancroft had been misled by an incident which had occurred to
another Colonel Heed; and the charge based upon that error was
withdrawn in the Centenary Edition of the History. The language
used with reference to Schuyler and Sullivan was also in each case
modified in this edition, but the judgments of the men remain
substantially as before. In the Preface to the sixth volume, he
says, *'I hope at least it will appear that I have written with
candor, neitlier exaggerating vices of character, nor reviving na-
tional animosities, but rendering a just tribute to virtue wherever
found." It was in this spirit that he approached the subject; and
upon the judgment of posterity as to how far he was able to live up
to it his reputation as an historian must stand. The majority
of readers at the present day will give liim credit for the exercise
of a judicial spirit in reaching the conclusions for which he was
attacked. His History would be worthless if he had not been
manly enough to express his opinions. His conclusions have
seldom been doubted, save where they conflicted with the esti-
mates of kindred.

His style has been condemned as redundant and pompous ; but
his capacity for marshalling events in narrative form has been
admitted even by those disposed to criticise. Colonel Higginson's
opinion on this point is vividly set forth in the following language :
''The reader is compelled to admit that his resources in the way
of preparation are inexhaustible, and that his command of them is
astounding. One must follow him minutely through the history of
the war for independence, to appreciate in full the consummate
grasp of a mind which can deploy military events in a narrative
as a general deploys brigades in a field." The same writer calls
attention to a fault '' which Bancroft shared with his contempora-
ries, but in which he far exceeded any of tliem, — an utter ignoring
of the very meaning and significance of a quotation mark." This
criticism is based upon an obvious defect, the existence of which
cannot be denied. Instances can be found in Bancroft's works in
which quotation marks are used enclosing paragraphs in which
there were abridgments and insertions for the purpose of making
the thought continuous, without the typograj)hical marks essen-
tial to denote these changes. He does not, however, need especial

366 GEORGE BANCROFT.

defence for pursuing the practice of his contemporaries. It is
logical perhaps to say that, when tried by Mr. Bancroft's literary
methods, the mass of correspondence quoted in the History of the
Constitution, "though valuable as suggestion, is worthless as au-
thority," but it is severe. The same test would compel the rejec-
tion of Sparks's Washington, if indeed it would not carry with it
the entire collection of Washington's manuscripts in the State
Department; for we know that some of the letter-book copies of the
correspondence, on file in that office were revamped b}^ Washing-
ton himself. There is no hesitation in quoting Sparks, and perfect
confidence is felt that he made no change, either by alteration or
omission, that could in his judgment affect the sense of the text,
nor is it to be believed that Washington ever changed a word in his
letter-books which he thought could modify their interpretation.
The point must be indeed narrow and technical, a question of close
construction or of the use of a word under circumstances demand-
ing for an estimate of its value all that immediately precedes or
follows it, that would call for the verification of any of Bancroft's
quotations by comparison with original documents.

Marginal references in the first six volumes, although not copious,
are frequently met with. In the seventh volume they are entirely
omitted, and thenceforward through the History are rarely to be
seen. This change is much regretted by students. An explana-
tion of why the references were omitted in the seventh volume is
to be found in the Preface. The reason stated was "the variety
and multitude of the papers which have been used, and which could
not be intelligently cited, without burdening the pages with a
disproportionate commentary." It was apparently Bancroft's in-
tention at that time to cull out for publication such letters as
would confirm his narrative, " and possess an intrinsic and gen-
eral interest by illustrating the character and sentiments of the
people during the ten or twelve years preceding the 4th of July,
1776." This purpose he did not carry out.

The extracts from the opinions of reviewers previously quoted
show that nearly all of the writers were of opinion that Bancroft
was destined to hold permanently his position as the historian of
the United States. How far the charges of partisanship made by
such writers as Griswold and Lecky, and later by pamphleteers in
this country, have affected this position, may be measured somewhat
by the expressed opinions of historical students. Robert C. Win-
throp, at the first meeting of the Massachusetts Historical Society

GEORGE BANCROFT. 367

after Bancroft's death, said: ''You have well said, Mr. President,
that Bancroft was foremost as the historian of the United States.
His great work in all its varied editions will always he read and
recognized as the leading authority on American history for the
period which it includes. His style may he criticised, and cen-
sured as redundant or rhetorical. His philoso^jhy may be discarded
as partaking occasionally of that German mysticism which he im-
bibed in his youth. A vein of partisanship too, may sometimes be
detected amid all his professions of impartiality. It could hardly
be otherwise. No one in writing history, or in doing anything
else, can escape from himself, or can wholly conceal, even should
he try to do so, his own preconceived opinions, his own individual
peculiarities and idiosyncrasies." Further on in his remarks
Mr. Winthrop added, " The triith of history was upj)ermost in his
aims and efforts from first to last." Two months later Mr. Win-
throp attended the semiannual meeting of the American Anti-
quarian Society. Bancroft had for twenty years been the First
Vice-President of this Society. A memorial of his life was read
by Samuel S. Green. Mr. Winthrop, alluding to Bancroft's death,
then said, "1 do not forget that he and I shared so long the dis-
tinction of being the oldest members of the Society, and that is
now left to me alone." He then added, with great earnestness of
manner, " I paid my little tribute to him at our Massachusetts
Historical Society, and I have nothing to add to it, and nothing
to detract from it." Mr. Winsor, in the Narrative and Critical
History of America deals with the question of the probability of
Bancroft retaining his position in the following language: "His
learning and the extraordinary resources of his material are likely
to make his work necessary for the student, till another with equal
or better facilities shall compass the subject in a way to gain wider
sympathy." In other words, the writer who shall supplant Ban-
croft must command at least equal facilities, and deal with the sub-
ject in a more attractive way. To measure the probabilities of the
occurrence of these events, it is necessary to review the resources
at his command.

It has been seen that no especial claim can be made for him in
regard to the facilities at his disposal for the preparation of his
Colonial History. The researches of topical students have placed
within easy reach of the writer of to-day much that was to be
learned only by diligent study when Bancroft wrote. To examine
the Colonial records of the original States, he was obliged to travel

368 GEORGE BANCROFT.

from State to State. Many of these records have since been pub-
lished, and are now to be found in all the great libraries. When
we reach the Revolutionary period, however, it is easy to compre-
hend why Colonel Higginson speaks of Bancroft's resources as inex-
haustible, and why Mr. Winsor characterizes them as extraordinary.
Bancroft himself has given a detailed recapitulation of his manu-
script sources of authority in the Prefaces to his sixth and ninth
volumes. There is not space at command to give this informa-
tion in full. The following extract from Mr. Winsor's abridged
statement of the information furnished by these Prefaces will serve
to show how remarkable were his opportunities : —

" Nothing was refused hiin in the English State Paper Office, nor at the
Ti'easury. The manuscripts of the British Museum and the Royal Insti-
tution, such of the Chatham Papers as had not been printed, the Shelburne
Papers, including the letters of Shelburne and tlie King, an autobiography
of the third Duke of Grafton, a journal of the Earl of Dartmouth, the
letters which passed from the King to Lord North, not to mention others
of lesser importance, were placed at his disposal. In France the archives
were thrown open to his search without restraint, and the treasures of the
Marine and War Departments were largely drawn upon. On the nego-
tiations for peace, the French archives offered liini the richest material.
From Cierniany his acquisitions were peculiarly valuable, as Sparks had
scarcely reaped anything from that field. He found the archives of
Hesse-Cassel closed to him as to others, but through the instrumentality
of Friedrich Kapp and others he secured the possession of private journals
and reports of the Hessian officers, and caused searches to be made in the
wide field of the contemporary publications in Germany for letters and
criticisms on the part of the German auxiliaries in the war which he con-
siders 'in the main the most in)portant of all that have been preserved.'
From Berlin, he got the reports made to the Duke of Brunswick by his
officers which have finally found a lodgment in the Russian archives; and
he also secured the collection which I\Iax von Eelking, the writer on the
Hessian story, had amassed in his studies. He likewise obtained copies
of the correspondence of Frederick the Great with his foreign ministers,
so far as it touched upon the affairs of America. From Moscow and
Vienna, from Holland and from Spain, other documents came to swell
the record, which have enabled him to make his account of the foreign
relations of the Confederacy the best by far which had been prepared.

" His wealth of American papers is probably from their scope unsur-
passed in private hands. He iiad of course at his command the resources
of the government archives, and those of the original States; he could
examine the papers of the Revolution gathered in public libraries, and in
the cabinets of historical societies; and besides these he had his own
gatherings ; the correspondence of the agents of the various Colonies in

GEORGE BANCROFT. 369

London prior to the outbreak of actual war, like Bollan, Jasper Mauduit,
Richard Jackson, Arthur Lee, Franklin, W. S. Johnson, and others; the
papers more or less extensive of Hutchinson, Israel Mauduit, Pownall,
Hollis, JNIayhew, Andrew Eliot, Golden, Bernard; and, above all, the papers
of Samuel Adams, which passed into Bancroft's hands some years ago.

" He speaks also of two volumes of papers of Greene, and the papers
of Anthony Wayne, which were submitted to his inspection."

If ever the opportunity should occur for one man to command
such resources as these, he still must, in the contest for supremacy,
measure swords with Bancroft in the treatment of the subject.

Mr. Winsor in the "final statement " in his History speaks of
the value of monograph, as rounding the treatment of any phase of
history in a way rarely accomplished in more comprehensive work.
One of the criticisms which has been made upon Bancroft's work is.
that he did not keep up with the times, and that in his last revision
he did not devote himself to a more detailed investigation of the
work of specialists in the several topics covered by his History, in
preference to confining his labor mainly to the elimination of re-
dundancies and the condensation of material. A glance at the
manner in which he carried the scheme of his History into execu-
tion wall furnish a partial answer to this criticism. He treated his
subject by topics exhaustively, and as he progressed he devoted his,
time to the investigation of the new field which w'as before him.
It was impossible that he should be constantly at work where he
had already concluded his labors, and equally impossible but that
from time to time his attention should be called to errors which
notwithstanding his vigilance had crept in. As edition after
edition of his works came out, he eliminated such errors as came
under his observation; but his main labor was devoted to the perfec-
tion of his scheme. When the History proper was concluded, and
afterward when the volumes on the Constitution were jiublished,
he was confronted with the question, whether, in a revision of the
published volumes, he should merely try to condense them, or
whether he should attack the subject anew and attempt to treat
it as a whole, taking up the stud^^ of each part where he had pre-
viously dropped it. At his time of life the latter course was prac-
tically impossible. He chose the former, and, while opinions may
differ as to the wisdom of the choice, it will commend itself to the
majority.

Mr. Bancroft's career has been presented as statesman, as
historian, and as citizen. Whether his memory is longer to be
VOL. XXVI. Is. s. xviii.) 24

370 JULIUS ERASMUS HILGARD.

preserved tlirougli his founding the Kaval School, and through his
connection with the acquisition of California and the enlargement
of the rights of naturalized citizens, in the first capacity, through
his recognized pre-eminence as the historian of his country; or
through the grateful recognition of his forethought in founding
exhibitions at Exeter, Harvard, and Worcester, posterity alone
can tell.

JULIUS EKASMUS HILGARD.

The father of the subject of this sketch was a man of brilliant
mind and of high attainments. He gave up a promising judicial
career in Bavaria, against the wishes of that government, because
of his republican sentiments, and of his desire to find a congenial
sphere of action in the ideal republic which he expected to find on
this side of the ocean. Accredited to the good will of the American
people by a letter of introduction from Lafayette, he emigrated to
this country with a large family of children, and arrived at New
Orleans in the winter of 1835. Going up the Mississippi, he
purchased a farm near Belleville, St. Clair County, Illinois, and
there devoted his time to farming and to the education of his chil-
dren, when the latter were not engaged in the manual labor in-
separable from a pioiieer life in the West. His son Julius, born
in Zweibriicken, January 7, 1825, received from his father instruc-
tion in the classics and in the modern languages, but soon out-
stripped his teacher in mathematics, and at the age of eighteen
went to Philadelphia to study engineering. There, at the house
of a mutual friend, he met Professor Bache.

In tracing the history of his connection with the Coast Survey,
to which his life was devoted, we find a letter from young Hilgard
to Bache, dated January, 1844, calling the attention of the latter
to errors in the formulas used by the Coast Survey in the computa-
tion of geographical positions, and giving his own development of
correct formulas. In reply. Superintendent Bache wrote to the
youth of nineteen, saying, ^'You have overriden two of our most
experienced computers, and have shown that they were seriously in
error." To an offer from Bache of employment in the Survey in a
subordinate capacity and at small pay, he responded tliat he would
rather do '* high work at low pay than low work at higli pay," and
gladly accepted the position. He entered at once upon his duties,
but his formal appointment was dated December 28, 1846. His

JULIUS ERASMUS HILGARD. 371

ability as a field officer in astronomical work and triangulation at
once secured liim the commendation of his immediate superiors
and of the Superintendent. In 1851, 1852, and 1853 he was in
charge of the computing division, and there again his ability was
made manifest, as is shown by the report of Captain Benham,
U. S. Eng., who, on assuming charge of the Coast Survey Office
in 1853, reported to the Superintendent that he found the com-
puting division the best organized division in the office.

The annual reports of the Survey contain the evidence of Hil-
gard's scientific activity up to 18G1, when for a brief period he
severed his direct connection with the Survey to embark in private
enterprise. But at the outbreak of the War, when the very exist-
ence of the Survey was endangered by threatened legislation, Bache
appealed to him to return to Washington and help save the Survey,
Responding at once, Hilgard called upon Schuyler Colfax and
Roscoe Conkling, eminent as leaders in the dominant party, and
made a strong argument showing how necessary such an organiza-
tion as the Coast Survey must be to the country in time of war.
So searching was Roscoe Conkling's examination of Hilgard's
argument that the latter mistook his earnestness for an evidence
of a spirit hostile to the Survey; but to his relief and satisfaction
it was soon shown that he had found a most earnest supporter in
Conkling no less than in Colfax. This incident is here recorded
in order to recall to the mind that, in summing up the labors of
Hilgard in the cause of science, it should be remembered that it is
a sufficient outcome of a life to have borne a large part in main-
taining the efficiency of so great and useful an institution as the
Coast Survey. In 1862, Hilgard assumed charge of the Coast and
Geodetic Survey Office, and this during the war involved heavy
responsibilities.

Professor Bache having become incapacitated through mental
disease, the duty of directing the Survey in all its branches de-
volved upon Hilgard in the autumn of 1864, in addition to his
labors as assistant in charge of the office. Thus he was virtually
Superintendent until the appointment of a successor to Bache, in
February, 1867, without reaping the reward of a formal appoint-
ment, which he richly deserved. As an estimate of his services by a
critic by no means partial, and yet competent, the opinion of Bache's
successor, Benjamin Peirce, is cited here: "During the illness
of my lamented predecessor, the administration of the Survey fell
upon the shoulders of the assistant in charge, J. E. Hilgard. The

372 JULIUS ERASMUS HILGARD.

distinguished ability with which this difficult service was dis-
charged was manifest to all. He has extended to me the bene-
fits of this experience liberally and loyally. While I willingly
acknowledge myself under deep and lasting obligations to him for
the aid thus rendered me, I can also testify that in all respects he
has been equally true to my predecessor, the greatness of whose
reputation has not been diminished in his keeping."

Until his appointment to the superintendency, in 1881, Hilgard
continued in charge of the Coast Survey Office. In addition to this
duty, he virtually conducted the office of Weights and Measures,
and helped in shaping the legislation in regard to the legalization
of the metric system in this country. The metric standards for
the States of the Union were prepared under his supervision. It
was therefore fitting that he should be appointed one of the scientific
delegates to represent this government in Paris at an international
conference having for its object the construction of a new meter as
an international standard of length. In 1872 he attended the con-
ference and actively participated in its deliberations, and when it
had been decided to establish an International Bureau of Weights
and Measures at Paris, its directorship was offered to him, but de-
clined. While in Europe on the mission just referred to, he made
a determination of Transatlantic longitudes, including in his opera-
tions Paris and Greenwich. It is rather singular that the first
successful telegraphic longitude determination between these great
ob^jervatories should have been made by an American, and his suc-
cess in this undertaking was always looked upon by him as also a
diplomatic triumph.

Hilgard was a charter member of the National Academy of
Sciences, and took an active part in many of the investigations
made by the Academy for the government. He conducted a mag-
netic survey of the country, under the auspices of the Academy,
with means supplied out of the Bache Fund. In 1874 he was elected
President of the American Association for the Advancement of
Science.

His life Avas in many respects burdened by misfortune; for of
his four children three died young and only one lived to man's
estate and then died, leaving him childless and overwhelmed by
grief at a time when a fatal disease had already begun its inroads
on his mental and physical strength. This disease had already
seriously impaired his health when he was appointed to the super-
intendency in 1881, which, to use his own words, came ''too late

J)

CHRISTIAN HEINRICH FRIEDRICH PETERS. 373

There can be no doubt that he was conscious of his failing strength
and ability while still occupying the position of Superintendent,
for on more than one occasion he gave expression to the wish that
the burden of his duties might be shifted to other shoulders.

His retirement took place in 1885, and from that time on his
lingering illness entailed great sufferings, and several times brought
him to the point of death. From each of these attacks he rallied,
with less power of resistance, until death relieved him of his suf-
ferings, on May 8, 1891.

CHRISTIAN HEINRICH FRIEDRICH PETERS.

Christian Heinrich Friedrich Peters was born on September
19, 1813, at Coldenbiittel, in the Duchy of Schleswig. He studied
mathematics and astronomy at Berlin between 1832 and 1836, and,
after attaining the degree of Doctor of Philosophy, continued his
education under the celebrated Gauss at Gottingen. In 1838 he
was engaged to assist in a scientific expedition to Sicily for in-
vestigations on Mount Etna, and was employed, after the conclu-
sion of this work, to direct the trigonometrical survey undertaken
by the Neapolitan government.

Upon the outbreak of the revolutionary movements of 1848, it
became impossible for him to retain his situation without sup-
pressing his own sentiments in favor of freedom. Unable to con-
ceal his liberal sympathies, he was dismissed and banished. But
he soon returned to take an active part in the Sicilian insurrec-
tion, during which he served under Mieroslawski, first as captain
and then as major of engineers. The suppression of the insurrec-
tion left him in imminent peril of capture by the Royalists ; but
he finally effected his escape to France. He next went to Turkey,
in hopes of obtaining scientific employment in that country; but
the outbreak of the Crimean war put an end to these expectations,
and he came to the United States in 1854. Here he found em-
ployment in the Coast Survey, and subsequently, in 1858, was
appointed Professor of Astronomy at Hamilton College, Clinton,
New York, and Director of the Observatory connected with that
institution. He retained this position until his sudden death, in
1890. On the morning of July 19 in that year, his body was
found on the steps leading to the building where he had lived, and
it appeared that he had died on his way home from his customary
work in the Observatory.

374 CHARLES JOHN BIAXIMOWICZ.

Notwithstanding his duties as a teacher, he found time during
his life at Clinton for a great amount of astronomical observation.
His principal work was that of determining the places of faint
stars, with a view to the preparation of an extensive series of
charts, part of which he published at his own expense in 1882.
In the course of these observations he discovered many new vari-
able stars and forty-eight new asteroids. He also made a long
series of observations of solar spots. In 1874 he was chief of the
expedition sent by the United States government to observe the
transit of Venus. In 1876 he was chosen a member of the Na-
tional Academy of Sciences.

His studies of the ancient catalogues of stars, such as that of
Ptolemy, were extensive and profound, and part of his published
work relates to these and similar subjects. His frequent journeys
to Europe maintained his acquaintance with his professional col-
leagues of Germany and other countries, among whom he was
always cordially welcomed. While the combative temperament
which had formerly made him a soldier in the cause of Sicilian
independence occasionally led him into controversies with regard
to the extent of his personal rights, he made many friends, by
whom he was greatly beloved.

FOREIGN HONORARY MEMBERS.

CPIAELES JOHN MAXIMOWICZ.

Russia has been fruitful during the last seventy years in bot-
anists of more than ordinary ability, as is shown by the mention
from among them of such names as Besser, Bongard, Bunge and
the Fischers, Herder, Ledelx>ur, Maximowicz, Meyer and Kegel,
Ruprecht, Trautvetter, and Trinius, all well known to the botani-
cal world. Of these this Academy has numbered among its For-
eign Honorary Membei-s only the subject of the present notice,
C J. Maximowicz, who was elected on October 10, 1888, and died
on the 16th of February last.

Maximowicz was born on November 23, 1827, in the town of
Toula in Central Russia, though most of his boyhood was spent in
St. Petersburg. In 1844 he entered the University at Dorpat,
where Dr. Bunge was then Professor of Botany, and upon gradu-

CHARLES JOHN MAXIMOWICZ. 375

ation he received the appointment of Director's Assistant at the
Dorpat Botanic Garden, whence he was removed in 1852 to the
curatorship of the Imperial Botanic Garden in St. Petersburg.
The next year he was commissioned as botanist and collector
for the Garden to accompany the frigate Diana upon an expedi-
tion around the world, but the voyage was interrupted by the
breaking out of the war with France and England, and Maxi-
mowicz left the ship upon reaching the Russian colony that had
been recently established near the mouth of the Amur on the
coast of Mandshuria. He here devoted himself to a botanical
exploration of the then little known region bordering the Amur
River, which he carried on assiduously under many difficulties for
over two years, returning to St. Petersburg through Siberia in
1857. The critical study of his collections, and of such other
material as had been received from the same territory, occupied
him for two years longer, the results being embodied in his Primi-
tice FlorcB Amurensis. In this elaborate work he gave not only
a detailed account of the plants, but a general view of the physical
and botanical features of the country, the distribution of trees, and
a comparison of the flora with others most nearly related to it. In
acknowledgment of its scientific merits, he was awarded the Demi-
doff prize of five thousand roubles banco. He was now again sent
to Eastern Asia to continue his botanical researches, and for four
years travelled through Mandshuria, reaching the frontiers of
Korea, and through the Jaj)anese islands of Jesso, ISTipou, and
Kiusiu, returning to Europe in 1864.

From this time till the end of his life his main purpose was the
preparation of a Flora of Japan and Eastern Asia. As Chief Bot-
anist and Curator of the Herbarium at the Imperial Botanic Gar-
den, and, after the death of Ruprecht, as Director of the Museum
and Herbarium of the Imperial Academy of Sciences, he was
burdened with official duties which continually interrupted and
delayed the carrying out of this design, but it was never given up.
Many contributions were published, chiefly in the Memoirs and
Bulletins of the Imperial Academy, which were more or less di-
rectly related to this work, and which are often of interest to
American botanists on account of the close relationship of the East
Asiatic and the North American floras, and the consequent neces-
sity of his taking into consideration American as well as Asiatic
forms. Among these contributions may be noted revisions of the
Asiatic Ehamnece (1866), HijdranrjecB (1867), and Rhododendrece

376 KARL WILHELM VON NAEGELI.

(1870), of the genus Lespedeza (1873)/ of the Spirccacece (1879),
and of Coriaria, Ilex, etc.. (1881) ; also a series of twenty papers
entitled Diagnoses Plantarum Novarum Japonioi et Mandshurice
(1866-187(3), and another series entitled Diagnoses Plantarum
Novarum Aslaticarum (1877-1890). In 1873 he visited Finland
and Sweden, especially to consult the herbarium of Thunberg at
Upsal, and most of the summer of 1875 was spent in a A-isit to
Scotland, Kew, Paris, and Germany. At about this time he was
also expending much critical labor ujDon Japanese plants in aid
of Franchet and Savatier in the preparation of their Enumeratio
Plantarwn Japonlcarum, which owes its value verj largely to this
co-operation.

The last ten years of his life were occupied chiefly in the study
of large and important collections from the previously almost
inaccessible regions of Central Asia, especially those of Przewalski
and Potanin from Tangout (Northern Tibet) and Mongolia, and in
the elaboration of an extended report which was to be illustrated
with a hundred or more finely engraved plates. Much of this was
completed and ready for the press, but only the first parts are as
yet published. The general results, as showing the characteristics
of the flora of the region, were ably summarized by him in a paper
read before the Botanical and Horticultural Congress held at St.
Petersburg in 1884. To the great loss of botanical science he was
cut off most unexpectedly in the midst of his labors, dying on
February 16, 1891, of an attack of influenza, after a short illness.

The work of Maximowicz, as a botanist, is remarkable through-
out for its extreme thoroughness and most scrupulous exactness in
all its details, for good judgment and a purely scientific spirit,
and he must always rank as a high authority in the department to
which he devoted himself. As a man he was most estimable, of
noble and spotless character, a scholar of high culture, and a most
courteous and genial correspondent.

KARL WILHELM VON NAEGELL

Karl Wiliiklm von Naegeli was born on March 27, 1817, at
Kilchberg, near Zurich, and died at Munich, May 10, 1S91. His
education, begun in a private school at Zurich, was continued in
the Gymnasium of that city until he entered the University of
Zurich, where he received his doctor's degree in 1840. JI(> had
at first intended to study medicine, but his taste for natural science

KARL WILIIELM VON NAKGELI.

377

was so decided that lie was allowed to go to Geneva, where he stud-
ied with De C-andulle for a time. He afterwards weiil to Berlin,
where his attention was turned to philosophical studies. He then
proceeded to Jena, where he was associated with Schleiden, whose
influence is clearly seen in the earlier writings of Naegeli. He was
married in 1845, and soon after returned to Zurich, where he be-
came Privatdocent, and afterwards Professor Extraordinarius. In
1852, Naegeli, after having declined a call to Giessen, was ap-
pointed Professor at Freiburg, in Breisgau. In 1855 he again
returned to Zurich, where he was made Professor of General Botany
in the new Polytechnic School. He resigned this position in 1857,
and accepted an appointment as Professor of Botany and Director
of the Botanic Garden at Munich, where he remained until his
death. Por the last twenty years of his life Naegeli's health was
feeble, but he was nevertheless able to continue his scientific work
during most of that period. In his feeble condition he was unable .
to rally from au attack of the influenza during the epidemic of
1889-90, and gradually succumbed to the disease.

During his long and active scientific career, Naegeli's influence
was seen mainly in his writings, for, as a university lecturer, he
did not succeed to the same extent as some of his contemporaries
in attracting numbers of enthusiastic followers. Owing to certain
peculiarities of temperament he was not personally popular with the
botanists of Germany, and few of the younger botanists sought his
instruction. That he was, however, capable of stimulating oth-
ers to work of the highest grade is evident, if we consider that
C. Cramer and Schwendener were his pupils and associates.

As a writer and investigator probably no botanist of the present
century has had greater influence in shaping the course of modern
botany than Naegeli. His botanical career began at a time when
the influence of Schleiden was predominant, and naturally the
early work of Naegeli bore the mark of Schleiden's peculiar views.
But Naegeli was a man of decided originality, and united great
accuracy as an observer with a genius for speculation and philo-
sophical inquiry, and he soon freed himself from the limitations ol
Schleidenian conceptions. If, at the present day, we are obliged
to admit that some of Naegeli's own theories have not stood the
test of time, we must also admit that they were very suggestive
and fruitful of results in their day, while, as an observer of facts,
we can only admire his uniform accuracy and truthfulness. It is
only natural that the theoretical views of Naegeli, formulated at a

378 KARL WILHELM YON NALGELI.

time when botanists were comparatively few and accurate observa-
tions scant;^, should, in the light of accumulated modern observa-
tions, be superseded by other more tenable theories ; but we must
still continue to acknowledge that we are indebted to Naegeli for
the solid foundation of more than one of the branches of botany
which are now regarded as among the most interesting and imjjortant
fields of modern research.

Naegeli was neither a pure sj'^stematist, nor strictly a physi-
ologist in the modern sense. If he pursued systematic studies
to some extent, it was with the view of preparing himself to dis-
cuss the abstract questions of the nature of species and the theory
of descent. His detailed work on the genera Cirslum and Hiera-
c'lum was undertaken with very much the same purpose as that of
Darwin in his work on Cirrhipeds; namely, by mastering the
specific differences to be found in a few large and variable genera,
to prepare himself for the intelligent discussion of the relations
and probable origin of species in general. Beyond this he felt
little interest in systematic work. So far as his work on the na-
ture of cell structure, the formation of the cell wall, the method of
reproduction in Cryptogams, and the phenomena of fermentation
is concerned, it was certainly physiological rather than systematic;
but, using the terminology of the present day, it may be said that
Naegeli was pre-eminently an histologist, and that the greater part
of his theories and general views, so far as they were derived from
his own work, had a histological basis. Unlike De Bary, Prings-
heim, and the younger generation of German botanists, he did not
attempt, to any extent, to study what may be called the life-history
of any special group by means of cultures.

The histological work of Naegeli was admirable, and he was
practically the first to introduce histological methods into the study
of algte and other groups. In his ''Die neueren Algensysteme
und Versuch zur Begriindung eines eigenen Systems der Algen und
Florideen " (1847) he accumulated a large mass of facts, and was
the first to give an accurate account of the thallus of different
species, and to show the necessity for studying tlie apical growth
as a means of classification in this group. In his attempt to form
a new system of classification he was less successful. He excluded
the Floridece from algae; but although he was accurate as far as
concerned the microscopic structure which lie studied, he failed to
recognize the true sexual relations of the FIoridecB. In his i)aper
" P)eitr{ige zur Morphologic und Systomatik der Ccramiacoa; "

KARL WILHELM VON NAEGELI. 379

(1861) his accuracy as an observer is shown, for in this paper he
first figured correctly the young female condition of ilio' Floridece,
but again failed to comprehend the true significance of his obser-
vation, and it was left to Bornet and Thuret, in 1867, to give the
proper explanation and fix the true position of the Floridece. It
may be said that had Naegeli studied living instead of alcoholic
material, he might perhaps have avoided his error. Naegcli's
paper on the cell division in Delesserla Ili/jjoglossum and on the
structure of Caulerpa proUfera were also valuable contributions to
our knowledge of algse, and his " Gattungen einzelliger Algen "
(1849), more purely systematic than his other works on algye, still
remains a classic monograph on the subject.

In his first histological paper (1841) on the development of pol-
len, and in later papers on the development of stomata and the
structure of the root-apex, Naegeli proved himself to be a better
observer than his teacher, Schleiden, and in two important papers
published in 1844 and 1846 on nuclei and the formation and growth
of vegetable cells, he showed emphatically that cell division is the
true mode of vegetative cell formation. Although in his earlier
paper on the growth of the leaf, Naegeli had been led to errone-
ous views on the nature of stems and leaves, nothing but praise
can be said of his paper on the ''Growth of the Stem and Root in
Vascular Plants and the Arrangement of the Vascular Bundles "
(1858). This paper is regarded by botanists as his most important
histological work, and is the basis of the countless works of more
recent writers on the subject. The enormous work of over six
hundred pages on starch grains (1858) is full of important obser-
vations, and served as the foundation of the micellar theory, which
has been alternately attacked and defended up to the present time,
one phase of the discussion being the method of formation of cell
walls by intussusception as opposed to apposition. For some time
the preponderance of opinion rather favored Naegeli 's theory of
intussusception, but, although the question cannot as yet be said
to be settled fully, the advocates of the theory of apposition have
of late appeared to have the strongest evidence on their side, and
physicists, as a rule, do not regard the micellar theory with much
favor.

Naegeli's writings on fermentation, ''Theorie der Gahrung "
(1879), with which may be classed properly his "Die niederen
Pilze " (1877), represent rather his theoretical views based on
the work of others than conclusions founded on his own work.

380 KARL WILHELM VON NAEGELI.

Although interesting and suggestive, they hardly possess the same
weight as his other writings. He differs with many recent writers
in believing that, among forms like bacteria, it is dovibtful whether
definite species exist, morphologically speaking, as in higher plants.
It is still too soon to say whether his view on this point is correct
or not; for, although most bacteriologists do not now agree with
Naegeli, it must be admitted that the question is still an open one,
and it Avould be rash to predict what would be the general verdict
on this point a decade hence.

Up to this point we have spoken only of the special work which
entitles Naegeli to be regarded as one of the foremost botanists of
his time, unsurpassed and jDerhaps unequalled in his own special
field. But his influence was felt beyond purely botanical circles,
and he acquired by his writings on evolution a wide-spread repu-
tation among all scientific men. In his doctor's thesis (1840) on
the Swiss species of Cirsium, Naegeli foreshadowed a line of study
which he afterwards worked out more elaborately in his later work
on Hieracium, in which he made a minute and critical study of the
variable species of a large genus serve as a groundwork for a con-
sideration of the theory of descent. Early in life he believed in
the absolute difference of species, although he urged the necessity
of the study of development as that which gives real value to the
knowledge of mature forms. In a later work, however, he stated
that this early belief in the absolute difference of species 'Mid
not prevent his believing even then in the origin of species by
descent." In his important paper, '*Die Entstehung und Begriff
der naturhistorischen Art " (1865), he discussed the Darwinian view
of the origin of species, and stated that his own belief in the origin
of species by descent had been definitely expressed in a paper pub-
lished in 1856. He differed with Darwin in believing that varia-
tion occurred in a definite direction, — a view similar to that held
by Gray, — and he was unable to accept natural selection as a
sufiicient explantion of evolution. His views were peculiar in
that he believed that the different groups of plants originated inde-
pendently from what he called " Urzellen," and, taking the differ-
ent groups as they now exist, he failed to recognize a gradual
development of the higher forms from the lower. Without stop-
ping to consider his papers on the influence of external conditions
in the formation of varieties and the theory of hybrid formation, we
need here only refer to ''Die mechanisch-physiologische Theorie
der Abstammungslehre " (1884), a work in which he states the

EDUARD SCHONPELD. 381

conclusions reached as the result of an unusually long and profound
study. This work must be regarded as Naegeli's greatest contri-
bution to speculative science, worthy to be classed with the mas-
terpieces of the great writers on evolution. He maintains that
variation arises from internal, not external causes, and that the
transmission of hereditary characters dejiends not on the general
protoplasm, but on a limited and definite part of it, the idioplasm.
The enunciation of the general principle in the Abstammungslehre
is forcibly and even brilliantly stated, and Naegeli's presentation
of the subject has exerted and will continue to exert a marked in-
fluence on modern thought, although in some details he allows
himself to indulge in views which are too purely speculative, and
not borne out by the more exact microscopic work of a younger
generation of workers.*

EDUARD SCHONFELD.

Eduard Schonfeld was born on December 22, 1828, at Hild-
burghausen, Germany. The comprehensive activity of his mind
was early displayed in the course of his education, for he studied
architecture and chemistry before finally selecting astronomy as
his special field of work. His astronomical studies were begun at
the University of Marburg, and continued, in 1852 and later, at
Bonn, under the guidance of the illustrious Argelander. He took
his degree in 1854, but had already in the previous year been
appointed Assistant in the Observatory.

At this time, Argelander was entering upon the execution of
his plan for the formation of a catalogue which should exhibit the
approximate positions, and also the magnitudes, of all stars in the
northern hemisphere not fainter than the ninth magnitude. He
had made some preliminary observations for tliis purpose in 1852,
but the work was definitely begun only after he had been joined
by Schonfeld, who took a prominent part both in the observations
themselves and in their reduction. The catalogue itself, which
comprises 324,198 stars, was mainly drawn up by him.

In 1859, Schonfeld was appointei Director of the Observatory
at Mannheim. Here he undertook the systematic observation of
variable stars, and his two successive catalogues of these objects

* A detailed account of Naegeli's life and work is to be found in the " Neue
Zurcher Zeitung" for May 16, 1891, and in " Nature" for October 15, 1891.

382 EDUARD SCHONFELD.

are well known and higlil}' valued among astronomera. With the
Mannheim telescope he observed many nebulte, and published a
catalogue of these also.

Upon the death of Argelander, in 1875, Schonfeld was recalled
to Bonn to succeed him. He then undertook the extension of the
great catalogue above mentioned into the southern hemisphere, as
far as the twenty-third degree of south declination. The number
of stars thus added to the catalogue was 133,659. This additional
catalogue, as well as that previously prepared under Argelander's
direction, is accompanied by an atlas, showing the places of all
the stars observed. The value of the entire work to astronomers
can hardly be described in terms which would not seem extrava-
gant to readers of other professions.

But the mere enumeration of the official occupations which
formed the regular business of Schonfeld's life would give a very
imperfect idea of his claim to the esteem and gratitude of his col-
leagues and contemporaries. As has already been suggested, his
mind was distinguished by its breadth and versatility. An aston-
ishing memory, very extensive reading, and pre-eminent ability in
teaching and in writing, combined with a perfectly unpretentious
and amiable disposition, made him, as may easily be imagined,
a leader in his profession, commanding the admiration and good
will of all who knew him personally or only by his jiiiblications.
He was one of the most active and valuable members of the As-
tronomische Gesellschaft, and, as one of its secretaries in 1875 and
later, was largely concerned in the publication of its journal, the
** Vierteljahrsschrift," to which he frequently contributed articles
of more than ordinary interest.

He died on May 1, 1891, after a prolonged illness, which, in the
opinion of some of his friends, was to be ascribed to the unremit-
ting labor to which he subjected himself in the observations for
the extension of his great catalogue of stars, after his return to
Bonn. His interest in astronomy continued unabated to the last,
and the number of the '' Astronomische Nachrichten " subsequent
to that in which his death is announced contains a final article
from his pen on a point in the history of astronomy, in explaining
which his learning and ingenuity are, as usual, made evident.

REPORT OP THE COUNCIL. 383

The Academy has received an accession of eight mem-
bers, — five Fellows, one Associate Fellow, and two Foreign
Honorary Members.

Two Resident Fellows have resigned.

The roll of the Academy corrected to date includes the
names of 186 Fellows, 89 Associate Fellows, and 69 Foreign
Honorary Members.

May 26, 1891.

LIST

OF THE FELLOWS AND FOREIGN HONORARY MEMBERS.
(Corrected to January 1, 1892.)

RESIDENT FELLOWS. — 18G.

(Number limited to two himdred.)

Class I. — Alathematical and Physical Sciences.

78.

Section L — 7.

Mathematics.

Gustavus Hay,
Benjamin O. Peirce,
James M. Peirce,
John D. Runkle,
T. H. Safford,
William E. Story,
Henry Taber,

Boston.
Cambridge.
Cambridge.
Brookline.
Williamstown.
Worcester.
Worcester.

Section H. — 10.
Practical Astronomy and Geodesy.

Seth C. Chandler, Cambridge.
Alvan G. Clark, Cambridgeport.
George B. Clark, Cambridgeport.
J. Rayner Edmands, Cambridge.
Henry Mitchell, Boston.

Edward C. Pickering, Cambridge.

John Ritchie, Jr.,
Edwin F. Sawyer,
Arthur Searle,
O. C. Wendell,

Boston.
Brighton.
Cambridge.
Cambridge.

Section HL — 46.

Physics and

Chemistry.

A. Graham Bell,

Washington.

Clarence J. Blake,

Boston,

Francis Blake,

Weston.

John H. Blake,

Boston.

Arthur M. Comey,

Somerville

Josiah P. Cooke,

Cambridge.

James M. Crafts,

Boston.

Charles R. Cross,

Boston.

Amos E. Dolbear,

Somerville.

Thos. M. Drown,

Boston.

Charles W. Eliot, Cambridge.

Moses G. Farmer, Eliot, Me.

Thomas Gaffield, Boston.

Wolcott Gibbs, Newport, R. L

Edwin H. Hall, Cambridge.

Henryr B. Hill, Boston.

X. D. C. Hodges, New York.

Silas W. Holman, Boston.

William L. Hooper, Somerville.

Eben N. Horsford, Cambridge.

Henry M. Howe, Boston.

T. Sterry Hunt, New York.

Charles L. Jackson, Cambridge.

William W. Jacques, Newton.

Alouzo S. Kimball, Worcester.

Leonard P. Kinnicutt, Worcester.
William R. Livermore, Newjiort, R.L

Joseph Lovering,

Charles F. Mabery,

Arthur Michael,

A. A. Michelson,

Charles E. Munroe,

John U. Nef,

Lewis M. Norton,

Robert H. Richards,

Theodore W. Richards, Cambridge.

Edward S. Ritchie, Brookline.

A. Lawrence Rotch, Boston.

Charles R. Sanger, Cambridge.

Stephen P. Sharpies, Cambridge.

Cambridge.

Cleveland.

'NA'orcester.

Worcester.

Newport, R.I.

Worcester.

Newton.

Boston.

Francis H. Storer,
Eliliu Thomson,
John Trowl)ridge,
Harold Whiting,
Charles H. Wing,
Edward S. Wood,

Boston.

Lynn.

Camliridge.

Cambridge.

Boston.

Cambridge.

RESIDENT FELLOWS.

385

Section IV. — 15.
Technology and Engineering.

John M. Batchelder, Cambridge.
Wiutield S. Chaplin, St. Louis.
Eliot C. Clarke, Boston.

James B. Francis, Lowell.
Gaetano Lanza, Boston.

E. D. Leavitt, Cambridgeport.

William R. Lee,
Hiram F. Mills,
Cecil n. Peabody,
Alfred P. Rockwell,
Peter Schwamb,
Charles S. Storrow,
George F. Swain,
AVilliam Watson,
Morrill Wyman,

Roxbury.

Lawrence.

Boston.

Boston.

Arlington.

Boston.

Boston.

Boston. ,

Cambridge.

Class IL — Natural and Physiological Sciences. — 55.

Section I. — 9.

Geology, Mineralogy, and Physics of
the Globe.

Thomas T. Bouve,
Algernon Coolidge,
William O. Crosby,
William M. Davis,
O. W. Huntington,
Jules Marco u,
William H. Niles,
Nathaniel S. Shaler,
Warren Upham,

Boston.

Boston.

Boston.

Cambridge.

Cambridge.

Cambridge.

Cambridge.

Cambridge.

Somerville.

Section H. — 6.

Botany.

William G. Farlow,
George L. Goodale,
H. H. Hunnewell,
Charles S. Sargent,
Charles J. Sprague,
Sereno Watson,

Cambridge.

Cambridge.

Wellesley.

Brookline.

Boston.

Cambridge.

Section HL — 20.
Zoology and Physiology.
Alex. E. R. Agassiz, Cambridge.

Robert Amory,
James M. Barnard,
Henry P. Bowditch,
Wm. Brewster,

Boston.
IVIilton.
Boston.
Cambridge.

Louis Cabot,
Harold C. Ernst,
J. Walter Fewkes,
Edw. G. Gardiner,
Samuel Henshaw,
Alpheus Hyatt,
Theodore Lyman,
Edward L. Mark,
Charles S. Minot,
Edward S. Morse,
James J. Putnam,
Samuel H. Scudder,
William T. Sedgwick,
Henry Wheatland,
James C. White,

Brookline.

Boston.

Boston.

Boston.

Cambridge.

Cambridge.

Brookline.

Cambridge.

Boston.

Salem.

Boston.

Cambridge.

Boston.

Salem.

Boston.

Section IV. — 20.

Medicine and Surgery.

Samuel L. Abbot, Boston.
Henry I. Bowditch, Boston.
Edward H. Bradford, Boston.
Arthur T. Cabot, Boston.
David W. Cheever, Boston.
Benjamin E. Cotting, Roxbury.
Frank W. Draper, Boston.
Thomas Dwight, Boston.

Reginald H. Fitz, Boston.
Charles F. Folsom, Boston.
Richard M. Hodges, Boston.
Oliver W. Holmes, Boston.

VOL. XXVI. (n. S. XVIII.)

25

386

RESIDENT FELLOWS.

Frederick I. Knight, Boston.

Francis Minot, Boston.

Samuel J. Mixter, Boston.

Wm. L. Richardson, Boston.

George C. Shattuck,
Henry P. Walcott,
John C. Warren,
Henry W. Williams,

Boston.
Cambridge.
Boston.
Boston.

Class III. — Moral and Political Sciences. — 53.

Section I. — 10.

Philosophy and Jurisprudence.

James B. Ames,
Phillips Brooks,
Charles C. Everett,
Horace Gray,
John C. Gray,
Nathaniel Holmes,
John Lowell,
Henry W. Paine,
Josiah Royce,
James B. Thayer,

Cambridge.

Boston.

Cambridge.

Boston.

Boston.

Cambridge.

Newton.

Cambridge.

Cambridge.

Cambridge.

Section H. — 17.

Philology and Archceology.

William S. Appleton, Boston.

Lucien Carr, Cambridge

Franklin Carter,

Joseph T. Clarke,

Henry G. Denny,

Epes S. Dixwell,

William Everett,

William W. Goodwin, Cambridge.

Henry W. Haynes, Boston.

Bennett H. Nash, Boston.

Frederick W. Putnam, Cambridge.

F. B. Stephenson, Boston.

Joseph H. Thayer,

Crawford II. Toy,

John W. Wliite,

Justin Winsor,

Edward J. Young,

Williamstown .
Boston.
Boston.
Cambridge.
Quiucy.

Cambridge.
Cambridge.
Cambridge.
Cambridge.
Waltham.

Section III. — 17.
Political Economy and History.

Charles F. Adams,
Edward Atkinson,
John Cummings,
Charles F. Dunbar,
Samuel Eliot,
A. C. Goodell, Jr.,
Henry C. Lodge,
Augustus Lowell,
Edward J. Lowell,
Francis Parkman,
Andrew P. Peabody,
John C. Ropes,
Denman W. Ross,
F. W. Taussig,
Henry W. Torrey,
Francis A. Walker,
Robert C. Winthrop,

Quincy.

Boston.

Woburn.

Cambridge.

Boston.

Salem.

Boston.

Boston.

Boston.

Boston.

Cambridge.

Boston.

Cambridge.

Cambridge.

Cambridge.

Boston.

Boston.

Section IV. — 9.
Literature and the Fine Arts.

George S. Boutwell,
Martin Brimmer,
J. Elliot Cabot,
Francis J. Child,
Charles G. Loring,
Charles Eliot Norton,
Horace E. Scudder,
Barrett Wendell,
John G. Whittier,

Groton.

Boston.

Brookline.

Cambridge.

Boston.

Cambridge.

Cambridge.

Boston.

Amesbury.

ASSOCIATE FELLOWS.

387

ASSOCIATE FELLOWS. — 89.

(Number limited to one hundred.)

Class L — Mathematical and Physical Sciences. — 33.

Section L — 4.
Mathematics.

Simon Newcomb, Washington.
H. A. Newton, New Haven.
James E. Oliver, Ithaca, N.Y.
J. N. Stockwell, Cleveland, Ohio.

Section IT. — 12.

Practical Astronomy and Geodesy.

W.H.C.Bartlett,
S. W. Burnham,
Geo. Davidson,
Wm. H. Emory,
Asaph Hall,
George AV. Hill,
E. S. Holden,
Sam. P. Langley,
T. C. Mendenhall,
William A. Rogers
George M. Searle,
Chas. A. Young,

Yonkers, N.Y.
San Jose, Cal.
San Francisco.
Washington.
Washington.
Washington.
San Jose, Cal.
Washington.
Washington.
, Waterville, Me.
NevF York.
Princeton, N.J.

Section IH. — 11.
Physics and Chemistry.

Carl Barus,
J. WillardGibbs,
Frank A. Gooch,
S. W. Johnson,
M. C. Lea,
J. W. Mallet,
A. M. Mayer,
Ira Remsen,
Ogden N. Rood,
H. A. Rowland,
L.M.Rutherfurd,

Washington.
New Haven.
New Haven.
New Haven.
Philadelphia.
Ch arlottesvi lie , Va.
Hoboken, N. J.
Baltimore.
New York.
Baltimore.
New York.

Section IV. — 6.

Technology and Engineering.

Henry L. Abbot,
Geo. W. Cullum,
Geo. S. Morison,
John Newton,
William Sellers,
W. P. Trowbridge,

New York,
New York.
New York.
New York.
Philadelphia.
New Haven.

Class IL — Natural and Physiological Sciences. — 28.

Section I. — 14.

Geology., Mineralogy , and Physics of
the Globe.

Cleveland Abbe,
George J. Brush,
James D. Dana,
Sir J.W. Dawson,
F. A. Genth,

Washington.
New Haven.
New Haven.
Montreal.
Philadelphia.

James Hall,
F. S. Holmes,
Clarence King,
Joseph Le Conte,
J. Peter Lesley,
J. S. Newberry,
J. W. Powell,
R. Pumpelly,
Geo. C. Swallow,

Albany, N.Y.
Charleston, S.C
New York.
Berkeley, Cal.
Philadelphia.
New York.
Washington.
Newport, R.I.
Columbia, Mo.

388

ASSOCIATE FELLOWS,

Section II. — 2.

Botany.

A. W. Chapman, Apalachicola, Fla.
D. C. Eatou, New Haven.

Section III. — 7.

Zoology and Physiology.
Joel A. Allen, New York.

G. B. Goode,
O. C. Marsh,
H. N. Martin,

AVashington.
New Haven.
Baltimore.

S. Weir Mitchell,
A. S. Packard,
A. E. Verrill,

Philadelphia.
Providence.
New Haven.

Section IV. — 5.

Medicine and Surgery.

John S. Billings, Washington.
Jacob M. Da Costa, Philadelphia.
W. A. Hammond, New York.
Alfred Stille, Philadelphia.

H. C. Wood, Philadelphia.

Class III. — Moral and Political Sciences. — 28.

Section I. — 8.

Philosophy and

T. M. Cooley,

D. R. Goodwin,
A. G. Haygood,
James McCosh,
Charles S. Peirce,
Noah Porter,

E. G. Robinson,
Jeremiah Smith,

Jurisprudence.
Ann Arbor, Mich.
Philadelphia.
Oxford, Ga.
Princeton, N.J.
New York.
New Haven.
Providence.
Dover, N.H.

Section II. — 7.

Philology and

A. N. Arnold,
Timottiy Dwight,
I). C. Gilman,
A. C. Kcndrick,
E. E. Salisbury,
A. D. White,
W. D. Whitney,

Archceology.

Pawtuxet, R.I.
New Haven.
Baltimore.
Rochester, N.Y.
New Haven.
Ithaca, N.Y.
New Haven.

Section III. — 7.

Political Economy and History.

Washington.
New Haven.
Cincinnati.

Henry Adams,
Geo. P. Fisher,
M. F. Force,
Henry C. Lea,
Edward J. Phelps,
W. G. Sumner,
J. H. Trumbull,

Philadelphia.
Burlington, Vt.
New Haven.
Hartford.

Section IV. — 6.

Literature and the Fine Arts.
James B. Angell, Ann Arbor, Mich.
L. P. di Cesnola, New York.
F. E. Church, New York.

R. S. Greenough, Florence.
William W. Story, Rome.
Wm. R. Ware, New York.

FOREIGN HONORARY MEMBERS.

389

FOREIGN HONORARY MEMBERS. — 69.

(Elected as vacancies occur.)

Class I. — Mathematical and Physical Sciences. — 24.

Section I. — 6.

Mathematics.

John C. Adams,
Sir George B. Airy,
Francesco Brioschi,
Arthur Cayley,
Charles Hermite,
J. J. Sylvester,

Cambridge.

London.

Milan.

Cambridge.

Paris.

Oxford.

Section II. — 4.

Practical Astronomy and Geodesy.

Arthur Auwers,
J. H. W. Dbllen,
H. A. E. A. Faye,
Otto Struve,

Berlin.
Pulkowa.
Paris.
Pulkowa.

Section III. — 11.
Physics and Chemistry.

Adolf Baeyer,

JVIarcellin Berthelot,

R. Bunsen,

H. L. F. Helmholtz,

A. W. Hofmann,

Mendeleeff,

Marignac,

Lord Rayleigh,

Sir H. E. Roscoe,

Sir G. G. Stokes,

Julius Thomseu,

Munich.
Paris.

Heidelberg.
Berlin.
Berlin.
St. Petersburg.

Geneva.

Witham.

London.

Cambridge.

Copenhagen.

Section IV. — 3.

Technology and Engineering.

Marquis de Caligny, Versailles.
F. M. de Lesseps, Paris.
Sir Wm. Thomson, Glasgow.

Class II. — Natural and Physiological Sciences. — 25.

Section I. — 6.

Geology, Mineralogy, and Physics of
the Globe.

H. Ernst Beyrich, Berlin.
Alfred Des Cloizeaux, Paris.
A. E. Nordenskiold, Stockholm.
C. F. Rammelsberg, Berlin.
Sir A. C. Ramsay, Beaumaris.
Heinrich Wild, St. Petersburg.

Section II. — 5.

Botany.

J. G. Agardh, Lund.

Alphonsede Candolle, Geneva.
Sir Joseph D. Hooker, London.
Julius Sachs, Wiirzburg.

Marquis de Saporta, Aix.

390

FOREIGN HONORARY MEMBERS.

Section III. — 10.

Zoology and Physiology.

P. J. Van Beneden, Louvain.
Du Bois-Reymond, Berlin.
Thomas H. Huxley, London.
Albrecht KoUiker, Wlirzburg.
Lacaze-Duthiers, Paris.
Rudolph Leuckart, Leipsic.
C. F. W. Ludwig, Leipsic.
Sir Richard Owen, London.

Louis Pasteur, Paris.

J. J. S. Steenstrup, Copenhagen.

Section IV. — 4.

Medicine and Surgery.

Sir Wm. Bowman, London.
C. E. Brown- Sequard, Paris.
Sir James Paget, London.

Rudolph Virchow, Berlin.

Class III. — Moral and Political Sciences. — 20.

Section I. — 3.
Philosophy and Jurisprudence.

James Martineau, London.

Henry Sidgwick, Cambi'idge.

Sir James F. Stephen, London.

Section II. — 7.

Philology and Archceology.
John Evans, Hemel Hempstead.
Pascual de Gayangos, Madrid.
Benjamin Jowett, Oxford.
J. W. A. Kirchhoff, Berlin.
Cx. C. C. Maspero, Paris ?
Max Miiller, Oxford.

Sir H. C. Rawlinson, London.

Section III. — 6.

Political Economy and History.

Due de Broglie, Paris.

Ernst Curtius, Berlin.

W. Ewart Gladstone, Hawardea.
Charles Merivale, Ely.

Theodor Mommsen, Berlin.
Jules Simon, Paris.

Section IV. — 4.

Literature and the Fine Arts.
Jean Leon Gerome, Paris.
John Ruskin, Coniston.

Leslie Stephen, London.

Lord Tennyson,

Isle of Wight.

INDEX.

A.

Acalypha dissitiflora, 148.
flavescens, 149.
(Linostachys) longipes, 149.
multispicata, 148.
Acetoacetic ether on quinones, the

action of, 29.5.
Agave (Manfreda) brunnea, 156.

(Littasa) Hartmani, 156.
American Botany, contributions to,

by Sereno Watson, 124.
Amidoxyoxindol, properties of the

chloride of, 96.
Anilidotrinitrophenylmalonic ester,
88.
nitrite of, 77.
properties, 78.
Anilidotrinitrophenyltartronic ester,
82.
red modification, 83.
yellow modification, 84.
Anilidotrinitrotartrouic ester, salts

of, 86.
Anilidotrinitrotoluol, sodium salt

of, 79.
Arabis Macounii. 124.
Archaetogeron linearifolius, 139.
Arethusa grandiflora, 1.34.
Argenione Mexicana, 162.
Aristolochia nana, 145.
Arracacia Mariana, 136.

multifida, 136.
Arsenic, quantitative determination
of same in wall papers and
fabrics, 24.
Ascension, notes upon a collection
of plants from the island of,
161.
Aspleniiim Ascensionis, 163.
Aster carnerosanus, 139.
Engelmanni, Gray, 176.

Atomic weight of copper, a revision

of the, 240.

Ayena Berlandieri, 133.

Jaliscana, 133.

B.

Bacterium of Kern's milk-ferment,

Dispora Caucasica, 104.
Bahia Schaffneri, 142.
Bassovia Mexicana, 171.
Begonia (Weilbachia) Pringlei, 136.
Beuzofurfuran derivative, 298, 299.
derivatives from quinone and

acetoacetic ether, 306.
derivatives, synthesis of, 295.
Benzo-jr>difurfuran a-dimethyl /3-di-

carboxylic ether, 307.
Bidens dahlioides, 142.

jiilosa, 162.
Biographical notices, list of, -337.
George Bancroft, 355.
Henry Jacob Bigelow, 339.
Charles Otis Boutelle, 351.
Julius Erasmus Hilgard, 370.
Alfred Hosmer, 351.
Charles John Maximowicz, 374.
Karl Wilhelmvon Naegeli,376.
Christian Heinrich Friedrich

Peters, 373.
Eduard Schdnfeld, 381.
Bletia Palmeri, 153.
Boerhaavia octandra, 145.
Botany, American, contributions

to, by Sereno Watson, 124.
Bromdinitrophenylmalonic ester, ni-
trite of, 93.
Bromtrinitrophenylmalonic ester, on
the products obtained by the
action of nitric acid upon,
67.

392

INDEX.

Bromtrinitrophenylmalonic ester,
preparation of the nitrite of,
72.
properties of the nitrite of, 74.

Bromtrinitrophenyltartronic ester,
80.

Buddleia Chapalana, 169.

Bunchosia Pringlei, 133.

Burrillia, gen. uov., 18.

Cacalia (Conophora) poculifera, 143.
Carapylopus introflexus, Brid., 103.
Castilicia macrostigma, 173.
Chamsedorea Pringlei, 157.
Chimelia Pringlei, 137.
Chloracetoacetic ether, 299.
Choreocolax Polysiphoniag, on the
structure and development
of, 46.
Chromite in Kiowa County pallas-

ite, 6.
Citharexylum Berlandieri, 174.
Cladothrix cryptantha, 125.
Cnicus linearifolius, 143.

velatus, 143.
Cologania Jaiiscana, 136.
Commelina nudiflora, L. 162.
Communications, —

Carl Barns, 313.

Oliver Whipple Huntington, 1.

M. Ikuta, 295.

C. Loring Jackson and W. B.
Bentley, 67, 98.

Charles L. Mix, 102.

B. (). Peirce, 20. 218.

Herbert Maule Richards, 46.

Theodore William Richards,
240.

B. L. Robinson, 164.

Charle.s R. Sanger, 24.

William Albert Setchell, 13, 177.

Henry Taber, 64.

John Trowbridge, 115.

Sereno Watson, 124.
Copper, a revision of the atomic

weight of, 240.
Corelia Pringlei, 109.
Cornuella, gen. nov., 19.
Croton (Eucroton) calve.scens, 147.

(Eutropia) elajagnoides, 147.
Crusea megalocarpa, 137.
Cupric oxide, occluded gas in, 284.

the analysis of, 276.

Cupric sulphate, action of heat upon,
265.
the analysis of, 244.
Cyperus umbellatus, Benth., 162.

D.

Dahlia dissecta, 141.

pubescens, 142.
Dasylirion interme, 157.
Desmodium amans. 135.

Jaliscanum, 164.

subspicatum, 135.
Diacetylsuccinic ether, 298.
Dichlorbenzo-/?-difurfuran a-dirae-
thyl /3-dicarboxylic ether, 302.

acid, 303.
Dichlorhydroquinonediacetoacetic

ether, 30 1, 302.
Dichlorquinonediacetoacetic ether,
300, 305.

addition products of, 303.
Dicranella, 103.
Disodium salt, 86.
Doassansia, Cornu, preliminary notes
on the species of, 13.

Alismatis, Cornu, 16.

Comari, De Toni, 18.

deformans, sp. nov., 17.

Epilobii, Farlow, 15.

HottoniiB, De Toni, 15

Lythropsidis, Lagerh., 18.

MartianofBana, Schroeter, 17.

obscura, 10.

occulta, var. Farlowi, 17.

opaca, sp. nov., 15.

punctiformis, Winter, 18.

Sagittaripe, Fisch., 15.
Doassausiopsis, 16.

E.

Echeandia nodosa, 156.
Ehretia IMexicana. 144.
Electric flow in flat circular plates,

on some simple cases of, 218.
Electrical oscillations on iron wires,

dampening of, 115.
Equation, matrical, on the, 04.
Eriocaulon Jaliscanum, 1.57.
Eriogonum destricola, 125.

minutiflorum, 125.
Eryngiuni Mexicanura, 136.
Erysium arenicola, 124.

INDEX.

393

Erythroniura, Linn., 126.

albidum, Nutt., 128.

Ainericiumin, Ker., 127.

Bolaiideri, 129.

citrinum, l-JO.

gigaiiteiim, 129.

giandilloruin, 128.

llartwegi, Watson, 129.

Hendersoni, Watson, 130.

Howellii, AVatson, 130.

mesochoreum, 128.

montanum, 180.

paiviflornm, 129.

propullans, 128.

purpurascens, Watson, 130.

revoliitum, Smith, 129.
Ester, anilidotriniphenylmalonic,
88.

anilidotrinitrophenyltartronic,
82.

anilidotrinitrotartronic, 70.

bromtrinitrophenylraalonic, 67.

bromtrinitrophenyltartronic, 80.

bromtrinitrotartronic, 67.
Eudoassansia, 14.
Eupatoiium Cliapalense, 138.

espinosarum, Gray, 165.

Madrense, 137.
Euphorbia digitata, 146.

inisella, 116.

origanoides, L., 162.

subpeltata, 146.

F.

Fabrics, quantitative determination

of arsenic in, 24.
Fellows, Associate, deceased, —

George Bancroft, 334.

Jolni Charles Fremont, 334.

Christian Henry Friedrich Pe-
ters, 334.
Fellows, Associate, elected, —

Henry Newell Martin, 327.
Fellows, Associate, list of, 387.
Fellows, Resident, deceased, —

Henry Jacob Bigelow, 334.

Charles Otis Boutelle, 334.

William Pr^scott Dexter, 334.
Fellows, Resident, elected, —

Arthur INIessinger Comey, 332.

Charles Edward Munroe, 332.

John Ulric Nef , 332.

Theodore William Richards,
332.

Charles Robert Sanger, 332.

Fellows, Resident, list of, 384.
Ferment, milk, Kern's, 102.
Ficus fasciculata, Watson, 152.

Guadalajarana, 151.

Jaliscana, 150.

Pringlei, 150.

radulina, 151.
Foreign Honorary Members, elected.

Sir William Bowman, Bart., 330.

Sir Henry Enfie<d Roscoe, 327.
Foreign Honorary Members, list of,
389.

G.

Gamopetalfe, descriptions of new
plants in Mexico, chiefly, 164.
Gerardia punctata, 172.
Gonolobus parviflorus, 169.
Govenia elliptica, 1.53.
Gymnogramme Ascensionis, 163.
Gymnolomia decumbens, 105.

H.

Habenaria filifera, 1.54.
Hechtia pedicellata, 155.
Heliotropium Pringlei, 170.
Herpestis auriculata, 172.

Integrals, line and sui-face, on some
theorems which connect, 20.
Ipomoea Leonensis, 170.
Isocarbopyrotritaric ether, 298.
Isometrics, the, 320.

J.

Juglans Mexicana, 1.52.

Justicia Pringlei,

173.

K.

Kephir-like yeast, found in the

United States, 102.
Kern's milk ferment, 102.
Kiowa County pallasites, Prehistoric

and, 1.
compared with Krasnojarsk, 8.
Krasnojarsk pallasite, compared

with Prehistoric and Kiowa

County, 8.

394

INDEX.

L.

Lactuca scariola, L., 162.
Laurentia ovatifolia, 16G.
Lejeunia pterota, Taylor, 163.
Lobelia novella, 1G7.
Lycopodium cernuum, L., 162.

M.

Manihot Pringlei, 148.
Matrical equation, on the, 64.
Melampodium (Unxia) bibractea-
tum, 140.

glabvum, 139.
Microstylis (Dienia) tenuis, 152.
Milk-ferment, Kern's, 102.
Mimulus Congdonii. 175.

(Eumimulus) filicaulis, 125.

gracilipes, 176.
Molecular pressure, note on the va-
riation of, 313.
Monopotassium salt, 86.
Myriocarpa Mexicana, 152.

N.

Nasturtium bracteatum, 131.
Nemacladus oppositifolius, 168.
Neopringlea, 134.

integrifolia, 135.
Nephrodium (?) viscidum, 163.

o.

Oligonema, 138.

heterophylla, 138.
OMvine crystals in pallasites, 2.
Omphalodes acuminata, 170.
Oscillations, electrical, dampening

of, on iron wires, 115.
Otopappus acuminatus, 140.

alternifolius, 165.
Oxalis corniculata, L., 162.
Oxide, cupric, the analysis of, 276.

occluded gas in, 284.
Oxybenzofurfuran a-dimethyl /3-car-
boxylic ether, 309.

Pallasites, the Prehistoric and Kiowa
County, 1.
table of, 9.

Parmelia saxatilis, Ach., 163.

Peperomia Jaliscana, 145.

Perezia collina, 144.

Phyllanthus Pringlei, 147.

Pilea glabra, 152.

Pimpinella Mexicana, 164.

Piper (Enckea) Jaliseanum, 145.

Pogonia (Triphora) INIexicaua, 154.

Polygala sublata, 132.

Prehistoric pallasite, having the old-
est authentic record, 1.
compared with those of Krasno-
jarsk and Kiowa County, 8.

Pseudodoassansia. 16.

Psilactis tenuis, 139.

Pteris flabellata, Thunb., 162.
incisa, Thunb., 162.

Q.

Quinone ;>difurfuran a-dimethyl /3-
dicarboxylic ether dihydro-
chloride, 304.

Quinones, tiie action of acetoacetic
ether on, 295.

R.

Ranunculus vagans, 131.
Rhacopilum gracile, Mitt., 163.
Rubus nanus. 162.
Ruellia Bourgsei, Hemsley, 173.

S.

Saccharomvces cerevisiae, INIeyen,
111,'113.

galacticola, Pirotta, 110, 113.

kefyr, Beyerinck, 107, 110.

lactis, 110.

Tyrocola, 110.
Saccorhiza dermatodea, concerning

the life-historj' of, 177.
Salraea Palmeri, 141.
Sargentia Pringlei, 134.
Schultesia Mexicana, 144.
Scutellaria hispidula, 174.
Sebastiania Pringlei, 149.
Senebiera didyma, Pers., 162.
Senecio Guadalajarensis, 166.

Jaliscana, 143.
Sida Alamosana, 133.
Silene Macounii, 124.
Sisymbrium multiracemosum, 132.

INDEX.

395

Sisyriuchium platyphyllum, 155.
Sodium pheiiolate, 299.
Spilanthes Botterii, lil.
Spiranthes Jaliscana, 153.

Pringlei, 153.
Styrax Jaliscana, 144.
Sulphur, the atomic weight of, 268.
Symplocos Pringlei, 168.

T.

Talinum Coahuilense, 132.
Theloschistes crysophthalma, 163.
Thermal capacity, volume relations

of, 323.
Thymol, compressibility of, 321.
expansion of, 315.
specific heat of solid and liquid,

317.
Tillandsia cylindrica, 155.

Pringlei, 155.
Tithonia macrophylla, 140.
Tradescantia Pringlei, 157.
Tribrommononitrobenzol, note on,

98.
Tribromtrinitrobenzol, preparation

of, 71.
Trichlorhydroquinoneacetoacetic

ether, 296, 298.
Trichlor />oxybenzofurfuran a-me-

thyl /3-carboxylic ether, 297.
acid, 300.

Trichlorquinoneacetoacetic ether,

295.
Trinitroi:)henylenedimalonic ester,

nitrite of, 90.

Viguiera leptocaulis, 140.

W.

Wall papers, quantitative determi-
nation of arsenic in, 24.

Widmanstattian figures in pallas-
ites, 6.

Withania melanocystis, 171.

X.

Xanthoxylum Pringlei, 134.
Xylosma Pringlei, 164.

Y.

Yeast, Kephir-like, found in the
United States, 102.

Zea, a wild species from Mexico, 158.

MBL WHOl LIBRARY 0^

UH lAfiT Y

^ f/^ s

.)ti\

i^wuakjUv